Purification of the inlB Gene Product of Listeria monocytogenes and Demonstration of Its Biological Activity

Abstract

Entry of Listeria monocytogenes into nonphagocytic cells requires the inlAB gene products. InlA and InlB are bacterial cell wall-associated polypeptides that can be released by sodium dodecyl sulfate treatment. By applying more gentle extraction methods, we have purified InlB in its native form. Treatment of bacteria with various nondenaturating agents including mutanolysin, thiol reagents, sodium chloride, and detergents like Triton X-100 or 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate did not release substantial amounts of InlB from the bacterial cell wall. Instead, InlB was nearly quantitatively extracted in a solubilized form by treatment of bacteria with 1 M Tris-Cl or other protonated amines at pH 7.5. However, the reduced solubility of the extracted InlB in low-salt buffers hampered further biochemical purification. A panel of monoclonal antibodies against listerial Tris-Cl extracts containing InlB was therefore produced to generate reagents for use in affinity chromatography. One of the monoclonal antibodies enabled purification of the InlB protein to homogeneity with relatively high yields. When added externally, purified InlB associated with the surface of noninvasive bacteria such as Listeria innocua or an L. monocytogenes inlB2 mutant, where it promoted entry of these strains into Vero cells >300- and 17-fold, respectively. This effect was even more dramatic for HeLa cells, where the observed invasion was increased about 9,000- and 4,000-fold, respectively. The availability of purified native, invasion-competent InlB will allow analysis of the molecular basis of InlB-mediated entry into tissue culture cell lines in greater detail.

Listeria monocytogenes is a gram-positive, facultative intracellular bacterium that causes food-borne infections in animals and humans with severe implications, especially for newborns and immunocompromised individuals. The initial site of entry into the host normally occurs in the gut following ingestion of Listeria-contaminated food. Subsequently invading bacteria can breach the intestinal, placental, or blood-brain barrier, leading to systemic infections. During this process, the bacterium must also be capable of invading nonphagocytic parenchymal, epithelial, and endothelial cells (14, 19, 21). Early electron microscopic studies demonstrated the ability of this pathogen to invade epithelial cells of the cornea and the intestine in vivo (33, 34). A variety of nonphagocytic cell lines from different tissues, including epithelial cells, fibroblasts, and hepatocytes, can be infected in vitro with L. monocytogenes (5, 17, 26). Like Salmonella, Shigella, and Yersinia species, L. monocytogenes actively triggers its entry into these nonphagocytic cells. This process, also termed induced phagocytosis, involves host cell signalling pathways leading to rearrangements of the cortical actin cytoskeleton (2, 13).

Transposon-induced mutagenesis enabled the isolation of noninvasive mutants of L. monocytogenes and subsequently led to the identification of a genetic locus coding for the internalin A (InlA) and internalin B (InlB) polypeptides, which were identified as proteins with molecular weights of 88,000 and 65,000, respectively (17, 26). Monoclonal antibodies (MAbs) generated against either internalin detected both the InlA and InlB polypeptides in sodium dodecyl sulfate (SDS) cell wall extracts and culture supernatants of L. monocytogenes. The expression of both polypeptides was shown to be strongly dependent on growth temperature and the PrfA regulator protein. Transcription analysis of the inlAB locus revealed that these genes are transcribed both in an operon as well as individually by PrfA-dependent and -independent mechanisms (9, 10, 26).

Evidence that InlA is involved in invasion of nonphagocytic cells stems from genetic complementation studies, in which InlA when expressed in noninvasive Listeria innocua rendered this strain invasive for the human enterocyte cell line Caco-2 (17). InlA mediates entry into Caco-2 and other cell lines expressing its receptor, the cell adhesion molecule E-cadherin (29). Entry of bacteria requires the surface-bound form of InlA, which is tethered to the bacterial cell wall by a 20-amino-acid C-terminal region harboring an LPXTG motif followed by a membrane-spanning region of about 20 amino acids and a few positively charged amino acid residues (9, 35).

Unlike InlA, InlB is highly enriched in cell wall extracts and only weakly detectable in culture supernatants of L. monocytogenes (26). Despite its presence in cell wall extracts of these bacteria, InlB is unusual because its primary sequence harbors neither a C-terminal membrane anchor nor a cell wall anchoring motif, both of which are present in the InlA polypeptide (8, 9, 17). Recently, it has been shown that the 230-amino-acid C-terminal region comprising about three 80-amino-acid repeats that start with the motif Gly-Trp (GW) is responsible for the association of InlB with the bacterial cell wall (3).

By constructing isogenic chromosomal deletion mutants, it was recently demonstrated that InlB is also a crucial virulence factor for L. monocytogenes. In mice infected intraperitoneally with inlB deletion mutants, such strains were attenuated for virulence in comparison to the wild-type strain (26). Dramsi and colleagues (8) reported that the InlB polypeptide was essential for entry into hepatocytes but not for invasion of epithelial Caco-2 cells. Nevertheless, heterologous expression of inlB in L. innocua failed to promote entry of this recombinant strain into hepatocytic cell lines, suggesting that additional products of L. monocytogenes are involved in the uptake (8). Also, significant impairment of inlB2 deletion mutants was observed with respect to entry into different epithelial-like cells, such as the human HEp-G2, HeLa, or A549 cells (7, 26), and human umbilical vein endothelial cells (32).

In this study, we sought independent experimental evidence that the InlB polypeptide does indeed mediate bacterial adherence and internalization. Here we report on a simple procedure to purify the native inlB gene product of L. monocytogenes in large quantities for biochemical and functional analysis. Purified InlB was found to be highly active and promoted entry into two cell lines when added externally to noninvasive Listeria strains.

MATERIALS AND METHODS

Bacterial strains, cultivation, and reagents.

The wild-type L. innocua strain (NCTC 11288), L. monocytogenes EGD (serotype 1/2) and the isogenic EGD inlB2 deletion mutant, and the L. monocytogenes actA2/pERL3 50-1 strain have been described previously (20, 26). All Listeria strains were grown in brain heart infusion broth (Difco, Detroit, Mich.) overnight at 37°C and with erythromycin (5 μg/ml) in the case of L. monocytogenes actA2/pERL3 50-1. All chemical reagents were purchased from Sigma, Deisenhofen, Germany, unless indicated otherwise.

The African green monkey kidney cell line Vero (ATCC CCL81) and the human epithelial-like cell line HeLa (ATCC CCL 2) were cultured in Dulbecco’s modified Eagle’s medium (Gibco, Life Technologies, Eggenstein, Germany) supplemented with 10% fetal calf serum (Gibco), 2 mM l-glutamine (Sigma), and 1% nonessential amino acids (Gibco) at 37°C in 10% CO₂.

Extraction of bacterial cell wall proteins with SDS, Triton X-100, CHAPS, and Tris-Cl.

Exponentially growing bacterial cultures were harvested by centrifugation (6,000 × g for 10 min) and washed with phosphate-buffered saline (PBS) twice at room temperature. Pelleted bacteria were immediately resuspended in approximately 0.5% of the original culture volume, using PBS containing either 2% (wt/vol) SDS, 1% (vol/vol) Triton X-100, or 16.2 mM 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS) (27) or in Tris-Cl buffer at different concentrations and pH values. Resuspended bacteria were incubated for 15 min at 37°C with gentle shaking, except in the case of Tris-Cl extractions, for which bacteria were incubated for 60 min on ice. Bacterial suspensions were centrifuged at 12,000 × g for 10 min; the supernatants were aliquoted and stored at −70°C.

Extraction with mutanolysin.

The bacterial pellet was resuspended in an excess of ice-cold acetone following 10 min of incubation on ice (15). After centrifugation (12,000 × g, 10 min), the pellets were resuspended in 50 mM Tris-Cl (pH 6.8); 10 U of mutanolysin per ml was added, followed by overnight incubation at 37°C with gentle shaking. After centrifugation (12,000 × g, 10 min), the supernatants were aliquoted and stored at −70°C.

SDS-PAGE and immunoblotting.

SDS-polyacrylamide gel electrophoresis (PAGE) was done with 10% polyacrylamide gels. For staining of the gels, a silver staining kit (Bio-Rad, Munich, Germany) was used. Immunoblotting was performed by a semidry method using Immobilon P membranes (Millipore, Eschborn, Germany). After incubation with horseradish peroxidase-conjugated secondary antibodies (Dianova, Hamburg, Germany), the blots were reacted by using a sensitive enhanced chemiluminescence-based immunoblot assay (Amersham Buchler, Braunschweig, Germany) as instructed by the vendor.

MAbs.

Tris-Cl extracts of L. monocytogenes actA2/pERL3 50-1 (see above) were dialyzed against PBS, and approximately 100 μg of protein was used for repeated immunizations of three BALB/c mice. The immunization and fusion protocols have been described previously (31). Hybridoma supernatants were tested by enzyme-linked immunosorbent assay (ELISA) on separate microtiter plates coated with Tris-Cl extracts of EGD or the isogenic mutant EGD inlB2. Clones that were positive on EGD extracts and negative on EGD inlB2 extracts were further analyzed by immunoblotting prepared with Tris-Cl extracts of the EGD strain. Positive clones were subcloned twice by limiting dilution, and immunoglobulin subclasses were determined by using an isotyping kit (Medac, Hamburg, Germany). Five InlB-specific MAbs of the immunoglobulin G (IgG) subclass were further processed (see below). The InlB-specific MAb IC100 has been described previously (26).

Purification of InlB. (i) Gel filtration and phenyl-Sepharose chromatography.

Tris-Cl extracts were concentrated by ultrafiltration (cutoff, 30 kDa) and dialyzed extensively against 20 mM Tris-Cl (pH 7.6) containing 1 M NaCl. After high-speed centrifugation (100,000 × g, 30 min, 4°C), the sample (3 ml) was loaded onto a gel filtration column at 5 ml/min, using Sephacryl S-100 (XK; 1.9 by 90 cm; Pharmacia, Uppsala, Sweden) equilibrated with the same buffer. For phenyl-Sepharose chromatography, concentrated Tris-Cl extracts were dialyzed against 3 M NaCl–20 mM Tris-Cl (pH 7.6). The sample (5 ml) was loaded on a 1 ml phenyl-Sepharose HP column (Pharmacia) equilibrated with the same buffer. InlB was eluted by a continuous salt gradient starting with 3 M NaCl.

(ii) Affinity purification of InlB.

The InlB-specific MAbs were purified from hybridoma culture supernatants on protein A-Sepharose, immobilized on CNBr-activated Separose 4B (Pharmacia), and incubated overnight at 4°C on a rotating device (batch method) with Tris-Cl extracts of L. monocytogenes actA2/pERL3 50-1 previously dialyzed against PBS. After several PBS washes, portions of the Sepharose affinity matrix were boiled with SDS sample buffer and analyzed by SDS-PAGE as previously described (31). For scaling up the affinity purification, the affinity matrix was filled into an appropriate column, washed with PBS, and then washed with 1.5 M NaCl in PBS to remove nonspecifically bound material. Before the elution was started, the Sepharose was washed with 5 volumes of PBS. Finally, the bound antigen was eluted with 0.1 M sodium citrate (pH 4.0). The eluted fractions were immediately neutralized with 1 M Tris-Cl (pH 8.9) and analyzed by SDS-PAGE. Relevant fractions were pooled, dialyzed against PBS, and concentrated by repeated centrifugation in Centriprep-10 concentrators (Amicon, Witten, Germany) at 4°C, using PBS.

Amino acid sequence analysis.

For N-terminal amino acid sequence analysis, samples of trichloroacetic acid-precipitated fractions were separated by SDS-PAGE and blotted onto polyvinylidene difluoride membranes (Problot; Applied Biosystems, Weiterstadt, Germany) with a mini trans-blot electrophoretic transfer cell (Bio-Rad) for 2 h at 150 mA. The blot was then stained with amido black, and the regions corresponding to the proteins of interest were excised and subjected to analysis on an Applied Biosystems gas-phase sequenator (model A470) equipped with an on-line phenylthiohydantoin amino acid analyzer.

Analytical ultracentrifugation analysis.

To determine the native form of purified InlB, sedimentation-diffusion equilibrium runs were performed in a Beckman model F analytical ultracentrifuge. Samples with InlB at 230, 490, and 700 μg/ml in PBS were run to equilibrium in a 12-mm charcoal-filled Epon double-sector cell with sapphire windows in a two-place AN-H rotor. Two independent runs were performed. After 20 and 24 h, the samples were scanned at 280 nm and evaluated as previously described (16).

Reassociation of externally added InlB to bacteria.

One milliliter of an overnight culture of bacteria grown in brain heart infusion broth was pelleted by centrifugation in an Eppendorf tube. The resulting pellet was washed three times in PBS (pH 7.4) before addition of various concentrations of purified InlB (4 and 10 μg/ml). The mixture was incubated for 30 min at 30°C with continuous shaking (Thermomixer; Eppendorf, Hamburg, Germany), pelleted and washed five times in PBS, resuspended in Dulbecco modified Eagle medium, and used in the invasion assay as described previously (7).

RESULTS AND DISCUSSION

Selective extraction of the native InlB from bacterial pellets.

The inlB gene product was easily identified as a distinct protein band of 65 kDa in SDS extracts of L. monocytogenes EGD and was also present in small amounts in concentrated culture supernatants, where it showed extensive degradation (26). We tested various buffer combinations containing nondenaturing detergents or enzymes to extract the inlB gene product in its native form from the surface of L. monocytogenes.

For this purpose, sedimented bacteria were resuspended with small volumes of buffers containing the respective reagents and incubated for various times. Subsequently the supernatants were analyzed by SDS-PAGE (Fig. 1A) and immunoblotting (Fig. 1B) using the InlB-specific MAb IC100 (26). Figure 1 compares different extraction methods using SDS, Triton X-100, CHAPS, and mutanolysin. In comparison with the SDS treatment (Fig. 1, lane 1), neither nonionic detergents such as CHAPS (Fig. 1, lane 2), Triton X-100 (Fig. 1, lane 3), nor mutanolysin (Fig. 1, lane 4) were not effective; only small quantities of InlB were released, compared to a number of proteins other than InlB that were extracted in relatively large amounts. Treatment of Listeria pellets with high-salt solutions such as 1 M NaCl was also ineffective, as were treatments with reducing agents such as 10 mM dithiothreitol (data not shown). To our surprise, InlB was efficiently and very selectively released from the bacterial surface upon incubation with buffers containing high Tris-Cl concentrations (Fig. 1, lane 5). Triethanolamine chloride, imidazole, and ammonium chloride were also effective, but uncharged hydroxylamine was not (data not shown). As shown in lane 5 of Fig. 1, only small quantities of other listerial surface proteins were extracted by Tris-Cl, and the most predominant contaminating polypeptide showed a molecular weight slightly higher than that of InlB when analyzed by SDS-PAGE.

FIG. 1.

Analysis of extracts generated by treatment of L. monocytogenes with different detergents and other reagents. (A) Silver-stained SDS-gel (10%). Extracts were obtained by treating bacterial pellets (strain actA2/pERL3 50-1) with 2% SDS (lane 1), 16.2 mM CHAPS (lane 2), 1% Triton X-100 (lane 3), mutanolysin (lane 4), and 1 M Tris-Cl (pH 7.5) (lane 5). (B) Corresponding immunoblot reacted with InlB MAb IC100. Note that InlB was selectively extracted by Tris-Cl in lane 5. The arrow indicates the InlB protein band. Molecular mass markers (bars on the left), from top to bottom: 97, 66, 45, and 31 kDa.

To optimize the extraction conditions, we varied both the Tris-Cl concentration and the pH. The best results were obtained with 1 M Tris-Cl within a pH range of 7.25 to 8.0, where Tris is protonated (Fig. 2). As described above, the InlB was not extractable with 1 M NaCl, indicating that it was not the high ionic strength but rather a Tris effect that resulted in efficient release of InlB from the bacterial surface. The Tris-effect seems to be specific for amine functions, and although amines could act through various mechanisms, we assume that specific amino-carboxylate salt bridges may be involved in anchoring InlB to components of the cell wall. The high isoelectric point of InlB of 10.1 suggests that Tris might compete with the free amino groups of InlB that can interact with constituents of the cell wall. Similar Tris effects were described by Keen et al. (25), using 0.5 M Tris-Cl (pH 6.5) for reversibly dissociating the components of isolated coated vesicles. In that case, the Tris effect has been successfully used to identify and purify the various protein constituents of coated vesicles (36).

FIG. 2.

Effects of varying pH values on the extraction of InlB from bacterial pellets. The pH of the Tris-Cl buffer was adjusted to 7.0 (lane 1), 7.25 (lane 2), 7.5 (lane 3), 8.0 (lane 4), and 9.0 (lane 5). (A) Silver-stained extracts; (B) corresponding immunoblot reacted with InlB-specific MAb IC100. Molecular mass markers (bars on the left) indicate 66 kDa.

To verify the identity of the InlB polypeptide, we determined the N-terminal amino acid sequence of the 65-kDa band (Fig. 1). The resulting sequence, ETITVSTP, corresponded to amino acid residues 35 to 42 of the InlB sequence (17, 26). In our initial experiments, we used L. monocytogenes EGD/pERL3 50-1 for extractions with Tris-Cl. This recombinant strain harbors additional copies of the prfA gene and expressed larger amounts not only of the InlB polypeptide but also of the other listerial virulence factors (20). The N-terminal sequence analysis revealed contamination of the InlB protein band with ActA degradation products (data not shown), which prompted us to use the isogenic L. monocytogenes EGD actA deletion mutant, complemented with prfA (actA2/pERL3 50-1), for further purification of InlB.

Purification of InlB and generation of new InlB-specific MAbs.

When we tried to isolate InlB from the bacterial Tris-Cl extracts with biochemical methods, the rather unusual nature of the protein complicated the application of standard chromatography material. To apply ion-exchange chromatography (Mono-Q and Mono-S), the Tris-Cl extract was dialyzed against buffers with low ionic strength (20 mM Tris-Cl or 20 mM sodium phosphate [pH 7.5]). Under these conditions, InlB was readily sedimentable by low-speed centrifugation.

By testing several standard buffers in combination with high-speed centrifugation, we found that InlB remained soluble in solutions of moderate ionic strength such as PBS. However, interaction of InlB with standard ion-exchange material was drastically reduced under these conditions. When Tris-Cl extracts were chromatographed on a Sephacryl S-100 column equilibrated with 1 M NaCl–20 mM Tris-Cl (pH 7.6), low-molecular-weight proteins were easily separable from the InlB fraction, which still contained a protein with a slightly higher molecular weight of approximately 70,000. Even though the use of a phenyl-Sepharose matrix enabled the separation of InlB from the 70-kDa polypeptide (data not shown), a combination of both chromatography procedures was not satisfactory due to the low yield of purified InlB. We therefore tried to establish a more convenient purification protocol for InlB.

We previously succeeded in purifying the ActA polypeptide from concentrated crude Listeria culture supernatants by affinity chromatography using ActA-specific MAbs (31). Since the available InlB-specific MAb IC100 poorly reacted with the native InlB protein (data not shown), we generated a new panel of InlB-specific MAbs by immunizing mice with Tris-Cl extracts of the InlB-overproducing actA2/pERL3 50-1 strain. Hybridoma supernatants were screened by ELISA using microtiter plates coated with Tris-Cl extracts derived from the EGD wild-type strain and its isogenic inlB2 mutant, respectively. Supernatants only positive on the EGD extracts were then further analyzed by immunoblotting using Tris-Cl extracts from the EGD wild-type strain and from the isogenic inlB deletion mutant. We selected InlB-specific hybridomas that produced IgGs and purified them from culture supernatants on protein A-Sepharose. Purified IgGs were immobilized on activated Sepharose 4B and incubated with the Tris-Cl extracts that had been dialyzed extensively against PBS. Washing and elution conditions were tested as recently described for the immunoaffinity purification of ActA (31). One of the MAbs tested, clone IF32, was found to recognize and to bind the native InlB polypeptide present in the Tris-Cl extracts. Following overnight incubation of dialyzed Tris-Cl extracts with immobilized IF32 MAb, the affinity matrix was processed as described in Materials and Methods. InlB was not dissociated from the immobilized MAbs by extensively washing the affinity matrix with buffer containing 1.5 M NaCl. This rather stringent treatment enabled the removal of nonspecifically adsorbed contaminants. Elution conditions were optimized in terms of ionic strength and pH (data not shown), and InlB was quantitatively eluted with 0.1 M citrate at pH 4.0. Positive fractions were pooled and revealed a single band at 65 kDa on silver-stained SDS-gels (Fig. 3, lane 2). N-terminal amino acid sequencing of the purified protein immobilized on a blot membrane gave the sequence ETITVSTPIKQIF, identical to that of the predicted sequence of secreted InlB (17, 26). The immunoaffinity purification procedure yielded approximately 1 mg of pure InlB from 4.5 liters of stationary-growth-phase bacterial cultures. Due to the relatively mild elution conditions at pH 4.0, the affinity matrix was reusable for several purification cycles without any measurable loss of activity. Thus, the affinity chromatography using the InlB-specific MAb IF32 turned out to be a very convenient purification procedure. Within one purification step, this rapid method yielded large quantities of homogeneous InlB from crude Tris extracts, enabling further characterization of this polypeptide.

FIG. 3.

Purification of InlB from Tris-Cl extracts of L. monocytogenes by immunoaffinity chromatography. (A) Silver-stained SDS-gel (10%) of the Tris-Cl extract of L. monocytogenes actA2/pERL3 50-1 (lane 1) which was incubated with immobilized MAb IF32. After extensive washing of the matrix, InlB was eluted with 0.1 M sodium citrate, pH 4.0 (lane 2). (B) Corresponding immunoblot reacted with MAb IC100. Molecular mass markers (bars on the left), from top to bottom: 97, 66, and 45 kDa.

To determine the native form of purified InlB under physiological conditions, we performed sedimentation-diffusion equilibrium runs in PBS at different protein concentrations, using an analytical ultracentrifuge. At protein concentrations between 100 and 700 μg/ml, no self-association of InlB was detectable (data not shown), suggesting that under physiological conditions the purified InlB polypeptide is a stable, monomeric protein.

Purified InlB promotes efficient internalization of noninvasive listeriae.

It was recently shown that prior incubation of spent culture supernatants of InlB-expressing bacteria confers invasiveness to an inlB mutant (3). This study was recently extended by assessing the ability of purified recombinant InlB proteins expressed in Escherichia coli to promote invasion of an inlB mutant and noninvasive L. innocua (4). We used this assay to assess the activity of the purified InlB protein by examining the ability of externally added purified InlB to promote the invasivity of the wild-type L. monocytogenes EGD strain, an isogenic inlB2 mutant, and a noninvasive L. innocua strain. Preincubation of bacteria with purified InlB resulted in strong association of the protein with the bacterial cell wall, which was readily visible in Coomassie blue-stained SDS extracts of the inlB2 mutant strain (Fig. 4A). Specific association of InlB to the bacterial surface was confirmed by using MAb IF32, specific for InlB (Fig. 4B). In invasion assays, the addition of increasing amounts of purified InlB to noninvasive strains rendered them highly competent for entry into the Vero tissue culture cell line. Thus, in the experiment presented, prior addition of InlB at 4 μg/ml increased invasion of the inlB2 mutant 17-fold and that of a noninvasive L. innocua strain greater than 300-fold (Fig. 5). No increase in invasion was observed using the wild-type strain for this cell line. The results were more dramatic when invasion was examined in the epithelial-like cell line HeLa. Thus, externally added InlB at 4 μg/ml increased invasion of the wild-type strain, the inlB2 mutant, and L. innocua 6-, 3,700-, and 9,000-fold, respectively (Fig. 5). These results demonstrate that the purified native InlB isolated in this study is highly active and promotes invasion of noninvasive bacteria into two different cell lines.

FIG. 4.

Purified InlB binds to the surface of L. monocytogenes. Bacteria (10⁸/ml) of the L. monocytogenes EGD inlB2 mutant were incubated with 4 μg of purified InlB for 30 min at room temperature as described in Materials and Methods and analyzed following SDS-PAGE by Coomassie blue staining (A) or immunoblotting using InlB-specific MAb IC100 (B). SDS extracts from the L. monocytogenes EGD inlB2 mutant (lane1) and the same mutant preincubated with 4 μg of purified InlB (lane 2). The position of InlB in panel A is indicated by an arrow. Molecular mass markers (bars on the left), from top to bottom: 97, 66, 45, 31, 21, and 14 kDa.

FIG. 5.

Externally added purified InlB protein enhances entry of invasive and noninvasive Listeria strains into Vero (A) and HeLa (B) cell lines. Bacteria were either not treated or preincubated with InlB at 4 μg/ml before performance of the invasion assay as described previously (10). Values along the vertical axis are given relative to the invasion of wild-type strain EGD, which is arbitrarily fixed at 100. Bars indicate standard deviations.

Braun et al. (4) previously reported very similar results, e.g., that purified recombinant InlB expressed in E. coli promoted, when added externally, entry of noninvasive listeriae into Vero cells. Compared with our entry rates of reconstituted, formerly noninvasive Listeria strains, InlB purified from L. monocytogenes seems to be more active than that expressed in E. coli. Further evidence that InlB is sufficient to promote entry into several nonphagocytic cell lines stems from experiments in which inert latex beads coated with purified InlB were internalized by Vero, HEp-2, and HeLa cells (4).

The invasion of nonphagocytic cells by bacterial pathogens has extensively been studied for the gram-negative genera Salmonella, Yersinia, and Shigella (11, 18, 24, 28, 30; for further references, see references 2 and 13). In the case of Yersinia, these studies have led to the identification of a bacterial ligand, invasin, and its cellular receptor, β1 integrin, both of which serve as paradigms for studying bacterial invasion into nonphagocytic cells (23). Among gram-positive bacteria, the pathogenic mechanisms of L. monocytogenes have been studied extensively in the past few years. Analysis of the invasion process of L. monocytogenes has recently led to the identification of the internalin gene family (17) and E-cadherin as the cellular receptor of the InlA polypeptide (29).

Unlike InlA, no protocols have yet been established for the purification of InlB from L. monocytogenes except that for recombinant InlB expressed in E. coli (3, 4). For InlA, purification of the protein was achieved by using concentrated supernatant cultures. However, InlB is hardly detectable in supernatant fluids but could be nearly quantitatively extracted from the bacterial cell wall by using 1 M Tris-Cl. Such extracts were almost devoid of other contaminating proteins, suggesting that this extraction procedure exploits an unusual property of the InlB protein. Purified InlB is a monomeric protein that is capable of conferring invasive properties to normally noninvasive bacteria for two different cell lines. This finding provides strong evidence that InlB can by itself bind to the surface of eucaryotic cells and promote invasion. We found that InlB did not promote invasion of bacteria into the Caco-2 enterocytic cell line (data not shown), indicating that entry into Vero and HeLa cells are InlB dependent, a result that is consistent with previously published data for deletion and complementation mutants of L. monocytogenes (10, 26).

The observed association of purified InlB with the cell wall of either L. monocytogenes and L. innocua suggests that it interacts with a surface molecule that is common to both pathogenic and nonpathogenic Listeria species. The association of externally added InlB to its potential receptor on the bacterial surface was extremely strong and could be released only by SDS extraction (Fig. 4) or the Tris-based extraction method developed in this study (data not shown). Similar association of cell wall proteins from gram-positive bacteria to receptors on the bacterial surface have been previously demonstrated. Thus, the C-terminal repeat domains of Streptococcus pneumoniae amidase promotes binding to choline within the bacterial cell wall (6). The secreted bacteriocin, lysostaphin of Staphylococcus simulans, binds specifically to the surface of Staphylococcus aureus, where it cleaves peptidoglycan. In this case, the target cell specificity is conferred by a 92-amino-acid C-terminal region (1) that had previously been shown to harbor some homology to the 80-amino-acid GW repeat in InlB of L. monocytogenes (3). Since externally added InlB promotes invasion, it will be of great interest to identify the bacterial InlB cell wall receptor required for this interaction.

The simple and rapid purification of InlB by immunoaffinity chromatography should now stimulate more studies on its biological properties. We have examined its ability to associate with only two cell lines here; these studies can now be expanded to include both primary cell cultures and various tissue culture cell lines. Thus, it will be possible to determine the region(s) on the InlB molecule required for interaction with host cell receptors. The simple invasion assay used in this study can now also be applied to examine the adhesive and invasive properties of the other leucine-rich-repeat-containing proteins of L. monocytogenes, such as InlA and IrpA (7, 12). Thus, events within the host cell, such as increase in phosphoinositide 3-kinase activity, that are required for bacterially mediated entry (22) can now be directly examined by assessing the ability of purified recombinant InlB mutant proteins to promote bacterial invasion. Finally, purification of the InlB protein now makes it possible to identify the ligand that it binds to on the surface of the eucaryotic cell.