Clustering of antigenic determinants on H-2 molecules

The spatial relationship of individual antigenic determinants on H-2Kk and H-2Db molecules was investigated with seven different monoclonal anti-H-2Kk and seven anti-H-2Db antibodies. In these studies the binding of radiolabeled monoclonal anti-H-2 to target cells was competed by addition of various cold anti-H-2 antibodies. The results indicate that on both H-2Kk and H-2Db molecules the antigenic determinants are arranged in two spatially separated clusters. Thus, antibodies to determinants within a cluster show mutual inhibition of binding but do not block the binding of antibodies to the other cluster, and vice versa. Furthermore, in the case of H-2Db antigens it was observed that binding of antibodies to one cluster would considerably enhance the binding of antibodies to the other cluster. A preliminary Scatchard analysis indicated that the enhancing antibody did not alter the affinity of the radiolabeled antibody, but led to an increase of available binding sites on the cell membrane. In addition, binding inhibition studies revealed that the conventional private specificity H-2.2 of H-2Db consists of at least two independent sites on the molecule.     

Endocytosis and recycling of MHC-encoded class II molecules by mouse B lymphocytes

The movements of mouse MHC-encoded class II (H-2E) and class I (H-2K), transferrin receptor and surface Ig molecules of B lymphocytes were studied using radiolabeled mAb and electron microscopy. A total of 10 to 20% of antibodies specific for H-2E molecules were gradually internalized with a t 1/2 of 15 min, reaching a plateau after 30 min at 37 degrees C. Equivalent results were obtained either with the whole antibody or Fab' fragments, suggesting that the internalization of class II molecules was spontaneous. Similar results were obtained with antibodies specific for the transferrin receptor, of which 50% were internalized with t 1/2 of 5 min, reaching a plateau after 30 min. In contrast to antibodies specific for H-2E molecules and the transferrin receptor, antibodies specific for H-2K were not internalized. Reappearance of internalized H-2E-specific antibodies at the cell surface was observed at 37 degrees C. When compared to antibodies specific for surface Ig, degradation of antibodies specific for H-2E molecules was limited even after 5 h incubation. Neither ammonium chloride nor cycloheximide inhibited internalization and recycling. Electron microscopy showed that internalization of H-2E molecules occurred via coated pits/coated vesicles. These results indicate that class II molecules are spontaneously internalized and recycled by B lymphocytes.     

Separation of spleen colony-forming units and prothymocytes by use of a monoclonal antibody detecting an H-2K determinant

The density of H-2K antigens was determined on both the mouse hemopoietic stem cell, using an assay for spleen colony-forming units (CFU-S), and the prothymocyte, using a thymus repopulation assay. This was done by light-activated cell sorting of bone marrow cells labeled first with a biotinylated antibody against H-2Kk and then with avidin-fluorescein isothiocyanate. Almost all CFU-S were found to be present among the 4% bone marrow cells with high forward light scatter (FLS), low perpendicular light scatter (PLS), and bright immunofluorescence. Thymus regeneration by this brightly fluorescent fraction was delayed 3 days compared to thymus regeneration by unsorted cells, although the same number of CFU-S was present in each cell suspension. This delay indicates that differentiation from CFU-S to prothymocytes takes 3 days. The fraction of cells in the FLS/PLS window with dull anti-H-2Kk fluorescence contained few CFU-S and gave rise to a transient thymus regeneration. These findings indicate that the prothymocyte carries fewer H-2K antigens than does the CFU-S. The H-2K antigen is a marker with which CFU-S and prothymocytes can be separated. Therefore, during early T-cell differentiation, the number of H-2K molecules on the cell surface decreases (CFU-S----prothymocyte----cortical thymocyte). During maturation of T cells, a reexpression of H-2K molecules occurs, since lymph node cells and spleen cells were shown to be brightly positive for H-2K antigen.     

Detection of very low receptor numbers on cells by flow cytometry using a sensitive staining method

We describe a staining method for flow cytometry that resolves with a high degree of sensitivity very low numbers of cell surface molecules, which are normally too few to detect using the conventional fluorescein-conjugated reagents. We took advantage of the fact that liposomes can be constructed to contain hundreds of thousands of fluorochrome molecules per vesicle; antigen specificity can be conferred by covalently conjugating them to antibodies or protein A. Unlike fluorochromes such as fluorescein isothiocyanate (FITC) that are directly conjugated to protein ligands with a fluorochrome to protein ratio of about 2 to 1 on the average, their large encapsulating capacity gives liposomes a tremendous potential for signal amplification. In an indirect immunofluorescence study using liposomes that contained the fluorochrome carboxyfluorescein (CF) and that were covalently conjugated to protein A, we were able to obtain up to 50 times the fluorescence signal over background that could be detected with FITC-conjugated protein A. Scatchard analysis showed that the thymoma cell line RDM4 expresses 23,000 and 2,600 binding sites for monoclonal antibodies (mAb) against H-2K and H-2D, respectively. When RDM4 cells were treated with anti-H-2K mAb followed by FITC-conjugated protein A, at best we were able to obtain a fluorescence signal that was only 7 times above background. However, when these cells were treated with the same antibody followed by protein A conjugated to small unilamellar liposomes or large unilamellar liposomes, the fluorescence signals were 110 and 335 times above background, respectively.(ABSTRACT TRUNCATED AT 250 WORDS)     

Interactions of immunoliposomes with target cells

We have covalently attached a monoclonal antibody (11-4.1) against the murine major histocompatibility antigen, H-2Kk, on the surface of liposomes. The interaction of these antibody-coated liposomes (immunoliposomes) with target cells, RDM-4 lymphoma (H-2Kk), was investigated. About 90% of the immunoliposomes taken up by target cells at 4 degrees C could be removed by a mild protease treatment of the cells, whereas only 30% of the uptake at 37 degrees C was labile to the same treatment. Furthermore, the uptake of immunoliposomes at 37 degrees C was inhibitable by cytochalasin B or by a combination of 2-deoxyglucose and NaN3. These results suggest that immunoliposome binding to the target cell surface is the primary uptake event at 4 degrees C and that the surface-bound liposomes are rapidly internalized by the cells at 37 degrees C, probably via an endocytic pathway. Studies with fluorescence microscopy of target cells treated with immunoliposomes containing carboxyfluorescein also supported this conclusion. If endocytosis is the mechanism by which immunoliposomes gain entry into target cells, the efficacy of a cytotoxic drug encapsulated would depend on the resistance of the drug to lysosomal inactivation and its ability to escape from the lysosomal system. Consistent with this notion, we observed that methotrexate encapsulated in liposomes bearing 11-4.1 antibody specifically inhibited deoxy[6-3H]uridine incorporation into DNA in target RDM-4 cells but not in P3-X63-Ag8 myeloma cells (H-2Kd) at the same doses. The observed cytotoxic effect of encapsulated methotrexate could be reversed by the treatment of cells with a lysosomotropic amine, chloroquine, which has been shown to increase the intralysosomal pH of mammalian cells. On the other hand, cytosine-beta-D-arabinofuranoside encapsulated in immunoliposomes showed no target-specific killing, probably because the drug is readily inactivated in the lysosomal system. These results are discussed in terms of the drug carrier potential of immunoliposomes.     

Endocytosis of liposomes and intracellular fate of encapsulated molecules: Encounter with a low pH compartment after internalization in coated vesicles

We have compared the intracellular fate of several fluorescent probes and colloidal gold entrapped in negatively charged liposomes. Weakly acidic molecules (carboxyfluorescein) appear in the cytoplasm of CV-1 cells in 30 min; agents that raise lysosomal pH block this process. Highly charged molecules (calcein) and large molecules (FITC-dextran: 18 kd) remain confined to extra-or intracellular vesicles. Thin section electron micrographs show gold-containing liposomes bound to coated pits, in intracellular coated and uncoated vesicles, and in secondary lysosomes, including dense bodies. Free gold was not observed in the cytoplasm. We conclude that negatively charged liposomes are endocytosed and processed intracellularly by the coated vesicle pathway, and acidification of the endocytic vesicle, rather than liposome fusion, permits escape of certain molecules to the cytoplasm.     

Multiple CTL subsets generated by H-2.L locus products stimulation: evidence for an antigenic determinant private to H-2.Ld

Analysis of the fine specificity of CTL subpopulations raised by an H-2.L locus products stimulation (H-2dm2 anti-H-2d) was performed by absorption experiments by using monolayers of macrophages of H-2m, H-2q, H-2b, and H-2k haplotypes. The results show the existence of four CTL subsets. The pattern of reactivity of three of them could be correlated with that of antibodies present in H-2dm2 anti-H-2d antisera (anti-H-2.64, anti-H-2.65, and anti-H-2.Kk). The fourth CTL subset reacted with a specificity unique to H-2.Ld molecules (a private specificity?), absent on cells from H-2m, H-2q, H-2b, and H-2k haplotypes, and undescribed as yet by serologic methods. These data support the hypothesis that the H-2.L locus products are comparable in their antigenic properties to those of the H-2.K and H-2.D loci.     

Interferon sensitive and insensitive MHC variants of a murine thymoma differentially resistant to methotrexate-containing antibody-directed liposomes and immunotoxin

We evaluated the ability of methotrexate-containing liposomes or a ricin alpha-chain immunotoxin, both associated with monoclonal antibodies specific for the major histocompatibility complex-encoded class I molecule H-2Kk, to kill cells of the murine k haplotype thymoma RDM4. Cells were incubated with liposomes or immunotoxin in the presence or absence of interferon-gamma, which is known to augment the expression of the target class I molecules. The great majority of cells were killed by either of these reagents. Two types of mutant cells were obtained: type 1 cells, selected by methotrexate-containing liposomes, failed to express sufficient target H-2k molecules to be killed by liposomes in the absence of interferon-gamma. In the presence of interferon-gamma, these cells increased expression of all H-2 class I molecules and could be killed by targeted liposomes. Type 2 cells were immunoselected from cloned type 1 cells by liposomes in the presence of interferon. These cells failed to respond to interferon with expression of the H-2Kk molecule, but continued to augment H-2Dk expression in response to interferon. A third variant (type 3) selected from the wild type population by an H-2Kk specific immunotoxin in the absence of interferon phenotypically resembled type 1 cells. Type 1 but not type 2 cells respond to interferon by augmented synthesis of H-2Kk specific mRNA. The results suggest that for interferon-sensitive cell surface molecules of tumor cells, use of interferon improves the efficacy of targeted chemotherapy, but does not prevent development of mutants lacking the target molecule.     

Intracellular transport and localization of major histocompatibility complex class II molecules and associated invariant chain

The intracellular transport and location of major histocompatibility complex (MHC) class II molecules and associated invariant chain (Ii) were investigated in a human melanoma cell line. In contrast to the class II molecules, which remain stable for greater than 4 h after synthesis, the associated Ii is proteolytically processed within 2 h. During or shortly after synthesis the NH2-terminal cytoplasmic and membrane-spanning segment is in some of the Ii molecules cleaved off; during intracellular transport, class II associated and membrane integrated Ii is processed from its COOH terminus in distinct steps in endocytic compartments. Immunocytochemical studies at the light and electron microscopic level revealed the presence of class II molecules, but not of Ii on the cell surface. Intracellularly both Ii and class II molecules were localized in three morphologically and kinetically distinct compartments, early endosomes, multivesicular bodies, and prelysosomes. This localization in several distinct endosomal compartments contrasts with the localization of class II molecules in mainly one endocytic compartment in B lymphoblastoid cell lines. As in these lymphoblastoid cell lines Ii is known to be rapidly degraded it is conceivable that the rate of proteolysis of the class II associated Ii and its dissociation from class II molecules modulates the retention of the oligomeric complex in endocytic compartments, and as a consequence the steady-state distribution of these molecules within the endosomal system.     

Cytotoxic T-lymphocyte tolerance to minor H-43a alloantigen is induced exclusively in the context of the self major histocompatibility complex class I H-2Kb molecules

We elucidated previously that cytotoxic T lymphocyte precursors (CTLp) against H-43a allo-antigen, which we had discovered as a new mouse minor H antigen, were primed in H-43b mice only in the context of self H-2Kb restriction element, and that anti-H-43a CTLp tolerance was induced in H-43b mice by injection with H-43a spleen cells (SC) from H-43 congenic mice, i.e., under the condition of disparity at only the H-43 locus. The present study attempted to determine whether the H-2Kb restriction element for anti-H-43a CTLp priming is also implicated in the induction of anti-H-43a CTLp tolerance. For this purpose, we used a newly established H-43b C3W (H-2k) strain which is H-43 congenic to H-43a C3H/HeN. When (C3W X B10.MBR)F1 (H-43b, H-2Kk/b, Ik/k, Dk/q) mice were injected with H-43a-bearing (C3H/HeN X B10.AKM)F1 (H-43a/b;H-2Kk/k,Ik/k,Dk/q)SC, their selfH-2Kb-restricted anti-H-43a CTLp were were primed (cross-priming). By contrast, injection of H-43a-bearing (C3H/HeN X B10.MBR)F1 (H-43a/b; H-2Kk/b,Ik/k, Dk/q)SC, which differ from (C3H/HeN x B10.AKM) F1 SC solely at H-2K and possess H-2Kb molecules, did not prime but specifically inactivated the anti-H-43a CTLp of (C3W x B10.MBR)F1 mice. These results indicate clearly that anti-H-43a CTLp tolerance is induced exclusively in the context of the H-2Kb element expressed on the antigenic H-43a SC.     

Association between H-2 and vaccinia virus-induced antigens on the surface of infected cells.

The relationship between H-2 molecules and vaccinia virus-induced antigens on the surface of H-2d infected cells was investigated by the differential redistribution method and by the blocking capacity of monospecific anti-H-2 sera on an anti-vaccinia cell-mediated cytotoxicity (CMC). Capping of either H-2K or H-2D molecules upon addition of monospecific and anti-H-2 sera was followed by the complete redistribution of viral antigens, suggesting the formation, on the cel membrane, of complexes of H-2K, H-2D molecules and vaccinia virus-induced antigens. However, not all H-2 molecules were involved in this association since i) free H-2K and H-2D molecules still moved independently on the cell surface, and ii) capping of vaccinia virus-induced antigens failed to induce the redistribution of all the H-2K and H-2D molecules. In addition, either monospecific anti-H-2K or anti-H-2D antiserum was found to exert potent blocking activity on anti-vaccinia CMC, indicating also a close topographical relationship between H-2K, H-2D molecules and vaccinia virus-induced antigens.     

Lateral diffusion and patch formation of H-2Kk antigens on mouse spleen lymphocytes

We have studied the diffusion and aggregation of H-2Kk antigens labeled with a fluorescent anti-H-2Kk monoclonal antibody (IgG) on mouse splenic lymphocytes, employing fluorescence photobleaching recovery and fluorescence microscopy. The H-2Kk antigens were initially distributed homogeneously on all lymphocytes. Upon antibody binding, sub-micron patches were formed on 50-60% of the cells. A lateral diffusion coefficient, D, of 7.1 X 10(-10) cm2/s and a mobile fraction of 0.73 were found for H-2Kk antigens on diffusely-labeled cells, while these antigens were immobile (D less than or equal to 5 X 10(-12) cm2/s) on patched cells. The patched and nonpatched sub-populations did not correspond to B- and T-lymphocytes. Subjection to low temperature or treatment with NaN3 or cytoskeleton-disrupting drugs did not affect the diffusion or patching of H-2Kk, indicating no involvement of metabolic energy or drug-sensitive cytoskeletal components. These findings could be related to the interactions of H-2 antigens on the cell surface, and to the different susceptibilities of various cells to lysis by cytotoxic T-cells.     

Acid intracellular vesicles and the cytolysis of mammalian target cells by Entamoeba histolytica trophozoites

Entamoeba histolytica kills mammalian target cells in a multi-step sequential process with separate adherence, cytolytic, and phagocytic events. In the studies reported here, we used fluorescein isothiocyanate linked to dextran to label the endocytic vesicles of the HM1 strain of E. histolytica and measure vesicle pH (5.1 +/- 0.2 by spectrofluorimetry). Concentrations of NH4Cl (1.0-10.0 mM) sufficient to increase vesicle pH to greater than or equal to 5.7 inhibited amebic killing of target Chinese hamster ovary (CHO) cells as assayed by trypan blue staining, by the release of 3H-thymidine previously incorporated into CHO cell monolayers, and by the release of 111indium oxine from radiolabeled CHO cells. Similar effects were also observed with two other weak bases, primaquine and chloroquine (both 50 microM). In contrast, NH4Cl (10 mM) did not affect either the adherence or phagocytic events, as measured by amebic adherence to CHO cells at 4 degrees C and by the binding and ingestion of 3H-leucine-labeled bacteria. In the presence of NH4Cl and the carbohydrate ligand asialofetuin, there was no evidence of intracellular trapping of the amebic galactose-inhibitable lectin; inhibition of adherence by cycloheximide (10 micrograms/ml for 3 h) suggested rapid turnover of the surface lectin. Prolonged exposure to NH4Cl for 48 h (which had no effect on amebic protein synthesis) or shorter exposure to cycloheximide (10 micrograms for 3 h) produced persistent inhibition of cytolysis. These results indicate that an uninterrupted acid pH in intracellular endocytic vesicles is necessary for the cytolysis of target cells by E. histolytica trophozoites.     

Weak base amines can inhibit class I MHC-restricted antigen presentation

This report describes the effects of NH4Cl, CH3NH2, and chloroquine on class I and II MHC-restricted Ag presentation. OVA-specific T-T hybridomas were used to detect processed OVA in association with class I, H-2Kb, and class II, I-Ad/b, molecules on a B lymphoblastoid APC. OVA, internalized by APC under hypertonic conditions, was presented in association with class I and II MHC molecules. Treating the APC with NH4Cl or CH3NH2 inhibited class I- and II-restricted Ag presentation. In contrast, chloroquine markedly inhibited class II, but not class I-restricted Ag presentation. Controls indicated that drug-treated APC were fully competent to interact with T cells and present processing-independent antigenic peptides in association with both class I and II MHC molecules. NH4Cl and CH3NH2 did not inhibit the uptake of radiolabeled Ag by the APC. After the proteolytic removal of H-2Kb from the surface of APC, NH4Cl and CH3NH2-treated and control APC regenerated identical amounts of surface H-2Kb and this regeneration required de novo protein synthesis. These latter results indicate that NH4Cl and CH3NH2 can inhibit Ag presentation without affecting the synthesis, transport, or surface expression of H-2Kb. Also, NH4Cl did not affect the transport of H-2Db to the surface of mutant RMA-S cells that were cultured with exogenous peptides. Taken together these results strongly suggest that NH4Cl and CH3NH2 but not chloroquine can inhibit a critical and early intracellular step in class I-restricted Ag presentation while simultaneously inhibiting class II-restricted Ag presentation.     

The mechanism of antigen presentation by dendritic cells and splenic adherent cells in the induction of an allogeneic cytotoxic T-lymphocyte response to H-2Kk liposomes

Induction of an allogeneic cytotoxic T-lymphocyte (CTL) response to purified alloantigen is partially dependent on uptake and processing of the class I alloantigen by antigen-presenting cells (APC) followed by recognition of the alloantigen and self Ia by helper T cells (TH). The activated TH provides the helper signal(s) to the alloantigen-specific CTL for proliferation and differentiation into an active effector CTL. The role of antigen processing and presentation of major histocompatibility complex alloantigens was examined and the ability of different types of APC to present purified H-2Kk liposomes was investigated. Splenic adherent cells (SAC), splenic dendritic cells (DC), and B-cell lymphoblastoid lines were all shown to be effective in the presentation of H-2Kk liposomes. The relative ability of these cells to serve as APC was determined to be DC greater than B-cell tumors greater than SAC. The role of processing of H-2Kk liposomes by SAC and DC was examined by investigating the effect of weak bases on pulsing of the APC. These experiments suggest that presentation of alloantigen by both SAC and DC involves a step which is sensitive to inhibition by weak bases. We examined whether the TH were activated by similar mechanisms when stimulated by the various APC. The functional involvement of the T-cell surface marker L3T4 was demonstrated in the induction of TH. In contrast, L3T4 was not involved in the subsequent generation of CTL since monoclonal antibody (MAb) specific for L3T4 was not effective in blocking CTL function in the presence of nonspecific T helper factor (THF). Similarly, Ia on the APC was shown to be involved in the stimulation of the TH pathway but not directly in the differentiation of the CTL. Thus, DC and B cells in addition to SAC can present H-2Kk to TH. The presentation of alloantigen by both cell types may involve an intracellular route as demonstrated by the blocking of the TH response by weak bases. Both Ia and L3T4 are required on the APC for induction of the TH response. The minimal requirements for activation of the CTL were H-2Kk liposomes and a source of THF.     

Constitutive internalization of murine MHC class I molecules

The total number of cell surface glycoprotein molecules at the plasma membrane results from a balance between their constitutive internalization and their egress to the cell surface from intracellular pools and/or biosynthetic pathway. Constitutive internalization is net result of constitutive endocytosis and endocytic recycling. In this study we have compared spontaneous internalization of murine major histocompatibility complex (MHC) class I molecules (K(d), D(d), full L(d), and empty L(d)) after depletion of their egress to the cell surface (Cycloheximide [CHX], brefeldin A [BFA]) and internalization after external binding of monoclonal antibody (mAb). MHC class I alleles differ regarding their cell surface stability, kinetics, and in the way of internalization and degradation. K(d) and D(d) molecules are more stable at the cell surface than L(d) molecules and, thus, constitutively internalized more slowly. Although the binding of mAbs to cell surface MHC class I molecules results in faster internalization than depletion of their egress, it is still slow and, thereby, can serve as a model for tracking of MHC class I endocytosis. Internalization of fully conformed MHC class I molecules (K(d), D(d), and L(d)) was neither inhibited by chlorpromazine (CP) (inhibitor of clathrin endocytosis), nor with filipin (inhibitor of lipid raft dependent endocytosis), indicating that fully conformed MHC class I molecules are internalized via the bulk pathway. In contrast, internalization of empty L(d) molecules was inhibited by filipin, indicating that non-conformed MHC class I molecules require intact cholesterol-rich membrane microdomains for their constitutive internalization. Thus, conformed and non-conformed MHC class I molecules use different endocytic pathways for constitutive internalization.     

The ability of the murine placenta to absorb monoclonal anti-fetal H-2K antibody from the maternal circulation.

In previous work we have found that partially purified 125I-labeled anti H-2 antibody is localized in the placenta when injected i.v. into females pregnant by males bearing the target haplotype. This led to the concept that the placenta is an H-2 antigen-bearing immunoabsorbent barrier between mother and fetus. In this report we have used an anti-H-2Kk monoclonal antibody of the IgG2a subclass, also labeled with 125I, to verify this concept, as well as to improve the genetic definition of the immunoabsorbent antigen. In addition we have prepared F(ab')2 fragments of this antibody, and these also show the immunoabsorbent effect. This indicates that transport into fetally derived tissues via Fc binding is not a prerequisite for reaction of the antibody with paternal strain H-2 K antigens.     

Defining cytolytic T lymphocyte recognition of chemically modified self. I. Response to trinitrophenyl-H-2Kk

We used purified class I antigen incorporated into liposomes to examine the response of secondary cytolytic T lymphocytes (CTL) to chemically modified self. By generating the secondary response in the presence of T cell helper factor, the level of CTL response was limited by CTL recognition of added antigen rather than by helper cell generation of lymphokines. We found a strong secondary response against chemically modified self when spleen cells from trinitrophenyl (TNP)-primed C3H/HeJ mice were stimulated with a) TNP-modified liposomes containing H-2Kk, b) liposomes containing H-2Kk purified from TNP-modified RDM-4 (H-2k) cells, or c) liposomes containing the limited trypsin proteolysis product of H-2Kk that had been directly modified with TNP. In contrast, we were not able to generate a significant CTL response with unmodified H-2Kk incorporated into vesicles along with TNP-modified membrane components lacking H-2Kk. These results suggest that TNP-modified H-2Kk is a major antigenic site recognized by CTL from C3H/HeJ mice after priming against TNP-modified self.     

Endocytosis in filter-grown Madin-Darby canine kidney cells

In this paper, we have characterized the apical and basolateral endocytic pathways of epithelial MDCK cells grown on filters. The three- dimensional organization of the endocytic compartments was analyzed by confocal microscopy after internalization of a fluorescent fluid-phase marker from either side of the cell layer. After 5 min of internalization, distinct sets of apical and basolateral early endosomes were observed lining the plasma membrane domain from which internalization had occurred. At later time points, the apical and the basolateral endocytic pathways were shown to converge in the perinuclear region. Mixing of two different fluorescent markers could be detected after their simultaneous internalization from opposite sides of the cell layer. The extent of the meeting was quantitated by measuring the amount of complex formed intracellularly between avidin internalized from the apical side and biotinylated horseradish peroxidase (HRP) from the basolateral side. After 15 min, 14% of the avidin marker was complexed with the biotinylated HRP and this value increased to 50% during a subsequent chase of 60 min in avidin-free medium. We also determined the kinetics of fluid internalization, recycling, transcytosis, and intracellular retention using HRP as a marker. Fluid was internalized with the same rates from either surface domain (1.2 x 10(-4) microns 3/min per microns 2 of surface area). However, significant differences were observed for each pathway in the amounts and kinetics of marker recycled and transcytosed. The content of apical early endosomes was primarily recycled and transcytosed (45% along Bach route after 1 h internalization), whereas delivery to late endocytic compartments was favored from the basolateral early endosome (77% after 1 h). Our results demonstrate that early apical and basolateral endosomes are functionally and topologically distinct, but that the endocytic pathways converge at later stages in the perinuclear region of the cell.