Cell Motility and the Cytoskeleton 17:46-58 (1990)
Effects of Trypsin and Low Ca*+ on Zonulae Adhaerentes Between Chick Retinal Pigment Epithelial Cells in Organ Culture Martin Sandig, Greg J. Hergott, and Vitauts I. Kalnins
Department of Anatomy, University of Toronto, Toronto, Ofifario,Canada The junctional complexes in chick retinal pigment epithelial (RPE) cells in situ contain unusually large zonulae adhaerentes (ZAs) composed of subunits termed zonula adhaerens complexes ( Z A G ) . To determine whether the properties of the ZAs differ between RPE cells which contain Z A G , and MDCK cells which lack ZACs, we investigated the effects of treatment with trypsin and/or low Caz+ by transmission electron microscopy and staining for F-actin. Treatment of RPE cells for I h with trypsin alone has no apparent effect on the morphology of the ZA in either MDCK or RPE cells. In contrast to the ZAs in MDCK cells, which split after 3 min in low C a 2 + , the ZAs in chick RPE cells stay intact even after 2 h, although the intermembrane discs, i.e., the extracellular components of the ZACs, are no longer visible. After 30 min of treatment with trypsin and low C a 2 + , the ZAs split in both cell types. The CMBs start to contract, translocate toward the cell interior, and eventually disappear. This process continues even when the RPE cells are returned to normal medium. New ZAs, composed of ZACs, form between RPE cells 3 h after return to normal medium. These findings suggest that the ZACs in the ZAs of RPE cells are not directly responsible for the increase in resistance to low Ca2+. They also show that the ZA-junctions in RPE cells are not only structurally different from those previously examined, but also behave differently in response to experimental manipulation. Key words: circumferential microfilament bundles, intercellular adhesion, cytoskeleton, junctional complex
INTRODUCTION
Intercellular adhesion is an important characteristic of epithelial cells [Edelman and Thiery, 1985; Obrink, 1986; Fristrom, 19881. In order to maintain the integrity of the epithelial sheet, cells have acquired specialized structures, such as desmosomes [Cowin et al., 19851 and zonula adhaerens (ZA)-junctions [Geiger et al., 19851, each of which contains specific sets of proteins, and they are associated with intermediate filaments and microfilaments respectively and, therefore, link the cytoskeletons of adjacent cells. In most epithelia, both desmosomes and ZA-junctions participate in intercellular adhesion [Farquhar and Palade, 1963; Staehelin, 19741. In birds, however, the retinal pigment epithelial (RPE) cells that form a monolayer between the choroid and the neural retina [Nguyen-Legros, 1978; Zinn and Marmor, 0 1990 Wiley-Liss, Inc.
19791 lack desmosomes [Docherty et al., 19841 but contain extremely large ZA-junctions associated with circumferential microfilament bundles (CMBs) [Kuwabara, 1979; Owaribe and Masuda, 1982, 1986; Sandig and Kalnins, 19881. Furthermore, during embryonic development, the ZA-junctions of chick RPE cells become organized into subunits, the zonula adhaerens complexes
Received January 16, 1990; accepted May 31, 1990. Address reprint requests to V.I. Kalnins, Department of Anatomy, University of Toronto, Medical Sciences Building, Toronto, Ontario, Canada. M5S lA8. Martin Sandig’s present address is Lady Davis Institute for Medical Research, 3755 Chemin C6te Ste-Catherine, Montreal, Quebec, Canada, H3T lE2.
Splitting of ZA-Junctions in RPE Cells (ZACs) [Sandig and Kalnins, 1988, 19901. Each ZAC is composed of an extracellular intermembrane disc (ID) in the space between the two junctional membranes, and of two cytoplasmic plaques, one in either of the two adjacent cells [Sandig and Kalnins, 19881. The IDS probably represent the extracellular domains of adhesion molecules [Geiger et al., 19851, like A-CAM [Volk and Geiger, 1984, 1986a,b], whereas the cytoplasmic plaques contain proteins such as a-actinin and vinculin, which are involved in binding the microfilaments in the CMB to the cell membrane [Geiger et al., 1980, 1981; Craig and Pardo, 1980; Geiger, 1983; Opas and Kalnins, 1985; Blanchard et a]. , 19891. Most studies investigating the mechanism of intercellular adhesion [Takeichi, 1977, 1988; Grunwald et al., 1980; Edelman, 19851 and the assembly and organization of ZA-junctions employed cell lines and cultured cells that, in the case of epithelial cells, express both ZA-junctions and desmosomes [Kartenbeck et al., 1982; Volberg et al., 1986; Green et a]., 1987; O’Keefe et a]. , 19871. Such studies led to the identification of two adhesion systems, a Ca’+-dependent and a Ca’+-independent one [Edelman and Thiery, 1985; Takeichi, 19881. The adhesion molecules so far characterized belong to either one or the other of these two adhesion systems. In contrast to the Ca’ +-independent adhesion molecules, the cadherins, the Ca2+-dependent adhesion molecules, are affected by treatment with trypsin and low Ca2+, and with low Ca*+ alone. Their proteolytic degradation by trypsin, however, can be protected with Ca2+ [Takeichi, 19881. Because of the morphological differences between the ZA-junctions in RPE cells and those in other cell types [Sandig and Kalnins, 198817we wished to establish which type of adhesion system operates in the ZA-junctions of RPE cells, and whether ZA-junctions with and without ZACs respond differently to trypsin and/or low CA2+ treatment. Sheets of chick RPE cells, maintained in an organ culture system [Hergott et al., 19891 were, therefore, compared with MDCK cells. The organ culture system was chosen to avoid exposure to agents that disrupt intercellular junctions prior to experimental treatment. Our results show that both Ca’+-dependent and Ca2 -independent adhesion systems operate in the ZAjunctions of RPE cells. Thus the ZA-junctions in chick RPE cells are not only structurally different from those in MDCK cells, but also behave differently upon experimental manipulation. +
MATERIALS AND METHODS Organ Culture
Organ cultures were established according to a method previously described [Hergott et al., 19891. Eyes
47
were excised from newly hatched chicks and placed in culture medium. The anterior segments, vitreous, and neural retina were gently removed from the enucleated eyes, taking care to avoid damage to the RPE. The eyecups were then cut into smaller pieces (3 mm’), from which the RPE together with the choroid were removed with fine forceps. While submerged in the culture medium, the sheets of RPE on choroid were placed, choroid-side down, onto sterile HA-type Millipore filters (pore size 0.45 pm). Such explants firmly adhered to the filters. They were then transferred to culture dishes containing 2.5 ml of medium, where they were supported at the air liquid interface by sterile wire mesh screens. The cultures were maintained at 37’C in a minimal essential medium containing 10% fetal calf serum, I00 IU penicillin, 0.25 p g fungizone, and 100 pg streptomycin per ml of medium and kept in a humidified atmosphere containing 5% C 0 2 . Treatment With Low Ca2+ and/or Trypsin
In general we followed protocols used by other investigators, who had previously studied the splitting and the reformation of ZA-junctions of MDBK cells [Kartenbeck et al., 1982; Volberg et al., 19861 and keratinocytes [Green et al., 1987; O’Keefe et al., 19871, in order to compare their results with ours. Briefly, after 1 h of incubation, cultures were placed in medium containing either 4 mM or 10 mM EGTA (to chelate Ca*+), 0.05% trypsin, or 0.05% trypsin and 0.53 mM EDTA. Fetal calf serum was omitted from the medium containing the trypsin. After 15 rnin, 30 rnin, 1 h, and 2 h of treatment some of the cultures and untreated controls were fixed for fluorescence and electron microscopy. Other cultures were returned to normal medium and fixed for fluorescence and electron microscopy 30 rnin, 1 h, 3 h, and 6 h later. MDCK cells (ATCC, CCL 34) were grown on round glass coverslips (@7 mm) and treated as above. Fluorescence Microscopy
To label F-actin or vinculin, cells were fixed in 3.5% formaldehyde in phosphate-buffered saline (PBS) for 15 min, briefly rinsed in PBS, and extracted in a buffer containing 0.1 mM PIPES, 1 mM EGTA, 4% polyethylene glycol (PEG 8000), and 0.1% Triton X100, pH 6.9, for 5 rnin at room temperature. After rinsing in PBS, cells were incubated for 30 min with 0.3 p M tetramethylrhodamine isothiocyanate (TR1TC)-conjugated phalloidin (Molecular Probes) in PBS, or with mouse monoclonal anti-vinculin IgG (lot #2, ICN Immunobiologicals, Lisle, IL) at a dilution of 1:lO in PBS. For labelling of vinculin7 cells were then washed in PBS and further incubated for 30 rnin at room temperature with fluorescein isothiocyanate (F1TC)-conjugated
48
Sandig et al.
goat anti-mouse F(ab’)’ (Cappel). After a final wash in PBS, all preparations were mounted on slides in Vinol containing 0.25% 1,4-diazabicycIo-(2,2,2)-octane(DABC0)to reduce photobleaching . The preparations were examined with a ZEISS Photomicroscope I11 equipped with epifluorescence optics and TRITC and FITC specific filters. Electron Microscopy
The cells were fixed for 1 h at room temperature in 2.5% glutaraldehyde containing 2% tannic acid and 0.05 mg/ml saponin in 0.1 M phosphate buffer, postfixed for 30 min at room temperature in 0.5% osmium tetroxide in 0.1 M phosphate buffer, dehydrated in ethanol, and embedded in an Araldite/Epon mixture. Ultrathin sections were cut with a Sorvall MT2 Ultramicrotome, collected on copper grids, stained with uranyl acetate and lead citrate, and viewed with a Hitachi 7000 transmission electron microscope. RESULTS Control Cultures
To verify that our culture conditions did not affect the integrity and ultrastructural appearance of the ZAjunctions, we maintained intact sheets of RPE on choroid for up to 7 h in culture (the time necessary for experimental treatment) and compared their ZA-junctions with those of RPE cells in situ. The F-actin in these control cultures (Fig. lb) is distributed in a hexagonal pattern, corresponding to the CMBs associated with the ZA-junctions. This pattern is identical with that seen in en face preparations of RPE cells (Fig. la) [Turksen and Kalnins, 19871 and similar to that seen in the MDCK cells (Fig. lc). It indicates that the ZA-junctions in the cultured RPE cells have remained intact. The ultrastructure of the apically located ZA-junctions in cultured RPE cells (Fig. lf,g) is also identical to that seen in RPE cells in situ (Fig. ld,e) [Sandig and Kalnins, 19881. The IDS of ZACs are clearly visible in the ZA-junctions of cultured RPE cells (Fig. lg), and the size of the junctions in culture (Fig. lg) corresponds to that seen in RPE cells of an equivalent age in situ (Fig. le) [Sandig and Kalnins, 1988, 19901. Furthermore, the general morphology of the cultured RPE cells (Fig. lf) is similar to that of RPE cells in situ (Fig. ld) [Sandig and Kalnins, 1988, 19901. In cultures the RPE cells were often more columnar than those in situ, and the basement membrane under the RPE cells had frequently folded (Fig. l f ) , whereas in situ it appears fairly flat and straight (Fig. ld). These changes are probably due to elastic recoil of the extracellular matrix during the preparation of the tissue. Although these differences in general morphology were observed, the culture conditions used did not alter the integrity or
the ultrastructural appearance of the ZA-junctions between RPE cells. Treatment With Trypsin at a Normal Ca2+ concentration
To determine whether a Ca’ -dependent adhesion system, suggested by the presence of A-CAM [Volk and Geiger, 19841, is present in the ZA-junctions of RPE cells, we incubated RPE and MDCK cells, where a similar Ca2+-dependentadhesion system is found [Imhof et al., 1983; Behrens et al., 1985; Volberg et al., 19861, for up to 2 h in a medium containing 0.05% trypsin and normal Ca’ levels. This treatment should leave Ca’ dependent adhesion molecules intact and disrupt the Ca’+-independent ones. The F-actin in both the RPE cells (Fig. 2a), and the MDCK cells (Fig. 2b) is distributed in a pattern identical to that seen in control cultures (Fig. lb) and in en face preparations (Fig. la). No dissociation of cells was observed. Moreover, electron micrographs show no change in the morphology of ZAjunctions (Fig. 2c,d), and distinct IDS of ZACs are observed in the ZA-junctions of RPE cells treated with trypsin at normal Ca2+ levels (Fig. 2d). In addition, adjacent RPE cells have maintained contact all along their lateral borders (Fig. 2c). This suggests that in both RPE cells and MDCK cells Ca’+ -dependent adhesion molecules are present in the region of the ZA-junctions, as well as along other parts of their lateral membranes, and that the ZACs in the ZA-junctions of RPE cells are components of this adhesion system. +
+
+
Treatment With Low Ca2+
In order to test 1) whether a Ca’+-independent adhesion system exists in addition to the Ca2+-dependent one, 2) whether the ZACs are indeed components of the Ca’+ -dependent adhesion system, and 3) whether the presence of ZACs in the ZA-junctions between RPE cells influences their response to low Ca’+, we incubated sheets of RPE and MDCK cells in medium containing 4 mM and 10 mM EGTA. This treatment should disrupt the Ca2+-dependent adhesion system, but leave the Ca’+ -independent one intact. Whereas in MDCK cells the distribution of F-actin showed that cells had dissociated after 5 min of treatment with low Ca’+ (Fig. 3b), in RPE cells maintained for up to 2 h in low Ca2+ (Fig. 3a) the distribution of F-actin has a hexagonal pattern similar to that seen in control cultures (Fig. lb) and in en face preparations (Fig. la). Electron micrographs of RPE cells, maintained for 1 h in low Ca’+, show that although the cells have dissociated along most of their lateral borders, their ZAjunctions stay intact (Fig. 3c,d). The IDS of ZACs, however, are no longer visible in the ZA-junctions of treated RPE cells, but are replaced by granular electron-dense
Fig. I . Comparison of RPE in situ (a,d,e) and after 7 h of culture (b,f,g). F-actin labelling of RPE cells in situ (a), in culture (b), and of MDCK cells (c) shows regular hexagonal CMBs in RPE cells (a,b) and more irregular ones in MDCK cells (c). Electron micrographs of RPE cells in situ (d,e) and in culture (f,g) show the localization of ZA-junctions ( m o w s in d,f) and the IDS (arrows in e,g) of ZACs forming these junctions. NR = neural retina. Bar in a and b = 10 pm, in c = 20 km, in d and f = 4 p m , in e and g = 0.1 pm.
50
Sandig et a].
Fig. 2. F-actin distribution in RPE (a) and in MDCK cells (b) after I h of treatment with trypsin at normal Ca2+ levels shows no dissociation in the junctional region. Electron micrographs of RPE cells (c,d) show intact apical ZA-junctions (arrows in c) with IDS (arrows in d) of ZACs forming these junctions. Bar in a = 10 pm, in b = 20 pm, in c = I pm, in d = 0.1 pm.
material (Fig. 3d). Similar results were obtained with RPE sheets cultured for 30 min and for 2 h in low Ca2+ (data not shown). After treatment with low Ca2+, some of the RPE cultures were returned to normal culture medium for 1 h, 3 h, and 5 h to determine whether the ZACs would reform. Even after 5 h of culturing in normal medium, distinct ZACs are not observed in the ZA-junctions of RPE cells (Fig. 3e,f), regardless of the length of our previous treatment with low Ca*+. Instead, dense gran-
ular material (Fig. 3f), similar to that seen in the cells fixed during treatment (Fig. 3d), persists in the junctional space between RPE cells. Treatment With Trypsin at a Low ca2+ concentration
In order to study the splitting and the reformation of ZA-junctions, sheets of RPE cells were incubated in medium containing 0.05% trypsin and 0.53 mM EDTA. This treatment is normally used to dissociate cells (in-
Splitting of ZA-Junctions in RPE Cells
Fig. 3 . F-actin distribution in RPE cells after 1 h (a) and in MDCK cells after 5 min (b) of treatment with low Caz+ shows that only the MDCK cells have dissociated. Electron micrographs of RPE cells after I h of treatment (c,d) and after 1 h of treatment followed by 5 h of incubation in normal medium (e,9 show intact apical ZA-junctions (arrows in c,e), which lack IDS of ZACs (d,f). Note that the RPE cells have separated in regions basal to the ZA-junctions. Bar in a = 10 pm, in b = 20 pm, in c and e = 2 pm, in d and f = 0.1 bm.
51
52
Sandig et a].
cluding MDCK cells) for subculturing [Freshney, 19861, and is known to disrupt both Ca’+-dependent and Ca2+independent adhesion systems. As expected, treatment with trypsin and low Ca’+ results in the splitting of ZA-junctions in RPE cells. This was suggested by results obtained by fluorescence microscopy and confirmed by electron microscopy. The distribution of F-actin after a 30 min treatment suggests that in some areas of the epithelial sheet junctional splitting has just started, since the CMBs have separated from one another and become circular in shape (Fig. 4a), instead of being polygonal as in the intact epithelial sheet (Fig. 4f). In other areas of the RPE sheet, the separation of CMBs has progressed farther, and they have contracted into smaller rings. The CMBs in these regions (Fig. 4b) have separated from those of adjacent cells to varying degrees, resulting in a pattern that consists of rows and islands of rings separated from one another by larger gaps. This pattern, and the presence of fainter lines of F-actin staining along lateral cell membranes, suggests that the cells have dissociated apically, but still remain in contact basal to the former ZA-junctions. After staining of similar cultures with antibodies to vinculin, a pattern identical to that seen after F-actin labelling was obtained (data not shown). This observation indicates that the vinculin-containing cytoplasmic plaques in the ZA-junctions, which anchor the CMBs to the junctional membranes, remain closely associated with the CMBs during their separation after treatment with trypsin at a low Ca2+ concentration. The splitting of the ZA-junctions by this treatment was confirmed at the EM level (Fig. 5a). Higher magnifications show that the CMBs are still attached to the separating membranes in the junctional regions, as the CMBs together with this membrane material gradually contract and translocate away from one another (Fig. 5b). The IDS of ZACs are visible in those parts of the ZA-junctions, which are still intact (Fig. 5b). Even after further separation (Fig. 5c,d), the CMBs remain in contact with membranous material. Some cells can share a ZA-junction with one of the neighbouring cells after having completely separated in the junctional region from others (Fig. 5c), indicating that the splitting of the ZAjunctions is both asynchronous and asymmetricai. Reformation of ZA-Junctions
To determine whether new ZA-junctions, containing ZACs, can form after junctional splitting, sheets of RPE cells were returned to normal medium after treatment with trypsin at a low Ca’+ concentration, During the first hour in normal medium, the preexisting CMBs further contract into small rings from which fine MF bundles appear to radiate, together forming aster-like structures (Fig. 4c).
Electron micrographs show that the RPE cells at this stage (Fig. 6a) are in lateral contact with their neighbours, but have not formed new ZA-junctions (Fig. 6a). In the apical cytoplasm electron-dense areas can be seen (Fig. 6a,b). These correspond to sections through the preexisting contracted CMBs, and are seen in Figure 4c and in the inset of Figure 6a in an en face view at the light microscope level. The electron-dense material is closely associated with membranous vesicles and invaginations of the plasma membrane (Fig. 6b,c). Apical cytoplasmic processes, containing MF bundles, fan out from the apical regions constricted by the CMBs (Fig. 6). These MF-containing apical processes might be the rays that appear to radiate from the contracted rings seen after F-actin labelling (Fig. 4c). Stress fibers were not detected in these cells. By 3 h in normal medium, following 30 min treatment with trypsin at a low Ca2+ concentration, F-actin labelling shows that almost all of the preexisting contracted CMBs have disappeared. Instead, straight fluorescent lines are seen (Fig. 4d), suggesting the formation of new ZA-junctions between many of the adjacent cells. After 5 h in normal medium7 similar F-actin staining is seen around all of the cells in the culture, giving a polygonal pattern (Fig. 4e). This pattern, however, differs from that seen in control cultures (Fig. 4f) in that the cells are irregularly (Fig. 4e) rather than hexagonally (Fig. 4f) shaped. Electron micrographs confirm that in RPE cells after 5 h in normal medium, following 30 rnin and 1 h treatments with trypsin at a low Ca’+ concentration, apical ZA-junctions had indeed reformed (Fig. 7a). Distinct IDS of Z A G , however, are seen only in the newly formed ZA-junctions of cells, which had been previously treated for 30 rnin (Fig. 7b), but not in ZA-junctions of RPE cells, which had been treated for 1 h (Fig. 7c). DISCUSSION
In summary, we have shown that, whereas the ZAjunctions of MDCK cells split upon treatment both with low Ca2+ and with a mixture of trypsin and low Ca2+, the ZA-junctions of RPE cells only split when treated with a mixture of trypsin and low Ca2+. Epithelial ZAjunctions can, therefore, be split and reformed in cells maintained on their native basement membrane, i.e. in cells which, unlike MDCK cells, have not been previously dissociated and plated on glass. Treatment with trypsin alone or with low Ca2+ alone is insufficient to split the ZA-junctions in RPE cells, although the IDS of ZACs disappear upon treatment with low Ca2+. We have further shown that after junctional splitting in RPE cells, the CMBs contract asymmetrically into small rings, which stay associated with membranous material
Fig. 4. RPE treated with trypsin at low Caz+ levels showing the distribution of F-actin after a 30 rnin treatment (a). RPE cells start to dissociate apically, and their CMBs retract (arrows). In other areas (b) the dissociation has progressed asymmetrically and the CMBs have contracted into smaller rings (arrows). After a 30 rnin treatment followed by 1 h in normal medium (c) the contraction of the CMBs(arrows) continues. After a 30 rnin treatment followed by 3 h in nor-
ma1 medium (d) new ZA-junctions (arrows) have started to form. Note the remnant of the preexisting CMB in one of the RPE cells (open arrow). After a 30 rnin treatment followed by 5 h in notmal medium (e) the new CMBs can be seen around all of the RPE cells forming an irregular pattern compared to the hexagonal one in the untreated controls (0.Bar = 10 +m.
54
Sandig et al.
Fig. 5 . Electron micrographs of RPE cells and their junctional regions after treatment with trypsin at low Ca*+ levels at early (a,b) and later (c-e) stages of junctional splitting. Note that IDS of ZACs are still visible in those regions of the junction, which have not separated (arrowhead in b). The CMBs, associated with membranous material, progressively retract from one another, and are seen as electron-dense
areas (arrows in bd).Similar material is seen in cells that have lost some of their neighbours (arrow in e). The insets in d and e show the corresponding distribution of F-actin in an en face view at the light microscope level. Bar in a and e = 3 pm, in b = 0.2 pm, in c = 0.4 pm, in d = 0.5 pm.
Splitting of ZA-Junctions in RPE Cells
Fig. 6. Electron micrographs of RPE cells after 1 h incubation in normal medium following 30 min treatment with trypsin at low Ca2+ levels. Electron-dense areas in the apical cytoplasm of RPE cells correspond to grazing (a,b) and cross (c) sections through the contracted CMBs (inset in a). Many of these electron-dense areas (arrows) are
55
closely associated with membranous vesicles and invaginations from the cell surface. Numerous apical processes radiate from a cytoplasmic protrusion above the contracted CMB. Note that ZA-junctions have not reformed at this stage. The area indicated in a is shown at a higher magnification in b. Bar in a = 2 pm, in b and c = 0.5 pm.
56
Sandig et al.
Fig. 7 . Electron micrographs of RPE cells after 3 h (a,b) and 5 h (c) of incubation in normal medium, following 30 min (a,b) and I h (c) treatment with trypsin at low Caz+ levels. In the newly formed apical ZA-junctions (arrows in a, and panels b and c) the IDS (arrows in b)
of ZACs are present after treatment for 30 min (b), but not after 1 h (c) of treatment. The inset in a shows the corresponding distribution of F-actin in an en face view at the light microscope level. 3ar in a = 2 p n , in b and c = 0.1 pm.
as they are internalized. After junctional splitting, new ZA-junctions, composed of Z A G , form upon return to normal medium. These results can be explained by hypothesizing that the Ca’+-dependent adhesion molecules in RPE cells are less sensitive to low Ca2+ than those in MDCK cells. It is more likely, however, that two adhesion systems, a Ca’+ -dependent and a Ca2+-independent one, are differentially expressed along the lateral cell membranes of RPE cells. Such Ca*+-dependent and Caz+independent adhesion systems have been identified in other cell types [Takeichi, 1977; Grunwald et al., 1980; Obnnk, 1986; Takeichi, 19881. The Ca2+-dependent adhesion system can be disrupted by treating cells with trypsin and Ca2+-chelators, or with Ca2+-chelators alone. The proteolytic degradation of the adhesion molecules in this system by trypsin, however, can be prevented by addition of Ca2+. The Ca’+-independent system, in contrast, can be disrupted by incubating cells with trypsin alone [Takeichi, 1977, 19881. We have shown that upon treatment with trypsin at normal Ca2+ levels, RPE cells and MDCK cells do not
dissociate, and that the ZA-junctions in RPE cells remain intact and are morphologically identical to those in situ. These results indicate that Ca’ -dependent adhesion molecules are present along the lateral cell membranes and within the ZA-junction, since in this case proteolytic degradation is prevented by the presence of exogenous Ca2+ [Takeichi, 1977, 19881. Since the IDS of ZACs remain in ZA-junctions of RPE cells treated with trypsin, the ZACs may be components of this Ca2+-dependent adhesion system. It has been reported previously that A-CAM, a Caz+-dependent adhesion molecule [Volk and Geiger, 1986a,b], is indeed present in embryonic chick RPE cells [Volk and Geiger, 19841. We have confirmed these findings in the RPE cells of newly hatched chicks (data not shown). Furthermore, in MDCK cells a different Ca’+-dependent adhesion molecule, Arc- 1, which seems to be identica1 to L-CAM and E-cadherin [Takeichi, 19881, has been localized [Imhof et al., 1983; Behrens et al., 19851. We have also shown that in contrast to ZA-junction in MDCK cells, the ZA-junctions in RPE cells do not split when maintained for up to 2 h in low Ca2+. The IDS of ZACs, however, are no longer +
Splitting of ZA-Junctions in RPE Cells
visible and, instead, irregular electron-dense material is present in the junctional space. In addition, upon exposure to low Ca2+ cells dissociate laterally basal to the ZA-junctions, so that the ZA-junctions appear to be the only remaining contact areas. These findings can be explained as follows. The exposure to a low Ca2+ concentration disrupts the binding by the Ca2+-dependent adhesion molecules in both RPE and MDCK cells. As a result RPE cells dissociate along their lateral borders and lose the IDS (components of the Ca2+-dependent adhesion molecules in the ZA-junctions) of ZACs in their ZA-junctions. However, in addition to the Ca2+-dependent adhesion system (the IDS of ZACs) Ca2+-independent adhesion molecules are present in the ZA-junctions of RPE cells, which keep the ZA-junctions intact during the low Ca2+ treatment. This suggests that in RPE cells the Ca2 -independent adhesion molecules are present only in the ZA-junctions, whereas the Ca2+-dependent ones are distributed along the entire length of the lateral membranes. In contrast MDCK cells appear to lack the Ca2+-independent adhesion molecules, since these cells dissociate completely on exposure to low Ca2+. In order to split the ZA-junctions of RPE cells, treatment with both trypsin and low C a 2 + , a mixture normally used to dissociate cells for subculturing [Freshney, 19861, is necessary. This observation further suggests that in RPE cells both the Ca2+-dependent and the Ca’ -independent adhesion systems, disrupted by this treatment, are present. These findings are consistent with the results obtained in a study in which E7 chick RPE was subjected to various treatments to establish optimal conditions for cell dissociation [Vielkind and Crawford, 19881. This study showed that treatments with low Ca2+ alone, or trypsin alone, do not result in a high yield of dissociated cells for plating. Although the pattern of junctional splitting in RPE cells was generally similar to that seen in MDBK [Kartenback et al., 1982; Volberg et al., 19861 and MDCK cells, the time and the type of treatment required were quite different. The ZA-junctions in MDCK and MDBK cells split very rapidly upon treatment with low Ca2+ (after 3-5 min), whereas RPE cells required trypsin in addition to low Ca2+ and had to be treated between 30 min and I h for junctions to split. In MDBK cells, the CMBs together with the junctional plaques completely detach from the cell membrane, contract, and translocate toward the cell interior [Volberg et al., 19861. The CMBs of RPE cells also contract after junctional splitting induced by treatment with trypsin and low Ca2+. However, since they are still attached at least partially to the cell membrane, this contraction leads to the constriction of the apical portions of the cells, causing the apical processes to fan out. The contracted CMBs of RPE cells were associated with irregularly shaped ves+
+
57
ides, some of which may be cross sections of invaginations of the cell membrane. Similar vesicles associated with electron-dense junctional plaque material have been observed in MDBK cells near the internalized CMBs [Kartenbeck et al., 19821. In MDBK cells many of these, however, appeared to be associated with desmosomal plaque proteins and originated from the internalization of desmosomes [Kartenbeck et a]., 19821, absent in chick RPE cells [Docherty et al., 19841. In both RPE cells and in MDBK cells, the contraction of the CMBs continues upon return to normal medium after treatment with trypsin and low C a 2 + , and with low Ca2+, respectively. The CMBs together with the vinculin-containing plaque material finally disappear after approximately 3 h in normal medium [Volberg et a]., 19861. In both RPE cells and MDBK cells new ZA-junctions form after recovery from treatment. Junction formation in MDBK cells occurred within 30-45 min after return to normal medium and was essentially completed after 2 h [Volberg et a]., 19861. In RPE cells, the formation of new junctions did not start before 1 h, and was not completed until 4-5 h, after return to normal medium. These differences in the time course of the formation of new ZA-junctions might be due to a higher speed of spreading of MDBK cells on glass compared to the rate of spreading of RPE cells on basement membrane, or to differences in the rate of synthesis and/or assembly of new junctional proteins between the two cell types. It is interesting that after recovery, IDS of ZACs are present in the newly formed ZA-junctions (after previous treatment with trypsin in low Ca2+),whereas they do not reform in the intact ZA-junctions after loss during exposure to low Ca2+ alone. These observations suggest that junctional splitting might be a signal for activating the synthesis and assembly of the molecular components of Z A G . Alternatively, the turnover of molecules in the intact junction may be slower than the time required to assemble new ZA-junctions, so that we were unable to observe the eventual appearance of ZACs in the ZAjunctions of cells, which had been treated with low Ca2+, in the time frame examined. In conclusion, we have shown that the ZA-junctions in chick RPE cells are resistant to treatment with low Ca2+. In order to split these junctions, treatment with both low Ca2+ and trypsin is required. These observations suggest that in chick RPE cells the ZA-junctions, composed of Z A G , might mediate stronger adhesion compared to the ZA-junctions in other epithelial cells. Stronger adhesion between RPE cells might be necessary since they lack desmosomes [Docherty et al., 19841, which provide the structural basis for strong adhesion in other epithelia.
58
Sandig et al.
ACKNOWLEDGMENTS M.S. is an Ontario Graduate Scholar, and G.J.H. is a recipient of a Dental Fellowship from the Medical Research Council of Canada. This work was supported by a grant from the RP Eye Research Foundation to V.I.K. REFERENCES Behrens, J., Birchmeier, W., Goodman, S.L., and Imhof, B.A. (198s): Dissociation of Madin-Darby canine kidney epithelial cells by the monoclonal antibody anti-arc-I: Mechanistic aspects and identification of the antigen as a component related to uvomorulin. J. Cell Biol. lOl:I307-13lS. Blanchard, A., Ohmian, V., and Critchley, D. (1989): The structure and function of a-actinin. J. Muscle Res. Cell Motil. 10:280289. Cowin, P., Franke, W.W., Kapprell, H.P., and Kartenbeck, J. (1983: The desmosome intermediate filament complex. In Edelman, G.M., and Thiery, J.P. (eds.): “The Cell in Contact.” New York: Neuroscienc Inst. Publ., J . Wiley & Sons, pp. 427-460. Craig, S.W., and Pardo, J.V. (1980): Alpha-actinin localization in the junctional complex of intestinal epithelial cells. J. Cell Biol. 80:203 -2 10. Docherty, R.J., Edwards, J.G., Garrod, D.R., and Mattey, D.L. (1984): Chick embryonic pigmented retina is one of the group of epithelioid tissues that lack cytokeratins and desmosomes and have intermediate filaments composed of vimentin. J. Cell Sci. 71:61-74. Edelman, G.M. (198s): Specific cell adhesion in histogenesis and morphogenesis. In Edelman, G.M., and Thiery, J.P. (eds.): “The Cell in Contact.” New York: Neurosciences Publ. J. Wiley & Sons Inc., pp. 139-168. Edelman, G.M., and Thiery, J.P. (1985): “The Cell in Contact.” New York: Neurosci. Inst. Publ., J. Wiley & Sons. Fxquhar, M.G., and Palade, G.E. (1963): Junctional complexes in various epithelia. J. Cell Biol. l7:375-409. Freshney, R.I. (1986): “Animal Cell Culture.” Oxford: IRL Press. Fristrom, D. (1988): The cellular basis of epithelial morphogenesis, A review. Tissue Cell 20:645-690. Geiger, B, (1983): Membrane-cytoskeleton interaction. Biochim. Biophys. Acta 737:305-341. Geiger, B., Tokuyasu, K.T., Dutton, A.E., and Singer, S.J. (1980): Vinculin, an intercellular protein localized at specialized sites where microfilament bundles terminate at cell membranes. F’roc. Natl. Acad. Sci. U.S.A. 77:4127-4131. Geiger, B., Dutton, A.H., Tokuyasu, K.T., and Singer, S.J. (1981): Immunoelectron microscope studies of membrane-microfilament interactions: Distribution of alpha-actinin, tropomyosin, and vinculin in intestinal epithelial brush border and chicken gizzard smooth muscle cells. .I. Cell Biol. 91:614-628. Geiger, B., Avnur, Z., Volberg, T., and Volk, T. (1985): Molecular domains of adherens junctions. ln Edelman, G.M., and Thiery, J.P. (eds.): “The Cell in Contact.” New York: Neuroscience lnst. Publ. J. Wiley & Sons, pp. 461-490. Green, K .J, , Geiger, B ., Jones, .l.CR., TaIian, J.C., and Goldman, R.D. (1987): The relationship between intermediate filaments and microfilaments before and during the formation of desmosomes and adherens-type junctions in mouse epidermal keratinocytes. J. Cell Biol. 104:1389-1402. Grunwald, G.B., Geller, R.L., and Lilien, J. (1980): Enzymatic dissection of embryonic cell adhesive mechanisms. J. Cell Biol. 8S1766-776. Hergott, G.J., Sandig, M., and Kalnins, V.1. (1989): Cytoskeletal organization of migrating retinal pigment epithelial cells during wound healing in organ culture. Cell Motil. Cytoskeleton 13: 83-93.
Imhof, B . A . , Vollmers, H.P., Goodman, S.L., and Birchmeier, W. 91983): Cell-cell interaction and polarity of epithelial cells: Specific perturbation using a monoclonal antibody. Cell 35: 667-675. Kartenbeck, J., Schmid, E., Franke, W.W., and Geiger, B. (1982): Different modes of internalization of proteins associated with adherens junctions and desmosomes: Experimental separation of lateral contacts induces endocytosis of desmosomal plaque material. EMBO J. 6:725-732. Kuwabara, T. (1979): Species differences in the retinal pigment epithelium. In Zinn, K.M., and Marmor, M.F. (eds.): “The Retinal Pigment Epithelium.” Cambridge: Harvard University Press, pp. 58-82. Nguyen-Legros, J. (1978): Fine structure of the pigment epithelium in the vertebrate retina. lnt. Rev. Cytol. 7:287-328. Obrink, B. (1986): Epithelial cell adhesion molecules. Exp. Cell Res. I 63: 1-2 I . O’Keefe, E., Briggaman, R.A., and Herman, B. (1987): Calciuminduced asembly of adherens junctions in keratinocytes. J. Cell Biol. lOS:807-817. Opas, M., and Kalnins, V.I. (198s): Spatial distribution of cortical proteins in cells of epithelial sheets. Cell Tissue Res. 239: 4SI-4S4. Owaribe, K . , and Masuda, H. (1982): Isolation and characterization of circumferential microfilament bundles from retinal pigment epithelial cells. J. Cell Biol. 9S:310-315. Owaribe, K., and Masuda, H. (1986): Organization of microfilaments and intermediate filaments in retinal pigmented epithelial cells. In Ishikawa, H., Hatano, S . , and Sato, H. (eds.): “Cell Motility: Mechanism and Regulation.” New York: Alan R. Liss, Inc., pp. S077.514. Sandig, M., and Kalnins, V.1. (1988): Subunits in zonulae adhaerentes and striations in the associated circumferential microfilament bundles in chicken retinal pigment epithelial cells in situ. Exp. Cell Res. 17S:1-14. Sandig, M., and Kalnins, V.1. (1990): Morphological changes in the zonula adhaerens during embryonic development of chick retinal pigment epithelial cells. Cell Tissue Res. 2S9:45S-461. Staehelin, L.A. (1974): Structure and function of intercellular junctions. Int. Rev. Cytol. 39:191-283. Takeichi, M. (1977): Functional correlation between cell adhesive properties and some cell surface proteins. J. Cell Biol. 7.5: 464 -447. Takeichi, M. (1988): The cadherins: Cell-cell adhesion molecules controlling animal morphogenesis. Development 102:63945S. Turksen, K., and Kalnins, V.I. (1987): The cytoskeleton of chick retinal pigment epithelial cells in situ. Cell Tissue Res, 248: 9s-101, Vielkind, U., and Crawford, B.J. (1988): Evaluation of different procedures for the dissociation of retinal pigmented epithelium into single viable cells. Pigment Cell Res. 1:419-433. Volberg, T., Geiger, B., Kartenbeck, J., and Franke, W.W. (1986): Changes in membrane-microfilament interaction in intercellular adherens junctions upon removal of extracellular Ca ions. J . Cell Biol. l02:1832-l842. Volk, T., and Geiger, B. (1984): A 135-kD membrane protein of intercellular adherens junctions. EMBO J. 3:2249-2260. Volk, T., and Geiger, B. (1986a): A-CAM: A 135-kD receptor of intercellular adherens junctions. I. lmmunoelectron microscopic localization and biochemical studies. J. Cell Biol. 103: 1441-1450. Volk, T., and Geiger, B. (1986b): A-CAM: A 135-kD receptor of intercellular adherens junctions. 11. 4ntibody-mediated modulation of junction formation. J. Cell Biol. 103:14S1-1464. Zinn, K.M., and Marmor, M.F. (1979): ‘‘The Retinal Pigment Epithelium, ” Cambridge: Harvard University Press.
Effects of Trypsin and Low Ca*+ on Zonulae Adhaerentes Between Chick Retinal Pigment Epithelial Cells in Organ Culture Martin Sandig, Greg J. Hergott, and Vitauts I. Kalnins
Department of Anatomy, University of Toronto, Toronto, Ofifario,Canada The junctional complexes in chick retinal pigment epithelial (RPE) cells in situ contain unusually large zonulae adhaerentes (ZAs) composed of subunits termed zonula adhaerens complexes ( Z A G ) . To determine whether the properties of the ZAs differ between RPE cells which contain Z A G , and MDCK cells which lack ZACs, we investigated the effects of treatment with trypsin and/or low Caz+ by transmission electron microscopy and staining for F-actin. Treatment of RPE cells for I h with trypsin alone has no apparent effect on the morphology of the ZA in either MDCK or RPE cells. In contrast to the ZAs in MDCK cells, which split after 3 min in low C a 2 + , the ZAs in chick RPE cells stay intact even after 2 h, although the intermembrane discs, i.e., the extracellular components of the ZACs, are no longer visible. After 30 min of treatment with trypsin and low C a 2 + , the ZAs split in both cell types. The CMBs start to contract, translocate toward the cell interior, and eventually disappear. This process continues even when the RPE cells are returned to normal medium. New ZAs, composed of ZACs, form between RPE cells 3 h after return to normal medium. These findings suggest that the ZACs in the ZAs of RPE cells are not directly responsible for the increase in resistance to low Ca2+. They also show that the ZA-junctions in RPE cells are not only structurally different from those previously examined, but also behave differently in response to experimental manipulation. Key words: circumferential microfilament bundles, intercellular adhesion, cytoskeleton, junctional complex
INTRODUCTION
Intercellular adhesion is an important characteristic of epithelial cells [Edelman and Thiery, 1985; Obrink, 1986; Fristrom, 19881. In order to maintain the integrity of the epithelial sheet, cells have acquired specialized structures, such as desmosomes [Cowin et al., 19851 and zonula adhaerens (ZA)-junctions [Geiger et al., 19851, each of which contains specific sets of proteins, and they are associated with intermediate filaments and microfilaments respectively and, therefore, link the cytoskeletons of adjacent cells. In most epithelia, both desmosomes and ZA-junctions participate in intercellular adhesion [Farquhar and Palade, 1963; Staehelin, 19741. In birds, however, the retinal pigment epithelial (RPE) cells that form a monolayer between the choroid and the neural retina [Nguyen-Legros, 1978; Zinn and Marmor, 0 1990 Wiley-Liss, Inc.
19791 lack desmosomes [Docherty et al., 19841 but contain extremely large ZA-junctions associated with circumferential microfilament bundles (CMBs) [Kuwabara, 1979; Owaribe and Masuda, 1982, 1986; Sandig and Kalnins, 19881. Furthermore, during embryonic development, the ZA-junctions of chick RPE cells become organized into subunits, the zonula adhaerens complexes
Received January 16, 1990; accepted May 31, 1990. Address reprint requests to V.I. Kalnins, Department of Anatomy, University of Toronto, Medical Sciences Building, Toronto, Ontario, Canada. M5S lA8. Martin Sandig’s present address is Lady Davis Institute for Medical Research, 3755 Chemin C6te Ste-Catherine, Montreal, Quebec, Canada, H3T lE2.
Splitting of ZA-Junctions in RPE Cells (ZACs) [Sandig and Kalnins, 1988, 19901. Each ZAC is composed of an extracellular intermembrane disc (ID) in the space between the two junctional membranes, and of two cytoplasmic plaques, one in either of the two adjacent cells [Sandig and Kalnins, 19881. The IDS probably represent the extracellular domains of adhesion molecules [Geiger et al., 19851, like A-CAM [Volk and Geiger, 1984, 1986a,b], whereas the cytoplasmic plaques contain proteins such as a-actinin and vinculin, which are involved in binding the microfilaments in the CMB to the cell membrane [Geiger et al., 1980, 1981; Craig and Pardo, 1980; Geiger, 1983; Opas and Kalnins, 1985; Blanchard et a]. , 19891. Most studies investigating the mechanism of intercellular adhesion [Takeichi, 1977, 1988; Grunwald et al., 1980; Edelman, 19851 and the assembly and organization of ZA-junctions employed cell lines and cultured cells that, in the case of epithelial cells, express both ZA-junctions and desmosomes [Kartenbeck et al., 1982; Volberg et al., 1986; Green et a]., 1987; O’Keefe et a]. , 19871. Such studies led to the identification of two adhesion systems, a Ca’+-dependent and a Ca’+-independent one [Edelman and Thiery, 1985; Takeichi, 19881. The adhesion molecules so far characterized belong to either one or the other of these two adhesion systems. In contrast to the Ca’ +-independent adhesion molecules, the cadherins, the Ca2+-dependent adhesion molecules, are affected by treatment with trypsin and low Ca2+, and with low Ca*+ alone. Their proteolytic degradation by trypsin, however, can be protected with Ca2+ [Takeichi, 19881. Because of the morphological differences between the ZA-junctions in RPE cells and those in other cell types [Sandig and Kalnins, 198817we wished to establish which type of adhesion system operates in the ZA-junctions of RPE cells, and whether ZA-junctions with and without ZACs respond differently to trypsin and/or low CA2+ treatment. Sheets of chick RPE cells, maintained in an organ culture system [Hergott et al., 19891 were, therefore, compared with MDCK cells. The organ culture system was chosen to avoid exposure to agents that disrupt intercellular junctions prior to experimental treatment. Our results show that both Ca’+-dependent and Ca2 -independent adhesion systems operate in the ZAjunctions of RPE cells. Thus the ZA-junctions in chick RPE cells are not only structurally different from those in MDCK cells, but also behave differently upon experimental manipulation. +
MATERIALS AND METHODS Organ Culture
Organ cultures were established according to a method previously described [Hergott et al., 19891. Eyes
47
were excised from newly hatched chicks and placed in culture medium. The anterior segments, vitreous, and neural retina were gently removed from the enucleated eyes, taking care to avoid damage to the RPE. The eyecups were then cut into smaller pieces (3 mm’), from which the RPE together with the choroid were removed with fine forceps. While submerged in the culture medium, the sheets of RPE on choroid were placed, choroid-side down, onto sterile HA-type Millipore filters (pore size 0.45 pm). Such explants firmly adhered to the filters. They were then transferred to culture dishes containing 2.5 ml of medium, where they were supported at the air liquid interface by sterile wire mesh screens. The cultures were maintained at 37’C in a minimal essential medium containing 10% fetal calf serum, I00 IU penicillin, 0.25 p g fungizone, and 100 pg streptomycin per ml of medium and kept in a humidified atmosphere containing 5% C 0 2 . Treatment With Low Ca2+ and/or Trypsin
In general we followed protocols used by other investigators, who had previously studied the splitting and the reformation of ZA-junctions of MDBK cells [Kartenbeck et al., 1982; Volberg et al., 19861 and keratinocytes [Green et al., 1987; O’Keefe et al., 19871, in order to compare their results with ours. Briefly, after 1 h of incubation, cultures were placed in medium containing either 4 mM or 10 mM EGTA (to chelate Ca*+), 0.05% trypsin, or 0.05% trypsin and 0.53 mM EDTA. Fetal calf serum was omitted from the medium containing the trypsin. After 15 rnin, 30 rnin, 1 h, and 2 h of treatment some of the cultures and untreated controls were fixed for fluorescence and electron microscopy. Other cultures were returned to normal medium and fixed for fluorescence and electron microscopy 30 rnin, 1 h, 3 h, and 6 h later. MDCK cells (ATCC, CCL 34) were grown on round glass coverslips (@7 mm) and treated as above. Fluorescence Microscopy
To label F-actin or vinculin, cells were fixed in 3.5% formaldehyde in phosphate-buffered saline (PBS) for 15 min, briefly rinsed in PBS, and extracted in a buffer containing 0.1 mM PIPES, 1 mM EGTA, 4% polyethylene glycol (PEG 8000), and 0.1% Triton X100, pH 6.9, for 5 rnin at room temperature. After rinsing in PBS, cells were incubated for 30 min with 0.3 p M tetramethylrhodamine isothiocyanate (TR1TC)-conjugated phalloidin (Molecular Probes) in PBS, or with mouse monoclonal anti-vinculin IgG (lot #2, ICN Immunobiologicals, Lisle, IL) at a dilution of 1:lO in PBS. For labelling of vinculin7 cells were then washed in PBS and further incubated for 30 rnin at room temperature with fluorescein isothiocyanate (F1TC)-conjugated
48
Sandig et al.
goat anti-mouse F(ab’)’ (Cappel). After a final wash in PBS, all preparations were mounted on slides in Vinol containing 0.25% 1,4-diazabicycIo-(2,2,2)-octane(DABC0)to reduce photobleaching . The preparations were examined with a ZEISS Photomicroscope I11 equipped with epifluorescence optics and TRITC and FITC specific filters. Electron Microscopy
The cells were fixed for 1 h at room temperature in 2.5% glutaraldehyde containing 2% tannic acid and 0.05 mg/ml saponin in 0.1 M phosphate buffer, postfixed for 30 min at room temperature in 0.5% osmium tetroxide in 0.1 M phosphate buffer, dehydrated in ethanol, and embedded in an Araldite/Epon mixture. Ultrathin sections were cut with a Sorvall MT2 Ultramicrotome, collected on copper grids, stained with uranyl acetate and lead citrate, and viewed with a Hitachi 7000 transmission electron microscope. RESULTS Control Cultures
To verify that our culture conditions did not affect the integrity and ultrastructural appearance of the ZAjunctions, we maintained intact sheets of RPE on choroid for up to 7 h in culture (the time necessary for experimental treatment) and compared their ZA-junctions with those of RPE cells in situ. The F-actin in these control cultures (Fig. lb) is distributed in a hexagonal pattern, corresponding to the CMBs associated with the ZA-junctions. This pattern is identical with that seen in en face preparations of RPE cells (Fig. la) [Turksen and Kalnins, 19871 and similar to that seen in the MDCK cells (Fig. lc). It indicates that the ZA-junctions in the cultured RPE cells have remained intact. The ultrastructure of the apically located ZA-junctions in cultured RPE cells (Fig. lf,g) is also identical to that seen in RPE cells in situ (Fig. ld,e) [Sandig and Kalnins, 19881. The IDS of ZACs are clearly visible in the ZA-junctions of cultured RPE cells (Fig. lg), and the size of the junctions in culture (Fig. lg) corresponds to that seen in RPE cells of an equivalent age in situ (Fig. le) [Sandig and Kalnins, 1988, 19901. Furthermore, the general morphology of the cultured RPE cells (Fig. lf) is similar to that of RPE cells in situ (Fig. ld) [Sandig and Kalnins, 1988, 19901. In cultures the RPE cells were often more columnar than those in situ, and the basement membrane under the RPE cells had frequently folded (Fig. l f ) , whereas in situ it appears fairly flat and straight (Fig. ld). These changes are probably due to elastic recoil of the extracellular matrix during the preparation of the tissue. Although these differences in general morphology were observed, the culture conditions used did not alter the integrity or
the ultrastructural appearance of the ZA-junctions between RPE cells. Treatment With Trypsin at a Normal Ca2+ concentration
To determine whether a Ca’ -dependent adhesion system, suggested by the presence of A-CAM [Volk and Geiger, 19841, is present in the ZA-junctions of RPE cells, we incubated RPE and MDCK cells, where a similar Ca2+-dependentadhesion system is found [Imhof et al., 1983; Behrens et al., 1985; Volberg et al., 19861, for up to 2 h in a medium containing 0.05% trypsin and normal Ca’ levels. This treatment should leave Ca’ dependent adhesion molecules intact and disrupt the Ca’+-independent ones. The F-actin in both the RPE cells (Fig. 2a), and the MDCK cells (Fig. 2b) is distributed in a pattern identical to that seen in control cultures (Fig. lb) and in en face preparations (Fig. la). No dissociation of cells was observed. Moreover, electron micrographs show no change in the morphology of ZAjunctions (Fig. 2c,d), and distinct IDS of ZACs are observed in the ZA-junctions of RPE cells treated with trypsin at normal Ca2+ levels (Fig. 2d). In addition, adjacent RPE cells have maintained contact all along their lateral borders (Fig. 2c). This suggests that in both RPE cells and MDCK cells Ca’+ -dependent adhesion molecules are present in the region of the ZA-junctions, as well as along other parts of their lateral membranes, and that the ZACs in the ZA-junctions of RPE cells are components of this adhesion system. +
+
+
Treatment With Low Ca2+
In order to test 1) whether a Ca’+-independent adhesion system exists in addition to the Ca2+-dependent one, 2) whether the ZACs are indeed components of the Ca’+ -dependent adhesion system, and 3) whether the presence of ZACs in the ZA-junctions between RPE cells influences their response to low Ca’+, we incubated sheets of RPE and MDCK cells in medium containing 4 mM and 10 mM EGTA. This treatment should disrupt the Ca2+-dependent adhesion system, but leave the Ca’+ -independent one intact. Whereas in MDCK cells the distribution of F-actin showed that cells had dissociated after 5 min of treatment with low Ca’+ (Fig. 3b), in RPE cells maintained for up to 2 h in low Ca2+ (Fig. 3a) the distribution of F-actin has a hexagonal pattern similar to that seen in control cultures (Fig. lb) and in en face preparations (Fig. la). Electron micrographs of RPE cells, maintained for 1 h in low Ca’+, show that although the cells have dissociated along most of their lateral borders, their ZAjunctions stay intact (Fig. 3c,d). The IDS of ZACs, however, are no longer visible in the ZA-junctions of treated RPE cells, but are replaced by granular electron-dense
Fig. I . Comparison of RPE in situ (a,d,e) and after 7 h of culture (b,f,g). F-actin labelling of RPE cells in situ (a), in culture (b), and of MDCK cells (c) shows regular hexagonal CMBs in RPE cells (a,b) and more irregular ones in MDCK cells (c). Electron micrographs of RPE cells in situ (d,e) and in culture (f,g) show the localization of ZA-junctions ( m o w s in d,f) and the IDS (arrows in e,g) of ZACs forming these junctions. NR = neural retina. Bar in a and b = 10 pm, in c = 20 km, in d and f = 4 p m , in e and g = 0.1 pm.
50
Sandig et a].
Fig. 2. F-actin distribution in RPE (a) and in MDCK cells (b) after I h of treatment with trypsin at normal Ca2+ levels shows no dissociation in the junctional region. Electron micrographs of RPE cells (c,d) show intact apical ZA-junctions (arrows in c) with IDS (arrows in d) of ZACs forming these junctions. Bar in a = 10 pm, in b = 20 pm, in c = I pm, in d = 0.1 pm.
material (Fig. 3d). Similar results were obtained with RPE sheets cultured for 30 min and for 2 h in low Ca2+ (data not shown). After treatment with low Ca2+, some of the RPE cultures were returned to normal culture medium for 1 h, 3 h, and 5 h to determine whether the ZACs would reform. Even after 5 h of culturing in normal medium, distinct ZACs are not observed in the ZA-junctions of RPE cells (Fig. 3e,f), regardless of the length of our previous treatment with low Ca*+. Instead, dense gran-
ular material (Fig. 3f), similar to that seen in the cells fixed during treatment (Fig. 3d), persists in the junctional space between RPE cells. Treatment With Trypsin at a Low ca2+ concentration
In order to study the splitting and the reformation of ZA-junctions, sheets of RPE cells were incubated in medium containing 0.05% trypsin and 0.53 mM EDTA. This treatment is normally used to dissociate cells (in-
Splitting of ZA-Junctions in RPE Cells
Fig. 3 . F-actin distribution in RPE cells after 1 h (a) and in MDCK cells after 5 min (b) of treatment with low Caz+ shows that only the MDCK cells have dissociated. Electron micrographs of RPE cells after I h of treatment (c,d) and after 1 h of treatment followed by 5 h of incubation in normal medium (e,9 show intact apical ZA-junctions (arrows in c,e), which lack IDS of ZACs (d,f). Note that the RPE cells have separated in regions basal to the ZA-junctions. Bar in a = 10 pm, in b = 20 pm, in c and e = 2 pm, in d and f = 0.1 bm.
51
52
Sandig et a].
cluding MDCK cells) for subculturing [Freshney, 19861, and is known to disrupt both Ca’+-dependent and Ca2+independent adhesion systems. As expected, treatment with trypsin and low Ca’+ results in the splitting of ZA-junctions in RPE cells. This was suggested by results obtained by fluorescence microscopy and confirmed by electron microscopy. The distribution of F-actin after a 30 min treatment suggests that in some areas of the epithelial sheet junctional splitting has just started, since the CMBs have separated from one another and become circular in shape (Fig. 4a), instead of being polygonal as in the intact epithelial sheet (Fig. 4f). In other areas of the RPE sheet, the separation of CMBs has progressed farther, and they have contracted into smaller rings. The CMBs in these regions (Fig. 4b) have separated from those of adjacent cells to varying degrees, resulting in a pattern that consists of rows and islands of rings separated from one another by larger gaps. This pattern, and the presence of fainter lines of F-actin staining along lateral cell membranes, suggests that the cells have dissociated apically, but still remain in contact basal to the former ZA-junctions. After staining of similar cultures with antibodies to vinculin, a pattern identical to that seen after F-actin labelling was obtained (data not shown). This observation indicates that the vinculin-containing cytoplasmic plaques in the ZA-junctions, which anchor the CMBs to the junctional membranes, remain closely associated with the CMBs during their separation after treatment with trypsin at a low Ca2+ concentration. The splitting of the ZA-junctions by this treatment was confirmed at the EM level (Fig. 5a). Higher magnifications show that the CMBs are still attached to the separating membranes in the junctional regions, as the CMBs together with this membrane material gradually contract and translocate away from one another (Fig. 5b). The IDS of ZACs are visible in those parts of the ZA-junctions, which are still intact (Fig. 5b). Even after further separation (Fig. 5c,d), the CMBs remain in contact with membranous material. Some cells can share a ZA-junction with one of the neighbouring cells after having completely separated in the junctional region from others (Fig. 5c), indicating that the splitting of the ZAjunctions is both asynchronous and asymmetricai. Reformation of ZA-Junctions
To determine whether new ZA-junctions, containing ZACs, can form after junctional splitting, sheets of RPE cells were returned to normal medium after treatment with trypsin at a low Ca’+ concentration, During the first hour in normal medium, the preexisting CMBs further contract into small rings from which fine MF bundles appear to radiate, together forming aster-like structures (Fig. 4c).
Electron micrographs show that the RPE cells at this stage (Fig. 6a) are in lateral contact with their neighbours, but have not formed new ZA-junctions (Fig. 6a). In the apical cytoplasm electron-dense areas can be seen (Fig. 6a,b). These correspond to sections through the preexisting contracted CMBs, and are seen in Figure 4c and in the inset of Figure 6a in an en face view at the light microscope level. The electron-dense material is closely associated with membranous vesicles and invaginations of the plasma membrane (Fig. 6b,c). Apical cytoplasmic processes, containing MF bundles, fan out from the apical regions constricted by the CMBs (Fig. 6). These MF-containing apical processes might be the rays that appear to radiate from the contracted rings seen after F-actin labelling (Fig. 4c). Stress fibers were not detected in these cells. By 3 h in normal medium, following 30 min treatment with trypsin at a low Ca2+ concentration, F-actin labelling shows that almost all of the preexisting contracted CMBs have disappeared. Instead, straight fluorescent lines are seen (Fig. 4d), suggesting the formation of new ZA-junctions between many of the adjacent cells. After 5 h in normal medium7 similar F-actin staining is seen around all of the cells in the culture, giving a polygonal pattern (Fig. 4e). This pattern, however, differs from that seen in control cultures (Fig. 4f) in that the cells are irregularly (Fig. 4e) rather than hexagonally (Fig. 4f) shaped. Electron micrographs confirm that in RPE cells after 5 h in normal medium, following 30 rnin and 1 h treatments with trypsin at a low Ca’+ concentration, apical ZA-junctions had indeed reformed (Fig. 7a). Distinct IDS of Z A G , however, are seen only in the newly formed ZA-junctions of cells, which had been previously treated for 30 rnin (Fig. 7b), but not in ZA-junctions of RPE cells, which had been treated for 1 h (Fig. 7c). DISCUSSION
In summary, we have shown that, whereas the ZAjunctions of MDCK cells split upon treatment both with low Ca2+ and with a mixture of trypsin and low Ca2+, the ZA-junctions of RPE cells only split when treated with a mixture of trypsin and low Ca2+. Epithelial ZAjunctions can, therefore, be split and reformed in cells maintained on their native basement membrane, i.e. in cells which, unlike MDCK cells, have not been previously dissociated and plated on glass. Treatment with trypsin alone or with low Ca2+ alone is insufficient to split the ZA-junctions in RPE cells, although the IDS of ZACs disappear upon treatment with low Ca2+. We have further shown that after junctional splitting in RPE cells, the CMBs contract asymmetrically into small rings, which stay associated with membranous material
Fig. 4. RPE treated with trypsin at low Caz+ levels showing the distribution of F-actin after a 30 rnin treatment (a). RPE cells start to dissociate apically, and their CMBs retract (arrows). In other areas (b) the dissociation has progressed asymmetrically and the CMBs have contracted into smaller rings (arrows). After a 30 rnin treatment followed by 1 h in normal medium (c) the contraction of the CMBs(arrows) continues. After a 30 rnin treatment followed by 3 h in nor-
ma1 medium (d) new ZA-junctions (arrows) have started to form. Note the remnant of the preexisting CMB in one of the RPE cells (open arrow). After a 30 rnin treatment followed by 5 h in notmal medium (e) the new CMBs can be seen around all of the RPE cells forming an irregular pattern compared to the hexagonal one in the untreated controls (0.Bar = 10 +m.
54
Sandig et al.
Fig. 5 . Electron micrographs of RPE cells and their junctional regions after treatment with trypsin at low Ca*+ levels at early (a,b) and later (c-e) stages of junctional splitting. Note that IDS of ZACs are still visible in those regions of the junction, which have not separated (arrowhead in b). The CMBs, associated with membranous material, progressively retract from one another, and are seen as electron-dense
areas (arrows in bd).Similar material is seen in cells that have lost some of their neighbours (arrow in e). The insets in d and e show the corresponding distribution of F-actin in an en face view at the light microscope level. Bar in a and e = 3 pm, in b = 0.2 pm, in c = 0.4 pm, in d = 0.5 pm.
Splitting of ZA-Junctions in RPE Cells
Fig. 6. Electron micrographs of RPE cells after 1 h incubation in normal medium following 30 min treatment with trypsin at low Ca2+ levels. Electron-dense areas in the apical cytoplasm of RPE cells correspond to grazing (a,b) and cross (c) sections through the contracted CMBs (inset in a). Many of these electron-dense areas (arrows) are
55
closely associated with membranous vesicles and invaginations from the cell surface. Numerous apical processes radiate from a cytoplasmic protrusion above the contracted CMB. Note that ZA-junctions have not reformed at this stage. The area indicated in a is shown at a higher magnification in b. Bar in a = 2 pm, in b and c = 0.5 pm.
56
Sandig et al.
Fig. 7 . Electron micrographs of RPE cells after 3 h (a,b) and 5 h (c) of incubation in normal medium, following 30 min (a,b) and I h (c) treatment with trypsin at low Caz+ levels. In the newly formed apical ZA-junctions (arrows in a, and panels b and c) the IDS (arrows in b)
of ZACs are present after treatment for 30 min (b), but not after 1 h (c) of treatment. The inset in a shows the corresponding distribution of F-actin in an en face view at the light microscope level. 3ar in a = 2 p n , in b and c = 0.1 pm.
as they are internalized. After junctional splitting, new ZA-junctions, composed of Z A G , form upon return to normal medium. These results can be explained by hypothesizing that the Ca’+-dependent adhesion molecules in RPE cells are less sensitive to low Ca2+ than those in MDCK cells. It is more likely, however, that two adhesion systems, a Ca’+ -dependent and a Ca2+-independent one, are differentially expressed along the lateral cell membranes of RPE cells. Such Ca*+-dependent and Caz+independent adhesion systems have been identified in other cell types [Takeichi, 1977; Grunwald et al., 1980; Obnnk, 1986; Takeichi, 19881. The Ca2+-dependent adhesion system can be disrupted by treating cells with trypsin and Ca2+-chelators, or with Ca2+-chelators alone. The proteolytic degradation of the adhesion molecules in this system by trypsin, however, can be prevented by addition of Ca2+. The Ca’+-independent system, in contrast, can be disrupted by incubating cells with trypsin alone [Takeichi, 1977, 19881. We have shown that upon treatment with trypsin at normal Ca2+ levels, RPE cells and MDCK cells do not
dissociate, and that the ZA-junctions in RPE cells remain intact and are morphologically identical to those in situ. These results indicate that Ca’ -dependent adhesion molecules are present along the lateral cell membranes and within the ZA-junction, since in this case proteolytic degradation is prevented by the presence of exogenous Ca2+ [Takeichi, 1977, 19881. Since the IDS of ZACs remain in ZA-junctions of RPE cells treated with trypsin, the ZACs may be components of this Ca2+-dependent adhesion system. It has been reported previously that A-CAM, a Caz+-dependent adhesion molecule [Volk and Geiger, 1986a,b], is indeed present in embryonic chick RPE cells [Volk and Geiger, 19841. We have confirmed these findings in the RPE cells of newly hatched chicks (data not shown). Furthermore, in MDCK cells a different Ca’+-dependent adhesion molecule, Arc- 1, which seems to be identica1 to L-CAM and E-cadherin [Takeichi, 19881, has been localized [Imhof et al., 1983; Behrens et al., 19851. We have also shown that in contrast to ZA-junction in MDCK cells, the ZA-junctions in RPE cells do not split when maintained for up to 2 h in low Ca2+. The IDS of ZACs, however, are no longer +
Splitting of ZA-Junctions in RPE Cells
visible and, instead, irregular electron-dense material is present in the junctional space. In addition, upon exposure to low Ca2+ cells dissociate laterally basal to the ZA-junctions, so that the ZA-junctions appear to be the only remaining contact areas. These findings can be explained as follows. The exposure to a low Ca2+ concentration disrupts the binding by the Ca2+-dependent adhesion molecules in both RPE and MDCK cells. As a result RPE cells dissociate along their lateral borders and lose the IDS (components of the Ca2+-dependent adhesion molecules in the ZA-junctions) of ZACs in their ZA-junctions. However, in addition to the Ca2+-dependent adhesion system (the IDS of ZACs) Ca2+-independent adhesion molecules are present in the ZA-junctions of RPE cells, which keep the ZA-junctions intact during the low Ca2+ treatment. This suggests that in RPE cells the Ca2 -independent adhesion molecules are present only in the ZA-junctions, whereas the Ca2+-dependent ones are distributed along the entire length of the lateral membranes. In contrast MDCK cells appear to lack the Ca2+-independent adhesion molecules, since these cells dissociate completely on exposure to low Ca2+. In order to split the ZA-junctions of RPE cells, treatment with both trypsin and low C a 2 + , a mixture normally used to dissociate cells for subculturing [Freshney, 19861, is necessary. This observation further suggests that in RPE cells both the Ca2+-dependent and the Ca’ -independent adhesion systems, disrupted by this treatment, are present. These findings are consistent with the results obtained in a study in which E7 chick RPE was subjected to various treatments to establish optimal conditions for cell dissociation [Vielkind and Crawford, 19881. This study showed that treatments with low Ca2+ alone, or trypsin alone, do not result in a high yield of dissociated cells for plating. Although the pattern of junctional splitting in RPE cells was generally similar to that seen in MDBK [Kartenback et al., 1982; Volberg et al., 19861 and MDCK cells, the time and the type of treatment required were quite different. The ZA-junctions in MDCK and MDBK cells split very rapidly upon treatment with low Ca2+ (after 3-5 min), whereas RPE cells required trypsin in addition to low Ca2+ and had to be treated between 30 min and I h for junctions to split. In MDBK cells, the CMBs together with the junctional plaques completely detach from the cell membrane, contract, and translocate toward the cell interior [Volberg et al., 19861. The CMBs of RPE cells also contract after junctional splitting induced by treatment with trypsin and low Ca2+. However, since they are still attached at least partially to the cell membrane, this contraction leads to the constriction of the apical portions of the cells, causing the apical processes to fan out. The contracted CMBs of RPE cells were associated with irregularly shaped ves+
+
57
ides, some of which may be cross sections of invaginations of the cell membrane. Similar vesicles associated with electron-dense junctional plaque material have been observed in MDBK cells near the internalized CMBs [Kartenbeck et al., 19821. In MDBK cells many of these, however, appeared to be associated with desmosomal plaque proteins and originated from the internalization of desmosomes [Kartenbeck et a]., 19821, absent in chick RPE cells [Docherty et al., 19841. In both RPE cells and in MDBK cells, the contraction of the CMBs continues upon return to normal medium after treatment with trypsin and low C a 2 + , and with low Ca2+, respectively. The CMBs together with the vinculin-containing plaque material finally disappear after approximately 3 h in normal medium [Volberg et a]., 19861. In both RPE cells and MDBK cells new ZA-junctions form after recovery from treatment. Junction formation in MDBK cells occurred within 30-45 min after return to normal medium and was essentially completed after 2 h [Volberg et a]., 19861. In RPE cells, the formation of new junctions did not start before 1 h, and was not completed until 4-5 h, after return to normal medium. These differences in the time course of the formation of new ZA-junctions might be due to a higher speed of spreading of MDBK cells on glass compared to the rate of spreading of RPE cells on basement membrane, or to differences in the rate of synthesis and/or assembly of new junctional proteins between the two cell types. It is interesting that after recovery, IDS of ZACs are present in the newly formed ZA-junctions (after previous treatment with trypsin in low Ca2+),whereas they do not reform in the intact ZA-junctions after loss during exposure to low Ca2+ alone. These observations suggest that junctional splitting might be a signal for activating the synthesis and assembly of the molecular components of Z A G . Alternatively, the turnover of molecules in the intact junction may be slower than the time required to assemble new ZA-junctions, so that we were unable to observe the eventual appearance of ZACs in the ZAjunctions of cells, which had been treated with low Ca2+, in the time frame examined. In conclusion, we have shown that the ZA-junctions in chick RPE cells are resistant to treatment with low Ca2+. In order to split these junctions, treatment with both low Ca2+ and trypsin is required. These observations suggest that in chick RPE cells the ZA-junctions, composed of Z A G , might mediate stronger adhesion compared to the ZA-junctions in other epithelial cells. Stronger adhesion between RPE cells might be necessary since they lack desmosomes [Docherty et al., 19841, which provide the structural basis for strong adhesion in other epithelia.
58
Sandig et al.
ACKNOWLEDGMENTS M.S. is an Ontario Graduate Scholar, and G.J.H. is a recipient of a Dental Fellowship from the Medical Research Council of Canada. This work was supported by a grant from the RP Eye Research Foundation to V.I.K. REFERENCES Behrens, J., Birchmeier, W., Goodman, S.L., and Imhof, B.A. (198s): Dissociation of Madin-Darby canine kidney epithelial cells by the monoclonal antibody anti-arc-I: Mechanistic aspects and identification of the antigen as a component related to uvomorulin. J. Cell Biol. lOl:I307-13lS. Blanchard, A., Ohmian, V., and Critchley, D. (1989): The structure and function of a-actinin. J. Muscle Res. Cell Motil. 10:280289. Cowin, P., Franke, W.W., Kapprell, H.P., and Kartenbeck, J. (1983: The desmosome intermediate filament complex. In Edelman, G.M., and Thiery, J.P. (eds.): “The Cell in Contact.” New York: Neuroscienc Inst. Publ., J . Wiley & Sons, pp. 427-460. Craig, S.W., and Pardo, J.V. (1980): Alpha-actinin localization in the junctional complex of intestinal epithelial cells. J. Cell Biol. 80:203 -2 10. Docherty, R.J., Edwards, J.G., Garrod, D.R., and Mattey, D.L. (1984): Chick embryonic pigmented retina is one of the group of epithelioid tissues that lack cytokeratins and desmosomes and have intermediate filaments composed of vimentin. J. Cell Sci. 71:61-74. Edelman, G.M. (198s): Specific cell adhesion in histogenesis and morphogenesis. In Edelman, G.M., and Thiery, J.P. (eds.): “The Cell in Contact.” New York: Neurosciences Publ. J. Wiley & Sons Inc., pp. 139-168. Edelman, G.M., and Thiery, J.P. (1985): “The Cell in Contact.” New York: Neurosci. Inst. Publ., J. Wiley & Sons. Fxquhar, M.G., and Palade, G.E. (1963): Junctional complexes in various epithelia. J. Cell Biol. l7:375-409. Freshney, R.I. (1986): “Animal Cell Culture.” Oxford: IRL Press. Fristrom, D. (1988): The cellular basis of epithelial morphogenesis, A review. Tissue Cell 20:645-690. Geiger, B, (1983): Membrane-cytoskeleton interaction. Biochim. Biophys. Acta 737:305-341. Geiger, B., Tokuyasu, K.T., Dutton, A.E., and Singer, S.J. (1980): Vinculin, an intercellular protein localized at specialized sites where microfilament bundles terminate at cell membranes. F’roc. Natl. Acad. Sci. U.S.A. 77:4127-4131. Geiger, B., Dutton, A.H., Tokuyasu, K.T., and Singer, S.J. (1981): Immunoelectron microscope studies of membrane-microfilament interactions: Distribution of alpha-actinin, tropomyosin, and vinculin in intestinal epithelial brush border and chicken gizzard smooth muscle cells. .I. Cell Biol. 91:614-628. Geiger, B., Avnur, Z., Volberg, T., and Volk, T. (1985): Molecular domains of adherens junctions. ln Edelman, G.M., and Thiery, J.P. (eds.): “The Cell in Contact.” New York: Neuroscience lnst. Publ. J. Wiley & Sons, pp. 461-490. Green, K .J, , Geiger, B ., Jones, .l.CR., TaIian, J.C., and Goldman, R.D. (1987): The relationship between intermediate filaments and microfilaments before and during the formation of desmosomes and adherens-type junctions in mouse epidermal keratinocytes. J. Cell Biol. 104:1389-1402. Grunwald, G.B., Geller, R.L., and Lilien, J. (1980): Enzymatic dissection of embryonic cell adhesive mechanisms. J. Cell Biol. 8S1766-776. Hergott, G.J., Sandig, M., and Kalnins, V.1. (1989): Cytoskeletal organization of migrating retinal pigment epithelial cells during wound healing in organ culture. Cell Motil. Cytoskeleton 13: 83-93.
Imhof, B . A . , Vollmers, H.P., Goodman, S.L., and Birchmeier, W. 91983): Cell-cell interaction and polarity of epithelial cells: Specific perturbation using a monoclonal antibody. Cell 35: 667-675. Kartenbeck, J., Schmid, E., Franke, W.W., and Geiger, B. (1982): Different modes of internalization of proteins associated with adherens junctions and desmosomes: Experimental separation of lateral contacts induces endocytosis of desmosomal plaque material. EMBO J. 6:725-732. Kuwabara, T. (1979): Species differences in the retinal pigment epithelium. In Zinn, K.M., and Marmor, M.F. (eds.): “The Retinal Pigment Epithelium.” Cambridge: Harvard University Press, pp. 58-82. Nguyen-Legros, J. (1978): Fine structure of the pigment epithelium in the vertebrate retina. lnt. Rev. Cytol. 7:287-328. Obrink, B. (1986): Epithelial cell adhesion molecules. Exp. Cell Res. I 63: 1-2 I . O’Keefe, E., Briggaman, R.A., and Herman, B. (1987): Calciuminduced asembly of adherens junctions in keratinocytes. J. Cell Biol. lOS:807-817. Opas, M., and Kalnins, V.I. (198s): Spatial distribution of cortical proteins in cells of epithelial sheets. Cell Tissue Res. 239: 4SI-4S4. Owaribe, K . , and Masuda, H. (1982): Isolation and characterization of circumferential microfilament bundles from retinal pigment epithelial cells. J. Cell Biol. 9S:310-315. Owaribe, K., and Masuda, H. (1986): Organization of microfilaments and intermediate filaments in retinal pigmented epithelial cells. In Ishikawa, H., Hatano, S . , and Sato, H. (eds.): “Cell Motility: Mechanism and Regulation.” New York: Alan R. Liss, Inc., pp. S077.514. Sandig, M., and Kalnins, V.1. (1988): Subunits in zonulae adhaerentes and striations in the associated circumferential microfilament bundles in chicken retinal pigment epithelial cells in situ. Exp. Cell Res. 17S:1-14. Sandig, M., and Kalnins, V.1. (1990): Morphological changes in the zonula adhaerens during embryonic development of chick retinal pigment epithelial cells. Cell Tissue Res. 2S9:45S-461. Staehelin, L.A. (1974): Structure and function of intercellular junctions. Int. Rev. Cytol. 39:191-283. Takeichi, M. (1977): Functional correlation between cell adhesive properties and some cell surface proteins. J. Cell Biol. 7.5: 464 -447. Takeichi, M. (1988): The cadherins: Cell-cell adhesion molecules controlling animal morphogenesis. Development 102:63945S. Turksen, K., and Kalnins, V.I. (1987): The cytoskeleton of chick retinal pigment epithelial cells in situ. Cell Tissue Res, 248: 9s-101, Vielkind, U., and Crawford, B.J. (1988): Evaluation of different procedures for the dissociation of retinal pigmented epithelium into single viable cells. Pigment Cell Res. 1:419-433. Volberg, T., Geiger, B., Kartenbeck, J., and Franke, W.W. (1986): Changes in membrane-microfilament interaction in intercellular adherens junctions upon removal of extracellular Ca ions. J . Cell Biol. l02:1832-l842. Volk, T., and Geiger, B. (1984): A 135-kD membrane protein of intercellular adherens junctions. EMBO J. 3:2249-2260. Volk, T., and Geiger, B. (1986a): A-CAM: A 135-kD receptor of intercellular adherens junctions. I. lmmunoelectron microscopic localization and biochemical studies. J. Cell Biol. 103: 1441-1450. Volk, T., and Geiger, B. (1986b): A-CAM: A 135-kD receptor of intercellular adherens junctions. 11. 4ntibody-mediated modulation of junction formation. J. Cell Biol. 103:14S1-1464. Zinn, K.M., and Marmor, M.F. (1979): ‘‘The Retinal Pigment Epithelium, ” Cambridge: Harvard University Press.
