The toxoplasma micronemal protein MIC4 is an adhesin composed of six conserved apple domains.

The initial stage of invasion by apicomplexan parasites involves the exocytosis of the micronemes-containing molecules that contribute to host cell attachment and penetration. MIC4 was previously described as a protein secreted by Toxoplasma gondii tachyzoites upon stimulation of micronemes exocytosis. We have microsequenced the mature protein, purified after discharge from micronemes and cloned the corresponding gene. The deduced amino acid sequence of MIC4 predicts a 61-kDa protein that contains 6 conserved apple domains. Apple domains are composed of six spacely conserved cysteine residues which form disulfide bridges and are also present in micronemal proteins from two closely related apicomplexan parasites, Sarcocystis muris and Eimeria species, and several mammalian serum proteins, including kallikrein. Here we show that MIC4 localizes in the micronemes of all the invasive forms of T. gondii, tachyzoites, bradyzoites, sporozoites, and merozoites. The protein is proteolytically processed both at the N and the C terminus only upon release from the organelle. MIC4 binds efficiently to host cells, and the adhesive motif maps in the most C-terminal apple domain.

nonsporulated, partially sporulated, and fully sporulated preparations, were obtained from cats infected with the VEG strain of T. gondii (kindly provided by Dr. Michael White, Montana State University). A clonal isolate of the RH hxgprt Ϫ of T. gondii was used as the recipient strain for all the transfection experiments.
Cloning of MIC4 Genomic Locus and DNA Sequencing and Analysis-The cosmid library used the SuperCos vector modified with SAG1/ble Toxoplasma selection cassette inserted into its HindIII site. The library was prepared from a Sau3AI partial digestion of RH genomic DNA ligated into the BamHI cloning site (kindly provided by D. Howe).
Inserts from TgEST phage clones were amplified using T3 and T7 primers and cloned into pCR2.1 (Invitrogen). DNA sequencing was conducted by cycle-sequencing using ABI Prism Big Dye terminator cycle sequencing reaction kits (ABI, Foster City, CA) and resolved on ABI 377 DNA sequencers. Sequence analysis was conducted with the Genetics Computer Group programs (13), programs available through the National Center for Biotechnology, and programs at the ExPasy site.
Construction of Expression Plasmids-The vector pGEXMIC4A1 was constructed by cloning a PCR product corresponding to the N-terminal region of MIC4 and encompassing the first apple domain A1 into pGEX-4T vector (Amersham Pharmacia Biotech). The sense and antisense primers used for PCR amplification are 5Ј-cgcggatcctggtttggagtggctaaagccc-3Ј and 5Ј-ggttaattaagtggatcccaacacccctcgttccttaa-3Ј, respectively. In parallel, the same A1 fragment was cloned into a pET vector, and the nonfusion protein could also be produced in native soluble form in Escherichia coli BL21.
The expression vector for T. gondii pTMIC4mycHXGPRT was obtained by cloning a PCR product corresponding to the complete coding sequence of MIC4 between the EcoRI and PacI sites of the pTmycHXGPRT vector previously described (14). The primers for PCR were 5Ј-cggaattccctttttcgacaaaatgagagcgtcgctccc-3Ј and 5Ј-ccttaattaaaatgcatcttctgtgtctttcgcttc-3Ј. An additional epitope tag was introduced at the N terminus of MIC4 to generate pTty1MIC4mycHXGPRT. The 11 amino acid Ty-1 tag coding sequence (LEVHTNQDPLD) was inserted as double-stranded oligonulcleotides into the unique PstI site at amino acid 47 of MIC4: 5Ј-gaggtccacacgaaccaggacccgctcgaccatgca-3Ј and 5Ј-tggtcgagcgggtcctggttcgtgtggacctctgca-3Ј (15). A vector expressing MIC4 with a stretch of eight histidine residues at the C terminus was obtained by inserting double-strand oligonucleotides 5Ј-gggcaccaccatcaccaccatcaccattaat-3Ј and 5Ј-taatggtgatggtggtgatggtggtgccctgca-3Ј into the unique PstI and PacI sites of the expression vector pT-HXGPRT. The vector pTMIC4⌬C12 corresponds to a deletion mutant at the C terminus of MIC4 lacking the last 12 amino acids, which was generated by PCR using the antisense primer 5Ј-ccttaattaatcaggatccattgtcacagaaagtatagggtc-3Ј.
Production of Polyclonal Antibodies-The preparation of the glutathione S-transferase (GST) fusion protein was obtained by cloning a fragment encompassing the first apple domain of the predicted MIC4 protein. The DNA sequence coding for the amino acids 19 -205 was cloned into the pGEX-4T vector (Amersham Pharmacia Biotech) for production as a fusion protein in E. coli. Expression of the recombinant MIC4 fragment fused to GST was achieved in the E. coli strain BL21 after a 4-h induction with isopropyl-␤-D-thiogalactopyranoside. The protein was purified under native conditions according to the manufacturer. Rabbit polyclonal sera were made against the purified GST-A1-2 fusion. The initial immunization was performed with 500 g of protein with complete Freund's adjuvant, whereas 300 g of protein in Gerbu adjuvant (LQ) were used for the subsequent boosts.
SDS-Polyacrylamide Gel Electrophoresis and Western Blotting-SDS-PAGE was performed according to Laemmli (16). Freshly released tachyzoites were harvested and washed in PBS. Samples were boiled in SDS sample buffer with (reduced) or without (nonreduced) 144 mM ␤-mercaptoethanol and separated on 8.5 or 10% polyacrylamide gels. Gels were stained with Coomassie or transferred to nitrocellulose membranes for Western blotting and to polyvinylidene difluoride nylon membranes for N-terminal sequencing. Western blots were probed with antibodies to Toxoplasma proteins followed by goat anti-mouse or goat anti-rabbit IgG peroxidase and developed by chemiluminescence using the ECL system (Roche Molecular Biochemicals) or SuperSignal (Pierce). Western blots were quantified by exposure of blots to GS-250 imaging screen CH using a model 363 Molecular Imaging System (Bio-Rad) and analyzed using the Molecular Analyst software.
Selection of Stable Transformants Using HXGPRT as a Selectable Marker-To generate stable transformants, 5 ϫ 10 7 extracellular RHhxgprt Ϫ parasites were transfected and selected as previously described (17), with the following modifications. Parasites were transfected with 80 -100 g of linearized plasmid. Twenty-four hours later, parasites were subjected to mycophenolic acid/xanthine exposure and cloned 3 to 5 days later by limiting dilution in 96-well microtiter plates containing HFF cells in the presence of mycophenolic acid/xanthine. Stable transformants were analyzed for the presence of the recombinant protein by IFA.
IFA-All manipulations were carried out at room temperature. Tachyzoite-infected HFF cells on glass coverslips were fixed with 3% paraformaldehyde, 0.05% glutaraldehyde or 4% paraformaldehyde only for 20 min followed by a 3-min incubation with 0.1 M glycine in PBS. Fixed cells were permeabilized with 0.2% Triton X-100 in PBS for 20 min and blocked in 2% fetal calf serum or BSA in PBS for 20 min. The cells were then stained with the primary antibodies followed by Cy2and Cy3-conjugated goat anti-mouse antibodies (Bio-Rad). Confocal images were collected with a Leica laser-scanning confocal microscope (TCS-NT DM/IRB) using a 100 ϫ Plan-Apo objective with a numerical aperture of 1.30. Single optical sections were recorded with an optimal pinhole of 1.0 (according to Leica instructions) and 16ϫ averaging. All other micrographs were obtained with a Zeiss Axiophot equiped with a camera (Photometrics Type CH-250). Adobe Photoshop (Adobe Systems, Mountain View, CA) was used for image processing.
Electron Microscopy-Thin sections of paraformaldehyde-fixed, LR White-embedded materials were mounted on nickel grids. The grids with sections of enteric forms, tachyzoites, or tissue cysts were floated on drops of 1% BSA in Tris/HCl buffer, pH 7.2, to reduce nonspecific staining followed by the rabbit anti-MIC4 antibodies appropriately diluted in Tris buffer. After washing, the grids were floated on secondary antibody conjugated to either 5-or 10-nm colloidal gold particles. Sections were stained with uranyl acetate before examination in the electron microscope.
Preparation of Micronemal Proteins-For large scale preparation of excretory-secretory antigens (ESA), ϳ5 ϫ 10 9 tachyzoites were resuspended in 1 ml of HHE and stimulated to discharge micronemes by the addition of ethanol to a final concentration of 1.0% and warming to 37°C for 30 min (8). Cells were removed by centrifugation at 2,000 ϫ g, and the supernatant was kept for binding experiments. To purify the contents of micronemes, ϳ5 ϫ 10 9 tachyzoites were harvested in HHE as above and subjected to sonication and cell fractionation as described previously (8). Briefly, parasites were resuspended in cold HHE at ϳ10 9 /ml and sonicated while on ice (3 ϫ 30-s pulses at setting 35 on a BioSonik III microprobe sonicator (Bronwill Scientific, Rochester, NY). After sonication, large cellular debris was removed by centrifugation at 2,000 ϫ g, 10 min, 4°C, and the supernatant was further clarified by spinning at 8,000 ϫ g for 20 min, 4°C. The micronemes were recovered from the supernatant by centrifugation at 30,000 ϫ g for 30 min at 4°C. The 30,000 ϫ g pellet was resuspended in PBS, pH 6.0, containing a mixture of protease inhibitors (1 g/ml E64, 10 g/ml (4-amidinophenyl)methanesulfonyl fluoride (APMSF), 10 g/ml TLCK, 1 g/ml leupeptin) and subjected to three rapid freeze/thaw cycles. The suspension was then sonicated 3 ϫ 15 s using the maximum setting for the microprobe sonicator (550 Sonic Dismembrator, Fisher). The suspension was centrifuged at 100,000 ϫ g for 1 h to remove unbroken micronemes and membranes, and the supernatant containing soluble micronemal proteins was kept for cell binding experiments.
Cell Binding Assays-Confluent monolayers of HFF cells grown in 6-well plates were rinsed in PBS and blocked for 30 min at 12°C with 1% BSA in PBS containing 1 mM CaCl 2 and 0.5 mM MgCl 2 (CM-PBS). Excess BSA was removed by rinsing in CM-PBS, and micronemal proteins (20 g/ml total) were added in a volume of 1 ml of CM-PBS and incubated at 12°C for 1 h. The unbound fraction (referred to as supernatant) was removed, and the monolayers were rinsed four times in cold CM-PBS (referred to as W1, W2, W3, and W4). The cell-bound fraction (CBF) was collected by lysing the monolayer in 1 ml of radioimmune precipitation buffer (50 mM Tris, pH 7.5, 1% Triton X-100, 0.5% sodium deoxycholate, 0.2% SDS, 100 mM NaCl, 5 mM EDTA). Fractions were acetone-precipitated and resuspended in SDS sample buffer containing 2% ␤-mercaptoethanol. Dilution standards of the micronemal protein preparations were loaded in parallel to simulate the contents of 1, 5, and 10% of the total input material. ESA treated with 2-mercaptoethanesulfonic acid (MESNA, Sigma) was preincubated at 37°C for 30 min in the presence of 50 mM MESNA and then diluted 50 times in CM-PBS for the binding assay.

Characterization of Cellular and Secreted Forms of MIC4
and N-terminal Microsequencing of the Secreted Form-To characterize secretory proteins of Toxoplasma, we generated mAbs to the ESA fraction released by extracellular tachyzoites. When the mAb 5B1 was used to probe Western blots, it recognized a sharp band that migrated at 72 kDa in tachyzoite cell lysates and 70 kDa in ESA resolved under reducing conditions. Both bands migrated more rapidly and diffusely in the absence of reduction, suggesting the presence of internal disulfide bonds (Fig. 1A). To identify the gene corresponding to the protein recognized by mAb 5B1, we isolated ESA on a large scale and resolved the proteins by SDS-PAGE. Parallel strips were transferred to nitrocellulose for Western blotting to identify the band recognized by mAb 5B1 (Fig. 1B) and to polyvinylidene difluoride membranes for microsequencing. N-terminal sequencing of the band corresponding to the 72-kDa form yielded a partially degenerate sequence of 12 residues (X(G/S)-E(P/N)(D/A)(K/P)LDLA(P/L)V).
Identification and Sequencing of the MIC4 Gene-Comparison of the N-terminal sequence against the Toxoplasma dbEST data base using BLAST identified one hit that matched at 9 of 12 residues to the clone TgESTzy06c08.r1. The sequence of this clone was used to identify other overlapping ESTs. A radioactive probe derived from the TgESTzy26b09.r1 clone was used to screen a cosmid library, and an extended genomic sequence of the locus was determined by primer walking across 5000 base pairs on a positive clone (GenBank TM accession number AF143487). The sequence of the gene codes for a protein of 580 amino acids with a predicted mass of 61 kDa. The short hydrophobic stretch that follows the start codon has the hallmark of a putative signal peptide (Fig. 1C). Hydropathy analysis indicates no other hydrophobic stretch on the protein.
Comparison of the complete coding sequence against the nonredundant GenBank TM data base using BLAST revealed that the gene was homologous to several micronemal antigens previously described from S. muris (18,19) and to the recently reported E. tenella micronemal protein EtMIC5 (9). These proteins share the feature of containing conserved, cysteine-rich domains known as apple motifs, which were detected using Prosite (ExPasy). Such a domain contains six half-cystine residues at highly conserved positions that form a structure resembling an apple (20,21). The consensus for the internal four cysteine residues of this sequence is CX 3 CX 5 CX 11 C. The six apple domains of MIC4 are arranged as follows A1 (amino acids 67-139), A2 (amino acids 140 -230), A3 (amino acids 231-303), A4 (amino acids 304 -417), A5 (amino acids 418 -490), and A6 (amino acids 491-580) (for alignment of the six apple domains, see Ref. 9). In the case of human plasma prekallikrein, it has been shown that three highly conserved disulfide bonds are linking the first and sixth, second and fifth, and third and fourth half-cystine residues in each domain (14). Since these cysteine residues are conserved in MIC4, it is likely that disulfide bond formation in MIC4 is similar to prekallikrein. The sequence analysis of MIC4 predicts a signal peptide cleavage site between residues Ala 25 and His 26 (ExPasy). Although we have not been able to verify the N-terminal sequence of the 72-kDa form of the protein found in cells, it likely corresponds to the mature N terminus generated within the secretory pathway, with removal of an additional 32 residues occurring at the time of secretion into the medium.
Features of MIC4 Gene-The 1743-base pair open reading frame of MIC4 contains no introns. The putative transcription start site of MIC4 was determined by sequencing several clones obtained by 5Ј rapid amplification of cDNA ends PCR. A sequence analysis of the promoter region revealed the lack of TATA box and no element resembling an initiator element (Inr) (22). However, a consensus sequence (heptamer motif) found multiple times in the 5Ј-flanking sequences of several T. gondii genes (23) is present in the promoter region of MIC4 (Fig. 1C). Two heptamer motifs are positioned at Ϫ716 (AGAGACG) and Ϫ496 (TGAGACG) from the transcription start site. These elements have been previously mapped and shown to be critical for transcription of the family of GRA genes (23) and are also included in the 27-base pair repeat element of SAG1 gene (24). A single in-frame ATG lies 58 residues upstream of the Nterminal sequence that was obtained from the purified protein. This ATG probably serves as the translational initiation codon based on the facts that 1) it is the first in-frame ATG, 2) the six nucleotides preceding the ATG (CACAAA) are consistent with the consensus sequence for translational initiation in T. gondii (GNCAAA) (25), and 3) the ATG immediately precedes a sequence predicted to encode a hydrophobic signal peptide. Northern blot and Southern blot analyses confirmed that MIC4 is present as a single copy gene in T. gondii genome that produces a single transcript of the expected size in tachyzoites (data not shown).
Subcellular Localization and Pattern of Expression of MIC4 -We previously reported that the antigen recognized by mAb 5B1 is secreted from Toxoplasma in a manner consistent with it originating from micronemes (8,26). To confirm that the gene described here corresponds to a micronemal protein, we produced a bacterial recombinant GST-A1 fusion of the Nterminal 186 amino acids of MIC4 encompassing the first apple domain (A1) and raised polyclonal antibodies against the purified protein.
The rabbit antisera obtained were tested on immunoblots loaded with the recombinant nonfusion A1 and GST-A1 fusion expressed in E. coli, T. gondii tachyzoites, and Vero cell lysates ( Fig. 2A). The sera recognized specifically a 72-kDa protein in tachyzoites. Neither anti-GST antibodies nor the preimmune rabbit sera reacted with T. gondii proteins on Western blots (data not shown).
To determine the pattern of expression of MIC4 in the different life stages of the parasite, cell lysates of tachyzoites, bradyzoites, and oocysts were resolved by SDS-PAGE and probed by Western blotting (Fig. 2B). MIC4 was not detected in unsporulated or partially sporulated oocysts but was present at approximately equal levels in fully sporulated oocysts (sporozoites), bradyzoites, and tachyzoites.
As anticipated, the mAb 5B1 specifically recognized micronemes by indirect immunofluorescence and immunoelectron microscopy (data not shown). IFA studies with rabbit polyclonal antibodies raised against GST-A1 confirmed a typical staining pattern for proteins located in the apical microneme organelles of the parasites (punctate fluorescence pattern at the apical pole).
MIC4 colocalized perfectly with the other micronemal protein MIC2 (27) (Fig. 2B). IFA performed on extracellular parasites only stained parasites permeabilized with Triton X-100 before incubation with the first antibody (data not shown). This observation suggests that MIC4 is predominantly localized in the micronemes and absent from the cell surface. Ultrastructural examination confirmed that the polyclonal antibodies recognized a protein located within the micronemes of tachyzoites (Fig. 2C,  a) and bradyzoites (Fig. 2C, b). The few micronemes in the merozoite of mature schizonts were also labeled (Fig. 2C, c). In keeping with the nomenclature of previously established Toxoplasma proteins (27), we named this antigen MIC4.
MIC4 Is Proteolytically Cleaved Only after Release by the Micronemes-The T. gondii micronemal proteins characterized so far are subjected to extensive proteolytic remodeling during their transport and/or secretion (27). Comparison of the amino acids sequence deduced from the MIC4 gene with the information obtained from N-terminal sequencing of the mature pro- tein is indicative of proteolytic cleavage. To determine whether MIC4 is proteolytically cleaved during its transport to the micronemes, we generated recombinant parasites expressing MIC4 tagged at both ends with epitopes. The construct pTMIC4mycHXGPRT produced MIC4 with epitope tags at the N terminus and/or the C terminus. An additional Ty-1 epitope tag was introduced 10 amino acids upstream of the cleavage site mapped previously by the N-terminal sequencing of secreted form to generate pTty1MIC4mycHXGPRT. Both constructs were stably integrated into T. gondii tachyzoites, and expression of MIC4myc or Ty-1MIC4myc was examined by Western blot and by IFA. The mAbs anti-Myc and anti-Ty-1 recognized the 72-kDa form of MIC4 on Western blot (Fig. 3A) and gave a typical microneme staining on IFA (Fig. 3B). These results demonstrate that the form of MIC4 stored in the micronemes is not proteolytically cleaved beside the cotranslational removal of the signal peptide and, as for MIC2, the processing on MIC4 occurred uniquely post-exocytosis.
MIC4 Is Processed at Both Ends after Release by the Micronemes, and MPP2 Is Likely to Be Responsible for the Cterminal Cleavage-Immunoblot analysis of tachyzoite lysates and ESA material using mAb 5B1 and the rabbit polyclonal antisera revealed the existence of an additional proteolytic processing at the C terminus that generates the two products of ϳ50 and 15 kDa (Fig. 3C, lanes 1 and 2, upper and lower  panels). The 72-kDa precursor form of MIC4 is present in the micronemes, whereas the processed forms are uniquely detectable in ESA. The polyclonal antibodies were raised against the apple domains A1 and A2 and recognized the 50-kDa form only, whereas the 15-kDa product was detected exclusively by the mAb 5B1. Together, these results document that processing occurs at the surface of the parasite only after release by the micronemes and allowed the mapping of the epitope recognized by the mAb 5B1 within the A6 at the C terminus of MIC4. Recently, distinct protease activities for MIC2 have been described using a variety of protease inhibitors (28). Upon release from the micronemes, MIC2 is proteolytically modified at multiple sites by two distinct enzymes, microneme protein protease 1 (MPP1) and microneme protein protease 2 (MPP2), which probably operate on the parasite surface (28). A subset of serine and cysteine protease inhibitors was shown to block MPP2 activity. Similarly, examination of MIC4 in ESAs from parasites pretreated with various protease inhibitors revealed that the processing of MIC4 into 50-and 15-kDa species was blocked by chymostatin, N-acetyl-L-leucinyl-L-leucinyl-L-norleucinal, and N-acetyl-L-leucinyl-L-leucinyl-methioninal but not any of the other protease inhibitors tested (Fig. 3C). The profile of sensitivity to protease inhibitors strongly suggests that MPP2 processes both MIC2 and MIC4.
MIC4 Binds to Host Cells-Several previous reports indicate that micronemal proteins bind to host cells and may participate in parasite/cell attachment (6,7,28,29). To determine whether MIC4 binds to host cells, we incubated HFF monolayers with mixtures of micronemal proteins isolated from intact cells or from ESAs. Binding assays were conducted at 12°C to prevent internalization by endocytosis as previously described (28). After incubation, monolayers were washed, and the cell-bound fraction (CBF) was obtained by detergent lysis. MIC4 present in ESA (70 kDa) or in parasite lysates (72 kDa) bound tightly to HFF cells and was recovered in the CBF (Fig. 4A). Comparison of the input fraction with the recovered material by phosphorimaging analysis revealed that 11% of MIC4 in the microneme preparation and 26.7% of MIC4 in ESA preparations remained bound to the cell surface (Fig. 4B). Collectively, these data indicate that MIC4 binds substantially to the surface of human fibroblasts. As control, GRA1, which is abundantly present in the ESA, does not bind detectably to host cells. The processed forms of MIC4 were examined separately for their ability to bind to host cells using the mAb 5B1 after resolution on 15% SDS-PAGE or the rabbit antiserum (Fig. 4A, right panel). Interestingly, the 15-kDa form but not the 50-kDa form of MIC4 bound to host cells. The mAb 5B1 recognizing the 15-kDa FIG. 4. MIC4 binds to host cells, and the adhesive motif is restricted to the C terminus encompassing the A6 domain. A, Western blot analysis of cell binding. The micronemal protein MIC4 bound specifically to host cells, whereas binding of the dense granule protein GRA1 was not detected. After incubation with total micronemal proteins (Microneme prep.) or ESA, the unbound fraction was removed (Sup), and cells were washed (W1 and W4 correspond to the first and fourth wash, respectively). The CBF was recovered by detergent extraction, and samples were resolved by SDS-PAGE and Western-blotted for MIC4 (mAb 5B1) and GRA1 (Tg17-43). Standards correspond to 10, 5, and 1% of the input material from the microneme prep. The loading of the SUP represents 5%, whereas all other fractions represent 20% of the total material recovered.
To determine the host cell binding activity of the processed forms of MIC4, the CBF was analyzed on a 15% SDS-PAGE using the rabbit anti-MIC4 or the mAb-5B1. The 70-and 15-kDa forms bound to host cells, whereas no binding was detected with 50-kDa form. B, quantification of the binding of MIC4 to host cells. Approximately 11% of MIC4 in the microneme preparation and 26% of MIC4 in ESA was bound to the host cell (*). In contrast, GRA1 was not detected in the CBF. Values are plotted as relative intensity as determined by phosphorimage analysis and compared with loading standards for 1, 5, and 10% of the starting material. C, ESA prepared from wild type RH and TyMIC4Myc were analyzed in the cell binding assay. The ESA from RH was preincubated for 30 min at 37°C in absence or presence of 50 mM strong reducing agent MESNA. The treated ESA was then diluted 50 times in CM-PBS before incubation with host cells. D, ESAs prepared from RH and parasites expressing MIC4His, TyMIC4Myc, or two independent clones of MIC4⌬C12 were tested and compared in host cell binding assays. The wild type and mutated forms of MIC4 are indicated by an arrow. The rabbit serum anti-MIC4 used in this experiment showed a cross-reaction with host cells, indicated by an asterisk. This signal was also detectable in the sample of HFF cells, which was not incubated in presence of ESA product inhibits about 50% of cell binding by the 70-kDa form. 2 From these results, we concluded that the 15-kDa C-terminal product, which corresponds approximately to the A6 domain, carries the adhesive properties of MIC4.
As mentioned above, the apple structure is maintained by the formation of three disulfide bridges. To test whether these disulfide bridges are necessary for MIC4 binding to host cells we pretreated ESA with the strong reducing agent MESNA. This treated completely abolished MIC4 binding to host cells, providing additional evidence that the adhesive properties of MIC4 are dependent on the presence of intact cystine residues (Fig. 4C). MESNA is not acting on a host cell receptor since preincubation of the ESA with MESNA at room temperature or the addition of the reducing agent to host cells during the binding assay did not impair MIC4 adhesiveness (data not shown). The generation of recombinant parasites expressing a truncated form of MIC4 confirmed these observations. We compared the binding activity of MIC4 with Ty-1MIC4Myc, MIC4his carrying a stretch of eight histidine residues at the C terminus, and a deletion mutant of MIC4 lacking the last 12 amino acids (MIC4⌬C12). MIC4his and MIC4⌬C12 were expressed in a clone of mic4ko mutant parasites lacking MIC4 gene that had been deleted by double homologous recombination. 3 Western blot and IFA analysis of the transformed parasites confirmed that MIC4 mutants were of the expected size and appropriately targeted to the micronemes (data not shown). ESAs prepared from these parasites were tested in host cell binding assays. As for the endogenous MIC4, the TyMIC4myc and MIC4his proteins bound substantially to host cells. In contrast, ESA corresponding to MIC4⌬C12 failed to bind to host cells (Fig. 4D). Two additional deletion mutants of MIC4 with deletion of apple domains A5-6 or A3-6 were generated and failed to bind to host cells (data not shown), consistent with the 50 kDa protein lacking adhesive properties. Truncation of the last 12 C-terminal amino acids abrogated the adhesiveness of MIC4, possibly by compromising the proper folding of the domain A6.
The Nature of the Interaction between MIC4 and the Host Cells-The MIC4-related major micronemal protein of S. muris contains two apple domains and has been shown previously to function as a dimeric lectin with high affinity for galactose (10). To test if the A5-6 domains of MIC4 exhibit similar properties, we have undertaken host cell binding assays in the presence of increasing concentrations of galactose, N-acetylgalactosamine, or N-acetylglucosamine. These competition experiments revealed that the interaction of MIC4 with host cells could be enhanced in the presence of a low concentration of carbohydrates (1 mg/ml), whereas high concentrations (50 mg/ml) are inhibitory (Fig. 5). The carbohydrates tested here showed similar competitive effect, with galactose slightly more potent compared with the others. Our findings on MIC4 structure, processing, and adhesive activities are summarized in the Fig. 6.

DISCUSSION
Micronemal proteins are thought to be critical ligands determining host cell specificity at the time of invasion. Recent studies provide strong evidence that the transmembrane micronemal proteins of the TRAP family contribute not only to attachment but also to gliding motility and, thus, actively participate in the invasion process (2,31). To ensure delivery of ligands at the right time and optimal place, micronemes exocytose adhesins and other factors in a regulated fashion onto the parasite surface during an early phase of invasion (8). This apical secretion is sensitive to the kinase inhibitor staurosporine and can be stimulated by calcium ionophore or ethanol treatment (8,26). These characteristics were used to explore the content of micronemes and to develop a strategy for the identification of novel micronemal proteins and cloning of their corresponding gene (32).
We present here the identification and characterization of a novel micronemal protein identified by this approach. The gene corresponding to MIC4 revealed the existence of a distinct type of adhesive motif called an "apple domain." MIC4 contains six apple domains and shows a high degree of homology with a small major micronemal protein of S. muris containing two apple domains (18,19) and the much larger micronemal protein from E. tenella EtMIC5 (9) with 11 apple domains. The S. muris protein is proteolytically processed and released at the apical tip of invading merozoites (33). This protein, called SML, was shown to form noncovalent homodimers and to recognize N-acetylgalactosamine as the dominant sugar (10). Apple do- mains have been described previously on plasma proteins such as factor XI and prekallikrein (20,34) and is composed of six half-cystine residues at highly conserved positions. Several studies show that apple domains are implicated in specific interactions between factors of the blood coagulation cascade (35,36). A single apple domain can exhibit a very specific affinity, as illustrated by the interaction of the third domain A3 of activated factor XI with factor IX (34).
MIC4 has a calculated molecular mass of 61 kDa, and the deduced amino acid sequence from the gene predicts the presence of a signal peptide and six apple domains. We showed that the protein is localized to the micronemes of all infective stages of the parasite. One surprising incidental finding was that the polyclonal anti-MIC4 stained a sub-population of dense granules (wall-forming bodies, type 1) in the macrogametocyte and the outer veil of the early oocyst in the cat intestine. However, it was not possible to identify the molecule recognized as MIC4 or a closely related MIC4-like protein (37).
MIC4 is synthesized and stored in the parasites as a fulllength 72-kDa form. Upon discharge from the micronemes, MIC4 is rapidly cleaved at the N terminus to produce a 70-kDa form and less efficiently at the C terminus. The C-terminal cleavage of the 70-kDa species into 50-and 15-kDa products causes a gap in size, as the processed forms do not add up to form 70 kDa. We can not exclude additional cleavage events, but the most likely explanation is inaccuracies in the size estimates or a change in conformation that effects migration. The C-terminal cleavage probably results from the protease activity of MPP2, which mediates the N-terminal processing of MIC2 at the surface of the parasite (28). In the case of MIC2, processing at the C terminus by another protease (MPP1) released the protein from the surface of the parasites and alters drastically the adhesive properties of MIC2. MIC4 binds efficiently to host cells, and the analysis of the diverse processed forms revealed that the adhesive properties of the molecule are confined within the apple domain at the C terminus. In contrast to MIC2, cleavage of MIC4 does not appear to influence the binding properties of MIC4. The 72-kDa precursor as well as 70 and 15 kDa processed forms of MIC4 bind to host cells, and therefore, the biological significance of MIC4 processing is not clear yet. The fact that the 50-kDa form of MIC4 failed to bind to host cells suggested that the adhesive properties of MIC4 are confined strictly to the last 15-kDa form of the protein, which corresponds to the domain A6. A deletion of 12 amino acids at the C terminus of MIC4 confirmed the importance of this region of the molecule for binding. In addition, a pretreatment of ESA with the reducing agent MESNA at 37°C (but not at room temperature) abbrogates completely MIC4 binding, suggesting that an intact apple structure hold by disulfide bridges is prerequisite for adhesion.
In T. gondii, several studies point to a crucial role of sugarbinding proteins in host cell recognition. The glycoprotein, BSA-glucosamide, competitively blocks infection of human fibroblasts by tachyzoites and depends on the presence of the major surface antigen SAG1 (38). Incubation of tachyzoites in the presence of gold-labeled albumin-N-acetyl-D-glucosamine or albumin-galactose but not in the presence of albumin-mannose led to labeling of the rhoptries in a pattern similar to that observed with the lectins (39). More recent studies suggest that host recognition by T. gondii is mediated by parasite lectins (40) and that sulfated proteoglycans are one determinant used for substrate and cell recognition by MIC2 (30). In competition experiments, MIC4 binding to host cells in the presence of increasing concentrations of carbohydrates showed a diphasic effect. Host cell binding was enhanced at lower concentrations of competitors, but minimal binding was observed at higher doses. The relatively high doses imposed to induce competition suggest that the specificity of the lectin has not yet been identified, and possibly multivalent or more complex carbohydrate structures are involved. A previous study reported the identification of 45, 65, and 71 kDa lectins in T. gondii tachyzoites (40). The association of the 65-and 71-kDa proteins with host cells was abolished in presence of fucoidan, and a biphasic effect was reported in the competition experiments, similar to our observations. MIC4 binding to host cells in the presence of increasing amounts of fucoidan showed no competition, which likely rules out that MIC4 corresponds to the 71-kDa protein described by Ortega-Barria and Boothroyd (40). Alternatively, like the apple domain studied in the coagulation factors, a yet unknown very specific protein-protein interaction might be responsible for binding.
Since MIC4 lacks a transmembrane or lipid anchor, it likely contributes to parasite adhesion by acting as a bridge between a receptor on the parasite and a receptor on the host cell. Indeed, a recent study has revealed that MIC4 forms a complex with two other micronemal proteins, MIC1 and MIC6. 3 MIC6 is a transmembrane protein that functions as a cargo receptor and ensures proper sorting of MIC1 and MIC4 to the micronemes. MIC6 also likely retains these soluble adhesins at the surface of the parasite during invasion. In this complex, MIC1 is directly and stably associated to MIC4, since the two proteins coimmunoprecipitate even in absence of MIC6. The region of MIC4 interacting with MIC1 is currently being investigated. MIC1, like MIC4, has previously been shown to bind to host cells (6) and does it also in the absence of MIC4 in mutant mic4ko. 4 Therefore MIC1 association with MIC4 represents an interfering parameter in our competition studies that might explain the high doses of galactose necessary to abolish completely host cell binding. For further studies, the nature of the interaction between MIC4 and host cells will have to be examined in the absence of MIC1.
As also observed for other types of micronemal proteins, structural homologues of MIC4 exist in other Apicomplexa (1). The major micronemal antigen of S. muris contains two domains, whereas EtMIC5 exhibits 11 apple domains (Gen-Bank TM accession number AJ245536). A search through the current status of the genome sequencing project of Plasmodium falciparum and the ESTs available for Plasmodium vivax and Plasmodium berghei failed to reveal the presence of a homologue in these members of Apicomplexa. Intriguingly S. muris, T. gondii, and Eimeria infect their hosts via the digestive tract. In contrast, Plasmodium species that are transmitted by an insect vector enter their mammalian host directly by injection into the blood. The existence of several distinct types of adhesins in T. gondii, including MIC1, MIC2, MIC3, and now MIC4, illustrates the diversity of strategies used by the parasite to establish interactions with the host. This diversity confers either a functional redundancy or might accommodate the broad range of host cell type specificity. It will be interesting to examine the possible role of MIC4 in the context of tissue specificity and to determine the nature of the receptor on host cells.