| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | ADVANCED SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Division of Cellular Interactions, Institute of Molecular Embryology and Genetics, Kumamoto University, Kumamoto 860-0811, Japan
2 Department of Pediatrics, Faculty of Medical and Pharmaceutical Sciences, Kumamoto University, Kumamoto 860-8556, Japan
3 Department of Cell Biology, Faculty of Medicine, Kyoto University, Kyoto 606-8501, Japan
4 Laboratory for Cellular Morphogenesis, RIKEN Center for Developmental Biology, Kobe, Hyogo 650-0047, Japan
5 KAN Research Institute, Inc., Shimogyo-ku, Kyoto 600-8815, Japan
 |
Abstract |
|---|
 Top Abstract IntroductionResultsDiscussionExperimental proceduresReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionResultsDiscussionExperimental proceduresReferences |
|---|
The apical membrane is an epithelial cell-specific domain that possesses unique features. This membrane domain is separated from the basolateral domain by the junctional complex in epithelial cells (for review, see Rodriguez-Boulan & Nelson 1989). Epithelial cells segregate their plasma membrane proteins into apical and basolateral domains to perform vectorial functions. The segregation process involves polarized exocytosis, which is probably regulated by complex vesicle transport machinery (Keller et al. 2001). Syntaxin 3, a t-SNARE, is a component of the plasma membrane fusion machinery and is localized to apical plasma membranes in polarized MDCK cells (Kreitzer et al. 2003). It has been reported that anti-syntaxin 3 antibodies or over-expression of syntaxin 3 specifically inhibited or decreased apical transport (Low et al. 1998). Another special feature of the apical membrane region is the development of microvilli. Intestinal epithelial cells, in particular, possess well-developed microvilli called brush borders. In most simple epithelia, however, more sparse microvilli were observed on the apical cell surface. Ezrin is a major constituent of the intestinal brush border and is commonly localized at the apical membrane with microvilli (Berryman et al. 1993). This protein is a member of the ERM protein family, which links membrane proteins to the actin cytoskeleton (for review, see Tsukita & Yonemura 1999). The loss of ezrin expression causes disorganization of mouse intestinal epithelium (Saotome et al. 2004). Thus, syntaxin 3 and ezrin play a crucial function in apical formation or maintenance, and are frequently used as markers of the apical membrane domain.
The molecular mechanisms pertaining to junctional complex and apical membrane formation have been extensively investigated using epithelial cell lines such as MDCK cells. The use of differentiated epithelial cell lines, however, was unsuitable in elucidating the initial processes involving junctional complex and apical membrane formation during simple epithelium differentiation. For this purpose, nonepithelial cells, the epithelial differentiation of which could be induced under monolayer culture conditions, would be needed. F9 is a mouse cell line derived from teratocarcinoma in 129 strains (Sherman & Miller 1978). It was reported that when F9 cell aggregates were cultured in the presence of retinoic acid (RA), outer-layer cells could differentiate into visceral endoderm (VE)-like epithelial cells (Hogan et al. 1981). Electron microscopy analysis confirmed the development of apical microvilli and junctional complex formation in differentiated outer-layer cells. Recent DNA chip analysis demonstrated that the expression of several epithelial genes including keratin 8, keratin 18, and claudin-6 was induced by RA (Harris & Childs 2002; Shibamoto et al. 2004). Thus, the F9 cell system has been successfully used as a viable in vitro model of epithelial differentiation. Although immunohistochemical analysis demonstrated that DS and TJ could form under monolayer culture conditions (Trevor & Steben 1992; Chiba et al. 2003), it has been reported that aggregate formation is necessary for the VE differentiation of F9 cells (Casanova & Grabel 1988). Aggregate culture conditions prevented this system from being used for a detailed analysis of epithelial differentiation processes.
In the present study, we investigated epithelial specific gene expression, and the expression and subcellular localization of junctional complex components and apical membrane markers in monolayer cultured F9 cells. The morphology of monolayer cultured F9 cells in the presence of RA was also investigated by the use of electron microscopy. Our data indicated that monolayer cultured F9 cells could differentiate into epithelial cells in the presence of RA. Using this system, we could demonstrate that the junctional complex and apical membrane were formed under strict regulation during epithelial differentiation.
|
Results |
|---|
|
TopAbstractIntroductionResultsDiscussionExperimental proceduresReferences |
|---|
Epithelial differentiation of F9 cells is usually observed when cell aggregates are cultured in the presence of RA (Casanova & Grabel 1988). In an effort to determine whether F9 cells could differentiate even in monolayer culture, we performed comparative gene expression analyses of F9 cells grown under aggregate and monolayer culture conditions prior to and following treatment with RA. To analyze the initial state associated with RA induction, we compared F9 cells cultured for 3 days in the absence or presence of RA. RNA isolated from the cells was used to prepare probes for oligonucleotide array analysis using the MU74Av2 Affymetrix chip. In the comparison that followed, we considered only those genes whose signals were greater than 10 and were judged as present according to the absolute analysis algorithm. When the gene expression signals obtained from an RA-induced aggregate sample were compared to those obtained from a control aggregate sample, 147 genes were induced fivefold or more. More than 90% of these genes were also induced 2.5-fold or greater under monolayer culture conditions following treatment with RA. Some of these genes, including GATA4, GATA6, SOX17, type IV collagen, laminin B1, keratin 8, and keratin 18, have previously been reported to be induced by RA in F9 cells (Grover et al. 1983; Rogers et al. 1990; Arceci et al. 1993; Shibamoto et al. 2004). Judging from the signal strength of the DNA chip data, the induction level of these genes was similar under both aggregate and monolayer culture conditions (Table 1, Fig. 1A). These results demonstrated that at least at the gene expression level, monolayer, and aggregate cultured F9 cells were similarly induced by RA.
|
|
-Fetoprotein (AFP) is commonly used as a final differentiation marker of VE (Hogan et al. 1981). Villin is a component of microvilli and is reported to be a protein, the expression of which is induced by RA in F9 cells (Buc-Caron et al. 1989). Our DNA chip data indicated that AFP and villin expression was not detected 3 days following RA treatment, even in aggregate cultured F9 cells (data not shown), suggesting that the expression of AFP and villin was induced 4 days or more following RA treatment. In an effort to determine whether the expression of AFP and villin was induced by RA even in monolayer culture, we performed RT-PCR using the mRNA derived from monolayer cultured F9 cells prior to and 5 days following RA treatment (Fig. 2). For semiquantitative purposes, we compared the data derived from the 30-cycle and 35-cycle amplification. The expression of AFP and villin was barely detected in undifferentiated F9 cells, even after a 35-cycle amplification. Five days following RA treatment, the expression of AFP and villin was detected even with a 30-cycle amplification. These data indicated that RA could induce AFP and villin gene expression even under monolayer culture conditions.
|
Using data derived from the DNA chip analysis, we investigated the gene expression of junctional complex components in monolayer cultured F9 cells. Using the quick search program located on the Affymetrix home page (http://www.affymetrix.com), we found many genes associated with several key words including "tight junction,""claudin,""adherens junction,""cadherin,""catenin," and "desmosome." Of these, we selected those genes that showed a signal greater than 50, and were judged as present in the data derived from the aggregate cultured F9 cells following RA treatment, as adhesion-related proteins highly expressed in differentiated F9 cells (Table 1). The DNA chip analysis data clearly showed that all of these genes were expressed even in monolayer cultured F9 cells 3 days following RA treatment. TJ and DS were formed following RA treatment. Consistent with these observations, some of the TJ and DS genes including claudin-6, claudin-7, and desmocollin 2 were induced fivefold or more following RA treatment (Table 1). The induction of these genes was also observed even in monolayer cultured F9 cells following RA treatment (Fig. 1B). The expression level of ZO-1, which is known as a TJ protein but co-localizes with cadherin in nonepithelial cells lacking TJ (Itoh et al. 1993), was unaffected following RA treatment (Fig. 1C). The expression of AJ components including vinculin, E-cadherin, -catenin, and ß-catenin remained unchanged following RA treatment (Fig. 1C). These data clearly showed that monolayer cultured F9 cells could express junctional complex components that F9 cell aggregates usually express in the presence of RA.
Induction of the ZO-1+ isoform by RA in monolayer cultured F9 cells
It has been reported that an isoform of ZO-1 with exon (ZO-1+) was induced during simple epithelial differentiation (Sheth et al. 1997). We set out to examine the expression of ZO-1+ in F9 cells. RT-PCR analysis showed that a low level of exon expression was detected in undifferentiated F9 cells and that the expression was highly induced during differentiation of monolayer cultured F9 cells (Fig. 2). Western blot analysis showed that the commonly used anti-ZO-1 antibody, T8-754, detected an additional polypeptide with a slightly higher molecular weight in F9 cells following RA treatment (Fig. 3A). We then attempted to clarify whether or not this additional polypeptide is an + variant of ZO-1. Commercially available anti-ZO-1 antibodies were initially screened and it was found that anti-ZO-1 mAb purchased from Chemicon International, Inc. mainly detected the additional polypeptide (Fig. 3A). This antibody detected, albeit weakly, the same band even in undifferentiated F9 cells. To confirm that this antibody specifically recognized the ZO-1+ isoform, we constructed expression vectors for ZO-1-GFP and ZO-1(+)-GFP fusion proteins without and with + exon, respectively. The vectors were introduced into F9 cells and stable transfectants that expressed ZO-1-GFP or ZO-1(+)-GFP fusion molecules were isolated. Western blot analysis showed that commonly used anti-ZO-1 antibodies detected expressed fusion proteins in addition to endogenous ones (Fig. 3B). The Chemicon anti-ZO-1 antibodies clearly detected an additional polypeptide in the sample derived from the transfectant expressing ZO-1(+)-GFP. In nontransfected cells and cells transfected with ZO-1-GFP expression vector, however, only endogenously expressed ZO-1+ was detected. Taken together, we concluded that the Chemicon anti-ZO-1 antibody is specific for the ZO-1+ isoform. It was also demonstrated that specific alternative splicing to produce an + variant was prominent in F9 cells during RA treatment.
|
The subcellular localization of junctional complex components in monolayer cultured F9 cells was then investigated. We chose claudin-6, ZO-1+, ZO-1, E-cadherin, and desmoplakin as major representatives of junctional complex components. The localization of actin filaments as a cytoskeleton component of AJ was also investigated. In undifferentiated F9 cells, the signals of claudin-6, ZO-1+ and desmoplakin were barely detected (Fig. 4Aa, c and e). E-cadherin, an AJ component, was expressed and localized at cell to cell contact sites, which are not linear but fragmental at cell to cell boundaries (Fig. 4Ab). ZO-1 was also localized at cell to cell contact sites even in undifferentiated F9 cells (Fig. 4Ad). Because actin staining showed strong signals in the cytoplasm, its localization at cell to cell contact sites was unclear (Fig. 4Af).
|
+ and ZO-1 showed sharp, linear localization (Fig. 4Aa'–d'). Condensation of the actin filaments was prominent at cell to cell contact sites, and probably reflected epithelial-type AJ (zonula adherens) formation (Fig. 4Af'). Despite the linear localization of the TJ and AJ components, desmoplakin showed typical dot-like localization at cell to cell boundaries, indicating the formation of desmosomes (Fig. 4Ae'). Ultrastructural analysis confirmed the formation of junctional complexes in monolayer cultured F9 cells following RA treatment. When cells with apical-like structures were closely examined, junctional complexes composed of TJ, AJ, and DS were observed at the uppermost part of cell to cell boundaries (Fig. 5A). Because TJ was not clearly separated from but intermingled with AJ, both TJ and AJ seemed to be supported by actin-based cytoskeletons (Fig. 5B). These data clearly showed that differentiated F9 cells could develop junctional complexes even under monolayer culture conditions following RA treatment.
|
A morphological analysis employing electron microscopy confirmed the epithelial differentiation of F9 cells in monolayer culture. Seven days following RA treatment, two types of cells in the F9 monolayer culture were identified. The morphology of one cell type was similar to that of undifferentiated F9 cells (Fig. 5C). The cell was thick at the level of the nucleus, but thin at cell to cell boundaries. Only a few microvilli and cytoplasmic vacuoles were observed (Fig. 5C,E). The second cell type showed clear epithelial morphology (Fig. 5D). The number of microvilli increased at the apical free surface. Various vacuoles were present mainly above the nucleus, that is, under the apical surfaces (Fig. 5F). These morphological observations provided evidence of epithelium formation under monolayer culture conditions.
Subcellular localization of junctional complex components during epithelial differentiation
The subcellular localization of junctional proteins in F9 cells treated with RA for 3 days was investigated in an effort to delineate the initial step of junctional complex formation. Among the junctional components whose expression was induced by RA, the expression of ZO-1+ and desmoplakin was clearly detected in F9 cells 3 days following RA treatment. The expression of ZO-1+ was restricted in only certain groups of cells, and the expressed molecules were localized at cell to cell boundaries in a linear fashion (Fig. 4Ba'). Although fasciculation of actin filaments was observed, the claudin-6 signals were barely detected in most of the ZO-1+ positive junctions in this period (data not shown). Expression of desmoplakin was detected in many cells irrespective of the expression of ZO-1+ (Fig. 4Ba and a'). The localization of desmoplakin to cell to cell contact sites was occasionally observed, but it was weak and unclear. Expressed desmoplakin was mainly localized in the cytoplasm and showed a filamentous distribution. These data indicated that the expression and localization of ZO-1+ was closely synchronized with the formation of a linear junction, probably an epithelial-type AJ. It was also shown that the expression of desmoplakin was not synchronized with DS formation.
Gene expression of apical-related components during RA treatment of monolayer cultured F9 cells
Using data derived from the DNA chip analysis, the gene expression of apical-related components was investigated. Using the quick search program on the Affymetrix home page, we found several genes associated with the key word "apical." Of these, three genes were selected, calbindin-D9K, legumain, and cubilin, as apical-related genes highly induced by RA. The signals of these genes were greater than 50 and judged as present in the data derived from the RA-treated aggregate sample. Moreover, the expression was induced fivefold or more when the gene expression signals obtained from the RA-treated aggregate sample were compared to those obtained from the control aggregate sample (Table 1). Cubilin, also known as an intrinsic factor-cobalamin receptor, is a multiligand scavenger receptor (Fyfe et al. 2004). Legumain is a lysosomal endopeptidase (Shirahama-Noda et al. 2003) and calbindin-D9K is an intracellular calcium transport (Bouillon et al. 2003). All of these proteins seem to be involved in absorption or degradation of some materials from the apical membrane. The DNA chip data showed that the expression of these apical-related genes was also highly induced by RA in monolayer culture (Table 1, Fig. 1D). These data indicated that the expression of several genes involved in the absorptive function at the apical surface is induced during the differentiation of F9 cells under monolayer culture conditions.
Apical membrane formation in RA-treated F9 cells in monolayer culture
Ezrin, a member of the ERM family proteins, and syntaxin 3, a t-SNARE protein, are localized at the apical surface in many types of simple epithelial cells (Berryman et al. 1993; Mostov et al. 2000). DNA chip data suggested that RA increased expression of the aforementioned genes under aggregate and monolayer culture conditions (Table 1, Fig. 1). Even in the absence of RA, however, these genes were significantly expressed. Consistent with the DNA chip data, Western blot analysis showed that ezrin and syntaxin 3 protein expression was detected even in undifferentiated F9 cells (Fig. 6A). The expression level of these proteins increased significantly following RA treatment (Fig. 6A). Immunocytochemical analysis demonstrated a change in the subcellular distribution of ezrin following RA treatment (Fig. 4B,b). Despite protein being expressed, ezrin was barely detected and did not show any specific subcellular localization in undifferentiated F9 cells (data not shown). Following RA induction, the subcellular localization of ezrin protein in F9 monolayer cultures changed dramatically. When we focused on cells that had differentiated into epithelial cells judging from the linear localization of ZO-1, ezrin seemed to be localized at cell to cell boundaries and the upper surface of cells (Fig. 4B,b). In certain cells, a dotty distribution was observed, which reflected microvilli formation. Confocal microscopic analysis demonstrated that F9 cells had differentiated into epithelial cells with uniform height, and the localized signals of ezrin appeared at the upper plasma membrane (Fig. 4Da). The immunostaining signal with anti-syntaxin 3 antibody was too weak to provide any information concerning the subcellular localization. To confirm the apical localization of syntaxin 3, we constructed an expression vector for a syntaxin 3-GFP fusion, introduced the expression vector into F9 cells, and examined the expression and subcellular localization of the fusion molecules prior to and following RA treatment of monolayer cultures. Western blot analysis showed that syntaxin 3-GFP fusion proteins with the expected molecular weight were expressed even in undifferentiated F9 cells (Fig. 6B). The expressed fusions, however, were not localized at the cell surface but were irregularly localized in the cytoplasm (Fig. 4Ca). Five days following RA treatment, the subcellular localization of the fusion proteins changed dramatically compared to the situation observed for ezrin (Fig. 4Cb,d, see Fig. 4Bb). When we focused on cells expressing ZO-1+, the syntaxin 3-GFP fusion seemed to be localized at the upper cell surface. Confocal microscopic analysis again demonstrated that the syntaxin 3-GFP signals were localized at the upper plasma membrane (Fig. 4Db). Interestingly, the expression level of both endogenous and exogenously expressed syntaxin 3 proteins increased following RA treatment (Fig. 6). Because transcription from the introduced expression vector appeared not to be affected by RA treatment, syntaxin 3 protein may be stabilized during the epithelial differentiation of F9 cells. Recently it was well accepted that PAR-3 complex are involved in the epithelial polarity formation in mammalian cells (Ohno 2001). The localization of PAR-3 and its membrane partner JAM-A (Itoh et al. 2001) in differentiated F9 cells were then examined. When cells expressing ZO-1+ were examined, both PAR-3 and JAM-A were localized at cell to cell boundaries in a linear fashion and co-localized with ZO-1+ (Fig. 7). These results indicated that the apical membrane was formed and polarity related-proteins were localized at epithelial junctions in monolayer cultured F9 cells following RA treatment. It was also suggested that apical membrane localization of syntaxin 3 was strictly regulated during epithelial differentiation.
|
|
Finally, using the advantages of monolayer culture conditions, the time course of epithelial differentiation of F9 cells was examined (Fig. 8). The expression and linear localization of ZO-1+, which correlated with epithelial AJ formation and apical localization of ezrin and syntaxin 3, appeared to represent an optimal marker for epithelial differentiation of F9 cells. The number of nuclei that were surrounded by ZO-1+ positive linear contacts was used as representing the number of epithelial differentiated F9 cells. Up until 2 days following RA induction, the number of cells present that displayed a linear localization of ZO-1+ at cell to cell contact sites was negligible. Three days following RA treatment, a linear localization of ZO-1+ at cell to cell boundaries in a group of cells was detected. From 4 to 5 days following RA induction, the number of epithelial-differentiated F9 cells present increased markedly. Because the size of epithelial-differentiated F9 cells was larger than that of nonepithelial cells, epithelial sheets covered 60–70% of the bottom surface of the dishes 5 days following RA induction. From 7 days following induction, the number of epithelial F9 cells settled to around 70% of the total number of cells, which covered 80–90% of the bottom surface. These data suggested that dynamic processes of epithelial morphogenesis were frequently observed 4 days following RA induction.
|
|
Discussion |
|---|
|
TopAbstractIntroductionResultsDiscussionExperimental proceduresReferences |
|---|
Epithelial differentiation of F9 cells under monolayer culture conditions
The molecular mechanisms of epithelium formation in multicellular organisms are only beginning to be understood. For a detailed analysis, it is useful to establish an epithelial differentiation system in monolayer culture. It is, however, a little difficult to distinguish epithelial cells from nonepithelial cells in vitro, as even nonepithelial cells form a monolayer on the surface of a culture dish. Consequently, some definitions of epithelial cells are necessary when employing the monolayer culture system.
The use of a gene expression profile is now available to define the state of differentiation. It has already been established that F9 cells differentiate into epithelial cells when cell aggregates are cultured in the presence of RA (Hogan et al. 1981). We compared the gene expression profile of monolayer cultured F9 cells with that of aggregate cultured cells prior to and following RA treatment. Most of the genes induced by RA in aggregate cultured cells were also induced in monolayer cultured cells. In particular, keratin 8 and keratin 18, which are commonly expressed in simple epithelial cells (Owens & Lane 2003), were highly induced in the presence of RA under monolayer and aggregate culture conditions. Aside from the genes known thus far, the expression of which is regulated during epithelial differentiation, DNA chip analysis showed that the expression of many genes was induced or reduced following RA treatment of F9 cells. For example, 177 genes were induced and 34 genes were reduced threefold or more following RA induction under both aggregate and monolayer culture conditions. Whether these genes are involved in epithelial differentiation remains to be determined. Further studies employing DNA chip analysis of epithelial and nonepithelial cells would assist in the identification of novel epithelial differentiation regulatory genes.
Junctional complex formation is also a characteristic of epithelial cells. It has been reported that TJ and DS were induced even in monolayer cultured F9 cells in the presence of RA (Trevor & Steben 1992; Chiba et al. 2003). Here we also demonstrated the reorganization and fasciculation of actin filaments at cell to cell boundaries, suggesting that AJ is converted from a nonepithelial type to an epithelial type in the presence of RA under monolayer culture conditions. Moreover, electron microscopy analysis showed junctional complexes composed of TJ, AJ and DS were formed in RA-treated F9 cells under monolayer culture conditions.
Another characteristic of epithelial cells relates to the establishment and maintenance of the apico-basal epithelial polarity, which is required for epithelial tissue functions such as the directional transport of several materials (Rodriguez-Boulan & Nelson 1989). RA treatment of F9 cells under monolayer culture conditions induced a major aspect of epithelial polarity, the correct sorting of apical plasma membrane markers, ezrin (Saotome et al. 2004) and syntaxin 3 (Kreitzer et al. 2003), and the junctional localization of polarity-related proteins JAM-A (Itoh et al. 2001) and PAR-3 (Ohno 2001). Ezrin staining also showed microvilli formation, which is commonly observed at apical surfaces of epithelial cells. Electron microscopy analysis confirmed the typical apical membrane structure with microvilli and clearly discerned the apical and basal cytoplasm. Thus, judging from the epithelial-specific gene expression, junctional complex, and apical membrane formation, F9 cells were able to differentiate into epithelial cells under monolayer culture conditions.
The earlier study reported that differentiated F9 cells could not maintain their epithelial morphology under monolayer culture conditions. This appears to be inconsistent with our present data. It is, however, known that epithelial morphology can be lost in cultured epithelial cells following reseeding. Under normal monolayer culture conditions, it takes about 5 or more days to recover original epithelial morphology. It was recently reported that activation of LKB/PAR-4 induced this morphological change in a single epithelial cell (Baas et al. 2004). Epithelial-differentiated F9 cells might require 3 days or more to establish their epithelial morphology under monolayer culture conditions.
Although F9 cells could form a simple epithelium, the differentiation of F9 cells into VE cells seemed to be incomplete under monolayer culture conditions. So far, AFP secretion is commonly used as a final differentiation marker of VE derived from F9 cells (Hogan et al. 1981). Although it was reported that F9 cells could not secrete AFP under monolayer culture conditions even in the presence of RA (Casanova & Grabel 1988), we detected the expression of AFP mRNA. It was likely that aggregate formation may be necessary for the proper secretion of this protein, but not for gene expression. Recently it was reported that the cubilin/amnionless complex, which is involved in the endocytosis/transcytosis of one or more ligands from the apical membrane, is expressed exclusively in VE during the early postimplantation stage (Fyfe et al. 2004; Strope et al. 2004). Although strong induction of cubilin in F9 cells was detected both under monolayer and aggregate culture conditions, induced expression of amnionless was not observed in monolayer cultured F9 cells following RA treatment (M.S. & A.N., unpublished observation). Microvilli formation increased in RA-treated F9 cells under monolayer culture conditions. The total number and length of microvilli, however, seemed fewer and shorter than those observed on the VE-differentiated outer surface of F9 cell aggregates (Casanova & Grabel 1988). These results might show that F9 cells only partially differentiate into VE cells under monolayer culture conditions and that aggregate formation is a prerequisite for complete differentiation into the VE.
Apical membrane formation during epithelial differentiation
Well-developed microvilli, such as the brush border of intestinal cells, are regarded as typical apical structures. Recently it was reported that the definition of apical vs. basolateral membrane domains might result directly from establishment of the brush border (Baas et al. 2004). Our observation, however, demonstrated that the density of microvilli was relatively sparse compared to the intestinal microvilli in differentiated F9 cells. It was also shown that apical localization of ezrin, a major constituent of microvilli, was observed despite the density of microvilli. These observations suggested that ezrin is first localized at the apical surface, then some signal(s) induce microvilli formation at the membrane domain with ezrin.
Because syntaxin 3 is involved in vesicle transport to the apical membrane in simple epithelial cells (Mostov et al. 2000), it could be that this molecule is a determinant of the apical membrane domain. This, however, is not the case because the endogenous syntaxin 3 is expressed even in undifferentiated F9 cells. Moreover the exogenously expressed syntaxin 3 showed irregular localization in the cytoplasm of undifferentiated F9 cells. How are apical membrane domains determined and established during the differentiation of F9 cells? The cellular model described here may be useful in delineating the molecular mechanisms involved in the initial step of apical membrane formation during epithelial differentiation.
Junctional complex formation during epithelial differentiation
The process concerning junctional complex formation of simple epithelia in mammals has been extensively investigated using the calcium switch model of epithelial cell lines. In this model system, a cadherin-based cell adhesion system is completely inactivated prior to reformation of the junctional complex. Cadherin-mediated cell adhesion, however, is active even in nonepithelial cells prior to epithelial differentiation in vivo. Because cadherin-mediated cell adhesion is always active in the F9 model used here, the process involving junctional complex formation was expected to be more physiological.
DSs are newly formed in F9 cells following RA treatment. Consistent with this, the expression of desmoplakin protein was induced by RA. Interestingly, the desmoplakin signals were originally observed in the cytoplasm and showed a filamentous localization. It was previously reported that this cytoplasmic desmoplakin co-localized with cytokeratin filaments (Trevor & Steben 1992). Cells expressing cytoplasmic desmoplakin did not necessarily overlap with those expressing ZO-1+. Cytoplasmic desmoplakin disappeared in cells with well-developed desmosomes following 5 days differentiation. These data indicated that the expression of desmosome components and cytokeratin are not necessarily synchronized with epithelial-type AJ formation. It was also demonstrated that desmoplakin could decorate cytokeratin filaments if it was not included in proper DSs. In epithelial junctional complexes, DSs are usually positioned more basally to AJs (Farquhar & Palade 1963). The molecular cues employed to localize DSs remains unknown. Detailed analysis of the behavior of DS components during simple epithelial differentiation may provide new insights into the regulatory mechanisms involved in junctional complex formation.
We clearly showed that the expression of ZO-1+, the alternative splicing product of ZO-1, was induced during epithelial differentiation of F9 cells. The expression of ZO-1+ was also induced during epithelial differentiation in mouse early embryos (Sheth et al. 1997). Since localization of the ZO-1+ isoform at cell to cell boundaries in early mouse embryos is well synchronized with TJ membrane proteins, it was suggested that the ZO-1+ isoform is involved in TJ formation. The role of the + domain and the physiological meaning of + isoform expression during epithelial differentiation, however, remain unknown. Aside from a physiological function, expression of the ZO-1+ isoform could be used as a kind of marker for epithelial differentiation in terms of alternative splicing. Because the subcellular localization was restricted to newly organized epithelial-type AJ lacking TJ and was roughly synchronized with the apical localization of ezrin, ZO-1+ protein also seems to be suitable for use as an epithelial differentiation marker in an immunocytochemical analysis of the F9 cell model. Using this marker and advantages of the monolayer culture system, the time course of epithelial differentiation of F9 cells was estimated. Following RA induction, epithelial-differentiated cells began to be observed on the third day, the number of which increased markedly on the fourth and fifth day. The final number of epithelial differentiated cells reached 60–70% of the total cell population. The first appearance of epithelial differentiated cells on the third day was consistent with a previous report that indicated that treatment with RA for 2 days was sufficient to determine epithelial differentiation of F9 cells (Sato et al. 1985). It was also consistent with our present DNA chip data, which showed that many epithelial differentiation marker genes were induced in the sample prepared from F9 cells cultured for 3 days in the presence of RA. Our data also indicated that more than half of the epithelial cells had differentiated on the fourth and fifth days. During this active differentiation period, complicated processes involved in junction and apical membrane formation could be examined under monolayer culture conditions.
In this paper, we have shown that monolayer cultured F9 cells represent a viable model system for simple epithelium formation. Using the advantages of the monolayer culture system as a model, we demonstrated that apical membrane and junctional complex formation were highly regulated during epithelial differentiation, although the precise mechanisms remain unclear. It was also revealed that F9 cells differentiated into epithelial cells predominantly on the fourth and fifth day following RA induction. Employment of gene disruption strategies can also provide useful information concerning the molecular mechanisms pertaining to F9 cells (Stephens et al. 1993; Volberg et al. 1995; Clifford et al. 1996; Chiba et al. 1997; Maeno et al. 1999). For example, when we isolated genetically arranged F9 cells lacking genes for junctional proteins or epithelial polarity regulating proteins, the roles of these genes in epithelial formation could be clearly determined.
|
Experimental procedures |
|---|
|
TopAbstractIntroductionResultsDiscussionExperimental proceduresReferences |
|---|
Mouse F9 teratocarcinoma cells and their derivatives were cultured in Dulbecco's modified Eagle medium containing 10% heat-inactivated fetal calf serum (Gibco BRL). Culture dishes, cover slips or Transwell (CORNING) were precoated with 0.2% gelatin for 15 min. To induce epithelial differentiation in monolayer cultures, cells were placed at a density of 6 x 103 cells/cm2 in the presence of 50 nM or 1 µM all-trans-RA (Sigma Chemical Co.). In suspension, F9 cells were allowed to form an aggregate composed of 1000 cells using the hanging drop method and cultured in EOG-sterilized polystyrene Petri dishes (IWAKI). To induce epithelial differentiation, cell aggregates were cultured in the presence of 50 nM all-trans-RA.
Oligonucleotide array analysis
Total RNA was prepared from F9 cells treated with or without RA for 3 days using RNeasy Mini columns (QIAGEN). Double-stranded cDNA was prepared from approximately 12 µg of total RNA using Superscript II (Life Technologies), as recommended by Affymetrix. The cDNA was used as a template in a biotin-labeled in vitro transcription reaction to form cRNA according to the manufacturer's protocol (Enzo Bioarray, Affymetrix). Biotinylated cRNA was collected on RNeasy columns (QIAGEN) and used for preparation of the hybridization reaction. The mouse genome U74Av2 microarray from Affymetrix was used in all hybridizations. This array contains probe sets for approximately 6000 full-length genes and approximately 6000 EST clusters. Biotinylated target cRNA was hybridized on to the array and then processed using the Affymetrix gene chip Fluidics Workstation 400. Following binding with phycoerythrin-coupled streptavidin, microarrays were scanned on a gene array scanner (Agilent Technologies).
Microarray data was analyzed with Affymetrix MICROARRAY SUITE version 5.0. The absolute analysis algorithm was used to calculate whether genes were "absent" or "present" based on default parameters. To compare signals derived from different samples, the data from the F9 cell aggregates cultured in the absence of RA were used as a baseline file, and "all probe sets" was selected for normalization.
Reverse transcriptional PCR
Total RNA was prepared from F9 cells treated with or without RA for 5 days using a RNeasy Mini Kit (Qiagen). First-strand cDNA synthesis was conducted in a total volume of 20 µL using the SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen) with 2 µg of total RNA. PCR amplifications were performed in a 25-µL reaction volume containing 1x reaction buffer, 200 µM dNTPs, 10 pmol of each primer, 0.5 µL of cDNA and 1.25 U of Ex Taq DNA polymerase (Takara Bio Inc.). Each cycle consisted of denaturation at 95 °C for 30 s, annealing at 57 °C for 30 s, followed by extension at 72 °C for 30 s. Primers were designed as follows:
AFP (upstream, 5'-GACCTCAGCAGAGCTGATCGA-3'; downstream, 5'-CCTTCTGGAGATGTTTAAACGCC-3')
villin (upstream, 5'-CATCACTCCTCGGCTCTTCG-3'; downstream, 5'-GTCATGCCCAAGGCCCTAGT-3')
ZO-1+ splicing variant (upstream, 5'-CCTGGACTTAAGCCAGC-3'; downstream, 5'-CCTTCCTGTACACCTTTGC-3')
GAPDH (upstream, 5'-CCATCACCATCTTCCAGGAG-3'; downstream, 5'-CCTGCTTCACCACCTTCCTTG-3')
Immunological analysis
The following primary antibodies were used for the immunocytochemical and immunoblot analyses: mouse anti-rat ZO-1 monoclonal antibody (mAb) (T8-754; Itoh et al. 1993), rat anti-mouse ZO-1 mAb (Chemicon International, Inc.), anti-GFP polyclonal antibody (pAb) (Molecular Probes), rabbit anti-claudin-6 pAb (Morita et al. 1999), rat anti-mouse E-cadherin mAb (ECCD-2; Shirayoshi et al. 1986), mouse anti-desmoplakin I & II (Multi-epitope cocktail to desmoplakin I & II, Progen Biotechnik), rat anti-mouse ezrin mAb (M11; Takeuchi et al. 1994), rabbit anti-syntaxin 3 pAb (Synaptic Systems), and rabbit anti-PAR-3 pAb (C2-3AP; Izumi et al. 1998). The anti-JAM-A pAb was raised in rabbits against a synthesized peptide corresponding to the mouse JAM-A sequence located at amino acid positions 282-300 (SQPSTRSEGEFKQTSSFLV). The anti-JAM-A pAb was then affinity-purified using this peptide. FITC- or Cy3-conjugated donkey anti-rabbit, anti-mouse, and anti-rat IgGs (Jackson Immunoresearch, West Grove, PA) were used as secondary antibodies. Alexa Fluor 488-labeled phalloidin (Molecular Probes) was used for the actin staining.
SDS-PAGE and immunoblotting were performed as previously described. Samples were solubilized in SDS sample buffer and then separated by SDS-PAGE. For the immunoblotting, proteins were electrophoretically transferred onto nitrocellulose sheets. Nitrocellulose membranes were then incubated with primary antibody. Antibody detection was performed using an Amersham biotin-streptavidin kit.
For the immunofluorescent staining, cells cultured on 15-mm gelatin-coated cover slips or 12-mm Transwell (for confocal images) were washed, fixed, and then incubated with primary and secondary antibodies as previously described (Imamura et al. 1999). Cells were fixed with ice-cold methanol for 10 min (for anti-desmoplakin I & II) or 1% paraformaldehyde for 15 min. For the actin staining, samples were incubated with Alexa Flour 488-labeled phalloidin instead of the secondary antibody. Samples were embedded using a SlowFade Light Antifade Kit (Molecular Probes). Images were acquired using the AxioVision 3.0.6 equipped with a microscope (model Axiovert 200; Carl Zeiss MicroImaging, Inc.), and AxioCAM cooled CCD camera, with a 40/0.75 NA plan-Neofluar objective. An Olympus confocal microscope (model FLUOVIEW FV500; Olympus) with a 40/1 NA UPlanApo objective was used to acquire the confocal image. Adjustments of brightness, contrast, color balance, and final image size of images were achieved using Adobe PHOTOSHOP version 6.0.
Expression vectors and transfection
pCAG-CGFP (Matsuda et al. 2004) was used for the construction of GFP fusion protein expression vectors. In pCAG-CGFP, the EcoRI-ApoI fragment of a GFP cDNA was inserted into the EcoRI site of the pCAGGSneodelEcoRI vector (Niwa et al. 1991). In the GFP cDNA, EcoRI and XhoI sites were inserted just before an initiation codon, and an ApoI site was inserted just after the termination codon using PCR methods. For the construction of pCAG-ZO-1-GFP and pCAG-ZO-1( +)-GFP, the expression vectors for ZO-1-GFP fusions, the EcoRI fragments of ZO-1 cDNA were inserted into the EcoRI site of pCAG-CGFP. The ZO-1 cDNA fragments were derivatives from the full-length ZO-1 cDNA without or with + exon (Itoh et al. 1993). In both cDNAs, the EcoRI site was inserted just before an initiation methionine codon. Stop codons were also replaced with the EcoRI recognition sequence. To construct pCAG-Syn3-GFP, the expression vector for the syntaxin 3-GFP fusion, the EcoRI-XhoI fragment of syntaxin 3 cDNA was inserted into the EcoRI-XhoI site of pCAG-CGFP. This full-length syntaxin 3 cDNA was isolated by RT-PCR using mRNA derived from 3-day differentiated F9 cells using the primer pairs (upstream, 5'-GAATTCGGGCCAGGATGAAGGACCGG-3'; downstream, 5'-CTCGAGTTTCAGCCCAACGGACAGTC-3'; bold letters show the EcoRI site, initiation methionine codon, and the XhoI site, respectively) (Bennett et al. 1993). In this cDNA, the EcoRI site was inserted eight bases before the initiation methione codon. The stop codon was replaced with an XhoI site.
Expression vectors were transfected into F9 cells using the Lipofectamin 2000 system (Gibco BRL). Cells were subjected to G418 selection at 400 µg/mL for 2 weeks. At least five clones were isolated for each transfectant and one clone, which showed a common phenotype, was chosen as the representative clone.
Electron microscopy
Cells were fixed with 2% fresh formaldehyde and 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.4) for 2 h followed by postfixation with 1% OsO4 in the same buffer for 1 h on ice. To prevent disruption of membranes or extraction of the cytoplasmic matrix, 0.2% tannic acid and 0.05% saponin were added in the fixation buffer when the high magnification image of the junctional complex was taken (Maupin & Pollard 1983). Samples were then dehydrated with ethanol and embedded in Epon 812. Thin sections were cut using a diamond knife, doubly stained with uranyl acetate and lead citrate, then examined using a JEM 1010 electron microscope (JEOL, Tokyo, Japan) at an accelerating voltage of 100 kV.
Quantification of epithelial differentiated F9 cells
Cells were doubly stained with anti-ZO-1+ antibody and DAPI. To estimate number of epithelial differentiated cells, we chose randomly 10 fields of a microscope and counted number of nuclei that were surrounded by linear ZO-1+ signals. The ratio of number of nucleus between epithelial differentiated cells and total cells was presented as a rate of epithelial differentiation. For each experiments, at least 400 nuclei were counted.
|
Acknowledgements |
|---|
Part of this work at the Department of Cellular Interactions of the Institute of Molecular Embryology and Genetics of Kumamoto University was supported by a grant to A.N. from the Core Research for Evolutional Science and Technology (CREST); Grants-in-Aid for Scientific Research and Cancer Research and a Grant-in-Aid for 21st Century COE Research ‘Cell Fate Regulation Research and Education Unit’ from the Ministry of Education, Culture, Sports, Science, and Technology of Japan. H.L. had been a JSPS postdoctoral fellow.
|
Footnotes |
|---|
These authors contributed equally to this work. 
* Correspondence: E-mail: naga-san{at}kaiju.medic.kumamoto-u.ac.jp |
References |
|---|
|
TopAbstractIntroductionResultsDiscussionExperimental proceduresReferences |
|---|
Baas, A.F., Kuipers, J., van der Wel, N.N., et al. (2004) Complete polarization of single intestinal epithelial cells upon activation of LKB1 by STRAD. Cell 116, 457–466.[CrossRef][Medline]
Bennett, M.K., Garcia-Arraras, J.E., Elferink, L.A., et al. (1993) The syntaxin family of vesicular transport receptors. Cell 74, 863–873.[CrossRef][Medline]
Berryman, M., Franck, Z. & Bretscher, A. (1993) Ezrin is concentrated in the apical microvilli of a wide variety of epithelial cells whereas moesin is found primarily in endothelial cells. J. Cell Sci. 105 (4), 1025–1043.[Abstract]
Bouillon, R., Van Cromphaut, S. & Carmeliet, G. (2003) Intestinal calcium absorption: molecular vitamin D mediated mechanisms. J. Cell Biochem. 88, 332–339.[CrossRef][Medline]
Buc-Caron, M.H., Lamblin, D. & Kellermann, O. (1989) F9 teratocarcinoma aggregates express villin upon differentiation into visceral endoderm-like cells. Biol. Cell 65, 195–198.[CrossRef][Medline]
Casanova, J.E. & Grabel, L.B. (1988) The role of cell interactions in the differentiation of teratocarcinoma-derived parietal and visceral endoderm. Dev. Biol. 129, 124–139.[CrossRef][Medline]
Chiba, H., Clifford, J., Metzger, D. & Chambon, P. (1997) 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, 735–747.
Chiba, H., Gotoh, T., Kojima, T., et al. (2003) Hepatocyte nuclear factor (HNF)-4 triggers formation of functional tight junctions and establishment of polarized epithelial morphology in F9 embryonal carcinoma cells. Exp. Cell Res.
286, 288–297.[CrossRef][Medline]
Clifford, J., Chiba, H., Sobieszczuk, D., Metzger, D. & Chambon, P. (1996) RXR-null F9 embryonal carcinoma cells are resistant to the differentiation, anti-proliferative and apoptotic effects of retinoids. EMBO J.
15, 4142–4155.[Medline]
Farquhar, M.G. & Palade, G.E. (1963) Junctional complexes in various epithelia. J. Cell Biol.
17, 375–412.
Fyfe, J.C., Madsen, M., Hojrup, P., et al. (2004) The functional cobalamin (vitamin B12)-intrinsic factor receptor is a novel complex of cubilin and amnionless. Blood
103, 1573–1579.
Green, K.J. & Gaudry, C.A. (2000) Are desmosomes more than tethers for intermediate filaments? Nat. Rev. Mol. Cell Biol. 1, 208–216.[CrossRef][Medline]
Grover, A., Oshima, R.G. & Adamson, E.D. (1983) Epithelial layer formation in differentiating aggregates of F9 embryonal carcinoma cells. J. Cell Biol.
96, 1690–1696.
Harris, T.M. & Childs, G. (2002) Global gene expression patterns during differentiation of F9 embryonal carcinoma cells into parietal endoderm. Funct. Integr. Genomics 2, 105–119.[CrossRef][Medline]
Hogan, B.L., Taylor, A. & Adamson, E. (1981) Cell interactions modulate embryonal carcinoma cell differentiation into parietal or visceral endoderm. Nature 291, 235–237.[CrossRef][Medline]
Imamura, Y., Itoh, M., Maeno, Y., Tsukita, S. & Nagafuchi, A. (1999) Functional domains of -catenin required for the strong state of cadherin-based cell adhesion. J. Cell Biol.
144, 1311–1322.
Itoh, M., Nagafuchi, A., Yonemura, S., et al. (1993) The 220-kD protein colocalizing with cadherins in non-epithelial cells is identical to ZO-1, a tight junction-associated protein in epithelial cells: cDNA cloning and immunoelectron microscopy. J. Cell Biol.
121, 491–502.
Itoh, M., Sasaki, H., Furuse, M., et al. (2001) Junctional adhesion molecule (JAM) binds to PAR-3: a possible mechanism for the recruitment of PAR-3 to tight junctions. J. Cell Biol.
154, 491–497.
Izumi, Y., Hirose, T., Tamai, Y., et al. (1998) 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, 95–106.
Keller, P., Toomre, D., Diaz, E., White, J. & Simons, K. (2001) Multicolour imaging of post-Golgi sorting and trafficking in live cells. Nat. Cell Biol. 3, 140–149.[CrossRef][Medline]
Kreitzer, G., Schmoranzer, J., Low, S.H., et al. (2003) Three-dimensional analysis of post-Golgi carrier exocytosis in epithelial cells. Nat. Cell Biol. 5, 126–136.[CrossRef][Medline]
Low, S.H., Chapin, S.J., Wimmer, C., et al. (1998) The SNARE machinery is involved in apical plasma membrane trafficking in MDCK cells. J. Cell Biol.
141, 1503–1513.
Maeno, Y., Moroi, S., Nagashima, H., et al. (1999) -catenin-deficient F9 cells differentiate into signet ring cells. Am. J. Pathol.
154, 1323–1328.
Matsuda, M., Kubo, A., Furuse, M. & Tsukita, S. (2004) A peculiar internalization of claudins, tight junction-specific adhesion molecules, during the intercellular movement of epithelial cells. J. Cell Sci.
117, 1247–1257.
Maupin, P. & Pollard, T.D. (1983) Improved preservation and staining of HeLa cell actin filaments, clathrin-coated membranes, and other cytoplasmic structures by tannic acid-glutaraldehyde-saponin fixation. J. Cell Biol.
96, 51–62.
Morita, K., Sasaki, H., Furuse, M. & Tsukita, S. (1999) Endothelial claudin: claudin-5/TMVCF constitutes tight junction strands in endothelial cells. J. Cell Biol.
147, 185–194.
Mostov, K.E., Verges, M. & Altschuler, Y. (2000) Membrane traffic in polarized epithelial cells. Curr. Opin. Cell Biol. 12, 483–490.[CrossRef][Medline]
Niwa, H., Yamamura, K. & Miyazaki, J. (1991) Efficient selection for high-expression transfectants with a novel eukaryotic vector. Gene 108, 193–199.[CrossRef][Medline]
Ohno, S. (2001) Intercellular junctions and cellular polarity: the PAR-aPKC complex, a conserved core cassette playing fundamental roles in cell polarity. Curr. Opin. Cell Biol. 13, 641–648.[CrossRef][Medline]
Owens, D.W. & Lane, E.B. (2003) The quest for the function of simple epithelial keratins. Bioessays 25, 748–758.[CrossRef][Medline]
Rodriguez-Boulan, E. & Nelson, W.J. (1989) Morphogenesis of the polarized epithelial cell phenotype. Science
245, 718–725.
Rogers, M.B., Watkins, S.C. & Gudas, L.J. (1990) Gene expression in visceral endoderm: a comparison of mutant and wild-type F9 embryonal carcinoma cell differentiation. J. Cell Biol.
110, 1767–1777.
Saotome, I., Curto, M. & McClatchey, A.I. (2004) Ezrin is essential for epithelial organization and villus morphogenesis in the developing intestine. Dev. Cell 6, 855–864.[CrossRef][Medline]
Sato, M., Ozawa, M., Hamada, H., et al. (1985) Cell surface markers to monitor the process of visceral endoderm differentiation from embryonal carcinoma cells: identification of the stage sensitive to high concentration of retinoic acid. J. Embryol. Exp. Morph. 88, 165–182.[Medline]
Sherman, M.I. & Miller, R.A. (1978) F9 embryonal carcinoma cells can differentiate into endoderm-like cells. Dev. Biol. 63, 27–34.[CrossRef][Medline]
Sheth, B., Fesenko, I., Collins, J.E., et al. (1997) Tight junction assembly during mouse blastocyst formation is regulated by late expression of ZO-1+ isoform. Development
124, 2027–2037.[Abstract]
Shibamoto, S., Winer, J., Williams, M. & Polakis, P. (2004) A blockade in Wnt signaling is activated following the differentiation of F9 teratocarcinom
