Infection of Epithelial Cells by Pathogenic Neisseriae Reduces the Levels of Multiple Lysosomal Constituents

Abstract

Members of our group reported recently that neisseria infection of human epithelial cells results in accelerated degradation of the major lysosomal integral membrane protein LAMP1 and that this is due to hydrolysis of this glycoprotein at its immunoglobulin A1 (IgA1)-like hinge by the neisseria type 2 IgA1 protease (L. Lin et al., Mol. Microbiol. 24:1083–1094, 1997). We also reported that the IgA1 protease plays a major role in the ability of the pathogenic neisseriae to survive within epithelial cells and hypothesized that this is due to alteration of lysosomes as a result of protease-mediated LAMP1 degradation. In this study, we tested the hypothesis that neisseria infection leads to multiple changes in lysosomes. Here, we report that neisseria infection also reduces the levels of three other lysosomal markers: LAMP2, lysosomal acid phosphatase (LAP), and CD63. In contrast, neither the epidermal growth factor receptor level nor the β-tubulin level is affected. A detailed examination of LAMP2 indicated that the reduced LAMP2 levels are not the result of an altered biosynthetic rate or of cleavage by the IgA1 protease. Nevertheless, the protease plays a role in reducing LAMP2 and LAP activity levels, as these are partially restored in cells infected with an iga mutant. We conclude that neisseria infection results in multiple changes to the lysosomes of infected epithelial cells and that these changes are likely an indirect result of IgA1 protease-mediated cleavage of LAMP1.

The pathogenic neisseriae Neisseria meningitidis (meningococcus [MC]) and Neisseria gonorrhoeae (gonococcus [GC]) are closely related gram-negative bacteria that share many genetic and biological traits. At the mucosa, they initially form a loose association with the apical surfaces of epithelial cells, an interaction which subsequently develops into tight contact between the bacterial and host cell plasma membranes. The bacteria subsequently invade the cell, transcytose, exit the cell, and enter the subepithelial matrix, where they initiate the symptoms of disease. Studies using infected organ cultures (15, 27, 28) and a model epithelium (16, 24) indicate that transcellular trafficking by the pathogenic neisseriae is a lengthy process and that bacterial transcytosis does not destroy the barrier functions of the monolayer.

The immediate environment of intracellular neisseriae is unclear at present. Some studies indicate the presence of a phagosomal membrane surrounding intracellular MC (24, 28) and GC (31). Others suggest that intracellular neisseriae have access to the host cell cytoplasm (25, 32). Recently, the neisserial type 2 immunoglobulin A1 (IgA1) protease was shown to play a role in intracellular survival of MC and GC (14).

All pathogenic neisseriae constitutively secrete one of two closely related types of IgA1 proteases which cleave at different sites within the hinge of the human IgA1 (hIgA1) subclass of immunoglobulins (19, 21, 23). Type 1 protease cleaves at a specific proline-serine (P-S) bond, while type 2 protease cleaves at a proline-threonine (P-T) bond in the hIgA1 hinge. The specificity of this enzyme for hIgA1 and the presence on infected mucosa of hIgA1 fragments of the sizes predicted for IgA1 protease products (18) strongly suggest a role for this enzyme in bacterial colonization. Recently, a second biological function was identified for the neisseria type 2 IgA1 protease: that of altering the levels of a major lysosomal protein, thereby promoting intracellular survival of the bacteria (14).

Lysosomes are terminal degradative compartments in the endocytic route. They perform key functions within a eukaryotic cell, among them the digestion of foreign compounds and macromolecules that have been endocytosed. Sequestered in the lysosome lumen are numerous hydrolases that degrade a wide range of biological materials, including proteins, carbohydrates, lipids, and nucleic acids. These enzymes have pH optima that reflect the acidic pH of the lysosome. Associated with the lysosomal membrane are enzymes that participate in the acidification of the lumen, selective transport of metabolites from the lumen to the cytoplasm, and fusion of the lysosome with other compartments and organelles (11, 13). Located in the lysosomal membrane is a unique class of glycoproteins known as lysosome-associated membrane proteins (LAMPs), of which LAMP1 and LAMP2 are members. LAMP1 and LAMP2, both with M_rs of approximately 120,000, have 37% amino acid identity and consist of two heavily glycosylated domains roughly equal in size, a single transmembrane domain and a short cytoplasmic tail (2–5, 7, 8, 13). The luminal domains of LAMP1 are separated by a proline-rich hinge with striking similarities to the hIgA1 hinge (6, 29). The function of LAMPs is unknown, although they have been hypothesized to play a role in protecting the lysosome from its associated hydrolases (7, 13).

Human epithelial cells infected by the pathogenic neisseriae were recently observed to contain significantly reduced levels of LAMP1. The decrease in LAMP1 levels is due to accelerated degradation, which in turn is brought about by hydrolysis of this glycoprotein at its hIgA1-like hinge by the neisserial type 2 IgA1 protease (14). In vitro cleavage of LAMP1 by the protease has also been demonstrated (10, 14). In addition, the IgA1 protease was shown to play an important role in the ability of neisseriae to survive within epithelial cells, as a mutant in which the iga gene was deleted was unable to replicate within epithelial cells, unlike its isogenic wild-type (WT) parent. Based on these results, it was proposed that intracellular survival of the pathogenic neisseriae is due to an alteration of the lysosomes via IgA1 protease-mediated accelerated LAMP1 turnover (14).

In this study, we tested the hypothesis that neisseria infection leads to multiple changes in lysosomes. We present evidence that the levels of three lysosomal constituents other than LAMP1, LAMP2, lysosomal acid phosphatase (LAP), and CD63, are decreased significantly in neisseria-infected cells. In contrast, the levels of two nonlysosomal components, epidermal growth factor receptor (EGFR) and β-tubulin, are unaffected. We show that the reduction in LAMP2 levels is not due to perturbations in the biosynthetic rate of this protein or to direct hydrolysis by the IgA1 protease. Finally, we present evidence that the IgA1 protease plays an indirect role in reducing both LAMP2 and LAP levels. We conclude that neisseriae cause multiple changes to occur in the lysosomes of infected epithelial cells. Our data strongly suggest that the alterations in LAMP2, LAP, and CD63 levels are due to IgA1 protease-mediated cleavage of LAMP1.

MATERIALS AND METHODS

Strains and culture and infection methods.

All neisseria strains used in this study produce type 2 IgA1 protease (14). N. meningitidis 8013.6 (20) belongs to serogroup C, is piliated, and expresses an uncharacterized Opa (15a). N. gonorrhoeae GCM740 is piliated and does not express Opa, as demonstrated by immunoblots with the pan-Opa monoclonal antibody (MAb) 4B12 (a generous gift of M. Blake). GCM740Δ4 is an isogenic derivative of GCM740 from which 93% of the iga gene has been deleted. GCM740Δ4 has no detectable IgA1 protease activity (26). MS11A307 [Δ(pilE1 pilE2)] is a nonpiliated, Opa⁻, very low adherence mutant derived from MS11A (16). Bacteria were propagated on supplemented GCB agar. A431 human epidermoid carcinoma epithelial cells (ATCC CRL 1555) were maintained in Dulbecco’s minimum essential medium (Gibco) plus 5% fetal bovine serum (Gibco), and infections were carried out in the same medium plus 50 μg of human transferrin per ml. Unless otherwise stated, A431 cells were infected with MC at a multiplicity of infection (MOI) of 1 and with GC at an MOI of 5. Unless otherwise stated, cells were infected 16 h prior to the assays.

Reagents.

Polyclonal antibodies against LAMP1 and LAMP2 were generated as described previously (2). MAbs H4B4 (anti-LAMP2), E7 (anti-β-tubulin), and H5C6 (anti-CD63) were obtained from the Developmental Studies Hybridoma Bank (maintained by the Department of Pharmacology and Molecular Sciences, Johns Hopkins University School of Medicine, Baltimore, Md., and the Department of Biological Sciences, University of Iowa, Iowa City, under contract N01-HD-6-2915 from the National Institute of Child Health and Human Development). The anti-EGFR MAb was purchased from Santa Cruz Biotech. Tetramethyl rhodamine isothiocyanate-conjugated goat anti-rabbit antibody was purchased from Pierce, and BODIPY-conjugated goat anti-mouse antibody was purchased from Molecular Probes.

Quantitation of cellular components.

LAMP2 and β-tubulin levels were determined by immunoblotting (Western blotting) as described elsewhere (14). Primary anti-β-tubulin MAb was detected with goat anti-mouse IgG–alkaline phosphatase (Boehringer Mannheim). Polyclonal anti-LAMP2 antibody was detected with either goat anti-rabbit IgG–horseradish peroxidase (Super Signal chemiluminescence kit; Pierce) or anti-rabbit IgG–alkaline phosphatase (Boehringer Mannheim) developed by a colorimetric reaction with nitroblue tetrazolium–5-bromo-4-chloro-3-indolylphosphate toluidinium (Boehringer Mannheim). Signals from the blots were scanned and quantitated with the NIH-image V.1.60 program. The LAMP2 values for N. meningitidis 8013.6-, N. gonorrhoeae GCM740-, and N. gonorrhoeae GCM740Δ4-infected cultures were normalized to their respective internal β-tubulin values and expressed relative to the normalized value for uninfected cultures. To quantitate EGFR steady-state levels, cultures were infected for 15 h, washed extensively, and labelled for 6 h with 200 μCi of [³⁵S]methionine-[³⁵S]cysteine (NEN). EGFR was immunoprecipitated from the lysates with the anti-EGFR MAb and resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The EGFR signals were scanned and quantitated with the NIH-image V.1.60 program, and the resultant value for each sample was normalized to the number of viable cells in parallel cultures as determined by trypan blue exclusion assays (see next section). To quantitate LAP, cultures were infected for 16 h, washed three times with phosphate-buffered saline (PBS), and lysed with PBS–Triton X-100 (1%), and the lysates were cleared by centrifugation at 16,000 × g for 15 min. The supernatants were assayed for LAP activity with the Acid Phosphatase, Alkaline Phosphatase, and Prostatic Acid Kit (Sigma Diagnostics) as specified by the manufacturer.

Double immunofluorescence microscopic detection of LAMP2 and bacteria in infected A431 cells.

Infected cells were processed for microscopy as described previously (14). Coverslips containing infected cells were fixed for 25 min in Zamboni’s fixative (picric acid paraformaldehyde) (33) at room temperature, then washed three times with PBS, and permeabilized and blocked in the immunofluorescence blocking buffer IFB (PBS with 3% goat serum, 0.01% azide, and 0.01% saponin). Primary antibodies were added at the appropriate dilutions, and the cells were incubated for 3 h at room temperature. Cells were next washed three times with PBS, treated with IFB for 20 min, and incubated with secondary antibodies in the dark for 30 min. Coverslips were finally washed five times with PBS and mounted in mounting buffer (5 mM Tris [pH 8.0], 20-mg ml⁻¹ n-propyl gallate, 90% glycerol). Cells were examined with a Leica confocal laser scanning microscope equipped with a Leitz Fluovert-FU inverted microscope and an argon-krypton laser with a transmitted light detector DMIR.

Determination of LAMP2 biosynthesis rates.

LAMP2 biosynthesis rates were determined essentially as described for LAMP1 (14). Infected and uninfected cells were labelled with 200 μCi of [³⁵S]methionine-[³⁵S]cysteine (New England Nuclear). At various times, the cells were washed and lysed with ice-cold NET buffer (0.01 M Tris [pH 7.4], 0.15 M NaCl, 0.005 M EDTA) containing 1% Triton X-100 and 1 mM phenylmethylsulfonyl fluoride. Lysates were centrifuged at 27,200 × g for 5 min, LAMP2 was immunoprecipitated from the lysates with the anti-LAMP2 MAb and protein A-Sepharose, and the precipitate was counted and resolved by SDS-PAGE. LAMP2 precipitable counts for each sample were expressed relative to the number of viable cells as determined by trypan blue exclusion assays with parallel unlabelled cultures.

IgA1 protease cleavage of LAMP1 and LAMP2.

Cleavage of the LAMP proteins by type 2 IgA1 protease was performed as described previously for LAMP1 (14). Three nanograms of purified LAMP1 or LAMP2 (2) was incubated with 300 ng of purified neisseria type 2 IgA1 protease (Boehringer Mannheim) in buffer (pH 7.5, 6.5, or 5.0), or with buffer alone, at 37°C for 4 h. The reaction products were resolved by SDS-PAGE and detected by immunoblotting using the anti-LAMP1 or anti-LAMP2 polyclonal antibody.

RESULTS

LAMP2 steady-state levels in neisseria-infected cultures.

If neisseria infection alters lysosomes, it can be predicted that lysosomal constituents other than LAMP1 are affected. Therefore, the LAMP2 levels in neisseria-infected cultures were determined. A431 human epithelial cells were infected with N. meningitidis 8013.6 or N. gonorrhoeae GCM740, two of the strains which were previously shown to reduce LAMP1 levels in infected cells (14). Equal amounts of total cell proteins were immunoblotted with a polyclonal antibody against LAMP2 and a MAb against β-tubulin or immunoprecipitated with a MAb against EGFR, a plasma membrane receptor which shares biosynthetic pathways with LAMPs (12). Previous studies have demonstrated that β-tubulin levels are unaltered in neisseria-infected cells (14). The results indicated that the steady-state levels of EGFR (Fig. 1C) and β-tubulin (data not shown) were identical in infected and uninfected cultures. In contrast, LAMP2 levels were noticeably decreased in MC- and GC-infected cultures compared to those in uninfected controls (Fig. 1A and B). Normalization of LAMP2 signals to their corresponding internal β-tubulin controls revealed a 31% decrease in LAMP2 levels in infected cultures (relative levels in MC-infected and GC-infected cultures, 69% ± 8% and 71% ± 5%, respectively; means and standard deviations from four independent experiments). Similar values were obtained when LAMP2 signals were normalized to EGFR signals (data not shown).

FIG. 1.

LAMP2 and EGFR levels in neisseria-infected A431 human epithelial cells. (A) LAMP2 and β-tubulin signals in representative immunoblots of total cell proteins from uninfected cultures and cultures infected with piliated bacteria. U, uninfected cultures; GC, WT piliated GC (strain GCM740); GC iga, GCM740Δ4, a protease-deficient but otherwise isogenic derivative of GCM740; MC, WT piliated MC (strain 8013.6). (B) LAMP2 levels in infected A431 cultures relative to that in the uninfected control culture (means and standard deviations [vertical bar] from four independent experiments). (A′) LAMP2 and β-tubulin signals in representative immunoblots of total cell proteins from A431 cells infected with MS11A307, a nonpiliated, Opa⁻, low-adherence strain. (B′) LAMP2 levels in A431 cultures infected with piliated MS11A (GC P⁺) and nonpiliated MS11A307 (GC P⁻) relative to the uninfected control culture. (C) EGFR level in A431 cultures infected with strain GCM740 (GC) expressed relative to that in the uninfected (control) culture (U) (mean and standard deviation [vertical bar] from three independent experiments).

To determine whether the effect of infection on LAMP2 levels is due to bacterial contact with the epithelial cell, LAMP2 levels were examined in A431 cells infected with a nonpiliated mutant. MS11A307, an otherwise isogenic derivative of the piliated, adherent MS11A strain, has deletions in both of its pilE loci [Δ(pilE1 pilE2)] and is therefore nonpiliated (16). After 10 h of infection, >50% of the total population of MS11A cells and ∼2% of the total population of MS11A307 cells are cell associated (16). Results from this experiment indicated that both MS11A and MS11A307 were able to reduce LAMP2 levels in A431 cells (Fig. 1A′ and B′). Relative to those in uninfected cultures, LAMP2 signals were decreased ∼37% in MS11A-infected cultures and ∼50% in MS11A307-infected cultures (after normalization to the appropriate internal β-tubulin values). These results suggest that the ability of neisseriae to reduce LAMP2 steady-state levels is unlikely to be due to extensive contact of bacteria with the epithelial cells.

The degree of LAMP2 reduction in an infected culture is apparently moderate. A likely explanation for this result is that the pathogenic neisseriae do not uniformly infect all cells in a given culture. The degree of infection can range from 40 to 90%, depending on the age of the inoculum, the MOI, and various undefined conditions. In our experiments, the level of infection ranged from 60 to 70%. Thus, in a given infected cell, the LAMP2 steady-state level should be significantly lower that those in uninfected cells. This effect can be seen in GC-infected A431 cell cultures double-stained for bacteria and LAMP2 (Fig. 2). Cells infected with bacteria had very low levels of LAMP2. In contrast, LAMP2 signals were readily apparent in a group of uninfected cells in the same field. These results demonstrate that LAMP2 levels in infected cells are indeed greatly reduced.

FIG. 2.

Confocal laser scanning microscopy of GC-infected A431 cells double-stained for LAMP2 (green) and bacteria (red). Panels A and B show one horizontal 4-μm-thick optical section taken of the same field of cells. The arrow indicates the location of a cell demonstrated by phase-contrast microscopy (data not shown) and by its fluorescence properties to be infected.

LAMP2 biosynthetic rates in neisseria-infected cultures.

To determine whether the reduction in LAMP2 levels was the result of decreased biosynthesis, LAMP2 biosynthetic rates in infected and uninfected cultures were next compared. A431 cells were infected with GCM740, washed extensively, and labelled for various lengths of time. LAMP2 was immunoprecipitated from lysates using the anti-LAMP2 MAb, and the precipitate was counted and resolved by SDS-PAGE. LAMP2 precipitable counts for each sample were expressed relative to the number of viable cells as determined by trypan blue exclusion assays on parallel unlabelled cultures. The results indicated that the rates of incorporation of label into LAMP2 were identical in infected and uninfected cultures (Fig. 3B). A typical autoradiogram from one such experiment is shown in Fig. 3A. Thus, neisseria infection does not alter the rate of synthesis of this glycoprotein.

FIG. 3.

Biosynthesis rates of LAMP2 in uninfected A431 cell cultures (U) and cultures infected with WT GC (strain GCM740) (I). (A) A representative autoradiogram of LAMP2 immunoprecipitated from radiolabelled cultures. (B) Rates of incorporation of [³⁵S]Met-[³⁵S]Cys into LAMP2. Data are from three independent experiments.

LAMP2 levels in cultures infected with an iga mutant.

Neisseria infection of epithelial cells was shown to result in reduced steady-state levels of LAMP1 (14). This reduction was not due to an altered biosynthetic rate but to hydrolysis of the glycoprotein by the bacterial IgA1 protease (14) at its proline-rich hinge, presumably at the P-T bond (19, 29). While the LAMP2 hinge bears little resemblance to either the LAMP1 or hIgA1 hinge, it does contain three P-T bonds: hIgA1 hinge, CPVPSTPPTPSPSTPPTPSPSCC; hLAMP1 hinge, PSPTTAPPAPPSPSPSPVPKSPS; and hLAMP2 hinge, TSTVAPTIHTTVPSPTTTPTP.

An experiment was performed to determine whether the IgA1 protease also plays a role in reducing LAMP2 levels in infected cultures. A431 cells were infected with N. gonorrhoeae GCM740 or GCM740Δ4, and their LAMP2 levels were determined by immunoblotting. GCM740Δ4 is an otherwise isogenic derivative of GCM740 in which ∼93% of the iga coding sequence has been deleted. It has no detectable IgA1 protease activity. Its outer membrane protein profile has been extensively characterized and shown to be identical to that of its isogenic parent strain, GCM740 (26). LAMP2 levels in GCM740Δ4-infected cultures were clearly higher than those in GCM740-infected controls (Fig. 1B), indicating that the IgA1 protease plays a role in altering the levels of this glycoprotein. However, the LAMP2 levels in GCM740Δ4-infected cultures were somewhat lower than those in uninfected controls (relative LAMP2 levels in GCM740Δ4-infected cells, 83% ± 4%; mean ± standard deviation from four independent experiments). A similar situation was observed for LAMP1: its levels in GCM740Δ4-infected A431 cell cultures were slightly reduced (∼93% of those in uninfected cell controls [14]).

Cleavage of purified LAMP2 by purified type 2 IgA1 protease.

It was next determined whether LAMP2 is a substrate for the IgA1 protease. Purified LAMP1 or LAMP2 (2) was incubated with purified neisseria type 2 IgA1 protease at 37°C for 4 h in pH 7.0, 6.5, or 5.0 buffer at an enzyme-to-substrate ratio of 100:1. The products were separated by SDS-PAGE and visualized by immunoblotting using the anti-LAMP1 or anti-LAMP2 polyclonal antibodies. In control reactions, IgA1 protease was effective in hydrolyzing purified hLAMP1 at pH 7.5 and 6.5 and was partially active at pH 5.0 (Fig. 4A). The two LAMP1 cleavage products, comigrating at ∼60 kDa, could be seen. Decreased protease activity on the LAMP1 substrate at pH 5.0 has been observed previously (14). In pH 7.5 buffer, the enzyme also hydrolyzed >95% of human IgA1 in <1 h at an enzyme-to-substrate ratio of 1:1 (data not shown; see reference 14). In contrast, the IgA1 protease did not hydrolyze LAMP2 under these conditions (Fig. 4B). LAMP2 migrated slightly more diffusely in samples containing the protease than in samples containing buffer alone. Densitometric analysis indicated that the amounts of unit-length LAMP2 were identical in all samples. Furthermore, no LAMP2 cleavage products were detected. Thus, LAMP2 is unlikely to be a substrate for the IgA1 protease. These results indicate that the decrease in LAMP2 levels in neisseria-infected cells is unlikely to be the direct result of IgA1 protease hydrolysis.

FIG. 4.

Purified human LAMP2 is not cleaved by neisseria type 2 IgA1 protease in vitro. Three nanograms of LAMP1 (A) or LAMP2 (B) was incubated with 300 ng of purified neisseria type 2 IgA1 protease in buffer at pH 7.5, 6.5, or 5.0, or with buffer alone, at 37°C for 4 h. The reaction products were resolved by SDS-PAGE, and the cleavage products were detected by immunoblotting with the anti-LAMP1 or anti-LAMP2 polyclonal antibody. The arrow indicates the position of the LAMP1 cleavage products.

LAP activity in neisseria-infected cultures.

Neisseria-infected cultures were also assessed for the activity of LAP, a glycoprotein which is membrane associated during transport to lysosomes (30) and proteolytically processed into soluble mature enzyme upon delivery to this compartment (9). A431 cells were infected with N. meningitidis 8013.6 or N. gonorrhoeae GCM740 or GCM740Δ4, the iga mutant derivative of GCM740, and cell lysates were assayed for LAP activity. Results from three independent experiments indicated that cells infected by WT MC and GC had significantly lower LAP activity (Fig. 5) than uninfected cells. Like LAMP2 levels, LAP activity levels were significantly higher in cells infected with the iga mutant GCM740Δ4. These results demonstrate that neisseria infection affects LAP and strongly suggest that the IgA1 protease plays a role in this process.

FIG. 5.

LAP activities in cultures of A431 cells infected with MC strain 8013.6 (MC) or GC strains GCM740 (GC) and GCM740Δ4 (GC iga) expressed relative to that in control uninfected cells (U). Results represent the means and standard deviations from three independent experiments.

CD63 levels in neisseria-infected cells.

Finally, neisseria-infected cells were examined for their levels of CD63, a lysosomal membrane glycoprotein of unknown function (17). A431 cells were infected with N. meningitidis 8013.6 for 12 h at an MOI of 15, fixed, and stained for immunofluorescence microscopy. The results indicated that the CD63 levels were significantly reduced in MC-infected cells compared to those in uninfected cells (Fig. 6). In uninfected cells, CD63 signals appeared as punctate dots evenly distributed over the entire cell. An intense signal was also present in the perinuclear region, the region corresponding to the normal location of lysosomes. In infected cells, punctate signals were much less apparent and the intensity of the perinuclear signal was significantly reduced. As observed for LAMP2 (Fig. 2), uninfected cells had normal levels of CD63 (center of field in Fig. 6C and D), while infected cells in the same field had much lower CD63 levels. The decrease in CD63 signals could not be quantitated, as the anti-CD63 MAb did not recognize the protein processed for immunoblotting.

FIG. 6.

CD63 levels in uninfected A431 cells (A and B) and cells infected with N. meningitidis 8013.6 (C and D), visualized by double-label immunofluorescence microscopy. (A and C) Cells stained for CD63. (B and D) The same field stained for MC. Stained cells were examined and photographed with a Nikon Microphot FX at the same magnifications. Note that the uninfected cell within the center of this field of infected cells (panel D) had normal levels of CD63 (panel C).

DISCUSSION

We reported previously that neisseria infection of epithelial cells results in significantly reduced LAMP1 levels (14). In the present study, we tested the hypothesis that neisseria infection leads to multiple changes in lysosomes. Our data indicate that the steady-state levels of LAMP2 and CD63 and the activity of LAP are also reduced in these cells. In contrast, the biosynthetic rate of neither LAMP1 (14) nor LAMP2 (Fig. 3B) is affected by neisseria infection. Moreover, the steady-state levels of β-tubulin (14) and EGFR (Fig. 1C) in infected cells are normal. Thus, reductions in the levels of these lysosomal glycoproteins are unlikely to be due to a general decrease in cell viability.

The decrease in LAMP1 levels in neisseria-infected cells is due to cleavage of the glycoprotein at its IgA1-like hinge by the bacterial IgA1 protease (14). Reductions in the levels of the other lysosomal markers are unlikely to be the direct result of protease hydrolysis, as these glycoproteins do not appear to be substrates for the enzyme. Neither CD63 (17) nor LAP (22) contains an IgA1-like hinge. A proline-rich region is present in LAMP2; however, it bears little resemblance to the hIgA1 or hLAMP1 hinges. Furthermore, IgA1 protease does not cleave purified LAMP2 in vitro (Fig. 4B). Nevertheless, IgA1 protease is indirectly involved in reducing LAMP2 level and LAP activity. Cells infected with GCM740Δ4, a genetically defined, protease-deficient neisseria mutant, have significantly higher LAMP2 levels (Fig. 1B) and LAP activity (Fig. 5) than cells infected with its WT isogenic parent, GCM740.

The function of LAMP1 is unknown. It has been hypothesized that it protects the lysosomal membrane from digestion by the hydrolytic enzymes within this compartment (7, 13). If this is the case, then even a partial reduction in LAMP1 levels, as is observed in neisseria-infected cells, may exert a negative effect on the lysosome, leading to a reduction in the total number of functional lysosomes in a cell as well as in the stability of such compartments. Indeed, rapid degradation of LAMPs was observed when N glycosylation of these proteins was inhibited (1). If the hypothesis regarding LAMP1 function is correct, the IgA1 protease may indirectly affect the levels of LAMP2, LAP, and CD63 by accelerating LAMP1 degradation. The IgA1 protease has also been shown to promote intracellular survival of the pathogenic neisseriae (14). This function is also likely to be linked directly to the action of the protease on the lysosome.

Interestingly, LAMP1 (14) and LAMP2 (Fig. 1B) levels and LAP activity (Fig. 5) are moderately reduced in cells infected with GCM740Δ4, the protease-deficient neisseria mutant, compared to levels in uninfected cells. Thus, the reduction in the levels of these three lysosomal markers is not due entirely to the IgA1 protease. These results suggest that at least one other neisseria factor contributes to these alterations.

In summary, we have shown that infection of epithelial cells by the pathogenic neisseriae reduces the levels of multiple lysosomal constituents. Further work is required to determine the exact molecular bases for these changes; however, our data are consistent with the notion that IgA1 protease-mediated degradation of LAMP1 is responsible in part for them.