The human AC133 antigen and mouse prominin are structurally related plasma membrane proteins. However, their tissue distribution is distinct, with the AC133 antigen being found on hematopoietic stem and progenitor cells and prominin on various epithelial cells. To determine whether the human AC133 antigen and mouse prominin are orthologues or distinct members of a protein family, we examined the human epithelial cell line Caco-2 for the possible expression of the AC133 antigen. By both immunofluorescence and immunoprecipitation, the AC133 antigen was found to be expressed on the surface of Caco-2 cells. Interestingly, immunoreactivity for the AC133 antigen, but not its mRNA level, was down-regulated upon differentiation of Caco-2 cells. The AC133 antigen was specifically located at the apical rather than basolateral plasma membrane. An apical localization of the AC133 antigen was also observed in various human embryonic epithelia including the neural tube, gut, and kidney. Electron microscopy revealed that, within the apical plasma membrane of Caco-2 cells, the AC133 antigen was confined to microvilli and absent from the planar, intermicrovillar regions. This specific subcellular localization did not depend on an epithelial phenotype, because the AC133 antigen on hematopoietic stem cells, as well as that ectopically expressed in fibroblasts, was selectively found in plasma membrane protrusions. Hence, the human AC133 antigen shows the features characteristic of mouse prominin in epithelial and transfected non-epithelial cells, i.e. a selective association with apical microvilli and plasma membrane protrusions, respectively. Conversely, flow cytometry of murine CD34+ bone marrow progenitors revealed the cell surface expression of prominin. Taken together, the data strongly suggest that the AC133 antigen is the human orthologue of prominin. The human AC133 antigen and mouse prominin are structurally related plasma membrane proteins. However, their tissue distribution is distinct, with the AC133 antigen being found on hematopoietic stem and progenitor cells and prominin on various epithelial cells. To determine whether the human AC133 antigen and mouse prominin are orthologues or distinct members of a protein family, we examined the human epithelial cell line Caco-2 for the possible expression of the AC133 antigen. By both immunofluorescence and immunoprecipitation, the AC133 antigen was found to be expressed on the surface of Caco-2 cells. Interestingly, immunoreactivity for the AC133 antigen, but not its mRNA level, was down-regulated upon differentiation of Caco-2 cells. The AC133 antigen was specifically located at the apical rather than basolateral plasma membrane. An apical localization of the AC133 antigen was also observed in various human embryonic epithelia including the neural tube, gut, and kidney. Electron microscopy revealed that, within the apical plasma membrane of Caco-2 cells, the AC133 antigen was confined to microvilli and absent from the planar, intermicrovillar regions. This specific subcellular localization did not depend on an epithelial phenotype, because the AC133 antigen on hematopoietic stem cells, as well as that ectopically expressed in fibroblasts, was selectively found in plasma membrane protrusions. Hence, the human AC133 antigen shows the features characteristic of mouse prominin in epithelial and transfected non-epithelial cells, i.e. a selective association with apical microvilli and plasma membrane protrusions, respectively. Conversely, flow cytometry of murine CD34+ bone marrow progenitors revealed the cell surface expression of prominin. Taken together, the data strongly suggest that the AC133 antigen is the human orthologue of prominin. Chinese hamster ovary bone marrow mononuclear cells Hanks' balanced salt solution phycoerythrin sulfosuccinimidyl-6(biotinamido)-hexanoate biotinylating agent phosphate-buffered saline monoclonal antibody A characteristic feature of epithelial cells, which is a prerequisite for their function, is the selective association of certain proteins with specific subdomains of the plasma membrane (1.Rodriguez-Boulan E. Nelson J. Science. 1989; 245: 718-725Crossref PubMed Scopus (961) Google Scholar, 2.Nelson W.J. Science. 1992; 258: 948-955Crossref PubMed Scopus (174) Google Scholar, 3.Simons K. Dupree P. Fiedler K. Huber L.A. Kobayashi T. Kurzchalia T. Olkkonen V. Pimplikar S. Parton R. Dotti C. Cold Spring Harbor Symp. Quant. Biol. 1992; 57: 611-619Crossref PubMed Google Scholar). An example is the recently identified protein prominin, a 115-kDa five transmembrane domain protein found to be expressed on the apical surface of neuroepithelial cells and several other embryonic epithelia and on brush border membranes of adult kidney proximal tubules (4.Weigmann A. Corbeil D. Hellwig A. Huttner W.B. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12425-12430Crossref PubMed Scopus (502) Google Scholar). Within the apical plasma membrane domain, prominin is selectively associated with microvilli and other related plasma membrane protrusions rather than the planar subdomain of the membrane. Studies with prominin-transfected Madin-Darby canine kidney cells have shown that this selective association is due to a combination of apical sorting and retention in microvilli (5.Corbeil D. Röper K. Hannah M.J. Hellwig A. Huttner W.B. J. Cell Sci. 1999; 112: 1023-1033Crossref PubMed Google Scholar). Remarkably, a plasma membrane protrusion-specific localization of prominin was also observed in prominin-transfected non-epithelial cells (i.e. CHO cells),1 showing that the mechanism underlying the selective association of membrane proteins with plasmalemmal protrusions is conserved between epithelial and non-epithelial cells (4.Weigmann A. Corbeil D. Hellwig A. Huttner W.B. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12425-12430Crossref PubMed Scopus (502) Google Scholar). These studies were conducted with prominin derived from one species, the mouse. In independent studies, a novel marker of human hematopoietic stem and progenitor cells was identified, characterized, and referred to as AC133 antigen (6.Yin A.H. Miraglia S. Zanjani E.D. Almeida-Porada G. Ogawa M. Leary A.G. Olweus J. Kearney J. Buck D.W. Blood. 1997; 90: 5002-5012Crossref PubMed Google Scholar, 7.Miraglia S. Godfrey W. Yin A.H. Atkins K. Warnke R. Holden J.T. Bray R.A. Waller E.K. Buck D.W. Blood. 1997; 90: 5013-5021Crossref PubMed Google Scholar). The human AC133 antigen appears to be related to mouse prominin in size (120 kDa), membrane topology (five transmembrane domains, two large extracellular loops), and sequence (8.Corbeil D. Röper K. Weigmann A. Huttner W.B. Blood. 1998; 91: 2625-2626Crossref PubMed Google Scholar, 9.Miraglia S. Godfrey W. Buck D. Blood. 1998; 91: 4390-4391Crossref PubMed Google Scholar). However, the tissue distribution reported for the human AC133 antigen (6.Yin A.H. Miraglia S. Zanjani E.D. Almeida-Porada G. Ogawa M. Leary A.G. Olweus J. Kearney J. Buck D.W. Blood. 1997; 90: 5002-5012Crossref PubMed Google Scholar, 7.Miraglia S. Godfrey W. Yin A.H. Atkins K. Warnke R. Holden J.T. Bray R.A. Waller E.K. Buck D.W. Blood. 1997; 90: 5013-5021Crossref PubMed Google Scholar) is remarkably distinct from that reported for mouse prominin (4.Weigmann A. Corbeil D. Hellwig A. Huttner W.B. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12425-12430Crossref PubMed Scopus (502) Google Scholar). In contrast to the epithelial expression of prominin (4.Weigmann A. Corbeil D. Hellwig A. Huttner W.B. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12425-12430Crossref PubMed Scopus (502) Google Scholar), the AC133 antigen was found to be selectively expressed on human CD34bright progenitors derived from bone marrow, fetal liver, and peripheral blood (6.Yin A.H. Miraglia S. Zanjani E.D. Almeida-Porada G. Ogawa M. Leary A.G. Olweus J. Kearney J. Buck D.W. Blood. 1997; 90: 5002-5012Crossref PubMed Google Scholar) and on subsets of CD34+leukemia cells, i.e. non-epithelial cells, but not in epithelial tissues such as colon and kidney (7.Miraglia S. Godfrey W. Yin A.H. Atkins K. Warnke R. Holden J.T. Bray R.A. Waller E.K. Buck D.W. Blood. 1997; 90: 5013-5021Crossref PubMed Google Scholar). Moreover, the sequence identity between mouse prominin and the human AC133 antigen is relatively low (average 60%), in particular in the three extracellular domains (54%), which comprise more than 70% of the polypeptide. This raises a major question. Is the AC133 antigen the human orthologue of prominin? If so, an explanation would be required as to why the AC133 antigen is apparently not expressed in epithelial tissues such as adult kidney (7.Miraglia S. Godfrey W. Yin A.H. Atkins K. Warnke R. Holden J.T. Bray R.A. Waller E.K. Buck D.W. Blood. 1997; 90: 5013-5021Crossref PubMed Google Scholar), which in the case of mouse does express prominin (4.Weigmann A. Corbeil D. Hellwig A. Huttner W.B. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12425-12430Crossref PubMed Scopus (502) Google Scholar), although mRNA for the AC133 antigen is well detectable (7.Miraglia S. Godfrey W. Yin A.H. Atkins K. Warnke R. Holden J.T. Bray R.A. Waller E.K. Buck D.W. Blood. 1997; 90: 5013-5021Crossref PubMed Google Scholar). In addition, one would then expect the AC133 antigen to be selectively associated with plasma membrane protrusion upon expression in fibroblasts, as previously observed in the case of prominin (4.Weigmann A. Corbeil D. Hellwig A. Huttner W.B. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12425-12430Crossref PubMed Scopus (502) Google Scholar). Conversely, prominin should then be expressed also in murine hematopoietic progenitor cells. Or, alternatively, are prominin and the AC133 antigen distinct members of a protein family? In the support of this possibility is the occurrence of at least three open reading frames in theCaenorhabditis elegans genome related to mouse prominin and the human AC133 antigen (8.Corbeil D. Röper K. Weigmann A. Huttner W.B. Blood. 1998; 91: 2625-2626Crossref PubMed Google Scholar). In that case, one would not necessarily expect the AC133 antigen to be selectively associated with plasma membrane protrusions. Answering this question is particularly important in light of a set of recent findings (10.Maw M.A. Corbeil D. Koch J. Hellwig A. Wilson-Wheeler J.C. Bridges R.J. Kumaramanickavel G. John S. Nancarrow D. Röper K. Weigmann A. Huttner W.B. Denton M.J. Hum. Mol. Genet. 2000; 9: 27-34Crossref PubMed Scopus (229) Google Scholar) that are as follows. First, a frameshift mutation in the gene encoding the AC133 antigen, which results in a truncated protein, causes retinal degeneration in humans. Second, a similar truncated version of mouse prominin does not reach the cell surface. Third, consistent with the detection of AC133 immunoreactivity in human retinoblastoma cells (6.Yin A.H. Miraglia S. Zanjani E.D. Almeida-Porada G. Ogawa M. Leary A.G. Olweus J. Kearney J. Buck D.W. Blood. 1997; 90: 5002-5012Crossref PubMed Google Scholar), in mouse, prominin is found in photoreceptor cells where it is selectively associated with the plasma membrane evaginations at the base of the rod outer segment that are intermediates in disc formation. If the AC133 antigen was the human orthologue of prominin rather than a distinct member of the prominin family, this would suggest that the lack of appearance of this protein in the plasma membrane evaginations of the rod outer segment may lead to impaired disc formation and, eventually, to retinal degeneration. Here we report several lines of evidence strongly suggesting that the AC133 antigen is the human orthologue of prominin. The evidence includes the expression of the AC133 antigen in certain human epithelial cells. This apparent contradiction of previous observations (6.Yin A.H. Miraglia S. Zanjani E.D. Almeida-Porada G. Ogawa M. Leary A.G. Olweus J. Kearney J. Buck D.W. Blood. 1997; 90: 5002-5012Crossref PubMed Google Scholar, 7.Miraglia S. Godfrey W. Yin A.H. Atkins K. Warnke R. Holden J.T. Bray R.A. Waller E.K. Buck D.W. Blood. 1997; 90: 5013-5021Crossref PubMed Google Scholar) is very likely explained by our finding that the epitope recognized by monoclonal antibody AC133, which is thought to be a glycosylated structure (7.Miraglia S. Godfrey W. Yin A.H. Atkins K. Warnke R. Holden J.T. Bray R.A. Waller E.K. Buck D.W. Blood. 1997; 90: 5013-5021Crossref PubMed Google Scholar), is down-regulated upon differentiation of epithelial cells. Caco-2 cells (kindly donated by Drs. S. Robine and D. Louvard, CNRS-Institut Curie, Paris, France) were maintained in Dulbecco's minimal essential medium supplemented with 10% fetal calf serum, 1% non-essential amino acids (Life Technologies, Inc.), 2 mml-glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin and were cultured in a humidified incubator at 37 °C under 10% CO2atmosphere. Cultures were used between passages 8 and 18. CHO cells were cultured in Ham's F-12 medium supplemented with 10% fetal calf serum and 50 μg/ml gentamycin at 37 °C under 5% CO2. For experiments studying polarized cell surface delivery, Caco-2 cells were plated at a density of 4.5 × 105cells/cm2 on permeable membranes (24-mm TranswellTM-COL chambers, 0.4-μm pore size). Media were changed every day, and the experiments were performed 11 days after seeding to allow the development of a tight monolayer. CHO cells were transfected with the pCR3.1-Uni plasmid containing the AC133 antigen-coding cDNA sequence under the control of the cytomegalovirus promoter (7.Miraglia S. Godfrey W. Yin A.H. Atkins K. Warnke R. Holden J.T. Bray R.A. Waller E.K. Buck D.W. Blood. 1997; 90: 5013-5021Crossref PubMed Google Scholar), using the LipofectAMINE reagent (Life Technologies, Inc.) according to the supplier's instructions. Cells expressing the neomycin resistance gene were then selected by introducing 600 μg/ml of G418 into the incubation medium. Two weeks later, G418-resistant colonies were pooled and expanded. To enhance the expression of the transgene, cells were incubated for 17 h with 10 mm sodium butyrate (11.Gorman C. Howard B. Reeves R. Nucleic Acids Res. 1983; 11: 7631-7648Crossref PubMed Scopus (282) Google Scholar). All subsequent steps were performed in medium lacking butyrate. Under these conditions 10–30% of neomycin-resistant cells expressed the recombinant AC133 antigen. Human bone marrow was aspirated into preservative-free heparin from the posterior iliac crest of normal adult volunteers according to procedures approved by the ethics committee of the Royal Adelaide Hospital. Bone marrow mononuclear cells (BMNC) were obtained following separation over Ficoll (Lymphoprep, Nygaard) and washed twice at 4 °C in HHF buffer (comprising calcium and magnesium-free Hanks' balanced salt solution (HBSS) supplemented with 5% human AB serum and 5% fetal calf serum). CD34+ cells were isolated from BMNC suspensions by means of immunomagnetic beads (Dynal Pty. Ltd). Briefly, BMNC were pelleted by centrifugation, and the pellet was suspended in 100 μl of CD34 Dynabeads to give a final cell:bead ratio of 4:1. After gentle vortexing for 2 min, the cells were resuspended in 4 ml of HHF and placed on a rotator for a further 30 min at 4 °C to ensure complete capture of CD34+ cells. Bead-rosetted CD34+ cells were isolated and purified by five sequential rounds of washing with HHF followed by capture on a magnet. The CD34 antibody-coated beads were released from the CD34+ cells by the addition of 200 μl of the CD34 Detachabead reagent (supplied with the kit) followed by incubation in a shaking waterbath for 45 min at 37 °C. After removal of the detached beads using the magnet, the cells were washed three times in HHF, and the aliquots were analyzed for purity using HPCA-2-PE (Becton Dickinson) and for expression of AC133 antigen using AC133-PE conjugates (Amcell). In all experiments, the purity of the CD34+ cell preparations obtained in this manner was >98%. Mouse bone marrow cells were obtained from 8–10-week-old BALB/c mice by flushing dissected femurs, using a syringe with 27G needle, with Hanks' balanced salt solution supplemented with 0.5% bovine serum albumin (Sigma). Red blood cells were lysed by a 5-min incubation with red cell lysing buffer (Sigma), and BMNC were isolated on a Histopaque 1077 gradient (Sigma). Interface cells were harvested and washed twice with HBSS/bovine serum albumin at 4 °C and adjusted to a concentration of 2 × 107/ml prior to staining for flow cytometric analysis. Antibodies were purified from murine ascites or hollow fiber culture harvests (Cellmax QUAD artificial capillary system, Cellco Inc. Germantown, MD) by protein A affinity chromatography and phycoerythrin (PE) conjugates prepared as described previously (6.Yin A.H. Miraglia S. Zanjani E.D. Almeida-Porada G. Ogawa M. Leary A.G. Olweus J. Kearney J. Buck D.W. Blood. 1997; 90: 5002-5012Crossref PubMed Google Scholar). 106 murine BMNC were stained for 30 min at 4 °C with biotinylated anti-mouse CD34 (RAM 34) and, after washing with HBSS/bovine serum albumin, stained with streptavidin-Cy-Chrome (Pharmingen). Cells were washed twice with HBSS/bovine serum albumin before incubation with 150 ng/test of PE-conjugated 13A4 antibody (anti-mouse prominin). Stained cells were analyzed on a FACScan flow cytometer (BDIS, San Jose, CA) with the scatter gates set on the lymphoid population. 50,000 events were acquired and analyzed using Cell Quest software (BDIS). Total RNA from Caco-2 cells was prepared using the acidic phenol/chloroform phase separation method (12.Chomczynski P. Sacchi N. Anal. Biochem. 1987; 162: 156-159Crossref PubMed Scopus (66733) Google Scholar). Northern blot analyses were performed usingCLONTECH multiple tissue Northern blot membrane and by resolving 20 μg of total RNA from Caco-2 cells on formaldehyde-agarose gels followed by RNA transfer to nylon membranes and hybridization with the EcoRI-EcoRI fragment (nucleotides 756–1195) derived from the human AC133 hematopoietic stem cell antigen cDNA (GenBankTM accession numberAF027208). The probe was labeled with [α-32P]dCTP using the Rediprime kit (Amersham Pharmacia Biotech). Hybridization and washing were performed either at 65 °C using Church buffer as described elsewhere (13.Church G.M. Gilbert W. Proc. Natl. Acad. Sci. U. S. A. 1984; 81: 1991-1995Crossref PubMed Scopus (7898) Google Scholar) or at 68 °C using the Expresshyb hybridization solution (CLONTECH) according to the manufacturer's instructions. Blots were analyzed using a Fuji phosphoimager. All steps of cell surface biotinylation and immunoprecipitation were carried out at 4 °C. Just prior to use, the membrane-impermeable sulfo-NHS-LC-biotin agent (Pierce) was dissolved in Ca/Mg-PBS (PBS containing 1 mm CaCl2 and 0.5 mmMgCl2) to a final concentration of 0.2 mm. After repeated washing (three times) with Ca/Mg-PBS, Caco-2 cells (on 100-mm dishes) were incubated for 1 h in 3 ml of the biotin solution. The cells were washed three times with Ca/Mg-PBS and then incubated for 10 min with Ca/Mg-PBS containing 20 mmglycine to quench the residual biotin. Cells were then lysed in ice-cold solubilization buffer (1% Triton X-100, 0.1% SDS, 150 mm NaCl, 5 mm EGTA, 50 mm Tris/HCl, pH 7.5, 1 mm phenylmethylsulfonyl fluoride, 2 μg/ml leupeptin, 10 μg/ml aprotinin). Cell lysates obtained after centrifugation (10 min, 10,000 × g) were subjected to immunoprecipitation. For domain-selective cell surface biotinylation (14.Le Bivic A. Real F.X. Rodriguez-Boulan E. Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 9313-9317Crossref PubMed Scopus (183) Google Scholar), 11-day-old monolayers on duplicate Transwell filters were washed three times with Ca/Mg-PBS and then biotinylated from either the apical (1.5 ml) or basolateral (2.6 ml) chamber compartment. The compartment not receiving the sulfo-NHS-LC-biotin was filled with Ca/Mg-PBS. After quenching, cells were scraped into ice-cold solubilization buffer, and cell lysates were prepared as described above. Cell lysates corresponding to one confluent 100-mm dish or half of one 24-mm filter were diluted 4-fold in immunoprecipitation buffer (1% Triton X-100, 0.1% SDS, 150 mm NaCl, 10 mmEDTA, 50 mm Tris/HCl, pH 7.8, 10 μg/ml aprotinin, 2 μg/ml leupeptin, and 1 mm phenylmethylsulfonyl fluoride), and monoclonal antibody (mAb) AC133 (10 μg/ml) was added. Samples were incubated overnight on an end-over-end shaker. Immune complexes were collected with protein A-Sepharose CL-4B beads (Amersham Pharmacia Biotech), pre-adsorbed with rabbit anti-mouse IgG, and used for deglycosylation, SDS-polyacrylamide gel electrophoresis, and NeutrAvidin blotting (see below). Immune complexes containing the cell surface-biotinylated proteins were eluted from protein A-Sepharose beads using glycosidase buffer (1% Triton X-100, 0.1% SDS, 25 mm EDTA, 1% 2-mercaptoethanol, 50 mm sodium phosphate, pH 7.2) at 95 °C. Eluates were incubated for 6 h at room temperature in the absence or presence of 1 unit PNGase F (Roche Molecular Biochemicals). Proteins were then analyzed by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes (Schleicher and Schuell, pore size 0.45 μm) using standard procedures. After transfer, nitrocellulose membranes were incubated for 2 h in PBS containing 5% low fat milk powder and 0.3% Tween 20 and then washed extensively with PBS. Biotinylated proteins were identified with horseradish peroxidase-conjugated NeutrAvidin (Pierce) (diluted in PBS containing 0.1% Tween 20) followed by incubation with the SuperSignal ULTRA chemiluminescent substrate (Pierce). Cell surface immunofluorescence was carried out as described previously (5.Corbeil D. Röper K. Hannah M.J. Hellwig A. Huttner W.B. J. Cell Sci. 1999; 112: 1023-1033Crossref PubMed Google Scholar). Cells grown on glass coverslips were washed with Ca/Mg-PBS, first at room temperature and then on ice, and surface-labeled for 30 min at 4 °C by the addition of the mAb AC133 (10 μg/ml) or mAb AC141 (10 μg/ml) or double surface-labeled by an additional incubation for 10 min with fluorescein isothiocyanate-conjugated wheat germ agglutinin (Sigma). Both antibodies and wheat germ agglutinin were diluted in ice-cold immunofluorescence buffer (Ca/Mg-PBS containing 0.2% gelatin). Unbound antibodies or free wheat germ agglutinin were removed by five washes with ice-cold immunofluorescence buffer. Fixative (3% paraformaldehyde (Sigma) in PBS) was added to the cells on ice, and the coverslips were placed at room temperature for 30 min. The fixative was removed by three washes with Ca/Mg-PBS, and the residual formaldehyde was quenched for 30 min with 0.1 m glycine in PBS. Cells were then incubated for 30 min at room temperature with lissamine rhodamine-conjugated goat anti-mouse IgG (H + L) (Dianova). Coverslips were rinsed sequentially with immunofluorescence buffer, PBS, and distilled water and mounted in Mowiol 4.88 (Calbiochem). In experiments with permeabilized cells, cells grown on glass coverslips were washed with PBS and fixed with 3% paraformaldehyde in PBS for 30 min at room temperature. Coverslips were then rinsed with and incubated for 10 min in PBS containing 50 mm ammonium chloride. Cells were permeabilized and blocked with 0.2% saponin/0.2% gelatin in PBS (blocking solution) for 30 min. Cells were then incubated sequentially with mAb AC133 (10 μg/ml), 1 mg/ml RNase A for 10 min at room temperature (to digest cytoplasmic mRNA), and CyTM2-conjugated goat anti-mouse IgG (H + L) (Dianova), all in blocking solution. Nuclei were labeled with propidium iodide (0.3 μg/ml) during the incubation of secondary antibody. Coverslips were rinsed and mounted as described above. Cells were observed with a Leica TCS4D confocal laser scanning microscope. The confocal microscope settings were such that the photomultipliers were within their linear range. The images shown were prepared from the confocal data files using Adobe photoshop software. A human embryo at 32 days of development was obtained immediately after voluntary termination of pregnancy induced with the RU 486 antiprogestative compound. Informed consent for the use of the embryo in research was obtained from the patient, and the embryo was collected according to the guidelines and with the approval of both national and institutional ethics committees. Human embryonic tissue, fixed in 4% paraformaldehyde (v/v) in PBS for 1 h at 4 °C, was rinsed in PBS for several hours and then twice in 15% sucrose in PBS for at least 24 h. The tissue was then embedded in 15% sucrose, 7.5% gelatin in PBS, and frozen at −70 °C. Five-micron-thick frozen sections were thawed, hydrated in PBS, and endogenous peroxidases were inhibited for 20 min in PBS containing 0.2% hydrogen peroxide. Sections were then washed with 0.25% Triton X-100 in PBS and the mAb AC133 (20 μg/ml) was added overnight at 4 °C. After washing with PBS-Triton X-100, incubation was performed for 1 h at room temperature with biotinylated rabbit anti-mouse antibody (DAKO) and subsequently with peroxidase-coupled streptavidin (DAKO). Peroxidase activity was revealed with 0.025% (v/v) 3.3-diaminobenzidine (Sigma) in PBS containing 0.015% hydrogen peroxide. Slides were counterstained with Harris' hematoxylin, and mounted in XAM neutral medium (BDH). Caco-2 cells or CHO cells stably transfected with the human AC133 gene cultured on collagen-coated 60-mm Petri dishes (Collagen R, Serva) were washed with 200 mmHepes-NaOH, pH 7.4, and fixed with 4% paraformaldehyde (Caco-2 cells) or 8% paraformaldehyde (CHO cells) in Hepes buffer for 4 h at 4 °C. CHO cells were embedded in 10% gelatin in PBS. Ultrathin cryosections were prepared as described previously (5.Corbeil D. Röper K. Hannah M.J. Hellwig A. Huttner W.B. J. Cell Sci. 1999; 112: 1023-1033Crossref PubMed Google Scholar). Immunogold labeling of cryosections using either mAb AC133 (0.5–1.5 mg/ml) or mAb AC141 (0.15–2.3 mg/ml) followed by goat anti-mouse antibody coupled to 15-nm gold particles (Aurion) was performed according to the manufacturer's instruction (Aurion) with minor modifications. Human CD34+hematopoietic progenitor cells were incubated for 45 min at 4 °C with saturating concentrations of AC133 antibody and with an equivalent concentration of an IgG1 isotype-matched nonbinding control antibody, 1B5 (generously supplied by Dr. Graham Mayerhofer, Hanson Center for Cancer Research). After two washes with HHF buffer, the cells were incubated with goat anti-mouse IgG coupled to colloidal gold particles of average diameter of 10 nm (GAM-Au10, Janssen Pharmaceutica) for 1 h at 4 °C with gentle rocking, washed three times with HHF buffer, and then fixed in a 3% (v/v) solution of EM-grade glutaraldehyde (Agar Aids) in 150 mm cacodylate buffer (pH 7.3). Additional aliquots of CD34+ cells were stained in parallel using an identical protocol with mAb DREG-56 (anti-CD62L) and with mAb 7E10 (an antibody prepared in P. J. Simmons' laboratory, which identifies a class 3 epitope on the CD34 molecule (15.Nishio H. Tada J. Hashiyama M. Hirn J. Ingles-Esteve J. Suda T. Kishimoto T. Kikutani H. von dem Borne A.E.G.K. Goyert S.M. Miyasaka M. Moretta L. Okumura K. Shaw S. Springer T.A. Sugamura K. Zola H. Sixth International Workshop and Conference, Kobe, Japan, November 10–14, 1996. Garland Publishing, Inc., New York1997: 974-976Google Scholar)). Immunolabeled cells were then washed three times in cacodylate buffer, post-fixed for 1 h in 1% osmium tetroxide (Johnson Matthey Chemical Ltd.) in the same buffer, and then dehydrated by exposure to a graded series of ethanol. Following dehydration in 100% ethanol, the cells were transferred to propylene oxide (BDH) prior to infiltration with epon-araldite resin overnight at room temperature. The resin was polymerized for 8 h at 80 °C after which ultrathin sections were cut and examined without counterstaining in a transmission electron microscope (Jeol) operating at 80 kV. The qualitative and quantitative statements about the distribution of the various antigens between protrusions and planar regions of the plasma membrane are based on the examination of 80–100 cells and the counting of at least 250 gold particles for each antibody. Both previously described (6.Yin A.H. Miraglia S. Zanjani E.D. Almeida-Porada G. Ogawa M. Leary A.G. Olweus J. Kearney J. Buck D.W. Blood. 1997; 90: 5002-5012Crossref PubMed Google Scholar) mouse mAb against the human AC133 hematopoietic stem cell antigen, mAb AC133 (Fig. 1 A) and mAb AC141 (Fig. 1 B) were found to stain, on cell surface immunofluorescence, Caco-2 cells, a human intestine-derived epithelial cell line (16.Fogh J. Fogh J.M. Orfeo T. J. Natl. Cancer Inst. 1977; 59: 221-226Crossref PubMed Scopus (1353) Google Scholar). The staining showed a punctate pattern, which is characteristic of a microvilli-associated antigen. Caco-2 cells did not show any staining when primary antibodies were omitted (Fig.1 C). The antigen recognized by mAbs AC133 and AC141 on the surface of Caco-2 cells was characterized. Both mAb AC133 (7.Miraglia S. Godfrey W. Yin A.H. Atkins K. Warnke R. Holden J.T. Bray R.A. Waller E.K. Buck D.W. Blood. 1997; 90: 5013-5021Crossref PubMed Google Scholar) and mAb AC141 2D. W. Buck, unpublished data. recognize a glycosylated structure. We therefore labeled the surface of Caco-2 cells with the membrane-impermeant sulfo-NHS-LC-biotin and subjected the antigen immunoprecipitated with mAb AC133 to incubation with and without PNGase F, followed by SDS-polyacrylamide gel electrophoresis and blotting with horseradish peroxidase-conjugated NeutrAvidin (Fig.2 A). Without PNGase F treatment, a band with an apparent molecular mass of ≈120 kDa was detected (Fig. 2 A, lane 1). Deglycosylation yielded a 95-kDa band (Fig. 2 A, lane 2). These apparent molecular masses are those expected for the AC133 antigen (7.Miraglia S. Godfrey W. Yin A.H. Atkins K. Warnke R. Holden J.T. Bray R.A. Wa