Characterization of a >900,000-Mr Cryptosporidium parvumSporozoite Glycoprotein Recognized by ProtectiveHyperimmune Bovine Colostral Immunoglobulin
CAROLYN PETERSEN,1 2* JIRI GUT,1'2 PATRICIA S. DOYLE,"2 JOSEPH H. CRABB,3RICHARD G. NELSON,"12'4 AND JAMES H. LEECH" 2
Parasitology Laboratory, San Francisco General Hospital,' and Departments ofMedicine2 andPharmaceutical Chemisty, University of California, San Francisco, San Francisco, California 94143, and
ImmuCell Corp., Portland, Maine 041033
Received 1 July 1992/Accepted 14 September 1992
Cryptosporidium parvum, a zoonotic Apicomplexan pathogen, causes profound diarrhea, malnutrition, anddehydration in patients with AIDS. A less severe, self-limited disease occurs in immunocompetent individuals,particularly children, animal handlers, and residents of the developing world. Very little is known about thebiology of the organism, the pathophysiology of the disease process, or the mechanism of protective immunity.There is no effective therapy for cryptosporidiosis, but hyperimmune bovine colostrum raised againstCryptosporidium oocysts and sporozoites has ameliorated infection and disease in some patients with AIDS, anda variety of monoclonal antibodies, as well as hyperimmune bovine colostrum, have significantly reducedcryptosporidial infection of mice and calves. We report here the identification and initial characterization of a>900,000-Mr Cryptosporidium sporozoite glycoprotein (GP900) that is a prominent antigen recognized byprotective hyperimmune bovine colostral immunoglobulin. Three of six murine anticryptosporidial monoclonalantibodies reacted with GP900, indicating that the molecule is highly immunogenic in mice as well as in cows.GP900 is Triton X-100 soluble and N glycosylated. Western blotting of the N-deglycosylated protein, detectedwith antibodies eluted from recombinant clones expressing a partial GP900 fusion protein, suggested that thepolypeptide backbone of the glycoprotein has an Mr of < 190,000. GP900 is encoded by a single-copy gene thatresides on the largest Cryptosporidium chromosome.
Cryptosporidium parvum is a coccidian parasite of thegastrointestinal tract that causes a clinical syndrome ofdiarrhea for which there is currently no effective treatment.Infection of immunocompetent individuals is usually self-limited (36). However, infection of immunocompromisedpatients, especially those with AIDS, is often persistent andmay produce life-threatening malabsorption and dehydration(6, 24, 27). Infection is acquired by the oral ingestion ofoocysts which rupture in the upper intestine and releasesporozoites that invade the microvillus border of epithelialcells. The intracellular parasites multiply asexually, produc-ing merozoites which emerge from the host cell and infectadjacent cells. Some merozoites differentiate sexually intogametes, and fusion of gametes produces new oocysts thatare shed into the intestinal lumen. Many oocysts are ex-creted with the feces, but some rupture in the intestinallumen, releasing invasive sporozoites and producing a cycleof autoinfection. Autoinfection may contribute to the heavyparasite burdens and the persistent infections that occur inimmunocompromised individuals.Although naturally acquired immunity to C. parvum ap-
pears to involve both the cellular and humoral arms of theimmune response (6, 18, 24, 27, 35), studies with animals andhumans indicate that antibody alone can prevent new infec-tion, can ameliorate established infection, and may be effi-cacious in the treatment of immunocompromised individualswith cryptosporidiosis (1, 3, 4, 8, 10, 19, 20, 25, 26, 30,
* Corresponding author.
32-34). Preincubation of sporozoites with monoclonal anti-bodies (MAbs) to sporozoite antigens reduced sporozoiteinfectivity 80 to 100% in a mouse model of C. parvuminfection (25, 26), and the passive transfer of one of theseantisporozoite MAbs to animals with established infectionreduced the numbers of infected intestinal epithelial cells (3).Hyperimmune bovine colostrum (HBC), prepared by immu-nizing cows with oocysts, also neutralized sporozoite infec-tivity in animals and reduced clinical symptoms and oocystexcretion in some patients with AIDS and cryptosporidiosis(8, 10, 19, 20, 32-34). The immunoglobulin fraction of HBC(HBC Ig) appeared to contain the neutralizing activity (9,29). These studies indicate that intraluminal administrationof antibodies to C. parvum may be an effective treatment forcryptosporidiosis in immunocompromised individuals. Themost likely targets for these antibodies are the extracellular,intraluminal life cycle stages of the parasite, the sporozoitesand merozoites.
Important steps in the development of antibody-mediatedimmunotherapy for cryptosporidiosis are the identificationand characterization of specific protective antigens. In addi-tion, molecular cloning and in vitro expression of protectiveantigens by recombinant DNA techniques would facilitatethe preparation of large amounts of polyclonal antibody andwould yield a more uniform product than immunizationswith oocysts. We report here the identification and biochem-ical characterization of GP900, a >900,000-Mr Cryptosporid-ium glycoprotein and the initial molecular biological charac-terization of its gene. GP900 is a very prominent sporozoiteantigen recognized by protective HBC Ig.
Parasites. Cryptosporidium sp. oocysts isolated from SanFrancisco General Hospital patients with AIDS were used inthe production of polyclonal antibodies and MAbs and forelectrophoretic karyotype analysis. C. parvum, AUCP-1isolate (Byron Blagburn, Auburn University, Auburn, Ala.),was propagated in Holstein calves and was used for immuno-precipitations, Western blots (immunoblots), indirect immuno-fluorescent-antibody (EFA) studies, and genomic Southern andelectrophoretic karyotype analyses. Qocysts were isolatedfrom feces as previously described (11). A calf isolate of C.panvum was used to prepare HBC Ig at ImmuCell Corp.
Preparation of polyclonal antibodies, MAbs, and REA.Polyclonal antibodies to oocyst and sporozoite proteins wereproduced as previously described (21). MAbs were preparedby using a similar immunization regimen followed by a finalintravenous booster dose with the supernatant from soni-cated oocysts given through the tail vein 3 days beforefusion. Spleens were removed from each mouse for fusionwith cells of the mouse myeloma cell line X63.Ag8.653, anon-Ig secretor (13), by a technique previously described (7).Six sporozoite-reactive hybridomas were produced. MAbsfrom three of these hybridomas (10C6, 7B3, and E6) reactedwith very-high-Mr comigrating antigens on Western blotsand were determined by enzyme-linked immunosorbent as-say (Zymed) to be of the IgGl subclass. Hybridoma super-natants were used for immunoprecipitations, Western blots,and IFA studies.
Antibodies, designated S34 recombinant eluted antibodies(S34 REA), were affinity purified from HBC Ig antibodies(bovine S34 REA) or mouse polyclonal antioocyst/antisporo-zoite antibodies (murine S34 REA) on isopropyl-I3-D-thioga-lactopyranoside (IPTG)-induced confluent plaque lifts ofS34, a Cryptosporidium lambda gtll genomic expressionlibrary clone, and eluted with 10 mM glycine (pH 2.6)-150mM NaCl as previously described (2, 5, 22). In previouswork, using murine S34 REA, we have identified a very largesporozoite protein as the protein partially encoded by cloneS34 (21).
Preparation and assessment of in vivo efficacy ofHBC. HBCwas prepared and its efficacy in preventing disease in calveswas determined by ImmuCell Corp. (5a). Briefly, Holsteincows were repeatedly parenterally immunized during thepreparturation interval. The resulting colostrum (lot 40529),compared with nonimmune colostrum, showed protectiveefficacy in a newborn calf model of acute cryptosporidiosis.In this study, calves were fed 100 ml of control or immunecolostrum at 4 to 6 h after birth and daily thereafter. Thecalves were challenged at 12 h after birth with a virulent calfisolate of C. parvum. Statistically significant protection wasevident from diarrhea scoring, dehydration scoring, andfecal oocyst counts.
Preparation of HBC Ig. Pooled lyophilized colostral wheyIg (HBC Ig lot 40529) was prepared from the protectivecolostrum by ImmuCell's proprietary purification methods.The preparation was 51% pure IgG protein.IFA detection of cryptosporidial antigens in fixed sporozo-
ites, oocysts, and intracellular meronts. Air-dried, acetone-fixed sporozoites and oocysts or intracellular meronts inMadin-Darby canine kidney (MDCK) cells (11) were probedwith MAbs for 1 h in a humidified chamber, washed withphosphate-buffered saline, pH 7.4, and incubated with affin-ity-purified goat anti-mouse IgG-IgA-IgM antibody conju-gated with fluorescein isothiocyanate (Zymed). Slides werecounterstained with Evans blue, coverslipped, and observed
and photographed with a Nikon Optiphot microscopeequipped for immunofluorescence with fluorescein (21).Western blot of oocyst/sporozoite proteins. Sporozoites
were excysted as previously described (11), solubilized insodium dodecyl sulfate-polyacrylamide gel electrophoresis(SDS-PAGE) reducing sample buffer and fractionated bySDS-5% PAGE (17). Oocysts containing sporozoites weresuspended in a cocktail of protease inhibitors, lysed by fivefreeze-thaw cycles (21), and fractionated directly, afterextraction in an equal volume of 2% Triton X-100, or afterextraction of the Triton X-100-insoluble pellet in equalvolumes of the protease inhibitor cocktail and 4% SDS (12).To ensure efficient transfer, proteins were electroblotted tonitrocellulose membranes (31) for at least 7 h at 700 mA at8°C. Western blots were probed with MAbs, S34 REA (21),or HBC Ig. Antibodies were detected by using 1"I-protein G(Amersham) followed by autoradiography or by using alka-line phosphatase-labeled murine anti-IgG (Protoblot) devel-oped with the colorometric reagents nitroblue tetrazoliumand 5-bromo-4-chloro-3-indolyl phosphate (see Fig. 5). Mo-lecular mass standards for the >900,000-Mr glycoproteinwere apolipoprotein B (Apo B) (500 kDa) (15), nebulin (900kDa) (28), and titin (2,500 kDa) (16). Apo B was a gift fromMark Wardell, The Gladstone Institute, San Francisco,Calif.); nebulin and titin were gifts from Kuan Wang, Uni-versity of Texas, Austin. The Western blot in Fig. 5 is adigitized reproduction obtained by scanning with a Hewlett-Packard flatbed scanning densitometer in the reflectancemode.
Immunoprecipitations. One percent Triton X-100 parasiteprotein extracts and 2% SDS extracts of the remaining 1%Triton X-100-insoluble proteins were immunoprecipitated aspreviously described for malaria parasites (12). Immunecomplexes were collected with protein A-agarose beads,separated by reducing SDS-5% PAGE, Western blotted, anddetected as described above.
Deglycosylation of N-linked carbohydrate moieties. Prior togel electrophoresis and Western blotting, a parasite lysate,prepared in the presence of protease inhibitors, was incu-bated overnight at 37°C with N-glycosidase F (BoehringerMannheim) according to the manufacturer's instructions. Tocontrol for proteolysis during incubation, the parasite lysatewas also incubated under identical conditions in the absenceof N-glycosidase F.Genomic Southern and molecular karyotypic analyses. Re-
striction of the lambda gtll clone S34 with EcoRI resulted inthe excision of a 1.2-kb foreign DNA insert (data not shown).Genomic Southern analysis and [32P]dATP labeling of the1.2-kb S34 DNA insert cloned into pGem was carried out aspreviously described (23). After hybridization, membraneswere washed with 0.1x SSPE (lx SSPE is 0.15 M NaCl, 10mM NaPO4, and 1 mM EDTA [pH 7.7])-0.25% Sarkosyl at65°C and autoradiographed. Figure 6A is the digitized repro-duction of the autoradiogram obtained by scanning with aHewlett-Packard flatbed scanning densitometer in the reflec-tance mode. Southern molecular karyotypic analysis wasperformed after contour-clamped homogeneous electric fieldfractionation of Cryptosporidium chromosomes of both hu-man and animal origin, transfer of the fractionated chromo-somes to nylon membranes, and hybridization with labeledpGem S34 (14).
RESULTS
Ig from protective HBC reacts with a >900000-Mr antigenof C. parvum sporozoites. We used HBC (lot 40529) which
FIG. 1. Western blot of Triton X-100-soluble oocyst/sporozoiteproteins fractionated by SDS-PAGE and probed with MAb 10C6(lane 1), Cryptosporidium HBC IgG at 1:250, 1:500, and 1:1,000dilutions (lanes 2, 3, and 4, respectively), and sham HBC Ig at 1:250,1:500, and 1:1,000 dilutions (lanes 5, 6, and 7, respectively). 1"I-protein G was used to detect immune complexes. The dye front isindicated by an arrow. The large protein detected by MAb 10C6comigrated with a large Triton X-100-soluble protein detected byHBC Ig. Sham HBC Ig did not identify any proteins.
prevented diarrhea and reduced oocyst excretion by calveswhen given by mouth daily after oocyst administration.Oocyst excretion was reduced from 108 oocysts (controlcalves) to fewer than 103 oocysts (treated calves) on days 5to 9 postinfection (7a). To identify the protein antigensrecognized by Ig from HBC (lot 40529), we performed aWestern blot analysis with Triton X-100-extracted oocysts.HBC Ig reacted specifically with more than 10 distinctantigens with Mrs higher than 46,000 (Fig. 1, lanes 2 to 4).Smaller proteins were not resolved by our SDS-5% PAGEsystem. Inspection of the autoradiogram indicated that a>900,000-Mr antigen was the most intensely labeled.Coomassie blue staining of oocyst/sporozoite proteins frac-tionated by SDS-PAGE indicated that it was also the mostintensely stained protein expressed in these stages (data notshown).To investigate whether the >900,000-Mr antigen recog-
nized by HBC Ig was from oocysts or sporozoites, we used
-46 kD-14.3 kD
1 2 3 4 1 2 3FIG. 2. (A) Western blot of Triton X-100-soluble oocyst/sporo-
zoite proteins (lanes 1 and 3) and oocyst/sporozoite proteins immu-noprecipitated with MAb 10C6 (lanes 2 and 4) probed with HBC Ig(1:500) (lanes 1 and 2) and rabbit anti-mouse IgG (1:1,000) (lanes 3and 4). "2I-protein G was used to detect antibodies. The largeprotein immunoprecipitated by MAb 10C6 (lane 2) was detected byHBC Ig and comigrated with the protein most strongly detected byHBC Ig in the blot of Triton X-100-soluble proteins (lane 1),indicating that both reagents detected the same protein. The West-ern blot of Triton X-100-soluble proteins with a 1:500 dilution ofHBC Ig (lane 1), autoradiographed for a shorter time than the blot inFig. 1, indicated that the large protein was prominent, but no otherproteins were detected by HBC Ig under these conditions. Controllanes 3 and 4 indicated that the antigens identified in the lowerportions of lanes 2 and 4 are mouse heavy and light immunoglobulinchains from the original immunoprecipitate recognized by '"I-protein G. (B) Western blot of Triton X-100-extracted oocyst/sporozoite proteins immunoprecipitated with MAb 10C6 and probedwith MAb 10C6 (lane 1), MAb 7B3 (lane 2), and MAb E6 (lane 3).MAbs 10C6, 7B3, and E6 recognized the same molecule.
MAbs that were selected for their immunofluorescenceactivity with sporozoites. The MAbs were the products of asingle fusion, using spleen cells from a mouse immunizedwith sonicated oocysts and sporozoites. Of six sporozoite-reactive MAbs, three (MAbs 7B3, 1OC6, and E6) reacted onWestern blots with a >900,000-Mr antigen that comigratedwith the large antigen recognized by HBC Ig (Fig. 1, lane 1,data for MAb 1OC6 shown). To confirm that the comigratingantigens recognized by HBC Ig and the MAbs were thesame, we immunoprecipitated oocyst/sporozoite antigenswith MAb 1OC6, transferred the MAb 10C6 antigen tonitrocellulose, and probed with HIBC Ig (Fig. 2A, lane 2) orthe MAbs (Fig. 2B). HBC Ig, as well as MAbs 7B3 and E6,reacted specifically with the >900,000-Mr antigen that wasimmunoprecipitated by MAb 1OC6, indicating that BBC Igand the three MAbs recognize the same antigen. The MLAbsrecognized at least two distinct epitopes on the >900,000-Mrmolecule. MAb 7B3, but not MLAb 1OC6 or E6, recognized a38,000-Mr molecule present in oocysts but not sporozoites,as well as the >900,000-Mr protein, indicating that MAb 7B3
FIG. 3. (A) IFA study of fixed oocysts and sporozoites probed with MAb 7B3. The anterior portions of sporozoites leaving an excystingoocyst were stained green with fluorescein. The oocyst was counterstained red with Evans blue. (B) IFA study of fixed meronts in MDCKcells with MAb 10C6. The linear speckled pattern appeared around intracellular merozoites.
recognizes a different epitope than MAbs 1OC6 and E6 (datanot shown).MAbs 1OC6, 7B3, and E6 reacted by immunofluorescence
with acetone-fixed sporozoites but not with oocysts (Fig. 3,data for MAb 7B3 shown). The immunofluorescence reac-tivity had a globular distribution over the anterior region offixed sporozoites, but there was no reactivity with unfixedsporozoites. Thus, the >900,000-Mr protein is from sporo-zoites, not oocysts, and appears to be primarily intracellular.As previous studies have indicated that C. parvum sporozo-ites and merozoites have morphological and immunologicalsimilarities (4, 30), we also performed immunofluorescenceexperiments with MAb 1OC6 and acetone-fixed intracellularmerozoites from the MDCK cell in vitro model of C. parvuminfection. MAb 1OC6 reacted with intracellular merozoites ina circumferential pattern around the periphery of the mero-zoites (Fig. 3). Thus, MAb 1OC6 recognized an epitopecommon to sporozoites and merozoites.The >900,000-Mr protein is N glycosylated and Triton
X-100 soluble. To determine whether the >900,000-Mr anti-gen is a glycoprotein, we performed deglycosylation exper-
iments with N-glycosidase F, which removes N-linked car-bohydrates, and then Western blotting and probing with anantibody to the protein portion of the >900,000-Mr protein.The polypeptide-reactive antibody, REA, was affinity puri-fied from HBC Ig (Fig. 4) or polyclonal antisporozoite mouseserum (Fig. 5) on a lambda gtll genomic library clone (S34)expressing part of a >900,000-Mr antigen. S34 REA recog-nized the >900,000-Mr antigen that was immunoprecipitatedby MAb 1OC6 (Fig. 4) and, as previously reported (21),reacted, like the MAbs, with the anterior portion of fixedsporozoites by immunofluorescence. These results indicatethat clone S34 encodes a portion of the highly immunogenicprotein identified by the three MAbs and by HBC IgG. SinceS34 REA used for the experiment of Fig. 4 was affinitypurified from HBC Ig, these results also indicate that HBC Igcontains abundant antibodies which recognize the polypep-tide backbone of the >900,000-Mr protein.
Following treatment of oocyst/sporozoite proteins withN-glycosidase F, the three MAbs, visualized with alkalinephosphatase-conjugated secondary antibody, did not detectan antigen on Western blots (Fig. 5, lane 4, data for MAb
FIG. 4. Western blot of Triton X 100-extracted oocyst/sporozo-ite proteins inmmunoprecipitated with MiAb 10C6 and probed with
control antibody (rabbit anti-mouse IgG) (lane 1) or bovine S34 REA
(lane 2). '5I-protein G was used for detection. Bovine S34 REA
recognized the large protein immunoprecipitated by MAb 10C6.
10C6 shown). However, murine S34 REA recognized a
pattern of multiple antigens with Mr5 of <190,000 (Fig. 5,lane 2). To control for the sensitivity of the detection
method, we incubated the Western blot sections again with
the primary reagents and detected the target proteins with
125I-protein A, a much more sensitive detection method than
that using alkaline phosphatase-conjugated antibody. When
125I-protein A was used and the Western blot was autorad-
iographically overexposed, several molecules, detected byMAb 10C6 in lane 4 and comigrating with the deglycosylatedproteins detected by S34 REA and alkaline phosphatase-conjugated antibody in lane 2, began to appear (data not
shown). This result suggests that some of the MlAb 10C6
epitopes were not destroyed under the N deglycosylationconditions we used but were not visible with the less
sensitive alkaline phosphatase detection method. Taken
together these observations indicate that (i) the large proteinis a glycoprotein, (ii) the MiAbs each react with an epitopethat is dependent on intact N glycosylation, and (iii) the Mrof the N-deglycosylated protein(s) is <190,000. We have
called the >900,000-Mr glycoprotein GP900. Inmmunoprecip-itation of Triton X-100-soluble oocyst/sporozoite proteins(Fig. 5, lane 6) and the SDS extract of the Triton X-100-
insoluble pellet (Fig. 5, lane 7) indicated that almost all of
GP900 is Triton X-100 soluble.
GP900 is encoded by a single-copy gene that resides on the
largest C. parvum chromosome. Investigation of the protec-
200 ---
aVWm
__"00
97---FIG. 5. Western blot of SDS-5% PAGE gel of oocyst/sporozoite
proteins probed with murine S34 REA (lanes 1 and 2) and MAb 10C6(lanes 3 to 8). Proteins were incubated overnight with N-glycosidaseF (lanes 2 and 4) or in the absence of N-glycosidase F (lane 5) priorto SDS-PAGE. Proteins from 1% Triton X-100-extracted parasites(lane 6) or the 2% SDS-extracted Triton X-100-insoluble pellet (lane7) were immunoprecipitated with MAb 10C6. MAb 1OC6 was alsoused for immunoprecipitation in the absence of parasite proteins(lane 8). Murine S34 REA reacted weakly with a large protein (lane1) which comigrated with the protein detected by MAb 1OC6 (lane3); murine S34 REA reacted more strongly with a series of proteinswith Mrs of <190,000 after N deglycosylation (lane 2), but theseproteins were not detected by MAb 1OC6 (lane 4). The large proteinis mostly extracted in 1% Triton X-100 (lanes 6 and 7). TheMr of thelarge glycoprotein was estimated by comparing its migration withthose of other large proteins of known Mrs. GP900 migrated morerapidly than titin, a 2.5-MDa protein (16), and more slowly thannebulin, a 900-kDa protein (28), indicating that its Mr is greater than900,000.
tive efficacy of antibody to GP900 would be facilitated by itsmolecular cloning and in vitro expression using recombinantDNA techniques. As an initial step toward this goal, weisolated the 1.2-kb S34 insert which encodes a portion ofGP900. Genomic Southern analysis showed that the S34insert hybridized with single BstEII, ApaI, BclI, ClaI, andSspI fragments of approximately 9.0, >23, >23, 6, and 4 kb,respectively, and with two restriction fragments generatedby BanII (Fig. 6A). These results indicated that GP900 wasencoded by a single-copy gene. Molecular karyotypic anal-ysis revealed that the gene is located on the largest detectedCryptosporidium chromosome, a chromosome of approxi-mately 1,400 kb, and is present in both bovine and humanisolates (Fig. 6B).
DISCUSSION
A plausible approach to the treatment of cryptosporidiosisin immunocompromised individuals is passive immunother-apy with antibodies directed against C. parvum sporozoitesand/or merozoites in the lumen of the gastrointestinal tract.To identify potentially protective antigens, we performed aWestern blot analysis of oocyst/sporozoite antigens withpartially purified Ig from an HBC preparation that reducedinfection and inhibited parasite development in vivo. Wereport here the identification and biochemical characteriza-tion of GP900 and the initial molecular biological investiga-tion of its gene. The abundance and immunogenicity ofGP900 suggested that it might be a protective antigen and ledus to initiate this biochemical and molecular characteriza-tion.
1,020/945-FIG. 6. (A) Genomic Southern analysis of C. parvum DNA
restricted with BstEII (lane 1), ApaI (lane 2), BanII (lane 3), BclI(lane 4), ClaI (lane 5), or SspI (lane 6) and probed with pGem S34.The results indicated that GP900 is encoded by a single-copy gene.(B) Molecular karyotype analysis of Cryptosporidium DNA ofhuman (lanes 1 and 4) and bovine (AUCP-1) (lanes 2 and 5) originseparated by contour-clamped homogeneous electric field fraction-ation. Lanes 3 and 6 contain Saccharomyces cerevisiae chromo-somal DNA size standards. Lanes 1 to 3 are stained with ethidiumbromide. Lanes 4 to 6 are the corresponding lanes after transfer toa nylon membrane, hybridization with pGem S34, and autoradiog-raphy. The S34 gene was present on a chromosome of the same size,approximately 1,400 kb, in the karyotypes of both human andbovine C. parvum isolates. In addition, all of the separated chromo-somes of these two isolates appeared to be of the same size.
Our inspection of Coomassie blue-stained gels after elec-trophoretic separation of sporozoite proteins suggested thatGP900 was abundant. GP900 reacted more strongly withHBC Ig than did any other Triton X-100-soluble protein.GP900 was the target of three of six sporozoite-reactivemurine MAbs from one fusion, suggesting that GP900 ishighly immunogenic in mice as well as in cows. Previousinvestigations of HBC-reactive antigens did not identifyGP900, perhaps because of its very large size and itsmigration near the top of SDS gels containing higher per-
centages of acrylamide (29, 35). We performed SDS-PAGEof sporozoite proteins in gels containing 5% acrylamide,which facilitated the resolution and identification of very
large proteins.Our deglycosylation experiments with N-glycosidase F
indicated that the polypeptide backbone of GP900 had an Mr
of <190,000, suggesting that a very large fraction of thenative protein was carbohydrate. GP900-reactive polyclonalantibodies (S34 REA) detected multiple proteins with Mrs of<190,000 after N deglycosylation of GP900, whereas thespecific MAbs detected comigrating N-deglycosylated mol-ecules only when the more sensitive "2I-protein A methodand prolonged autoradiography were used. These proteinsappear to represent stepwise N deglycosylation of a proteinthat is either heavily glycosylated or glycosylated with verylarge complex mucopolysaccharides. Alternatively, theymay represent partially N-deglycosylated glycoproteins ofdifferent sizes which migrate together in the compressedhigh-Mr region of the gel. The latter explanation seems lesslikely as GP900 is the product of a single-copy gene. Themarkedly reduced reactivity of all three MAbs with N-degly-cosylated GP900 suggests that the N-linked carbohydratesare the targets of the MAbs and that they elicit an intenseimmunologic response.GP900 was detected in C. parvum isolated from calves by
MAbs raised against C parvum isolated from humans. ADNA probe derived from a library constructed from a bovineisolate hybridized under stringent conditions with genomicDNA of a human Cryptosporidium isolate. Taken together,these data indicate that the protein and its gene are presentin isolates from both sources and that there is significantconservation at the amino acid and DNA sequence levels.Both the anti-GP900 MAbs and REA reacted by immuno-
fluorescence with acetone-fixed sporozoites but not withunfixed sporozoites, suggesting that most of GP900 is intra-cellular. The immunofluorescence was localized over theanterior region of the sporozoite. No reactivity was seenwith excysted oocysts. The immunofluorescence studiesconfirmed that GP900 is a sporozoite protein, but the func-tion of GP900 in the biology of C. parvum has not beendetermined. Recent experiments, however, suggest thatGP900 and a smaller molecule, which is a product of orcross-reacts with GP900, can be surface labeled with 1"I(7a).We also observed that the anti-GP900 MAbs reacted with
intracellular C. parvum merozoites, with a circumferentialdistribution around the entire merozoite. This result indi-cates that GP900 or an immunologically cross-reactive mol-ecule is expressed by merozoites as well as sporozoites, butthe different immunofluorescence patterns with MAb 10C6for merozoites and sporozoites are unexplained. Othershared epitopes of C. parvum sporozoites and merozoiteshave been described previously (4, 30). Immunotherapy withantibodies against antigens common to both sporozoites andmerozoites might be more effective than with antibodies thatreact with only one or the other of these two life cyclestages.We conclude that GP900 is an abundant and immunogenic
glycoprotein of C. parvum sporozoites and possibly ofmerozoites and that it is a potential target for passiveimmunotherapy against cryptosporidiosis. We are attempt-ing to determine whether monospecific antibodies to recom-binant or purified GP900 are effective against C. parvuminfection in vivo and in vitro and to identify the biologicalfunction of GP900.
ACKNOWLEDGMENTSWe thank Joel Ernst and Ross Coppel for their interest and helpful
comments during this investigation.This work was supported by grants from the California Universi-
tywide AIDS Task Force (R90SF200), the University of California-San Francisco School of Medicine Research Evaluation and Allo-
cation Committee, and the National Institutes of Health (AI29882and AI129886).
REFERENCES1. Arrowood, M. J., J. R. Mead, J. R. Mahrt, and C. R. Sterling.
1989. Effects of immune colostrum and orally administeredantisporozoite monoclonal antibodies on the outcome of Cryp-tosporidiumparvum infections in neonatal mice. Infect. Immun.57:2283-2288.
2. Beall, J. A., and G. F. Mitchell. 1986. Identification of aparticular antigen from a parasite cDNA library using antibodiesaffinity purified from selected portions of Western blots. J.Immunol. Methods 86:217-223.
3. Bjorneby, J. M., B. D. Hunsaker, M. W. Riggs, and L. E.Perryman. 1991. Monoclonal antibody immunotherapy in nudemice persistently infected with Cryptosponidium parvum. In-fect. Immun. 59:1172-1176.
4. Bjorneby, J. M., M. W. Riggs, and L. E. Perryman. 1990.Cryptosporidiumparvum merozoites share neutralization sensi-tive epitopes with sporozoites. J. Immunol. 145:298-304.
5. Coppel, R. L., F. M. Smith, and C. Petersen. Antibody screeningof expression libraries. Methods Mol. Biol., in press.
5a.Crabb, J. H. Unpublished data.6. Current, W. L., N. C. Reese, J. V. Ernst, W. S. Bailey, M. B.
Heyman, and W. M. Weinstein. 1983. Human cryptosporidiosisin immunocompetent and immunodeficient persons-studies ofan outbreak and experimental transmission. N. Engl. J. Med.308:1252-1257.
7. Danforth, H. D., G. H. Campbell, M. F. Leef, and R. L.Beaudoin. 1982. Production of monoclonal antibodies by hybrid-omas sensitized to sporozoites of Plasmodium berghei. J.Parasitol. 68:1029.
7a.Doyle, P. S., and C. Petersen. Unpublished data.8. Fayer, R., C. Andrews, B. L. P. Ungar, and B. Blagburn. 1989.
Efficacy of hyperimmune bovine colostrum for prophylaxis ofcryptosporidiosis in neonatal calves. J. Parasitol. 75:393-397.
9. Fayer, R., A. Guidry, and B. L. Blagburn. 1990. Immunothera-peutic efficacy of bovine colostral immunoglobulins from ahyperimmunized cow against cryptosporidiosis in neonatalmice. Infect. Immun. 58:2962-2965.
10. Fayer, R., L. R. Perryman, and M. W. Riggs. 1989. Hyperim-mune bovine colostrum neutralizes Cryptosporidium sporozo-ites and protects mice against oocyst challenge. J. Parasitol.75:151-153.
11. Gut, J., C. Petersen, R. G. Nelson, and J. Leech. 1991. Crypto-spondium parvum: in vitro cultivation in Madin-Darby caninekidney cells. J. Protozool. 386:S72-S73.
12. Howard, R. J., J. A. Lyon, S. Uni, A. J. Saul, S. B. Aley, F.Klotz, L. J. Panton, J. A. Sherwood, K. Marsh, M. Aikawa, andE. P. Rock. 1987. Transport of an Mr approximately 300,000Plasmodiumfalciparum protein (Pf EMP 2) from the intraeryth-rocytic asexual parasite to the cytoplasmic face of the host cellmembrane. J. Cell Biol. 104:1269-1280.
13. Kearney, J. F., A. Radbruch, B. Liesegang, and K. Rajewsky.1979. A new mouse myeloma cell line that has lost immunoglob-ulin expression but permits the construction of antibody-secret-ing hybrid cell lines. J. Immunol. 123:1548-1550.
14. Kim, K., L. Gooze, C. Petersen, J. Gut, and R. G. Nelson. 1992.Isolation, sequence and molecular karyotype analysis of theactin gene of Cryptosporidium parvum. Mol. Biochem. Parasi-tol. 50:105-114.
15. Knott, T. J., R. J. Pease, L. M. Powell, S. C. Wallis, S. C. Rall,Jr., T. L. Innerarity, B. Blackhart, W. H. Taylor, Y. Marcel, R.Milne, D. Johnson, M. Fuller, A. J. Lusis, B. L. McCarthy,R W. Mahley, B. Levy-Wilson, and J. Scott. 1986. Completeprotein sequence and identification of structural domains ofhuman apolipoprotein B. Nature (London) 322:734-738.
16. Kurzban, G. P., and K. Wang. 1988. Giant polypeptides ofskeletal muscle titin: sedimentation equilibrium in guanidinehydrochloride. Biochem. Biophys. Res. Commun. 150:1155-1161.
17. Laemmli, U. K. 1970. Cleavage of structural proteins during the
assembly of the head of bacteriophage T4. Nature (London)227:680-685.
18. Lasser, K. H., K. J. Lewin, and F. W. Ryning. 1979.Cryptosporidial enteritis in a patient with congenital hypogam-maglobulinemia. Hum. Pathol. 10:234-240.
19. Nord, J., P. Ma, D. DiJohn, S. Tzipori, and C. 0. Tacket. 1990.Treatment with bovine hyperimmune colostrum of crypto-sporidial diarrhea in AIDS patients. AIDS 4:581-584.
20. Perryman, L. E., M. W. Riggs, P. H. Mason, and R. Fayer. 1990.Kinetics of Cryptosporidium parvum sporozoite neutralizationby monoclonal antibodies, immune bovine serum, and immunebovine colostrum. Infect. Immun. 58:257-259.
21. Petersen, C., J. Gut, J. H. Leech, and R. G. Nelson. 1992.Identification and initial characterization of five Cryptosporid-ium parvum sporozoite antigen genes. Infect. Immun. 60:2343-2348.
22. Petersen, C., R. Nelson, J. Leech, J. Jensen, W. Wollish, and A.Scherf. 1990. The gene product of the Plasmodium falciparum11.1 locus is a protein larger than one megadalton. Mol. Bio-chem. Parasitol. 42:189-196.
23. Petersen, C., R. Nelson, C. Magowan, W. Wollish, J. Jensen, andJ. Leech. 1989. The mature erythrocyte surface antigen ofPlasmodium falciparum is not required for knobs or cytoadher-ence. Mol. Biochem. Parasitol. 36:61-66.
24. Pitlik, S. O., V. Fainstein, D. Garza, L. Guarda, R. Bolivar, A.Rios, R. L. Hopfer, and P. A. Mansell. 1983. Humancryptosporidiosis: spectrum of disease-report of six patientsand review of the literature. Arch. Intern. Med. 143:2269-2275.
25. Riggs, M. W., T. C. McGuire, P. H. Mason, and L. E. Perryman.1989. Neutralization-sensitive epitopes are exposed on the sur-face of infectious Cryptosporidium sporozoites. J. Immunol.143:1340-1345.
26. Riggs, M. W., and L. E. Perryman. 1987. Infectivity andneutralization of Cryptosporidium parvum sporozoites. Infect.Immun. 55:2081-2087.
27. Soave, R., R. L. Danner, C. L. Honig, P. Ma, C. C. Hart, T.Nash, and R. B. Roberts. 1984. Cryptosporidiosis in homosexualmen. Ann. Intern. Med. 100:504-511.
28. Stedman, H., K. Browning, N. Oliver, M. Oronzi-Scott, K.Fischbeck, S. Sarkar, J. Sylvester, R. Schmickel, and K. Wang.1988. Nebulin cDNAs detect a 25-kilobase transcript in skeletalmuscle and localize to human chromosome 2. Genomics 2:1-7.
29. Tilley, M., R. Fayer, A. Guidry, S. J. Upton, and B. L. Blagburn.1990. Cryptosporidium parvum (Apicomplexa: Cryptosporidi-idae) oocyst and sporozoite antigens recognized by bovinecolostral antibodies. Infect. Immun. 58:2966-2971.
30. Tilley, M., S. J. Upton, R. Fayer, J. R. Barta, C. E. Chnsp, P. S.Freed, B. L. Blagburn, B. C. Anderson, and S. M. Barnard.1991. Identification of a 15-kilodalton surface glycoprotein onsporozoites of Cryptosporidium parvum. Infect. Immun. 59:1002-1007.
31. Towbin, H., T. Stahelin, and J. Gordon. 1979. Electrophoretictransfer of proteins from polyacrylamide gels to nitrocellulosesheets: procedure and some applications. Proc. Natl. Acad. Sci.USA 76:4350-4354.
32. Tzipori, S., D. Robertson, and C. Chapman. 1986. Remission ofdiarrhea due to cryptosporidiosis in an immunodeficient childtreated with hyperimmune bovine colostrum. Br. Med. J. 293:1276-1277.
33. Tzipori, S., D. Robertson, D. A. Cooper, and L. White. 1987.Chronic cryptosporidiosis and hyperimmune cow colostrum.Lancet ii:344-345.
34. Ungar, B. L. P., D. J. Ward, R. Fayer, and C. A. Quinn. 1990.Cessation of Cryptosporidium-associated diarrhea in an ac-quired immunodeficiency syndrome patient after treatment withhyperimmune bovine colostrum. Gastroenterology 98:486-489.
35. Whitmire, W. M., and J. A. Harp. 1990. Characterization ofbovine cellular and serum antibody responses during infectionby Cryptosporidium parvum. Infect. Immun. 59:990-995.
36. Wolfson, J. S., J. M. Richter, M. A. Waldron, D. J. Weber,D. M. McCarthy, and C. C. Hopkins. 1985. Cryptosporidiosis inimmunocompetent patients. N. Engl. J. Med. 312:1278-1282.
![PyramidDilatedDeeper ConvLSTM for Video Salient Object ......PyramidDilatedDeeper ConvLSTM for Video Salient Object Detection Hongmei Song 1⋆, Wenguan Wang ⋆[0000−0002−0802−9567],](https://static.documents.pub/doc/80x56/60de1b2f8027db74bd1b57cf/pyramiddilateddeeper-convlstm-for-video-salient-object-pyramiddilateddeeper.jpg)
