INTRODUCTION

Epithelial cells cover almost all internal and external body surfaces. All types of epithelia are polarized with apical and basolateral surfaces that have different membrane proteins and thereby different functions. The cells thus compartmentalize the tissues of which they are a part. The cells exist in sheets in which individual cells are interconnected by various types of cell-cell junctions, the regulation of which is at play during development, wound healing, and pathological processes such as cancer or chronic inflammation. Therefore, the control of cell-cell junctions and their disruption and reformation is of interest in several fields.

The LLC-PK₁ cell line (1) offers an advantageous model system for studying this process. The cell line grows subconfluently as islands of adherent cells. Upon reaching confluence, it forms a differentiated epithelial monolayer with functionally intact tight junctions. Treatment of subconfluent cultures with the tumor promoter 12-O-tetradecanoylphorbol-13-acetate (TPA)¹ causes a breakdown of the cell junctions resulting in a more scattered growth pattern. Acute exposure of LLC-PK₁ confluent cell sheets to TPA causes a rapid decrease in the transepithelial resistance (TER) to less than 15% of its initial value (2) because of its effect on one type of cell junction known as tight junctions.

We have been interested in the molecular events in this process and have concentrated much of our efforts on protein kinase C (PKC), the molecular target of TPA. PKC has been shown to be one of the process's key components. Treatment of LLC-PK₁ cells with the PKC inhibitor GF109203X inhibits the effects of TPA (3). PKC mediates calcium-induced tight junction assembly (4), and its inhibition prevents the proper distribution of tight junction-associated proteins such as ZO-1 and cingulin (5). One of the issues in PKC-mediated cell junction regulation is which specific isoform(s) is at play. PKC is not a single protein, as the name implies, but a family of proteins with at least 11 different members (6). Although each member or isoform is encoded by a separate gene, they all consist of conserved regulatory and catalytic regions. The isoforms can be divided into four classes based on regulatory properties. The first class, the conventional PKCs , , and , is regulated by calcium. The second, nonconventional or novel class, is made up of PKCs , , , and , which are not regulated by calcium. The third class, the atypical PKCs  and (/), is not regulated by calcium or TPA. A recently discovered isoform, PKC-µ, represents a fourth class, which is topologically related to the PKC family but possesses a kinase domain more related to calcium/calmodulin-dependent kinases. The complexity of the PKC family and the differential expression of the isoforms suggest that the members serve distinct roles in signal transduction processes. To examine this issue, as it relates to PKC-mediated cell junction regulation in LLC-PK₁ epithelial cells, we have begun to modulate the activities of individual isoforms by exogenously expressing wild type and dominant negative versions of the genes in an inducible expression system. LLC-PK₁ cells express the , , , and  isoforms of PKC, and, in this study, we report that alteration of the activity of the  isoform of PKC modulates the integrity of cell-cell junctions as measured by sensitivity to TPA.

EXPERIMENTAL PROCEDURES

Expression vectors comprising the tetracycline-responsive expression system described by Gossen and Bujard (7) were used to express both wild-type and dominant negative versions of PKC-. cDNA encoding the entire reading frame of wild-type PKC- was excised from its original vector (8) by digestion with EcoRI and recloned into the EcoRI site of the tetracycline-responsive vector pUHD10-3 creating pUHD10-3.

A dominant negative version of the gene was created by in vitro mutagenesis replacing the conserved lysine in the ATP binding domain in position 368 with an alanine. The XcmI-BsiWI fragment spanning the Lys  Ala mutation was then excised from its expression vector to replace the analogous fragment in pUHD10-3 vector creating pUHD10-3_DN.

LLC-PK₁ cells were maintained in minimum essential medium  supplemented with 10% fetal bovine serum (HyClone Laboratories). When tetracycline was present, its concentration was 1 µg/ml. For transfection, cells were harvested by trypsinization, washed, and 5 × 10⁶ cells were resuspended in 0.8 ml of phosphate-buffered saline in an electroporation cuvette (0.4 cm). 20 µg of pUHD15-1, which encodes the bacterial transactivator tTA, along with 5 µg of a plasmid conferring resistance to hygromycin, were added to the suspension. This was subjected to one pulse (300 V; 250 microfarads) in a Bio-Rad electroporator. Cells were then plated onto three 10-cm tissue culture plates, incubated for 1 day, then adjusted to 400 µg/ml hygromycin. 2-4 weeks later, drug-resistant clones were isolated, expanded, and analyzed by transient transfection for tetracycline-responsive expression of a luciferase reporter gene cloned into the pUHD10-3 expression vector. One of the resulting subclones, LLC-PK₁tTA, was selected for subsequent constructions, which were to stably transfect the PKC- expression vectors described above. This was performed by the same electroporation procedure utilizing 5 µg of pSVzeo (Invitrogen) and 1 mg/ml zeocin for selection. Expression of wild-type and dominant negative PKC- was assessed by Western analysis.

Cells were harvested by trypsinization followed by centrifugation. Cell pellets were then lysed in Tris-Cl buffer containing 0.1% sodium dodecyl sulfate and proteinase inhibitors. DNA was sheared to reduce viscosity, and the protein concentration was determined with Bio-Rad protein assay dye reagent. 50 µg of total protein was mixed with an equal volume of 2 × sample buffer (0.125 M Tris-Cl, pH 6.8, 4% sodium dodecyl sulfate, 20% glycerol, 10% 2-mercaptoethanol, 0.002% bromphenol blue). Samples were heated in a boiling water bath for 5 min and then loaded onto a denaturing 6% polyacrylamide electrophoresis gel.

After separation, proteins were electrophoretically transferred to a nitrocellulose membrane. The filters were first treated with Tris-buffered saline (50 mM Tris-Cl, pH 7.5, 0.15 M NaCl) with 0.05% Tween 20 (TBST) containing 3% nonfat milk. This was followed by hybridization in 5 ml of TBST with 1% bovine serum albumin and 1 µg of anti-PKC- antibody (Upstate Biotechnology, Inc.) for 5 h at room temperature. The blots were subsequently washed in TBST and then incubated with anti-mouse peroxidase-conjugated antibody (Amersham) for 1 h. After washing in TBST, the blots were developed using the enhanced chemiluminescence Western blotting system (Amersham).

Electrical equipment and protocols for electrical measurement have been described before (9). Briefly, 1 × 10⁶ cells were seeded onto Falcon 3102 filter rings (4.2 cm²). On day 4, after which electrical resistance across the cell sheet had stabilized, cells were refed with fresh media, incubated for 1-2 h, and then their initial TER was measured. Cell sheets were then refed with media containing 1.0 × 10⁸ M TPA, and electrical resistance across the cell sheet was determined through the course of resistance decreases. Measurements were done in triplicate, and all results are representative of at least three experiments.

For assessment of cell scattering 3 × 10³ cells were seeded in 24-well plates in 1 ml of medium containing 10⁸ M TPA. Growth patterns were followed by light microscopy and photographed 3 days after seeding.

RESULTS

Since we have noted previously that TPA treatment of LLC-PK₁ cells is associated with a translocation of PKC- from the cytosolic fraction of the cells to the membrane fraction (9), we decided to modulate the activity of PKC- by exogenous expression of the gene. Because randomly isolated sublines can vary from the parental in any given property, there is consequent difficulty in ascribing any phenotypic change to the expression of an exogenous gene. We chose to avoid the problems of clonal variation by utilizing the tetracycline-responsive expression system of Gossen and Bujard (7) to overexpress PKC-. In the first step, we cotransfected into LLC-PK₁ cells a plasmid expressing tTA and a plasmid conferring hygromycin resistance. After screening drug-resistant clones for tetracycline-repressible expression of tTA, we selected a subline, LLC-PK₁tTA, for additional experiments. Next, pUHD10-3 PKC- was transfected into LLC-PK₁tTA using pSVzeo (Invitrogen) as a coselectable marker. 20 drug-resistant clones were analyzed for expression of PKC-, and one was found to have elevated levels. Fig. 1 shows Western analysis of extracts of cells cultured with and without tetracycline, expressing exogenous PKC-. Parental cells exhibited a single protein of approximately 75 kDa. LLC-PK₁ cells grown in the absence of tetracycline exhibited PKC- levels that were 10-20 times higher. The same subline grown in the presence of tetracycline showed levels that were almost equal to levels in the parental LLC-PK₁ tTA line. Phenotypically, the cells resemble the parental line, growing as islands of coherent cells. Upon reaching confluence, both the transfectant and the parental line remain a single layer and differentiate into a polar epithelial-like cell sheet with apical and basolateral surfaces. Both lines form dome-like or cystic structures as a result of the vectorial transepithelial transport and the tight junctions of the cell sheet. However, as seen in Fig. 2, not only do domes collapse when treated with TPA, but the transfected cells rapidly round up and begin to detach from the dish. When expression of the exogenous PKC- was down-regulated by the addition of tetracycline, the normal dome collapse occurred on TPA treatment but not the cell rounding and detaching. The disruption of tight junctions was assessed quantitatively in measurements of TER decreases on TPA treatment. As seen in Fig. 3, cells overexpressing PKC- showed a more rapid response to all concentrations of TPA tested compared with wild type LLC-PK₁. However, when these cells were grown in the presence of tetracycline, TER decreases matched that of the parental line. The presence of tetracycline had no effect on TER decreases in wild-type LLC-PK₁ (data not shown).

Western analysis of LLC-PK₁.

Phase-contrast micrographs of LLC-PK₁ and its parental line before and after TPA treatment.

7

11

10

9

8

7

8

To generate a subline of LLC-PK₁ cells expressing an inducible dominant negative version of PKC-, we transfected pUHD10-3_DN into LLC-PK₁tTA by the same procedure as described above. Fig. 4 shows the Western analysis of four clones that were isolated. Each showed an exogenous protein running 4-5 kDa below the endogenous PKC-. In all four clones, the exogenous protein was down-regulated significantly by tetracycline. PKC is first synthesized as an inactive unphosphorylated polypeptide. It is first phosphorylated by an unknown PKC kinase and then undergoes two autophosphorylations to generate the active species (10). The kinase-dead dominant negative protein is obligatorily unable to autophosphorylate and therefore displayed a faster electrophoretic migration compared with the endogenous protein.

Western analysis of LLC-PK_1DN.

Physiologically, the effect of PKC-_DN was the opposite of its wild-type counterpart. LLC-PK_1DN cells grown in the absence of tetracycline showed a much slower response to 10⁸ M TPA than did parental counterparts. As seen in Fig. 5, at 1 h of treatment, TERs of LLC-PK_1DN were 80% of their initial value, whereas the parental cells had decreased to 25%. The degree of tetracycline responsiveness of the system is also seen in electrical measurements with LLC-PK_1DN cells grown in the presence of tetracycline which showed responses to TPA nearly identical to those of parental LLC-PK₁tTA.

TER of LLC-PK_1DN as a function of duration of exposure to 10⁸ M TPA.

We next assessed the effects of PKC-_DN expression on growth patterns of the line. LLC-PK₁, like most epithelial lines, grow as islands of coherent cells. Treatment of subconfluent cultures with TPA disrupts the cell junctions that produce this compact architecture. The result is a more fibroblastic growth pattern with individual cells migrating away from the island and a more scattered appearance. Fig. 6 shows the results of TPA treatment of LLC-PK_1DN cells grown in the presence and absence of tetracycline. With expression of PKC-_DN turned off, treatment of cells with TPA induced a scattered growth pattern identical to that of wild type cells treated with TPA (panel B). However, with the expression of PKC-_DN turned on, TPA-induced cell scattering was completely inhibited; that is, growth patterns were identical to either wild type LLC-PK₁ or LLC-PK_1DN in the absence of TPA.

Phase-contrast micrographs of LLC-PK_1DN.

DISCUSSION

Protein phosphorylation is the primary mechanism for controlling diverse and complex cellular processes that affect structure, growth, and differentiation. Among the many kinases that are at play in these events is PKC, which often plays a pivotal role in the signal transduction process. Additionally, because PKC is the major receptor for tumor-promoting phorbol esters such as TPA, the long history of tumor promotion research has led to a wealth of information indicating that the modulation of PKC activity is a key step in tumorigenesis as well. The development of an invasive tumor involves a disruption of normal cell-cell junctions that serve to maintain tissue architecture. Additionally, more recent studies have indicated that the mechanism of action of many growth factor receptors involves activation of PKC. This occurs via phospholipase C, which produces diacylglycerol, which binds to and activates the enzyme in the same manner as TPA. The receptor for scatter factor or hepatocyte growth factor, c-Met, is an example of such a receptor, indicating that PKC plays a role in producing a response to this cytokine.

Soon following the initial discovery of PKC, the activity was found to consist of more than one species. With the advent of techniques in molecular biology, the family of isoforms has since grown to include 11 members (6). Differences in tissue distribution, subcellular distribution, and substrate specificity suggest that there is a divergence of function among the isoforms. However, despite the abundance of PKC research, little information is available on the roles of individual isoforms. Since inhibitors of the enzyme are not specific for individual isoforms, we and others have begun to address this issue by the use of exogenous expression studies. This technique, which utilizes both wild-type and dominant negative versions of the gene to modulate the activity of specific individual isoforms, is beginning to indicate which isoforms are involved in particular processes.

We have begun examining the role of the  isoform in the disruption of cell junctions. Calcium plays an established role in the maintenance of various cell junctions including tight junctions (4, 11), and PCK is the major calcium-dependent isoform expressed in LLC-PK₁ cells. Its activity is essential for v-ras transformation of keratinocytes (12), and we have previously noted that, upon TPA treatment, PKC- translocates from the cytosolic to the membrane fraction. This occurs in a time-dependent manner, which correlates with increased leakiness of tight junctions. Furthermore, in cells chronically treated with TPA, which form uneven cell sheets of single and multilayered areas, PKC- stays up-regulated in areas in which there is multilayering, which also is where tight junctions are most leaky (9). Ellis et al. (13) report that an LLC-PK₁ subline that decreases and then rapidly recovers its TER in response to TPA also rapidly down-regulates its PKC activity, whereas a subline whose TER does not recover in the presence of TPA does not down-regulate its PKC.

Our results show that a direct correlation exists between activity levels of PKC- and the sensitivity of LLC-PK₁ cell junctions to TPA; that is, up-regulation by expression of exogenous wild-type PKC- enhances sensitivity, and down-regulation by expression of the dominant negative form decreases sensitivity. Although in vitro studies of PKC show a lack of specificity toward certain substrates, several literature studies have reported that the technique of exogenous expression of individual members of the PKC family produces results specific for individual isoforms. For example, Li et al. (14) expressed six different isoforms of PKC in 32D cells and reported that only the expression of the  isoform rendered the cells sensitive to TPA-induced differentiation. In experiments with NIH 3T3 cells, the  isoform inhibited cell growth, whereas the  isoform rendered the cells tumorigenic (15). Similarly, Baier-Bitterlich et al. (16) reported that only the  isoform was competent in stimulation of AP-1 activity in T-lymphocytes. This specificity might be explained in part by the fact that isoforms are compartmentalized differently in cells, conferring specificity to given substrates by spatial accessibility. For example, PKC- has been localized to the cytoskeleton of HL-60 cells (17), and PKC- has been localized by immmunoelectron microscopy in Madin-Darby canine kidney epithelia to the region of the zonula occludens and/or zonula adherens (18). Transfection of endothelia with antisense PKC- blocked the phorbol ester-induced increase in tight function permeability normally seen in these cell sheets (19). Similarly, overexpression of PKC- yields an endothelial cell sheet with a dramatically increased effect of phorbol ester on tight junction permeability (20). We have begun to examine other isoforms of PKC in terms of the effect they have on regulating cell junctions in LLC-PK₁ and are finding similar results.

Our work represents a significant step toward identifying the relevant isoforms of PKC in regulating cell junctions. More work is necessary to identify the mechanism of this regulation. Many putative targets of PKC have been identified which are associated with cell junctions. Among these are c-Raf (21), vinculin (22), talin (22), MARCKS (23), glycogen synthase kinase-3 (24) and focal adhesion kinase (25). PKC-mediated modulation of any one of these might conceivably play a role in reorganizing cytoskeletal structures during TPA-induced cell migration and disruption of cell junctions.