Identification and Characterization of a Neospora caninum Microneme-Associated Protein (NcMIC4) That Exhibits Unique Lactose-Binding Properties

Abstract

Microneme proteins have been shown to play an important role in the early phase of host cell adhesion, by mediating the contact between the parasite and host cell surface receptors. In this study we have identified and characterized a lectin-like protein of Neospora caninum tachyzoites which was purified by α-lactose-agarose affinity chromatography. Upon separation by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, this lactose-binding protein migrated at 70 and 55 kDa under reducing and nonreducing conditions, respectively. Immunofluorescence and immunogold electron microscopy with affinity-purified antibodies showed that the protein was associated with the tachyzoite micronemes. Mass spectrometry analyses and expressed sequence tag database mining revealed that this protein is a member of the Neospora microneme protein family; the protein was named NcMIC4 (N. caninum microneme protein 4). Upon two-dimensional gel electrophoresis, NcMIC4 separated into seven distinct isoforms. Incubation of extracellular parasites at 37°C resulted in the secretion of NcMIC4 into the medium as a soluble protein, and the secreted protein exhibited a slightly reduced M_r but retained its lactose-binding properties. Immunofluorescence was used to investigate the temporal and spatial distribution of NcMIC4 in tachyzoites entering their host cells and showed that reexpression of NcMIC4 took place 30 min after entry into the host cell. Incubation of secreted fractions and purified NcMIC4 with Vero cells demonstrated binding of NcMIC4 to Vero cells as well as binding to chondroitin sulfate A glycosaminoglycans.

Neospora caninum is a protozoan parasite causing bovine abortion or stillbirth and neuromuscular disorders in dogs (7, 8, 14, 19). Although closely related to Toxoplasma gondii, N. caninum exhibits clear differences in ultrastructure, antigenic composition, natural host range, and host cell interaction (20, 21, 27, 28, 36). N. caninum is an obligatory intracellular parasite, and upon lysis of the host cell, the invasive stages need to enter a new cell in order to survive and proliferate. Extracellular maintenance, even for short periods of time, results in a loss of infectivity of the tachyzoites (13, 29). Therefore, adhesion and invasion are essential steps in the life cycle of N. caninum.

The process of invasion is apparently relatively conserved throughout the phylum Apicomplexa (9). It involves the sequential exocytosis of three different secretory organelles, namely, micronemes, rhoptries, and dense granules (2). Microneme proteins, often containing adhesive domains, are released first at the apical end of the parasite at the onset of the initial contact with the target cell and are involved in host cell recognition and invasion (37), as well as in parasite motility (31). Rhoptry proteins are exocytosed next during host cell invasion and are responsible for the formation of the parasitophorous vacuole. As invasion is completed and the intracellular parasite is enclosed by the parasitophorous vacuole, dense granule proteins are secreted and modify the vacuolar membrane (9).

Toxoplasma and Neospora tachyzoites utilize sulfated proteoglycans to adhere to host cells (4, 28, 29, 32). In addition, some of the adhesive domains present in microneme proteins have been shown to bind to host cell surface receptors including heparan sulfate and chondroitin sulfate glycosaminoglycans (28, 30, 34). Therefore, it is reasonable to consider that protein-carbohydrate interactions play a major role during the physical interaction between parasite and host cell.

In other apicomplexan parasites, lectin-like proteins have been identified by lactose affinity chromatography. For instance, a lectin-like microneme protein was identified in Sarcocystis muris, exhibiting a high affinity for N-acetylgalactosamine (23). T. gondii microneme protein 2 (TgMIC2) has been shown to bind to host cell glycosaminoglycans via its thrombospondin (TSP)- and integrin-like domains (3). Other T. gondii microneme proteins such as TgMIC1 and TgMIC4 were isolated from a Toxoplasma extract by using lactose affinity chromatography, though only TgMIC1 was reported to possess lectin-like properties (24). Nevertheless, TgMIC1 and TgMIC4 efficiently bind to host cells (1, 10). In order to identify potential carbohydrate-binding proteins in N. caninum, we also employed lactose affinity chromatography of parasite extracts. This resulted in the purification of a single lactose-binding protein. This paper reports on the purification, localization, identification by mass spectrometry (MS), and functional characterization of this molecule.

MATERIALS AND METHODS

Unless otherwise stated, all reagents and tissue culture media were purchased from Sigma (St. Louis, Mo.).

Tissue culture and parasite purification.

Cultures of Vero cells were maintained in RPMI 1640 medium (Life Technology, Basel, Switzerland) supplemented with 7% fetal calf serum, 2 mM glutamine, 50 U of penicillin per ml and 50 μg of streptomycin per ml at 37°C under 5% CO₂ in tissue culture flasks. Cultures were trypsinized at least once a week. N. caninum tachyzoites of the Nc-Liverpool isolate were used and maintained in Vero cell monolayers (15). Intracellular parasites were harvested by trypsinization of infected Vero cells, followed by repeated passages through a 25-gauge needle at 4°C, followed by separation on PD10 columns filled with Sephadex G-25 (Pharmacia) as previously described (13, 14).

Production of polyclonal antiserum directed against N. caninum tachyzoites.

Anti-Neospora antiserum was generated as previously described (13). Briefly, a serologically negative rabbit was inoculated with 10⁷ freshly purified tachyzoites, and 10⁸ parasites were administered orally. This procedure was repeated two times at intervals of 10 days. Fourteen days after the last boost, the animal was bled, and serum was stored at −80°C.

In vitro culture of N. caninum bradyzoites.

For in vitro tachyzoite-to-bradyzoite stage conversion, monolayers of keratinocytes grown on glass coverslips were infected with tachyzoites (10⁵/ml) of the Nc-Liverpool strain in combination with 70 μM sodium nitroprusside, all in 24-well tissue culture plates. The infected monolayers were cultured at 37°C under 5% CO₂ for 8 days with a daily addition of sodium nitroprusside as described by Vonlaufen et al. (38).

Mouse brain tissue infected with N. caninum bradyzoites.

Paraffin blocks of tissue harboring N. caninum bradyzoites had been used in previous studies (11, 35) and were kindly provided by Milton McAllister, University of Illinois at Urbana.

Detergent extraction of N. caninum tachyzoites and purification of a lactose-binding protein.

A total of 10⁹ freshly purified N. caninum tachyzoites were incubated in 20 ml of phosphate-buffered saline (PBS) containing 1% Triton X-100 and 1 mM phenylmethylsulfonyl fluoride for 15 min at 4°C, followed by centrifugation at 2,000 × g for 30 min at 4°C. After centrifugation, supernatants (12,000 × g for 20 min at 4°C) were collected and submitted to affinity chromatography on a 1-ml α-lactose-agarose column previously equilibrated with ice-cold PBS containing 0.5 M NaCl. After extensive washing with equilibration buffer, the adsorbed material was eluted with 3 ml of 0.1 M lactose in equilibration buffer as described by Lourenço et al. (24). In some experiments, secreted fractions (see below) were subjected to lactose affinity chromatography, yielding identical results.

SDS-PAGE, silver staining, immunoblotting, and affinity purification of antibodies.

Triton X-100-soluble N. caninum extracts and fractions of the lactose column (flowthrough and eluates) were precipitated in methanol-chloroform (40) and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) under reducing and nonreducing conditions, respectively. The amounts of protein loaded per lane corresponded to 10⁷ tachyzoites. Gels were silver stained, stained with Coomassie blue, or transferred to nitrocellulose filters as described previously (15, 16). After the blocking of unspecific binding sites in Tris-buffered saline-3% bovine serum albumin (BSA)-0.3% Tween 20, the blots were labeled either with anti-N. caninum antiserum, diluted 1:2,000 in antibody dilution buffer (Tris-buffered saline-0.3% BSA-0.3% Tween), or with affinity-purified antibodies (see below) at dilutions of 1:30 to 1:50. The bound antibodies were visualized by using goat anti-rabbit alkaline phosphatase-conjugated immunoglobulin G (IgG) (Promega). Specific affinity-purified antibodies against the lactose-binding protein eluted by α-lactose affinity chromatography were prepared by the separation of column eluates by a preparative SDS-PAGE and transfer to nitrocellulose filters, and further steps were performed as previously described (14).

Two-dimensional gel electrophoresis, MS analysis of purified N. caninum lactose-binding protein, and database search.

Two-dimensional electrophoresis was performed essentially as described by Cohen et al. (6) for proteins of T. gondii. Briefly, purified lactose-binding protein (1 μg) was separated in the first dimension by isoelectrofocusing by using immobilized pH gradient (pH 3 to 10) strips (Amersham Biosciences, Chalfont St. Giles, Buckinghamshire, United Kingdom). After rehydration of the strips, proteins were focused at 500 V for 1 h, at 1,000 V for 1 h, and at 4,000 V for 2 h by gradient and finally at 8,000 V of constant voltage for 6 h. For separation in the second dimension, the immobilized pH gradient strips were equilibrated in 10 ml of a solution of 50 mM Tris-HCl (pH 8.8), 6 M urea, 30% glycerol, and 2% SDS containing 100 mg of dithiothreitol for 15 min, followed by incubation in 10 ml of a solution of 50 mM Tris-HCl (pH 8.8), 6 M urea, 30% glycerol, and 2% SDS containing 250 mg of iodoacetamide for a further 15 min. Strips were then embedded on 18-cm precast SDS-12% PAGE gels (Amersham Biosciences) and electrophoresed at 110 V for 24 h at 15°C. Gels were stained with either Coomassie blue or Sypro Orange (Bio-Rad, Hemel Hempstead, Hertfordshire, United Kingdom). Spots were picked and subjected to in-gel trypsin digestion and analyzed by using a Qstar pulsar electrospray mass spectrometer (Applied Biosystems, Foster City, Calif.). The resulting MS spectra were searched against the National Center for Biotechnology Information nonredundant and expressed sequence tag (EST) databases by using the MASCOT search engine (Matrix Science, London, United Kingdom), with Alveolata set as the taxonomy. MS data were extracted by using the Mascot script within AnalystQS, and the databases were searched in the tandem MS (MS/MS)-ion mode. Fixed carbamidomethyl modification and variable methionine oxidation were allowed for. Tolerances were 250 ppm for peptide mass and 35 Da for MS/MS data, allowing for one missed cleavage by trypsin. Hits were considered to be significant if P was <0.05. Hits to a number of contiguous, overlapping ESTs were obtained, and these were aligned to obtain the final sequence. The deduced amino acid sequence was investigated for potential phosphorylation and glycosylation sites by using the Expasy molecular biology server (http:/us.expasy.org; programs used were NetPhos, NetOGlyc, NetNglyc, and YinOYang), N-terminal signal peptide analysis (SignalP program), and identification of specific domains (SMART: simple modular architecture research tool [http://smart.embl-heidelberg.de]).

Immunofluorescence.

Freshly purified N. caninum tachyzoites, grown in Vero cells on coverslips, and in vitro-generated bradyzoite tissue cysts were fixed, permeabilized, and blocked as described earlier (13, 35, 40). Immunolabeling was carried out by using affinity-purified antibodies directed against eluted NcMIC4 (N. caninum microneme protein 4), diluted 1:2 in PBS-0.5% BSA. The secondary antibody was a fluorescein isothiocyanate-conjugated anti-rabbit IgG. Tachyzoites were then stained with anti-N. caninum antiserum (13), and N. caninum tissue cysts were labeled with monoclonal antibody (MAb) CC2, which stains a cyst wall-associated component in both N. caninum and T. gondii (12). These antibodies were used at dilutions of 1:300 and 1:250, respectively. They were detected by tetramethyl rhodamine isocyanate-conjugated secondary antibodies. Immunofluorescence staining on paraffin-embedded mouse brain tissue sections containing N. caninum bradyzoite tissue cysts was performed according to Keller et al. (22). For double staining, antibodies against NcMIC4 and the MAb CC2 were used as described above. Nuclei were stained with Hoechst 33258 (25 μg/ml) in PBS at a 1:300 dilution. All specimens were viewed on a Nikon Eclipse E800 digital fluorescence microscope, and images were processed by employing the Openlab version 2.0.7 software (Improvision, Heidelberg, Germany).

Immunogold labeling and TEM.

LR White embedding and on-section labeling of N. caninum-infected Vero cell cultures were performed as previously described (11, 16, 18). Sections were labeled with affinity-purified NcMIC4 antibodies diluted 1:2 in PBS-0.1% BSA, and bound antibodies were detected by using goat anti-rabbit antibody conjugated to 10-nm gold particles (Amersham, Zürich, Switzerland). Grids were stained with lead citrate and uranyl acetate (17) and were finally viewed by transmission electron microscopy (TEM) on a Phillips 300 instrument operating at 60 kV.

Secretion assays.

Freshly purified tachyzoites (10⁸ ml⁻¹) were resuspended in 2 ml of RPMI 1640 medium and were incubated at 37°C for 10 min (27). Following secretion, parasites were centrifuged (2,000 × g for 10 min at 4°C), and supernatants were collected and recentrifuged (10,000 × g for 30 min at 4°C). The parasite pellet and the supernatant from the second centrifugation were processed for SDS-PAGE and Western blotting by using affinity-purified antibodies directed against NcMIC4. Secreted fractions were stored at −80°C. Prior to binding experiments, secretory proteins were centrifuged at 20,000 × g for 30 min at 4°C to remove possible aggregates.

Binding of NcMIC4 to Vero cells and glycosaminoglycans.

NcMIC4-Vero cell coprecipitation assays were performed as previously described (22, 27). Solid-phase binding assays were also performed as described earlier (28). Ninety-six-well enzyme-linked immunosorbent assay (ELISA) plates were coated with lactose, heparin, and chondroitin sulfates A, B, and C (CSA, CSB, and CSC, respectively), each at a concentration of 5 mg/ml, overnight at 4°C. Following a washing step, unspecific binding sites were blocked with bovine hemoglobin for 2 h at room temperature and washed again. The wells were then incubated with 100 μl of either the secreted fractions (1 μg of protein) or purified NcMIC4 (0.2 μg of protein). After a further washing step, bound NcMIC4 was detected by employing the anti-NcMIC4 antibody. By using p-nitrophenylphosphate as a substrate, absorbance values were measured at 405 nm on a Dynatech MR7000 ELISA reader. Each assay was carried out in triplicate, and each value represents the mean and standard deviation.

Nucleotide sequence accession number.

Nucleotide sequence data reported in this paper are available in the EMBL, GenBank, and DDJB databases as a Third Party Annotation under the accession number BK005222.

RESULTS

Isolation and localization of an N. caninum lactose-binding protein by affinity chromatography.

In previous studies (28) it was shown that N. caninum tachyzoites bind to the Vero cell surface by preferential interaction with sulfated proteoglycans as receptors. Chondroitin sulfates had been shown to be crucially involved, although tachyzoites also bind to heparin and heparan sulfate. To exploit the carbohydrate binding nature of parasite proteins, we applied α-lactose-agarose affinity chromatography to isolate lectin-like proteins out of a Triton X-100-soluble N. caninum extract. Elution with 0.1 M lactose resulted in the isolation of a major lactose-binding protein, which upon separation by an SDS gel migrated at 70 kDa under reducing conditions and at 55 kDa under nonreducing conditions (Fig. 1A). Typically, loading 10⁹ parasites onto a lactose-agarose column resulted in a yield of 6 μg of lactose-binding protein in 3 ml of elution buffer (data not shown). Affinity-purified antibodies directed against the eluted protein exhibited a specific reactivity when assessed by immunoblotting on N. caninum extracts (Fig. 1B). The purification of lactose-binding proteins from T. gondii tachyzoites was also attempted by using the same procedure, yielding numerous (4 to 6) lactose-binding bands (data not shown). When the Toxoplasma lactose-binding proteins were subjected to immunoblotting and labeled with antibodies directed against the 70-kDa N. caninum lactose-binding protein, no reaction could be detected (data not shown). In addition, antibodies directed against the N. caninum lactose-binding protein did not bind to T. gondii tachyzoites, based on results of immunoblotting and immunofluorescence (data not shown) assays.

FIG. 1.

(A) Silver-stained SDS-PAGE of two independent purifications, showing different fractions loaded onto and eluted from α-lactose affinity chromatography columns. Lane 1, N. caninum Triton X-100-soluble extract loaded on the column; lane 2, unbound fraction (flowthrough) of the column; lanes 3 to 5, eluted fractions. In lanes 1 to 4, SDS-PAGE was performed under reducing conditions; in lane 5, eluted fractions were separated under nonreducing conditions. (B) Immunoblot of the same fractions as in panel A, labeled with the affinity-purified anti-NcMIC4 antibody and detected by goat anti-rabbit antibody conjugated to alkaline phosphatase.

Affinity-purified antibodies were used for immunofluorescence studies. Both intracellular N. caninum tachyzoites grown in Vero cells (Fig. 2A to C) and freshly purified tachyzoites (Fig. 2D) exhibited intense staining almost exclusively at the apical tip. Immunogold EM of LR White-embedded N. caninum tachyzoites showed that the 70-kDa lactose-binding protein was largely associated with the micronemes of the parasites (Fig. 2E and F).

FIG. 2.

Localization of the lactose-binding protein. Immunofluorescence labeling of N. caninum tachyzoites grown in Vero cells (A to C) and purified parasites (D). Parasites were stained with anti-NcMIC4 antibody and fluorescein isothiocyanate-conjugated anti-rabbit IgG (green or yellow) and with anti-Neospora serum and tetramethyl rhodamine isocyanate-conjugated anti-rabbit IgG (red). Nuclei were labeled with Hoechst stain. Panels E and F show TEM immunogold labeling of LR White-embedded N. caninum tachyzoites. Sections were stained with anti-NcMIC4 antibodies and secondary goat anti-rabbit antibodies conjugated to 10-nm-diameter gold particles. Arrows point towards labeled micronemes. Bar, 0.2 μm.

The lactose-binding protein is an N. caninum homologue of TgMIC4.

Purified lactose-binding protein was subjected to two-dimensional gel electrophoresis, which showed that the 70-kDa band separated into two distinct families of spots (Fig. 3). These spots were excised and subjected to MS/MS analysis. The resulting spectra of the seven spots (marked by arrows in Fig. 3) exhibited striking homologies to TgMIC4 sequences. The three spots separated by a larger gap (marked by arrowheads in Fig. 3) were identified as horse serum albumin, which was present in the culture medium and was copurified as a minor contaminant along with NcMIC4.

FIG. 3.

Two-dimensional analysis of purified NcMIC4 stained with Sypro Orange. Spots marked with arrows correspond to NcMIC4 isoforms, while those marked with arrowheads were identified as horse serum albumin.

Inspection of the N. caninum EST database revealed several clones with overlapping sequences corresponding to NcMIC4. The deduced amino acid sequence of a number of aligned overlapping EST fragments composed of 579 amino acids, is shown in Fig. 4 in comparison to the corresponding T. gondii homologue TgMIC4. All nine peptides identified through MS are found on the sequence (Fig. 4). The N-terminal, putatively cleaved signal sequence is comprised of 16 amino acids. Overall, TgMIC4 and NcMIC4 are highly homologous, with 54% identity, 67% conservative amino acid substitutions, and 33 of 34 cysteine residues located at identical positions. The major part of NcMIC4, similar to TgMIC4 (1), is comprised of six consecutive, potentially adhesive Apple domains. The deduced NcMIC4 sequence contains 28 sites, a high number, for potential serine (11), threonine (9), and tyrosine (8) phosphorylation. In addition, 17 sites for potential serine/threonine O-glycosylation could be identified (data not shown).

FIG. 4.

NcMIC4 protein sequence deduced from alignment of overlapping EST sequences (GenBank accession no. BF717084, BF824356, BG235284, CD537254, CD679931, CD537254, CD667806, CD680175, and CF274679), aligned with TgMIC4. The putative N-terminal signal peptides are marked in italics, with a cleavage site between amino acids 16 and 17 for NcMIC4. The NcMIC4 protein sequence contains 34 cysteines and six consecutive Apple domains (underlined). The peptides identified by MS/MS are shown in bold.

Differential expression of NcMIC4 in tissue cysts that contain N. caninum bradyzoites.

Paraffin sections of mouse brains containing N. caninum bradyzoite tissue cysts were labeled with anti-NcMIC4 antibody and MAb CC2. Double immunofluorescence revealed two types of staining within these tissues, corresponding to different degrees of cyst maturity (22). In tissue cysts exhibiting intravacuolar MAb CC2 labeling, NcMIC4 was still strongly expressed at the apical tip of the parasite (Fig. 5A). In contrast, tissue cysts exhibiting strictly peripheral MAb CC2 staining were regarded as more mature cysts and showed no reactivity with anti-NcMIC4 antibodies (Fig. 5B). The expression of NcMIC4 in N. caninum bradyzoites cultured in vitro (38, 39) was also analyzed, and the in vitro situation was quite different. After 8 days of in vitro culture, bradyzoites still expressed NcMIC4, although peripheral staining with MAb CC2 was evident (Fig. 5C).

FIG. 5.

Expression of NcMIC4 investigated by immunofluorescence. (A) A largely immature N. caninum tissue cyst in a paraffin section of mouse brain, showing intracystic NcCC2-staining by using MAb CC2 (red) and punctated staining pattern in parasites. (B) Mature N. caninum tissue cyst in a paraffin-embedded mouse brain, exhibiting peripheral, cyst wall-associated NcCC2 labeling (red) but severely downregulated expression of NcMIC4. (C) In vitro-cultured N. caninum bradyzoites exhibiting cyst wall staining with MAb CC2 (red) and marked expression of NcMIC4 (green). (D) Time course of NcMIC4 expression (green) in N. caninum tachyzoites entering and developing within their host cells. Note that at 30 min following incubation of the parasite with host cells, NcMIC4 expression is not detectable, but expression is restored after 4 h and at late time points.

Secretion and resynthesis of NcMIC4 following host cell entry.

As a microneme protein, NcMIC4 was bound to be a component of the secreted fractions of N. caninum tachyzoites. In order to follow the kinetics of secretion and resynthesis of NcMIC4 following host cell invasion, Vero cell monolayers were allowed to interact with freshly purified tachyzoites at 37°C for different periods of time (Fig. 5D). Immunofluorescence staining revealed that most parasites had released NcMIC4 during the initial 30 min of host cell interaction, and the protein was not detectable at this time point. At 4 h, some parasites had started to resynthesize NcMIC4, and after 24 and 48 h, reexpression of NcMIC4 was fully restored, as revealed by distinct staining at the apical tip.

In order to investigate whether NcMIC4 was released as a soluble protein, freshly purified tachyzoites were incubated at 37°C to induce secretion, and control tachyzoites were maintained at 4°C. Following centrifugation, the supernatants were processed for SDS-PAGE and immunoblotting. Figure 6 shows that NcMIC4 was secreted into the medium supernatant. In order to investigate whether NcMIC4 would be modified during the secretion process, intracellular NcMIC4 (in a Triton X-100-soluble extract) and secreted NcMIC4 (secreted fraction) were separated by SDS-PAGE and transferred to nitrocellulose. Comparative staining with anti-NcMIC4 antibody demonstrated a slight difference in the M_rs of intra- and extracellular NcMIC4 (Fig. 6B). However, secretion of NcMIC4 did not alter its unique lactose-binding properties. This was shown by comparative SDS-PAGE of NcMIC4 obtained through lactose affinity chromatography of Triton X-100 soluble fractions and secreted fractions. Again, NcMIC4 purified from secreted fractions exhibited a reduced M_r (Fig. 6C).

FIG. 6.

Secretion of NcMIC4. (A) Immunoblots of fractions secreted by tachyzoites when incubated at either 4 or 37°C for 10 min. (B) Comparison of an N. caninum Triton X-100-soluble extract (iNcMIC4) and the fraction secreted (sNcMIC4) at 37°C, stained with anti-NcMIC4 antibodies. (C) Silver-stained SDS-PAGE of purified NcMIC4 obtained from Triton X-100 extracts (iNcMIC4) and secreted fractions (sNcMIC4). Note the different M_rs of intracellular and secreted NcMIC4 proteins.

NcMIC4-host cell interaction.

To investigate whether NcMIC4 could possibly be involved in tachyzoite-host cell binding, a secreted fraction of N. caninum tachyzoites was incubated with Vero cells, and following centrifugation, the pellet containing Vero cells and bound parasite proteins and the supernatant with the unbound proteins were separated by SDS-PAGE and transferred to nitrocellulose filters. Immunoblotting revealed that a substantial portion of NcMIC4 coprecipitated with Vero cells (Fig. 7A). As a control the same fractions were stained with anti-NcMIC1 (positive control [22]) and anti-NcGRA7 antibodies (negative control [18]) (Fig. 7B and C, respectively). Further resuspension of Vero cell-NcMIC4 complexes and subsequent centrifugation did not result in dissociation of NcMIC4 binding, and coprecipitation assays using purified NcMIC4 produced similar results (data not shown). Solid-phase binding experiments were performed in order to obtain information on which molecules on the Vero cell surface could be involved in the binding of NcMIC4. ELISA wells were coated with lactose, CSA, CSB, CSC, and heparin. As shown in Fig. 8, besides its strong affinity for lactose, NcMIC4 exhibited a significantly increased binding to CSA-glycosaminoglycan, while CSB, CSC, and heparin did not bind NcMIC4.

FIG. 7.

Immunoblots of parasite proteins coprecipitated with prefixed Vero cells that were incubated with secreted fractions of N. caninum tachyzoites. Western blots were stained with anti-NcMIC4 (A), anti-recombinant NcMIC1 (B), and anti-GRA7 (C) antibodies. Lanes P (pellet) represent the protein fraction bound to Vero cells; lanes SN (supernatant) are nonbound proteins.

FIG. 8.

Binding of purified NcMIC4 to solid-phase-bound lactose, CSA, CSB, CSC, and heparin. Binding of NcMIC4 was assessed by labeling with anti-NcMIC4 antibodies and alkaline phosphatase-conjugated secondary antibodies. Note that NcMIC4 binds preferentially to CSA but not the other glycosaminoglycans. OD 405, optical density at 405 nm.

DISCUSSION

In this study we report on the purification and characterization of a lectin-like protein (NcMIC4) from N. caninum tachyzoites, which was found to be a homologue of the previously described TgMIC4 (1). Both proteins are composed of six potentially adhesive Apple domains, of which the most C-terminal one in TgMIC4 has been shown to be responsible for host cell binding (1). TgMIC4 is abundantly expressed in tachyzoites, bradyzoites, sporozoites, and merozoites and is proteolytically processed upon secretion (1). In contrast to TgMIC4, NcMIC4 exhibits unique lactose-binding properties, allowing it to be efficiently purified out of a complex N. caninum tachyzoite Triton X-100-soluble extract by simple α-lactose-agarose affinity chromatography.

Purified NcMIC4 migrates at different M_rs when separated by SDS-PAGE under reducing and nonreducing conditions (Fig. 1). This finding indicates that the numerous cysteine residues found within the deduced amino acid sequence could be important determinants of the tertiary structure of this molecule. This is confirmed by analyzing the deduced amino acid sequence, which shows that 33 of 34 cysteine residues are located in conserved positions in both NcMIC4 and TgMIC4 (Fig. 4). Two-dimensional gel electrophoresis showed that the band corresponding to NcMIC4 separates into seven distinct spots with an isoelectric point of pH 6 to 7 (Fig. 3). The pattern of distribution of these spots was regular, as often observed when posttranslational modifications, such as phosphorylation, occur or when different glycosylation patterns are present. MS identified within all of these protein spots distinct peptide fragments of NcMIC4, suggesting that this protein is extensively modified. Indeed, sequence analysis revealed the presence of multiple phosphorylation and glycosylation sites.

In order to be functionally active during events that lead to host cell entry, microneme proteins need to be secreted. We have previously demonstrated that both NcMIC1 (22) and NcMIC3 (27) are secreted from N. caninum tachyzoites upon elevating the temperature of the medium to 37°C. While NcMIC3 appears to remain bound to the surface of tachyzoites, NcMIC1 is secreted as a soluble protein. Similarly, NcMIC4 was shown to be efficiently secreted from most tachyzoites (Fig. 5D) and released into the medium supernatants (Fig. 6). The secreted protein exhibited a slightly lower M_r than the intracellular protein, suggesting that this molecule could be proteolytically processed during secretion, as has been shown, e.g., for TgMIC2 (5) and TgMIC4 (1). However, we cannot rule out the possibility that NcMIC4 could also be proteolytically cleaved at times after secretion through proteases released from rhoptries or dense granules, for example. In addition, other types of modification could be responsible for the slightly altered migration of NcMIC4 following secretion. This aspect needs further investigation. Despite this modification, secreted NcMIC4 retains its lactose-binding capacity; thus, we have been able to purify this molecule from medium supernatants with an efficiency similar to that of the intracellular protein (see Fig. 6).

In contrast to tachyzoites, we found that NcMIC4 expression is largely down-regulated in N. caninum tissue cysts containing bradyzoites, which were generated in mice. The mature tissue cysts were identified through peripheral staining with MAb CC2, which is specific for a cyst wall-associated antigen (12, 38, 39). In contrast, TgMIC4 was reported to be expressed in all invasive stages of the T. gondii life cycle (tachyzoites, bradyzoites, and sporozoites) (1). In vitro-cultured N. caninum bradyzoites still exhibited prominent apical NcMIC4-specific staining, despite distinct MAb CC2 labeling at the periphery of the cyst. This corresponds to earlier findings with NcMIC1 (22) and demonstrates the completely different environmental situation as it occurs during in vitro culture compared to the in vivo formation of N. caninum tissue cysts in mouse brain tissue.

Several N. caninum microneme proteins identified to date possess adhesive domains that could potentially interact with receptors on the surface of target cells, similar to related domains found in vertebrate extracellular matrix proteins. These adhesive motifs include TSP-like domains in NcMIC1 (22), integrin- and TSP type I-like domains in NcMIC2 (25), epidermal growth factor-like domains in NcMIC3 (27, 35), and Apple domains in NcMIC4 (this study). Thus, as for other apicomplexans, N. caninum microneme secretion is believed to confer a more stable association between parasite and host cell and increases the adhesive capacity of N. caninum tachyzoites.

NcMIC4 present in secreted fractions from N. caninum tachyzoites was, therefore, assessed for its capacity to interact with the Vero cell surface. Secreted NcMIC4 was found to coprecipitate with Vero cells at about the same efficiency as has been observed earlier for NcMIC1 (22). However, we did not know whether this interaction was based on a direct physical contact between host cell surface glycosaminoglycans or other receptors and NcMIC4 or whether other components, providing some means of indirect binding or binding enhancement, are involved. However, TgMIC4 was also shown to bind efficiently to host cells in competitive inhibition assays, interacting with different carbohydrates, including galactose, N-acetylglucosamine, and N-acetylgalactosamine (1).

Solid-phase binding assays (Fig. 8) suggest that purified NcMIC4 exhibits a high affinity to CSA. Interestingly, N. caninum tachyzoites have been shown previously to interact with host cells preferentially through chondroitin sulfates, and the same accounts for recombinant NcMIC3, which also binds to chondroitin sulfate residues on the host cell surface (28). Thus, beside NcMIC3, NcMIC4 is now the second N. caninum microneme protein reported to exhibit preferential binding to chondroitin sulfate glycosaminoglycans. However, it is not clear to what extent this reflects the in vivo situation.

These doubts are reasonable, as in the case of T. gondii it has been demonstrated by several groups that microneme proteins act as complexes (26, 33). Several such complexes have been identified and characterized. One of them is composed of the two soluble adhesins TgMIC1 and TgMIC4, which are anchored in the parasite membrane by TgMIC6, a protein containing transmembrane domains (33). TgMIC4 binds to host cells and coprecipitates with TgMIC1, even in the absence of TgMIC6. Although it is conceivable that NcMIC4 would also exert its function in a similar complex, NcMIC1 was never identified within lactose-binding fractions. Clearly, the receptor-ligand interactions in relation to NcMIC4 and its binding to host cells need to be investigated in more detail in the future.