ABSTRACT
Cryptosporidium parvum is a waterborne enteric coccidian that causes diarrheal disease in a wide range of hosts. Development of successful therapies is hampered by the inability to culture the parasite and the lack of a transfection system for genetic manipulation. The glycoprotein products of the Cpgp40/15 gene, gp40 and gp15, are involved in C. parvum sporozoite attachment to and invasion of host cells and, as such, may be good targets for anticryptosporidial therapies. However, the function of these antigens appears to be dependent on the presence of multiple O-linked α-N-acetylgalactosamine (α-GalNAc) determinants. A eukaryotic expression system that would produce proteins bearing glycosylation patterns similar to those found on the native C. parvum glycoproteins would greatly facilitate the molecular and functional characterization of these antigens. As a unique approach to this problem, the Cpgp40/15 gene was transiently expressed in Toxoplasma gondii, and the expressed recombinant glycoproteins were characterized. Antisera to gp40 and gp15 reacted with the surface membranes of tachyzoites expressing the Cpgp40/15 construct, and this reactivity colocalized with that of antiserum to the T. gondii surface protein SAG1. Surface membrane localization was dependent on the presence of the glycophosphatidylinositol anchor attachment site present in the gp15 coding sequence. The presence of terminal O-linked α-GalNAc determinants on the T. gondii recombinant gp40 was confirmed by reactivity with Helix pomatia lectin and the monoclonal antibody 4E9, which recognizes α-GalNAc residues, and digestion with α-N-acetylgalactosaminidase. In addition to appropriate localization and glycosylation, T. gondii apparently processes the gp40/15 precursor into the gp40 and gp15 component glycopolypeptides, albeit inefficiently. These results suggest that a surrogate system using T. gondii for the study of Cryptosporidium biology may be useful.
Cryptosporidium parvum, a ubiquitous intestinal apicomplexan parasite, is the etiological agent of many waterborne outbreaks of diarrheal disease around the world (5). In immunocompetent hosts, this parasite causes a self-limiting diarrheal illness, but in hosts with compromised immune systems, cryptosporidiosis can develop into a severe, chronic, and life-threatening infection (14). The paucity of data concerning Cryptosporidium pathogenesis hampers the development of effective vaccines and treatments. Identification of virulence factors and understanding the natural biology of C. parvum are especially difficult because of the lack of an in vitro culture system for propagation of the parasite and, consequently, the inability to manipulate the genome by transfection.
C. parvum completes its whole life cycle within the host gut epithelium (30). When oocysts present in the environment are ingested, sporozoites are released in the small intestine, where they attach to and invade intestinal epithelial cells. Parasite replication occurs in a unique intracellular location that is beneath the host cell microvillous membrane but is segregated from the cytoplasm. After two rounds of merogony, sexual stages are produced that fuse and form oocysts that are either released into the environment or excyst and reinitiate infection. C. parvum isolates that are classified as genotype II have an unrestricted host range, commonly infecting neonatal ruminants and humans, whereas genotype I isolates are restricted to human hosts (22). Despite an apparent lack of reservoir hosts, most human infections are caused by genotype I isolates (7).
One approach to the development of anticryptosporidial agents has been to identify sporozoite and merozoite surface antigens involved in recognition, attachment, and invasion of the host epithelial cells in order to block these interactions. The glycoprotein products of the Cpgp40/15 gene are two of several antigens that are implicated in these processes (1, 2, 23, 29, 34). gp40 (also referred to as gp45 [29] and S45 [34]) and gp15 (also referred to as Cp17 [23] and S16 [34]) are produced as a single preprotein, proteolytically processed and localized to the surface of zoite stages, from where they are shed in trails during sporozoite locomotion. The soluble gp40 adhesin appears to attach to the sporozoite surface membrane through association with the glycophosphatidylinositol (GPI)-anchored gp15 (34; O'Connor and Ward, unpublished data). Antibodies to gp40 block infection in vitro, and partially purified gp40 binds to intestinal epithelial cells in a manner suggestive of a specific receptor-ligand interaction (2). gp40 and gp15 display O-linked α-N-acetylgalactosamine (αGalNAc) determinants (1, 9, 34). The importance of these residues in attachment and invasion is highlighted by the observation that lectins that recognize these determinants block sporozoite attachment (1) and completely and irreversibly eliminate sporozoite infectivity for intestinal epithelial Caco-2 cells (10). Underscoring the importance of these antigens to Cryptosporidium infection is the observation that the Cpgp40/15 gene exhibits extensive polymorphism among genotype I isolates (19, 29). Although the functional significance of this variation is unknown, these data suggest that the antigens may be important targets of anticryptosporidial immunity.
Since these data support the hypothesis that gp40 and gp15 are integral to the establishment of C. parvum infection, further molecular characterization of the antigens is warranted. However, because of the difficulty of obtaining large quantities of native antigen, particularly from type I isolates, an appropriate eukaryotic expression system that could generate recombinant glycoproteins that would mimic the function of the native antigens is needed. To address this problem, we explored the possibility of expressing these antigens in Toxoplasma gondii. T. gondii is a closely related apicomplexan that is easily propagated and genetically manipulated and has been shown to produce glycoproteins that display glycans similar to those observed on the gp40 and gp15 antigens (35). If a system to successfully exploit T. gondii for heterologous expression of the Cpgp40/15 gene could be developed, then not only could localization, processing, and identification of adhesive domains be investigated, but the potential functional significance of Cpgp40/15 diversity could also be explored.
In this paper, we report the expression of the Cpgp40/15 gene in T. gondii and glycosylation, partial processing, and localization of its protein products to the tachyzoite surface. These experiments suggest that a heterologous expression system using T. gondii for the study of C. parvum glycoproteins and their role in host-parasite interactions may prove useful.
MATERIALS AND METHODS
Parasites.Iowa isolate (genotype II) C. parvum oocysts were obtained from Pleasant Hill Farm, Troy, Idaho, and GCH1 oocysts (genotype II) were obtained from Saul Tzipori, Tufts University School of Veterinary Medicine, Grafton, Mass. C. parvum lysates were prepared in phosphate-buffered saline, pH 7, containing 1% octyl glucoside and protease inhibitors [10 μM trans-epoxysuccinyl-l-leucylamido-(4-guanidino)butane, 20 μM leupeptin, and 2 mM phenylmethylsulfonyl fluoride], as previously described (1). T. gondii RH strain tachyzoites were maintained by passage through human foreskin fibroblast (HFF) cells.
Antibodies and lectins.4E9 is a mouse immunoglobulin M (IgM) monoclonal antibody (MAb) that identifies an αGalNAc-containing epitope present on C. parvum sporozoite antigens gp40 and gp900 (1). C. parvum GCH1 gp40 and gp15 sequences were cloned into the pET 32 Xa/LIC vector (Novagen, Madison, Wis.) and overexpressed in E. coli AD494 as described previously (2). Recombinant fusion proteins (rgp40 and rgp15) were purified by metal-ion affinity chromatography by using His bind (Novagen) or Talon (Clontech, Palo Alto, Calif.) kits according to the manufacturer's directions. Purified recombinant fusion proteins were excised from sodium dodecyl sulfate (SDS)-polyacrylamide gels and used to immunize BALB/c mice. Polyclonal antisera to recombinant gp40 and gp15 (anti-rgp40 and anti-rgp15) obtained from these mice reacted specifically with the cognate native antigen from C. parvum lysates upon Western blot analysis (K. Rogers, R. M. O'Connor, A. M. Cevallos, and H. D. Ward, unpublished data). The GalNAc-specific lectin Helix pomatia agglutinin (HPA) was obtained from EY Laboratories (San Mateo, Calif.). Anti-p30 (SAG1) polyclonal rabbit antiserum and anti-GRA3 MAb T63H11 were obtained from Lloyd Kasper, Dartmouth University, and Jean Francois Dubremetz, Institut de Biologie de Lille, Lille, France, respectively. Anti-MIC2 MAb 6D10 and anti-ROP1 MAb Tg49 were obtained from David Sibley, Washington Univeristy, and Joe Schwartzman, Dartmouth University, respectively.
Plasmids.All C. parvum proteins were expressed from the T. gondii expression vector pGRA1PKR1-N177GFPGRA2. This vector is a modified version of pGRA1GFP5S65TGRA2 (18) with an NcoI site encompassing the ATG initiator methionine and 177 amino acids of the T. gondii PKR1 (J. Tang and K. Kim, unpublished data). This vector contains the GRA1 promoter, the green fluorescent protein (GFP5 S65T) sequence, and the GRA2 3′ untranslated region (3′UTR) in Bluescript SK (Stratagene, La Jolla, Calif.). Cloning of the C. parvum sequences into the NcoI-PacI or NcoI-SacI sites removed the GFP and PKR1 sequences.
A schematic of the various constructs is shown in Fig. 1. Inserts were amplified from GCH1 genomic DNA. The GCH1 Cpgp40/15 sequence (GenBank accession no. AF155624 ) is identical to the Iowa isolate sequence (GenBank accession no. AF164489 ) except for two additional serines in the polyserine domain. The Cpgp40/15 coding sequence, including the signal sequence, was amplified by PCR by using the forward primer NcoI-40 F (5′ GGC CAT GGA AAC CAG TGA AGC TGC TGC A 3′) and the reverse primer PacI-15 R (5′ GGT TAA TTA AAG CAC GAA TAA GGC TGC AAA G 3′), digested with the appropriate enzymes, and cloned into the NcoI-PacI sites of pGRA1PKR1-N177GFPGRA2 (Fig. 1, pGRA1Cpgp40/15). The sequence change needed to engineer the NcoI site resulted in an amino acid change in the second amino acid of the signal sequence from an arginine to a glutamic acid. An additional construct was generated containing the Cpgp40/15 coding sequence and the full-length Cpgp40/15 3′UTR (21) cloned into the NcoI-SacI sites of pGRA1PKR1-N177GFPGRA2, replacing the GRA2 3′UTR.
gp40 and gp15 expression vectors. Cpgp40/15 sequences were cloned into the T. gondii expression vector pGRA1PKR1-N177GFPGRA2 under the control of the GRA1 promoter.
The coding sequences for gp40 and gp15 were also cloned separately into the NcoI-PacI sites of pGRA1PKR1-N177GFPGRA2. gp40 was cloned by using the forward primer, NcoI-40 F, and a reverse primer homologous to the end of the gp40 coding sequence (E222) and containing a PacI site that provided the stop codon (5′ GGT TAA TTA ATC TGA GAG TGA TCT TCT TGA 3′) (Fig. 1, pGRA1gp40). In order to clone gp15 with the Cpgp40/15 signal sequence at the N terminus, a hybrid sequence was constructed by overlap extension as described previously (13). The 90-bp signal sequence was amplified by PCR with primer NcoI-40F and the reverse primer 5′ TTC ACT GGT TTC CTT TAA AGT TCC TCT 3′. The gp15 coding sequence was amplified with the forward primer 5′ GGA ACT TTA AAG GAA ACC AGT GAA GCT 3′ and PacI-15R. The two amplicons generated had homologous sequences at the 3′ end of the signal sequence and the 5′ end of the gp15 sequence. The amplicons were gel purified and combined together with deoxynucleoside triphosphates and Pfu Taq (Stratagene), and an extension reaction was performed for 10 cycles. Two microliters of the extension reaction was then amplified with NcoI-40F and PacI-15R to generate the final product (Fig. 1, pGRA1gp15). By using the same method, a construct was generated that contained a truncated gp15 insert from which the putative GPI anchor attachment site and downstream sequences (amino acids N295 to L326) (24) had been removed (reverse primer PacI-15ΔGPI R, 5′ GGT TAA TTA ATC CTT ATC AAC CAA GTC TCC 3′) (Fig. 1, pGRA1gp15ΔGPI).
All the Cpgp40/15 inserts were sequenced by the dye-terminator method on a Perkin-Elmer ABI 377 sequencer at the Tufts University School of Medicine core facility. Plasmids containing the correct sequences were purified either by cesium chloride density gradient centrifugation or by using a QIAGEN Megaprep kit (QIAGEN, Valencia, Calif.).
Transfections. T. gondii tachyzoites were transiently transfected as previously described (26) by using a Bio-Rad Gene Pulser II electroporator (Bio-Rad, Hercules, Calif.) at settings of 1.9 kV and 50 μF of capacitance. After a 10-min incubation at room temperature, the transfected parasites were added to HFF monolayers in 25-mm2 flasks and in 8-well chamber slides (Becton Dickinson Labware, Franklin Lakes, N.J.). Twenty-four hours postinfection, the monolayers were processed for immunofluorescence assays (IFAs) and Western blot analysis.
Immunoassays.IFAs were performed on T. gondii intracellular stages as described by Khan et al. (17). Primary antibodies were detected with Alexa Fluor-488-conjugated goat anti-mouse IgG or Alexa Fluor-594-conjugated goat anti-rabbit IgG, as appropriate. Lectin reactivity was detected directly with fluorescein isothiocyanate (FITC)-conjugated HPA lectin (EY Laboratories). The slides were mounted with Vectashield mounting medium containing 4′,6′-diamidino-2-phenylindole (DAPI; Vector Laboratories, Burlingame, Calif.) and examined by differential interference contrast (DIC) and fluorescence microscopy by using a Zeiss Axioplan 2 microscope (Carl Zeiss Microscopy, Jena, Germany). Images were captured with an IEEE1394 digital CCD camera (Hamamatsu, Hamamatsu-City, Japan), and colocalization of the fluorescent labels was determined by merging the images with OpenLab software (Improvision Inc., Lexington, Mass.).
For Western blot analysis, infected monolayers were scraped from the flasks, pelleted by centrifugation at 400 × g for 15 min, and lysed in phosphate-buffered saline (pH 7.0) containing 1% octyl glucoside and protease inhibitors. After incubation for 20 min on ice, the supernatants were collected by centrifugation at 16,000 × g for 30 min and separated by SDS-polyacrylamide gel electrophoresis (PAGE). Proteins were transferred to polyvinylidene difluoride membranes and probed for antibody and lectin reactivity as described previously (1).
Triton X-114 extraction. C. parvum oocysts (that had been hypochlorite-treated and allowed to spontaneously excyst at 37°C for 1 h) and T. gondii-infected HFF cells were lysed for 1 h on ice in Tris-buffered saline containing 2% precondensed Triton X-114. The insoluble components were pelleted at 10,000 × g for 10 min at 4°C, and the supernatant was collected and warmed to 37°C. The supernatant was centrifuged at 1,000 × g for 10 min at room temperature, and the aqueous and detergent phases were collected into different tubes. The samples were precipitated with 8 volumes of acetone overnight at −20°C and centrifuged at 3,000 × g for 15 min at 4°C, and the pellets were resuspended in SDS-PAGE sample buffer and analyzed by Western blotting.
α-N-acetylgalactosaminidase digestion. C. parvum- and T. gondii-infected HFF cell lysates were digested overnight at 37°C with 2,000 U of α-N-acetylgalactosaminidase (New England Biolabs, Beverly, Mass.) per ml in 50 mM sodium phosphate, pH 7.5, supplemented with 100 mg of bovine serum albumin (digestion buffer) per ml. Lysates treated with digestion buffer alone under the same conditions were included as controls. The specificity of the α-N-acetylgalactosaminidase digestion was ascertained by preincubation of the enzyme with the substrate p-nitrophenyl-N-acetyl-α-d-galactosaminide at a final concentration of 5 mM.
RESULTS
Cpgp40/15-encoded glycoproteins are expressed on the surface of transfected T. gondii tachyzoites. T. gondii tachyzoites, transiently transfected with the pGRA1Cpgp40/15 construct, displayed a surface-membrane-type reactivity with antibodies raised against E. coli recombinant gp40 and gp15 (Fig. 2). In contrast, none of the parasites transfected with the vector alone reacted with the antibodies (data not shown). Expression of the gene was not toxic to T. gondii tachyzoites, as all parasites expressing the gene (approximately 5 to 10% of the population) were morphologically identical to the nontransfected parasites as determined by DIC microscopy (Fig. 2A).
Cpgp40/15 proteins expressed in T. gondii are localized to the tachyzoite surface membrane and are glycosylated. T. gondii RH strain tachyzoites were transfected with pGRA1Cpgp40/15 and allowed to infect HFF monolayers. IFAs were performed 24 h postinfection with anti-rgp40 and -rgp15 sera or with MAb 4E9 followed by Alexa Fluor-488-conjugated secondary antibodies and FITC-conjugated HPA lectin. The top panels show DIC fields, and the bottom panels show a merged fluorescence image of the same field. DAPI-stained nuclei are in blue, and Alexa Fluor-488- and FITC-stained nuclei are in green.
Native C. parvum gp40 and gp15 are decorated with O-linked αGalNAc residues (1, 9). To determine if the Cpgp40/15 products expressed by T. gondii were similarly glycosylated, IFAs were performed with the anti-gp40 MAb 4E9, which recognizes an αGalNAc-containing epitope unique to the C. parvum glycoproteins gp40 and gp900 (1). In addition, reactivity with the GalNAc-specific lectin, HPA, was examined. Both these reagents reacted with the surfaces of tachyzoites expressing Cpgp40/15 products (Fig. 2). HPA also reacted (at a lower intensity) with material in the parasitophorous vacuole of nontransfected (Fig. 2) and control-transfected parasites (data not shown). This reactivity is likely due to the presence of O-glycosylated T. gondii dense granule glycoproteins such as GRA2, which have previously been shown to display αGalNAc residues (35).
Surface localization of Cpgp40/15 products is dependent upon the presence of the gp15 GPI anchor attachment site.In C. parvum, gp15 has been shown to be membrane-associated via a GPI anchor (24). To determine if surface localization of the C. parvum proteins expressed in T. gondii tachyzoites was dependent on the presence of the GPI anchor attachment sequence, IFA analysis was performed on tachyzoites transfected with pGRA1gp40, containing the gp40 coding sequence; pGRA1gp15, containing only the gp15 coding sequence; and pGRA1gp15ΔGPI, containing the same insert as pGRA1gp15 but without the sequence including and downstream of the putative GPI anchor attachment site (Fig. 1). To confirm localization to the tachyzoite surface membrane, parasites were labeled with rabbit anti-SAG1 antiserum as well as the anti-rgp40 and anti-rgp15 mouse sera. The glycoproteins expressed from the pGRA1Cpgp40/15 and pGRA1gp15 constructs colocalized with SAG1 on the tachyzoite surface membrane (Fig. 3A and C). gp40 expressed from the pGRA1gp40 construct was found in the tachyzoite cytoplasm and the parasitophorous vacuole and did not colocalize with SAG1 (Fig. 3B). The surface localization of T. gondii recombinant gp15 was disrupted by the absence of the GPI anchor attachment sequences. Similarly to that of gp40, the truncated gp15 product was found in the cytoplasm of the tachyzoite as well as in the parasitophorous vacuole (Fig. 3D).
The GPI anchor attachment site is required for surface expression of Cpgp40/15 proteins in T. gondii tachyzoites. Tachyzoites were transfected with pGRA1Cpgp40/15 (top panel), pGRA1gp40 (second panel), pGRA1gp15 (third panel), and pGRA1gp15ΔGPI (fourth panel), and IFAs were performed on infected HFF cells 24 h postinfection. The T. gondii surface protein SAG1 was detected with a rabbit anti-SAG1 serum and Alexa Fluor-594-conjugated secondary antibody, and the C. parvum proteins were detected with mouse antisera to recombinant gp40 and gp15 and Alexa Fluor-488-conjugated secondary antibody.
T. gondii Cpgp40/15 is partially processed into the gp40 and gp15 mature polypeptides.In C. parvum, the Cpgp40/15 gene is expressed as a precursor protein that is then proteolytically processed into the glycopolypeptides gp40 and gp15 (2, 29, 34). To determine if processing occurred when the Cpgp40/15 gene was heterologously expressed in T. gondii, tachyzoites were transfected with pGRA1Cpgp40/15, pGRA1gp40, pGRA1gp15, and pGRA1gp15ΔGPI, as well as the pGRA1PKR1-N177GFPGRA2 vector as a negative control, and protein expression was evaluated by Western blotting. The results are shown in Fig. 4.
Cpgp40/15, expressed in T. gondii, is proteolytically processed. Lysates of HFF cells infected for 24 h with transfected T. gondii were analyzed by Western blotting with anti-rgp40 and -rgp15 sera, MAb 4E9, and HPA lectin. The constructs used are the vector alone (lane1), pGRA1gp40/15 (lane 2), pGRA1gp40 (lane 3), pGRAgp15 (lane 4), and pGRA1gp15ΔGPI (lane 5). Lane Cp contains C. parvum oocyst lysates. The asterisks indicate partially glycosylated products, the arrowhead indicates processed gp15, and the arrows indicate native gp40 and gp15 HPA reactivity. Molecular mass is shown on the left in kilodaltons.
MAb 4E9, HPA lectin, anti-rgp40, and anti-rgp15 sera all identified a 50-kDa precursor protein in lysates from parasites expressing the full-length Cpgp40/15 (Fig. 4, all panels, lane 2) but not in the control-transfected parasites (Fig. 4, all panels, lane 1). Both anti-rgp40 and -rgp15 sera also reacted with a band of approximately 45 kDa (indicated in the figure by an asterisk) that may represent an incompletely glycosylated or unglycosylated form of the precursor, as this band was not detected with MAb 4E9 or with HPA. Likewise, anti-rgp40 serum strongly recognized two bands of ∼42 and 40 kDa in lysates of the pGRA1gp40 transfectants (Fig. 4, first panel, lane 3), but only the larger band was recognized by MAb 4E9 and HPA (Fig. 4, third and fourth panels, lane 3), suggesting that the smaller band may represent an unglycosylated species. The glycoprotein produced from pGRA1gp40 migrated at a slightly higher molecular mass than native gp40 (Fig. 4, first panel, lanes 3 and Cp, respectively). Two bands of approximately 11 to 13 kDa were identified by anti-rgp15 serum in the lysates of pGRA1gp15 transfectants, and a band of approximately 10 kDa was identified in the pGRA1gp15ΔGPI transfectants (Fig. 4, second panel, lanes 4 and 5, respectively).
Anti-rgp15 serum identified an 11-kDa band in lysates of parasites transfected with pGRA1Cpgp40/15 (Fig. 4, second panel, lane 2, arrowhead) that comigrated with the ∼13-kDa band identified in the pGRA1gp15 transfectants (Fig. 4, second panel, lane 4), suggesting that some of the gp40/15 precursor protein is processed. A glycosylated, processed form of gp40 would be expected to migrate just below the glycosylated precursor and thus may not be distinguishable from the partially glycosylated or unglycosylated 40- and 42-kDa forms of the precursor (Fig. 4, first panel, lane 2, asterisk). However, MAb 4E9 does not appear to identify any band that could represent processed gp40 (Fig. 4, third panel, lane 2).
HPA lectin identifies both gp40 and gp15 from C. parvum lysates (Fig. 4, fourth panel, lane Cp, arrows). While HPA reacted with the gp40/15 precursor and gp40 expressed alone (Fig. 4, fourth panel, lanes 2 and 3, respectively), it did not react with any form of gp15 expressed by T. gondii (Fig. 4, fourth panel, lanes 2, 4, and 5).
Since 3′UTRs can affect expression of gene products, a construct was made that exchanged the GRA2 3′UTR for the 1,003-bp Cpgp40/15 3′UTR (pGRA1Cpgp40/15-3′UTR). However, expression of Cpgp40/15 from the pGRA1Cpgp40/15-3′UTR construct did not alter processing or glycosylation of the gene product (data not shown).
T. gondii recombinant gp40/15 and gp15 partition into the Triton X-114 detergent phase, confirming an association with the T. gondii tachyzoite membrane.The glycoprotein products expressed from pGRA1Cpgp40/15 and pGRA1gp15 colocalized with SAG1, indicating that the products were associated with the tachyzoite surface membrane (Fig. 3). This localization was confirmed by Triton X-114 extraction of HFF cells infected with transfected parasites followed by phase separation. The detergent and aqueous phases from the extractions were analyzed by Western blotting with anti-rgp15 serum (Fig. 5). C. parvum gp15 (Fig. 5, lane Cp, d), Cpgp40/15 precursor and processed gp15 (Fig. 5, lane 2, d), and gp15 expressed alone (Fig. 5, lane 4, d) were all extracted into the detergent phase, as would be expected for membrane-bound proteins. Conversely, most of the gp15ΔGPI polypeptide was extracted into the aqueous phase (Fig. 5, lane 4, a, arrowhead). The gp15ΔGPI lanes had to be loaded with three times the amount of sample due to the difficulty of visualizing this recombinant on Western blots, probably due to proteolytic degradation of the truncated polypeptide.
Cpgp40/15 and gp15 expressed by T. gondii partition into the detergent phase in Triton X-114 extractions. HFF cells infected with transfected T. gondii for 24 h were extracted with Triton X-114, and the detergent and aqueous phases were analyzed by Western blotting with anti-rgp15 serum. Lanes: Cp, excysted oocysts; 1, vector only; 2, pGRA1Cpgp40/15; 3, pGRA1gp15; 4, pGRA1gp15ΔGPI. (d) Detergent phase; (a) aqueous phase. The lanes for gp15ΔGPI contain a threefold concentration of sample. Only gp15ΔGPI partitions into the aqueous phase (arrowhead). Molecular masses are shown on the left in kilodaltons.
T. gondii recombinant gp40 contains terminal O-linked αGalNAc residues.Reactivity of the T. gondii recombinants with MAb 4E9 and HPA lectin suggested the presence of O-linked αGalNAc residues (Fig. 2 and 4), as previously shown for native C. parvum gp40 (1). To confirm this observation, T. gondii recombinant gp40 was digested with α-N-acetylgalactosaminidase, which cleaves terminal O-linked αGalNAc residues linked to serine or threonine residues of the polypeptide backbone. The digestions were analyzed by Western blotting with MAb 4E9, anti-rgp40 serum, and HPA lectin (Fig. 6). After digestion with the glycosidase, neither protein was reactive with MAb 4E9 or the lectin HPA on Western blots, demonstrating loss of the epitope (Fig. 6, first and third panels, C.p. and T.g., lane 2). Probing of the blots with anti-rgp40 serum demonstrated that the glycosidase caused a size shift from 45 to approximately 36 kDa for the T. gondii recombinant antigen (Fig. 6, second panel, T.g., lane 2). The specificity of the reaction was confirmed by the ability to block digestion with the substrate p-nitrophenyl-N-acetyl-α-d-galactosaminide (Fig. 6, all panels, lane 3).
The presence of O-linked αGalNAc residues on T. gondii recombinant gp40 is confirmed by digestion with N-acetylgalactosaminidase. Lysates of HFF cells infected with pGRA1gp40-transfected T. gondii (T.g.) were digested with N-acetylgalactosaminidase, Western blotted, and probed with anti-rgp40, MAb 4E9, and HPA. C. parvum lysates (C.p.) and T. gondii transfected with the vector alone [T.g.(con)] were run in parallel. Treatments include no enzyme (lane 1), 2,000 U of N-acetylgalactosaminidase per ml (lane 2), and 2,000 U of N-acetylgalactosaminidase per ml inhibited with 5 mM p-nitrophenyl-N-acetyl-α-d-galactosaminide (lane 3). Molecular mass is shown on the left in kilodaltons.
DISCUSSION
Of all the apicomplexans, Toxoplasma is the most easily cultured and genetically manipulated. Because of these features, it has been proposed as a model system for studying the function of molecules from other apicomplexa that cannot be propagated in vitro or are less amenable to genetic manipulation (25). Our investigations suggest that T. gondii may prove useful as a surrogate system for the study of Cryptosporidium glycoproteins. Although C. parvum is somewhat AT rich (33), it is much less so than Plasmodium falciparum (31), and thus resynthesis of the Cpgp40/15 gene was not necessary for expression. Tachyzoites that had been transiently transfected with an expression vector containing the Cryptosporidium Cpgp40/15 gene expressed, glycosylated, and localized the antigen in a manner analogous to native expression of the gene. Recently, T. gondii has been used successfully to clone the C. parvum IMP-dehydrogenase gene by using a genetic complementation strategy (28). Both this study and ours illustrate that the similarities between Toxoplasma and Cryptosporidium can be exploited to advance the understanding of Cryptosporidium biology.
In C. parvum, gp15 has been shown to be attached to the membrane via a GPI anchor and the putative attachment site for the anchor has been identified (24). Targeting of the glycoproteins expressed from pGRA1Cpgp40/15 and pGRA1gp15 to the T. gondii tachyzoite surface membrane is dependent on the presence of the GPI attachment signal sequences, suggesting that T. gondii recognizes the cryptosporidial GPI attachment sequences. The addition of a GPI anchor has been shown to be sufficient to target proteins to the tachyzoite membrane (16), and it is possible that the Cpgp40/15 recombinants are targeted similarly to T. gondii GPI-anchored proteins such as SAG1. However, to confirm this hypothesis, the presence of the GPI moiety needs to be proved experimentally and the structure needs to be compared to that of the native antigen.
Targeting to the dense granules in T. gondii is considered to be the default pathway, and heterologous proteins that do not contain the specific sequences for microneme or rhoptry localization or for GPI anchor sites colocalize with dense granule proteins (15). However, neither gp40 nor gp15ΔGPI colocalized with GRA3 or with MIC2 or ROP1 (data not shown). gp40 expressed alone seems to be retained in the endoplasmic reticulum (ER). This might be part of an ER quality control mechanism by which proteins that are not properly folded or that lack signals necessary for exit cannot enter the secretory pathway appropriately and are retained in the ER (6). In addition, it is perhaps relevant that in other studies of heterologously expressed proteins a T. gondii N-terminal signal sequence was incorporated into the sequence (3, 4, 16). In our experiments, the native Cpgp40/15 signal sequence was used, and this, in the absence of the GPI anchor signal sequences, may have influenced the localization of gp40 and gp15ΔGPI.
Native gp40 and gp15 are decorated with O-linked αGalNAc determinants that may be essential for antigen function (1, 9, 34). The gp40/15 precursor and gp40 expressed in T. gondii displayed these glycotopes, as was demonstrated by reactivity with the GalNAc reactive lectin, HPA. More specifically, both these recombinant glycoproteins exhibited the 4E9 epitope. This epitope contains O-linked αGalNAc but is unique to the C. parvum glycoproteins gp900 and gp40 (1). The presence of this epitope suggests that the glycosylation of the T. gondii recombinant glycoproteins is very similar to that of the native glycoproteins. Appropriate glycosylation of T. gondii recombinant gp40 was also demonstrated by digestion of the antigen with α-N-acetylgalactosaminidase, which destroyed reactivity with HPA and MAb 4E9 and caused a significant reduction in the size of the antigen. However, the glycan structures need to be determined empirically on both the native C. parvum antigen and the T. gondii recombinant antigen in order to unequivocally confirm the similarity of glycosylation.
The size of the T. gondii Cpgp40/15 precursor appeared to be identical to that of the native C. parvum precursor. However, gp40 expressed alone migrated at a slightly larger molecular mass than that of the native antigen. It is possible that differences in glycosylation could account for this size difference. However, it is also possible that C. parvum gp40 may undergo additional processing after the initial cleavage, resulting in a size difference in the native protein that was not taken into account when constructing recombinant gp40. Sequencing of the N terminus of native gp15 suggested that proteolytic cleavage occurred between E222 and E223 (2, 23, 29). However, Winter et al. (34) observed that residues S216 to L220 were not found in any internal gp40 peptides, suggesting that the C terminus of native gp40 may undergo further proteolytic processing in C. parvum.
C. parvum gp15 is heterogeneous and can be identified as multiple bands or spots by Western blotting of one- or two-dimensional SDS-PAGE gels, respectively (23, 34). Only some of these are recognized by HPA lectin (34), raising the possibility of differential O glycosylation of the protein. However, T. gondii recombinant gp15 did not react with HPA either when it was expressed and processed from the full-length construct or when gp15 was expressed alone, suggesting that the protein may not be O glycosylated. Although the presence of αGalNAc residues on C. parvum gp15 has been proven empirically (34), none of the serine or threonine residues in gp15 are predicted to be O glycosylation sites (calculated by using the NetOGlyc 2.0 algorithm) (11). This raises the possibility that although O glycosylation may occur at specific serine or threonine residues in C. parvum, these sites may not be recognized by Toxoplasma. However, the apparent lack of αGalNAc residues in T. gondii recombinant gp15 needs to be confirmed experimentally.
The glycoprotein products of the Cpgp40/15 gene are expressed as a single precursor protein that is subsequently glycosylated and proteolytically processed into the gp40 and gp15 components (2, 29, 34). In T. gondii transfected with the pGRA1Cpgp40/15 construct, the majority of the recombinant gp40/15 remained in the precursor form, but some portion of the antigen was processed, suggesting the presence of a T. gondii homologue of the C. parvum protease. The Plasmodium knowlesi circumsporozoite protein expressed in T. gondii was also processed (4), indicating that proteases involved in processing surface antigens may be conserved across apicomplexa. Because of the presence of partially glycosylated or nonglycosylated forms of the Cpgp40/15 precursor, processing was unequivocally demonstrated only by probing Western blots for gp15. The size of the processed gp15 suggests that processing of T. gondii Cpgp40/15 occurs at or close to the same site as that in C. parvum. Given that processed gp15 is easily detected, it is unclear why MAb 4E9 or HPA does not identify processed gp40. Because gp40 expressed alone appears to be retained in the ER, a possible explanation is that the small amount of gp40 that is processed from the precursor is degraded. Additionally, the location and timing of gp40/15 processing in C. parvum are unknown. Processing of gp40/15 might be similar to that of the malarial antigen MSP1, a GPI-anchored surface antigen that is processed during merozoite maturation and invasion (12). In parasites that are stably expressing gp40/15, it may be possible to determine the timing of processing. Alternatively, the use of a less potent promoter than GRA1 (20) might result in lower expression levels and more complete glycosylation and processing. Promoters have also been shown to affect protein sorting in T. gondii, possibly via timing of expression (27); thus, the use of a more appropriate promoter may also improve gp40 and gp15 processing.
The 3′UTR of the Cpgp40/15 gene contains several polyadenylation sites that in C. parvum result in the expression of at least three mRNA transcripts from the single-copy gene; the largest of these is highly unstable (21). Since 3′UTRs have been shown to affect protein expression in protozoa (8, 32), a vector was constructed in which the Cpgp40/15 3′UTR was substituted for the GRA2 3′UTR. Expression of Cpgp40/15 from the pGRA1Cpgp40/15-3′UTR construct did not significantly alter expression, glycosylation, or processing for T. gondii. However, T. gondii recognition of the Cpgp40/15 polyadenylation sites was not confirmed.
The experiments described in this paper suggest that a system using T. gondii to study the biology of C. parvum antigens is appropriate and may prove indispensable for the molecular characterization of Cryptosporidium host-parasite interactions. This system may allow greater and more rapid progress to be made in identification of host cell receptors for sporozoite antigens, identification and characterization of C. parvum processing proteases, and understanding of the significance of antigenic diversity among genotype I isolates, greatly contributing to the development of effective vaccines and chemotherapies for C. parvum.
ACKNOWLEDGMENTS
This work was supported by grant AI46299 (to H.W.) from the National Institute of Allergy and Infectious Diseases of the National Institutes of Health (NIH) and by a pilot project grant (to R.O.C.) from Center Grant DK34928-18 from the National Institute of Diabetes and Digestive and Kidney Diseases of the NIH to the Center for Gastroenterology Research on Absorptive and Secretory Processes (GRASP) at the Tufts-New England Medical Center. R. M. O'Connor was supported by training grant AI07389 from the National Institute of Allergy and Infectious Diseases of the NIH.
We thank Jeffrey Priest and Jean Francois Dubremetz for helpful suggestions and comments; Anne Kane, Cheleste Thorpe, and Kathleen Rogers for careful review of the manuscript; Anne Kane and the GRASP Intestinal Microbiology Core for help with preparation of plasmids, media, and plates; Jean Francois Dubremetz, Lloyd Kasper, David Sibley, and Joe Schwartzman for antibodies; and Pat Mason and Saul Tzipori for oocysts.
FOOTNOTES
- Received 7 May 2003.
- Returned for modification 6 June 2003.
- Accepted 27 June 2003.
- Copyright © 2003 American Society for Microbiology
