INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Neospora caninum was isolated in 1988 by Dubey and colleagues from a dog suffering from severe neuromuscular disorders (18). Subsequently, extensive investigations (reviewed in references 19 and 29) showed that N. caninum represents the most important causative agent of bovine abortion worldwide. More recently, a second species in the genus Neospora was discovered, Neospora hughesi, which causes equine protozoal myeloencephalitis (34). Although N. caninum and N. hughesi exhibit many similarities to Toxoplasma gondii and the phylogenetic status of the genus Neospora is still a matter of discussion, several studies have now confirmed that N. caninum and N. hughesi represent species distinct from Toxoplasma spp.

As for all members of the phylum Apicomplexa, the invasive stages of N. caninum have acquired an obligatory intracellular lifestyle, without which these parasites could not survive, proliferate, or complete their life cycle. Processes which contribute to host cell adhesion and/or invasion of target cells are therefore of prime importance, and elucidation of respective mechanisms and characterization of the molecules involved could lead to novel means of intervention. One of the hallmarks of both Neospora and Toxoplasma, in contrast to other apicomplexan parasites, is the ability of these parasites to invade a wide range of mammalian cells and tissues; thus, the host cell specificity is very low (17). Many studies have shown that all apicomplexa employ a set of three secretory organelles, namely micronemes, rhoptries, and dense granules, which represent the key players in the invasion process. The sequential release of respective molecules from these secretory organelles is required to achieve and consolidate the physical interaction between the parasite and host cell surface and to ensure entry into the host cell, intracellular survival, and development (reviewed in reference 20).

With Plasmodium, Cryptosporidium, Eimeria, Sarcocystis, and Toxoplasma species it was shown that micronemes are crucially involved in mediating attachment and invasive stages of these parasites for the host cell, since microneme proteins are rapidly secreted right at the onset of establishing contact with the host cell surface. Adhesive modules identified include mucin-like domains of Cryptosporidium parvum, DBL domains in Plasmodium spp. merozoites, Cys-rich regions on microneme proteins of Theileria parva, thrombospondin type I regions, integrin insertion (I)-domains, epidermal growth factor (EGF)-like domains, and Apple domains (for a recent review see reference 42). Microneme secretion has been most extensively studied in T. gondii, where micronemes discharge their contents by fusing with the apical tip of the parasite, thus delivering microneme proteins to the apical surface. The discharge of micronemes was shown to be regulated by cytoplasmic Ca²⁺ (8).

For N. caninum, initial studies had shown earlier that adhesion and invasion represent two distinct processes, since not every N. caninum tachyzoite which adheres and binds to the host cell surface in vitro is also capable of invading its target cell (24). However, adhesion is clearly a prerequisite for successfully achieving host cell invasion. Two distinct, potentially adhesive Neospora microneme proteins have been identified so far. NcMIC2, identified by Lovett et al. (33), is homologous to TgMIC2 from T. gondii and represents a member of the thrombospondin type I family of adhesive proteins (36). The secretion of this protein was shown, as previously described for TgMIC2 (8), to be dependent on the mobilization of intracellular Ca²⁺ stores. The second microneme protein is NcMIC3 (41), a protein homologous to TgMIC3, which is an adhesin capable of binding to both the surface of the host cell and the surface of the parasite (22). Sequencing of the respective cDNA has indicated that NcMIC3 consists of three distinct domains, including (i) an N-terminal signal peptide sequence which could target the protein into the secretory pathway, (ii) two putative membrane-spanning regions, and (iii) a potentially adhesive part consisting of four consecutive EGF-like domains (amino acids [aa] 150 to 328). EGF-like domains are sequences of 30 to 40 aa residues in length and have been found in a more or less conserved form in a large number of animal extracellular matrix proteins and cell surface proteins. The functional significance of EGF domains with respect to their adhesive properties is reflected by the fact that they are present in extracellular domains of membrane-bound proteins or in proteins known to be secreted (4). EGF-like motifs have also been found in Plasmodium spp. Several merozoite surface proteins have been identified which contain one or several EGF-like domains (2, 12, 13, 35, 43). Antibodies directed against the EGF domains of Plasmodium falciparum MSP-1 were demonstrated to inhibit invasion of erythrocytes in vitro (3), and vaccination of mice with an Escherichia coli-produced recombinant protein containing the two EGF-like modules from MSP-1 of Plasmodium yoelii has protected mice against a lethal challenge with the same parasite strain (32). However, these are proteins which are constitutively expressed on the surface in invasive stages, and they are integrated into the surface membrane via a glycosylphosphatidyl inositol (GPI) anchor. Other microneme-associated proteins known to contain one or several EGF-like domains are SCRP in Cryptosporidium, MIC4 in Eimeria, and MIC3, MIC6, MIC7, and MIC8 in T. gondii (38, 42).

In this study, we investigate the subcellular localization of NcMIC3 prior to and at different time points after host cell invasion and show that NcMIC3 is transiently expressed on the surface of N. caninum tachyzoites once the parasite is set free from its host cell. In addition, we present data which point towards its possible role in host cell adhesion.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

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 (Gibco-BRL, Basel, Switzerland) supplemented with 7% fetal calf serum (FCS), 2 mM glutamine, 50 U of penicillin/ml, and 50 µg of streptomycin/ml at 37°C with 5% CO₂ in tissue culture flasks. Cultures were trypsinized at least once a week. N. caninum tachyzoites of the Nc-1 isolate (18) were maintained in Vero cell monolayers (25). Parasites were harvested when they were still intracellular by trypsinization of infected Vero cells, followed by repeated passages through a 25-gauge needle at 4°C, followed by separation on Sephadex-G25 columns as described previously (24).

In vitro secretion assays. Secretion of NcMIC3 from freshly liberated tachyzoites was assessed by resuspending 10⁷ purified parasites/ml in Earle's balanced salt solution (EBSS) (Gibco-BRL) and transferring them to a 37°C water bath for various time points (1 to 45 min). The effects of several components on NcMIC3 secretion were investigated by adding them to the incubation mixture. These reagents included 1 to 10% FCS, 1% ethanol (9), NH₄Cl (20 mM), thapsigargin (1 µM), A23187 (400 nM), ionomycin (400 nM) (8, 10), and cytochalasin D (10 µg/ml). Stock solutions of thapsigargin, A23187, ionomycin, and cytochalasin D were prepared in dimethyl sulfoxide, and control incubations in 1% dimethyl sulfoxide alone were also performed. Subsequently, the parasites were placed on ice for 5 min and centrifuged at 2,000 × g (5 min, 4°C), and the supernatants and pellets were collected. The supernatants were centrifuged again at a higher speed (10,000 × g, 4°C, 30 min), subjected to methanol-chloroform precipitation (44), and processed for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The pellets (containing the parasites) were each resuspended in 50 µl of cold phosphate-buffered saline (PBS). Five microliters of each specimen was placed into 50 µl of fixation solution (3% paraformaldehyde-0.05% glutaraldehyde), and tachyzoites were allowed to settle onto polylysine-coated coverslips for immunofluorescence surface staining (see below). The remaining 45 µl was subjected to methanol-chloroform precipitation and SDS-PAGE as for the supernatant.

Expression and purification of recombinant NcMIC3 (recNcMIC3FL) and NcMIC3-EGF-like domains (recNcMIC3EGF). Two PCR-amplified fragments were cloned in frame into the XhoI/KpnI-digested expression vector pTRC-HisA (Promega, Madison, Wis.), one encoding the full-length protein (recNcMIC3FL), employing primers F3-Sal-I (5'-AGG TCG ACC GTG GCG GGG CGT CCG CTC TCG TG-3') and R6-Kpn-I (5'-CTG GTA CCT CCC CTG TCC CAA AAT TCG AGG TTT ATC G-3') as described by Sonda et al. (41), and one encoding the stretch of four consecutive EGF-like domains (aa 149 to 328; recNcMIC3EGF), using the primers EGF-Bgl2 (5'-GCA GAT CTG CCG AAA GCC TCG TCG-3') and EGF-Hind-3 (5'-GCC TTC TCT TTA CGC ACA TTC GAA GCG-3'). The cloned sequences were expressed in E. coli as polyhistidine (6-His) fusion proteins (41). Bacteria were harvested by centrifugation, and the pellet was solubilized in sample buffer and processed for SDS-PAGE and immunoblotting. For the isolation of recombinant proteins, bacteria were resuspended in 50 mM sodium phosphate-300 mM NaCl (pH 7) and sonicated three times (30 s each) using a Branson 250 sonifier operating at 55W, and recNcMIC3FL and recNcMIC3EGF were purified by nickel-affinity chromatography. Following purification, the preparations were dialyzed against PBS for 48 h at 4°C, centrifuged at 10,000 × g for 30 min at 4°C, and analyzed by SDS-PAGE. Proteins were stored at 80°C prior to use.

Antibodies, SDS-PAGE, and immunoblotting. The polyclonal antibody directed against N. caninum tachyzoites was previously described (25). Monospecific antibodies against recNcMIC3FL, recNcMIC3EGF, and recNcSRS2 (26) were prepared by affinity purification following separation of corresponding E. coli lysates by SDS-PAGE and transfer to nitrocellulose (41). The monoclonal antibodies (MAbs) Ncmab-4, directed against NcSAG1, and MAb 9.5.2, directed against NcMIC2, were prepared as described earlier (1, 39). All antibodies were checked for specificity on N. caninum lysates by immunoblotting, where SDS-PAGE was carried out under reducing (polyclonal antibodies and Ncmab-4) or nonreducing (MAb 9.5.2) conditions. For assessment of secreted proteins in medium supernatants and parasite pellets, corresponding amounts of each fraction (representing 2 × 10⁶ parasites/lane) were separated by SDS-PAGE, and proteins were electrophoretically transferred to nitrocellulose. Blocking of unspecific binding sites was carried out for 4 h at 24°C in Tris-buffered saline (TBS) containing 3% bovine serum albumin (BSA) and 0.3% Tween 20. Polyclonal rabbit hyperimmune serum directed against N. caninum was applied at a dilution of 1:2,000, and affinity-purified antibodies were used at a 1:50 to 1:200 dilution in TBS-0.3% BSA-0.3% Tween 20. MAbs, originating from culture supernatants, were diluted 1:5. The filters were subsequently washed three times in TBS-0.3% Tween, and the bound antibodies were visualized using alkaline phosphatase-conjugated anti-rabbit and anti-mouse immunoglobulin antibodies, respectively (Promega). As a control, nitrocellulose filters containing E. coli and N. caninum lysates were also incubated with the antibody conjugates alone.

Immunofluorescence. For immunofluorescence, N. caninum tachyzoites (10⁷ parasites/ml), freshly fixed in a mixture of 3% paraformaldehyde and 0.05% glutaraldehyde in PBS, were applied to polylysine-coated glass coverslips. After 20 min, the coverslips were rinsed in PBS, and the specimen was placed into blocking buffer solution (PBS-1% BSA-50 mM glycine) for 30 min. For staining of intracellular binding sites, coverslips were placed into methanol and acetone at 20°C for 5 min each prior to blocking. For some experiments, Vero cells were grown on glass coverslips and were infected with purified N. caninum tachyzoites (24). After cultivation for 2 h to 3 days, infected cells were fixed and permeabilized as described above. The coverslips were then rinsed extensively in PBS and were subsequently incubated in blocking buffer. Affinity-purified antibodies directed against recNcMIC3FL and recNcMIC3EGF were applied diluted 1:4 for 30 min, followed by three washes in PBS. The secondary antibody was a goat anti-rabbit-fluorescein isothiocyanate (FITC) conjugate diluted at 1:100. The preparations were then stained either with the anti-N. caninum antiserum (1:500), followed by a goat anti-rabbit-tetramethyl rhodamine isocyanate (TRITC) conjugate (1:100), or nonpermeabilized and permeabilized cells were incubated with MAb 5-1-2 directed against alpha tubulin (41), followed by staining with a goat anti-mouse-TRITC secondary antibody. After extensive washing, specimens were embedded in Fluoroprep (bioMerieux S. A., Geneva, Switzerland) and were viewed on a Leitz Laborlux S fluorescence microscope.

Immunogold transmission electron microscopy (TEM). LR-White embedding and on-section labeling of N. caninum-infected Vero cell cultures, fixed and processed at various time points after infection, were performed essentially as previously described (28). Sections were incubated in affinity-purified antibodies directed against recNcMIC3FL diluted 1:4 in TEM blocking buffer for 1 h. As a negative control, incubations with the preimmune serum of the anti-N. caninum antiserum were performed. After washing in five changes of PBS, 2 min each, the goat anti-rabbit antibody conjugated to 10-nm gold particles (purchased from Amersham, Zurich, Switzerland) was applied (28). Finally, grids were stained with lead citrate and uranyl acetate (27) and were subsequently viewed on a Hitachi H-600 transmission electron microscope operating at 100 kV.

Parasite-Vero cell adhesion assays. In vitro adhesion assays were performed using prefixed Vero cells. For this, Vero cells (5 × 10⁵) were grown at 37°C with 5% CO₂ overnight, either directly in 24-well tissue culture plates or on poly-L-lysine-coated glass coverslips. The coverslips were then washed three times with ice-cold PBS and fixed in 3% paraformaldehyde-0.5% glutaraldehyde in 100 mM phosphate buffer for 1 h at 4°C. Subsequently, free aldehyde groups were blocked by incubation in 120 mM ethanolamine (pH 8) at 4°C overnight (23).

NcMIC3-Vero cell interactions. The binding of NcMIC3 to Vero cells was demonstrated by cosedimentation assays. Freshly split nonadherent Vero cells were fixed in 2.5% glutaraldehyde in EBSS for 30 min, followed by postfixation in 0.5% OsO₄ in 100 mM sodium phosphate buffer, pH 7.2. Subsequently, free aldehyde groups were blocked by incubation in 100 mM ethanolamine (pH 8) at 4°C overnight. Triton X-100 extracts of N. caninum tachyzoites were prepared by incubating 5 × 10⁸ freshly isolated N. caninum tachyzoites in 2 ml of PBS containing 1% Triton X-100 for 5 min at 4°C, followed by centrifugation at 10,000 × g for 30 min at 4°C. Prefixed Vero cells (10⁶) were then incubated in 250 µl of Triton X-100 extracts for 2 h at 4°C. In other experiments, Vero cells were incubated with recNcMIC3FL or recNcMIC3EGF. Subsequently the preparations were centrifuged at 10,000 × g for 5 min, and the supernatants were collected and processed for SDS-PAGE and immunoblotting. The pellets were washed in PBS three times and were finally taken up in SDS-PAGE sample buffer. Equal amounts of nonbound (supernatants) and bound (pellets) proteins were loaded onto gels. Immunoblotting was performed as described above.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Subcellular localization of NcMIC3. It was shown earlier that NcMIC3 is localized in the anterior micronemes of N. caninum tachyzoites (41). In this study, we investigated the subcellular distribution of NcMIC3 at different time points prior to and following host cell invasion by immunofluorescence and immunogold electron microscopy (Fig. 1 and 2). Immunofluorescence staining of both intracellular and extracellular tachyzoites in infected Vero cell cultures fixed at 1 h after addition of the parasites revealed that NcMIC3 was not located exclusively at the anterior tip of the parasite anymore but exhibited a more diffuse distribution, notably dispersed over the entire tachyzoite cell surface (Fig. 1a and b). At 3 h postinvasion (Fig. 1c), tachyzoites exhibited more pronounced NcMIC3-specific staining at the apical tip. The staining was found to be located at the anterior ends of the tachyzoites in inspection of the corresponding DNA staining, where both nuclear DNA and apicoplast-associated DNA (located at the apical part of the parasite) were detected (Fig. 1c). Furthermore, aggregated anti-NcMIC3-immunoreactive material, indicative of components shed by the parasite, could be seen on the surfaces of infected Vero cells (Fig. 1a and c). This material did not represent dead parasite organisms (as judged from the absence of nuclear DNA) but most likely originated from secretory parasite products. At 6 h postinvasion, tachyzoites underwent endodyogeny, and NcMIC3 was then observed to be more strictly localized at the anterior end of the intracellular tachyzoites (Fig. 1d). The anterior NcMIC3 labeling pattern of intracellular parasites persisted and became more pronounced during subsequent rounds of parasite replication (Fig. 1e).

View larger version (55K): [in this window] [in a new window]   FIG. 1.   Immunofluorescence staining of Vero cell cultures fixed and processed 1 h (a and b), 3 h (c), 6 h (d), and 36 h (e) following addition and subsequent culturing of N. caninum tachyzoites. NcMIC3 was detected using affinity-purified anti-recNcMIC3 antibodies followed by FITC-conjugated secondary anti-rabbit antibody. The entire parasite was stained with anti-N. caninum antiserum followed by an anti-rabbit-TRITC conjugate. Note the pronounced apical NcMIC3-specific staining in N. caninum tachyzoites appearing at 3 h postinfection. Arrows point towards NcMIC3-containing material deposited onto the host cell surface.

View larger version (158K): [in this window] [in a new window]   FIG. 2.   TEM immunogold labeling of LR-White embedded N. caninum tachyzoites (a and b) and infected Vero cells (c to e), fixed and processed immediately after purification of tachyzoites at 4°C (a and b), at 1 h following addition of tachyzoites to Vero cells (c), and at 24 and 48 h following host cell infection (d and e, respectively). Arrows in panel c point towards NcMIC3 which is secreted onto the surface of the parasite.

Secretion of NcMIC3. During subsequent studies we found that the localization of NcMIC3 in intracellular tachyzoites was significantly different from the distribution of NcMIC3 in tachyzoites obtained from lysed Vero cell cultures containing already extracellular parasites. A comparative assessment by immunofluorescence was performed. While NcMIC3 labeling of intracellular tachyzoites was located predominantly at the apical tips of the parasites, up to 80% of extracellular tachyzoites harvested from lysed Vero cell cultures exhibited labeling in other subcellular compartments, notably at the surface and/or at the posterior end, as well (data not shown). In order to determine whether this differential labeling pattern was due to secretion of NcMIC3 induced simply through extracellular maintenance of tachyzoites, all subsequent experiments were carried out using purified tachyzoites obtained strictly from cultures which did not exhibit any host cell lysis at all.

View larger version (46K): [in this window] [in a new window]   FIG. 3.   Double immunofluorescence surface staining of isolated N. caninum tachyzoites fixed and processed at the time points indicated (0 to 30 min) following incubation at 37°C in EBSS. NcMIC3 was detected using affinity-purified anti-recNcMIC3 antibodies followed by FITC-conjugated secondary anti-rabbit antibody. The entire parasite was stained with anti-N. caninum antiserum followed by an anti-rabbit-TRITC conjugate. Note the appearance of NcMIC3 on the surfaces of the parasites already after 2 min and its subsequent gradual distribution and clustering at the posterior ends.

View larger version (99K): [in this window] [in a new window]   FIG. 4.   Control experiments confirming the specificity of surface labeling and localization of surface-exposed NcMIC3 in N. caninum tachyzoites. Freshly purified N. caninum tachyzoites were incubated at 37°C for 15 min, followed by fixation in paraformaldehyde-glutaraldehyde (pFA/GA) as indicated in Materials and Methods. They were then labeled with anti-NcSRS2 (top panel) or anti-NcMIC3 antibodies (middle and bottom panels), respectively, followed by FITC conjugate. In the upper two panels, specimens were directly incubated with antitubulin antibodies and a TRITC conjugate; the bottom panel shows antitubulin staining following permeabilization of tachyzoites with methanol and acetone. Note that subpellicular tubulin staining is visible only following permeabilization of parasites and that tubulin staining (indicated by arrowheads), which serves as a marker for the apical part of the tachyzoites, is always located at the opposite end of NcMIC3 surface labeling (arrows).

View larger version (44K): [in this window] [in a new window]   FIG. 5.   Double immunofluorescence of N. caninum tachyzoites following secretion of NcMIC3 in the presence of cytochalasin D. Parasites were incubated at 37°C for 30 min in the presence of 10 µg of cytochalasin D/ml, fixed, and surface labeled using anti-NcMIC3 antibodies and a FITC conjugate. Specimens were then permeabilized and stained with antitubulin antibodies and a TRITC conjugate in order to visualize the apical ends of the parasites. Note that NcMIC3 is readily secreted but remains associated with the apical part.

View larger version (41K): [in this window] [in a new window]   FIG. 6.   NcMIC3 is not released into the medium following secretion. Immunoblotting of tachyzoites (pellets) and corresponding culture supernatants after incubation at 37°C is shown. Western blots were labeled with anti-NcMIC3 antibodies and with anti-N. caninum antiserum, respectively. Note the complete absence of NcMIC3 in the supernatant fractions.

View larger version (41K): [in this window] [in a new window]   FIG. 7.   Immunoblot of culture supernatants of N. caninum tachyzoites obtained after the addition of agents previously shown to stimulate microneme secretion in T. gondii for 10 min at 37°C. A representative immunofluorescence image is shown on the left side. Note the complete absence of NcMIC3 in the supernatants.

NcMIC3 and adhesion to host cells. Further experiments were performed in order to determine whether NcMIC3, and especially its four consecutive EGF-like domains, could be involved in parasite attachment to the host cell. Besides the recombinant NcMIC3 full-length protein recNcMIC3FL, another recombinant protein comprised of the four consecutive repeats fused to an N-terminal polyhistidine stretch was expressed in E. coli (recNcMIC3EGF). Immunoblotting (Fig. 8) showed that antibodies which were affinity purified on recNcMIC3EGF reacted exclusively with recNcMIC3EGF and recNcMIC3FL but not with any other E. coli proteins, and these antibodies labeled a distinct 38-kDa protein in N. caninum tachyzoite extracts. Immunofluorescence surface labeling of tachyzoites which had been previously incubated at 37°C for 10 min revealed a labeling pattern similar to that observed before using antibodies that were affinity purified on the corresponding recombinant full-length protein recNcMIC3FL (Fig. 8). This indicated that following secretion, these potentially adhesive EGF-like domains were surface exposed and accessible from the outside, and thus potentially they could readily interact with putative host cell surface receptors.

View larger version (66K): [in this window] [in a new window]   FIG. 8.   Immunoblot and immunofluorescence labeling obtained with affinity-purified anti-recNcMIC3EGF antibodies. Lane 1, E. coli extract expressing recNcMIC3EGF; lane 2, E. coli extract expressing recNcMIC3FL; lane 3, E. coli extract with the pTRCHis plasmid without insert; lane 4, N. caninum tachyzoite extract.

View larger version (25K): [in this window] [in a new window]   FIG. 9.   Vero cell adhesion assay carried out using N. caninum tachyzoites preincubated at 4°C (first three columns) and 37°C (following three columns). Bound parasites were detected by immunofluorescence. Note the increased adhesive capacity of tachyzoites incubated at 37°C and the inhibition of this effect mediated by the addition of anti-recNcMIC3FL and anti-recNcMIC3EGF antibodies.

View larger version (55K): [in this window] [in a new window]   FIG. 10.   Immunoblots of coprecipitation assays using prefixed Vero cells incubated in the presence of Triton X-100 extracts of N. caninum tachyzoites (panels 1, 2, and 5 to 7) and recNcMIC3EGF (panel 3) and recNcMIC3FL (panel 4). Immunoblots were stained with antibodies directed against NcMIC3, NcMIC2, NcSRS2, and NcSAG1. P indicates the cosedimenting fraction (pellet), and SN indicates the nonbound fraction (supernatant). Note the efficient cosedimentation of NcMIC3 and of recNcMIC3EGF compared to that of NcSRS2 and NcSAG1.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

These results suggested that during extracellular maintenance of the parasite and invasion of host cells, NcMIC3 appears on the tachyzoite surface, and its association with the tachyzoite plasma membrane is largely diminished during the subsequent rounds of intracellular proliferation.

In all apicomplexan parasites, important functions with respect to adhesion and invasion of host cells have been ascribed to microneme proteins (7, 10, 20, 21, 42). In order to carry out these functions, microneme contents are expelled from the apical tip of the invasive zoites, presumably at the time point of initial contact with the host cell or at the onset of extracellular maintenance once the parasite is liberated from the host cell. Microneme proteins are then distributed over the surface of the parasites, where (i) they are involved in tightening the interaction with host cell surface molecules, leading to the formation of a tight junction between the parasite and the host cell surface membrane, and (ii) due to their physical interaction with the pellicular actin/myosin motor proteins (14, 15), they provide the machinery and driving force to actively invade the host cells (42).

The recently described N. caninum microneme protein NcMIC3 had been identified as one of the immunodominant antigens recognized by a complex anti-N. caninum antiserum (41). The deduced polypeptide sequence of NcMIC3 is comprised of an N-terminal signal peptide, followed by two potential transmembrane domains and four consecutive EGF-like domains. Its overall structure and sequence closely resemble those of TgMIC3, the major part of which has been shown to be dominated by five consecutive EGF-like domains, two of which are overlapping (22). The presence of this stretch of four consecutive EGF-like domains has made NcMIC3 a likely Neospora candidate adhesion molecule, a hypothesis for which further evidence has been obtained in this study.

In order to carry out an adhesive function, NcMIC3 must be secreted onto the parasite surface. Previous results (41) and the data provided herein suggest that NcMIC3 is indeed localized within the micronemes in intracellular tachyzoites. Due to the fact that elevating the temperature is sufficient to induce secretion of this protein, it is very likely that NcMIC3 appears on the parasite as soon as tachyzoites are liberated from the host cell. Immunolocalization studies suggest that the distribution of NcMIC3 within the parasite cell is variable and is dependent on the time point of fixation of the specimen with regard to host cell invasion: N. caninum tachyzoites obtained from cultures containing exclusively intracellular parasites (isolated from the cellular interior at low temperatures) confirmed the association of NcMI3 with the micronemes (Fig. 2a and b). In contrast, immunofluorescence staining of Vero cell monolayers which had been fixed and processed 1 to 3 h following their exposure to freshly liberated N. caninum tachyzoites showed that NcMIC3 was not associated strictly with the micronemes anymore but was more diffusely distributed over the entire tachyzoite surface (Fig. 1). As evidenced by immunogold electron microscopy, surface-associated NcMIC3 was transported into the parasitophorous vacuole during invasion of tachyzoites. However, surface staining of these intracellular tachyzoites was largely diminished later at the onset of endodyogeny (6 h postinvasion) and progressively lost later on (see Fig. 1 and 2).

In order to investigate the NcMIC3 secretion process in more detail, a host cell-free system that was previously developed for studying T. gondii secretion (8) was used, where secretion could be monitored under defined conditions. Elevation of the temperature to 20°C or above was sufficient to induce NcMIC3 secretion (Fig. 3). The protein appeared on the surface by emerging at the apical tip, and it was further transported backwards, ending up at the posterior end of the tachyzoites. Retrograde surface movement was sensitive to cytochalasin D, indicating that a functional actin microfilament system is required (Fig. 5). This is in agreement with findings obtained in studies on microneme proteins from other apicomplexan species, including Sarcocystis, Toxoplasma, and Eimeria (reviewed in reference 42). The involvement of the actin microfilament system in this retrograde movement of surface constituents, most likely in conjunction with myosin motor proteins, appears to be one of the hallmarks of apicomplexan cell biology (6, 14, 40). Since it is the actin/myosin system which represents the driving force for host cell invasion, NcMIC3 not only could be involved in adhesion to the host cell but also could play an active role in the host cell entry process.

However, there are distinct differences between NcMIC3 and other microneme proteins identified to date, including the Neospora microneme protein NcMIC2. None of the Ca²⁺ modulators, previously shown to largely influence secretion of microneme proteins in Toxoplasma gondii (5, 8, 11, 16) and N. caninum (33) through elevation of intracellular Ca²⁺ levels, had any visible effect on the secretion kinetics of NcMIC3, most notably not on its appearance on the surface of N. caninum tachyzoites. Furthermore, once secreted and associated with the surface membrane, NcMIC3 appears to remain tightly bound to the parasite cell surface, as we were not able to detect this protein in the respective culture supernatants (Fig. 6 and 7).

It is interesting that the T. gondii homologue of NcMIC3, TgMIC3, exhibits similar properties: TgMIC3 secretion has also recently been shown not to be triggered by Ca²⁺ modulators, since it was not detected in culture supernatants following induction of microneme secretion (5). However, in contrast to NcMIC3, the TgMIC3 polypeptide sequence lacks any N-terminal hydrophobic putative membrane-spanning regions, and blot overlay assays have shown that this molecule can bind to host cells as well as to T. gondii tachzyoites. N-terminal peptide sequencing of TgMIC3 had revealed that this molecule is proteolytically processed downstream of the putative signal peptide cleavage site (22). At the moment we do not know whether the tight interaction between NcMIC3 and the parasite surface is due to the insertion of the two putative N-terminally located transmembrane domains into the lipid bilayer or whether other mechanisms, i.e., direct binding of the protein to the tachyzoite surface as suggested for TgMIC3 (22), are involved. Attempts to purify NcMIC3 by affinity chromatography using cross-reactive MAbs generated against TgMIC3 in order to perform N-terminal peptide sequencing of the native protein had failed so far, and thus no information is currently available with regard to similar proteolytical processing events which could affect NcMIC3. However, in order to determine how NcMIC3 is associated with the tachyzoite surface it will be important to address this question in the future.

Not surprisingly, the appearance of NcMIC3, and especially its EGF-like domains, on the parasite surface upon incubation at 37°C (Fig. 8) coincided with an increased efficiency with regard to adhesion to the surface of Vero cells (Fig. 9): temperature-treated tachyzoites were found to bind more efficiently to host cell monolayers than parasites which had not been incubated at 37°C and which lacked any detectable NcMIC3 on their surface. Although other secretory molecules will also contribute to this effect, we found that the addition of anti-recNcMIC3FL and anti-recNcMIC3EGF antibodies neutralized this increased adhesion efficiency, indicating that specific binding domains could be blocked (Fig. 9). Although not a conclusive proof, this is a strong indication for a functional involvement of NcMIC3 in host cell surface membrane adhesion. However, the actual repertoire of micronemal proteins and other adhesive molecules in N. caninum is far from being known. In T. gondii, at least nine microneme proteins have been identified to date (42), and it is very likely that a multitude of functional proteins are exposed on the surface of apicomplexan parasites upon secretion, proteins which form an adhesion and invasion machinery of high redundancy. This is supported by a recent study on T. gondii (38), which has shown that disruption of MIC1 or MIC6 genes does not result in phenotypes with reduced invasive capacities.

Cosedimentation assays were performed using prefixed Vero cells incubated in the presence of a Triton X-100-soluble fraction of N. caninum tachyzoites which contained, besides other proteins, NcMIC3 (41), NcMIC2 (33), and the two immunodominant surface antigens NcSRS2 and NcSAG1 (26, 30). In these assays, NcMIC3 and NcMIC2 bound to the Vero cell surfaces with a relatively high level of efficiency, while NcSRS2 and NcSAG1 coprecipitated with Vero cells at a much lower rate. Presently it is not known which molecules on the Vero cell surface could be involved in binding of NcMIC3. Further investigations are in progress which are targeted towards elucidating the role of host cell surface glycosaminoglycans, since it has been shown previously that T. gondii tachyzoites utilize sulfated glycosaminoglycans for substrate and host cell attachment (11, 37). Taking into account that Neospora and Toxoplasma are very closely related and considering the high degree of homology of those surface- and secretory organelle-associated molecules identified to date (29, 31), it is likely that these two parasite species employ similar mechanisms for achieving a physical interaction with their host cells.

Besides an increased capacity to adhere to the surface of the host cell, incubation of tachyzoites at 37°C resulted in the outward exposure of EGF-like domains on the surfaces of the parasites. Although we do not know whether these domains actually participate in the adhesion process, this assay shows that they are potentially available for interaction with a putative host cell surface receptor. The fact that recombinant EGF-like domains purified from E. coli lysates did also coprecipitate with Vero cells supports these findings. However, full-length recNcMIC3FL did not bind to Vero cells; thus, further studies are required to determine whether these EGF-like domains do indeed mediate the adhesive properties of this protein in vivo. For the Toxoplasma homologue TgMIC, it is also not clear whether the interaction with the host cell is indeed mediated through these EGF-like domains, since TgMIC3 also binds readily to mammalian cells lacking the EGF receptor (22). In addition, it was shown recently with T. gondii tachyzoites that the third EGF-like domain of TgMIC6 was crucially involved in the accurate sorting of TgMIC1 and TgMIC4 to the micronemes, showing that EGF-like domains carry out important functions related to intracellular protein-protein interactions (38). The correct folding of EGF-like domains could be crucial for retaining their biological activity. Thus, expression of NcMIC3 in a eukaryotic expression system will provide further information in future studies.