• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of eukcellPermissionsJournals.ASM.orgJournalEC ArticleJournal InfoAuthorsReviewers
Eukaryot Cell. Jan 2010; 9(1): 84–96.
Published online Nov 30, 2009. doi:  10.1128/EC.00288-09
PMCID: PMC2805294

Evidence for Mucin-Like Glycoproteins That Tether Sporozoites of Cryptosporidium parvum to the Inner Surface of the Oocyst Wall[down-pointing small open triangle]

Abstract

Cryptosporidium parvum oocysts, which are spread by the fecal-oral route, have a single, multilayered wall that surrounds four sporozoites, the invasive form. The C. parvum oocyst wall is labeled by the Maclura pomifera agglutinin (MPA), which binds GalNAc, and the C. parvum wall contains at least two unique proteins (Cryptosporidium oocyst wall protein 1 [COWP1] and COWP8) identified by monoclonal antibodies. C. parvum sporozoites have on their surface multiple mucin-like glycoproteins with Ser- and Thr-rich repeats (e.g., gp40 and gp900). Here we used ruthenium red staining and electron microscopy to demonstrate fibrils, which appear to attach or tether sporozoites to the inner surface of the C. parvum oocyst wall. When disconnected from the sporozoites, some of these fibrillar tethers appear to collapse into globules on the inner surface of oocyst walls. The most abundant proteins of purified oocyst walls, which are missing the tethers and outer veil, were COWP1, COWP6, and COWP8, while COWP2, COWP3, and COWP4 were present in trace amounts. In contrast, MPA affinity-purified glycoproteins from C. parvum oocysts, which are composed of walls and sporozoites, included previously identified mucin-like glycoproteins, a GalNAc-binding lectin, a Ser protease inhibitor, and several novel glycoproteins (C. parvum MPA affinity-purified glycoprotein 1 [CpMPA1] to CpMPA4). By immunoelectron microscopy (immuno-EM), we localized mucin-like glycoproteins (gp40 and gp900) to the ruthenium red-stained fibrils on the inner surface wall of oocysts, while antibodies to the O-linked GalNAc on glycoproteins were localized to the globules. These results suggest that mucin-like glycoproteins, which are associated with the sporozoite surface, may contribute to fibrils and/or globules that tether sporozoites to the inner surface of oocyst walls.

Cryptosporidium parvum and the related species Cryptosporidium hominis are apicomplexan parasites, which are spread by the fecal-oral route in contaminated water and cause diarrhea, particularly in immunocompromised hosts (1, 12, 39, 47). The infectious and diagnostic form of C. parvum is the oocyst, which has a single, multilayered, spherical wall that surrounds four sporozoites, the invasive forms (14, 27, 31). The outermost layer of the C. parvum oocyst wall is most often absent from electron micrographs, as it is labile to bleach used to remove contaminating bacteria from C. parvum oocysts (27). We will refer to this layer as the outer veil, which is the term used for a structure with an identical appearance on the surface of the oocyst wall of another apicomplexan parasite, Toxoplasma gondii (10). At the center of the C. parvum oocyst wall is a protease-resistant and rigid bilayer that contains GalNAc (5, 23, 43). When excysting sporozoites break through the oocyst wall, the broken edges of this bilayer curl in, while the overall shape of the oocyst wall remains spherical.

The inner, moderately electron-dense layer of the C. parvum oocyst wall is where the Cryptosporidium oocyst wall proteins (Cryptosporidium oocyst wall protein 1 [COWP1] and COWP8) have been localized with monoclonal antibodies (4, 20, 28, 32). COWPs, which have homologues in Toxoplasma, are a family of nine proteins that contain polymorphic Cys-rich and His-rich repeats (37, 46). Finally, on the inner surface of C. parvum oocyst walls are knob-like structures, which cross-react with an anti-oocyst monoclonal antibody (11).

Like other apicomplexa (e.g., Toxoplasma and Plasmodium), sporozoites of C. parvum are slender, move by gliding motility, and release adhesins from apical organelles when they invade host epithelial cells (1, 8, 12, 39). Unlike other apicomplexa, C. parvum parasites are missing a chloroplast-derived organelle called the apicoplast (1, 47, 49). C. parvum sporozoites have on their surface unique mucin-like glycoproteins, which contain Ser- and Thr-rich repeats that are polymorphic and may be modified by O-linked GalNAc (4-7, 21, 25, 26, 30, 32, 34, 35, 43, 45). These C. parvum mucins, which are highly immunogenic and are potentially important vaccine candidates, include gp900 and gp40/gp15 (4, 6, 7, 25, 26). gp40/gp15 is cleaved by furin-like proteases into two peptides (gp40 and gp15), each of which is antigenic (42). gp900, gp40, and gp15 are shed from the surface of the C. parvum sporozoites during gliding motility (4, 7, 35).

The studies presented here began with electron microscopic observations of C. parvum oocysts stained with ruthenium red (23), which revealed novel fibrils or tethers that extend radially from the inner surface of the oocyst wall to the outer surface of sporozoites. We hypothesized that at least some of these fibrillar tethers might be the antigenic mucins, which are abundant on the surface of C. parvum sporozoites. To test this hypothesis, we used mass spectroscopy to identify oocyst wall proteins and sporozoite glycoproteins and used deconvolving and immunoelectron microscopy (immuno-EM) with lectins and anti-C. parvum antibodies to directly label the tethers.

MATERIALS AND METHODS

Preparation of C. parvum oocysts for fluorescence microscopy.

C. parvum oocysts (Iowa strain), which were passaged through newborn calves, were purchased from Bunch Grass Farm, Dury, ID. Contaminating fecal bacteria were removed by two different methods. (i) C. parvum oocysts were enriched by centrifugation three times at 2,000 × g in cold phosphate-buffered saline (PBS), pH 7.5. The pellet containing the oocysts was resuspended in PBS, applied to a 1.2 M cesium chloride cushion, and centrifuged at 16,000 × g for 2 h. By this method, the outer veil on the surface of C. parvum oocysts was removed. (ii) Alternatively, contaminating bacteria were lysed by 5 to 10 pulses of sonication in PBS, and C. parvum oocysts were recovered in the pellet of a 2,000 × g spin for 10 min. By this method, which was used for fluorescence microscopy and electron microscopy in Fig. Fig.33 and and66 to to9,9, the outer veil was left intact. The wash was repeated three times, and oocysts were fixed in 2% paraformaldehyde in PBS for 10 min at 4°C.

FIG. 3.
(Set A) COWP1, which is present on the inner surface of the oocyst wall (stained red with MPA), is accessible to anti-COWP1 antibodies (green) only after oocyst walls are broken. In contrast, C. parvum walls of intact oocysts (labeled intact) are not ...
FIG. 6.
(Set A and panels B and C) A monoclonal antibody (Ab) to O-linked GalNAc on gp40 and gp900 (4E9 labeled green) binds to the inner surface of broken oocyst walls (Set A), the surface of intact sporozoites (B), and to vesicles within permeabilized sporozoites ...
FIG. 9.
(Set A) Polyclonal, mono-specific antibodies to recombinant gp15 (green) bind to the exterior of intact oocysts, where oocyst walls are labeled red with WGA. (Set B) Antibodies to gp15 (green) agglutinate intact C. parvum sporozoites, which are labeled ...

Sporozoites were induced to excyst by incubating oocysts in RPMI 1640 medium with 0.75% Na taurocholate (pH 7.5) at 37°C for 30 to 60 min. Sporozoites and broken oocyst walls were fixed in 2% paraformaldehyde in the presence or absence of 0.05% Triton X-100 to permeabilize membranes. For labeling the C. parvum surface, unfixed sporozoites were incubated for 60 min on ice in 200 mM carbonate-bicarbonate buffer (pH 9.3) with 20 μg/ml of the amine-reactive probe Alexa Fluor 610 5-tetrafluorophenyl (5-TFP) (Molecular Probes) (13). Alexa Fluor-labeled sporozoites were washed three times with PBS to remove excess fluorochrome, and parasites were fixed in paraformaldehyde before incubation with lectins or antibodies.

Plant lectins, including Maclura pomifera agglutinin (MPA), Ulex europaeus agglutinin I (UEA1), and Lens culinarus hemagglutinin (LCH) were purchased from EY Labs, Inc. The anti-retroviral, mannose-binding lectin cyanovirin-N (CVN) was a generous gift of Barry O'Keefe of the NCBI, Frederick, MD (2). Five hundred micrograms of each lectin was labeled with Alexa Fluors that have absorbance peaks at 488 nm or 610 nm, and unreacted fluorochromes were removed by dialysis versus PBS. Fixed oocysts, oocyst walls, and sporozoites were incubated with 4 μg/ml of each lectin for 1 h at room temperature (RT) in PBS and washed three times in PBS.

Mouse monoclonal antibodies to an N-linked epitope on gp900 (4G12) and to O-linked GalNAc on gp900 and gp40 (4E9), which have been described previously (7, 42), were diluted 1:250 and incubated with C. parvum oocysts, walls, and/or sporozoites for 1 h at RT in PBS. Bovine serum albumin (BSA) (1%) in PBS was used as a blocking reagent. Preparations were washed three times in PBS and then incubated with a secondary goat anti-mouse antibody conjugated with Alexa Fluor for 1 h at RT. Control incubation mixtures contained preimmune mouse sera or no primary mouse antibody. C. parvum parasites were washed three times with PBS.

Polyclonal mono-specific rabbit antibodies made against recombinant C. parvum gp40 and against recombinant C. parvum gp15 have been described previously (6, 26). A polyclonal rabbit antibody made against recombinant COWP1, which has been described previously (4), was a generous gift of Joe Crabb, Immucell Corp., Portland, MA. Rabbit antibodies were diluted 1:100 or 1:200 and incubated with C. parvum for 1 h at RT, and washed parasites were incubated with a secondary goat anti-rabbit antibody conjugated with Alexa Fluor dyes. BSA (1%) in PBS was used as a blocking reagent. Control incubation mixtures contained preimmune rabbit sera or no primary rabbit antibody.

Alexa Fluor-stained C. parvum parasites were incubated with 0.1 μg/ml 4′,6′-diamidino-2-phenylindole (DAPI) to stain nuclei (blue), and organisms were visualized with a DeltaVision deconvolving microscope (Applied Precision, Issaquah WA) with a channel for each fluorochrome. Images were taken at a primary magnification of ×100 and deconvolved using Applied Precision's softWoRx software.

Transmission electron microscopy (TEM) and immuno-EM of oocysts.

For TEM or immuno-EM, intact oocysts and oocyst walls after excystation were prepared as described above for fluorescence microscopy. In addition, “sonicated” oocyst walls were prepared as follows: C. parvum oocysts, which were separated from bacteria by cesium chloride centrifugation, were treated with 200 pulses (each for 10 s at 500 W) from a probe sonicator (model W250R; Ultrasonic, Inc.) in PBS on ice. Oocyst walls were separated from debris by centrifugation one or two times through a 60% sucrose cushion in PBS at 14,000 × g for 1 h. In addition, some oocyst walls were treated with pronase E (protease XIV; Sigma, St Louis, MO) in the presence of 50 mM sodium bicarbonate (pH 8.5) at 37°C for 20 h.

Intact oocysts, excysted walls, and sonicated walls were fixed in 2% paraformaldehyde and 1% glutaraldehyde buffered by 0.1 M sodium cacodylate (pH 7.3) for >12 hours at 4°C for conventional TEM. Intact oocysts were frozen and thawed three times after fixation to induce an ice artifact that allows fixatives to penetrate through the C. parvum wall (17). C. parvum oocysts were postfixed for >12 h at 4°C in 1% osmium tetroxide in the presence or absence of 0.5 mg/ml ruthenium red in cacodylate buffer (23). Water was removed from pellets in a graded series of solutions of ethanols and propylene oxide, and organisms were embedded in Epon at the Harvard Medical School Electron Microscopy Facility. Ultrathin sections (~60 to 80 nm) were cut on a Reichert Ultracut-S microtome, picked up and placed on copper grids, stained with 1% uranyl acetate and 0.2% lead citrate, and examined in a Tecnai G2 Spirit BioTWIN transmission electron microscope. Images were taken with a 2,000 Advanced Microscopy Techniques (AMT) charge-coupled-device (CCD) camera.

For immuno-EM, C. parvum parasites were fixed in 2% paraformaldehyde and 0.1% glutaraldehyde, postfixed in ruthenium red without osmium, dehydrated, and embedded in LR White resin. Ultrathin sections were picked up on copper grids, which were incubated with primary anti-C. parvum rabbit or mouse antibodies used for fluorescence microscopy. Primary rabbit antibodies were localized using protein A-gold, while primary mouse antibodies were incubated with rabbit anti-mouse antibodies prior to localization with protein A-gold. Control incubation mixtures omitted the primary mouse antibody.

Mass spectroscopy of purified oocyst walls and of total oocyst proteins purified on MPA affinity column.

For mass spectroscopy, oocysts were purified by cesium chloride centrifugation. Oocyst walls were prepared from excysted oocysts or sonicated oocysts by being centrifuged twice through a sucrose cushion. A portion of the C. parvum oocyst walls was examined by TEM for purity, and the rest of the excysted or sonicated C. parvum oocyst walls were treated with 10 μg of mass spectroscopy-grade trypsin (Promega), which was dissolved in 50 mM acetic acid and then diluted in 100 mM ammonium acetate (final pH 8) for 1 h at 37°C and then 16 h at 4°C. Undigested oocyst wall material was removed by centrifugation. Tryptic peptides from C. parvum oocyst walls were purified by reverse-phase chromatography on C18 Zip tip (Millipore) or PepClean (Pierce).

To affinity purify C. parvum glycoproteins, we sonicated oocysts, which contain walls and sporozoites, with 150 5-second pulses on ice and extracted proteins with 0.1% Triton X-100 and EDTA-free Complete protease inhibitor cocktail (Roche). Insoluble material was removed by centrifugation (>12,000 × g), and soluble proteins were applied to an MPA Sepharose column (EY Laboratories, Inc.) (41). Because there was nonspecific binding to this column, C. parvum glycoproteins were selectively eluted with 100 mM galactose rather than with sodium dodecyl sulfate (SDS). C. parvum glycoproteins were digested in solution with sequencing-grade trypsin.

Mass spectroscopy of tryptic peptides from oocyst walls or from MPA affinity-purified oocyst glycoproteins was performed using a nano-high-performance liquid chromatography (nano-HPLC) pump and autosampler (Surveyor and MicroAS, respectively; Thermo Finnigan, San Jose, CA) on a 10-cm by 100-μm inner diameter (ID) Magic C18 reverse-phase capillary column (Michrom, Auburn, CA) at the Boston University Proteomics Core Facility or at the Massachusetts Institute of Technology (MIT) Center for Cancer Research (CCR) Biopolymers Laboratory. Peptides, which were separated using gradients of 2% to 98% acetonitrile over 30 to 200 min in the presence of 0.1% formic acid (48), were analyzed using a LTQ ProteomeX ion trap mass spectrometer (Thermo Finnigan, San Jose, CA).

Mass spectra were compared to tryptic digestions of predicted C. parvum proteins from whole-genome sequencing (1, 47) using SEQUEST. Tryptic peptides with a SEQUEST XCorr score of >1.75, 2.5, or 3.5 for charge states +1, +2, and +3, respectively, a peptide probability of <0.01, and a protein probability of <0.001, and proteins with two or more high-scoring tryptic peptides were considered to be present in the sample. An estimate of the relative abundance of each protein in a sample was calculated from the area under the peak for all peptides belonging to that particular protein.

C. parvum proteins, which were identified by mass spectroscopy, were classified as secreted or membrane proteins if they contained either an N-terminal endoplasmic reticulum (ER)-targeting sequence and/or one or more predicted transmembrane helices (19, 24). Conversely, C. parvum proteins were classified as nucleocytosolic if they lacked both N-terminal sequences and transmembrane helices. C. parvum proteins were compared to the conserved domain database and to proteins in the NR database at the NCBI (22).

Analysis of C. parvum oocyst glycoproteins by lectin blots and nucleotide sugar transport assays.

Two orthogonal methods were used to characterize the carbohydrates on the glycoproteins of C. parvum oocysts. First, total oocyst proteins were obtained by sonicating (200 pulses) C. parvum oocysts in a solution of 0.1% Triton X-100, PBS (pH 7.5), and anti-fungal protease inhibitor cocktail (Roche). Walls, nuclei, and cell debris were removed by centrifugation, and half the sample was digested exhaustively with 2.5 μg of peptide-N-glycanase F (PNGase F from New England Biolabs) for 6 h at 37°C. C. parvum oocyst proteins, before or after PNGase F treatment, were run in parallel on SDS-polyacrylamide gels containing a 4 to 20% gradient of acrylamide, blotted onto polyvinylidene difluoride (PVDF) membranes, and probed with horseradish peroxidase (HRP) conjugated to MPA, UEA1, and cyanovirin-N.

Second, sporozoites from excysted oocysts were sonicated briefly in hypotonic buffer (10 mM HEPES-KOH [pH 7.2], 10 mM MgCl2, 25 mM KCl, and antifungal protease inhibitor cocktail [Roche]). C. parvum membranes were isolated by centrifugation at 100,000 × g for 45 min, and nucleotide sugar transport and transfer assays were performed with radiolabeled nucleotide sugars as described previously (3). In these assays, nucleotide sugar transport and transfer to glycoproteins are measured at the same time. The negative control is treatment of the membranes with 0.1% Triton X-100 to make open vesicles, which fail to concentrate nucleotide sugars (15). The amount of sugar transferred to N-glycans was determined by treating radiolabeled pellets with PNGase F, while the amount of sugar transferred to O-glycans was determined by treating pellets with 0.1 N NaOH (β-elimination).

RESULTS AND DISCUSSION

Fibrillar tethers extend in a radial manner from the inner surface of the oocyst wall to the surface of C. parvum sporozoites.

Ruthenium red, which is a polycationic dye, has previously been used to demonstrate a glycocalyx on the surfaces of C. parvum oocysts and sporozoites (23). Here we used ruthenium red to reveal an abundant set of fibrils, which extend radially from the inner surface of the oocyst wall to the outer surface of sporozoites (Fig. (Fig.1).1). These ruthenium red-stained fibrils, which are unbranched, were best seen when sporozoites in hypertonic fixative pull away from the intact oocyst wall or when sporozoites are in the process of excystation, suggesting the possibility that the fibrils are exaggerated by the method of preparing C. parvum oocysts for TEM. However, we do not believe that the ruthenium red-stained fibrils are a simple artifact of fixation. These ruthenium red-stained fibrils are attached to a moderately electron-dense inner layer of the oocyst wall, where COWP has been localized by immuno-EM (33). When the ruthenium red-stained fibrils are no longer attached to sporozoites, the fibrils appear to collapse into electron-dense globules, which resemble “knobs” stained by anti-C. parvum antibodies (11). Because these ruthenium red-stained fibrils appear to attach the sporozoites to the inner surface of the oocyst wall, we will refer to them as “tethers.”

FIG. 1.
Ruthenium red-stained fibrils appear to tether C. parvum sporozoites to the inner surface of oocyst walls. (A and B) Low- magnification (A) and higher-magnification (B) images of intact C. parvum oocysts reveal an intact outer veil (OV), which is separated ...

Ruthenium red also stained a narrow band, which is part of the outer veil on the outermost aspect of the C. parvum oocyst wall (Fig. (Fig.11 and and2).2). The C. parvum outer veil, which has the same appearance as that present on the surface of Toxoplasma oocysts, also contains an electron-translucent layer (10). The C. parvum outer veil is present on intact oocysts and on oocyst walls after excystation, but the outer veil is removed if oocysts are treated with bleach (27) or if the oocysts or walls are purified on CsCl2 gradients (observed here).

FIG. 2.
TEMs of excysted and sonicated C. parvum oocyst walls. (A) The oocyst wall (OW) of a sonicated oocyst, which has not been purified on cesium chloride gradient, contains the outer veil (OV). Some dense globules (Glob) are present on the inner surface of ...

COWP1 is by far the most abundant protein of purified C. parvum oocyst walls.

In order to characterize the proteins in the C. parvum oocyst wall by mass spectroscopy, we allowed sporozoites to excyst and purified excysted walls, which have characteristic curled ends at the site of the fracture (Fig. (Fig.2).2). Alternatively, we sonicated oocysts and purified walls, which curl into scrolls. Regardless of the method, the outer veil and numerous fibrillar tethers were lost from oocyst walls during the purification on CsCl2 gradients. Treatment of the purified oocyst walls with trypsin, which removed all layers of the wall except for the rigid double bilayer, did not change their curled ends after excystation or their scroll-like appearance after sonication.

By far the most abundant protein identified by mass spectroscopy from either excysted or sonicated C. parvum oocyst walls is COWP1, which makes up greater than 50% of the identified peptides (Table (Table1).1). This result is consistent with COWP1 being the most antigenic and the first C. parvum wall protein that was molecularly characterized (20, 28, 33, 37, 44). COWP6 and COWP8 are also relatively abundant in oocyst walls. COWP2, COWP3, and COWP4 are present in trace amounts, while COWP5, COWP7, and COWP9 were not detected. Together, COWPs, which are also present in the predicted proteome of Toxoplasma (37), account for ~75% of the C. parvum oocyst wall proteins identified by mass spectroscopy. The relative abundance of C. parvum oocyst wall proteins is in agreement with recent mass spectroscopic analysis of total oocyst proteins, which include proteins from oocyst walls and sporozoites (30).

TABLE 1.
Cryptosporidium oocyst wall proteins identified by mass spectroscopy

The mucin-like glycopeptide gp40/gp15 (7, 26), which localizes to the C. parvum sporozoite surface and to the inner surface of oocyst walls (see below), was a minor component of the oocyst walls by mass spectroscopy. Five other unique C. parvum proteins, which were present in much smaller amounts in purified C. parvum oocyst walls, are referred to as possible oocyst wall proteins (POWPs) (Table (Table1)1) (see Fig. S1 in the supplemental material). Possible oocyst wall protein 1 (POWP1), POWP2, and POWP3 lack any conserved domains and do not have homologues in other organisms. POWP3 is mucin-like in that it has a serine- and threonine-rich repeat near its C terminus. POWP4 and POWP5, which show a 20% amino acid identity with each other, each contain a glucose-methanol-choline (GMC) oxidoreductase domain that is also present in plant, fungal, and bacterial enzymes (e.g., cellobiose dehydrogenase) (22).

For two reasons, CCP1, CCP2, and CCP3, as well as FNPA, which has a carbohydrate-binding domain (38), (gene names in CryptoDB) were likely contaminants in C. parvum oocyst wall samples. First and most important, these C. parvum proteins are also present in all apicomplexa, including Plasmodium and Theileria, which are transmitted by insect vectors and so do not make an oocyst wall. Second, these glycoproteins were abundant in MPA affinity-purified proteins of C. parvum (below) and/or in mass spectroscopic analyses of sporozoite proteins (30, 32).

Proteins associated with N-glycan-independent quality control of protein folding in the ER, which include two chaperones (Hsp70 and Hsp90) and a protein disulfide isomerase, were also likely contaminants of the C. parvum oocyst wall preparations (Table (Table1).1). Similarly, nucleocytosolic proteins identified by mass spectroscopy were likely contaminants of the wall preparations.

The GalNAc-binding lectin (MPA) binds to oocyst walls and to peptide-N-glycanase-resistant carbohydrates on oocyst glycoproteins.

The goal here was to determine whether lectin affinity might be used to identify other oocyst wall glycoproteins, which were not identified by mass spectroscopy of cesium chloride-purified C. parvum walls. Both intact and fractured oocyst walls of C. parvum densely stain with the GalNAc-binding lectin Maclura pomifera agglutinin (MPA) (Fig. (Fig.3).3). MPA also binds to the surface of excysted sporozoites and agglutinates them (7, 35, 43). MPA binds to vesicles in permeabilized sporozoites, which are distinct from those labeled with the fucose-binding lectin Ulex europaeus agglutinin I (UEA1) (Fig. (Fig.3).3). MPA-labeled vesicles are also distinct from those labeled with the anti-retroviral lectin cyanovirin-N, which binds to high-mannose N-glycans of C. parvum (see Fig. S2 in the supplemental material) (2, 29). Cyanovirin-N binds weakly or not at all to C. parvum oocyst walls (data not shown).

Antibodies to COWP1 do not bind to the outer surface of intact oocysts but bind to the inner surface of broken oocyst walls (Fig. (Fig.3).3). This result is consistent with the localization of COWP1 to the inner, moderately electron-dense layer of the oocyst wall (33). COWP1 is for the most part absent from the surface of sporozoites but is abundant within vesicles of permeabilized sporozoites (Fig. (Fig.3).3). This result suggests that sporozoites retain some COWP1 within secretory vesicles (the most likely explanation) or have phagocytosed COWP1 (the less likely explanation).

The fucose-binding lectin UEA1 binds to broken oocysts, but not intact oocysts (data not shown), suggesting its localization on the inside oocyst walls (Fig. (Fig.3).3). A second fucose-binding lectin, Lens culinarus hemagglutinin (LCH), binds to vesicles in permeabilized sporozoites, which are distinct from those labeled with cyanovirin-N (see Fig. S3 in the supplemental material).

The binding of MPA to blots of total C. parvum oocyst proteins is not affected by treatment with peptide-N-glycanase (PNGase F) (Fig. (Fig.4).4). In contrast, the binding of UEA1 to blots of C. parvum oocyst proteins is affected by PNGase F, although it is not an all-or-nothing effect, as was the case with cyanovirin-N. The blotting results and the nucleotide sugar transport assays (next section) strongly suggest that the GalNac-binding lectin MPA is binding to O-linked glycans, while the fucose-binding lectin UEA1 is likely binding to both N- and O-linked glycans.

FIG. 4.
(A) Blotting shows that the GalNAc-binding lectin MPA binds to numerous C. parvum glycoproteins, and this binding is not affected by PNGase F treatment. (B) The fucose-binding lectin UEA1 also binds to numerous C. parvum glycoproteins, some of which are ...

C. parvum membranes have nucleotide sugar transporters for GDP-fucose, GDP-mannose, UDP-galactose, and UDP-GalNAc.

O-linked glycans, as well as complex N-linked glycans, are made in the Golgi apparatus using nucleotide sugars that are transported from the cytosol (15). Membranes of C. parvum sporozoites have nucleotide sugar transporter activities for UDP-GalNAc, UDP-galactose, GDP-fucose, and GDP-mannose (Fig. (Fig.5).5). Treatment of the radiolabeled glycoproteins with PNGase F and strong base to release N-glycans and O-glycans, respectively, suggest that GalNAc and galactose are transferred for the most part to O-linked glycans, while fucose and mannose may be transferred to both N- and O-linked glycans (Fig. (Fig.55).

FIG. 5.
(A) Membranes isolated from C. parvum sporozoites transport UDP-GalNAc, UDP-Gal, GDP-Fuc, and GDP-Man but fail to transport UDP-Glc, UDP-glucuronic acid (UDP-GlcA), and UDP-GlcNAc. Negative controls are membranes permeabilized with detergent, which fail ...

These results are consistent with the binding of the monoclonal antibody 4E9 and other GalNAc-binding lectins, such as MPA to O-linked GalNAc on the surface of C. parvum sporozoites (7). These results suggest the use of MPA affinity chromatography and mass spectroscopy to identify candidate glycoproteins, which might be part of the tethers that connect the outer surface of sporozoites to the inner surface of the oocyst walls (next section).

MPA affinity-purified oocyst glycoproteins include numerous mucin-like glycoproteins.

In the absence of any purification step, many of the proteins identified by mass spectroscopy from lysed C. parvum oocysts, which include oocyst walls and sporozoites, are nucleocytosolic, while secreted proteins are less abundant (our data not shown) (30, 32). In contrast, the vast majority of MPA affinity-purified proteins from C. parvum oocysts are secreted glycoproteins, which contain Ser- or Thr-rich repeats (gp900, gp40/15, and C. parvum MPA affinity-purified glycoprotein 1 [CpMPA1], CpMPA4, and CpMPA6) or are relatively rich in Ser and Thr (CpMPA2, CpMPA3, and CpMPA6) (Table (Table2)2) (see Fig. S4 in the supplemental material) (4, 7). Again this result is consistent with the presence of O-linked GalNAc on C. parvum glycoproteins (7).

TABLE 2.
MPA-enriched oocyst proteins identified by mass spectroscopy

MPA affinity-purified glycoproteins (CpMPA1 to CpMPA6), as well as gp40/gp15 and gp900, are unique C. parvum proteins that do not have conserved domains with the exception of CpMPA2, which has 12 KAZAL domains. KAZAL domains are small Cys-rich domains, which are present within many extracellular matrix proteins and trypsin inhibitors of metazoa but are absent from fungi, plants, and other protists (22). Other glycoproteins identified by MPA affinity chromatography and mass spectroscopy include a Gal/GalNAc-binding lectin (5), which binds to gp40 and gp900, and CpCCP2 and CpCCP3 that are likely contaminants in oocyst wall preparations (see above).

COWPs, which are abundant in oocyst wall preparations (above), are absent from MPA affinity preparations, consistent with the absence of mucin-like domains in COWPs (28). ER chaperones and protein disulfide isomerases, which are also present in oocyst wall preparations, are absent from MPA affinity-purified proteins, presumably because O-linked GalNAc is added in the Golgi apparatus rather than the ER.

The MPA affinity-purified glycoproteins from C. parvum oocysts are much less complex than glycoproteins purified by lectin affinity methods from Giardia, Entamoeba, and Trichomonas (our unpublished data). Three explanations may explain this result. First, it is possible that a terminal fucose on C. parvum glycoproteins blocks binding of MPA to O-linked GalNAc, as suggested by the distinct binding patterns of MPA and UEA1 in Fig. Fig.3.3. Second, the MPA affinity methods used here may detect only C. parvum glycoproteins, which contain Ser- and Thr-rich repeats (so-called mucins) or are rich in Ser and Thr (supported by the results in Table Table2).2). Third, we may only be detecting the most abundant C. parvum glycoproteins, which include gp900 and gp40/gp15 (30, 32). Indeed other recently characterized C. parvum mucins (C. parvum mucin 4 [CpMuc4] and CpMuc5), which are important for infectivity, are absent from the MPA affinity-purified proteins (25).

Antibodies to some C. parvum sporozoite mucins bind to the inner oocyst wall.

The hypothesis tested here is that some C. parvum mucins, which are present on the surface of C. parvum sporozoites by indirect immunofluorescence (4, 28), are also components of the fibrillar tethers that attach sporozoites to the inner surface of the oocyst wall. In support of this idea, the 4E9 monoclonal antibody (7) binds to O-linked GalNAc on the surface of C. parvum sporozoites and within small vesicles within permeabilized sporozoites (Fig. (Fig.6).6). The 4E9 antibody also binds in a punctate manner to the inside C. parvum oocyst walls. By immuno-EM, the 4E9 antibody bound to globules on the inside of the oocyst wall and to small vesicles within sporozoites.

Also in support of this idea, 4G12 monoclonal antibodies to the mucin gp900 bind to large vesicles within sporozoites and bind in a punctate manner to the inner surface of the oocyst wall (Fig. (Fig.7)7) (43). By immuno-EM, 4G12 antibodies to gp900 bind to fibrillar tethers and to large relatively electron-translucent vesicles within sporozoites.

FIG. 7.
(Set A and panel B) A monoclonal antibody (Ab) to gp900 (4G12 labeled green) binds to the inner surface of broken oocyst walls (Set A) and to the outer surface of a sporozoite (B). Oocyst walls are stained red with WGA, and the 4G12 antibody does not ...

Similarly, mono-specific antibodies to the recombinant gp40 (7) densely stain the surface of intact sporozoites and bind to relatively large vesicles within permeabilized sporozoites (Fig. (Fig.8).8). Anti-gp40 antibodies also bind in a punctate manner to the inner aspect of broken oocyst walls and bind to fibrillar tethers on the inner aspect of the oocyst wall (Fig. (Fig.8).8). Because we do not have antibodies to MPA1 to MPA6, we were unable to determine whether these C. parvum glycoproteins contribute to fibrillar tethers.

FIG. 8.
(Set A and panels B and C) Polyclonal, mono-specific antibodies to recombinant gp40 (green) bind to the inner surface of oocyst walls (Set A), the surface of intact sporozoites (B), and vesicles within permeabilized sporozoites (C). The walls of C. parvum ...

An exception to this idea is the binding of antibodies to gp15, which is cleaved from a precursor protein that also contains gp40 (6, 26, 42). Antibodies to gp15, which bind to their anterior ends and agglutinate the sporozoites, also bind to the surface of intact and broken oocysts (Fig. (Fig.9).9). By TEM, anti-gp15 antibodies bind to the outer veil, which is removed with high salt.

A revised model for the Cryptosporidium oocyst wall includes the fibrillar tethers.

Starting from the outside and working inwards (Fig. (Fig.10),10), the outer veil, which is labile to bleach, cesium chloride, and protease, contains gp15 and gp40 (6, 18, 26, 42). The inner, electron-translucent layer of the outer veil, as well as the rigid protease-resistant bilayer, remains for the most part uncharacterized. The weakly glycosylated, protease-sensitive, moderately electron-dense inner layer of the wall contains COWPs, as previously shown (33). The fibrillar tethers, which appear to connect the sporozoite to the oocyst wall, collapse into electron-dense globules that have also been called “knobs” (11). Antibodies to two mucins bind to the tethers (anti-gp900 and anti-gp40), while antibodies to O-linked GalNAc bind to globules.

FIG. 10.
A revised model for the C. parvum oocyst walls includes fibrillar tethers, which appear to attach sporozoites to the inner surface of the oocyst wall. The outer veil (khaki) stains with anti-gp15 antibodies, while the rigid protease-resistant bilayer ...

This model does not explain how the various proteins and carbohydrates are assembled in the C. parvum oocyst wall, and this model does not invalidate these mucins as vaccine candidates against C. parvum. The relatively simple protein composition of the C. parvum oocyst wall (composed mostly of COWPs) is similar to that of Entamoeba cyst walls, which contain just three types of chitin-binding lectins, or to that of Giardia cyst walls, which contain a small set of Leu-rich repeat proteins (called CWPs) (36, 40). In contrast, fungal and plant walls are much more complex than those of C. parvum oocysts (9, 16).

This model provides a novel tethering function for mucins in C. parvum oocysts, so that sporozoites are not “loose” within oocyst walls. Because we were unable to separate oocyst walls from sporozoites without removing the outer veil and most of the fibrillar tethers from the walls, the glycoprotein composition of the outer veil and fibrillar tethers is limited to those antigens, for which we have monoclonal antibodies or mono-specific polyclonal antibodies. We expect that as more antibodies become available (particularly to CpMPA1 to CpMPA6 or to C. parvum POWP1 [CpPOWP1] to CpPOWP5) or as better methods for purifying C. parvum oocyst walls are developed, other components of the outer veil, oocyst wall, and fibrillar tethers will be identified.

Supplementary Material

[Supplemental material]

Acknowledgments

This work was supported in part by NIH grants AI44070 (J.S.), GM31318 (P.W.R.), AI52786 (H.D.W.), and DK62816 (R.M.O.).

Thanks to Maria Ericsson for expert work with electron microscopy and Dick Cook for help with mass spectroscopy. Thanks to Joe Crabb for use of the anti-CWP1 monoclonal antibody.

Footnotes

[down-pointing small open triangle]Published ahead of print on 30 November 2009.

Supplemental material for this article may be found at http://ec.asm.org/.

REFERENCES

1. Abrahamsen, M. S., T. J. Templeton, S. Enomoto, J. E. Abrahante, G. Zhu, C. A. Lancto, et al. 2004. Complete genome sequence of the apicomplexan, Cryptosporidium parvum. Science 304:441-445. [PubMed]
2. Adams, E. W., D. M. Ratner, H. R. Bokesch, J. B. McMahon, B. R. O'Keefe, and P. H. Seeberger. 2004. Oligosaccharide and glycoprotein microarrays as tools in HIV glycobiology: glycan-dependent gp120/protein interactions. Chem. Biol. 11:875-881. [PubMed]
3. Banerjee, S., J. Cui, P. W. Robbins, and J. Samuelson. 2008. Use of Giardia, which appears to have a single nucleotide-sugar transporter for UDP-GlcNAc, to identify the UDP-Glc transporter of Entamoeba. Mol. Biochem. Parasitol. 159:44-53. [PubMed]
4. Barnes, D. A., A. Bonnin, J. X. Huang, L. Gousset, J. Wu, J. Gut, P. Doyle, J. F. Dubremetz, H. Ward, and C. Petersen. 1998. A novel multi-domain mucin-like glycoprotein of Cryptosporidium parvum mediates invasion. Mol. Biochem. Parasitol. 96:93-110. [PubMed]
5. Bhat, N., A. Joe, M. PereiraPerrin, and H. D. Ward. 2007. Cryptosporidium p30, a galactose/N-acetylgalactosamine-specific lectin, mediates infection in vitro. J. Biol. Chem. 282:34877-34887. [PubMed]
6. Cevallos, A. M., X. Zhang, M. K. Waldor, S. Jaison, X. Zhou, S. Tzipori, M. R. Neutra, and H. D. Ward. 2000. Molecular cloning and expression of a gene encoding Cryptosporidium parvum glycoproteins gp40 and gp15. Infect. Immun. 68:4108-4116. [PMC free article] [PubMed]
7. Cevallos, A. M., N. Bhat, R. Verdon, D. H. Hamer, B. Stein, S. Tzipori, M. E. Pereira, G. T. Keusch, and H. D. Ward. 2000. Mediation of Cryptosporidium parvum infection in vitro by mucin-like glycoproteins defined by a neutralizing monoclonal antibody. Infect. Immun. 68:5167-5175. [PMC free article] [PubMed]
8. Chen, X. M., S. P. O'Hara, B. Q. Huang, J. B. Nelson, J. J. Lin, G. Zhu, H. D. Ward, and N. F. LaRusso. 2004. Apical organelle discharge by Cryptosporidium parvum is temperature, cytoskeleton, and intracellular calcium dependent and required for host cell invasion. Infect. Immun. 72:6806-6816. [PMC free article] [PubMed]
9. de Groot, P. W., A. D. de Boer, J. Cunningham, H. L. Dekker, L. de Jong, K. L. Hellingwerf, C. de Koster, and F. M. Klis. 2004. Proteomic analysis of Candida albicans cell walls reveals covalently bound carbohydrate-active enzymes and adhesins. Eukaryot. Cell 3:955-965. [PMC free article] [PubMed]
10. Dubey, J. P., D. S. Lindsay, and C. A. Speer. 1998. Structures of Toxoplasma gondii tachyzoites, bradyzoites, and sporozoites and biology and development of tissue cysts. Clin. Microbiol. Rev. 11:267-299. [PMC free article] [PubMed]
11. Entrala, E., Y. Sbihi, M. Sánchez-Moreno, and C. Mascaró. 2001. Antigen incorporation on Cryptosporidium parvum oocyst walls. Mem. Inst. Oswaldo Cruz 96:233-235. [PubMed]
12. Fayer, R. 2004. Cryptosporidium: a water-borne zoonotic parasite. Vet. Parasitol. 126:37-56. [PubMed]
13. Ghosh, S., M. Frisardi, R. Rogers, and J. Samuelson. 2001. How Giardia swim and divide. Infect. Immun. 69:7866-7872. [PMC free article] [PubMed]
14. Harris, J. R., and F. Petry. 1999. Cryptosporidium parvum: structural components of the oocyst wall. J. Parasitol. 85:839-849. [PubMed]
15. Hirschberg, C. B., P. W. Robbins, and C. Abeijon. 1998. Transporters of nucleotide sugars, ATP, and nucleotide sulfate in the endoplasmic reticulum and Golgi apparatus. Annu. Rev. Biochem. 67:49-69. [PubMed]
16. Jamet, E., C. Albenne, G. Boudart, M. Irshad, H. Canut, and R. Pont-Lezica. 2008. Recent advances in plant cell wall proteomics. Proteomics 8:893-908. [PubMed]
17. Jenkins, M. C., C. Murphy, J. Trout, and R. Fayer. 2006. An improved electron microscopic technique for the immunolabeling of Cryptosporidium parvum oocysts. J. Parasitol. 92:403-405. [PubMed]
18. Jenkins, M. C., J. Trout, C. Murphy, J. A. Harp, J. Higgins, W. Wergin, and R. Fayer. 1999. Cloning and expression of a DNA sequence encoding a 41-kilodalton Cryptosporidium parvum oocyst wall protein. Clin. Diagn. Lab Immunol. 6:912-920. [PMC free article] [PubMed]
19. Krogh, A., B. Larsson, G. von Heijne, and E. L. Sonnhammer. 2001. Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J. Mol. Biol. 305:567-580. [PubMed]
20. Lally, N. C., G. D. Baird, S. J. McQuay, F. Wright, and J. J. Oliver. 1992. A 2359-base pair DNA fragment from Cryptosporidium parvum encoding a repetitive oocyst protein. Mol. Biochem. Parasitol. 56:69-78. [PubMed]
21. Leav, B. A., M. R. Mackay, A. Anyanwu, R. M. O'Connor, A. M. Cevallos, G. Kindra, N. C. Rollins, M. L. Bennish, R. G. Nelson, and H. D. Ward. 2002. Analysis of sequence diversity at the highly polymorphic Cpgp40/15 locus among Cryptosporidium isolates from human immunodeficiency virus-infected children in South Africa. Infect. Immun. 70:3881-3890. [PMC free article] [PubMed]
22. Marchler-Bauer, A., J. B. Anderson, M. K. Derbyshire, C. DeWeese-Scott, N. R. Gonzales, et al. 2007. CDD: a conserved domain database for interactive domain family analysis. Nucleic Acids Res. 35:D237-D240. [PMC free article] [PubMed]
23. Nanduri, J., S. Williams, T. Aji, and T. P. Flanigan. 1999. Characterization of an immunogenic glycocalyx on the surfaces of Cryptosporidium parvum oocysts and sporozoites. Infect. Immun. 67:2022-2024. [PMC free article] [PubMed]
24. Nielsen, H., S. Brunak, and G. von Heijne. 1999. Machine learning approaches for the prediction of signal peptides and other protein sorting signals. Protein Eng. 12:3-9. [PubMed]
25. O'Connor, R. M., P. B. Burns, T. Ha-Ngoc, K. Scarpato, W. Khan, G. Kang, and H. Ward. 2009. Polymorphic mucin antigens CpMuc4 and CpMuc5 are integral to Cryptosporidium parvum infection in vitro. Eukaryot. Cell 8:461-469. [PMC free article] [PubMed]
26. O'Connor, R. M., J. W. Wanyiri, A. M. Cevallos, J. W. Priest, and H. D. Ward. 2007. Cryptosporidium parvum glycoprotein gp40 localizes to the sporozoite surface by association with gp15. Mol. Biochem. Parasitol. 156:80-83. [PMC free article] [PubMed]
27. Petry, F. 2004. Structural analysis of Cryptosporidium parvum. Microsc. Microanal. 10:586-601. [PubMed]
28. Ranucci, L., H. M. Muller, G. La Rosa, I. Reckmann, M. A. Morales, F. Spano, E. Pozio, and A. Crisanti. 1993. Characterization and immunolocalization of a Cryptosporidium protein containing repeated amino acid motifs. Infect. Immun. 61:2347-2356. [PMC free article] [PubMed]
29. Samuelson, J., S. Banerjee, P. Magnelli, J. Cui, D. J. Kelleher, R. Gilmore, and P. W. Robbins. 2005. The diversity of protist and fungal dolichol-linked precursors to Asn-linked glycans likely results from secondary loss of sets of glycosyltransferases. Proc. Natl. Acad. Sci. U. S. A. 102:1548-1553. [PMC free article] [PubMed]
30. Sanderson, S. J., D. Xia, H. Prieto, J. Yates, M. Heiges, J. C. Kissinger, E. Bromley, K. Lal, R. E. Sinden, F. Tomley, and J. M. Wastling. 2008. Determining the protein repertoire of Cryptosporidium parvum sporozoites. Proteomics 8:1398-1414. [PMC free article] [PubMed]
31. Smith, H. V., R. A. Nichols, and A. M. Grimason. 2005. Cryptosporidium excystation and invasion: getting to the guts of the matter. Trends Parasitol. 21:133-142. [PubMed]
32. Snelling, W. J., Q. Lin, J. E. Moore, B. C. Millar, F. Tosini, E. Pozio, J. S. Dooley, and C. J. Lowery. 2007. Proteomic analysis and protein expression during sporozoite excystation of Cryptosporidium parvum (coccidia, Apicomplexa). Mol. Cell Proteomics 6:346-355. [PubMed]
33. Spano, F., C. Puri, L. Ranucci, L. Putignani, and A. Crisanti. 1997. Cloning of the entire COWP gene of Cryptosporidium parvum and ultrastructural localization of the protein during sexual parasite development. Parasitology 114:427-437. [PubMed]
34. Stein, B., L. Stover, A. Gillem, K. Winters, J. H. Leet, and C. Chauret. 2006. The effect of lectins on Cryptosporidium parvum oocyst in vitro attachment to host cells. J. Parasitol. 92:1-9. [PubMed]
35. Strong, W. B., J. Gut, and R. G. Nelson. 2000. Cloning and sequence analysis of a highly polymorphic Cryptosporidium parvum gene encoding a 60-kilodalton glycoprotein and characterization of its 15- and 45-kilodalton zoite surface antigen products. Infect. Immun. 68:4117-4134. [PMC free article] [PubMed]
36. Sun, C. H., J. M. McCaffery, D. S. Reiner, and F. D. Gillin. 2003. Mining the Giardia lamblia genome for new cyst wall proteins. J. Biol. Chem. 278:21701-21708. [PubMed]
37. Templeton, T. J., C. A. Lancto, V. Vigdorovich, C. Liu, N. R. London, K. Z. Hadsall, and M. S. Abrahamsen. 2004. The Cryptosporidium oocyst wall protein is a member of a multigene family and has a homolog in Toxoplasma. Infect. Immun. 72:980-987. [PMC free article] [PubMed]
38. Tosini, F., A. Agnoli, R. Mele, M. A. Gomez Morales, and E. Pozio. 2004. A new modular protein of Cryptosporidium parvum, with ricin B and LCCL domains, expressed in the sporozoite invasive stage. Mol. Biochem. Parasitol. 134:137-147. [PubMed]
39. Tzipori, S., and H. Ward. 2002. Cryptosporidiosis: biology, pathogenesis and disease. Microbes Infect. 4:1047-1058. [PubMed]
40. Van Dellen, K. L., A. Chatterjee, D. M. Ratner, P. E. Magnelli, J. Cipollo, M. Steffin, P. W. Robbins, and J. Samuelson. 2006. Unique posttranslational modifications of chitin-binding lectins of Entamoeba invadens cyst walls. Eukaryot. Cell 5:836-848. [PMC free article] [PubMed]
41. Wang, Y., S.-I. Wu, and W. S. Hancock. 2006. Approaches to the study of N-linked glycoproteins in human plasma using lectin affinity chromatography and nano-HPLC coupled to electrospray linear ion trap—Fourier transform mass spectrometry. Glycobiology 16:514-523. [PubMed]
42. Wanyiri, J. W., R. O'Connor, G. Allison, K. Kim, A. Kane, J. Qiu, A. G. Plaut, and H. D. Ward. 2007. Proteolytic processing of the Cryptosporidium glycoprotein gp40/15 by human furin and by a parasite-derived furin-like protease activity. Infect. Immun. 75:184-192. [PMC free article] [PubMed]
43. Ward, H., and A. M. Cevallos. 1998. Cryptosporidium: molecular basis of host-parasite interaction. Adv. Parasitol. 40:151-185. [PubMed]
44. Weir, C., G. Vesey, M. Slade, B. Ferrari, D. A. Veal, and K. Williams. 2000. An immunoglobulin G1 monoclonal antibody highly specific to the wall of Cryptosporidium oocysts. Clin. Diagn. Lab. Immunol. 7:745-750. [PMC free article] [PubMed]
45. Winter, G., A. A. Gooley, K. L. Williams, and M. B. Slade. 2000. Characterization of a major sporozoite surface glycoprotein of Cryptosporidum parvum. Funct. Integr. Genomics 1:207-217. [PubMed]
46. Xiao, L., J. Limor, U. M. Morgan, I. M. Sulaiman, R. C. Thompson, and A. A. Lal. 2000. Sequence differences in the diagnostic target region of the oocyst wall protein gene of Cryptosporidium parasites. Appl. Environ. Microbiol. 66:5499-5502. [PMC free article] [PubMed]
47. Xu, P., G. Widmer, Y. Wang, L. S. Ozaki, J. M. Alves, M. G. Serrano, D. Puiu, P. Manque, D. Akiyoshi, et al. 2004. The genome of Cryptosporidium hominis. Nature 431:1107-1112. [PubMed]
48. Yates, J. R., III, E. Carmack, L. Hays, A. J. Link, and J. K. Eng. 1999. Automated protein identification using microcolumn liquid chromatography-tandem mass spectrometry. Methods Mol. Biol. 112:553-569. [PubMed]
49. Zhu, G., M. J. Marchewka, and J. S. Keithly. 2000. Cryptosporidium parvum appears to lack a plastid genome. Microbiology 146:315-321. [PubMed]

Articles from Eukaryotic Cell are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...