European Journal of (eil Biology 70, 221-232 (1996, July) © Wissenschaftliche Verlagsgesellschaft . Stuttgart 221

Adhesiveness of the apical surface of uterineepithelial cells: the role of iunctional complexinteg rity

Michael Thie1)a*, Petra Fuchs"*, Stefan Butzb, Frank Sieckmannc, Heinz Haschützkyd, Ralf Kemlerd,Hans-Werner Denker"

a Institute of Anatomy, University of Essen Medical School, Essen/Germanyb Howard Hughes Medical Institute, University ofTexas, Dallas, TX!USAcInstitute of Radiation Biology, University of Essen Medical School, Essen/Germanyd Max Planck Institute for Immunobiology, Freiburg i.Br.lGermany

Received November 16, 1995

Accepted January 29, 1996

Uterine epithelium - epithelial polarization - junctionalcomplex - adhesiveness

Embryo implantation necessitates that the apical plasma membrane ofuterine epithelial cells acquires adhesiveness. Recent studies have indiocated that modulation of a major element of the epithelial phenotype,i.e. apical-basal cell polarity, might be critical in this respect. Here, weanalyze polar characteristics of nonadhesive vs. adhesive uterine epi­thelial cell Iines focusing on cytoskeletal-junctional interactions thatmay playa role in regulating adhesiveness of the apical plasma mem­brane.

HEC-l-A is a human uterine epithelial cellline exhibiting nonadhe­sive properties of its apical surface for trophoblast, whereas RL95-2represent another such cell line exhibiting adhesive properties en­abling trophoblast attachment. Homotypic intercellular contacts andfunctionally related proteins, i.e. ZO-l, E-cadherin, a-catenin, ß­catenin, plakoglobin, and desmoplakin I, were examined by trans­mission electron microscopy, immunocytochemistry, confocal laserscanning microscopy, and immunoprecipitation techniques. In addi­tion, details of actin filament architecture were studied after phalloidinlabeling. While nonadhesive HEC-l-A exhibited the well-knownpattern of cell-to-cell contacts of polarized epithelial cells, adhesiveRL95-2 showed a lack of ZO-l expression, tracer leakiness of the para­cellular pathway, and atypical features in adherens junctions: E­cadherin, a-catenin and plakoglobin were colocalized in all plasmamembrane domains and ß-catenin was localized in lateral membranedomains. Immunoprecipitations showed in both celllines the presenceof two different E-cadherin-catenin complexes, one composed of E­cadherin, a-catenin and ß-catenin, and the other of E-cadherin, a­catenin and plakoglobin. Concerning RL95-2 these data indicate thatE-cadherin/plakoglobin complexes are randomly distributed, whereasE-cadherin/ß-catenin complexes are laterally localized in these cells.Additionally, the actin-based cytoskeleton of RL95-2 lacked apolar

1) Dr. Michael Thie, Institut für Anatomie, Universitätsklinikum,

Hufelandstr. 55, D-45122 Essen/Germany." M. Thie and P. Fuchs contributed equally to this study.

organization. With respect to the intermediate filament-desmosomesystem, both cell types expressed desmoplakin I, but the vast majorityof RL95-2 lacked well-formed desmosomes as demonstrated by elec­tron microscopy.

It is concluded that modulation oftightjunctions and/or remodellingof adherens junctions, e.g. differential distribution of E-cadherin/pla­koglobin complexes and E-cadherin/ß-catenin complexes, are corre­lated with the development of apical adhesiveness of human uterineepithelial cells. This model system should allow to test experimentallywhether this correlation is due to any causal function in the develop­ment of epithelial cell polarity.

Introduction

Embryo implantation in mammals and in the human requiresadhesiveness of the apical plasma membrane of uterine epi­thelial cells thus enabling the initial contact between theembryo and maternal tissues. Human uterine epithelial cellsshow a polarized phenotype as typical für all simple epitheliaincluding ürganization of the surface membrane into biochem­ically distinct apical and basolateral plasma membranedomains. The basolateral membrane domain is studded withvarious adhesion molecules, including E-cadherin and inte­grins, mediating stable adhesion to neighboring cells or to theextracellular matrix. The apical domain is largely free of thesemolecules and nonadhesive for opposing uterine epithelialcells or embryonic cells such as trophoblast. Interestingly, theapical domain can be functionally reprogrammed towardsadhesiveness für trophoblast when cells are exposed to anappropriate hormonal milieu (for reviews, see [10, 16,38]).

The molecular details of how the surface of human uterineepithelial cells contacts trophoblast are still unclear. There isevidence that alterations in membrane compositions of uterinecells occur in a concerted way in both apical and basolateraldomains and are coupled somehow with a partialloss of epi­thelial characteristics thus leading to acquisition of apical

I

222 M. Thie, P. Fuchs, S. Butz, F. Sieckmonn et 01.

adhesiveness of epithelial cells [10, 11]. Für example, estab­lishment of apical adhesiveness of human uterine cells in vivois correlated with redistribution of a6-integrin subunits fromthe basal to the lateral subdomain and a significant increase ofCD44 expression in the lateral domain [1]. A reduction in thethickness of the glycocalyx and in the cell surface charge of theapical membrane domain seems to be related to these altera­tions [26]. Cell culture studies have shown that apical adhe­siveness for trophoblast of a human uterine-derived cellline,i.e. RL95-2 cells, correlates with the occurrence of a6-, ßl­and ß4-integrin subunits at the entire surface membrane,whereas polarized distribution of integrins correlates withnonadhesiveness of the apical membrane in HEC-I-A cells[51]. The common denominator of these changes might be a sofar non-identified mechanism modulatillg the apical-basal cellpolarity ultimately leadillg to adhesiveness of the apicalplasma membrane of uterine epithelial cells for trophoblast[10-12]. The present study was designed in order to get insightinto features involved in modulation of cell polarity in uterinecells.

Materials and methods

MaterialsAll chemicals were of analytical grade and were obtained from Merck,Darmstadt/Germany or Sigma-Aldrich, Deisenhofen/Germany.

Cell cultureHuman endometrial carcinoma cells were purchased from the Amer­ican Type Culture Collection (ATCC; Rockville, MD/USA), i.e. HEC­1-A cells (HTB 112) and RL95-2 cells (CRL 1671). Cell lines weregrown in plastic flasks in 5 % C02'95 % air at 37 oe. HEC-1- A cellswere seeded out in McCoy's 5A medium (Biochrom Seromed, BerlinlGermany) supplemented with 10 % fetal calf serum (Gibco-Life Tech­nologies, EggensteinlGermany), and RL95-2 cells in a 1+1 mixture ofDulbecco's modified Eagle's medium (Gib co) and Ham's F12 (Bio­chrom) supplemented with 10% fetal calf serum. All media were ad­ditionally supplemented with penicillin (100 lU/mi; Boehringer,Mannheim/Germany) and streptomycin (100 !-tg/ml; Boehringer,Mannheim/Germany). The growth medium was changed every 2 to 4days, and cells were subcultured by trypsinization (trypsin-EDTA solu­tion; Gibco) when they became confluent.

For metabolic labeling experiments, monolayers (2 x 10' cells)were grown in methionine-free medium (Gibco), supplemented with10 % dialyzed fetal calf serum, for 2 h prior to the addition of 50 !-tCi/ml [35S] methionine (Amersham, Braunschweig/Germany) for over­night incubation.

Adhesiveness of confluent monolayers for human JAR choriocarci­noma cell spheroids (ATCC: HTB 144) was routinely measured usinga centrifugal force-based adhesion assay [27]. Confirming recentlypublished data [51], RL95-2 monolayers allowed JAR cells to attachwhile HEC-1-A monolayers did not (not shown).

AntibodiesJunetional eomplexes. Rabbit polyclonal antibodies to human ZO-l(61-7300) were obtained from WAK Chemie, Bad Homburg/Germany.Mouse monoclonal antibody to human E-cadherin (6F9; [14]) was do­nated by Dr. J. Behrens (Max-Delbrück-Centrum, Berlin/Germany).Preparation of rabbit polyclonal antibodies to uvomorulin (E­cadherin; [41]), to peptides of a-catenin (aM12K; [21]), ß-catenin(ßP14L; [5]), and plakoglobin (D15A; [4]) has been described previ­ously. Mouse monoclonal antibody directed against desmoplakin (des­moplakin I; [7]) was a gift of Dr. J. Kartenbeck (DKFZ, Heidelberg/Germany).

Cytoskeleton. Mouse monoclonal antibodies to cytokeratin 7 (Ks7.18), to cytokeratin 8 (Ks 8.7), to cytokeratin 18 (Ks 18.04), and tocytokeratin 19 (Ks 19.1) were obtained from Progen, Heidelberg/Ger­many, and a mouse monoclonal antibody to vimentin (V9) fromSigma-Aldrich, Deisenhofen/Germany.

Preparation of cytoskeletal material andimmunoblottingIntermediate filament-enriched fractions were obtained from conflu­ent monolayers according to Eckert and Kartenbeck [13]. Monolayers(2 x 10' cclls) were washed three times with phosphate buifered saline(PBS) at room temperature and subsequently immersed in 2 ml highsalt buifer (10 mMTris-HCI, pH 7.5, 1.5 M KCI, 140 mM NaCI, 5 mMEDTA, 1 % Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 1 mMdithioerythritol, 1.5 !-tMpepstatin A). Cell extracts were prepared bysonication of cells and stirring of suspension for 20 min at 4°C fol­lowed by centrifugation at 5000g for 20 min at 4°C. The pelleted cyto­skeletal material was washed twice with PBS and used for sodiumdodecyl sulfate (SDS) polyacrylamide gel electrophoresis (PAGE)under reducing conditions [28].

For immunodetection, proteins separated by SDS-PAGE weretransferred electrophoretically to Immobilon membrane (Millipore,Bedford, MA/USA). Filters were immersed in 25 mM Tris-HCI,pH 7.4, 150 mM NaCI, 30 mM KCI, 0.1 % Tween 20, at 37°C for 1 h,incubated with 2 to 5 !-tg/mlprimary antibody and then with alkalinephosphatase-conjugated secondary antibody (Dianova, Hamburg/Ger­many). Bound antibodies were detected using Nitro blue tetrazoliumand 5-bromo-4-chloro-3-indolyl phosphate as substrates (Sigma­Aldrich).

Preparation of ceillysates andimmunoprecipitationCell Iysates were obtained from confluent monolayers according toButz and Kemler [3]. Monolayers (2 x 10' cells) were washed threetimes with PBS at room temperature and subsequently immersed in800 !-tlcold lysis buifer (20 mM imidazole, pH 6.8; 100 mM KCI, 2 mMMgCb, 10 mM ethylene gIycol-bis(ß-aminoethyl ether)N,N,N' ,N'­tetraacetic acid, 300 mM sucrose, 1 mM Na-vanadate, 1 mM Na­molybdate, 1 mM Na-fluoride, 0.2 % Triton X-100, 1 mM phenylmeth­yisulfonyl fluoride, 1 mM N-ethylmaleimide, and 1.5 !-tMpepstatin A)for 10 min at 4 oe. Celllysates were centrifuged at 16OOOgfor 10 minand supernatant was precleared by incubation for 1 h with 250 !-tI/misupernatant of 10 % (w/v) protein A-Sepharose beads (Pharmacia,Freiburg/Germany) preabsorbed with lysis buifer containing ovalbu­min (1 mg/mI). Finallysates were centrifuged at 1000g for 5 min fol­lowed by centrifugation at 16000g for 5 min.

For immunoprecipitation, precleared supernatants of cell Iysatesequivalent to 4 x 106 cells were incubated with the appropriate anti­body (5 !-tganti-uvomorulin, 10 !-tganti-a-catenin, 30 !-tganti-ß-cateninand 20 !-tganti-plakoglobin) and 50 !-tlprotein A-Sepharose for 1 h.

Fig.1. Ultrastructure of HEC-1-A and RL95-2 monolayers culti­vated on POlY-D-lysine-coatedcoverslips. HEC-I-A cells (a) form closelateral membrane contacts and are highly polarized, e.g. the nucleus islocated predominantly basally, organelles in the supranuclear region.Note numerous microvilli at the apical cell pole (arrows). RL95-2 cells(b) grow as monolayers of irregular shape and show little signs of polarorganization, e.g. the nucleus is located in the center, organelles tendto pile up perinuclearly. Mierovilli are largely laeking. - e, d. Photo­micrographs of ultrathin seetions of lateral plasma membran es ofHEC-I-A cells (e) and RL95-2 cells (d). Note that HEC-I-A eells butnot RL95-2 cells form morphologieally distinet tight junetions, inter­mediate junetions and desmosomes. - 00 Coverslip. - Stars: Growthmedium. - N Nucleus. - tj Tight junction. - ij Intermediate junetion. ­d Desmosome. - HI, H2 HEC-I-A cells. - RI, R2 RL95-2 cells. - Bars5 [tm (a, b), 0.25 [tm (e, d).

EJCB Remodelling of epithelial cell polority 223

I

224 M. Thie, P.Fuchs, S. Butz, F.Sieckmonn et 01. EJCB

Results

dehydrated in ethanol and propylene oxide, embedded in epoxy resinmixture and examined as deseribeel above omitting double-staining

with uranyl acetate and leael citrate.

Polarized vs. nonpolarized epithelial phenotypeUltrastructurc. Low magnification overviews of ultrastructureof HEC-I-A and of RL95-2 cells are shown in Figure 1. !-lECeells whieh are nonadhesive for trophoblast-type cells formedordered monolayers of cuboida1 to columnar cells (Fig. la).The apical surface was covered with numerous, relativelyshort microvilli. Nuclei of !-lEC cells were predominantly 10­

cated at the base of the eell. Mitochondria, Golgi apparatus,

and endoplasmie retieulum were mostly positioned in the su­pranuclear region of the cell. With respect to the distributionof these organelles, nonadhesive !-lEC eells showed a highlypolarized phenotype. Cells grew in close eontact to the sub­strate and to adjacent !-lEC cells.

RL cell monolayers which permit trophoblast attachmentformed irregular sheets (Fig. Ib). Single cells had a roundishshape. J n contrast to I-IEC cells, the apieal cell pole appeareddome-like and was largely free of microvilli. The cell nucleiwere predominantly located in the center of the cell, andorganelles tended to pile up perinuclearly without showingany sign of polar distribution. RL cells adhered to the sub­strate via large cytoplasmic extensions but did not form abroad eell-matrix eontact at the basal eell pole. At the lateralmembrane only primitive adherens junetions were seen; re­gions of interacting plasma membranes were alternating withregions of large intereellular spaees. Thus, RL eells laeked thestructural polarization as typieal for a simple epithelium.Analysis 01' cytoskclctaI proteins. To prove the epithelialnature 01' RL eells we performed cell typing with respect to the

expression 01' intermediatc filament proteins as shown in Fig­ure 2. SDS-PAGE 01' the cytoskeletal preparations showedfour major polypeptides of moleeular weight 54000 Dalton

8eads were washed 5 times with lysis buffer supplementeel with oval­

bumin. SDS-PAGE of the immunopreeipitates was performed underreelucing conditions [28J.

For fluorography, SOS-PAGEs were fixed in 10% acetic acid, incu­bated with 1 M Na-salicylatc and subsequently drieel. For immunode­

tection, the proteins scparated by SOS-PAGE were transferreel elec­trophoretically to Immobilon membrane. Transferred pro teins weredetecteel by appropriatc primat·y antiboclies as described above.

Immunofluorescence and phalloidin stainingMonolaycrs grown on polY-D-lysine-coated glass coverslips werc rinsedtwice in P8S, fixed and permeabilized by incubation in 96 %methanol-water for 7 min at - 20 oe. After sevcral washings with P8Sanel a final wash in PBS/O.5 % bovine serum albumin (BSA), cells werc

incubated for 40 min at room temperature with the primat·y antibody.Thereafter, cells were rinsed in P8S/0.5 % BSA (3 x 5 min) and

incubateel with the corresponding fluorescein isothiocyanate (FITC)­conjugateel seconelary antibody for 40 min at room temperature.

FITC-conjugated swine anti-rabbit secondary antibodies (1'205),FITC-conjugated rabbit anti-mouse seconelary antiboclies (F232) anelFITC-conjugated rabbit anti-rat secondary antibodies (1'234) wercobtaincel from Oako Diagnostika, I-Iamburg/Germany. After rinsingwith PBS/0.5 % ßSA, specimens were mounted with 90 % glycerol­

P8S, supplementcd with 1.0 % p-phenylenediamine as an anti­quenching agent and examineel with a Zeiss Axiophot microscope withepi-illumination (450-490 nm excitation; filterset 487909).

Far phalloidin staining, a solution of tetramethylrhoelamine isothio­cyanate (TRITC)-conjugated phalloielin (Sigma-Alelrich) at a concen­tration of 25 ,lg/ml in PBS was used. Monolayers grown on polY-D­Iysine-coatcel glass coverslips were rinsed in PBS, fixed with 3 % para­formalelehyde for 15 min at room temperature, permeabilized byineubation for 2 min in a solution of 0.05 % Triton X-IOO, and then

incubated for 15 min in a solution ofTRITC-phalloidin. After rinsing,the staineel cells were examined using a Zeiss Axiophot microscope

equipped with epi-illumination (530-585 nm excitation; filterset4879(0) .

Confocallaser scanning microscopyConfocal microscopy was carried out using a confoeal laser scanning

microscope (Lcitz DM RBE; Leica, Heielelberg/Germany) equippedwith an argon krypton ion laser. For experiments, 488 nm excitationlight was selected by a narrow bandpass interference filter. A 40-foleloil immersion objective (PL FLUOTAR) with a numerical aperturefrom 0.5 to 1.0 was chosen. In the laser scanning mode the theareticallateral resolution was calculateel to be 0.2 lW1 (NA = 1.0; ), = 488 nm)ami the axial resolution 0.22 [1m (NA = 1.0; "A = 488 nm). Yertical

optical sections were computed from x-z scans with 512 lines/image.Photographs were taken with an APX-100 film (Agfa Gevaert, Lever­kusen/Germany) from a black and white monitor.

11 2kDa -

84kDa -

COM VIM CK7 CK8 CK18 CK19

Fig. 2. SOS-PAGE and Western blots of cytoskelctal proteins ofI-IEC-I-A cells (H) and RL95-2 cells (R). Coomassie Blue staining(COM) and immunoreactions with antibodies to vimentin (VIM), to

cytokeratin 7 (CK7), 8 (CK8), 18 (CKI8) and 19 (CK19). Thc proteinswere resolved using 7.5 % acryl amide gels. l'restained rcference pro­teins used for coelectrophoresis are phosphorylase B (1l2 000 Da),bovine serum albumin (84000 Da), ovalbumin (53200 Da), and car­bonie anhydrase (34900 Da). Note that the same pallern of cytokera­tin polypeptides ami vimentin is found in both HEC-I-A eells andRLLJ5-2 cells.

R H R H

,..•.

...••

R H R H

- •••• 1..:,. _":"""':f I"

R H R H

53.2kDa -

34,9kDa -

Transmission electron microscopyCells grown as monolayers on polY-D-lysine-coated thermanox cover­

slips (Nunc, Naperville, [L/USA) were rinsed twice in PBS anel fixedin 2.5 % glutaraldehyde in 0.1 M cacodylate buffer, pI-I 7.4, for 30 minat room temperature. After repeated washings in cacodylate buffer,sampies were fixed with 1 % OS04 in cacodylate buffer, dehyelrateelwith graeled ethanol anel propylene oxiele, anel embedded in epoxyresin mixture [8]. The embedded monolayers were separated from the

thermanox coverslips by snap freezing in liquid nitrogcn. Ultrathin see­tions were mountcd on 200-mesh copper griels, elouble staineel withuranyl acetate and lead citrate, and examined with a Philips EM 400 at80 kY.

In some cultures ruthenium reel (4527.1; Roth, Karlsruhe/Gcrmany)

was added to the fixative applied to the apieal siele of confluent mono­layers. Sampies were washed twice in P8S and fixeel in 2.5 % glutaral­dehyele in cacodylate buffer, containing 0.1 '/"o ruthenium red for I h at4°C. After several washings (3 x 3 min) in eacodylate buffer, sampieswere postfixed with 1 % OS04 in cacodylate buffer containing 0.1 %ruthenium red for 3 h at room temperature [31]. Then sampies were

EJCB Remodelling of epithelial cell polarity 225

Fig.3. Immunostaining 01' HEC-I-A monolayers (a, eonventional

mieroseopy; c, eonfoeal vertieal image) and RL95-2 rnonolayers (h,eonventional mieroseopy) with antiboclies to ZO-I; note that I-IEC­I-A eells but not RL95-2 cells are positive for ZO-l. Arrolvs mark the

position 01' eell-to-eell contacts in monolaycrs (c). Photomierographs

01' ultrathin sections or I-IEC-I-A cells (d) am! RL95-2 cells (e) follow­ing ruthenium red labeling; ruthenium red penetrates into the lateralintercellular space in RL95-2 cells but not in IIEC-I-A cells. - 00Coverslip. - Stars: Growth medium. -c 1-11, 1-12I-IEC-I-A cells. - RI,R2 RL95-2 eclls. - Bars 10 etm (a, h), 10 ,Im (c), 0.5 [U11 (d, c).

I

226 M. Thie, P. Fuchs, S. Butz, F. Sieckmonn et 01. EJCB

EJCB Remodelling 01 epithelial cell polarity 227

(Da), 52000 Da, 45000 Da, and 40000 Da. In Western blotswith eytokeratin antibodies eytokeratins 7 (54000 Da), 8(52000 Da), 18 (45000 Da), and 19 (40000 Da) were iden ti­fied. Vimentin also gave a weak band (57000 Da) although notbeing deteeted by immunohistoehemistry formerly [511. Thesame pattern of eytokeratin polypeptides and in eontrast toformer studies (see above) also vimentin is found in eytoskele­tal residues from polarized !-IEC eells (Fig. 2). Thus, a eyto­keratin polypeptide pattern eonsisting of eomponents 7,8, 18,19 as weil as vimentin has been found for both, nonpolarizedRL eclls and polarized HEC eells, indieating their origin fromsingle-Iayered endometrial epithelium known to eontain thesecytoskeletal proteins [35, 37].

200kDo -

97.4kDo -

69kDo -

46kDo -

•• =-~~•-H R

Structure of epithelial junctional complexes'fight jUllctiolls. Thc lateral plasma mcmbrancs 01' adjaeentHEC eclls are aligncd in parallel amI form in their subapiealpart eharaeteristie tight junetionalmembrane eontaets eonsist­ing of aseries of fusion spots (Fig. Je). In freeze I'raetures, thetight junetion struetures appcared as smooth strands on theprotoplasmie face 01' the repliea and as furrows on the exo­plasmie face 01' the repliea (data not shown). Using immuno­histochemistry, the tight junetion protein ZO-I was likewiseI'ound at the subapieal part of thc lateral bordcr of adjaccnteells (Figs. 3a, e). Ruthenium red added to thc apieal siele ofHEC monolayers staincd intensely the apieal cell surfaeeincluding the membrane of luminal vesicles and the eontent ofmany apparently eytoplasmie vesicles, but no staining wasobserved beyond the level of the junctional contaet points(Fig. 3d) thus proving that thesc junclions were indccd ancffeetive penetration barrier.

In eontrast to HEC eclls, no eharaeteristie tight junctionswerc seen at the mcmbrane eontaets of adjacent RL eells intransmission eleett'on mieroseopy (Fig. Id). Also in frcczefraetures tight junetion strands wcrc nevcr obscrved (data notshown). In keeping with this, the tight junclion-assoeiatedprotein ZO-I was not demonstrable by immunohistoehemistry(Fig. 3b). The tracer ruthcnium rcd was found to be able topcnetrate in the lateral intercellular space from the apiealeompartmcnt bcyonel the subapieal region, indicating the lackof a fully devcloped light junetional barrier (Fig. 3e).AdhcrclIs jllllctiolls. Conventional immunofluoresecnce mi­eroseopy revea1ed that E-eadhcrin, a-eatcnin, r3-eatenin andplakoglobin are cxprcssed by H EC monolaycrs (data notshown). Confoeal mieroseopy showed that E-eaelherin(Fig. 4a), a-eatenin (Fig. 4e), r3-eatenin (Fig. 4e), ami plako­globin (Fig. 4g) wcre eonfined to ecll-to-eell eontaets.

Fig.4. Confocal vertical images 01' I-IEC-I-A monolayers (a, c, c, g)and RL95-2 monolayers (h, 11, f, h) after staining with monoclonalantibody to E-cadherin (a, h), a-catenin (c, 11), ß-catenin (c, f), andplakoglobin (g, h). Vertical sections revcal that E-cadherin, n-catenin,ß-catcnin, and plakoglobin were confined to cell-to-ccll contacts inHEC-I-A cells. In RL95-2 cells, only [3-catenin was confined to thelatcral plasma membrane, whilc E-cadhcrin, n-catcnin, and plakoglo­bin were loeated at all plasma membranc domains. Notc that n-eatcninfluorescence (11) as weil as plakoglobin fluoreseence (h) suggest anintracell\i1ar localization, too. JlrrulVs mark the position of cell-to-celleontacts in monolaycrs. - 00 Coverslip. - Stars: Growth medium. ­Bar 10 ,Im.

Fig. 5. Immunoprecipitations 01' E-eadherin/catenin complexes fromradiolabeled HEC-I-A cells (Ir) and RL95-2 cells (R). Cell extractswere immunoprecipitated with anti-E-cadherin (E-cadherin), anti-n­eatenin (n-catenin), anti-ß-catenin (ß-eatenin), ami anti-plakoglobin(plakoglobin). In both HEC-I-A cells and RL95-2 eells two differentE-eadherin/catenin complexes were detected, i.e. E-cadherin (E-cad)amI n-eatenin (n-eat) in association with ß-catenin (f3-eat) as weil as E­eadherin and n-catenin in association with plakoglobin (pg). Noteadditional proteins in I-IEC-I-A cell extracts when immunoprecipi­tated with anti-E-eadherin, anti-\1-catenin, ancl anti-plakoglobin. Theproteins were rcsolved by SOS-PAGE using 7.5 % acrylamide gels.Refercnce protcins used for eoeleetrophoresis are 14C-methylatcd myo­sin (200000 Da), "C-methylated phosphorylase B (97400 Da), 14C_methylated bovine serum albumin (69000 Da), and 14C-methylatedovalbumin (46000 Da).

When eell lysates I'rom metabolieally labeled HEC eellswere subjcetcel to immunopreeipitation using antiboclies spc­cific for E-eaelhcrin, four major protcin bands of moleClIlarwcight 120000 Da, 102000 Da, 88000 Da, and 80000 Da weredeteeted (Fig. 5). When sueh anti-E-eadherin immunopreeip­itate gels were bIotted anel subsequently stained with anti-E­cadhcrin, anti-a-catcnin, anti-ß-eatenin, ami anti­plakoglobin, it was possiblc to identify the major bands as E­eaelherin (120000 Da), a-eatenin (102000 Da), ['\-eatenin(88000 Da), and plakoglobin (80000 Da) (data not shown).Immunopreeipitates with antiboely speeifie to a-eatenin andsubsequent immunostaining with anti-E-eadherin, anti-a­eatenin, anti-ß-eatenin, anel anti-plakoglobin revealed E­eadherin, (1-eatcnin, ß-eatenin, ami plakoglobin (Fig. 5).

Immunopreeipitates eolleeted with anti-ß-catenin or anti­plakoglobin differed, howcvcr, insofar as either r)-eatenin orplakoglobin was not deteetable in the alternative immunopre­eipitate. This inelieates that two elifferent E-eadherin-catcnineomplexes are present in HEC eells, one composed of E­eadhcrin, a-eatenin and ß-eatenin, the other of E-eadherin, (1­eatenin anel plakoglobin. l3eyond that, immunopreeipitationseollected with anti-E-cadherin, anti-n-eatenin, anel plakoglo­bin revealed additional baneIs in the range oI 60000 Da to50000 Da whieh are, however, not identified so rar.

RL eells showeel, like I-IEC cells, thc presence of E­eadherin, a-eatenin, ß-eatenin and plakoglobin (data fromeonventional immunohistoehemistry; not shown). Confoealmieroseopy revealed that E-cadherin (Fig. 4b), (X-catenin(Fig. 4d), and plakoglobin (Fig. 4h) were colocalizcd in allplasma membrane elomains of these eells (in eontrast to I-IECeells), while r)-eatenin (Fig. 41') was loealized in lateral mem­brane domains of adjaeent eells. Moreover, plakoglobin l'Iuo-

I

228 M. Thie, P. Fuchs, S. Butz, F. Sieckmonn et 01.

rescence (Fig. 4h) and also a-catenin fluorescence (Fig. 4d)suggest an intracellular localization of plakoglobin and a­catenin, too.

When celllysates from metabolieally labeled RL cells wercsubjected to immunoprecipitations with antibodies against E­cadherin, a-catenin, ß-eatenin, and plakoglobin (Fig. 5) thesame pattern of bands was obtained as with HEC cells. In con­trast to !-lEC cells, however, RL eells laeked additional bands

in the range of 60000 Da to 50000 Da. Comparison of celllines leads to the conclusion that in both cell types two differ­ent E-cadherin-catenin complexes are present, one complexeontaining E-eadherin and a-catenin in association with ß­catenin, the other E-cadherin and a-catenin in association

with plakoglobin.

EJCB

Actin distribution. Actin filament arrangements as seen afterstaining with TRITC-conjugated phalloidin differed in RL and!-lEC cells as shown in Figure 6.

I-lEC eells showed spot-like actin staining associated withthe microvilli of the apical plasma membrane (Fig. 6a),peripheral bands whieh surrounded the margin of the cells(Figs. 6a, b, d) as weil as stress fibers which ran longitudinallyat the basc of thc cclls (Fig. 6b).

A different actin organization was observed in RL cells(Figs. 6e, e). In contrast to !-lEC cells, microvilli-assoeiatedstaining as weil as stress fibers could not be observed in RLeells (Fig. 6e). Ncvcrthclcss, cells showed actin staining alongthe entire cell surfacc anc! diffusc staining within the cyto­plasm. The appearance of actin staining beneath the plasma

Fig.6. Actin staining of HEC-I-A monolayers (a, h, convcntionalmicroscopy; a: focuscd apically; h: focused basally; d: confocal vertica!image) and RL95-2 monolayers (e, conventional microscopy; c, confo­cal vertical image) with TRITC-conjugated phalloidin. Note that

I-IEC-1-A cclls but not RL95-2 cells show fluoresccnce associated with

microvilli (a), peripheral bands, ami stress fibers (h). ArrolVs mark theposition of cell-to-cell contacts in vertical images. - 00 Coverslip. ­Stars: Growth medium. - Bars 10 run (a-e), 10 firn (d, c).

EJCB

membrane varied in different regions of the same eell, i.e. insome areas large clumps of Ouoreseent material were seen, inother areas only sm all spots (Fig. 6e).Dcsmosomcs. In HEC eells, transmission electron mieroseopyshowed typieal desmosomes loealized basally to the tight june­tions (Fig. le). 8y immunohistoehemistry, desmoplakin Ieould be identified in punetate arrays (Fig. 7a), restrietedalong eell-to-eell eontaets, i.e. at the lateral plasma membrane(Fig.7e).

In RL eells, lateral plasma membranes usually laeked typ­ieal desmosomes (Fig. Id) wh ich were only occasionallyfound. Desmoplakin I was demonstrablc immunhistochemi­cally (Fig. 7b) and, in contrast to HEC cells, was not restrictedto cell-to-eell eontaets but also seen in the luminal and basal

mcmbranes (data not shown). Further eleetron mieroseopystudies (e.g. [9]) will help understand the distribution andformation of desmosomes in these eells.

Discussion

RL95-2 eells represent a human uterine epithelial eell linewhieh displays adhesive properties of its apieal plasma mem­brane for trophoblast-type eells [27]. In this respeet, RL eellsmight aet as an in vitro model for the reeeptive uterine epithe-

Remodelling of epithelial cell polarity 229

lium. Reeent studies have indicated that apical adhesivenessof RL eells is associated with a low degree of their apical-basalcell polarity [50, 511. Focusing on steps leading to modulationof eell polarity in vivo, we mapped cell-to-cell contacts in RLcells and found defects of integrity of their junctional com­plex. RL eells showed a lack of ZO-l expression ami tracerleakiness of the paracellular pathway, and atypical features inadherens junetions. Additionally, the aetin-based eytoskelctonlacked apolar organization. Moreover, RL cells expresseddesmoplakin but lacked mostly well-formed desmosomes.

Although expressing E-cadherin, n-catenin, r3-catenin, andplakoglobin, RL eells lack well-organized adherens junetions.On the other hand, evidence was seen for an atypieal distribu­tion of these molecules, i.e. random loealization of E­

eadherin, a-eatenin and plakoglobin at all plasma membranedomains but restrietion of loealization of r3-eatenin to the later­al domains. E-cadherin is recognizcd as a master molecule forthe maintenanee of epithelial integrity and polarized organiza­tion [17,36]. The eytoplasmie domain of E-cadherin interactswith a family of submembranous plaque proteins of which arepresently known a-catenin, ß-catenin, plakoglobin [40] andp 120cas described recently [44]. These plaque proteins interactin a yet unidentifiecl way with structural components of theactin cytoskeleton [40, 48]. The ranclom localization of E­eaclherin, a-catenin ancl plakoglobin on the one side and the

Fig. 7. Immunostaining of I-IEC-l-A monolayers (a, conventionalmicroscopy; e, confocal vertical image) ami RL95-2 monolaycrs (h,conventional microscopy) with antibodies to desmoplakin 1. Desmo­plakin was found in punctate arrays along the cell-to-cell contacts inHEC-I-A cells. In RL95-2 cells staining was not only restrieted to late-

ralmembrane domains of adjacent eells but also seen in the apieal andbasal membrane domains. Arrows mark the position of cell-to-cellcontact in vertical images. - 00 Coverslip. - Slar: Growth mcdium. ­

Bars 10 ~lm (a, h), 10 ~lm (e).

I

230 M. Thie, P. Fuchs, S. Butz, F. Sieckmonn et 01.

laterallocalization of ß-catenin on the other side might be dueto (partial) disassembly of the complexes formed between E­cadherin, plaque proteins and the actin filament network [3,22, 24, 25, 39, 45]. The intracellular localization of plakoglo­bin and a-catenin (beyond their localization in plasma mem­branes) might point to internalization of distinct plaque com­ponents in RL cells. A decrease of function of adherens junc­tions is supposed to be controlled by the level of protein tyro­sine phosphorylation. For example, transformation of MDCKcells (an epithelial cellline) with v-src or inhibition of proteintyrosine phosphatase activity with pervanadate induced eleva­tion of tyrosine phosphorylation of catenins and functionalinactivation of cadherin-mediated cell-to-cell adhesion accom­

panied by dissociation of the junctional actin bundle from themembrane and followed by rounding up of the apical cell sur­faces and perturbation of epithelial monolayer morphology[2, 18, 34, 44, 53, 54]. Interestingly, complexes formedbetween E-cadherin, a-catenin and ß-catenin, and E-cadherin,a-catenin and plakoglobin are detectable in RL cells. Thus,combined evidence from confocal immunofluorescence

microscopy and from immunoprecipitations suggests that E­cadherin/plakoglobin complexes are located randomly and E­cadherin/ß-catenin complexes are located laterally in thesecells. The differential distribution of both complexes in thesame cell might point to differences in their association withthe actin-based cytoskeleton [32, 52]. Processes leading totyrosine phosphorylation of different catenins and thus inacti­vating the association of distinct catenins with actin filamentsmight be involved here [18, 23]. This leads to the questionwhether E-cadherin, a-catenin and ß-catenin or E-cadherin,a-catenin and plakoglobin might affect the epithelial pheno­type by directly or indirectly influencing the generation ofstructurally and functionally distinct membrane domains.Indeed, cells can differentially regulate the organization ofcadherin and associated cytoplasmic molecules at junctions inrelation to different functional requirements. In vascularendothelial cells assembly and disassembly of intercellularjunctions have been described to comprise several steps thatare regulated differently. Loose junctions as in early confluentcells are characterized by cadherin, a-catenin, and ß-catenindeposition at contact sites, and plakoglobin is not found here.Stable junctions as in tightly confluent cells are characterizedby plakoglobin localization at intercellular contact sites. Thesedata suggest that while cadherin, a-catenin, and ß-catenincomplexes can function as early recognition mechanismsbetween endothelial cells, the strength of the cell-to-cell con­tact can be modulated by plakoglobin accumulation at junc­tions [29]. Moreover, regulated expression of different cadhe­rin family members might lead to specialized cell phenotypesas shown in the retinal pigment epithelium system [33].

RL cells did not only show a modulation of adherens junc­tions but also alterations of tight junctions as they are charac­terized by a lack of ZO-1 expression and tracer leakiness ofthe paracellular pathway. It can be assumed that the level ofpro tein tyrosine phosphorylation influences adherens junc­tions as weil as tight junctions. Although pervanadate had noeffect on ZO-1 immunostaining in MDCK cells [54] andexpression of low levels of the tyrosine kinase v-src had noapparent effect on the thin section electron microscopicappearance of tight junctions [55], phosphorylation of tightjunction associated proteins caused an increase in the ionicpermeability in these cells [2, 47]. The apolar cytoskeletalarchitecture we observed in RL cells might point to an

involvement of the actin cytoskeleton in tight junctionmodulation. Indeed, disruption of tight junctions (and ofadherens junctions), possibly leading to a loss of apical-basalcell polarity, might be generated by disassembly of the actin­based cytoskeleton as shown by antimycin-A-generated ATPdepletion in kidney epithelium-derived celllines. Following aredistribution of the actin microfilament network and a dis­ruption of the cortical cytoskeleton, the surface membraneundergoes extensive alterations in microvillous morphology,disruption of cellular junctions and loss of membrane polarity.With regard to the opening of tight junctions, lateral migrationof domain-specific components could occur, resulting in loss ofsurface membrane polarity, i.e. membrane lipids and proteinsare able to redistribute across domains. Na+/K+-ATPase,untethered from the cortical cytoskeleton, redistributes intothe apical domain, while the apical marker protein leucineaminopeptidase redistributes into the basolateral domain [30,42]. Similarly, investigations using the renal epithelial celllineBS-C-l oxidatively stressed by H202 treatment have revealedthat disruption of focal adhesions with loss of talin from thebasal cell surface untethers the associated integrins, allowingthem to redistribute into the apical domain [15]. In vivo,moderate restructuring of the actin-based cytoskeletal net­work occurs in response to a variety of extracellular and intra­cellular signals. One of them might be growth factors like epi­dermal growth factor which cooperate with c-src tyrosinekin ase and GTPase-activating proteins [6].

An interesting question is whether the interdependent phe­nomena of a remodelling of cell-to-cell adhesion system andthe rearrangements of the actin cytoskeleton might depend ona master gene pro gram which, triggered by steroids, controlsthe development of the receptive state in the uterine epithe­lium [10-12]. Such master genes could be related to those pos­tulated to govern epithelium-to-mesenchyme transformationin embryology and cancer progression [19, 43, 46]. Apart fromE-cadherin mentioned earlier, the gene encoding fibroblast­specific protein 1 (FSPl) [49] has been proposed to representsuch a master gene [20]. Signals such as tumor-promotingphorbol esters, oncoproteins, and growth factors may in turncontrol master gene activation in these systems. It will be achallenge for further investigations on human uterine epi­thelial cells to look for direct evidence whether the gene pro­gram for the epithelial phenotype is indeed undergoing amodulation in connection with the acquisition of receptivityfor trophoblast.

Acknowledgements. The skillfull technical assistancc of Birgit Nowak,Dorothea Schünke, and Ulrike Tlolka is gratefully acknowledged. Wealso wish to thank Dr. J. Behrens and Dr. J. Kartenbeck für the gener­ous gifts of antibodies. We would like to thank Dr. G. Bruder and Dr.J. Kartenbeck für help with the choice of cytokeratin antibodies and toDr. R. Moll for helpful comments to intermediate filament analysis.We are also obliged to Prof. Dr. C. Streffer (Directür of the Institutefor Radiation Biology of the University, Essen/Germany) for makingavailable to us the confocal microscopy unit and für his interest in thiswürk.

References

[1] Albers, A., M. Thie, H.-P. Hohn, H.-W. Denker: Differentialexpression and localization of integrins and CD44 in the membranedomains of human uterine epithelial cell during the menstrual cyde.Acta Anat. 153, 12-19 (1995).

[2] Behrens, J., L. Vakaet, R. Friis, E. Winterhager, F. van Roy,M. M. Mareei, W. Birchmeier: Loss of epithelial differentiation andgain of invasiveness correlates with tyrosine phosphorylation of the E­cadherin/ß-catenin complex in cells transformed with a temperature­sensitive-v-SRC gene. J. Cell Bio!' 120,757-766 (1993).[3] Butz, S., R. Kemler: Distinct cadherin-catenin complexes in Ca 2+_

dependent cell-cell adhesion. FEBS Lett. 355, 195-200 (1994).[4] Butz, S., S. Rawer, W. Rapp, U. Birsncr: Immunization and affin­ity purification of antibodies using resin-immobilized lysine-branchedsynthetic peptides. Pept. Res. 7,20-23 (1994).[5] Butz, S. J. Stappert, H. Weissig, R. Kemler: Plakoglobin and ß­catenin: distinct but closely related. Science 257, 1142-1144 (1992).[6] Chang, J.-H., S. Gill, J. Settleman, S. J. Parsons: c-Src regulatesthe simultaneous rearrangement of actin cytoskeleton, pI90RhoGAP,and p120Ras GAP following epidermal growth factor stimulation. J.Cell Bio!' 130,355-368 (1995).[7] Cowin, P., H.-P. Kapprell, W. W. Franke: The complement of des­mosomal plaque proteins in different cell types. J. Cell Bio!' 101,1442-1454 (1985).[8] Cross, R. H. M.: A reliable epoxy resin mixture and its applicationin routine biological transmission electron microscopy. MicronMicrose. Acta 20, 1-7 (1989).[9] Demlehner, M. P., S. Schäfer, C. Grund, W. W. Franke: Continualassembly of half-desmosomal structures in the absence of cell contactsand their frustrated endocytosis: a coordinated sisyphus cycle. J. CellBio!' 131,745-760 (1995).[10] Denker, H.-W.: Trophoblast-endometrial interactions at embryoimplantation: a cell biological paradox. In: H.-W. Denker, J. D. Aplin(eds.): Trophoblast Invasion and Endometrial Receptivity. NovelAspects of the Cell Biology of Embryo Implantation. TrophoblastResearch Vo!. 4. pp. 3-29. Plenum Medical Book Comp. New York,London 1990.[11] Denker, H.-W: Implantation: a cell biological paradox. J. Exp.Zoo!. 266, 541-558 (1993).[12] Denker, H.-W: Cell biology of endometrial receptivity and oftrophoblast-endometrial interactions. In: S. R. Glasser, J. Mulholland,A. Psychoyos (eds.): Endocrinology of Embryo-Endometrium Inter­actions. pp. 17-32. Plenum Press. New York, London 1994.[13] Eckert, W, J. Kartenbeck: Methoden der Zell- und Molekular­biologie. Teil 1. Springer-Verlag. New York, London 1996.[14] Frixen, U. H., J. Behrens, M. Sachs, G. Eberle, B. Voss, A.Warda, D. Löchner, W. Birchmeier: E-cadherin-mediated cell-celladhesion prevents invasiveness of human carcinoma cells J. Cell Bio!'113,173-185 (1991).[15] Gailit, J., D. Coflesh, 1.Rabiner, J. Simone, M. S. Goligorsky:Redistribution and dysfunction of intcgrins in cultured renal epithelialcells exposed to oxidative stress. Am. J. Physio!. 264, F149-F157(1993).[16] Glasser, S., J. Mulholland: Receptivity is a polarity dependentspecial function of hormonally regulated uterine epithelial cells.Microsc. Res. Techn. 25, 106-120 (1993).[17] Gumbiner, B., B. Stevenson, A. Grimaldi: The role of the celladhesion molecule uvomorulin in the formation and maintenanceof the epithelial junctiona1 complex. J. Cell Bio!, 107, 1575-1587(1988).[18] Hamaguchi, M., N. Matsuyoshi, Y. Ohnishi, B. Gotoh, M. Takei­chi, Y. Nagai: p60v-src causes tyrosine phosphorylation and inactiva­tion of the N-cadherin-catenin cell adhesion system. EMBO J. 12,307-314 (1993).[19] Hay, E. D.: Epithelia1-mesenchymal transitions. Semin. Dev.Bio!, 1, 347-356 (1990).[20] Hay, E. D.: An overview of epithelio-mesenchymal transforma­tion. Acta Anat. 154 (1) (in press, 1996).[21] Herrenknecht, K., M.Ozawa, C. Eckerskom, F.Lottspeich,M. Lenter, R. Kemler: The uvomorulin-anchorage protein a catenin isa vincu1in homo1gue. Proc. Nat!. Acad. Sci. USA 88, 9156-9160(1991).[22] Hinck, L., 1. S. Näthke, J. Papkoff, W J. Nelson: Dynamics ofcadherin/catenin comp1ex formation: nove1 protein interactions andpathways of comp1ex assemb1y. J. Cell Bio!, 125, 1327-1340 (1994).

Remodelling of epithelial cell polarity 231

[23] Hoschützky, H., H. Aber1e, R. Kemler: ß-Catenin mediates theinteraction of the cadherin-catenin complex with epidermal growthfactor receptor. J. Cell Bio!, 127, 1375-1380 (1994).[24] Hü1sken, J., J. Behrens, W. Birchmeier: Tumor-suppressor geneproducts in cell contacts: the cadherin-APC-armadi1lo connection.Curr. Opin. Cell Bio!, 6, 711-716 (1994).[25] Hülsken, J., W Birchmeier, J. Behrens: E-cadherin and APCcompete for the interaction with beta-catenin and the cytoske1eton. J.Cell Bio!, 127, 2061-2069 (1994).[26] Jansen, R. P. S., M. Turner, E. Johannisson, B. M. Landgren,E. Dicfa1usy: Cyclic changes in human endometria1 surface glycopro­teins: a quantitative histochemica1 study. Ferti!. Steri!. 44, 85-91(1985).[27] John, N. J., M. Linke, H.-W Denker: Quantitation of humanchoriocarcinoma spheroid attachment to uterine epithelial cell mono­layers. In Vitro Ce1!. Dev. Bio!, 29A, 461-468 (1993).[28] Laemm1i, U. K.: C1eavage of structura1 pro teins during theassembly of the head of bacteriophage T4. Nature 227, 680-685(1970).[29] Lampugnani, M. G., M. Corada, L. Cavcda, F.Breviario, O.Ayalon, B. Geiger, E. Dejana: The mo1ecular organization of endothe­lial cell to cell junctions: differentia1 association of p1akoglobin, ß­catenin, and a-catenin with vascular endothelia1 cadherin (VE­cadherin). J. Cell Bio!, 129, 203-217 (1995).[30] Leiser, J., B. A. Molitoris: Disease processes in epithelia: the roleof the actin cytoskeleton and altered surface membrane po1arity.Biochim. Biophys. Acta 1225,1-13 (1993).[31] Luft, J. H.: Ruthenium red and violet. 1. Chemistry, purification,methods of use for e1ectron microscopy and mechanism of action.Anat. Rec. 171, 347-368 (1971).[32] Luna, E. J., A. L. Hitt: Cytoskeleton-plasma membrane interac­tions. Science 258, 955-964 (1992).[33] Marrs, J. A., C. Andersson-Fisone, M. C. Jeong, L. Cohen­Gou1d, C. Zurzolo, 1. R. Nabi, E. Rodriguez-Bou1an, W. J. Nelson:Plasticity in epithelia1 cell phenotype: modu1ation by expression of dif­ferent cadherin cell adhesion molecu1es. J. Cell Bio!, 129, 507-519(1995).[34] Matsuyoshi, N., S. Hamaguchi, S. Taniguchi, A. Nagafuchi,S. Tsukita, M. Takeichi: Cadherin-mediated cell-cell adhesion is per­tubed by v-src tyrosine phosphory1ation in metastatic fibrob1asts. J.Cell Bio!, 118,703-714 (1992).[35] Matthews, c.J.,C.P.F.Redfern, B.H.Hirst, E.J.Thomas:Characterization of human purified epithelia1 and stroma1 cell fromendometrium and endometriosis in tissue cu1ture. Ferti!. Steri!. 57,990-997 (1992).[36] McNei1l, H., M. Ozawa, R. Kemler, W. J. Nelson: Novel functionof the cell adhesion moleeule uvomoru1in as an inducer of cell surfacepolarity. Cell 62, 309-316 (1990).[37] Moll, R., R. Levy, B. Czernobilsky, P. Hohlweg-Majert, G.Dahlenbach-Hellweg, W. W. Franke: Cytokeratins of normal epithe1iaand some neop1asms of the fema1e genita1 tract. Lab. lnvest. 49,599-610 (1983).[38] Murphy, C. R.: The p1asma membrane of uterine epithelial cells:structure and histochemistry. Progr. Histochcm. Cytochem. 27, 1-66(1993).[39] Näthke, 1.S., L. Hinck, J. R. Swed1ow, J. Papkoff, W J. Ne1son:Defining interactions and distributions of cadherin and catenin com­1'!"\'" i" 1,,,I;,ri;"<I ,'pithelial cells. J. Cell Bio!, 125, 1341-1352 (1994).[4()] '\'a~afllchi. A .. S. Tsukita, M. Takeichi: Transmembrane contro1of cadherinmediated cell-cell adhesion. Semin. Cell Bio!, 4, 175-181(1993).[41] Ozawa, M., H. Baribault, R. Kem1er: The cytoplasmic domain ofthe cell adhesion mo1ecu1e uvomoru1in associates with three indepen­dent pro teins structurally related in different species. EMBO J. 8,1711-1717 (1989).[42] Paller, M. S.: Latera1 mobility of Na, K-ATPase and membranelipids in renal cells. Importance of cytoskeletal integrity. J. Membr.Bio!, 142, 127-135 (1994).[43] Reichmann, E.: Oncogencs and epithc1ia1 ccll transformation.Semin. Cancer Bio!, 5, 157-165 (1994).

232 M. Thie, P. Fuchs, S. Butz, F. Sieckmonn et 01.

[44] Reynolds, A. B., J. Daniel, P. McCrea, M. J. Wheelock, J. Wu,Z. Zhang: Identification of a new catenin: the tyrosine kinase sub­strate p120cas associates with E-cadherin complexes. Mo!. Cello Bio!'14,8333-8342 (1994).[45] Rubinfeld, B., B. Souza, 1. Albert, S. Munemitsu, P.Polakis: TheAPC protein and E-cadherin form similar but independent complexeswith alpha-catenin, beta-catenin, and plakoglobin. J. Bio!' Chem. 270,5549-5555 (1995).[46] Savagner, P., B. Boyer, A. M. Valles, J. Jouanneau, J. P.Thiery:Modulations of the epithelial phenotype during embryogenesis andcancer progression. Cancer Treat. Res. 71,229-249 (1994).[47] Staddon, J. M., K. Herrenknecht, C. Smales, L. L. Rubin: Evi­dence that tyrosine phosphorylation may increase tight junction per­me ability. J. Cell Sci. 108, 609-619 (1995).[48] Stappert, J., R. Kemler: Intracellular association of adhesionmolecules. Curr. Opin. Neurobio!. 3, 60-66 (1993).[49] Strutz, F., H.Okada, C. W. Lo, T. Danoff, R. L. Carone, J. E.Tomaszewski, E. G. Neilson: Identification and characterization of afibroblast marker: FSPl. J. Cell Bio!, 130, 393-405 (1995).[50] Thie, M., A. Albers, H.-W. Denker: Endometrial receptivity: cellbiological aspects. In: T. Mori, T. Aono, T. Tominaga, H. Hiroi (eds.):

Perspectives on Assisted Reproduction. Serono Symposia Series Fron­tiers in Endocrinology Vo!. 4. pp. 331-337. Ares-Serono Symposia.Christengraf S. r. 1. Rome 1994.[51] Thie, M., B. Harrach-Ruprecht, H. Sauer, P.Fuchs, A. Albers,H.-W. Denker: Cell adhesion to the apical pole of epithelium: a func­tion of cell polarity. Eur. J. Cell Bio!, 66, 180-191 (1995).

[52] Tsukita, S., M. Itoh, A. Nagafuchi, S. Tonemura, S. Tsukita: Sub­membranous junctional plaque proteins include potential tumor sup­pressor molecules. J. Cell Bio!, 123, 1049-1053 (1993).

[53] Volberg, T., B. Geiger, Y. R. Dror, Y. Zick: Modulation of inter­cellular adherenstype junctions and tyrosine phosphorylation of theircomponents in RSV transformed cultured chick lens cells. Cell Regu!.2,105-120 (1991).[54] Volberg, T., Y. Zick, R. Dror, LSabanay, C. Gilon, A. Levitzki,B. Geiger: The effect of tyrosine-specific protein phosphorylation onthe assembly of adherens-type junctions. EMBO J. 11, 1733-1742(1992).[55] Warren, S. L., W. J. Nelson: Nonmitogenic morphoregulatoryaction of pp60v-src on multicellular epithelial structures. Mo!. Cel!.Bio!, 7, 1326-1337 (1987).