Search

The binding and internalization of a model lysosomal enzyme, beta-galactosidase, was visualized by use of rabbit anti-beta-galactosidase and goat anti-rabbit IgG; the second antibody was labeled with rhodamine or fluorescein (for detection by fluorescence) or with horseradish peroxidase (for electron microscopy). Chinese hamster ovary cells were incubated with beta-galactosidase at 4 degrees C, and then were washed and sequentially incubated in the cold with the two antibodies. The beta-galactosidase was found primarily in coated pits. The binding of the enzyme was completely inhibited by 5 mM mannose 6-phosphate. After the reaction with enzyme and antibodies, the cells were warmed to 37 degrees C; within 1 minute, the beta-galactosidase--antibody complex had begun to move to uncoated vesicles (receptosomes). After 8 min, the beta-galactosidase--antibody complex was seen in receptosomes near tubular elements in the Golgi/GERL area, within such tubular elements and at times, in vesicular elements that may correspond to coated structures of the GERL system. After 15 min, the enzyme--antibody complex was found in lysosomes near the Golgi/GERL are and a half-hour later it was in lysosomes distributed throughout the cytoplasm. Double-label experiments using beta-galactosidase and gold/alpha 2-macroglobulin showed the presence of the two ligands in the same coated pits and receptosomes. Thus, the pathway for internalization of beta-galactosidase via the mannose 6-phosphate receptor is similar to the pathway established for other ligands such as low density lipoprotein and alpha 2-macroglobulin.

A tumor-specific transplantation antigen has been purified to homogeneity from the cytosol of a methylcholanthrene-induced tumor, Meth A. The purified antigen is highly immunogenic and specific against challenge with Meth A, providing greater than 95% inhibition of tumor growth in immunized syngeneic mice. Immunofluorescence analysis of Meth A showed that the antigen is a highly abundant cytosolic protein but that it is also present at the cell surface and, therefore, accessible to the host's immune system. The antigen consists of two polypeptide isoforms present in equimolar amounts, having similar masses (84 and 86 kDa), pI values (4.95 and 4.90), and amino acid compositions. Both are phosphoproteins, and neither is glycosylated. The NH2-terminal sequences of the two isoforms are identical except that each chain contains a portion of unique sequence. Comparison of the NH2-terminal and CNBr-fragment sequence data to the sequences of the yeast and Drosophila heat shock proteins (Hsp90 and Hsp83, respectively) reveals that 73 of 91 residues compared are identical. In addition, an anti-Meth A tumor antigen serum that defects the isoforms from a variety of tumors also immunoprecipitates proteins of identical mass and pI from both normal and heat-shocked mouse embryo cells.

Cytotoxic conjugates were constructed by chemically coupling Pseudomonas exotoxin to antitransferrin receptor antibodies. Toxicity of these conjugates, due to entry via the transferrin receptor, was enhanced 100- to 300-fold in the presence of adenovirus. By electron microscopy and immunofluorescence it was determined that antitransferrin receptor antibodies and the conjugates derived from them entered cells from coated pits into receptosomes. We believe that the enhanced toxicity resulted when adenovirus and the toxin conjugates were internalized into the same receptosomes. In the process of infection, adenovirus enters cells and brings about a virus-mediated disruption of receptosomes; and this disruption can liberate many more toxin molecules into the cytosol than is possible in the absence of virus.

4 beta-Phorbol 12-myristate 13-acetate (PMA) treatment of KB cells at 37 degrees C rapidly induces a 50% reduction in epidermal growth factor (EGF) binding that is maximal by 30 min. EGF binding activity returns to the original value by 1 hr and remains constant for 2 hr thereafter. Using a polyclonal antibody directed against the cytoplasmic domain of the EGF receptor (EGF-R), we examined the fate of the receptor after PMA treatment. Immunofluorescent and electron microscopic localization of the EGF-R after PMA treatment demonstrated that about 50% of the receptor became internalized into endocytic vesicles (receptosomes) and Golgi-associated structures. Unlike EGF-induced internalization, PMA-induced internalization did not cause delivery of EGF-R to lysosomes or receptor degradation. Rather, receptor reappeared on the cell surface. No stimulation of EGF-R synthesis was observed after 1 hr of PMA treatment. Loss of cell surface binding correlated with the internalization of the EGF-R observed morphologically. A possible explanation for these observations is that PMA, an activator of protein kinase C, confers a signal sufficient for EGF-R clustering and internalization but not for transport to lysosomes.

We have used electron microscopic immunocytochemistry to compare the distribution of LAMP-1, a marker for lysosomal membranes, with the intracellular localization of alpha 2-macroglobulin (alpha 2-M) and transferrin at various time points after their endocytosis into cultured NIH 3T3 cells. The purposes of this study were (a) to determine how soon endocytic ligands reach lysosomal organelles, (b) to examine whether the intermediate endocytic vesicles gained lysosomal markers gradually or in a precipitous, discrete event, and (c) to examine the relationship, if any, between the pathway of recycling ligands and lysosomes. At early time points (0-5 min) after initiation of endocytosis, most structures containing alpha 2-M labeled with colloidal gold (receptosomes) were not labeled by anti-LAMP-1 detected using ferritin bridge or peroxidase immunocytochemistry. At late time points (greater than or equal to 15 min), the structures containing alpha 2-M (lysosomes) were strongly labeled by anti-LAMP-1. In contrast, transferrin that was directly labeled with ferritin was mostly located in LAMP-1-negative structures at all time points studied. The proportion of alpha 2-M-gold containing vesicles strongly labeled for LAMP-1 roughly paralleled the proportion of alpha 2-M-gold-containing structures positive for cytochemically detectable acid phosphatase. Our data indicate that ligands such as transferrin that are internalized through coated pits and receptosomes, but not delivered to lysosomes, do not traverse a lysosomal organelle compartment as marked by LAMP-1 content. Ligands such as alpha 2-M that are destined for lysosomal delivery reach a LAMP-1-positive organelle compartment only after they traverse LAMP-1-negative, non-lysosomal vesicles previously described as receptosomes.

A monoclonal antibody, OVB1, was generated against a human ovarian carcinoma cell line, OVCAR-3. The antigen reacting with this antibody was strongly expressed on the external surface of the plasma membrane of OVCAR-3 cells and cells of 4/4 other ovarian carcinoma lines. Variable density and homogeneity of expression was found on cells from 5/5 breast carcinoma lines. Various ovarian tumor specimens and normal human tissues were frozen, cryostat-sectioned, and examined for OVB1 reactivity using immunoperoxidase methods. A strong, uniform, homogeneous reaction on 10/10 ovarian carcinoma specimens and variable, non-homogeneous reactions on breast tumors were seen. Normal tissues reacting with the antibody include thyroid, pituitary pars intermedia, breast ductal epithelium, Auerbach's plexus and neuronal processes in the GI tract, colonic mucosal epithelium, and salivary gland ductal epithelium. Polymorphonuclear leukocytes, eosinophils, and approximately 13% of peripheral lymphocytes, as well as cells around germinal centers in lymph nodes and spleen, showed strong reactivity by immunofluorescence and/or immunoperoxidase. Expression of the OVB1 antigen in the myeloid cells of normal human bone marrow occurred from the promyelocyte stage through to more mature cells in a subpopulation of myeloblasts. Indirect immunofluorescence of live peripheral blood cells showed localization to the surface of PMNs, eosinophils, and certain lymphocytes. Double-immunofluorescence studies (with a direct fluorescein-anti-lactoferrin antibody conjugate) showed co-localization of OVB1 and OKM1 (anti-C3bi receptor) antibodies to specific granules of PMNs. Localization of OVB1 and OKM1 antibodies to granular structures in the PMN was confirmed by electron microscopy using the ferritin bridge technique. The antigen reacting with the OVB1 antibody was shown to be neuraminidase sensitive, but protease insensitive. The OVB1 monoclonal antibody may be useful in identification of ovarian tumors and subclassification of myeloid leukemias.

Double-label fluorescence microscopy was used to demonstrate the efflux activity of the multidrug transporter in single cultured cells. NIH3T3 cells expressing a transfected MDR1 gene (NIH3T3-MDR) were treated with vinblastine or daunomycin. The accumulation of vinblastine was monitored by examining the morphology of tubulin in cells, using immunofluorescence. Overnight treatment of drug-sensitive cells caused disassembly of microtubules and formation of paracrystals; the absence of vinblastine effects was evident by the presence of intact microtubules. Daunomycin accumulation was detected in nuclei using the inherent fluorescence of the drug with rhodamine epifluorescence microscopy. Drug efflux in multidrug-resistant cells was inhibited with verapamil. When multidrug-resistant cells were treated overnight in vinblastine, an effect of 0.5 microM vinblastine on microtubules was seen only in the presence of verapamil. Similarly, when cells were treated with daunomycin, this drug accumulated in nuclei only when verapamil was present. When cells incubated with vinblastine and verapamil were washed free of drugs, they did not accumulate daunomycin in a subsequent incubation, indicating that the multidrug transporter was still active; this occurred even though the morphologic effects of vinblastine persisted. Cells incubated with vinblastine alone showed an immediate inhibition of efflux activity when verapamil was subsequently added with daunomycin. These results show that the efflux activity of the multidrug transporter can be rapidly manipulated by agents such as verapamil, despite a prior history of drug treatment, and that the effects of inhibition of the transporter are rapidly reversible.

The subcellular distribution of sialic acid was determined at the ultrastructural level using Limax flavus agglutinin (LFA). This lectin, which is specific for N-acetylneuraminic acid and N-glycolylneuraminic acid, was covalently conjugated to horseradish peroxidase (HRP). The conjugates (LFA-HRP) were applied to aldehyde-fixed, saponin-permeabilized 3T3 cells in pre-embedding labeling electron microscopy. Peroxidase label was detected in a patchy distribution at the cell surface, and in plasma-membrane-coated pits, endocytic vesicles (receptosomes), multivesicular bodies, and lysosomes. Smooth-surfaced tubular and vesicular structures, similar to those that participate in membrane recycling, were labeled. In the Golgi complex, more than half of the cisternae contained label--typically only one cisterna on the cis side was unlabeled. Heavily labeled structures of the trans Golgi included a reticular membranous system with coated regions--50-80 nm diameter vesicular or pit-like profiles and larger coated vacuoles. Smooth 200-300 nm vacuoles were labeled on the trans side of the Golgi stack. Similar structures have been previously shown to participate in the exocytosis of plasma membrane and secretory glycoproteins from the Golgi stacks. These findings identify those intracellular organelles that are functionally at the level of, or distal to, the sialyltransferase-containing membranes of the Golgi, and distinguish them from the pre-Golgi membranous structures. The LFA-HRP conjugate is an indicator for this functional trans domain of the cell, and should be applicable for ultrastructural double-label experiments as a cis versus trans marker of the exocytic pathway.

Concentrative receptor-mediated endocytosis of many specific ligands by cultured fibroblasts occurs through the coated pit-receptosome pathway. The formation of receptosomes was studied using two impermeant electron-dense labels for the cell surface, ruthenium red and concanavalin A-horseradish peroxidase. These studies show that at 4 degrees C, virtually all coated structures near the plasma membrane are in communication with the cell surface, and are not isolated coated vesicles. On warming cells to 37 degrees C for only 1 minute, a major portion of these structures become cryptic, that is, not labeled by these surface markers. However, on cooling cells immediately back to 4 degrees C, virtually all of these structures are again in communication with the surface. Many images showed that membrane of these cryptic pits to be continuous with the cell surface when caught in the appropriate plane of section; often there was a very narrow entrance that excluded extracellular label. At 37 degrees C, receptosomes could be occasionally seen forming as an invagination of membrane adjacent to the coated region. Mechanisms by which receptosomes may form and other evidence demonstrating the failure of coated pits to pinch off to form isolated coated vesicles during endocytosis are discussed.

Using an antibody prepared against performic acid-treated calmodulin, we have localized calmodulin in cultured fibroblastic cells by immunofluorescence and immunoelectron microscopy. In interphase cells, calmodulin was found to be diffusely distributed throughout the cytosol. An increased amount of calmodulin was found in the pericentriolar region of interphase cells. No significant aggregation of calmodulin was found in association with microfilaments, peripheral cytoplasmic microtubules or clathrin-coated structures. Calmodulin was present in moderate amounts in microvilli, ruffles, and zeiotic blebs of the cell surface. In motitic cells, calmodulin was found concentrated in the pericentriolar region, and appeared to concentrate along radiating spindle microtubules proximal to the centrioles. Redistribution of calmodulin was seen between early and late telophase, in which the pericentriolar pattern of calmodulin in early telophase shifted to an aggregation on the intercellular bridge, with a large part of the midbody portion of the bridge being devoid of calmodulin. These results show that calmodulin is distributed throughout the cytosol, but is markedly concentrated in the region of the microtubule organizing center in interphase cells, as well as in elements of the mitotic spindle apparatus. This distribution suggests that calmodulin has a regulatory role in the organization and function of microtubules during interphase, as well as during mitosis.

We performed genetic engineering to fuse enhanced green fluorescent protein (EGFP) to the N terminus of RTA, expressed the fusion protein in Escherichia coli, purified and reassociated EGFP−RTA with plant RTB, and purified EGFP−ricin by size exclusion HPLC. The fusion heterodimer was able to bind galactosides, intoxicate cells, and show strong fluorescence. Mammalian cells incubated with EGFP−ricin showed strong cell surface fluorescence at 4 °C and, on incubation at 37 °C, distributed initially to endosomes and then to Golgi vesicles. Variable sensitivity of mammalian cells to ricin and ricin fusion proteins may be due in part to different patterns of intracellular routing. Cells were incubated with ricin or EGFP−ricin, and inhibition of protein synthesis was measured. Human hepatocellular carcinoma Hep3B cells were 10-fold more sensitive to ricin and 85-fold more sensitive to EGFP−ricin than human epidermoid carcinoma KB cells. Epifluorescence microscopy of cells incubated with EGFP−ricin showed greater localization of the fluorescence signal in the Golgi compartments in Hep3B cells than in KB cells. These data support a model requiring a Golgi-dependent step in cell intoxication by ricin. The work further identifies the usefulness of green fluorescent protein fusions in the study of retrograde transport of internalized peptides.

Ricin toxin, the heterodimeric 65 kDa glycoprotein synthesized in castor bean seeds, contains a cell binding lectin subunit (RTB) disulfide linked to an RNA N-glycosidase protein synthesis-inactivating subunit (RTA). Investigations of the molecular nature of the lectin sites in RTB by X-ray crystallography, equilibrium dialysis, chemical modification, and mutational analysis have yielded conflicting results as to the number, location, and affinity of sugar-combining sites. An accurate assessment of the amino acid residues of RTB involved in galactose binding is needed both for correlating structure−function of a number of plant lectins and for the design and synthesis of targeted toxins for cancer and autoimmune disease therapy. We have performed oligonucleotide-directed mutagenesis on cDNA encoding RTB and expressed the mutant RTBs in insect cells. Partially purified recombinant proteins obtained from infected cell supernatants and cell extracts were characterized as to yields, immunoreactivities, asialofetuin binding, cell binding, ability to reassociate with RTA, and recombinant heterodimer cell cytotoxicity. Two single-site mutants (subdomain 1α or 2γ) and two double-site mutants (subdomains 1α and 2γ) were produced and studied. Yields varied by two logs with lower recoveries of double-site mutants. All the mutants showed immunoreactivity with a panel of anti-RTB monoclonal and polyclonal antibodies. Single-lectin site mutants displayed up to a 1 log decrease in asialofetuin binding avidity, while the double-site mutants showed close to a 2 log decrease in sugar binding. However, for each of the double-site mutants, residual sugar binding was demonstrated to both immobilized asialofetuin and cells, and this binding was specifically inhibitable with α-lactose. All mutants reassociated with RTA, and the mutant heterodimers were cytotoxic to mammalian cells with potencies 1000-fold or more times that of unreassociated wild-type RTA or RTB. These data support a model for three or more lectin binding subdomains in RTB.

DNA encoding ricin B chain was modified by site-directed mutagenesis, and eight separate mutant RTB cDNAs including four novel mutants were ligated into the baculovirus transfer vector, pAcGP67A. Cotransfection of S. frugiperda Sf9 cells with BaculoGold DNA was followed by limiting dilution isolation of recombinant baculoviruses. Infection of Sf9 cells at a multiplicity of infection of 5 in the presence of 25 mM lactose produced 0.05−1 mg/L of soluble, glycosylated 34 kDa proteins immunoreactive with monoclonal and polyclonal antibodies to ricin B chain. Mutant ricin B chains were partially purified by monoclonal antibody immunoaffinity chromatography to 10−50% purity in near milligram quantities. The mutant ricin B chains had decreased lectin binding relative to plant ricin B chain as measured by binding to immobilized lactose and asialofetuin and cell binding immunofluorescence. The mutant ricin B chains reassociated with plant RTA similarly to plant RTB, and the recombinant heterodimers had slightly reduced cell cytotoxicity relative to ricin.

The destruction of large pinosomes was examined with phase-contrast microscopy in cultured mouse fibroblasts. In areas of rapid pinosome breakdown, lysosomes were observed to repeatedly collide with pinosomes without fusing, tearing off small pieces until the pinosomes became smaller and denser. This segmentation of pinosomes by lysosomal collision has been named "piranhalysis."

Using the direct conjugate of epidermal growth factor (EGF) and horseradish peroxidase, we have followed the entry of EGF into KB (human carcinoma) cells. EGF initially was found bound diffusely to the entire cell surface at 4 degrees C; on warming to 37 degrees C, EGF was found clustered in clathrin-coated pits on the plasma membrane in 1 min or less. Within 1-2 min at 37 degrees C, EGF began to accumulate in receptosomes within the cell and remained there for up to 10 min. At 10-13 min after warming to 37 degrees C, EGF was found in thin reticular membranous elements of the Golgi system, as well as concentrated in the clathrin-coated pits present on these membranes. By 15 min after warming, EGF began to be delivered to lysosomes located near the Golgi system. These findings suggest that clathrin-coated pits in the Golgi reticular system accumulate EGF before delivery to lysosomes.

35S-Labeled adenovirus type 2 (Ad2) (10 ng/ml) was incubated with 1% Triton X-114 at various pH values varying from 3.0 to 8.0. The detergent phase was separated from the aqueous phase by centrifugation, and the amounts of Ad2 were determined in the two phases. At pH 7.0-8.0, less than 5% of Ad2 was associated with the detergent phase; at pH 5.0 or below, about 60% of Ad2 was associated with the detergent phase. When a mixture of 35S-labeled capsid proteins was used at pH 7.0, 60-70% of the total proteins were associated with the detergent at pH 5.0, but less than 5% of the proteins interacted with detergent at pH 7.0. Among the three major external proteins (hexon, penton base, and fiber), penton base had the highest association with Triton X-114 at pH 5.0. Both intact virus and the capsid proteins that were associated with Triton X-114 at pH 5.0 were released into the aqueous phase on subsequent incubation at pH 7.0. On the basis of these results, it is suggested that mildly acidic pH induces amphiphilic properties in adenovirus capsid proteins and may help Ad2 escape from acidic endocytic vesicles.

Selective enrichment of clathrin-coated membranes by anticlathrin immunoadsorption was used to examine the internalization of receptor-ligand complexes through coated pits. Using Staphylococcus aureus-anticlathrin antibody and [35S]methionine-labeled KB cells, the kinetics of association of the epidermal growth factor (EGF-R) and transferrin receptors (TF-R) with coated membranes were directly examined. The accumulation of EGF-R in coated pits at the cell surface was dependent upon EGF binding. EGF-R then passed sequentially through a compartment which did not react with anticlathrin antibody and a second clathrin-coated compartment. The EGF-R was degraded in lysosomes with a half-life of approximately 41-55 min. The tumor promoter, 4 beta-phorbol 12-myristate 13-acetate, appears to mimic the action of EGF in inducing EGF-R accumulation in coated pits at the cell surface and receptor internalization. In contrast to the results with EGF-R, the TF-R was found in clathrin-coated membranes in the presence or absence of TF, and the concentration of TF-R in clathrin-coated membranes did not significantly change with time. The method presented should be of great utility for examining the biochemical changes that occur during the receptor-mediated endocytosis and sorting of ligands and receptors.

An antibody against the human epidermal growth factor receptor (EGF), capable of activating its tyrosine kinase has been produced. Antibody 2913 recognizes only the cytoplasmic portion of the EGF receptor in A431 carcinoma cells, in normal human fibroblasts, and in a variety of other human tumor cell lines (Xu, Y.-A., Richert, N., Ito, S., Merlino, G. T., and Pastan, I. (1984) Proc. Natl. Acad. Sci. U. S. A. 81, 7308-7313). Indirect immunofluorescence and electron microscopy show that the antibody binds to intact cells only after membrane permeabilization. Moreover the antibody immunoprecipitates the v-erb-B gene product in avian myeloblastosis virus-infected cells but does not recognize the secreted form (105 kDa) of the A431 cell EGF receptor which lacks the cytoplasmic domain. Antibody 2913 activates the EGF receptor kinase in solubilized A431 membranes causing autophosphorylation on tyrosine residues only. Tryptic peptide maps suggest that antibody 2913 and EGF stimulate phosphorylation of the same amino acid residues. By electron microscopy, the cytoplasmic portion of the receptor was followed throughout its endocytotic pathway. The results show that the kinase domain is rapidly degraded in lysosomes with no accumulation in the cytoplasm or in the nucleus.

LysPE40 is a modified form of Pseudomonas exotoxin that lacks the cell-binding domain and has a chemically reactive lysine residue near the amino terminus. LysPE40 is made in Escherichia coli and secreted into the medium from which it is readily purified. Two immunotoxins were constructed by coupling LysPE40 to an antibody to the human transferrin receptor (TFR) or to an antibody to the human interleukin-2 receptor. These immunotoxins were selectively cytotoxic to receptor-bearing cells in tissue culture. Anti-TFR-LysPE40 given intraperitoneally to mice appeared rapidly in the blood and caused regression of A431 tumors growing as subcutaneous xenografts. These results show that it is possible to cause regression of a solid carcinoma by an immunotoxin if proper targeting can be achieved.

Two murine lysosome-associated membrane proteins, LAMP-1 of 105,000-115,000 D and LAMP-2 of 100,000-110,000 D, have been identified by monoclonal antibodies that bind specifically to lysosomal membranes. Both glycoproteins were distinguished as integral membrane components solubilized by detergent solutions but not by various chaotropic agents. The lysosome localization was demonstrated by indirect immunofluorescent staining, co-localization of the antigen to sites of acridine orange uptake, and immunoelectron microscopy. Antibody binding was predominantly located at the limiting lysosomal membrane, distinctly separated from colloidal gold-labeled alpha-2-macroglobulin accumulated in the lumen during prolonged incubation. LAMP-1 and LAMP-2 also appeared to be present in low concentrations on Golgi trans-elements but were not detected in receptosomes marked by the presence of newly endocytosed alpha-2-macroglobulin, or in other cellular structures. LAMP-1 and LAMP-2 were distinguished as different molecules by two-dimensional gel analysis, 125I-tryptic peptide mapping, and sequential immunoprecipitations of 125I-labeled cell extracts. Both glycoproteins were synthesized as a precursor protein of approximately 90,000 D, and showed a marked heterogeneity of apparent molecular weight expression in different cell lines. LAMP-2 was closely related or identical to the macrophage antigen, MAC-3, as indicated by antibody adsorption and tryptic peptide mapping. It is postulated that these glycoproteins, as major protein constituents of the lysosomal membrane, have important roles in lysosomal structure and function.

When KB cells labeled with 51Cr (1 muCi/ml) were treated with adenovirus type 2 (Ad2) at pH 6, 51Cr was released in a concentration-dependent manner with half-maximal release at 1 microgram/ml of virus. The 51Cr release was maximum at 10 micrograms/ml and represented nearly 20% of cell-associated 51Cr. 51Cr release depended on the length of incubation with Ad2 and the pH of the medium; maximum release was at pH 6 with very little release at pH 7.5. An antiserum against the penton base of Ad2 specifically neutralized the ability of Ad2 to release 51Cr. Chloroquine up to 200 microM did not block Ad2-dependent 51Cr release. DEAE-dextran stimulated Ad2-dependent 51Cr release; 4-5-fold stimulation was observed at 50 micrograms/ml of DEAE-dextran. We have compared the ability of Ad2 to release 51Cr with its ability to disrupt endocytic vesicles. Vesicle disruption was independent of pH of the medium in the range of pH 6 to 7.5 and was blocked by chloroquine, whereas, 51Cr release was much greater at pH 6 than at pH 7.5 and was not affected by chloroquine. These results suggest that Ad2 possess a lytic activity which ordinarily lyses acidic endocytic vesicles but, if the cells are maintained at acidic pH, can also directly disrupt the plasma membrane.

We have been developing a procedure for localizing intracellular antigens in cultured cells, by using peroxidase-labeled antibodies, that allows good morphologic preservation. Although useful, our previous technique did not preserve the morphology of membranes, and the location of the peroxidase reaction product was difficult to establish. In this paper, we report major improvements on the basic technique that markedly enhance the quality of localization and of morphology. Saponin is used to permeabilize membranes without destroying their morphology. The amount of reaction product is enhanced with a peroxidase-antiperoxidase label. The clarity of morphologic detail and contrast of reaction product density are increased by using postsectioning staining with the osmium/thiocarbohydrazide/osmium and uranyl acetate/lead citrate procedures. We have applied this technique to the ultrastructural localization of alpha2-macroglobulin and demonstrated that it is localized in membrane-limited vesicles. We have also used this method to improve the preservation of structures for localization by fluorescence microscopy.

When KB cells were labeled with either 51Cr (1 microCi/ml) or [35S]methionine (5 microCi/ml) and treated with 10 micrograms/ml of adenovirus type 2 (Ad2) at pH 6.0 for 60 min at 37 degrees C, about 25% of the cell-associated 51Cr and 5% of the [35S]methionine were released into the medium. The 51Cr was mainly associated with molecules of 1500 Da or less. When KB cells were labeled with either [3H] choline, alpha-[3H]aminobutyric acid, or [3H]deoxy-2-fluoro-D-glucose and exposed to Ad2, these molecules were released in amounts much higher than 51Cr. The Ad2-dependent release of choline was found to be dependent on Ad2 concentration, with maximum release (nearly 60%) at 10 micrograms/ml of Ad2, on the length of the incubation with Ad2, with maximum release at about 90 min, and on the medium pH with maximum activity at pH 6.0 to 6.5. Greater than 95% of the choline released was water-soluble and identified as choline phosphate. Less than 5% of the choline released was associated with lipids, and none was released as a phospholipid vesicle or micelle. The ability of Ad2 to release choline was abolished by incubating Ad2 for 10 min at 45 degrees C, whereas the binding of Ad2 to the cells was not affected. Fetal calf serum also blocked Ad2-dependent choline release.

Using nuclear envelopes from Chinese hamster ovary (CHO) cells as an antigen, a mouse monoclonal antibody (IgM; designated mAb CHON211) that specifically binds to components of the nuclear pore complex (nucleoporins) was isolated. Immunofluorescence localization of the antigen recognized by mAb CHON211 revealed a punctate pattern restricted to the nuclear envelope; this pattern changed dramatically during the cell cycle. When examined by electron microscopy, the antigen was largely restricted to the nucleoplasmic and cytoplasmic faces of the nuclear pore complex. Immunoblots showed that mAb CHON211 bound to a series of polypeptides enriched in the nuclear envelope fraction with apparent molecular masses ranging from 110 to 35 kDa. Using galactosyltransferase and the lectin wheat germ agglutinin as probes, we have previously shown that proteins bearing O-linked N-acetylglucosamine (GlcNAc) are restricted to the cytoplasmic and nucleoplasmic faces of the nuclear pore complex. The nuclear membrane proteins recognized by mAb CHON211 had properties very similar to those identified by galactosyltransferase and wheat germ agglutinin labeling. Removal of O-linked GlcNAc residues with beta-hexosaminidase greatly reduced the binding of mAb CHON211, strongly suggesting that O-linked GlcNAc moieties are part of the immunodeterminant. This monoclonal antibody defines a new family of antigens that are restricted to the nuclear pore and bear a common modification: O-linked GlcNAc. mAb CHON211 will be useful for defining the role of the nucleoporins in nucleo-cytoplasmic exchange.

We have utilized dark field microscopy to observe the surface microstructure of living cultured cells. Using this method, we have found that dibutyryl cAMP treatment causes regression of the numerous, long cell surface microvilli present on L929 cells. Thirty minutes after removal of dibutyryl cAMP, microvilli reappear. An inhibitor of phosphodiesterase (methylisobutylxanthine) and a stimulator of adenylate cyclase (prostaglandin E1), both of which raise cAMP levels, cause regression of microvilli in 15 min. Untransformed 3T3 cells show very few microvilli when viewed still attached to their substratum or after removal with EDTA. Treatment of these cells with trypsin causes the formation of numerous microvilli on their surface. When clumps of cells agglutinated by concanavalin A are examined by thin section electron microscopy, the cells are seen to be held together by a "forest" of interdigitating microvilli and only rarely is there apposition of the areas of membrane between microvilli. At the same time the distribution of surface-bound concanavalin A was examined using immunofluorescent light microscopy, and concanavalin A was found to be uniformly distributed over the cell surface. We propose that agglutinability of mouse and rat fibroblasts is regulated through the modulation of cell surface microvilli by cAMP, and that transformed cells are highly agglutinable because their low cAMP levels result in the formation of numerous surface microvilli.

AD6, a mutant derived from 3T3 Balb/c cells, is characterized by low adhesion to substratum, round shape, increase in surface microvilli, increase in agglutinability by concanavalin A, and loss of directional motility. These properties are often observed in transformed cells. However, the mutant has normal growth properties and anchorage-dependence of growth, and it does not form tumors. In AD6, the biosynthesis of complex carbohydrates and glycoproteins is impaired because of a block in the acetylation of GlcN-6-P. This defect is responsible for all the surface alterations because feeding of GlcNAc to AD6 cells corrects the defects in the synthesis of complex carbohydrates and the exposure of glycoproteins at the outer surface of the plasma membrane. Parallel to this biochemical reversion, there is full restoration of the altered biological properties. In contrast, GlcNAc has no effect on the morphologic features of two lines of transformed cells. Our results suggest that the carbohydrate portion of cell surface proteins has an important role in adhesion and related aspects of cell behavior. The fact that a defined alteration of the cell surface induces many properties often encountered in transformed cells, without affecting control of cell division, strongly suggests that these alterations in properties are not sufficient to account for the loss of growth regulation.

Monoclonal antibody 101 produced by a hybridoma obtained by fusion with NS-1 myeloma cells of spleen cells from a mouse immunized with the human epidermoid carcinoma cell line, A431, specifically precipitates epidermal growth factor receptor, a glycoprotein of 170,000 Mr solubilized from A431 cell membranes (Richert, N. D., Willingham, M. C., and Pastan, I. H. (1983) J. Biol. Chem. 258, 8902-8907). The antibody also binds to neutral glycolipids extracted from A431 cells as evidenced by solid phase radioimmunoassay and by autoradiography. Binding of antibody to its target is inhibited by lacto-N-fucopentaose I but not by 2‘-fucosyllactose or related oligosaccharides. Thus, antibody 101 is probably directed against the human blood group H type 1 sugar sequence Fuc alpha 1-2Gal beta 1-3GlcNAc . . . This sequence presumably occurs on the epidermal growth factor receptor. Monoclonal antibody 102 produced by another hybridoma from the same fusion has the same cell specificity as antibody 101 and also binds to neutral glycolipids. However, binding of antibody 102 to its target is inhibited by 2‘-fucosyllactose and not by lacto-N-fucopentaose I or related oligosaccharides. Thus, antibody 102 is probably directed against the human blood group H type 2 sugar sequence Fuc alpha 1-2Gal beta 1-4GlcNAc . . . Antibody 102 does not precipitate solubilized epidermal growth factor receptor. Both antibodies bind to neutral glycolipids extracted from human erythrocytes belonging to blood group O but not to neutral glycolipids extracted from human erythrocytes with the “Bombay” phenotype.

We report evidence that the enzyme transglutaminase (glutaminyl-peptide gamma-glutamyltransferase; R-glutaminyl-peptide:amine gamma-glutamyltransferase, EC2.3.2.13) participates in receptor-mediated endocytosis. Clustering and internalization of rhodamine-labeled alpha 2-macroglobulin (R alpha 2 M) in normal rat kidney (NRK) cells is inhibited by a wide spectrum of compounds that inhibit transglutaminases, including that from NRK cells. The pattern of clustering inhibition resembles the pattern of transglutaminase inhibition as follows: (i) The most potent transglutaminase inhibitors are dansylcadaverine and the transglutaminase-directed affinity label N-benzyloxy-carbonyl-5-diazo-4-oxonorvaline p-nitrophenyl ester; these were also the most potent inhibitors of clustering and internalization of R alpha 2M. (ii) The inhibition of clustering of R alpha 2M occurs in the same concentration range as that required for transglutaminase inhibition. (iii) Linear primary amines are more effective blockers than the iso-chain primary amines. (iv) The transglutaminase affinity label N-benzyloxycarbonyl-5-diazo-4-oxonorvaline p-nitrophenyl ester irreversibly inhibits a significant fraction of the NRK transglutaminase and the clustering and internalization of R alpha 2M. A closely related compound, N-trifluoroacetyl-6-diazo-5-oxonorleucine ethyl ester, does not significantly inhibit transglutaminase or clustering and internalization. (v) Clustering and internalization is inhibited 10-fold more effectively by the heptapeptide Ac-Gly2-LLeu-LLys-Gly3 than by the heptapeptides Ac-Gly2-LLeu-DLys-Gly3 or AcGly3-DLys-DLeu-Gly2. This is the pattern of stereospecificity for the inhibition of purified transglutaminases.

Using video intensification fluorescence microscopy and tetramethylrhodamine (Rho)-labeled 3,3',5-triiodo-L-thyronine (T3), we studied the uptake of T3 by cultured mouse fibroblasts. After incubation of cells with Rho-T3 for 30 min at 37 degrees C the fluorescent hormone was concentrated in many small bright accumulations. With a 1000-fold excess of unlabeled T3, only weak background fluorescence was seen. Furthermore, when cells were incubated with Rho or Rho-thyronine only background fluorescence was detected. These results indicate that the cellular uptake of Rho-T3 occurred through a T3-specific receptor-mediated process. Most of these accumulations underwent saltatory motion in living cells, indicating that the T3 was contained within endocytic vesicles. When cultured cells were incubated with Rho-T3 for 60 min at 4 degrees C, only diffuse fluorescence was observed, Rho-T3 became concentrated in vesicles upon warming of the cells to either 23 degrees C or 37 degrees C. Simultaneous incubation of cells with fluorescein-labeled alpha 2-macroglobulin and Rho-T3 showed that Rho-T3 was internalized in the same vesicles as alpha 2-macroglobulin. Furthermore, as previously reported for alpha 2-macroglobulin in the presence of methylamine, dansylcadaverine, or bacitracin, clustering and internalization were inhibited but the overall fluorescence intensity of the cells did not appear to be affected. Because it has been previously shown that receptor-mediated endocytosis of alpha 2-macroglobulin occurs through clustering of ligands in coated pits on the cell surface, these results indicate that Rho-T3 follows the same pathway. Thus it has now been demonstrated that a low-molecular weight hormone enters cells by this pathway.

The early events in the entry of ligands into cultured fibroblasts through plasma membrane clathrin-coated pits were examined by using serial sections and electron microscopy. Swiss 3T3 cells were labeled at 4 degrees C with concanavalin A-horseradish peroxidase and warmed to 37 degrees C for 1 or 2 min to label organelles participating in endocytosis. Cells were fixed in glutaraldehyde and further preserved with a membrane contrast-enhancement technique that utilized sequential osmium/ferrocyanide, thiocarbohydrazide, and osmium/ferrocyanide treatment in situ. This technique provides exceptional preservation and enhanced contrast of membranous structures. Serial sections taken both parallel to and perpendicular to the substratum were examined by electron microscopy, and 263 clathrin-coated structures containing horseradish peroxidase were evaluated for their connection to the plasma membrane. All of these coated structures were found to be connected to the plasma membrane either directly or via necks approximately equal to 170 A in width and up to 7,000 A in length. These necks were often surrounded by a dense filamentous network. These results strongly suggest that isolated coated vesicles do not form during the endocytosis process in Swiss 3T3 cells.

A conjugate of Pseudomonas exotoxin and epidermal growth factor (PE-EGF) inhibits proteins synthesis in KB cells, and this inhibition is increased by adenovirus. Protein synthesis inhibition is dependent on the amount of adenovirus and PE-EGF used and the time of incubation of cells with these agents. With 1 microgram of adenovirus and 0.5 micrograms of PE-EGF per ml, protein synthesis is inhibited about 80% in a 60-min experiment. Under these conditions neither adenovirus nor PE-EGF alone has any effect. In the presence of several weak bases or monensin, the enhancement of toxicity was substantially inhibited; half-maximal inhibition was achieved with 40 microM chloroquine, 10 mM ammonium chloride, 5 mM methylamine, 0.1 mM N-hexylamine and 1 microM monensin. At the concentrations employed, none of the inhibitors affected the amount of virus taken up or bound to the cell surface, and chloroquine had no effect on the amount of EGF taken up in 60 min. Chloroquine did not prevent the toxicity of the PE-EGF (5 micrograms/ml) alone. Because these compounds are known to elevate the pH in receptosomes, it seems likely that the acidification of the receptosome either enhances the lysis of the membrane by adenovirus or enhances some other step in the release of PE-EGF.

Antibodies were raised against the sequence Glu-Glu-Glu-Glu-Tyr-Met-Pro-Met -Glu, which represents a part of the middle T antigen of polyomavirus that is considered to be important in inducing the phenotype of transformed cells. The antibodies reacted with native as well as denatured middle T antigens. In addition, the antibodies immunoprecipitated a cellular protein with an apparent molecular weight of 130,000 (130K) from mouse and rat cells. In some cases, a 33K protein was also immunoprecipitated. Immunoprecipitation of middle T antigen as well as 130K and 33K proteins was blocked by the peptide. The antibodies labeled microfilaments of untransformed mouse, rat, human, and chicken cells by immunofluorescence. This labeling was also blocked by the peptide. The labeling pattern and distribution under a variety of conditions were indistinguishable from those of anti-actin antibodies, although no evidence has been obtained to indicate that the anti-peptide antibodies react with actin. The 130K protein migrated in sodium dodecyl sulfate-polyacrylamide gel electrophoresis slightly slower than chicken gizzard vinculin (130K) and slightly faster than myosin light-chain kinase of chicken smooth muscle (130K). Neither of these proteins absorbed the anti-peptide antibodies. The 33K protein does not seem to be tropomyosin (32K to 40K).

The binding of vesicular stomatitis virus (VSV) to Vero monkey cells was studied by using virus metabolically labeled with [35S]methionine. Under conditions where viral uptake did not occur (4 degrees C), apparent binding equilibrium was achieved within 12 h at a level representing 12% of the input virus. Two distinct forms of virus-cell interaction were found. At low concentrations of VSV, corresponding to multiplicities used for tissue culture studies, saturable binding was the major form of interaction. Saturation was complete at approximately 4,000 VSV virions per cell. At higher virus concentrations, nonsaturable binding prevailed. Trypsin treatment of Vero cells did not decrease the binding of VSV to the saturable binding sites. Internalization of VSV at 37 degrees C also displayed a saturable component which was directly comparable to that observed for binding. VSV binding to high-affinity, saturable sites on the plasma membrane may represent a receptor-mediated route of viral uptake.

Pseudomonas exotoxin (PE) kills mammalian cells in a complex process that involves cell surface binding, internalization by endocytosis, translocation to the cytosol, and ADP-ribosylation of elongation factor 2. PE is a three-domain protein in which domain I binds to the cell surface, domain II promotes translocation into the cytosol, and domain III carries out ADP-ribosylation. To determine how translocation occurs, we have mutated all the arginine residues in domain II and found that mutations at positions 276 and 279 greatly diminished the cytotoxicity of PE and mutations 330 and 337 substantially reduced cytotoxicity. Biochemical studies indicate that after internalization into an endocytic compartment, the PE molecule undergoes a specific and saturable intracellular interaction, and this interaction is deficient in an Arg276→Gly mutant. Our data suggest that the translocation process of PE involves a specific interaction of Arg276(and possibly Arg279, Arg330, and Arg337) with components of an intracellular compartment.

A monoclonal antibody to the epidermal growth factor (EGF) receptor of A431 cells was obtained after fusion of immunized BALB/c mouse spleen cells with NS-1 myeloma cells. Specific binding of the antibody to the plasma membrane of A431 cells was demonstrated by indirect immunofluorescence and electron microscopy. The antibody did not react with human KB cells, normal rat kidney cells, or Swiss 3T3 cells. The antibody is an IgG3K; it specifically immunoprecipitated a Mr approximately 170,000 protein from radiolabeled A431 cell extracts. This protein is phosphorylated in a EGF-dependent manner in intact A431 cells and in Triton X-100-solubilized plasma membranes. The specificity of the interaction of the antibody with the Mr = 170,000 protein was confirmed by electrophoretic transfer of A431 cell proteins to nitrocellulose followed by incubation with the antibody and 125I-protein A. When 125I-EGF was covalently cross-linked to its receptor, the 125I-EGF-receptor complex was specifically precipitated by the antibody. The monoclonal antibody did not inhibit the binding of 125I-EGF to its receptor in intact A431 cells and also failed to stimulate the phosphorylation of the Triton X-100-solubilized EGF receptor. The results indicate that the antibody and EGF bind to different sites on the EGF receptor. The antibody will be useful for isolating the EGF receptor in an unactivated form.

The use and properties of an alternative fixation procedure for ultrastructural cytochemistry is described that utilized primary fixation of cells with glutaraldehyde with considerable latitude in concentration and time of fixation. This was followed by treatment with sodium borohydride (as described by Weber, Rathke and Osborn (Proc Natl Acad Sci USA 75:1820, 1978)), which increases the accessibility of some cytoplasmic compartments to antibodies, apparently through the reduction of the Schiff bases induced by glutaraldehyde. Through the use of saponin membrane permeabilization, this primary fixation method allows good ultrastructural preservation with accessibility of most, but not all, cytoplasmic structures. Localization of tubulin, clathrin, alpha2-macroglobulin present in lysosomes, vesicular stomatitis virus (VSV) G protein, and myosin are demonstrated. This technique failed to expose the antigenic structure of actin in microfilament bundles or intranuclear antigens. This method should be useful for good preservation of ultrastructure and antibody localization of some, but not all, protein antigens in the cytoplasm of cultured cells, as well as in intact tissue. This technique is especially useful for cytosolic antigens in cultured cells.

Transferrin is taken up by receptor-mediated endocytosis into intracellular vesicles and tubules, and then recycles rapidly to the plasma membrane (diacytosis). We applied double-label cytochemistry to study whether the recycling structures containing transferrin fuse with the intracellular membranous structures that deliver newly synthesized membrane glycoproteins from the ER to the plasma membrane (exocytosis) or whether they remain independent. KB and Vero cells were infected with the temperature-sensitive transport mutant 0-45 of vesicular stomatitis virus (VSV). Temperature-regulated exocytosis of membrane glycoprotein "G" occurred simultaneously with diacytosis of transferrin. The exocytic "G" protein, as detected by immunoperoxidase electron microscopy, passed through the cisternal Golgi stacks and vacuolar, tubular, vesicular, and pit-like structures of the Golgi system. A transferrin-ferritin conjugate used in ultrastructural double-label experiments was detected in diacytic vesicles and tubules that accumulated in the proximal (trans-reticular) Golgi area of the cell. The ferritin-labeled vesicles/tubules were often close to and intermixed with the VSV-"G" containing membranous structures, but in most cases at early times (15-20 min) the transferrin and VSV-"G" containing vesicular structures remained distinct. At later times (30-45 min), the two labels were occasionally found in the same structures. These results indicate that rapid recycling of endocytosed materials and exocytosis of membrane glycoproteins to the cell surface usually occur in distinct vesicles, possibly along the same general morphologic exit pathway.

We have localized p55, a thyroid hormone-binding protein found in the endoplasmic reticulum in cultured cells, in samples of normal human and monkey tissues, using a monoclonal antibody with cryostat sections and immunoperoxidase histochemistry. Large amounts of p55 were found in many tissues, generally corresponding to the amount of endoplasmic reticulum contained in each cell type. Intense localization of p55 was found in cells of the anterior and intermediate pituitary lobes, in epithelial cells of thyroid follicles, in the glandular epithelium of mammary gland, in hepatocytes, in Paneth cells and Brunner's glands in duodenum, in acinar cells of pancreas, in adrenal cortical cells, and in scattered interstitial fibroblastic cells in many tissues. These results suggest a potential role for thyroid hormone and p55 in regulating protein synthesis or secretion in multiple organs.

Multidrug resistance to amphipathic natural product chemotherapeutic drugs is conferred on cancer cells by expression of the MDR1 gene, which encodes the 170-kDa multidrug transporter known as P glycoprotein. The P glycoprotein-mediated efflux of toxic chemotherapeutic drugs can be reversed by agents such as verapamil, which is a substrate for the multidrug transporter and appears to be a competitive inhibitor of the efflux pump. In this study, Bodipy-verapamil, a fluorescent derivative of verapamil, has been shown to be a substrate for the efflux pump activity of P glycoprotein. Single-cell fluorescence analysis reveals that Bodipy-verapamil accumulates in lysosomes of drug-sensitive NIH3T3 and KB cells but is rapidly effluxed from multidrug-resistant derivatives of these cell lines. Although Bodipy-verapamil is a substrate for the multidrug transporter, it is not an efficient inhibitor of the pump and does not reverse resistance to vinblastine and colchicine as effectively as does verapamil. This new derivative may be a useful tool for imaging of lysosomes in drug-sensitive cells and for rapid screening for the multidrug-resistant phenotype in other cell types.

Using transmission electron microscopy, we have studied the interaction of alpha 2 macroglobulin (alpha 2 M) with the surface of cultured fibroblasts. When cells were incubated for 2 h at 4 degrees C with ferritin-conjugated alpha 2 M, approximately 90% of the alpha 2 M was diffusely distributed on the cell surface, and the other 10% was concentrated in "coated" pits. A pattern of diffuse labeling with some clustering in "coated" pits was also obtained when cells were incubated for 5 min at 4 degrees C with alpha 2 M, fixed with glutaraldehyde, and the alpha 2 M was localized with affinity-purified, peroxidase-labeled antibody to alpha 2 M. Experiments in which cells were fixed with 0.2% paraformaldehyde before incubation with alpha 2 M showed that the native distribution of alpha 2 M receptors was entirely diffuse without significant clustering in "coated" pits. This indicates that some redistribution of the alpha 2 M-receptor complexes into clusters occurred even at 4 degrees C. In experiments with concanavalin A(Con A), we found that some of the Con A clustered in coated regions of the membrane and was internalized in coated vesicles, but much of the Con A was directly internalized in uncoated vesicles or pinosomes. We conclude that unoccupied alpha 2 M receptors are diffusely distributed on the cell surface. When alpha 2 M-receptor complexes are formed, they rapidly cluster in coated regions or pits in the plasma membrane and subsequently are internalized in coated vesicles. Because insulin and epidermal growth factor are internalized in the same structures as alpha 2 M (Maxfield, F.R., J. Schlessinger, Y. Schechter, I. Pastan, and M.C. Willingham. 1978. Cell, 14: 805--810.), we suggest that all peptide hormones, as well as other proteins that enter the cell by receptor-mediated endocytosis, follow this same pathway.

One form of multidrug resistance is due to the expression of a 170-kDa energy-dependent drug efflux pump called P-glycoprotein in the plasma membranes of human cancer cells. We have prepared conjugates of Pseudomonas toxin with the anti-P-glycoprotein monoclonal antibody MRK-16. These anti-P-glycoprotein-toxin conjugates specifically kill multidrug-resistant human KB cells. Similar conjugates could be useful in cancer therapy to reduce or eliminate multidrug-resistant tumor populations in tumors intrinsically resistant to chemotherapy or in populations that become resistant during combination chemotherapy.

A membrane-associated binding protein for 3,3',5-triiodo-L-thyronine (T3) was purified to apparent homogeneity from A431 human epidermoid carcinoma cells. A431 cells were specifically labeled with the N-bromoacetyl derivative of T3 labeled with 125I at the 3' position (BrAc[125I]T3) and were extracted with 3-[3-(cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), a zwitterionic detergent. The solubilized BrAc[125I]T3-labeled protein was successively purified by chromatography on Sephadex G-200 and QAE-Sephadex followed by NaDod-SO4/PAGE. Approximately 0.2 mg of purified protein was obtained from 2.5 X 10(9) cells, which represents a 3000-fold purification. The membrane-associated T3 binding protein is an acidic protein with a pI of 5.1 and an apparent molecular mass of 55,000 daltons determined by NaDodSO4/PAGE. Polyclonal antibodies against the 55-kDa protein were prepared and used in indirect immunofluorescence to show that the 55-kDa protein was mainly found in the nuclear envelope and endoplasmic reticulum.

An immunotoxin composed of Pseudomonas toxin coupled to an antibody to the human transferrin receptor was evaluated for its effect on ovarian cancer. In the tumor model employed, 60 million human ovarian cancer cells were injected into the peritoneal cavity of an immunodeficient nude mouse. By day 5, cancer cells were implanted and growing in small clusters throughout the peritoneal cavity. On days 5-8, 0.3-2 micrograms of immunotoxin was injected into the peritoneal cavity. Control mice died with malignant ascites at 34-58 days after the implantation of tumor cells, whereas immunotoxin-treated mice lived to 100 days or longer. Irrelevant immunotoxins or antibody alone had no antitumor activity. These findings suggest that intraperitoneal injection of immunotoxins may have a role in the treatment of ovarian cancer.

IL-2-PE40 is a chimeric protein composed of human interleukin 2 (IL-2) genetically fused to the amino terminus of a modified form of Pseudomonas exotoxin lacking its cell recognition domain. IL-2-PE40, which is extremely cytotoxic to IL-2 receptor-positive cells, was examined for its ability to prevent graft rejection in mice in which activation of T cells is prominent. We demonstrate that intraperitoneally administered IL-2-PE40 specifically and significantly prolongs the survival of vascularized heart allografts in mice. The chimeric toxin, IL-2-PE40, offers an alternative approach to the treatment of autoimmune diseases and transplant rejection in humans.

The pathway of transferrin uptake and recycling was investigated in KB cells to attempt to identify the organelles involved in the return of transferrin to the cell surface. Comparison was made with the pathway of internalization of epidermal growth factor (EGF), which has been shown to terminate in lysosomes. A horseradish peroxidase conjugate of transferrin (TF-HRP) was incubated with KB and at 4 degrees C and, at various times after warming to 37 degrees C, the location of TF-HRP was examined at the electron microscopic level. Transferrin, like EGF, was found to enter cells via coated pits and to move to the Golgi region in receptosomes (endosomes). Transferrin was located in the tubular elements of the transreticular portion of the Golgi but not in Golgi stacks. Interestingly, transferrin was not concentrated in the coated pits of the Golgi. In contrast, EGF was highly concentrated there. Transferrin was next detected in tubular elements (approximately equal to 600 A in width and up to 5,000 A in length) that were closely associated with microtubules and in dumbbell-shaped structures. In contrast, EGF was not detected in these structures. These results suggest that those organelles containing transferrin, but not EGF, participate in the return of receptor-bound transferrin to the cell surface.

The binding of 3,3',5-triiodo-L-thyronine (T3) to GH3 rat pituitary tumor cells was studied at 15 degrees C and was shown to be saturable, reversible, and stereospecific. Least-squares analysis of the binding data showed two classes of binding sites with dissociation constants of 1.8 +/- 0.2 nM and 260 +/- 30 nM and binding capacities of (5.2 +/- 0.2) X 10(4) and (1.6 +/- 0.2) X 10(6) sites per cell, respectively. Affinity labeling of intact cells was carried out by incubation of cells with 0.3 nM N-bromoacetyl-[125I]T3 at 15 degrees C for 1 hr. Analysis of the cellular extracts by sodium dodecyl sulfate gel electrophoresis showed three labeled protein bands with apparent molecular masses of 55, 47, and 33 kilodaltons (kDal) in a ratio of 86:7:7. The labeling of only the 55-kDal protein band was selectively reduced to 50% by 20 microM unlabeled T3. Highly purified plasma membranes of GH3 cells were prepared and shown to be free of nuclei. Affinity labeling of the purified plasma membranes gave the same labeling pattern as with intact cells. Peptide mapping by Staphylococcus aureus V8 digestion of the 55-kDal protein from cells or plasma membranes gave the identical peptide fragments. Thus the 55-kDal protein labeled from intact cells is the same protein as that from purified plasma membranes. These results together with our earlier findings [Horiuchi, R., Cheng, S.-y., Willingham, M. & Pastan, I. (1982) J. Biol. Chem. 257, 3139-3144] suggest that the 55-kDal protein may be involved in mediating the uptake of T3 in GH3 cells.

We characterized the transcriptional activity of the long terminal repeat (LTR) of Rous sarcoma virus by constructing a recombinant plasmid, pRSVcat, in which bacterial chloramphenicol acetyltransferase (CAT; acetyl-CoA:chloramphenicol 3-O-acetyltransferase, EC 2.3.1.28) coding sequences are placed under LTR control. We find that the LTR directs relatively high levels of CAT synthesis within 48 hr after calcium phosphate-mediated introduction of this plasmid into CV-1 monkey kidney cells, chicken embryo fibroblasts, Chinese hamster ovary cells, HeLa cells, or mouse NIH/3T3 cells. The level of CAT synthesis is 3-fold higher in CV-1 cells and up to 10-fold higher in HeLa and mouse NIH/3T3 cells than after transfection with a related vector, pSV2cat, carrying CAT sequences under control of the simian virus 40 early promoter. We have shown, by primer extension, that the amounts of CAT-specific mRNAs encoded by pRSVcat and pSV2cat correlate with the levels of CAT enzyme activity. By both S1 nuclease mapping and primer extension, we have demonstrated that the start site for RNA transcription within the LTR of pRSVcat corresponds to previous mapping data. We estimated transfection efficiencies by monitoring immunofluorescence induced by a rhodamine-labeled CAT antibody. Our results indicate that the Rous sarcoma virus LTR can direct synthesis of high levels of functional mRNA and has a wide expression range. The observed high transcriptional activity of the LTR is significant because it has been postulated that this LTR promotes activity of adjacent cellular oncogenes.

The human multidrug resistance gene (MDR1) encodes a drug efflux pump glycoprotein (P-glycoprotein) responsible for resistance to multiple cytotoxic drugs. A plasmid carrying a human MDR1 cDNA under the control of a chicken beta-actin promoter was used to generate transgenic mice in which the transgene was mainly expressed in bone marrow and spleen. Immunofluorescence localization studies showed that P-glycoprotein was present on bone marrow cells. Furthermore, leukocyte counts of the transgenic mice treated with daunomycin did not fall, indicating that their bone marrow was resistant to the cytotoxic effect of the drug. Since bone marrow suppression is a major limitation to chemotherapy, these transgenic mice should serve as a model to determine whether higher doses of drugs can cure previously unresponsive cancers.

Glycoproteins bearing a single N-acetylglucosamine (GlcNAc) residue attached by an O-glycosidic linkage to the polypeptide chain (Holt, G. D., and Hart, G. W. (1986) J. Biol. Chem. 261, 8049-8057) have been found to be enriched in the nuclear and soluble fractions of rat liver. Our goal was to determine the localization and membrane topography of proteins bearing O-linked GlcNAc using galactosyltransferase and wheat germ agglutinin (WGA) as membrane-impermeant probes. Latency of the enzyme mannose-6-phosphatase was used to quantitatively confirm the intactness of the nuclear envelope during incubations with galactosyltransferase or WGA. The O-linked GlcNAc residues of nuclei were fully accessible to modification by galactosyltransferase under conditions where the nuclear envelope mannose-6-phosphatase was 70% latent. Addition of detergent destroyed the permeability barrier but did not increase galactosylation of the O-linked GlcnAc. The major polypeptides bearing O-linked GlcNAc residues on nuclei were peripheral rather than integral membrane proteins with apparent molecular masses ranging from 210 to 54 kDa. The proteins were also detected on sealed nuclei using conjugates of WGA. WGA-rhodamine labeled intact nuclei when examined by immunofluorescence; WGA-peroxidase was used to identify the nuclear glycoproteins after transfer to nitrocellulose. WGA-ferritin selectively labels the cytoplasmic and nucleoplasmic faces of the nuclear pore complex when examined by electron microscopy. Taken together, these data strongly suggest that proteins bearing cytoplasmically oriented O-linked GlcNAc are components of the nuclear pore complex, thereby raising the possibility that cytoplasmic and nucleoplasmic glycoproteins are involved in the assembly or functioning of the nuclear pore.