Behavior of tight-junction, adherens-junction and cell polarity proteins during HNF-4α-induced epithelial polarization

Experimental Cell Research 310 (2005) 66 – 78 www.elsevier.com/locate/yexcr

Research Article

Behavior of tight-junction, adherens-junction and cell polarity proteins during HNF-4a-induced epithelial polarization Seiro Satohisa a,b, Hideki Chiba a,*, Makoto Osanai a, Shigeo Ohno c, Takashi Kojima a, Tsuyoshi Saito b, Norimasa Sawada a a

Department of Pathology, Sapporo Medical University School of Medicine, South-1, West-17, Chuo-ku, Sapporo 060-8556, Japan b Department of Obstetrics and Gynecology, Sapporo Medical University School of Medicine, Sapporo 060-8556, Japan c Department of Molecular Biology, Yokohama City University School of Medicine, Yokohama 236-0004, Japan Received 31 March 2005, revised version received 27 June 2005, accepted 30 June 2005 Available online 10 August 2005

Abstract We previously reported that expression of tight-junction molecules occludin, claudin-6 and claudin-7, as well as establishment of epithelial polarity, was triggered in mouse F9 cells expressing hepatocyte nuclear factor (HNF)-4a [H. Chiba, T. Gotoh, T. Kojima, S. Satohisa, K. Kikuchi, M. Osanai, N. Sawada. Hepatocyte nuclear factor (HNF)-4a triggers formation of functional tight junctions and establishment of polarized epithelial morphology in F9 embryonal carcinoma cells, Exp. Cell Res. 286 (2003) 288 – 297]. Using these cells, we examined in the present study behavior of tight-junction, adherens-junction and cell polarity proteins and elucidated the molecular mechanism behind HNF-4a-initiated junction formation and epithelial polarization. We herein show that not only ZO-1 and ZO-2, but also ZO-3, junctional adhesion molecule (JAM)-B, JAM-C and cell polarity proteins PAR-3, PAR-6 and atypical protein kinase C (aPKC) accumulate at primordial adherens junctions in undifferentiated F9 cells. In contrast, CRB3, Pals1 and PATJ appeared to exhibit distinct subcellular localization in immature cells. Induced expression of HNF-4a led to translocation of these tight-junction and cell polarity proteins to beltlike tight junctions, where occludin, claudin-6 and claudin-7 were assembled, in differentiated cells. Interestingly, PAR-6, aPKC, CRB3 and Pals1, but not PAR-3 or PATJ, were also concentrated on the apical membranes in differentiated cells. These findings indicate that HNF-4a provokes not only expression of tight-junction adhesion molecules, but also modulation of subcellular distribution of junction and cell polarity proteins, resulting in junction formation and epithelial polarization. D 2005 Elsevier Inc. All rights reserved. Keywords: HNF-4; Nuclear receptors; Cell polarity; Tight junctions; Adherens junctions; F9 cells; Claudin; JAM; Atypical protein kinase C; PAR

Introduction Tight junctions are the apical-most constituent of an intercellular junctional complex in mammalian epithelial cells. They are regarded to function as a barrier that allows the selective passage of ions and solutes through the paracellular pathway, as well as working as a fence dividing plasma membranes into apical and basolateral domains [1,2]. They also recruit various signaling proteins controlling gene expression, cell proliferation and cell polarity, * Corresponding author. Fax: +81 11 613 5665. E-mail address: [email protected] (H. Chiba). 0014-4827/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.yexcr.2005.06.025

thereby acting as a multifunctional complex [2 –5]. So far, three families of transmembrane proteins of tight junctions, occludin, claudins and junctional adhesion molecules (JAMs), as well as increasing numbers of their scaffold proteins, have been discovered [2,6 – 12]. Among them, claudins are the backbone of tight junctions because they can form continuous networks of intramembranous fibrils (tight-junction strands) [13]. They consist of over 20 members of a gene family in mice and humans, and two or more distinct claudins are generally expressed in a celland tissue-specific manner [2,14]. Recently, two cell polarity protein complexes, which are highly conserved throughout evolution, have been shown to

S. Satohisa et al. / Experimental Cell Research 310 (2005) 66 – 78

localize at tight junctions in vertebrate epithelia [5,15 – 25]. The first complex is PAR-3 (partitioning-defective-3)/PAR6/aPKC (atypical protein kinase C), of which PAR-3 is known to associate directly with JAM-A, JAM-B and JAMC [26 – 28]. The second complex is composed of CRB3 (Crumbs3), Pals1 (protein associated with Lin Seven 1) and PATJ (Pals1-associated tight-junction protein), and the sixth and eighth PDZ domains of PATJ bind to tight-junction proteins ZO-3 and claudin-1, respectively [29]. These two complexes participate in the establishment of apical – junctional complexes and apicobasal cell polarity [5,19,21,25], which are thought to be critical for epithelial cells to separate the distinct tissue compartment and control directed uptake and secretion of molecules between them. Moreover, they interact via a direct interaction of Par6 with either CRB3 or Pals1 [30 – 32], as well as binding of aPKC to both CRB and PATJ [33]. Thus, although molecular components of tight junctions and their interactions have been rapidly disclosed, it remains obscure which signals cue tightjunction formation and epithelial polarization. Hepatocyte nuclear factor-4a (HNF-4a) belongs to the nuclear receptor superfamily and transcriptionally controls expression of a large number of target genes involved in nutrient and drug metabolism, hematopoiesis and blood coagulation [34 – 38]. Whereas fatty acyl CoA thioesters are identified as endogenous ligands for HNF-4a [39], unlike traditional ligands for other nuclear receptors, they constitutively bind HNF-4a and induce neither conformational changes in HNF-4a nor alteration of its binding to corepressors or coactivators [40]. During early development, it is first detected in primitive endoderm (PrE) cells and subsequently expressed in visceral endoderm (VE) cells [41], which share many properties with hepatocytes [38]. In the adult, HNF-4a is known to be expressed in various organs containing epithelial cells, including not only the liver but also the kidney, intestine, pancreas and stomach [36]. Several lines of evidence have strongly suggested that HNF-4a plays important roles in morphological and functional differentiation of both VE cells and hepatocytes [42 – 45], but it has not been clarified by which mechanisms HNF-4a initiates epithelial differentiation. Mouse F9 embryonal carcinoma cells exhibit very little spontaneous differentiation when cultured in the absence of retinoic acid. On the other hand, those grown as monolayers and aggregates differentiate upon retinoic acid treatment into PrE and VE-like cells, respectively, both of which are polarized epithelial cells possessing junctional complexes and microvilli [46,47]. Based on these properties, F9 cells should provide an attractive system for studying not only early embryonic development, cell differentiation and retinoid signaling, but also the processes of establishment of both junctional complexes and epithelial polarity. In sharp contrast, to analyze steps of junction formation and epithelial polarization in mature epithelial cell lines such as MDCK and MTD1-A, junctions and/or polarity are once broken and subsequently repaired by experimental manip-

67

ulations including calcium switch ([48] and references therein) or wound healing assays [26,49 – 51]. To facilitate the analysis of gene functions in F9 cells, we previously generated cell line F9:rtTA:Cre-ERT L32T2 (also called F9 L32T2), which allows sequential inactivation of loxPflanked genes and tight control of gene expression, without impairing their general characteristics [52]. Using this line, we subsequently showed that two members of the nuclear receptor superfamily, retinoid receptors and HNF-4a, triggered expression of three tight-junction molecules, occludin, claudin-6 and claudin-7, as well as formation of functional tight junctions and epithelial polarity [53,54]. To gain a greater insight into the molecular mechanisms by which tight junctions and cell polarity are acquired, we, in the present study, examined whether the expression of any other cell adhesion and cell polarity molecules was activated in F9 cells by HNF-4a. We also analyzed subcellular localization of these proteins in undifferentiated and differentiated F9 cells in comparison with models of epithelial polarization in Drosophila and mammalian epithelial cells. Furthermore, we attempted to reevaluate potent functions of several nuclear receptors in the establishment of junctional complexes and epithelial polarity.

Materials and methods Cell lines and cell culture F9 cells possessing both doxycycline (Dox)-inducible gene expression (Tet-on; [55]) and tamoxifen-dependent Cre-mediated recombination [56] systems were established as described previously (F9 L32T2; [52]). The F9 L32T2 cells were electroporated with the expression vector pUHD10-3-rHNF-4a, in which the expression of rHNF4a1 is under the control of the tet-operator, along with the puromycin-resistant gene expression vector pHRLpuro1 [57], and puromycin-resistant clones were isolated as described [54,58]. F9 cells expressing Dox-inducible HNF-4a (F9 L32T2:HNF-4a) were selected by immunofluorescence staining and immunoblotting using an antiHNF-4a polyclonal antibody (pAb) (Santa Cruz Biotechnology, Santa Cruz, CA). Cells were plated in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (Sanko Junyaku, Tokyo, Japan), 100 U/ml penicillin and 100 Ag/ml streptomycin, and treated with the vehicle or Dox 1 day after plating. The medium was changed every 2 days. Generation of polyclonal antibodies To generate rabbit pAbs against claudin-6 and claudin-7, the following conjugated peptides, in which a cysteine residue was added at the amino-terminal ends, were used as antigens. Polypeptides, CSRGPSEYPTKNYV and

68

S. Satohisa et al. / Experimental Cell Research 310 (2005) 66 – 78

CYRAPRSYPKSNSSKEY, corresponding to the cytoplasmic domains of mouse claudin-6 (amino acids [aa] 207 – 219) and claudin-7 (aa 195– 210), respectively, were synthesized and coupled via the cysteine to keyhole limpet hemocyanin. The antisera were affinity purified with glutathione S-transferase fusion protein with the carboxylterminal cytoplasmic domains of the corresponding claudins. Specificity of these antibodies was verified by Western blot analysis using COS-7 cells transiently transfected with individual expression vectors containing mouse claudin-1 to claudin-8 cDNAs [59]. In brief, COS-7 cells were grown on 60-mm culture plates and transfected with 10 Ag of each expression vector using Lipofectamine 2000 reagent (Life Technologies, Gaithersburg, MD) according to the manufacturer’s protocol. Total cell extracts were isolated 24 h after transfection, followed by Western blot analysis.

nology), CRB3 (Santa Cruz Biotechnology), Pals1 (Santa Cruz Biotechnology), PATJ (Santa Cruz Biotechnology), nectin-2 (Santa Cruz Biotechnology) and nectin-3 (Santa Cruz Biotechnology) and rabbit pAbs against claudin-6, claudin-7, ZO-1 (Zymed), ZO-2 (Zymed), CAR (Santa Cruz Biotechnology), aPKCE (E3AP1; [63]), aPKC~ (Upstate Biotechnology, Lake Placid, NY), PAR-3 (C2-3; [15]) and PAR-6 (BC32AP; [64]). Rabbit pAbs against CRB3 and Pals1 were kindly provided by Dr. B. Margolis [29,65], and rabbit anti-JAM-A pAb was a gift from Dr. T. Kita [66]. They were then rinsed with PBS and incubated for 1 h at room temperature with appropriate secondary antibodies labeled with Alexa Fluor 488 (green) or 594 (red) (Molecular Probes, Eugene, OR). All samples were examined using a laser-scanning confocal microscope (MRC 1024; Bio-Rad, Hercules, CA).

RNA extraction and reverse transcription (RT)-PCR

Gel electrophoresis and immunoblotting

For analysis of gene expression, total RNA was isolated from cells using TRIZOL reagent (Gibco BRL), and RTPCR was performed as previously described [53,54,60]. The PCR primers for mouse cDNAs were as follows: PAR-3 (GenBank accession no. AY026057), 5V-CAGACTCAAGGCAGGAGACC-3V (nucleotide [nt] 1618– 1637) and 5VGGGTGTGAGAACAACGTCCT-3V (nt 1830 – 1849); PAR-6 (GenBank accession no. AF070970), 5V-TGACAGTGACGATGACAGCA-3V (nt 918– 937) and 5V-AGAGGCTGAATCCGCTAACA-3V (nt 1110 –1129); aPKCE (Gene Bank accession no. D28577), 5V-TATGGCTTCAGCGTTGACTG-3V (nt 1294 –1313) and 5V-CCTTTGGGTCCTTGTTGAGA-3V (nt 1494 –1513). The primers for occludin, claudin-6, claudin-7, JAM-A, JAM-B, JAM-C, CAR (coxsackievirus and adenovirus receptor) and 36B4 were described previously [53,61,62]. To confirm that amplifications were in the linear range, PCR was performed for 3 different cycles between 16 and 30 cycles depending on the genes analyzed. Aliquots of PCR products were loaded onto 2% agarose gel and analyzed after staining with ethidium bromide.

Cells were grown on two 60-mm tissue culture plates, washed twice with ice-cold PBS and scraped with 300 Al of ice-cold NaHCO3 buffer (1 mM NaHCO3 and 1 mM PMSF, pH 7.5). They were then collected into a microcentrifuge tube, sonicated for 10 s and put for 30 min on ice. Total cell lysates were resolved by one-dimensional SDS – PAGE and electrophoretically transferred onto a PVDF membrane (Immobilon; Millipore, Bedford, MA). The membrane was saturated with PBS containing 4% skim milk and incubated for 1 h at room temperature with primary antibodies in either PBS or PBS containing 4% skim milk. The antibodies described above and a rabbit pAb against actin (Sigma) were used as primary antibodies. After rinsing in PBS containing 0.5% Tween 20, the membrane was incubated for 1 h at room temperature with horseradish peroxidase-conjugated anti-mouse, anti-rabbit or anti-goat IgG (diluted 1:1000; Vector Laboratories, Burlingame, CA) in PBS containing 0.3% skim milk. They were then rinsed again and finally reacted using an enhanced chemiluminescence Western blotting system (Amersham Pharmacia Biotech, Buckinghamshire, UK). Some blots were stripped with Restore Western blot stripping buffer (Pierce, Rockford, IL) and immunoprobed according to the manufacturer’s recommendations.

Immunohistochemistry Cells grown on coverslips were fixed in 1% formaldehyde in PBS for 10 min. After being washed 3 times with PBS, they were permeabilized with 0.2% Triton X-100 in PBS for 10 min, rinsed again with PBS, and then preincubated in PBS containing 5% skim milk. They were subsequently incubated for 1 h at room temperature with primary antibodies, including mouse monoclonal antibodies (mAbs) against occludin (Zymed, San Francisco, CA), Ecadherin (BD Biosciences, San Diego, CA) and nectin-1 (Santa Cruz Biotechnology), rat mAb against JAM-C (R&D systems), goat pAbs against JAM-A (Santa Cruz Biotechnology), JAM-B (R&D systems, Minneapolis, MN), ZO-3 (Santa Cruz Biotechnology), PAR-6 (Santa Cruz Biotech-

Results Activation of gene expression of three families of tight-junction transmembrane molecules in F9 cells by HNF-4a Since we previously showed that expression of occludin, claudin-6 and claudin-7 genes was markedly activated in F9 cell lines expressing Dox-inducible HNF-4a (F9 L32T2:HNF-4a clones 2 and 8; [54]), we first determined, by RT-PCR analysis, whether the gene expression of any

S. Satohisa et al. / Experimental Cell Research 310 (2005) 66 – 78

other tight-junction components was also induced in the cells. As shown in Fig. 1A, gene expression of not only occludin, claudin-6 and claudin-7 but also JAM-A started to be induced in F9 L32T2:HNF-4a clone 8 cells after 12 – 24 h of 1 Ag/ml Dox treatment, at which time the nuclear localization of HNF-4a was observed in most cells as previously described [54]. The levels of these transcripts were maximally increased after 48 h of 1 Ag/ml Dox exposure and elevated by Dox in a dose-dependent manner, in terms of the amount of HNF-4a (Figs. 1A and B; for HNF-4a expression, see [54,58]). Induced gene expression of occludin, claudin-6, claudin-7 and JAM-A was also detected in F9 L32T2:HNF-4a clone 2 (data not shown). In contrast, expression of JAM-B, JAM-C and CAR mRNAs was not altered in the cells after Dox treatment. In addition, gene expression of PAR-3, PAR-6 and aPKCE was not influenced in the cells by HNF-4a.

69

Fig. 2. Anti-claudin-6 and anti-claudin-7 antibodies selectively recognize the corresponding claudins. COS-7 cells were transfected with either empty or claudin-1 to claudin-8 expression vectors. Five micrograms of total cell lysate was separated by SDS – PAGE and immunoblotted with anti-claudin6 or anti-claudin-7 pAbs, followed by chemiluminescence detection. The transfected mouse claudin cDNAs are indicated at the top of each lane. The mobility of molecular mass markers (kDa) is indicated on the left. Arrowheads indicate specific signals. NS, non-specific signal.

Specificity of anti-claudin-6 and anti-claudin-7 antibodies We next generated novel pAbs against claudin-6 and claudin-7 because these claudins are most probably major constituents of tight-junction strands in differentiated F9 cells [53,54]. To check the selectivity of these antibodies,

COS-7 cells were transiently transfected with either individual expression vectors containing mouse claudin1 to claudin-8 cDNAs or the empty vector, and their lysates were subjected to gel electrophoresis and immunoblotting. As shown in Fig. 2, the anti-claudin-6 pAb reacted with claudin-6, but not with claudin-1 to claudin5, claudin-7 or claudin-8. Similarly, the anti-claudin-7 pAb specifically recognized claudin-7. Neither claudin-6 nor claudin-7 was detected in COS-7 cells transfected with the empty vector. Effect of HNF-4a on expression of tight-junction proteins in F9 cells

Fig. 1. Expression of occludin, claudin-6, claudin-7 and JAM-A genes is activated in F9 cells expressing doxycycline (Dox)-induced HNF-4a. (A) F9 L32T2:HNF-4a clone 8 cells were treated with 1 Ag/ml Dox for 0, 12, 24, and 48 h (lanes 1, 2, 3 and 4, respectively). One microgram of total RNA from the cells was subjected to RT-PCR analysis for the indicated genes. PCR was performed for 16 (36B4), 22 (claudin-6), 24 (JAM-A), 25 (claudin-7), 26 (JAM-B, PAR-3 and PAR-6), 28 (JAM-C and aPKCE), 29 (occludin) or 30 (CAR) cycles. (B) Cells were treated for 48 h with 0, 50, 200 and 1000 ng/ml (lanes 1, 2, 3, and 4, respectively) Dox. PCR was performed as in panel A. Similar results were obtained for at least two independent experiments.

We subsequently checked, by Western blot analysis, expression levels of various tight-junction proteins in F9 L32T2:HNF-4a clone 8 cells. Induction of occludin, claudin-6, claudin-7 and JAM-A expression was detectable in the cells after 24 h of 1 Ag/ml Dox treatment, and their induction was increased and maintained after 48 and 72 h of treatment, respectively (Fig. 3A; for expression of occludin and claudin-6, see also [54]). Expression of these proteins was also increased by Dox, thereby HNF-4a, in a dosedependent fashion (Fig. 3B), in good agreement with the levels of these transcripts. By contrast, protein levels of JAM-B, JAM-C and CAR, as well as those of cell polarity proteins PAR-3, PAR-6 and aPKCE, were not changed by Dox treatment (Figs. 3A and B). Similar results were obtained in the clone 2 cells (data not shown). Taken together with the data of RT-PCR analysis, these results indicated that HNF-4a activated the expression of three families of tight-junction adhesion molecules. Modulation of subcellular localization of tight-junction and adherens-junction proteins in F9 cells by HNF-4a We next determined, by immunostaining, differences in the expression of both tight-junction and adherens-

70

S. Satohisa et al. / Experimental Cell Research 310 (2005) 66 – 78

Fig. 3. Expression of occludin, claudin-6, claudin-7 and JAM-A proteins is up-regulated in F9 cells expressing doxycycline (Dox)-induced HNF-4a. (A) F9 L32T2:HNF-4a clone 8 cells were treated with the vehicle (lanes 1, 3, and 5) or 1 Ag/ml Dox (lanes 2, 4, and 6) for 24 (lanes 1 and 2), 48 (lanes 3, and 4) and 72 (lanes 5 and 6) h. Twenty-five micrograms of whole cell extract from the cells was separated by SDS – PAGE and immunoblotted with the corresponding antibodies, followed by chemiluminescence detection. (B) Cells were treated for 72 h with 0, 50, 200, and 1000 ng/ml (lanes 1, 2, 3, and 4, respectively) Dox. Western blot analysis was performed as in panel A. Similar results were obtained for at least three independent experiments.

junction proteins between undifferentiated and differentiated F9 L32T2:HNF-4a clone 8 cells. In undifferentiated cells grown in the absence of Dox, adherens-junction molecules E-cadherin, nectin-1, nectin-2 and nectin-3 were observed in a discontinuous pattern along cell borders (Fig. 4, and data not shown). Tight-junction proteins ZO1, ZO-2, ZO-3, JAM-B and JAM-C were also localized in the same pattern at cell –cell contacts of immature cells (Fig. 4, and data not shown). On the other hand, JAM-A and CAR were primarily expressed in a dotlike manner, probably in the cytoplasm, and no immunoreactive signals of occludin, claudin-6 or claudin-7 were detected in the undifferentiated cells. However, in differentiated cells treated for 48 – 72 h with 1 Ag/ml Dox, all these molecules except for JAM-A were concentrated along cell boundaries in a continuous junctional pattern (Fig. 4, and data not shown). Similar observations were obtained for clone 2 (data not shown). These findings indicated that HNF-4a promoted junction formation and epithelial polarization.

Distribution of tight-junction, adherens-junction and cell polarity proteins in undifferentiated F9 cells We further examined, by double immunostaining, the colocalization among tight-junction, adherens-junction and cell polarity proteins in undifferentiated F9 L32T2:HNF-4a clone 8 cells cultured without Dox exposure. Nectin-1, nectin-2 and nectin-3, as well as ZO-1, ZO-2 and ZO-3, were colocalized with E-cadherin in a discontinuous manner along lateral membranes of undifferentiated F9 cells (Fig. 5, and data not shown). JAM-B and JAM-C were also observed at the same position as E-cadherin. In contrast, JAM-A and CAR were hardly observed with E-cadherin (Fig. 5, arrowheads) but detected as dots in the cytoplasm, often beside lateral membranes (Fig. 5, arrows). Interestingly, each member of PAR-3/PAR-6/aPKCE and CRB3/Pals1/PATJ complexes revealed distinct subcellular localization in undifferentiated F9 cells (Fig. 6). PAR-3 and PATJ were precisely colocalized with Ecadherin in a discontinuous pattern on lateral membranes

S. Satohisa et al. / Experimental Cell Research 310 (2005) 66 – 78

71

Fig. 4. Doxycycline (Dox)-induced expression of HNF-4a modulates expression and/or subcellular localization of various components of junctional complexes in F9 cells. F9 L32T2:HNF-4a clone 8 cells were treated for 72 h with either the vehicle or 1 Ag/ml Dox. They were subjected to immunostaining with the corresponding antibodies. E-cad, E-cadherin. Scale bar, 20 Am.

of the cells. Similarly, PAR-6, aPKCE and aPKC~ were coexpressed with E-cadherin (Fig. 6, arrowheads, and data not shown), although their immunoreactive signals were also detected on apical membranes and throughout the cytoplasm. On the other hand, CRB3 was predominantly localized on apical membranes (Fig. 6) but also expressed with E-cadherin (arrowheads) and in the cytoplasm (arrows). Pals1 was mainly observed in a dotlike pattern inside the cytoplasm (Fig. 6, arrows), but additional signals

were detected at the same sites as E-cadherin (arrowheads) and on apical surfaces. Localization of tight-junction, adherens-junction and cell polarity proteins in differentiated F9 cells We subsequently examined, using occludin as a marker for tight junctions, whether tight-junction and cell polarity proteins were coexpressed in differentiated F9 L32T2:HNF-

72

S. Satohisa et al. / Experimental Cell Research 310 (2005) 66 – 78

Fig. 5. Nectin-3, ZO-3, JAM-B and JAM-C, but little JAM-A or CAR, are colocalized with E-cadherin along lateral membranes in untreated F9 cells. F9 L32T2:HNF-4a clone 8 cells were grown for 72 h in the absence of Dox. They were subjected to double immunostaining with the corresponding antibodies. X – Y and Z-section images are indicated. Arrows and arrowheads indicate cytoplasmic signals and faint colocalization with Ecadherin (E-cad), respectively. Scale bar, 20 Am.

Fig. 6. PAR-3, PAR-6, aPKCE, CRB3, Pals1 and PATJ are differentially localized in undifferentiated F9 cells. F9 L32T2:HNF-4a clone 8 cells were cultured for 72 h without Dox treatment. They were subjected to double immunostaining with the corresponding antibodies. X – Y and Z-section images are indicated. Arrows and arrowheads correspond to immunoreactivity in the cytoplasm and at the same position as E-cadherin (E-cad), respectively. Scale bar, 20 Am.

S. Satohisa et al. / Experimental Cell Research 310 (2005) 66 – 78

Fig. 7. Claudin-6, ZO-1, JAM-B, JAM-C and CAR, but not JAM-A, are colocalized with occludin at the apical-most tips of lateral membranes in F9 cells expressing doxycycline (Dox)-induced HNF-4a. F9 L32T2:HNF-4a clone 8 cells were treated for 72 h with 1 Ag/ml Dox. They were subjected to double immunostaining with the corresponding antibodies. X – Y and Zsection images are indicated. Arrows indicate signals in the cytoplasm. Ecadherin (E-cad), respectively. Scale bar, 20 Am.

73

Fig. 8. PAR-6, aPKCE, CRB3 and Pals1, but not PAR-3 or PATJ, are concentrated on apical membranes in F9 cells expressing doxycycline (Dox)-induced HNF-4a. F9 L32T2:HNF-4a clone 8 cells were exposed for 72 h to 1 Ag/ml Dox. They were subjected to double immunostaining with the corresponding antibodies. X – Y and Z-section images are indicated. Scale bar, 20 Am.

74

S. Satohisa et al. / Experimental Cell Research 310 (2005) 66 – 78

4a clone 8 cells treated for 48– 72 h with 1 Ag/ml Dox. As expected, JAM-B and JAM-C, as well as ZO-1, ZO-2 and ZO-3, all of which were concentrated to primordial adherens junctions in undifferentiated F9 cells, were sorted out from E-cadherin-based adherens junctions and precisely colocalized with occludin at the apical-most sites of lateral membranes of the differentiated cells to form beltlike tight junctions (Fig. 7, and data not shown). Similarly, claudin-6, claudin-7 and CAR, the last of which showed additional intracytoplasmic distribution, were observed at the same position as occludin (Fig. 7, and data not shown). By contrast, JAM-A was not primarily colocalized with occludin but detected as dots in the cytoplasm of differentiated F9 cells (Fig. 7, arrows), as seen in immature cells. PAR-3, PAR-6, aPKCE and PATJ, which accumulated to premature adherens junctions in undifferentiated F9 cells, were colocalized with occludin at the apical-most tips of lateral membranes in differentiated F9 cells (Fig. 8). Similarly, CRB3 and Pals1 were coexpressed with occludin in the differentiated cells. Thus, HNF-4a appeared to initiate the recruitment of various cell adhesion and cell polarity molecules to tight junctions. Interestingly, PAR-6, aPKCE, CRB3 and Pals1, but not PAR-3 or PATJ, were also concentrated on apical membranes of the cells, suggesting that these 4 polarity proteins could form a complex on apical domains in mature epithelial cells.

Discussion The aim of the present study was to investigate the molecular mechanism underlying junction formation and apicobasal polarization in mammalian epithelial cells. To this end, we employed mouse F9 L32T2:HNF-4a cells, which can be differentiated into highly polarized epithelial cells bearing well-developed junctional complexes by conditionally induced expression of HNF-4a [54,58]. We first determined, by RT-PCR and immunoblot analyses, whether the expression of any cell adhesion and cell polarity molecules was up-regulated in F9 cells by HNF-4a. Second, using confocal immunofluorescence microscopy, we examined differences in behavior of tight-junction, adherensjunction and cell polarity proteins between undifferentiated and differentiated F9 L32T2:HNF-4a cells and compared our findings with models proposed from studies of junction assembly and epithelial polarization during early Drosophila development and wound healing of mature mammalian epithelial cell lines. We also discussed putative roles of nuclear receptors in epithelial morphogenesis. Possible mechanisms by which HNF-4a initiates establishment of junctional complexes and epithelial polarity in F9 cells are depicted in Fig. 9. We previously reported that HNF-4a, like retinoid receptors, induced the gene expression of cell adhesion molecules of tight junctions, occludin, claudin-6 and claudin-7 in F9 cells [53,54]. In this respect, it is interesting

Fig. 9. Models of HNF-4a-triggered epithelial polarization and junction formation in F9 cells. (A) Intermediate polarity and primordial junction in undifferentiated F9 cells. In immature F9 cells, not only ZO-1 and ZO-2, but also ZO-3, JAM-B and JAM-C, as well as PAR-3, PAR-6, aPKC and PATJ, are accumulated to premature adherens junctions together with E-cadherin (E-cad) and nectin-1 to nectin-3. In contrast, CRB3 is primarily distributed on apical membranes, and Pals1, JAM-A and CAR are predominantly localized in the cytoplasm, in particular beside lateral membranes. (B) Complete epithelial polarity and junctional complexes in differentiated F9 cells. In differentiated F9 cells expressing HNF-4a, a variety of tight-junction and cell polarity proteins, except for JAM-A, are translocated to beltlike tight junctions. PAR-6, aPKC, CRB3 and Pals1 are also concentrated to apical surfaces. HNF-4a induces gene expression of occludin (OCLN), claudin-6 (CLDN6), claudin-7 (CLDN7) and JAM-A.

to note that, in cingulin-deficient mouse embryoid bodies, mRNA levels of the same tight-junction components, as well as those of HNF-4a and its upstream modulators GATA-4 and GATA-6, are increased [67], though the molecular basis for cingulin to regulate HNF-4a expression is unclear. In the present work we demonstrated that the gene expression of not only occludin, claudin-6 and claudin-7 but also JAM-A was activated in F9 cells by HNF-4a. Since HNF-4a is expressed in liver, kidney, intestine, pancreas and stomach, it might also up-regulate gene expression of the 4 tight-junction transmembrane molecules in these organs. On the other hand, HNF-4a did not modulate expression of JAM-B, JAM-C or CAR, nor that of tight-junction scaffold proteins ZO-1 and ZO-2 [54], adherens-junction proteins E-cadherin, h-catenin and nectin-1 to nectin-3 ([54] and our unpublished results), or cell

S. Satohisa et al. / Experimental Cell Research 310 (2005) 66 – 78

polarity proteins PAR-3, PAR-6, aPKC, CRB3, Pals1 and PATJ in F9 cells (present study and our unpublished results). Hence, we conclude that HNF-4a can induce the expression of three families of tight-junction transmembrane molecules, occludin, claudin and JAM, during epithelial polarization. Adherens junctions are thought to play key roles in establishing epithelial polarity [68]. At the early stage of epithelial polarization, as well as in non-epithelial cells such as fibroblasts and cardiac muscle cells, primordial adherens junctions are formed at cell – cell contacts, where tightjunction proteins ZO-1 and ZO-2 are known to be colocalized with cadherins and nectins [49,50,69– 71]. We herein showed that, in undifferentiated F9 cells, ZO-1 and ZO-2, and also ZO-3, which is expressed in epithelium but not in endothelium or fibroblasts [72], were recruited to premature adherens junctions where E-cadherin was precisely colocalized with nectin-1, nectin-2 and nectin-3. We also demonstrated that JAM-B and JAM-C, but little JAMA or CAR, were colocalized with E-cadherin along lateral membranes in immature F9 cells. Although JAM-A is reported to be colocalized with E-cadherin and ZO-1 in nascent adherens junctions [26,51], at least in undifferentiated F9 cells, JAM-B and JAM-C, instead of JAM-A, appear to accumulate at premature adherens junctions together with ZO-1, ZO-2 and ZO-3. Our immunofluorescence analysis also revealed that cell polarity proteins PAR-3, PAR-6 and aPKC were recruited to primordial adherens junctions in F9 cells, consistent with findings obtained from studies of epithelial morphogenesis in Drosophila and using a wound healing assay of MDT1-A cells [51,73]. This is reasonable because JAM-B, JAM-C, nectin-1 and nectin-3, which directly interact with PAR-3 [28,74], were concentrated at premature junctions in F9 cells. Note that PAR-6 and aPKC exhibited additional apical and intracellular distribution in immature F9 cells. Interestingly, CRB3, which is preferentially expressed in epithelial tissues [65], Pals1 and PATJ possessed distinct subcellular localization in undifferentiated F9 cells. PATJ, like PAR-3, was exactly colocalized with E-cadherin at early adherens junctions, possibly via ZO-3, an interaction partner of PATJ [29]. The junctional localization of PATJ might be also associated with that of JAM-C, since the distribution of PATJ shows remarkable overlap with JAM-C at the Sertoli cell – spermatid junction in mice and the PATJ pattern is significantly altered in jam-C / mice [75]. In contrast, CRB3 and Pals1 were predominantly observed on apical membranes and the cytoplasm of undifferentiated F9 cells, respectively. Apical localization of CRB3 could be attributed to PAR-3, because its Drosophila orthologue Bazooka acts upstream of CRB to form the apical membrane [73,76]. CRB3 was also detected at early adherens junctions and inside the cytoplasm, whereas Pals1 showed additional distribution at primordial junctions and apical surfaces. Collectively, our results indicate that intermediate cell polarity, in terms of assembly of primordial junctions and

75

some identity of apical membranes, is formed in undifferentiated F9 cells (Fig. 9A). As expected, a variety of tight-junction proteins, ZO-1, ZO-2, ZO-3, JAM-B and JAM-C, all of which were localized at early adherens junctions in undifferentiated F9 cells, were sorted into occludin-positive beltlike tight junctions in differentiated F9 cells. Although JAM-C is localized at desmosomes but not tight junctions in human intestinal epithelial cells [77], our results revealed that both JAM-B and JAM-C were components of tight junctions in F9 epithelial cells expressing HNF-4a. CAR, claudin-6 and claudin-7, which were not concentrated to primordial junctions in immature F9 cells, also accumulated to tight junctions at the same position as occludin. In addition, we demonstrated that not only PAR-3, PAR-6, aPKC and PATJ localized primarily at premature junctions in F9 cells, but also CRB3 and Pals1, all of which are cooperatively involved in the maturation of nascent junctional complex into distinct adherens junctions and tight junctions as well as establishment of epithelial apicobasal polarity [5,19 – 21,25], were translocated to beltlike tight junctions. Thus, numerous tight-junction and cell polarity proteins, except for JAM-A, were recruited to tight junctions in HNF-4a-expressing F9 cells as far as we could determine (Fig. 9B). Interestingly, PAR-6, aPKC, CRB3 and Pals1 were concentrated not only to tight junctions with PAR-3 and PATJ, but also on the apical surface in the absence of PAR-3 or PATJ in F9 epithelial cells. Taken together with that the apical determinant CRB3 directly binds to PAR-6 and aPKC [31,33] and that the CRB3-binding partner Pals1 interacts with PAR-6 [30,32], these findings raise the interesting possibility that these four polarity proteins, but not PAR-3 or PATJ, form a complex on the apical domain in mature epithelial cells (Fig. 9B). Nuclear receptors are well known to play important roles in development and cell differentiation, but their molecular mechanisms remain incompletely defined. Taken collectively with the fact that the formation of the cell polarity and junctional complex is critical for differentiated epithelial cells, we should reevaluate the potent functions of nuclear receptors in epithelial cell polarization and junction formation. First, 1a,25(OH)2D3 induces expression of Ecadherin and occludin, as well as establishment of epithelial morphology, in human colon carcinoma cell line SW480 expressing vitamin D receptors [78]. Second, glucocorticoids, ligands for glucocorticoid receptors (GRs), accelerate organization of the apical junctional complex and cell polarity in rat intestinal and mammary epithelial cell lines ICEs and Con8 ([79,80] and references therein). Third, forced expression of HNF-4a in dedifferentiated hepatoma cells results in partial reexpression of E-cadherin, and dexamethasone, a synthetic ligand for GRs, initiates the establishment of single epithelial polarity in HNF-4atransfectants [43]. Fourth, chenodeoxycholic acid (CDCA), a ligand of the farnesoid X receptor/bile acid receptor (FXR/BAR) [81], cues cell polarization, namely the

76

S. Satohisa et al. / Experimental Cell Research 310 (2005) 66 – 78

appearance of well-organized microvilli and tight junctions, in rat hepatoma cell lines Fao and C2rev7 [82]. Fifth, in mammary glands of pregnant and lactating ERh / (estrogen receptor h / ) mice, weak and no immunoreactive signals of E-cadherin and occludin are detected, respectively, compared with those in wild-type mice [83]. Although the overall cell polarity is not altered in the mutant mammary glands, the interepithelial cell space is frequently increased, and the terminal differentiation, in terms of development of ductal side branching and alveoli, is impaired. It should also be mentioned that the ER signaling could activate the expression of MTA3, which is a component of Mi2/NuRD (nuclear remodeling and deacetylation) complex and suppress the expression of Snail, a master regulator of epithelial to mesenchymal transitions, leading to up-regulation of E-cadherin expression [84]. Taking all these observations into consideration, our findings strongly suggest that nuclear receptors expressed in epithelial cells are at least in part involved in induction of expression of cell adhesion molecules and acquisition of epithelial cell polarity and junctional complexes, thereby promoting epithelial differentiation. In summary, our results have indicated that HNF-4a activates the expression of three families of tight-junction adhesion molecules, occludin, claudin-6, claudin-7 and JAM-A, in F9 cells. We have also demonstrated that tightjunction proteins ZO-1, ZO-2, ZO-3, JAM-B and JAM-C, cell polarity proteins PAR-3, PAR-6, aPKC and PATJ, as well as some fraction of CRB3 and Pals1, but little JAM-A or CAR, are recruited to primordial adherens junctions in undifferentiated F9 cells possessing intermediate cell polarity. Moreover, HNF-4a apparently triggers sorting of the numerous molecules described above to beltlike tight junctions during epithelial polarization of the cells. Our observations have also suggested that PAR-6, aPKC, CRB3 and Pals1, but not PAR-3 or PATJ, form a complex on the apical surface in completely polarized epithelial cells. Since a wide variety of adherens-junction, tight-junction and cell polarity proteins can be expressed and localized at nascent and/or mature junctions in F9 cells, the F9 L32T2 [52,53] and F9 L32T2:HNF-4a [54,58] cell lines will indeed be very powerful tools to investigate functions of any gene products of interest in establishment of both junctional complexes and epithelial polarity.

Acknowledgments This work was supported by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan. We are grateful to Dr. T. Kita for the gift of the anti-JAM-A antibody; Dr. B. Margolis for the anti-CRB3 and anti-Pals1 antibodies; Immuno-Biological Laboratories for cooperation in generation of anti-claudin-6 and anti-claudin-7 antibodies; and Mr. K. Barrymore for help with the manuscript.

References [1] J.M. Anderson, M. Cereijido, Introduction: evolution of ideas on the tight junction, in: M. Cereijido, J.M. Anderson (Eds.), Tight Junctions, CRC Press, Boca Raton, 2001, pp. 1 – 18. [2] S. Tsukita, M. Furuse, M. Itoh, Multifunctional strands in tight junctions, Nat. Rev., Mol. Cell Biol. 2 (2001) 285 – 293. [3] A. Zahraoui, D. Louvard, T. Galli, Tight junction, a platform for trafficking and signaling protein complexes, J. Cell Biol. 151 (2000) F31 – F36. [4] K. Matter, M.S. Balda, Signalling to and from tight junctions, Nat. Rev., Mol. Cell Biol. 4 (2003) 225 – 236. [5] E.E. Schneeberger, R.D. Lynch, The tight junction: a multifunctional complex, Am. J. Physiol.: Cell Physiol. 286 (2004) C1213 – C1228. [6] M. Furuse, T. Hirase, M. Itoh, A. Nagafuchi, S. Yonemura, S. Tsukita, S. Tsukita, Occludin: a novel integral membrane protein localizing at tight junctions, J. Cell Biol. 123 (1993) 1777 – 1788. [7] M. Furuse, K. Fujita, T. Hiiragi, K. Fujimoto, S. Tsukita, Claudin-1 and -2: novel integral membrane proteins localizing at tight junctions with no sequence similarity to occludin, J. Cell Biol. 141 (1998) 1539 – 1550. [8] I. Martı`n-Padura, S. Lostaglio, M. Schneemann, L. Williams, M. Romano, P. Fruscella, C. Panzeri, A. Stoppacciaro, L. Ruco, A. Villa, D. Simmons, E. Dejana, Junctional adhesion molecule, a novel member of the immunoglobulin superfamily that distributes at intercellular junctions and modulates monocyte transmigration, J. Cell Biol. 142 (1998) 117 – 127. [9] K. Morita, M. Furuse, K. Fujimoto, S. Tsukita, Claudin multigene family encoding four-transmembrane domain protein components of tight junction strands, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 511 – 516. [10] C.J. Cohen, J.T.C. Shieh, R.J. Pickles, T. Okegawa, J.T. Hsieh, J.M. Bergelson, The coxsackievirus and adenovirus receptor is a transmembrane component of the tight junction, Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 15191 – 15196. [11] G. Bazzoni, The JAM family of junctional adhesion molecules, Curr. Opin. Cell Biol. 15 (2003) 525 – 530. [12] K. Ebnet, A. Suzuki, S. Ohno, D. Vestweber, Junctional adhesion molecules (JAMs): more molecules with dual functions? J. Cell Sci. 117 (2004) 19 – 29. [13] M. Furuse, H. Sasaki, K. Fujimoto, S. Tsukita, A single gene product, claudin-1 or -2, reconstitutes tight junction strands and recruits occludin in fibroblasts, J. Cell Biol. 143 (1998) 391 – 401. [14] K. Turksen, T.C. Troy, Barriers built on claudins, J. Cell Sci. 117 (2004) 2435 – 2447. [15] Y. Izumi, T. Hirose, Y. Tamai, S. Hirai, Y. Nagashima, T. Fujimoto, Y. Tabuse, K.J. Kemphues, S. Ohono, An atypical PKC directly associates and colocalizes at the epithelial tight junction with ASIP, a mammalian homologue of Caenorhabditis elegans polarity protein PAR-3, J. Cell Biol. 143 (1998) 95 – 106. [16] G. Joberty, C. Petersen, L. Gao, I.G. Macara, The cell-polarity protein PAR-6 links PAR-3 and atypical protein kinase C to Cdc42, Nat. Cell Biol. 2 (2000) 531 – 539. [17] D. Lin, A.S. Edwards, J.P. Fawcett, G. Mbamalu, J.D. Scott, T. Pawson, A mammalian PAR-3-PAR-6 complex implicated in Cdc42/Rac1 and aPKC signalling and cell polarity, Nat. Cell Biol. 2 (2000) 540 – 547. [18] R.G. Qiu, A. Abo, G. Steven Martin, A human homolog of the C. elegans polarity determinant PAR-6 links Rac and Cdc42 to PKCzeta signaling and cell transformation, Curr. Biol. 10 (2000) 697 – 707. [19] S. Ohno, Intercellular junctions and cellular polarity: the PAR-aPKC complex, a conserved core cassette playing fundamental roles in cell polarity, Curr. Opin. Cell Biol. 13 (2001) 641 – 648. [20] A. Suzuki, T. Yamanaka, T. Hirose, N. Manabe, K. Mizuno, M. Shimizu, K. Akimoto, Y. Izumi, T. Ohnishi, S. Ohno, Atypical protein kinase C is involved in the evolutionarily conserved PAR protein complex and plays a critical role in establishing epithelia-specific junctional structures, J. Cell Biol. 152 (2001) 1183 – 1196.

S. Satohisa et al. / Experimental Cell Research 310 (2005) 66 – 78 [21] E. Knust, O. Bossinger, Composition and formation of intercellular junctions in epithelial cells, Science 298 (2002) 1955 – 1959. [22] C. Lemmers, E. Me´dina, M.H. Delgrossi, D. Michel, J.P. Arsanto, A. Le Bivic, hINADl/PATJ, a homolog of Discs lost, interacts with Crumbs and localizes to tight junctions in human epithelial cells, J. Biol. Chem. 277 (2002) 25408 – 25415. [23] M.H. Roh, O. Makarova, C.J. Liu, K. Shin, S. Lee, S. Laurinec, M. Goyal, R. Wiggins, B. Margolis, The Maguk protein, Pals1, functions as an adapter, linking mammalian homologues of Crumbs and Discs Lost, J. Cell Biol. 157 (2002) 161 – 172. [24] M.H. Roh, S. Fan, C.J. Liu, B. Margolis, The Crumbs3-Pals1 complex participates in the establishment of polarity in mammalian epithelial cells, J. Cell Sci. 116 (2003) 2895 – 2906. [25] M.H. Roh, B. Margolis, Composition and function of PDZ protein complexes during cell polarization, Am. J. Physiol.: Renal Physiol. 285 (2003) F377 – F387. [26] K. Ebnet, A. Suzuki, Y. Horikoshi, T. Hirose, M.K. Meyer zu Brickwedde, S. Ohno, D. Vestweber, The cell polarity protein ASIP/PAR-3 directly associates with junctional adhesion molecule (JAM), EMBO J. 20 (2001) 3738 – 3748. [27] M. Itoh, H. Sasaki, M. Furuse, H. Ozaki, T. Kita, S. Tsukita, Junctional adhesion molecule (JAM) binds to PAR-3: a possible mechanism for the recruitment of PAR-3 to tight junctions, J. Cell Biol. 154 (2001) 491 – 497. [28] K. Ebnet, M. Aurrand-Lions, A. Kuhn, F. Kiefer, S. Butz, K. Zander, M.K. Meyer zu Brickwedde, B.A. Imhof, A. Suzuki, D. Vestweber, The junctional adhesion molecule (JAM) family members JAM-2 and JAM-3 associate with the cell polarity protein PAR3: a possible role for JAMs in endothelial cell polarity, J. Cell Sci. 116 (2003) 3738 – 3748 (3879-3891). [29] M.H. Roh, C.J. Liu, S. Laurinec, B. Margolis, The carboxyl terminus of zona occludens-3 binds and recruits a mammalian homologue of Discs lost to tight junctions, J. Biol. Chem. 277 (2002) 27501 – 27509. [30] T.W. Hurd, L. Gao, M.H. Roh, I.G. Macara, B. Margolis, Direct interaction of two polarity complexes implicated in epithelial tight junction assembly, Nat. Cell Biol. 5 (2003) 137 – 142. [31] C. Lemmers, D. Michel, L. Lane-Guermonprez, M.H. Delgrossi, E. Me´dina, J.P. Arsanto, A. Le Bivic, CRB3 binds directly to PAR-6 and regulates the morphogenesis of the tight junctions in mammalian epithelial cells, Mol. Cell. Biol. 15 (2004) 1324 – 1333. [32] R.R. Penkert, H.M. Divittorio, K.E. Prehoda, Internal recognition through PDZ domain plasticity in the PAR-6-Pals1 complex, Nat. Struct. Mol. Biol. 11 (2004) 1122 – 1127. [33] S. Sotillos, M.T. Dı´az-Meco, E. Camero, J. Moscat, S. Campuzano, DaPKC-dependent phosphorylation of Crumbs is required for epithelial cell polarity in Drosophila, J. Cell Biol. 166 (2004) 549 – 557. [34] F.M. Sladek, W. Zhong, E. Lai, J.E. Darnell Jr., Liver-enriched transcription factor HNF-4 is a novel member of the steroid hormone receptor superfamily, Genes Dev. 4 (1990) 2353 – 2365. [35] G.P. Hayhurst, Y.H. Lee, G. Lambert, J.M. Ward, F.J. Gonzalez, Hepatocyte nuclear factor 4a (nuclear receptor 2A1) is essential for maintenance of hepatic gene expression and lipid homeostasis, Mol. Cell. Biol. 21 (2001) 1393 – 1403. [36] F.M. Sladek, S.D. Seidel, Hepatocyte nuclear factor 4a, in: T.P. Burris, E.R.B. McCabe (Eds.), Nuclear Receptors and Genetic Disease, Academic Press, San Diego, 2001, pp. 309 – 361. [37] R.G. Tirona, W. Lee, B.F. Leake, L.B. Lan, C.B. Cline, V. Lamda, F. Parviz, S.A. Duncan, Y. Inoue, F.J. Gonzalez, E.G. Schuetz, R.B. Kim, The orphan nuclear receptor HNF4a determines PXR- and CAR-mediated xenobiotic induction of CYP3A4, Nat. Med. 9 (2003) 220 – 224. [38] A.J. Watt, W.D. Garrison, S.A. Duncan, HNF4: a central regulator of hepatocyte differentiation and function, Hepatology 37 (2003) 1349 – 1352. [39] R. Hertz, J. Magenheim, I. Berman, J. Bar-tana, Fatty acyl-CoA thioesters are ligands of hepatic nuclear factor-4a, Nature 392 (1998) 512 – 516.

77

[40] F.M. Sladek, Desperately seeking—Something, Mol. Cell 10 (2002) 219 – 221. [41] S.A. Duncan, K. Manova, W.S. Chen, P. Hoodless, D.C. Weinstein, R.F. Bachvarova, J.E. Darnell Jr., Expression of transcription factor HNF-4 in the extraembryonic endoderm, gut, and nephrogenic tissue of the developing mouse embryo: HNF-4 is a marker for primary endoderm in the implanting blastocyst, Proc. Natl. Acad. Sci. U. S. A. 91 (1994) 7598 – 7602. [42] W.S. Chen, K. Manova, D.C. Weinstein, S.A. Duncan, A.S. Plump, V.R. Prezioso, R.F. Bachvarova, J.E. Darnell Jr., Disruption of the HNF-4 gene, expressed in visceral endoderm, leads to cell death in embryonic ectoderm and impaired gastrulation of mouse embryos, Genes Dev. 8 (1994) 2466 – 2477. [43] G.F. Spa¨th, M.C. Weiss, Hepatocyte nuclear factor 4 provokes expression of epithelial marker genes, acting as a morphogen in dedifferentiated hepatoma cells, J. Cell Biol. 140 (1998) 935 – 946. [44] J. Li, G. Ning, S.A. Duncan, Mammalian hepatocyte differentiation requires the transcription factor HNF-4a, Genes Dev. 14 (2000) 464 – 474. [45] F. Parviz, C. Matullo, W.D. Garrison, L. Savatski, J.W. Adamson, G. Ning, K.H. Kaestner, J.M. Rossi, K.S. Zaret, S.A. Duncan, Hepatocyte nuclear factor 4a controls the development of a hepatic epithelium and liver morphogenesis, Nat. Genet. 34 (2003) 292 – 296. [46] B.L. Hogan, A. Taylor, E. Adamson, Cell interactions modulate embryonal carcinoma cell differentiation into parietal or visceral endoderm, Nature 291 (1981) 235 – 237. [47] S. Strickland, Mouse teratocarcinoma cells: prospects for the study of embryogenesis and neoplasia, Cell 24 (1981) 277 – 278. [48] L. Shoshani, R.G. Contreras, Biogenesis of epithelial polarity and tight junctions, in: M. Cereijido, J.M. Anderson (Eds.), Tight Junctions, CRC Press, Boca Raton, 2001, pp. 165 – 197. [49] S. Yonemura, M. Itoh, A. Nagafuchi, S. Tsukita, Cell-to-cell adherens junction formation and actin filament organization: similarities and differences between non-polarized fibroblasts and polarized epithelial cells, J. Cell Sci. 108 (1995) 127 – 142. [50] Y. Ando-Akatsuka, S. Yonemura, M. Itoh, M. Furuse, S. Tsukita, Differential behavior of E-cadherin and occludin in their colocalization with ZO-1 during the establishment of epithelial cell polarity, J. Cell Physiol. 179 (1999) 115 – 125. [51] A. Suzuki, C. Ishiyama, K. Hashiba, M. Shimizu, K. Ebnet, S. Ohno, aPKC kinase activity is required for the asymmetric differentiation of the premature junctional complex during epithelial cell polarization, J. Cell Sci. 115 (2002) 3565 – 3573. [52] H. Chiba, D. Metzger, P. Chambon, F9 embryonal carcinoma cells engineered for tamoxifen-dependent Cre-mediated site-directed mutagenesis and doxycycline-inducible gene expression, Exp. Cell. Res. 260 (2000) 334 – 339. [53] H. Kubota, H. Chiba, Y. Takakuwa, M. Osanai, H. Tobioka, G. Kohama, M. Mori, N. Sawada, Retinoid X receptor a and retinoic acid receptor g mediate expression of genes encoding tight junction proteins and barrier function in F9 cells during visceral endodermal differentiation, Exp. Cell Res. 263 (2001) 163 – 172. [54] H. Chiba, T. Gotoh, T. Kojima, S. Satohisa, K. Kikuchi, M. Osanai, N. Sawada, Hepatocyte nuclear factor (HNF)-4a triggers formation of functional tight junctions and establishment of polarized epithelial morphology in F9 embryonal carcinoma cells, Exp. Cell Res. 286 (2003) 288 – 297. [55] M. Gossen, H. Bujard, Tight control of gene expression in mammalian cells by tetracycline-responsive promoters, Proc. Natl. Acad. Sci. U. S. A. 89 (1992) 5547 – 5571. [56] D. Metzger, P. Chambon, Site- and time-specific gene targeting in the mouse, Methods 24 (2001) 71 – 80. [57] H. Chiba, J. Clifford, D. Metzger, P. Chambon, Specific and redundant functions of retinoid X receptor/retinoic acid receptor heterodimers in differentiation, proliferation, and apoptosis of F9 embryonal carcinoma cells, J. Cell Biol. 139 (1997) 735 – 747.

78

S. Satohisa et al. / Experimental Cell Research 310 (2005) 66 – 78

[58] H. Chiba, T. Itoh, S. Satohisa, N. Sakai, H. Noguchi, M. Osanai, T. Kojima, N. Sawada, Activation of p21CIP1/WAF1 gene expression and inhibition of cell proliferation by overexpression of hepatocyte nuclear factor-4a, Exp. Cell Res. 302 (2005) 11 – 21. [59] T. Ishizaki, H. Chiba, T. Kojima, M. Fujibe, T. Soma, H. Miyajima, K. Nagasawa, I. Wada, N. Sawada, Cyclic AMP induces phosphorylation of claudin-5 immunoprecipitates and expression of claudin-5 gene in blood – brain barrier endothelial cells via protein kinase A-dependent and -independent pathways, Exp. Cell Res. 290 (2003) 275 – 288. [60] H. Chiba, J. Clifford, D. Metzger, P. Chambon, Distinct retinoid X receptor-retinoic acid receptor heterodimers are differentially involved in the control of expression of retinoid target genes in F9 embryonal carcinoma cells, Mol. Cell. Biol. 17 (1997) 3013 – 3020. [61] M. Aurrand-Lions, C. Johnson-Leger, C. Wong, L. Du pasquier, B.A. Imhof, Heterogeneity of endothelial junctions is reflected by differential expression and specific subcellular localization of the three JAM family members, Blood 98 (2001) 3699 – 3707. [62] T. Ito, K. Tokunaga, H. Maruyama, H. Kawashima, H. Kitahara, T. Horikoshi, A. Ogose, Y. Hotta, R. Kuwano, H. Katagiri, N. Endo, Coxsackievirus and adenovirus receptor (CAR)-positive immature osteoblasts as targets of adenovirus-mediated gene transfer for fracture healing, Gene Ther. 10 (2003) 1623 – 1628. [63] K. Akimoto, M. Nakaya, T. Yamanaka, J. Tanaka, S. Matsuda, Q.P. Weng, J. Arvuch, S. Ohno, Atypical protein kinase C E binds and regulates p70 S6 kinase, Biochem. J. 335 (1998) 417 – 424. [64] T. Yamanaka, Y. Horikoshi, Y. Sugiyama, C. Ishiyama, A. Suzuki, T. Hirose, A. Iwamatsu, A. Shinohara, S. Ohno, Mammalian Lgl forms a protein complex with PAR-6 and aPKC independently of PAR-3 to regulate epithelial cell polarity, Curr. Biol. 13 (2003) 734 – 743. [65] O. Makarova, M.H. Roh, C.J. Liu, S. Laurinec, B. Margolis, Mammalian Crumbs3 is a small transmembrane protein linked to protein associated with Lin-7 (Pals1), Gene 302 (2003) 21 – 29. [66] H. Ozaki, K. Kishi, H. Horiuchi, H. Arai, T. Kawamoto, K. Okawa, A. Iwamatsu, T. Kita, Cutting edge: combined treatment of TNF-a and IFN-g causes redistribution of junctional adhesion molecule in human endothelial cells, J. Immunol. 163 (1999) 553 – 557. [67] L. Guillemot, E. Hammar, C. Kaister, J. Ritz, D. Caille, L. Jond, C. Bauer, P. Meda, S. Citi, Disruption of the cingulin gene does not prevent tight junction formation but alters gene expression, J. Cell Sci. 117 (2004) 5245 – 5256. [68] W.J. Nelson, Adaptation of core mechanisms to generate cell polarity, Nature 422 (2003) 766 – 774. [69] M. Itoh, A. Nagafuchi, S. Yonemura, T. Kitani-Yasuda, S. Tsukita, S. Tsukita, The 220-kD protein colocalizing with cadherins in nonepithelial cells is identical to ZO-1, a tight junction-associated protein in epithelial cells: cDNA cloning and immunoelectron microscopy, J. Cell Biol. 121 (1993) 491 – 502. [70] K. Tachibana, H. Nakanishi, K. Mandai, K. Ozaki, W. Ikeda, Y. Yamamoto, A. Nagafuchi, S. Tsukita, Y. Takai, Two cell adhesion

[71] [72]

[73]

[74]

[75]

[76]

[77]

[78]

[79]

[80]

[81] [82]

[83]

[84]

molecules, nectin and cadherin, interact through their cytoplasmic domain-associated proteins, J. Cell Biol. 150 (2000) 1161 – 1175. Y. Takai, H. Nakanishi, Nectin and afadin: novel organizers of intercellular junctions, J. Cell Sci. 116 (2003) 17 – 27. A. Inoko, M. Itoh, A. Tamura, M. Matsuda, M. Furuse, S. Tsukita, Expression and distribution of ZO-3, a tight junction MAGUK protein, in mouse tissues, Genes Cells 8 (2003) 837 – 845. D. Bilder, M. Schober, N. Perrimon, Integrated activity of PDZ protein complexes regulates epithelial polarity, Nat. Cell Biol. 5 (2003) 53 – 58. K. Takekuni, W. Ikeda, T. Fujito, K. Morimoto, M. Takeuchi, M. Monden, Y. Takai, Direct binding of cell polarity protein PAR-3 to cell – cell adhesion molecule nectin at neuroepithelial cells of developing mouse, J. Biol. Chem. 278 (2003) 5497 – 5500. G. Giki, K. Ebnet, M. Aurrand-Lions, B.A. Imhof, R.H. Adams, Spermatid differentiation requires the assembly of a cell polarity complex downstream of junctional adhesion molecule-C, Nature 431 (2004) 320 – 324. T.J. Harris, M. Peifer, Adherens junction-dependent and -independent steps in the establishment of epithelial cell polarity in Drosophila, J. Cell Biol. 167 (2004) 135 – 147. K. Zen, B.A. Babbin, Y. Liu, J.B. Whelan, A. Nusrat, C.A. Parkos, JAM-C is a component of desmosomes and a ligand for CD11b/CD18-mediated neutrophil transepithelial migration, Mol. Biol. Cell 15 (2004) 3926 – 3937. H.G. Pa´lmer, J.M. Gonza´lez-Sancho, J. Espada, M.T. Berciano, I. Puig, J. Baulida, M. Quintanilla, A. Cano, A.G. de Herreros, M. Lafarga, A. Munoz, Vitamin D3 promotes the differentiation of colon carcinoma cells by the induction of E-cadherin and the inhibition of hcatenin signaling, J. Cell Biol. 154 (2001) 369 – 387. A. Quaroni, J.Q. Tian, M. Go¨ke, D.K. Podolsky, Glucocorticoids have pleiotropic effects on small intestinal crypt cells, Am. J. Physiol.: Gastrointest. Liver Physiol. 277 (1999) G1027 – G1040. N.M. Rubenstein, Y. Guan, P.L. Woo, G.L. Firestone, Glucocorticoid down-regulation of RhoA is required for the steroid-induced organization of the junctional complex and tight junction formation in rat mammary epithelial tumor cells, J. Biol. Chem. 278 (2003) 10353 – 10360. J.A. Gustafsson, Seeking ligands for lonely orphan receptors, Science 284 (1999) 1285 – 1286. K.H. Ng, C. Le Goascogne, E. Amborade, B. Stieger, J. Deschatrette, Reversible induction of rat hepatoma cell polarity with bile acids, J. Cell Sci. 113 (2000) 4241 – 4251. C. Fo¨rster, S. Ma¨kela, A. Wa¨rri, S. Kietz, D. Becker, K. Hultenby, M. Warner, J.A. Gustafsson, Involvement of estrogen receptor h in terminal differentiation of mammary gland epithelium, Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 15578 – 15583. N. Fujita, D.L. Jaye, M. Kajita, C. Geigerman, C.S. Moreno, P.A. Wade, MTA3, a Mi-2/NuRD complex subunit, regulates an invasive growth pathway in breast cancer, Cell 113 (2003) 207 – 219.