Previous work has shown that cultured kératinocytes do not form desmosomes at low [Ca2+] (<0·l mM) but may be induced to do so by raising [Ca2+] to physiological levels (l·8–2mM). Here, fluorescent antibody staining with specific anti-desmosomal antibodies and electron microscopy have been used to determine whether Ca2+-induced desmosome formation also occurs in simple epithelial cells.

Both Madin-Darby canine and bovine kidney cells (MDCK and MDBK) exhibit Ca2+-induced desmosome formation, but there are significant differences between them. MDCK cells resemble kératinocytes in showing showing rapid desmosome formation characterized by the simultaneous appearance of four desmosomal antigens at the cell periphery within 15–20 min of raising the [Ca2+]. In contrast MDBK cells take between 7 and 8 h to form desmosomes after Ca2+ switching, and this is characterized by slow appearance of two desmosomal antigens, the 175–164 (×103)Mr glycoprotein and desmoplakin, at the cell periphery.

Differences in the pattern of staining for desmosomal antigens between the two cell types in low and high [Ca2+] are described and discussed in relation to desmosome formation and internalization. Triton X-100 extractability of desmosomal antigen staining is also considered. While most is non-extractable, staining for the glycoproteins known as desmocollins is completely extractable from MDCK cells in low [Ca2+], but that which reaches the cell periphery after Ca2+ switching becomes non-extractable. Although neither cell type forms desmosomes in low [Ca2+], both possess zonulae adhaerentes, suggesting a difference in Ca2+ requirement for formation of these two junctions.

Desmosomes are adhesive junctions of epithelial cells. Their ultrastructure has been well documented (Overton, 1962, 1974; Farquhar & Palade, 1963; Kelly, 1966; Campbell & Campbell, 1971; Kelly & Shienvold, 1976) and we now have some understanding of their molecular composition (Gorbsky & Steinberg, 1981 ; Miieller & Franke, 1983; Cowin & Garrod, 1983; Kapprell et al. 1985; Skerrow, 1985; Garrod, 1985; Garrod & Cowin, 1986).

Epidermal kératinocytes cultured in low [Ca2+] (<0·lmM) do not form desmosomes but may be stimulated to do so rapidly (5–15 min) by raising the external [Ca2+] to physiological levels (1–2mM)(Hennings et al. 1980; Jones et al. 1982; Watt et al. 1984; Jones & Goldman, 1985). We have shown, using immuno-fluorescent staining, that kératinocytes in low calcium medium (LCM) possess a diffuse distribution of all the desmosomal components. Raising the [Ca2+] causes rapid and synchronous accumulation of these components at the cell boundaries (Watt et al. 1984), corresponding with the first appearance of identifiable desmosomes by electron microscopy.

We have now investigated the possibility of using the Ca2+ ‘switch’ to examine desmosome formation in other epithelial cell types, MDBK (Madin-Darby bovine kidney) and MDCK (Madin-Darby canine kidney) cells. Both form polarized, transporting epithelia (Misfeldt et al. 1976; Cereijido et al. 1978; Richardson et al. 1981), which do not stratify. As they form a confluent monolayer, MDBK cells modify the distribution of desmosomal glycoproteins on their surfaces in a manner consistent with the development of polarized adhesive properties (Cowin et al. 19846; Garrod, 1985, 1986). We have also shown by immunoblotting that the desmosomal proteins of both MDBK and MDCK cells are similar or identical in molecular weight to those of bovine nasal epithelium, whereas the glycoproteins differ in heterogeneity and molecular weight (Suhrbier & Garrod, 1986).

We demonstrate that both MDBK and MDCK cells may be cultured in LCM and induced to form desmosomes by Ca2+ switching. Marked differences have been found between the cell types both in distribution of desmosomal components before the Ca2+ switch and in the time taken for desmosome formation,

Cell culture

Two kidney cell lines were used, Madin-Darby bovine kidney (MDBK) and Madin-Darby canine kidney (MDCK). These cells had been maintained in culture by serial passage or stored frozen in liquid nitrogen. High passage cells (>100 serial passages) were used in these experiments. They were routinely cultured in Eagle’s Minimal Essential Medium (MEM) with HEPES buffer (20mM) supplemented with 10% foetal calf serum (FCS), 100 i.u.ml-1 penicillin, 100 μgml-1 streptomycin, 2 mM-L-glutamine and non-essential amino acids. This will be referred to as standard medium (SM); its [Ca2+] is approximately l·8mM.

The low calcium medium (LCM) consisted of a 3:1 mixture of Dulbecco’s modified Eagle’s medium (DME) and Ham’s F12 nutrient medium without calcium salts (Imperial Laboratories, Salisbury) containing 10% foetal calf serum (FCS) depleted of divalent cations with Chele×100 resin (Bio-Rad Laboratories). The [Ca2+] in this medium was 0·04–0·05mM as determined by atomic absorption spectrophotometry. The medium was supplemented with hydrocortisone (0·5μgml-1), epidermal growth factor (10μgml-1) and cholera toxin (10−10M). These additives are necessary to support optimum growth in LCM.

For culture in LCM, cells in SM were dissociated with 0·25 % trypsin/1 mM-EDTA and pelleted in LCM. After centrifugation the cells were washed in calcium- and magnesium-free (CMF) Hanks’ salt solution, re-pelleted and dispersed in LCM. Cells were plated at a density of 3×105 to 5 ×106 ml-1 onto 13 mm glass coverslips, Thermanox coverslips (Lux) or plastic tissue culture dishes (Nunclon). Calcium-switching experiments were carried out on cells cultured for 4 days in LCM. The Ca2+ switch was accomplished by removing LCM and replacing it with SM.

Antibodies

Antibodies raised in guinea-pig against individual desmosomal components and cytokeratin from bovine nasal epithelium have been described previously (Billiget al. 1982; Cowin & Garrod, 1983; Cowin et al. 1984a,b; Suhrbier & Garrod, 1986). Antisera against desmosomal components were preabsorbed with bovine epidermal keratin to remove any contaminant keratin antibodies. The antisera are named according to the bands they recognize in bovine nasal epithelial desmosomal cores. These are as follows: anti-desmoplakin, anti-175–164K, anti-desmocollin and anti-83/75 K (K represents 103Mr). The anti-desmocollin used here was that referred to as anti-desmocollin I by Suhrbier & Garrod (1986). Each of the antibodies reacts exclusively with desmosomes in bovine nasal epithelium as determined by immunoelectron microscopy. The components recognized by the respective antibodies in MDBK and MDCK cells have been determined by Suhrbier & Garrod (1986).

For some experiments we also used a monoclonal antibody against desmoplakin I, kindly provided by Dr Douglas Hixson.

The anti-cytokeratin antibodies used were a polyclonal guinea-pig, anti-bovine nasal keratin (Billig et al. 1982), which precipitates bands of approximately 40000, 45 000 and 52000 Mr, from MDCK cells (unpublished observations), and monoclonal antibody LE61 (kindly provided by Dr E. B. Lane), which stains MDBK but not MDCK cells.

Fluorescent antibody staining

Cells on glass coverslips were either fixed with ice-cold methanol (5 min) or 3·5 % formaldehyde before staining, or extracted with detergent-containing buffer (cytoskeleton buffer, CSK) as described by Suhrbier & Garrod (1986) before formaldehyde fixation and staining. Cells were washed quickly in PBS then placed in CSK buffer for 15 min at 4°C. They were then washed in PBS and fixed for 20 min in 3·5 % formaldehyde in PBS. After further washing in PBS the cells were treated with 0·1 M-glycine/PBS for 30 min then washed again in PBS plus 0-2% gelatin.

The following staining schedule was used for both methanol-fixed and CSK-extracted cells. Incubation with primary antibody was carried out for 30 min at room temperature. For double labelling guinea-pig anti-prekeratin and mouse monoclonal anti-desmoplakin bodies were applied together. The cells were washed in PBS/gelatin and incubated with rabbit anti-guinea-pig immunoglobulin G (IgG) conjugated with fluorescein isothiocyanate (FITC) for 30 min. Singlelabelled cells were washed in PBS before mounting in PBS/glycerol (1/9, v/v). Double-labelled cells were washed with PBS/gelatin and incubated with sheep anti-mouse IgG conjugated with Texas Red (Amersham International). Fluorescence microscopy was performed with a Zeiss Photomicroscope III equipped with filters for fluorescein and rhodamine. Phase-contrast microscopy was carried out with a Nikon model M inverted microscope.

Electron microscopy

This was carried out as described previously for corneal epithelial cells by Mattey & Garrod (1984).

Cell culture in LCM

Both MDBK and MDCK cells plated in LCM attached and started to spread within 2h. At this stage they were similar in appearance to those plated out in SM. However, by 24 h there were marked differences in morphology. The cells in LCM were less regular in shape and size, and some possessed long overlapping cell processes. Membrane ruffling was apparent in some cells, particularly those at low densities. At high densities the cells quickly formed a confluent monolayer in which they became more regular in shape. Cells in LCM medium continued to proliferate and could be serially passaged in this medium. However, by the fourth or fifth passage they started to become more irregular and fibroblastic in appearance.

MDCK cells in the LCM: desmosomal antigens and electron microscopy

Cells cultured in LCM showed no staining with any of the anti-desmosomal antibodies at regions of intercellular contact. However, anti-desmoplakin, anti-175—164 K and anti-83/75 K gave cytoplasmic staining patterns consisting of a speckling of bright dots throughout the cytoplasm and a large perinuclear accumulation (Fig. 1A,D,G). The components were retained after the cells were extracted with CSK buffer. The smaller cytoplasmic spots were sometimes arranged in linear rows radiating towards the cell periphery (Fig. 2A), suggesting that they may be attached to filaments. Double labelling with anti-cytokeratin and monoclonal anti-desmoplakin I antibody revealed that much of the cytoplasmic desmoplakin staining was indeed associated with bundles of cytokeratin filaments. This was particularly noticeable on thicker cytokeratin bundles that ringed the nucleus (Fig. 2B,C).

Fig. 1.

Calcium-induced changes in desmosomal antigen and cytokeratin staining in MDCK cells. Cells were cultured in LCM for 4 days (A,D,G,J) and transferred into SM for 20 min (B,E,H,K) or 2h (C,F,I,L). They were then fixed in methanol, stained by the indirect immunofluorescence technique and viewed by fluorescence microscopy. The anti-desmosomal antibodies used were anti-desmoplakin (A-C), anti-175–164 K (D-F), anti-83/75 K (G-I) and anti-cytokeratin (J-L), all guinea-pig anti-bovine polyclonals. Note absence of peripheral staining for desmosomal antigens in LCM, and the appearance of peripheral staining within 20 min, which increases in brightness by 2h. Note also the distribution of cytoplasmic staining for all three desmosomal antigens, particularly the juxtanuclear concentrations in LCM, and in SM after 15 min. The cytokeratin also has a juxtanuclear concentration in LCM, from which cytokeratin filaments appear to radiate (J, arrow). Note the change in distribution of cytokeratin seen by 2 h in SM. Bar, 20 μm.

Fig. 1.

Calcium-induced changes in desmosomal antigen and cytokeratin staining in MDCK cells. Cells were cultured in LCM for 4 days (A,D,G,J) and transferred into SM for 20 min (B,E,H,K) or 2h (C,F,I,L). They were then fixed in methanol, stained by the indirect immunofluorescence technique and viewed by fluorescence microscopy. The anti-desmosomal antibodies used were anti-desmoplakin (A-C), anti-175–164 K (D-F), anti-83/75 K (G-I) and anti-cytokeratin (J-L), all guinea-pig anti-bovine polyclonals. Note absence of peripheral staining for desmosomal antigens in LCM, and the appearance of peripheral staining within 20 min, which increases in brightness by 2h. Note also the distribution of cytoplasmic staining for all three desmosomal antigens, particularly the juxtanuclear concentrations in LCM, and in SM after 15 min. The cytokeratin also has a juxtanuclear concentration in LCM, from which cytokeratin filaments appear to radiate (J, arrow). Note the change in distribution of cytokeratin seen by 2 h in SM. Bar, 20 μm.

Fig. 2.

Fluorescence photomicrographs showing association of desmoplakin staining with the cytoskeleton in MDCK cells cultured in LCM after CSK extraction. A. High power showing linear arrangement of some puncta stained with guinea-pig anti-desmoplakin (arrows). B,C. Pair of micrographs showing a cell stained with monoclonal anti-desmoplakin I (B) and guinea-pig anti-cytokeratin (C), showing an incomplete cytoplasmic ring of desmoplakin staining co-distributed with a ring of more intense cytokeratin staining. D,E. A similarly stained cell showing a single juxtanuclear accumulation of desmoplakin (D, arrow) associated with an intense region of anti-cytokeratin staining (E, arrow). Bars: A, 10 μm; B-E, 20μm.

Fig. 2.

Fluorescence photomicrographs showing association of desmoplakin staining with the cytoskeleton in MDCK cells cultured in LCM after CSK extraction. A. High power showing linear arrangement of some puncta stained with guinea-pig anti-desmoplakin (arrows). B,C. Pair of micrographs showing a cell stained with monoclonal anti-desmoplakin I (B) and guinea-pig anti-cytokeratin (C), showing an incomplete cytoplasmic ring of desmoplakin staining co-distributed with a ring of more intense cytokeratin staining. D,E. A similarly stained cell showing a single juxtanuclear accumulation of desmoplakin (D, arrow) associated with an intense region of anti-cytokeratin staining (E, arrow). Bars: A, 10 μm; B-E, 20μm.

The large perinuclear accumulation of desmoplakin often co-localized with a brighter region of cytokeratin staining (Fig. 2D,E). This was not an artifact caused by double labelling since a bright perinuclear region was also found in cells stained for cytokeratin only. It gave the appearance of a nucleation site towards which keratin bundles converged. It was not seen in all cells although it may have been hidden by the network of keratin filaments around the nucleus in some cases. Similar regions were not seen in cells maintained in SM.

When living or formaldehyde-fixed MDCK cells in LCM were stained with anti-desmocollin antibody bright dots of staining were found over the entire cell surface (Fig. 3A). However, treatment of cells with CSK buffer before fixation completely removed desmocollin staining (Fig. 3D). This suggests that the desmocollins on MDCK cells in LCM are not attached to the cytoskeleton. None of the other antibodies stained the surfaces of these cells.

Fig. 3.

Fluorescence micrographs showing anti-desmocollin staining of MDCK cells fixed with formaldehyde (A-C), or CSK-extracted (D-F). Cells were cultured in LCM for 4 days (A,D) and transferred to SM for 20 min (B,E) or 2h (C,F). Note that staining is distributed all over the cell surface in A, but concentrated towards the cell periphery in B and C. Desmocollin staining was completely removed by CSK extraction of cells in LCM (D), but 20 min after switching there was some non-extractable staining at regions of intercellular contact (E) and this was greatly increased by 2 h (F). Bar, 20 μm.

Fig. 3.

Fluorescence micrographs showing anti-desmocollin staining of MDCK cells fixed with formaldehyde (A-C), or CSK-extracted (D-F). Cells were cultured in LCM for 4 days (A,D) and transferred to SM for 20 min (B,E) or 2h (C,F). Note that staining is distributed all over the cell surface in A, but concentrated towards the cell periphery in B and C. Desmocollin staining was completely removed by CSK extraction of cells in LCM (D), but 20 min after switching there was some non-extractable staining at regions of intercellular contact (E) and this was greatly increased by 2 h (F). Bar, 20 μm.

At the ultrastructural level MDCK cells in LCM were quite rounded in shape and only flattened at the periphery (not shown). Occasional junctional zones with a belt of zonula adhaerens were found (Fig. 4A,B) although many cells had little or no contact with their neighbours. No desmosomes or desmosome-like structures were found at the plasma membrane during examination of hundreds of cells. Desmosomal remnants were occasionally seen in the cytoplasm in close association with the cytokeratin system. They usually appeared as small, dense vesicles or, very rarely, as paired plaque structures (Fig. 4C). The internalization of desmosomes in MDBK and MDCK cells is described in the accompanying paper (Mattey & Garrod, 1986).

Fig. 4.

Electron microscopy of MDCK cells in LCM. A. Low-power micrograph showing section cut parallel to the substratum. Note junction at regions (J) and intercellular spaces. B. High-power micrograph showing enlargement of a junctional region. C. High-power micrograph showing internalized desmosomal remnants (arrows). Note the paired plaque structure of the remnant on the left and apparent association of remnants with filaments. Bars: A, 2·0μ1; B, 0-3 μm; C, 0·4μm.

Fig. 4.

Electron microscopy of MDCK cells in LCM. A. Low-power micrograph showing section cut parallel to the substratum. Note junction at regions (J) and intercellular spaces. B. High-power micrograph showing enlargement of a junctional region. C. High-power micrograph showing internalized desmosomal remnants (arrows). Note the paired plaque structure of the remnant on the left and apparent association of remnants with filaments. Bars: A, 2·0μ1; B, 0-3 μm; C, 0·4μm.

MDCK cells transferred into SM: changes in desmosomal antigen distribution and desmosome formation

MDCK cells showed a rapid change in the distribution of desmosomal components when the [Ca2+] was raised. (Although [Ca2+] was routinely raised by placing cells in SM, results similar to those described below were obtained simply by adding T8mM-Ca2+ in CMF Hanks’ solution.) Staining of methanol-fixed cells showed that the desmoplakins, the 175–164 K glycoprotein and the 83/75 K component became detectable at the cell boundaries after 15–20 min although considerable cytoplasmic staining for these components remained (Fig. 1B,E,H). Surface staining also revealed a redistribution of the desmocollins to the periphery (Fig. 3B,C), although some staining remained on the upper surface. Staining of cell boundaries for all the components became gradually brighter over the next few hours (Fig. 1C,F,I) while the cytoplasmic and perinuclear staining gradually diminished. Four hours after the switch little cytoplasmic staining remained.

Extraction of cells with CSK buffer showed that desmocollin staining became non-extractable 15–20 min after raising the [Ca2+] (Fig. 3E,F). This corresponded with the time of first appearance of the other desmosomal components at the cell contact regions.

The keratin filaments also showed a redistribution from a predominantly perinuclear pattern to one that extended throughout the cytoplasm to the cell periphery (Fig. 1J,K,L) taking 24h to achieve a distribution comparable to that of cells maintained continually in SM.

Ultrastructurally, the first indication of desmosome formation occurred at 20 min after the Ca2+ switch when occasional cytoplasmic densities became evident at closely apposed membranes (Fig. 5A). Some intercellular material was seen in these regions as well as fine filaments in the cytoplasm. By 30 min the plaques were better developed, the intercellular space wider and the filaments more noticeable (Fig. 5B). Over the next 90 min the plaques increased in density and more intercellular material accumulated (Fig. 5C). Early stages in desmosome formation could still be seen during this period, but by 2h most of the desmosomes appeared fully mature.

Fig. 5.

A-C. Desmosome formation in MDCK cells cultured in LCM and transferred to SM for 20min (A), 30min (B) or 2h (C). Note parallel membranes and rudimentary plaques in A, denser plaques, intercellular material and intermediate filaments in B and fully mature structure in C. Bar, 0·2μm.

Fig. 5.

A-C. Desmosome formation in MDCK cells cultured in LCM and transferred to SM for 20min (A), 30min (B) or 2h (C). Note parallel membranes and rudimentary plaques in A, denser plaques, intercellular material and intermediate filaments in B and fully mature structure in C. Bar, 0·2μm.

MDBK cells in LCM: desmosomal antigens and electron microscopy

At 24 h after plating in LCM no punctate boundary staining was seen with anti-desmoplakin or anti-175—164 K. Instead bright spots were seen in the cytoplasm, sometimes in a ring around the nucleus (Fig. 6A,D). This staining pattern was retained in cells treated with CSK buffer before fixation. A similar staining pattern was seen with anti-83/75 K, although faint linear staining was also seen at the cell boundaries (Fig. 6G). Anti-desmocollin gave no punctate cytoplasmic staining but a linear boundary pattern was seen where cells came into contact (Fig. 6J). Des-mocollins were also located on the upper cell surface as revealed by staining live or after formaldehyde fixation (Fig. 7A). No surface staining for any other component was obtained. The cytoplasmic staining of anti-desmoplakin anti-175–164 K and anti-83/75 K gradually disappeared over a period of 4 days (Fig. 6B,E,H). However, the linear pattern of anti-83/75 K and anti-desmocollin remained at the cell boundaries (Fig. 6H,K). These antigens were retained after extraction with CSK buffer (Fig. 7B).

Fig. 6.

Calcium-induced changes in desmosomal antigen staining in MDBK cells. Cells were cultured in LCM for 24h (A,D,G,J) or 4 days (B,E,H,K), and the latter transferred into SM for 8h (C,F,I,L). They were then fixed in methanol, stained by the indirect immunofluorescence technique and viewed by fluorescence microscopy. The anti-desmosomal antibodies used were anti-desmoplakin (A-C), anti-175–164 K (D-F), anti-83/75 K (G-I) and anti-desmocollin (J—L), all guinea-pig anti-bovine polyclonals. Note the presence of punctate cytoplasmic staining in A,D and G, which disappears after 4 days in culture (B,E,H), and the presence of staining in intercellular contact regions in G and J, which persists after 4 days in culture (H,K). Note also the absence of any peripheral staining in A and D, and its appearance 8 h after transfer into SM (C,F). The pattern of peripheral staining for 83/75 K (H) and desmocollin (K) is not altered by transfer into SM (I,L). Bar, 20 μm.

Fig. 6.

Calcium-induced changes in desmosomal antigen staining in MDBK cells. Cells were cultured in LCM for 24h (A,D,G,J) or 4 days (B,E,H,K), and the latter transferred into SM for 8h (C,F,I,L). They were then fixed in methanol, stained by the indirect immunofluorescence technique and viewed by fluorescence microscopy. The anti-desmosomal antibodies used were anti-desmoplakin (A-C), anti-175–164 K (D-F), anti-83/75 K (G-I) and anti-desmocollin (J—L), all guinea-pig anti-bovine polyclonals. Note the presence of punctate cytoplasmic staining in A,D and G, which disappears after 4 days in culture (B,E,H), and the presence of staining in intercellular contact regions in G and J, which persists after 4 days in culture (H,K). Note also the absence of any peripheral staining in A and D, and its appearance 8 h after transfer into SM (C,F). The pattern of peripheral staining for 83/75 K (H) and desmocollin (K) is not altered by transfer into SM (I,L). Bar, 20 μm.

Fig. 7.

MDBK ceils cultured in LCM for 4 days and stained with anti-desmocollin antibody: A, after formaldehyde fixation; and B, after CSK extraction. Note the presence of staining over the entire surface as well as at the periphery of cells in A, but only at the periphery in B. Bar, 20μm.

Fig. 7.

MDBK ceils cultured in LCM for 4 days and stained with anti-desmocollin antibody: A, after formaldehyde fixation; and B, after CSK extraction. Note the presence of staining over the entire surface as well as at the periphery of cells in A, but only at the periphery in B. Bar, 20μm.

The staining of cytokeratin in these cells with monoclonal antibody LE61 was diffuse or very finely filamentous, extending throughout the cytoplasm to the cell periphery (Fig. 8A).

Fig. 8.

MDBK cells cultured in LCM for 4 days (A) and transferred to SM for 16 h (B), extracted with CSK buffer and stained with monoclonal anti-cytokeratin antibody, LE61. Note the dramatic alteration in cytokeratin distribution. Bar, 20μm.

Fig. 8.

MDBK cells cultured in LCM for 4 days (A) and transferred to SM for 16 h (B), extracted with CSK buffer and stained with monoclonal anti-cytokeratin antibody, LE61. Note the dramatic alteration in cytokeratin distribution. Bar, 20μm.

Examination of hundreds of cells by electron microscopy revealed no desmosomes. After 7 days in LCM cells still maintained their polarity with a junctional zone in the subapical region (Fig. 9A). Most of this region appeared to consist of a belt of adhaerens-type junctions (zonula adhaerens) (Fig. 9B). They were similar to those seen in control cells although the microfilaments did not occupy such a broad zone and the intercellular space seemed narrower in some regions. Below the zonula adhaerens the membranes of adjoining cells were often closely apposed although in some regions large spaces were found between cells (Fig. 9A). Intermediate filaments were found throughout the cell cytoplasm, but were not organized into thick bundles.

Fig. 9.

Electron microscopy of MDBK cells in LCM. A. Low-power micrograph of section cut perpendicular to the substratum showing apical junctional region (arrow). B. High-power micrograph of junctional region. Bars: A, 1 gm; B, 0·5 μm.

Fig. 9.

Electron microscopy of MDBK cells in LCM. A. Low-power micrograph of section cut perpendicular to the substratum showing apical junctional region (arrow). B. High-power micrograph of junctional region. Bars: A, 1 gm; B, 0·5 μm.

MDBK transferred into SM: appearance of desmosomal staining and desmosome formation

Cells were fixed and stained with anti-desmosomal antibodies at various time intervals after raising the [Ca2+] to 1·8 mM. No change in the pattern of staining was seen until 7–8 h when bright spots started to appear at the regions of contact of cells stained for anti-desmoplakin or anti-175–164 K (Fig. 6C,F). The number of spots gradually increased over the next 48 h. The linear boundary staining of anti-83/75 K and anti-desmocollin in LCM cells remained after switching (Fig. 61,L). The appearance of desmoplakin and anti-175–164 K staining at the cell boundaries was associated with a graduai change in the keratin pattern from a fine filamentous system (see Fig. 8A) to a network containing thicker bundles of filaments attached to desmosomes at the cell membrane (Fig. 8B).

Ultrastructurally, the sequence of events involved in desmosome formation in MDBK cells was identical to that found with MDCK cells, but the timing was very different. The first evidence of desmosome formation, small regions (0-2 μm long) where slight densities appeared on opposing membranes, was found between 7 and 8h after switching. There was also some intercellular material and small accumulations of fine filaments in the cytoplasm of these regions. Structures more recognizable as desmosomes appeared between 8 and 10 h after the Ca2+ switch. The plaque material increased and more filaments became associated with this region.

The intercellular space also widened and there was a further accumulation of intercellular material. By 16 h the desmosomes appeared fully mature and associated with prominent bundles of tonofilaments inserted perpendicularly into the plaque region. Desmosomes always appeared as paired symmetrical plaques in opposed cells; there was no evidence of asymmetrical plaque formation.

Apart from the reappearance of desmosomes the most noticeable feature of Ca2+- switched cells was a broadening of the microfilament network associated with the zonula adhaerens (Fig. 10).

Fig. 10.

Electron microscopy of MDBK cells in SM showing zonula adhaerens region 16 h after Ca2+ switching, and showing broadening of filamentous regions (compare with Fig. 9B). Bar, 0·5μm.

Fig. 10.

Electron microscopy of MDBK cells in SM showing zonula adhaerens region 16 h after Ca2+ switching, and showing broadening of filamentous regions (compare with Fig. 9B). Bar, 0·5μm.

We have demonstrated that neither MDBK nor MDCK cells form desmosomes when cultured at low [Ca2+] (0·05mm), but that desmosome formation is triggered when the [Ca2+] is raised to about T8mM. Regulation of desmosome formation by Ca2+ is therefore a feature of these simple polarized kidney epithelial cells as well as of kératinocytes, which are stratified, complex epithelial cells. Such regulation has also been reported for cultured rat mammary epithelial cells (Bologna et al. 1986) and may be a general phenomenon. However, there is a significant difference between MDBK and MDCK cells in time required for Ca2+-induced desmosome formation. MDCK cells, like kératinocytes (Watt et al. 1984), seem to possess all desmosomal components in LCM and to assemble them rapidly into desmosomes (15-20min) when transferred into SM. However, MDBK cells in LCM did not stain for desmoplakins and the 175-164 K glycoproteins, and may therefore need to synthesize them when transferred into SM, before they can assemble desmosomes. Biochemical experiments necessary to test this hypothesis are in progress. Preliminary evidence shows that desmoplakin synthesis starts about 4h after Ca2+ switching in MDBK cells.

Our results suggest a difference in Ca2+ requirement for formation of desmosomes and zonulae adhaerentes, since the latter were present between cells in LCM. The peripheral staining of MDBK cells in LCM with anti-83/75 K was rather similar to our results with retinal pigmented epithelial cells (Docherty et al. 1984). The latter cells do not possess desmosomes but are richly supplied with zonulae adhaerentes, as shown by electron microscopy (Middleton & Pegrum, 1976) and staining with anti-vinculin antibody (Docherty et al. 1984; Opas et al. 1985). The possibility of an association between the 83 000 MT protein and zonulae adhaerentes as well as with desmosomes is therefore worthy of further investigation.

Peripheral staining of low [Ca2+] MDBK cells with anti-desmocollin antibody was also found, in a pattern very similar to that obtained with anti-83/75 K. This surprising observation may suggest that in MDBK cells a transmembrane association between these two groups of molecules may exist even in LCM since the desmocollins in this peripheral ring were not removed by CSK extraction. The linear staining of the peripheries of MDCK with anti-desmocollin in SM also suggests that desmocollins are present in non-desmosomal membrane as well as in the desmosomes of these cells.

It appears that the desmocollins on the surface of low [Ca2+] MDCK cells are not attached to the cytoskeleton since they are removed by CSK buffer. We suggest that when the [Ca2+] is raised desmocollins undergo a type of patching behaviour (Garrod & Cowin, 1985; Garrod, 1985), concentrating in regions of intercellular contact and becoming attached to the cytoskeleton as the other desmosomal components arrive at the periphery. The cytoplasmic accumulations of the other desmosomal components found in MDCK cells in LCM gradually disappear after the Ca2+ switch. It is possible that these components are translocated to sites of desmosome assembly during rearrangement of the keratin system after the Ca2+ switch. However, it is more likely that the cytoplasmic components seen by immunofluorescence are the remnants of desmosomes internalized during cell passaging and are degraded only when the cells are placed in SM. We have found that pretreatment of low [Ca2+] MDCK cells with cycloheximide for 8h before the Ca2+ switch caused total inhibition of Ca2+-induced desmosome formation, although the cytoplasmic, filament-associated staining and surface staining still remained (unpublished observations). We therefore believe that much, if not all, of the cytoplasmic staining of these cells in LCM is due to internalized desmosomal remnants that are not re-utilized during new desmosome formation. We provide further evidence for this in the accompanying paper (Mattey & Garrod, 1986). This would mean that our fluorescent antibody staining of MDCK cells in LCM fails to demonstrate the desmosomal precursors that are rapidly deployed to the cell surface after Ca2+ switching. We find no evidence for desmosome formation from pre-assembled cytoplasmic aggregates of desmosomal components as suggested for mouse kératinocytes by Jones & Goldman (1985).

Ca2+-induced desmosome formation may not truly parallel the situation in -vivo, although other environmental and/or developmental triggers may lead to a similar sequence of events. The desmosome is a multimolecular complex that is assembled at the surface membrane and depends for its assembly upon the presence of all components, as well as on a permissive concentration of Ca2+. Under some circumstances (as in MDBK cells) the various components may not be synthesized concurrently or targeted simultaneously, in which case the rate of desmosome formation would be controlled by the rate at which the last component(s) appeared at the membrane. The number of desmosomes in a cell may then be reflected by the availability of one limiting component, for example the amount of desmocollin on the cell surface. It seems likely that MDBK and MDCK cells exhibit control of Ca2+- induced desmosome formation at different levels.

With regard to the mechanism by which Ca2+ controls desmosome formation, we have shown that the desmosomal glycoproteins of bovine nasal epithelium bind 45Ca2+ on Western blots, and that the desmocollins but not the 175 000-164 000Mr glycoproteins yield a Ca2+-protected tryptic fragment (Mattey et al. 1986). Furthermore, it has been shown that desmosomal plaques contain a high Mr calmodulin binding protein, desmocalmin (Tsukita & Tsukita, 1985). These results suggest that the role of Ca2+ in desmosome assembly may be complex, involving several molecular events.

This work was supported by the Cancer Research Campaign. We thank Dr Terry Kenny for criticism of the manuscript.

Bilug
,
D.
,
Nicol
,
A.
,
Mcginty
,
R.
,
Cowin
,
P.
,
Morgan
,
J.
&
Garrod
,
D.
(
1982
).
The cytoskeleton and substratum adhesion in chick embryonic corneal epithelial cells
.
J. Cell Sci
.
57
,
51
71
.
Bologna
,
M.
,
Allen
,
R.
&
Dulbecco
,
R.
(
1986
).
Organization of cytokeratin bundles by desmosomes in rat mammary epithelial cells
.
J. Cell Biol
.
102
,
560
567
.
Campbell
,
R. D.
&
Campbell
,
J. H.
(
1971
).
Origin and continuity of desmosomes
. In
Origin and Continuity of Cell Organelles
(ed. J. Reinert &
H.
Ursprung
), pp.
261
298
.
Berlin, Heidelberg, New York
:
Springer-Verlag
.
Cereijido
,
M.
,
Robbins
,
E. S.
,
Dolan
,
W. J.
,
Rotunno
,
C. A.
&
Sabbatini
,
D. D.
(
1978
).
Polarised monolayers formed by epithelial cells on a permeable and translucent support
.
J. Cell Biol
.
77
,
853
880
.
Cowin
,
P.
&
Garrod
,
D. R
, (
1983
).
Antibodies to epithelial desmosomes show wide tissue and species cross-reactivity
.
Nature, Land
.
302
,
148
150
.
Cowin
,
P.
,
Mattey
,
D.
&
Garrod
,
D.
(
1984a
).
Distribution of desmosomal components in the tissue of vertebrates, studied by fluorescent antibody staining
.
J. Cell Sci
.
66
,
119
132
.
Cowin
,
P.
,
Mattey
,
D.
&
Garrod
,
D.
(
1984b
).
Identification of desmosomal surface components (desmocollins) and inhibition of desmosome formation by specific Fab’
.
J. Cell Sci
.
70
,
41
60
.
Docherty
,
R. J.
,
Edwards
,
J. G.
,
Garrod
,
D. R.
&
Mattey
,
D. L.
(
1984
).
Chick embryonic pigmented retina is one of the group of epithelial tissues that lack cytokeratins and desmosomes and have intermediate filaments composed of vimentin
.
J. Cell Sci
.
71
,
61
74
.
Farquhar
,
M. G.
&
Palade
,
G. E.
(
1963
).
Junctional complexes in various epithelia. J’
.
Cell Biol
.
17
,
375
412
.
Garrod
,
D. R.
(
1985
).
The adhesions of epithelial cells
. In
Cellular and Molecular Control of Direct Cell Interactions in Developing Systems
(ed.
H.-J.
Marthy
), NATO Advanced Studies Institute, series A, vol. 99, pp.
43
83
.
New York
:
Plenum
.
Garrod
,
D. R.
(
1986
).
Desmosomes, cell adhesion molecules and the adhesive properties of tissue cells
.
J. Cell Sci. Suppl
.
4
,
221
237
.
Garrod
,
D. R.
&
Cowin
,
P.
(
1986
).
Desmosome structure and function
.
\n Receptors in Tumour Biology
(ed.
C. M.
Chadwick
), pp.
95
-
130
. Cambridge University Press.
Gorbsky
,
G.
&
Steinberg
,
M. S.
(
1981
).
Isolation of the intercellular glycoproteins of desmosomes
.
J. Cell Biol
.
90
,
243
248
.
Hennings
,
H.
,
Cheng
,
M. D.
,
Steinert
,
P.
,
Holbrook
,
K. S.
&
Yuspa
,
S. H.
(
1980
).
Calcium regulation of growth and differentiation of mouse epidermal cells in culture
.
Cell
19
,
245
254
.
Jones
,
J. C. R.
&
Goldman
,
R. D.
(
1985
).
Intermediate filamentsand the initiation of desmosome assembly
.
J. Cell Biol
.
101
,
506
517
.
Jones
,
J. C. R.
,
Goldman
,
A. E.
,
Steinert
,
P. M.
,
Yuspa
,
S.
&
Goldman
,
R. D.
(
1982
).
Dynamic aspects of the supramolecular organization of intermediate filament networks in cultured epidermal cells
.
Cell Motil
.
2
,
197
213
.
Kapprell
,
H.-P.
,
Cowin
,
P.
,
Franke
,
W. W.
,
Ponstingl
,
H.
&
Opferkuch
,
H. J.
(
1985
).
Biochemical characterization of desmosomal proteins isolated from bovine muzzle epidermis: amino acids and carbohydrate composition
.
Eur.J. Cell Biol
.
36
,
217
229
.
Kelly
,
D. E.
(
1966
).
Fine structure of desmosomes, hemidesmosomes and an adepidermal globular layer in developing newt epidermis
.
J. Cell Biol
.
28
,
51
72
.
Kelly
,
D. E.
&
Shienvold
,
F. L.
(
1976
).
The desmosome: Fine structural studies with freezefracture replication and tannic acid staining of sectioned epidermis
.
Cell Tiss. Res
.
172
,
309
323
.
Mattey
,
D. L.
&
Garrod
,
D. R.
(
1984
).
Organization of extracellular matrix by chick embryonic corneal epithelial cells in culture and the role of fibronectin in adhesion
.
J. Cell Sci
.
67
,
171
188
.
Mattey
,
D. L.
&
Garrod
,
D. R.
(
1986
).
Splitting and internalization of the desmosomes of cultured kidney epithelial cells by reduction in calcium concentration
.
J. Cell Sci
.
85
,
113
124
.
Mattey
,
D. L.
,
Suhrbier
,
A.
&
Garrod
,
D. R.
(
1986
).
Recognition, calcium and the control of desmosome formation
.
CibaFdn Symp. 125, Junctional Complexes of Epithelial Cells (in press)
.
Middleton
,
C. A.
&
Pegrum
,
S. M.
(
1976
).
Contacts between pigmented retina epithelial cells in culture
.
J. Cell Sci
.
22
,
371
383
.
Misfeldt
,
D. S.
,
Hamamoto
,
S. T.
&
Pitelka
,
D. R.
(
1976
).
Transepithelial transport in cell culture
.
Proc. natn. Acad. Sci. U.S.A
.
73
,
1212
1216
.
Müeller
,
H.
&
Franke
,
W. W.
(
1983
).
Biochemical and immunological characterization of desmoplakins I and II, the major polypeptides of the desmosomal plaque
.
J. niolec. Biol
.
163
,
647
671
.
Opas
,
M.
,
Turksen
,
K.
&
Kalnins
,
V. I.
(
1985
).
Adhesiveness and distribution of vinculin and spectrin in retinal pigmented epithelial cells during growth and differentiation in vitro
.
Devi Biol
.
107
,
269
280
.
Overton
,
J.
(
1962
).
Desmosome development in normal and reassociating cells of early chick blastoderm
.
Devi Biol
.
4
,
532
548
.
Overton
,
J.
(
1974
).
Cell junctions and their development
.
Prog. Surf. Membr. Sci
.
8
,
161
208
.
Richardson
,
J. C. W.
,
Scalera
,
V.
&
Simmons
,
N. L.
(
1981
).
Identification of two strains of MDCK cells which resemble separate nephron tubule segments
.
Biochim. biophys. Acta
673
,
26
36
.
Skerrow
,
C. J.
(
1985
).
Desmosomal proteins. In Biology of the Integument
, vol.
2
,
Vertebrates
(ed.
J.
Bereiter-Hahn
,
A. G.
Matoltsy
&
K. S.
Richards
), pp.
763
-
787
. Berlin, Heidelberg, New York,
Tokyo
:
Springer-Verlag
.
Suhrbier
,
A.
&
Garrod
,
D.
(
1986
).
An investigation of the molecular components of desmosomes in epithelial cells of five vertebrates. J’
.
Cell Sci
.
81
,
223
242
.
Tsukita
,
S.
&
Tsukita
,
S.
(
1985
).
Desmocalmin - a calmodulin-binding high molecular weight protein isolated from desmosomes
.
J. Cell Biol
.
101
,
2070
2080
.
Watt
,
F. M.
,
Mattey
,
D. L.
&
Garrod
,
D. R.
(
1984
).
Calcium-induced reorganization of desmosomal components in cultured human kératinocytes
.
J. Cell Biol
.
99
,
2211
2215
.