Structure and assembly of desmosome junctions: biosynthesis and turnover of the major desmosome components of Madin-Darby canine kidney cells in low calcium medium

Neither stratifying (primary keratinocytes) nor simple (Madin-Darby canine kidney [MDCK] and Madin-Darby bovine kidney [MDBK]) epithelial cell types from desmosomes in low calcium medium (LCM; less than 0.1 mM), but they can be induced to do so by raising the calcium level to physiological concentrations (standard calcium medium [SCM], 2 mM). We have used polyclonal antisera to the major bovine epidermal desmosome components (greater than 100 kD) in a sensitive assay involving immunoprecipitation of the components from metabolically labeled MDCK cell monolayers to investigate the mechanism of calcium-induced desmosome formation. MDCK cells, whether cultured in LCM or SCM, were found to synthesize the desmosome protein, DPI and desmosome glycoproteins DGI and DGII/III with identical electrophoretic mobility, and also, where relevant, with similar carbohydrate addition/processing and proteolytic processing. The timings of these events and of transport of DGI to the cell surface were similar in low and high calcium. Although the rates of synthesis of the various desmosome components were also similar under both conditions, the glycoprotein turnover rates increased dramatically in cells cultured in LCM. The half-lives decreased by a factor of about 7 for DGI and 12 for DGII/III and, consistent with this, MDCK cells labeled for 48 h in SCM had three and six times the amount of DGI and DGII/III, respectively, as cells labeled for 48 h in LCM. The rate of turnover and the levels of DPI were changed in the same direction, but to much lesser extents. Possible mechanisms for the Ca2+-dependent control of desmosome formation are discussed in the light of this new evidence.     

Chromosomal assignment of the human genes coding for the major proteins of the desmosome junction, desmoglein DGI (DSG), desmocollins DGII/III (DSC), desmoplakins DPI/II (DSP), and plakoglobin DPIII (JUP)

We have established PCR assays for the genes coding for the major proteins of the desmosome type of cell junction, the desmosomal cadherins DGI (desmoglein) and DGII/III (desmocollins), and the plaque proteins DPI/II (desmoplakin) and DPIII (plakoglobin) and used them to test human-mouse and human-rat somatic cell hybrids with different contents of human chromosomes. From these data we were able to assign DGI to chromosome 18 (DSG), DGII/III to chromosome 9p (DSC), DPI/II to chromosome 6p21-ter(DSP), and DPIII to chromosome 7 (JUP).     

Desmosomal glycoproteins II and III. Cadherin-like junctional molecules generated by alternative splicing

We have cloned the human genes coding for desmosomal glycoproteins DGII and DGIII, found in desmosomal cell junctions, and sequencing shows that they are related to the cadherin family of cell adhesion molecules. Thus a new super family of cadherin-like molecules exists which also includes the other major desmosomal glycoprotein, DGI (Wheeler, G. N., Parker, A. E., Thomas, C. L., Ataliotis, P., Poynter, D., Arnemann, J., Rutman, A. J., Pidsley, S. C., Watt, F. M., Rees, D. A., Buxton, R. S., and Magee, A. I. (1991) Proc. Natl. Acad. Sci. U.S.A., in press). DGIII differs from DGII by the addition of a 46-base pair exon containing an in-frame stop codon resulting in mature protein molecular weights of 84,633 for DGII and 78,447 for DGIII. The unique carboxyl-terminal region of DGII contains a potential serine phosphorylation site explaining why only DGII is phosphorylated on serine. The cadherin cell adhesion recognition sequence (His-Ala-Val) is replaced by Phe-Ala-Thr, suggesting that DGII/III may be adhesive molecules using a different mechanism.     

Regulation of desmosome assembly in epithelial cells: kinetics of synthesis, transport, and stabilization of desmoglein I, a major protein of the membrane core domain

Desmosomes are composed of two morphologically and biochemically distinct domains, a cytoplasmic plaque and membrane core. We have initiated a study of the synthesis and assembly of these domains in Madin-Darby canine kidney (MDCK) epithelial cells to understand the mechanisms involved in the formation of desmosomes. Previously, we reported the kinetics of assembly of two components of the cytoplasmic plaque domain, Desmoplakin I/II (Pasdar, M., and W. J. Nelson. 1988. J. Cell Biol. 106:677-685 and 106:687-699. We have now extended this analysis to include a major glycoprotein component of the membrane core domain, Desmoglein I (DGI; Mr = 150,000). Using metabolic labeling and inhibitors of glycoprotein processing and intracellular transport, we show that DGI biosynthesis is a sequential process with defined stages. In the absence of cell-cell contact, DGI enters a Triton X-100 soluble pool and is core glycosylated. The soluble DGI is then transported to the Golgi complex where it is first complex glycosylated and then titrated into an insoluble pool. The insoluble pool of DGI is subsequently transported to the plasma membrane and is degraded rapidly (t1/2 less than 4 h). Although this biosynthetic pathway occurs independently of cell-cell contact, induction of cell-cell contact results in dramatic increases in the efficiency and rate of titration of DGI from the soluble to the insoluble pool, and its transport to the plasma membrane where DGI becomes metabolically stable (t1/2 greater than 24 h). Taken together with our previous study of DPI/II, we conclude that newly synthesized components of the cytoplasmic plaque and membrane core domains are processed and assembled with different kinetics indicating that, at least initially, each domain is assembled separately in the cell. However, upon induction of cell-cell contact there is a rapid titration of both components into an insoluble and metabolically stable pool at the plasma membrane that is concurrent with desmosome assembly.     

Studies on the development and maintenance of epithelial cell surface polarity with monoclonal antibodies

We examined epithelial cell surface polarity in subconfluent and confluent Madin-Darby canine kidney (MDCK) cells with monoclonal antibodies directed against plasma membrane glycoproteins of 35,000, 50,000, and 60,000 mol wt. The cell surface distribution of these glycoproteins was studied by immunofluorescence and immunoelectron microscopy. At the ultrastructural level, the electron-dense reaction product localizing all three glycoproteins was determined to be uniformly distributed over the apical and basal cell surfaces of subconfluent MDCK cells as well as on the lateral surfaces between contacted cells; however, after formation of a confluent monolayer, these glycoproteins could only be localized on the basal-lateral plasma membrane. The development of cell surface polarity was followed by assessing glycoprotein distribution with immunofluorescence microscopy at selected time intervals during growth of MDCK cells to form a confluent monolayer. These results were correlated with transepithelial electrical resistance measurements of tight junction permeability and it was determined by immunofluorescence that polarized distributions of cell surface glycoproteins were established just after electrical resistance could be detected, but before the development of maximal resistance. Our observations provide evidence that intact tight junctions are required for the establishment of the apical and basal- lateral plasma membrane domains and that development of epithelial cell surface polarity is a continuous process.     

Expression of herpes simplex virus glycoproteins in polarized epithelial cells.

Members of the herpesvirus family mature at inner nuclear membranes, although a fraction of the viral glycoproteins is expressed on the cell surface. In this study, we investigated the localization of herpes simplex virus type 2 (HSV-2) glycoproteins in virus-infected epithelial cells by using a panel of monoclonal antibodies directed against each of the major viral glycoproteins. All of the HSV-2 glycoproteins were localized exclusively on the basolateral membranes of Vero C1008, Madin-Darby bovine kidney, and mouse mammary epithelial cells. Using a monoclonal antibody to HSV-2 gD which cross-reacts with HSV-1 strains, we could also localize HSV-1 gD on the basolateral membranes of Madin-Darby bovine kidney cells. These results indicate that these molecules contain putative sorting signals that direct them to basolateral membrane domains.     

Tumor promoter-induced disruption of junctional complexes in cultured epithelial cells is followed by the inhibition of cytokeratin and desmoplakin synthesis

The organization and synthesis of proteins involved in the formation and stabilization of desmosome-type junctions was investigated in cultured epithelial cells treated with a tumor promoter (12-O-tetradecanoyl-phorbol-13-acetate (TPA]. In Madin-Darby bovine (MDBK) and canine (MDCK) kidney cell colonies, TPA induced a rapid disruption of desmosomes and marked alterations in cell morphology. Within 4-6 h after TPA treatment, cell shape changed from cuboidal to highly irregular, with some very long extensions that contained cytokeratin fibrils, and many flat lamellar protrusions which were devoid of cytokeratin fibrils. These morphological changes in both MDBK and MDCK cells were followed by a dramatic and coordinated inhibition in the synthesis of all cytokeratins, 14-24 h after the addition of TPA, but without a similar effect on the synthesis of vimentin, which is coexpressed in these cells. In contrast, in dense cultures of MDBK and MDCK cells the synthesis of cytokeratins and the organization of desmosomal contacts were not affected by TPA. In an epithelial cell line derived from the bovine mammary gland (BMGE-H) the synthesis of an acidic cytokeratin of 45 kD, which was previously shown to be synthesized at high levels only in dense cultures, was dramatically inhibited by TPA treatment. Cell-free in vitro translation assays with mRNA from control and TPA-treated cells also demonstrated a decrease in the synthesis of cytokeratins in response to TPA. The inhibition of cytokeratin synthesis after TPA treatment was paralleled by a decrease in the synthesis of a high molecular weight (HMW) desmoplakin protein, which was abundant in dense MDBK and BMGE-H cells. The results with TPA-treated cells are suggestive of a coordinated down-regulation in the synthesis of only those cytokeratins and of a desmoplakin which were shown to be regulated by the extent of cell-cell contact. Cytokeratin phosphorylation in TPA-treated cells was low and reflected the decrease in their total mass, suggesting that it was not altered by TPA treatment. The possible linkage between the regulation of synthesis and organization of proteins involved in desmosome formation is discussed.     

Steady-state distribution and biogenesis of endogenous Madin-Darby canine kidney glycoproteins: evidence for intracellular sorting and polarized cell surface delivery

We used domain-selective biotinylation/125I-streptavidin blotting (Sargiacomo, M., M. P. Lisanti, L. Graeve, A. Le Bivic, and E. Rodriguez-Boulan. 1989 J. Membr. Biol. 107:277-286), in combination with lectin precipitation, to analyze the apical and basolateral glycoprotein composition of Madin-Darby canine kidney (MDCK) cells and to explore the role of glycosylation in the targeting of membrane glycoproteins. All six lectins used recognized both apical and basolateral glycoproteins, indicating that none of the sugar moieties detected were characteristic of the particular epithelial cell surface. Pulse-chase experiments coupled with domain-selective glycoprotein recovery were designed to detect the initial appearance of newly synthesized glycoproteins at the apical or basolateral cell surface. After a short pulse with a radioactive precursor, glycoproteins reaching each surface were biotinylated, extracted, and recovered via precipitation with immobilized streptavidin. Several basolateral glycoproteins (including two sulfated proteins) and at least two apical glycoproteins (one of them the major sulfated protein of MDCK cells) appeared at the corresponding surface after 20-40 min of chase, but were not detected in the opposite surface, suggesting that they were sorted intracellularly and vectorially delivered to their target membrane. Several "peripheral" apical proteins were detected at maximal levels on the apical surface immediately after the 15-min pulse, suggesting a very fast intracellular transit. Finally, domain-selective labeling of surface carbohydrates with biotin hydrazide (after periodate oxidation) revealed strikingly different integral and peripheral glycoprotein patterns, resembling the Con A pattern, after labeling with sulfo-N-hydroxy-succinimido-biotin. The approaches described here should be useful in characterizing the steady-state distribution and biogenesis of endogenous cell surface components in a variety of epithelial cell lines.     

Analysis of epithelial cell surface polarity with monoclonal antibodies

The hybridoma technique of K?hler and Milstein was utilized to isolate hybrid cell lines secreting monoclonal antibodies against cell surface proteins on the Madin-Darby canine kidney (MDCK) epithelial cell line. These antibodies were employed as high-affinity ligands to study the development and maintenance of epithelial cell polarity in MDCK cells and for the identification of nephron segment-specific proteins. Using standard procedures, we were able to immunoprecipitate glycoproteins with molecular weights of 25,000 ( 25K ), 35,000 ( 35K ), and 50,000 (50K). Immunofluorescence and immunoelectron microscopy of MDCK demonstrated that the 35K and 50K proteins could be localized on both the apical and basolateral membranes of subconfluent cells but primarily on the basolateral membranes of confluent cells. By determining the cell surface distribution of the 35K and 50K proteins on MDCK cells during growth into a confluent monolayer, and after the experimental disruption of tight junctions, evidence was obtained that the polarized distribution of these cell surface glycoproteins required the presence of tight junctions. We propose that confluent MDCK cells have a mechanism that is responsible for the establishment and maintenance of epithelial apical and basolateral membranes as distinct cell surface domains. These monoclonal antibodies were also used to localize the 25K and 35K glycoproteins in the kidney. The distribution of these proteins was mapped by immunofluorescence and immunoelectron microscopy and was determined to be on the basolateral membranes of epithelial cells in only certain tubular segments of the nephron. The possible functional implications of these distributions are discussed.     

Isolation of the intercellular glycoproteins of desmosomes

To characterize the desmosome components that mediate intercellular adhesion and cytoskeletal-plasma membrane attachment, we prepared whole desmosomes and isolated desmosomal intercellular regions (desmosomal "cores") from the living cell layers of bovine muzzle epidermis. The tissue was disrupted in a nonionic detergent at low pH, sonicated, and the insoluble residue fractionated by differential centrifugation and metrizamide gradient centrifugation. Transmission electron microscopic analyses reveal that a fraction obtained after differential centrifugation is greatly enriched in whole desmosomes that possess intracellular plaques. Metrizamide gradient centrifugation removes most of the plaque material, leaving the intercellular components and the adjoining plasma membranes. Sodium dodecyl sulfate polyacrylamide gel electrophoresis coupled with methods that reveal carbohydrate-containing moieties on gels demonstrate that certain proteins present in whole desmosomes are glycosylated. These glycoproteins are specifically and greatly enriched in the desmosome cores of which they are the principal protein constituents, and thus may function as the intercellular adhesive of the desmosome.     

Differential targeting of an epithelial plasma membrane glycoprotein in polarized Madin-Darby canine kidney cells

Using monoclonal antibodies directed against the plasma membrane of Madin-Darby canine kidney (MDCK) cells, we demonstrated previously that a glycoprotein with an Mr = 23,000 (gp23) had a non-polarized cell surface distribution and was observed on both the apical and basolateral membranes (Ojakian, G. K., Romain, R. E., and Herz, R. E. (1987) Am. J. Physiol. 253, C433-C443). However, in parallel studies on MDCK clonal lines (D11, D18) with high transepithelial electrical resistances and in kidney cells in vivo it was determined that gp23 had a polarized cell surface distribution, being localized only to the basolateral membrane. The cell surface distribution of other glycoproteins was identical in both MDCK and MDCK clonal lines, indicating that MDCK cells were not deficient in the ability to properly sort membrane glycoproteins. Metabolic labeling with radioactive substrates followed by immunopurification and gel electrophoresis demonstrated that gp23 from both MDCK and MDCK clone D11 had many biochemical similarities including electrophoretic mobility, glycosylation, and palmitate incorporation. However, proteolytic digestion of gp23 from MDCK and clone D11 cells produced unique peptide maps suggesting that these closely related glycoproteins may have different primary sequences. In this report, we present evidence that the differential targeting of gp23 may be due to differences between the primary sequences of the basolateral and non-targeted proteins. The possibility that the observed differences in gp23 targeting are due to the presence of a basolateral recognition signal in gp23 from clone D11 cells is discussed.     

Surface expression of viral glycoproteins is polarized in epithelial cells infected with recombinant vaccinia viral vectors.

In polarized epithelial cells, maturation sites of enveloped viruses that form by budding at cell surfaces are restricted to particular membrane domains. Recombinant vaccinia viruses were used to investigate the sites of surface expression in the Madin-Darby canine kidney (MDCK) cell line of the hemagglutinin (HA) of influenza virus, the G glycoprotein of vesicular stomatitis virus (VSV), and gp70/p15E of Friend murine leukemia virus (MuLV). These glycoproteins could be demonstrated by immunofluorescence on the surfaces of MDCK cells as early as 4 h post-infection. In intact MDCK monolayers, vaccinia recombinants expressing HA produced a pattern of surface fluorescence typical of an apically expressed glycoprotein. In contrast, cells infected with vaccinia recombinants expressing VSV-G or MuLV gp70/p15E exhibited surface fluorescence only when monolayers were treated with EGTA to disrupt tight junctions, as expected of glycoproteins expressed on basolateral surfaces. Immunoferritin labeling in conjunction with electron microscopy confirmed that MDCK cells infected with the HA recombinant exhibited specific labeling of the apical surfaces whereas the VSV-G and MuLV recombinants exhibited the respective antigens predominantly on the basolateral membranes. Quantitation of surface expression by [125I]protein A binding assays on intact and EGTA-treated monolayers confirmed the apical localization of the vaccinia-expressed HA and demonstrated that 95% of the VSV-G and 97% of the MuLV gp70/p15E glycoproteins were localized on the basolateral surfaces. These results demonstrate that glycoproteins of viruses that normally mature at basolateral surfaces of polarized epithelial cells contain all of the structural information required for their directional transport to basolateral plasma membranes.     

Ultrastructural and biochemical identification of con A receptors in the desmosome

Correlated ultrastructural and biochemical methods were used to identify and localize Concanavalin A (Con A) receptors in the desmosomes of bovine epidermis. Specific carbohydrate residues were labeled with ferritin-Con A in thin sections of tissue embedded in a hydrophilic resin. Quantitative mapping of ferritin distribution in labeled desmosomes revealed that Con A receptors are localized in the intercellular zone and concentrated along the desmosomal midline or central dense stratum. Labeling was almost entirely absent when sections were treated with ferritin-Con A in the presence of 0.1 M α-methyl mannoside, a hapten-inhibitor of Con A. “Whole” desmosomes and desmosomal intercellular regions (desmosomal “cores”) were purified from bovine muzzle epidermis. Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis reveals a limited number of major desmosomal protein constituents. Certain of these are glycoproteins and are greatly enriched in the core fraction. Almost all the desmosomal glycoproteins are intensely labeled when electrophoretic gels of whole desmosome or core fractions are exposed to fluorescent Concanavalin A.     

Glycosylation does not determine segregation of viral envelope proteins in the plasma membrane of epithelial cells

Enveloped viruses are excellent tools for the study of the biogenesis of epithelial polarity, because they bud asymmetrically from confluent monolayers of epithelial cells and because polarized budding is preceded by the accumulation of envelope proteins exclusively in the plasma membrane regions from which the viruses bud. In this work, three different experimental approaches showed that the carbohydrate moieties do not determine the final surface localization of either influenza (WSN strain) or vesicular stomatitis virus (VSV) envelope proteins in infected Madin-Darby Canine Kidney (MDCK) cells, as determined by immunofluorescence and immunoelectron microscopy, using ferritin as a marker. Infected concanavalin A- and ricin 1-resistant mutants of MDCK cells, with alterations in glycosylation, exhibited surface distributions of viral glycoproteins identical to those of the parental cell line, i.e., influenza envelope proteins were exclusively found in the apical surface, whereas VSV G protein was localized only in the basolateral region. MDCK cells treated with tunicamycin, which abolishes the glycosylation of viral glycoproteins, exhibited the same distribution of envelope proteins as control cells, after infection with VSF or influenza. A temperature-sensitive mutant of influenza WSN, ts3, which, when grown at the nonpermissive temperature of 39.5 degrees C, retains the sialic acid residues in the envelope glycoproteins, showed, at both 32 degrees C (permissive temperature) and 39.5 degrees C, budding polarity and viral glycoprotein distribution identical to those of the parental WSN strain, when grown in MDCK cells. These results demonstrate that carbohydrate moieties are not components of the addressing signals that determine the polarized distribution of viral envelope proteins, and possibly of the intrinsic cellular plasma membrane proteins, in the surface of epithelial cells.     

Regulation of desmosome assembly in MDCK epithelial cells: coordination of membrane core and cytoplasmic plaque domain assembly at the plasma membrane

Desmosomes are major components of the intercellular junctional complex in epithelia. They consist of at least eight different cytoplasmic and integral membrane proteins that are organized into two biochemically and structurally distinct domains: the cytoplasmic plaque and membrane core. We showed previously that in MDCK epithelial cells major components of the cytoplasmic plaque (desmoplakin I and II; DPI/II) and membrane core domains (desmoglein I; DGI) initially enter a pool of proteins that is soluble in buffers containing Triton X-100, and then titrate into an insoluble pool before their arrival at the plasma membrane (Pasdar, M., and W. J. Nelson. 1988. J. Cell Biol. 106:677-685; Pasdar. M., and W. J. Nelson. 1989. J. Cell Biol. 109:163-177). We have now examined whether either the soluble or insoluble pool of these proteins represents an intracellular site for assembly and interactions between the domains before their assembly into desmosomes at the plasma membrane. Interactions between the Triton X-100-soluble pools of DPI/II and DGI were analyzed by sedimentation of extracted proteins in sucrose gradients. Results show distinct differences in the sedimentation profiles of these proteins, suggesting that they are not associated in the Triton X-100-soluble pool of proteins; this was also supported by the observation that DGI and DPI/II could not be coimmunoprecipitated in a complex with each other from sucrose gradient fractions. Immunofluorescence analysis of the insoluble pools of DPI/II and DGI, in cells in which desmosome assembly had been synchronized, showed distinct differences in the spatial distributions of these proteins. Furthermore, DPI/II and DGI were found to be associated with different elements of cytoskeleton; DPI/II were located along cytokeratin intermediate filaments, whereas DGI appeared to be associated with microtubules. The regulatory role of cytoskeletal elements in the intracellular organization and assembly of the cytoplasmic plaque and membrane core domains, and their integration into desmosomes on the plasma membrane is discussed.     

Differential effect of monensin on enveloped viruses that form at distinct plasma membrane domains

We have observed a striking differential effect of the ionophore, monensin, on replication of influenza virus and vesicular stomatitis virus (VSV) in Madin-Darby canine kidney (MDCK) and baby hamster kidney (BHK21) cells. In MDCK cells, influenza virus is assembled at the apical surfaces, whereas VSV particles bud from the basolateral membranes; no such polarity of maturation is exhibited in BHK21 cells. A 10(-6) M concentration of monensin reduces VSV yields in MDCK cells by greater than 90% as compared with controls, whereas influenza virus yields are unaffected. In BHK21 cells, monensin also inhibits VSV production, but influenza virus is also sensitive to the ionophore. Immunofluorescent staining of fixed and unfixed MDCK monolayers indicates that VSV glycoproteins are synthesized in the presence of monensin, but their appearance on the plasma membrane is blocked. Electron micrographs of VSV-infected MDCK cells treated with monensin show VSV particles aggregated within dilated cytoplasmic vesicles. Monensin-treated influenza virus-infected MDCK cells also contain dilated cytoplasmic vesicles, but virus particles were not found in these structures, and numerous influenza virions were observed budding at the cell surface. These results indicate that influenza virus glycoprotein transport is not blocked by monensin treatment, whereas there is a block in transport of VSV G protein. Thus it appears that at least two distinct pathways of transport of glycoproteins to the plasma membrane exist in MDCK cells, and only one of them is blocked by monensin.     

Carbohydrate-mediated Golgi to cell surface transport and apical targeting of membrane proteins.

Polarized expression of most epithelial plasma membrane proteins is achieved by selective transport from the Golgi apparatus or from endosomes to a specific cell surface domain. In Madin-Darby canine kidney (MDCK) cells, basolateral sorting generally depends on distinct cytoplasmic targeting determinants. Inactivation of these signals often resulted in apical expression, suggesting that apical transport of transmembrane proteins occurs either by default or is mediated by widely distributed characteristics of membrane glycoproteins. We tested the hypothesis of N-linked carbohydrates acting as apical targeting signals using three different membrane proteins. The first two are normally not glycosylated and the third one is a glycoprotein. In all three cases, N-linked carbohydrates were clearly able to mediate apical targeting and transport. Cell surface transport of proteins containing cytoplasmic basolateral targeting determinants was not significantly affected by N-linked sugars. In the absence of glycosylation and a basolateral sorting signal, the reporter proteins accumulated in the Golgi complex of MDCK as well as CHO cells, indicating that efficient transport from the Golgi apparatus to the cell surface is signal-mediated in polarized and non-polarized cells.     

Epithelial cell interaction in air-liquid interface culture

Summary A novel culture method has been developed to study the interaction of epithelial cells in the absence of a solid substratum. Starting with either a single cell suspension or aggregates, cells were floated at the interface of air and liquid culture medium. Two epithelial cell lines have been studied in this system: Madin-Darby canine kidney cells (MDCK), and a rat bladder tumor cell line (NBT-II). Starting with a single cell suspension of MDCK, the floating cells coalesced in 24 h into sheets of cells. The cells were morphologically polarized with the apical surface facing the liquid medium. Domes were observed regularly in these sheets of cells. NBT-II cells migrated actively from aggregates at the air-liquid interface. In this floating culture, NBT-II cells produced extensive cell processes similar to those seen in cells grown on a solid surface. Because cells at the air-liquid interface lack a solid substratum for adhesion, cell membrane processes such as lamellapodia, retraction fibers, pseudopods, and long, intercellular connections can only exert a tension equal to or less than the surface tension of the liquid. Dimethyl sulfoxide 2% stimulated desmosome formation in floating NBT-II cells, resulting in a cribriform pattern in the sheet of cells. This method of interface can lead to new understanding of morphogenesis of epithelial cells, and the mechanism, of cell motility and formation of cell processess. This research was supported by research grant CA14137 from the National Institutes of Health, Bethesda, MD, and in part by the W. W. Smith Charitable Trust, Rosemont, PA     

Polarized endocytosis by Madin-Darby canine kidney cells transfected with functional chicken liver glycoprotein receptor

We have studied the expression of the chicken hepatic glycoprotein receptor (chicken hepatic lectin [CHL]) in Madin-Darby canine kidney (MDCK) cells, by transfection of its cDNA under the control of a retroviral promotor. Transfected cell lines stably express 87,000 surface receptors/cell with a kd = 13 nM. In confluent monolayers, approximately 40% of CHL is localized at the plasma membrane. 98% of the surface CHL is expressed at the basolateral surface where it performs polarized endocytosis and degradation of glycoproteins carrying terminal N-acetylglucosamine at a rate of 50,000 ligand molecules/h. Studies of the half-life of metabolically labeled receptor and of the stability of biotinylated cell surface receptor after internalization indicate that transfected CHL performs several rounds of uptake and recycling before it gets degraded. The successful expression of a functional basolateral receptor in MDCK cells opens the way for the characterization of the mechanisms that control targeting and recycling of proteins to the basolateral membrane of epithelial cells.     

Characterization of ZO-1, a protein component of the tight junction from mouse liver and Madin-Darby canine kidney cells

ZO-1, originally identified by mAb techniques, is the first protein shown to be specifically associated with the tight junction. Here we describe and compare the physical characteristics of ZO-1 from mouse liver and the Madin-Darby canine kidney (MDCK) epithelial cell line. The ZO-1 polypeptide has an apparent size of 225 kD in mouse tissues and 210 kD in canine-derived MDCK cells as determined by SDS-PAGE/immunoblot analysis. ZO-1 from both sources is optimally solubilized from isolated plasma membranes by either 6 M urea or high pH conditions; partial solubilization occurs with 0.3 M KCl. The nonionic detergents, Triton X-100 and octyl-beta-D-glucopyranoside, do not solubilize ZO-1. These solubility properties indicate that ZO-1 is a peripherally associated membrane protein. ZO-1 was purified to electrophoretic homogeneity from [35S]methionine metabolically labeled MDCK cells by a combination of gel filtration and immunoaffinity chromatography. Purified ZO-1 has an s20,w of 5.3 and Stokes radius of 8.6 nm. These values suggest that purified ZO-1 is an asymmetric monomeric molecule. Corresponding values for mouse liver ZO-1, characterized in impure protein extracts, were 6 s20,w and 9 nm. ZO-1 was shown to be a phosphoprotein in MDCK cells metabolically labeled with [32P]orthophosphate; analysis of phosphoamino acids from purified ZO-1 revealed only phosphoserine. ZO-1 epitope number was determined by Scatchard analysis of competitive and saturable binding of two different 125I-mAbs to SDS-solubilized proteins from liver and MDCK cells immobilized on nitrocellulose. Saturation binding occurs at 26 ng mAb/mg liver and 63 ng/mg of MDCK cell protein. This is equivalent to 30,000 ZO-1 molecules per MDCK cell assuming a single epitope/ZO-1 molecule.