INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

PROTEIN KINASE C (PKC) represents a family of serine-threonine protein kinases that are expressed to varying degrees in all mammalian cells and that are implicated in numerous biological events ranging from cell growth and differentiation to cytoskeletal organization, membrane trafficking, ion transport, and cell-cell communication (6, 19, 33). To date, at least 11 isoforms of PKC are known. These isozymes exhibit different patterns of tissue, cell, and subcellular distribution, suggesting that each subserves distinct intracellular functions (24). PKC isozymes are subclassified into three main groups depending on their requirements for 2,3-diacylglycerol (DAG) and Ca²⁺. Tumor-promoting phorbol esters such as phorbol 12-myristate 13-acetate (PMA) are well known to translocate, activate, and eventually downregulate conventional (DAG and Ca²⁺ dependent) and novel (DAG dependent, Ca²⁺ independent) isoforms by virtue of their high affinity for the DAG binding site of these PKC isozymes. The demonstration that PMA experimentally affects a given cellular property is often taken as evidence for the regulatory involvement of one or more classical or novel PKC isoforms in that process. However, because neither the specific isoform nor its downstream target(s) can be inferred from such observations, such data can yield only limited mechanistic insight into the precise role of PKC.

PMA complexly influences a number of fundamental properties of epithelial cells, including vectorial transport. For example, a number of investigators showed that PMA affects epithelial Cl secretion, the physiological process that accounts for mucosal surface hydration and the transport event whose disregulation accounts for diseases such as cystic fibrosis and secretory diarrhea. In several epithelial cell lines, PMA progressively inhibits cAMP-regulated Cl secretion (3, 30, 36, 43). The mechanism underlying this observation has not been clearly established. Initial studies in T84 cells suggested that PMA inhibition of Cl secretion might involve inhibition of apical membrane Cl conductance via reduced gene expression of the cystic fibrosis transmembrane conductance regulator (CFTR) (43). However, downregulation of CFTR mRNA and channel function by PMA was shown to lag several hours behind, and thus cannot account for, the loss of Cl secretion. A closer temporal correlation with inhibition of two basolateral transport sites, the basolateral Na⁺-K⁺-2Cl cotransporter (NKCC1) and the basolateral K⁺ conductance, was reported by several groups (3, 29, 30, 36). We showed that PMA induces a loss of binding sites for the NKCC1 inhibitor bumetanide (29) before any detectable change in steady-state levels of NKCC1 mRNA or protein (17) and suggested that PMA may reduce the surface expression of NKCC1. Moreover, we recently found (32) that PMA decreases the surface expression of another basolateral membrane transporter not directly involved in Cl secretion, specifically, a facilitated nucleoside transporter. A unifying explanation for these findings would be that PMA induces a generalized increase in the rate of basolateral endocytosis by PKC, resulting in the increased internalization of multiple membrane proteins including those involved in transmembrane transport.

Plasma membranes are remodeled continually by endocytosis, a tightly regulated multistep process that has been shown in diverse cell types to be profoundly influenced by PMA and, hence, presumably, by PKC (2, 11, 27). In epithelial cells, the plasma membrane is subdivided into distinct apical and basolateral domains and the process of membrane cycling appears to be regulated in a polarized fashion (9). Recent studies have shown that changes in fluid-phase endocytosis in epithelial cells may indeed correlate with changes in the surface expression of specific transport proteins, although to date, this has appeared to apply primarily if not exclusively to the apical membrane. For example, in T84 cells, apical but not basolateral endocytosis is inhibited by cAMP, an event that parallels a rapid increase in CFTR Cl channels (7, 8). Although PMA was reported to exert no effect on apical membrane endocytosis in T84 cells (6a), basolateral membrane effects have not been addressed. However, in Madin-Darby canine kidney epithelial cells, Mostov and co-workers (11) showed that activation of PKC increased basolateral endocytosis, transcytosis of membrane markers from the basolateral to the apical domain, and apical recycling.

Many fundamental questions about the role of PKC and specific PKC isoforms in the regulation of endocytosis remain unanswered. PMA exerts profound effects on cell shape and the organization of actin (16, 42). Although there are numerous reports linking cortical F-actin networks to various forms of endocytosis, the precise relationship between the effects of PKC on cytoskeletal organization and membrane traffic has yet to be established. In the present study, we examine the role of PKC in fluid-phase endocytosis in polarized T84 cells and address its isoform selectivity and potential cytoskeletal dependence. In particular, we examine the potential roles of F-actin and the F-actin cross-linking protein myristoylated alanine-rich C-kinase substrate (MARCKS) (21) in the regulation of endocytosis by PKC.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Cell culture. Human T84 intestinal epithelial cells obtained from American Type Culture Collection (Manassas, VA) and Dr. K. Barrett (University of California, San Diego) were grown to confluence at pH 7.4 in 162-cm² flasks (Corning Costar) with a 1:1 mixture of Dulbecco's modified Eagle's medium and F-12 nutrient mixture (Ham) supplemented with 6% fetal bovine serum (FBS), 15 mM HEPES, 14.3 mM NaHCO₃, and antibiotics-antimycotic. Flasks were passaged weekly and fed every 3 days. Cell monolayers for experiments were grown to confluence on collagen-coated Transwell inserts (Corning Costar, MA). Cells were fed every 3 days and used after stable transepithelial electrical resistance was achieved, ~7-14 days after plating.

Fluid-phase endocytosis. Uptake of FITC-dextran (mol wt 12,000, 0.73 mol fluorescein/mol dextran) from either the apical or the basolateral aspect of confluent T84 monolayers grown on collagen-coated permeable supports (4.7 cm², 3.0-µm pore size) was measured. Monolayers were incubated with 15 mg/ml of FITC-dextran in either apical or basolateral medium for 6 min at 4 or 37°C. Monolayers were then washed extensively with ice-cold HEPES-phosphate-buffered Ringer solution (HPBR) containing (in mM) 135 NaCl, 5 KCl, 3.33 NaHPO₄, 1 CaCl₂, 1 MgCl₂, 10 glucose, and 5 HEPES at pH 7.4. The membrane was excised from its plastic support, inserted into 430 µl of distilled water, sonicated for 20 s (setting 3, model 440 Sonic Dismembrator, Fisher Scientific) and spun twice for 5 min at 14,000 g. Fluorescence of the clear supernatant was measured by using a fluorimeter (Spex DM3000) with excitation and emission set at 495 and 565 nm, respectively. Nonspecific (temperature insensitive) binding represented at most 50% of total binding in control monolayers and was not affected by PMA treatment. Because it would have been impractical and costly to include parallel experiments at both 4 and 37°C for each experimental condition, the data reported here represent total uptake of FITC-dextran.

Peptide synthesis. A 25-amino acid tetra-serine (tet-Ser) peptide containing 4 serine residues and corresponding to the phosphorylation site domain (PSD) of MARCKS was synthesized by SynPep (Dublin, CA) and was >98% pure as determined by HPLC and mass spectrographic analysis. A tetra-alanine (tet-Ala) peptide that shares the same sequence as the tet-Ser, except that the four serines are replaced by alanines, was also synthesized as a negative control peptide. The sequences of the peptides are tet-Ser peptide: KKKKKRFSFKKSFKLSGFSFKKNKK and tet-Ala peptide: KKKKKRFAFKKAFKLAGFSFKKNKK. To render the peptides cell permeant, they were myristoylated at the NH₂ terminus.

Fluorescent microscopy. Fluorescent staining of F-actin was performed as previously described (22, 39). Polarized T84 cells grown to ~90% confluence on collagen-coated Anocell inserts (0.33 cm², 0.2-µm pore size, Whatman) were rinsed in PBS, fixed in 3.7% formaldehyde, and permeabilized in 20°C acetone. F-actin was detected by staining the cells with rhodamine-phalloidin for 1 h at room temperature. After being washed in PBS, filters were mounted on microscope slides in Vectashield mounting medium. Confocal images were acquired using a Zeiss inverted microscope equipped with MRC-1024 and Lasersharp software (Bio-Rad).

Subcellular fractionation. T84 cells grown to confluence on collagen-coated permeable supports were washed with ice-cold PBS three times and scraped into cold homogenization buffer (HB) containing (in mM) 20 Tris · HCl (pH 7.5), 250 sucrose, 4 EDTA, and 2 EGTA and complete protease inhibitor cocktail. The cells were homogenized on ice with 25 strokes of a glass tissue homogenizer. The resulting homogenate was ultracentrifuged at 86,000 g for 50 min at 4°C (TLA 45 rotor, TL-100 Ultracentrifuge, Beckman). The supernatant was designated as the cytosolic fraction. The pellet was resuspended in 800 µl of HB containing 0.5% (vol/vol) Triton X-100 by brief sonication and incubated in ice for 30 min. At the end of the incubation period, the samples were centrifuged at 14,000 g for 20 min at 4°C. The resulting supernatant was designated as the membrane fraction.

Immunoprecipitation of MARCKS. Equal amounts of protein (~2.5 mg/sample) extracted in HB were further solubilized in 1% Triton X-100, 0.5% NP-40, and 60 mM n-octyl -D-glucopyranoside. These samples were first precleared with 10% (vol/vol) protein A agarose for 30 min at 4°C and subjected to immunoprecipitation with 5 µg of monoclonal anti-human MARCKS overnight at 4°C. After overnight incubation, 10% (vol/vol) protein A agarose was added to each sample, incubated for 1.5 h at 4°C, and centrifuged for 2 min at 14,000 g. The pellet with bound MARCKS was washed three times with HB containing 1% Triton X-100 and 0.5% NP-40 and finally with HB containing no detergents. The pellets were mixed with 40 µl of Laemmli's sample buffer containing 5% (vol/vol) -mercaptoethanol and boiled for 5 min. After brief vortexing, the pellets were centrifuged at 14,000 g for 2 min and the supernatants were processed by SDS-PAGE as described in Gel electrophoresis and Western blotting.

Gel electrophoresis and Western blotting. Equal amounts (~50 µg/sample) of protein, as determined by the Bradford assay, were combined with Laemmli's sample buffer containing 5% (vol/vol) -mercaptoethanol and boiled for 5 min. Proteins were separated by electrophoresis on 10% SDS-PAGE gels and transblotted to nitrocellulose membranes. The protein-bound nitrocellulose sheets were first incubated for overnight at 4°C in a blocking buffer containing 20 mM Tris (pH 7.5), 500 mM NaCl, and 5% nonfat dry milk. Nitrocellulose sheets were then incubated with monoclonal anti-human MARCKS (1 g/ml) in the blocking buffer for 1 h at room temperature and rinsed for 30 min with a wash buffer containing 20 mM Tris, pH 7.5, 500 mM NaCl, and 0.2% Tween-20. Finally, the membranes were incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG antibody (1:3,000 dilution) for 1 h at room temperature and washed for 30 min with agitation, with the wash buffer changed every 5 min. MARCKS bands were visualized with enhanced chemiluminescence (ECL) detection reagents.

Materials. Tissue culture reagents were purchased from Life Technologies. Gel electrophoresis and Western blotting reagents were from Bio-Rad, with the exception of ECL detection reagent, which was purchased from Amersham. Complete protease inhibitor cocktail was from Boehringer Mannheim, and rhodamine-phalloidin was from Molecular Probes. Monoclonal antibodies against human MARCKS were purchased from Upstate Biotechnology. Antibodies against PKC- and -₁ were obtained from Sigma, anti-PKC- and - were obtained from Santa Cruz Biotechnology, and anti-PKC- was obtained from Alexis Biochemicals. Protein A agarose was from Life Technologies, and Vectashield mounting medium was purchased from Vector Laboratories. Calphostin C, myristoylated PKC pseudosubstrate, Gö-6976, rottlerin, and myristoylated coenzyme A were obtained from Calbiochem. Gö-6850 was purchased from Alexis Biochemicals, and staurosporine was from Sigma. All other chemicals were from Sigma.

Statistical analysis. Data are reported as means ± SE. Data were analyzed by Student's t-test for unpaired variates and by two-way ANOVA, where appropriate, with P < 0.05 considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

PKC selectively stimulates basolateral membrane fluid-phase endocytosis. Uptake of FITC-dextran from either the apical or the basolateral aspect of confluent T84 monolayers was time dependent at 37°C and completely blocked at 4°C (Fig. 1); by fluorescent microscopy, FITC-dextran was visualized in punctate distribution consistent with vesicular compartmentalization (not shown). On the basis of these experiments, a 6-min uptake period at 37°C was chosen. Treatment of monolayers with PMA increased the rate of basolateral but not apical uptake of FITC-dextran in a time- and dose-dependent fashion. At 10 nM PMA, a time-dependent increase in basolateral endocytosis is seen (Fig. 2A) that reaches a peak at ~60 min, with the earliest increase detectable at 15 min after exposure. At higher concentrations of PMA, substantial and dose-dependent increases in basolateral endocytosis were detectable within 10 min (Fig. 2B). The increase in endocytosis elicited by PMA was not uniformly sustained, an effect that was particularly evident at higher concentrations of PMA. For example, the peak stimulatory effect of 100 nM PMA was observed at 30 min (202 ± 24% control) and then declined thereafter such that after 120 min the basolateral uptake was 128 ± 26% control. We then examined whether the effect of PMA on basolateral endocytosis was exerted via PKC. As shown in Fig. 3, two additional phorbol esters also stimulated basolateral FITC-dextran uptake, whereas the PKC-inactive -isomer of PMA exerted no effect. Additionally, the DAG analog 1-oleoyl-2-acetylglycerol as well as the nonphorbol PKC agonist bryostatin 1 also markedly enhanced basolateral endocytosis.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Uptake of FITC-dextran (FD) across apical and basolateral aspect of confluent T84 monolayers. Time course of FITC-dextran uptake measured at 4 and 37°C from either apical or basolateral aspect of confluent T84 monolayers grown on 4.7-cm2 permeable supports is shown. Monolayers were incubated with 15 mg/ml FITC-dextran in either apical (A) or basolateral (B) medium for time indicated (n = 3 at each time point) at either 4 or 37°C.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Phorbol 12-myristate 13-acetate (PMA) selectively stimulates basolateral membrane fluid-phase endocytosis of T84 cells in time- and dose-dependent fashion. Confluent T84 monolayers grown on 4.7-cm2 permeable supports were treated with PMA, and the rate of FITC-dextran uptake was measured from either apical or basolateral aspect of confluent monolayers as described in EXPERIMENTAL PROCEDURES. A: monolayers were exposed to 10 nM PMA for time indicated (total n > 100), and 6-min uptake of FITC-dextran was measured and expressed as percentage of control. PMA increased basolateral but not apical uptake of FITC-dextran in time-dependent fashion (P < 0.001 by ANOVA). Peak effect was shown after 60-min exposure to 10 nM PMA. B: monolayers were exposed to indicated concentrations of PMA for 10 min. PMA dose-dependently increased basolateral uptake of FITC-dextran.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Activators of protein kinase C (PKC) enhance basolateral fluid-phase endocytosis. Monolayers were treated with 2 phorbol esters known to activate PKC as well as nonphorbol PKC agonists bryostatin 1 (bryo-1) and 2,3-diacylglycerol (DAG) analog 1-oleoyl-2-acetylglycerol (OAG), and uptake of basolateral FITC-dextran was measured. A 60-min exposure to 100 nM phorbol dibutyrate (PdBu) and phorbol didecanoate (PdDc) enhanced basolateral uptake (n = 5 and n = 12, respectively), whereas PKC-inactive -isomer of PMA exerted no effect (n = 3). Additionally, 60-min exposure to 100 nM bryo-1 (n = 5) and 20 µM OAG (n = 3) markedly stimulated basolateral endocytosis. * P < 0.0001.

PMA-stimulated basolateral endocytosis occurs via activation of PKC-. PMA-stimulated endocytosis was attenuated by the general PKC inhibitors staurosporine, calphostin C, and a myristoylated pseudosubstrate inhibitor (Fig. 4). Isoform selectivity to this effect was suggested by the finding that the conventional PKC inhibitor Gö-6976 (26) and the PKC--selective inhibitor rottlerin (27) did not block endocytosis at concentrations up to 10 µM. At 5 µM, Gö-6976 should completely inhibit PKC- (IC₅₀  2-6 nM for conventional isoforms) but exerts no effect against novel PKC isozymes (47). At 10 µM, rottlerin is rather specific for the novel isoform PKC- (IC₅₀  3-6 µM), weakly active against conventional isoforms (IC₅₀ ~ 30-40 µM), and inactive against PKC- (IC₅₀ > 80 µM) (20a). Moreover, the bisindoylmaleinimide Gö-6850, which inhibits both conventional and novel PKC isoforms at nanomolar concentrations (47), nearly completely inhibited PMA-stimulated endocytosis. Because Gö-6976 was completely ineffective at blocking PMA-induced endocytosis, it is unlikely that a conventional PKC isoform is responsible for the PMA effect. Similarly, PKC- is unlikely to be involved, because rottlerin failed to inhibit the PMA effect. In contrast, the ability of Gö-6850 to block the PMA effect strongly implicates PKC- as the isoform underlying PMA-stimulated endocytosis.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   General PKC inhibitors and the conventional and novel PKC inhibitor Gö-6850 attenuate PMA-stimulated endocytosis. Monolayers were treated with various PKC putative inhibitors for 1 h followed by 10 nM PMA for 1 h in presence of inhibitors. Basolateral uptake of FITC-dextran was subsequently measured and compared with control. Nonselective PKC inhibitors including 1 µM calphostin C, 100 µM staurosporine, and 100 µM myristoylated PKC pseudosubstrate peptide significantly attenuated PMA-stimulated basolateral endocytosis (each n > 3, P < 0.005). Conventional and novel PKC isoform inhibitor Gö-6850 attenuated PMA-elicited endocytosis at a concentration of 5 µM (n = 3, P < 0.01). However, conventional (but not novel) PKC inhibitor Gö-6976 enhanced PMA-stimulated endocytosis at a concentration of 5 µM (n = 5, P < 0.01). PKC- inhibitor rottlerin and atypical PKC inhibitor myristoylated coenzyme A (myr-coA) did not attenuate PMA-stimulated endocytosis at concentrations up to 10 µM. None of the agents that inhibited PMA-elicited endocytosis altered FITC-dextran uptake in absence of PMA (not shown). * P < 0.0001 for difference from control.

View larger version (26K): [in this window] [in a new window]   Fig. 5.   PMA induces membrane translocation of PKC-. Cytosolic (C) and membrane (M) fractions of T84 homogenates were obtained as described in EXPERIMENTAL PROCEDURES. Fractions were subjected to SDS-PAGE and blotted with PKC isoform-specific antibodies to examine redistribution of different PKC isoforms in response to PMA. Four different PKC isoforms, , , , and , were detected in T84 cells and were found to be dominantly associated with cytosolic fraction in inactive forms. PKC-1 was not found in these cells by Western blot. In response to 100 nM PMA, PKC- translocated to membrane fraction within 10 min, a time consistent with earliest detectable effect of PMA on endocytosis. Translocation of other PKC isoforms was not evident. Representative experiment shown was repeated 3 times.

PMA-stimulated endocytosis is associated with remodeling of F-actin cytoskeleton. Fluorescent imaging of rhodamine-phalloidin-labeled actin filaments in PMA-treated cells revealed attenuation of stress fibers and condensation of staining around the periphery of the cells, changes that were confined to the region below the perijunctional actin ring in monolayers viewed en face (Fig. 6). No changes were evident at the level of F-actin within the core of microvilli. Because this pattern resembled that previously reported for the microfilament disassembler cytochalasin D (28), we examined whether cytochalasin D also affected fluid-phase endocytosis. Treatment of T84 monolayers with 20 µM cytochalasin D (but not the F-actin-inactive analog chaetoglobosin in equimolar concentrations; not shown) enhanced basolateral uptake of FITC-dextran. The enhancement of endocytosis by PMA and cytochalasin D together was no greater than when maximal stimulatory concentrations of either agent were used (Fig. 7); at lower concentrations, however, there appeared to be additivity (not shown), observations that suggest a common target of action (e.g., the actin cytoskeleton). To further address whether PMA-stimulated endocytosis was associated with reorganization of the actin cytoskeleton, we used two structurally distinct F-actin stabilizers, phalloidin and jasplakinolide. In monolayers loaded by overnight incubation in medium containing 10 µM phalloidin and in monolayers treated with 1 µM jasplakinolide for 1 h, stimulation of FITC-dextran uptake by PMA was markedly attenuated (Fig. 7). Neither jasplakinolide nor phalloidin prevented actin reorganization induced by cytochalasin D, and neither agent blocked cytochalasin D-induced endocytosis (not shown).

View larger version (100K): [in this window] [in a new window]   Fig. 6.   PMA remodels basolateral F-actin. Confocal images of rhodamine-phalloidin labeled actin filaments were acquired from T84 monolayers as described in EXPERIMENTAL PROCEDURES. Three fields from each slide were chosen at random, and vertical (z) sections were collected at a step size of 1 µm. Cell heights of each monolayer were measured from tip of microvilli to basal membrane on the basis of staining of rhodamine-phalloidin. A: mean cell height of control monolayers was 31.9 ± 3.2 µm (n = 9). Basal stress fibers appeared as randomly dispersed homogeneous filaments at 6.0 ± 0.7 µm above basal membrane. B: after 1-h exposure to 100 nM PMA, mean cell height remained unchanged at 34.2 ± 3.4 µm (n = 8). At optical section of 6.6 ± 0.9 µm above basal membrane, F-actin was displaced toward periphery of cells and basal stress fibers were significantly attenuated. Representative images are shown. PMA-treated cells could be distinguished from control monolayers on the basis of these features in 100% of monolayers examined.

View larger version (13K): [in this window] [in a new window]   Fig. 7.   Effect of chemical manipulation of F-actin on basal and PMA-stimulated basolateral endocytosis. Monolayers were treated with 20 µM cytochalasin D (cyto D) for 1 h, and basolateral uptake of FITC-dextran was compared with control and PMA (100 nM, 1 h)-treated monolayers. Cytochalasin D alone enhanced basolateral uptake of FITC-dextran to a degree comparable to effect of PMA. In contrast, F-actin stabilizers phalloidin (phall) and jasplakinolide (jas) blocked PMA stimulation of endocytosis. Overnight treatment with 10 µM phalloidin and 1-h treatment with 1 µM jasplakinolide markedly attenuated PMA stimulation of FITC-dextran uptake (P < 0.0005 compared with PMA treatment). * P < 0.0001 for difference from control.

PMA and cytochalasin D elicit biochemical redistribution of MARCKS. MARCKS is a widely distributed PKC substrate that has been implicated in secretion and membrane trafficking in a number of cell types (1). It is a membrane-associated actin cross-linking protein (21) that is known to translocate from a membrane to a cytosolic location in response to PKC-dependent phosphorylation, an event associated with cytoskeletal remodeling (1). We therefore wondered whether stimulation of endocytosis by PMA and cytochalasin D was associated with effects on MARCKS dynamics in T84 cells. Immunoprecipitation and Western blotting of T84 cell lysates using a monoclonal antibody raised against recombinant human MARCKS detected a single band corresponding to a molecular mass of ~65 kDa after SDS-PAGE (Fig. 8). In control monolayers, MARCKS was found approximately equally distributed between the membrane and the cytosolic fraction of homogenized T84 cells. In response to PMA, however, MARCKS was observed to translocate from the membrane to the cytosolic fraction. Over this same time period, PKC- translocated from the cytosolic to the membrane fraction. Translocation of MARCKS was blocked by Gö-6850, whereas the conventional isoform inhibitor Gö-6796 and PKC--selective inhibitor rottlerin failed to do so (Fig. 9). Cytochalasin D also induced translocation of MARCKS in response to PMA (Fig. 10), consistent with its ability to enhance basolateral endocytosis as shown in Fig. 7.

View larger version (12K): [in this window] [in a new window]   Fig. 8.   PMA releases membrane-bound myristoylated alanine-rich C kinase substrate (MARCKS) into cytosol. A: immunoprecipitation of MARCKS from cell lysates using a monoclonal antibody against recombinant human MARCKS identified a single protein of mol wt ~ 65,000. B: T84 monolayers were exposed to 10 nM PMA for time indicated (min), and subcellular fractions were obtained as described in EXPERIMENTAL PROCEDURES. Both cytosolic (C) and membrane (M) fractions were subjected to SDS-PAGE and probed with antibodies to PKC- and MARCKS. In control monolayers, MARCKS was equally distributed between the membrane and the cytosolic fraction of homogenized T84 cells. In response to 100 nM PMA, however, MARCKS translocated from membrane to cytosolic fraction and by 60 min, MARCKS completely associated with cytosolic fraction. Over this same time period, PKC- migrated to membrane fraction.

View larger version (23K): [in this window] [in a new window]   Fig. 9.   Gö-6850, but not Gö-6976 or rottlerin, blocks PMA-induced translocation of MARCKS. Monolayers were pretreated with a panel of PKC isoform-specific inhibitors, and translocation of MARCKS was then determined in response to 100 nM PMA. A: MARCKS was immunoprecipitated from subcellular fractions of T84 homogenates and blot obtained from SDS-PAGE was probed with anti-human MARCKS. A 1-h pretreatment with 5 µM Gö-6850 blocked PMA-induced translocation of MARCKS from membrane (M) to cytosolic (C) fraction. However, conventional PKC inhibitor Gö-6976 and PKC- inhibitor rottlerin failed to block MARCKS translocation. B: densitometric analysis of data from A, expressing percentage of total MARCKS found in cytosolic fractions (cyto MARCKS). Similar data were obtained in duplicate experiment.

View larger version (19K): [in this window] [in a new window]   Fig. 10.   F-actin disassembler cytochalasin D enhances PMA-induced translocation of MARCKS. Monolayers were pretreated with 20 µM cytochalasin D for 1 h before subsequent exposure to PMA. MARCKS was immunoprecipitated from cytosolic (C) and membrane (M) fractions of cell homogenates and detected by SDS-PAGE and Western blotting with anti-human MARCKS antibody. A: in response to 20-min exposure to 100 nM PMA, MARCKS translocated from membrane to cytosol as shown in previous figures. A 1-h treatment with 20 µM cytochalasin D alone also showed effects similar to those of PMA on MARCKS translocation. B: distribution of MARCKS between cell fractions was determined by densitometric analysis, and amount of cytosolic MARCKS after each treatment is expressed in percentage of total. Similar results were confirmed in duplicate experiment.

A myristoylated peptide corresponding to PSD of MARCKS inhibits PMA-stimulated basolateral endocytosis. To determine whether the effects of PMA on MARCKS could be related to its effect on basolateral endocytosis, we used a synthetic 25-amino acid peptide containing 4 serine residues (tet-Ser) corresponding to the PSD of MARCKS. tet-Ser has been shown to inhibit PKC-dependent phosphorylation of MARCKS as well as a number of other PKC targets in vitro (20). Exposure of T84 monolayers to this peptide did not alter PMA-stimulated endocytosis (not shown). However, myristoylation of this peptide, which was hypothesized not only to enhance its membrane permeability but also to concentrate it at hydrophobic sites similar to MARCKS, enabled it to markedly inhibit PMA-stimulated endocytosis (Fig. 11). A similar 25-amino acid peptide in which the 4 serine residues were replaced by 4 alanine residues (tet-Ala) fails to block PKC-dependent MARCKS phosphorylation in vitro, although it functions as an effective inhibitor of phosphorylation of other PKC targets (20). This cognate tet-Ala peptide, modified in similar fashion by NH₂-terminal myristoylation, did not inhibit PMA-elicited endocytosis. The tet-Ser PSD peptide was also found to block translocation of MARCKS from the membrane to the cytosolic fraction in response to PMA, whereas tet-Ala did not (Fig. 12).

View larger version (14K): [in this window] [in a new window]   Fig. 11.   A myristoylated peptide corresponding to phosphorylation site domain (PSD) of MARCKS inhibits PMA-stimulated basolateral endocytosis. A myristoylated 25-amino acid peptide containing 4 serine residues and corresponding to PSD of MARCKS (tet-Ser) and a similar myristoylated 25-amino acid peptide in which the 4 serine residues are replaced by 4 alanine residues (tet-Ala) were synthesized. Monolayers were treated with 25 µM tet-Ser peptide for 1 h, and basolateral uptake of FITC-dextran was measured in response to PMA. A 1-h exposure to 10 nM PMA in absence of tet-Ser peptide increased basolateral endocytosis (n = 6). However, preincubation with tet-Ser peptide inhibited PMA-stimulated endocytosis by ~80% (n = 6, P < 0.005). Treatment with the tet-Ala peptide, however, did not alter PMA-stimulated endocytosis (n = 3). * P < 0.0001 for difference from control.

View larger version (18K): [in this window] [in a new window]   Fig. 12.   The tet-Ser peptide corresponding to PSD site of MARCKS inhibits PMA-induced translocation of MARCKS. A: monolayers were treated with either 25 µM tet-Ser peptide or 25 µM tet-Ala peptide for 1 h followed by 20-min exposure to 100 nM PMA. PMA-induced translocation of MARCKS from membrane (M) to cytosolic (C) fraction was blocked by tet-Ser peptide but not by tet-Ala peptide. B: MARCKS distribution was quantified by densitometric analysis and expressed as percent cytosolic MARCKS of total. Similar data were obtained in duplicate experiment.

Acetylcholine analog carbachol enhances basolateral endocytosis by a mechanism similar to PMA. Carbachol (CCH) acts via muscarinic receptors to stimulate Cl secretion in T84 monolayers and native tissue, an effect that involves an increase in intracellular Ca²⁺ (13). CCH is known to induce phospholipid turnover and generate DAG, thereby activating PKC (4, 13, 14). As shown in Fig. 13, CCH elicited an increase in basolateral endocytosis, an effect that persisted at least an hour after addition, well after termination of the associated Cl secretory response. However, the Ca²⁺-ATPase inhibitor thapsigargin, which is not known to activate PKC directly, did not increase endocytosis (data not shown), indicating that the effect of CCH on endocytosis can be dissociated from its effects on transepithelial secretion. As also shown in Fig. 13, CCH induced PKC- but not PKC- translocation within 30 min, consistent with its ability to induce basolateral endocytosis, and the effect on endocytosis was inhibited by Gö-6850 but not Gö-6796. Taken together, these data suggest that a physiological PKC stimulus such as muscarinic stimulation enhances basolateral endocytosis.

View larger version (15K): [in this window] [in a new window]   Fig. 13.   Carbachol (CCH) activates PKC- and elicits an increase in basolateral endocytosis. A: monolayers grown on 75-cm2 permeable supports were treated with 100 µM CCH for 30 min, and "cytosolic" (C) and "membrane" (M) fractions of homogenates were obtained as described in EXPERIMENTAL PROCEDURES. Fractions were subjected to SDS-PAGE and blotted with specific antibodies against PKC- and PKC- to examine redistribution of the 2 PKC isoforms in response to CCH. PKC-, but not PKC-, translocated to membrane fraction within 30 min of CCH treatment. Representative experiment shown was repeated 3 times. B: Monolayers grown on 4.7-cm2 permeable supports were treated with either 5 µM Gö-6850 or 5 µM Gö-6976 for 1 h followed by 100 µM CCH for 30 min in presence of inhibitors. Basolateral uptake of FITC-dextran was subsequently measured and compared with control. CCH increased basolateral endocytosis (n = 15, * P < 0.001), which was inhibitable by pretreatment with Gö-6850 (n = 3, P < 0.05) but not Gö-6976 (n = 3,  P < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

This study demonstrates that phorbol esters such as PMA selectively enhance basolateral but not apical endocytosis of a fluid-phase marker in polarized human T84 intestinal epithelial monolayers. This action is shared not only by bryostatin 1 and a cell-permeant DAG analog but also by CCH (an analog of acetylcholine, which is an endogenous regulator of intestinal transport function). The sensitivity of this response to a panel of putative isoform-selective inhibitors of PKC as well as the temporal correlation with PKC isoform translocation implicates a role for the novel PKC- isozyme. In response to PMA, PKC- translocated from the cytosolic to the membrane/particulate fraction within a time frame consistent with the early enhancement of endocytosis. Information regarding the role of specific PKC isoforms in plasma membrane dynamics is scant, but the available evidence suggests that individual isoforms may positively or negatively affect membrane traffic. For example, PKC- was shown to associate with caveolae (31), implicating a role for this isozyme in clathrin-independent endocytosis; however, activation of PKC- by phorbol esters appears to inhibit rather than enhance their internalization (41). Consistent with this notion, the conventional PKC inhibitor Gö-6796 slightly enhanced rather than inhibited PMA-elicited basolateral endocytosis in our experiments. The specific function(s) of PKC- are poorly understood, but available data have suggested that it plays a role in, among other processes, cell differentiation (34), oncogenesis (10), growth factor-stimulated proliferation (5), transduction of mechanical shear stress (44), synaptic function, and mucin exocytosis (23).

Surface expression of a given membrane protein represents a dynamic balance between its rate of exocytic insertion and endocytic retrieval. Activation of PKC has been shown to increase the internalization of a number of membrane markers, including receptors for transferrin, asialoglycoprotein, lactate dehydrogenase, the chemokine CXCR4, and various growth factors. Whether PKC enhances the endocytic retrieval of NKCC1 or other transport proteins from the basolateral membrane of secretory intestinal epithelial cells remains to be directly examined, although functional evidence to date supports such an effect. A similar hypothesis was recently proposed in A6 renal epithelia on the basis of the observation that PMA decreased basolateral binding of the Na⁺-K⁺-ATPase inhibitor ouabain in parallel with a decrease in basolateral membrane surface area as well as an increase in fluid-phase endocytosis. NKCC1 contains several potential PKC phosphorylation sites, raising the possibility that phosphorylation of NKCC1 at a specific PKC site could alter its affinity for the recycling process. However, it has been convincingly demonstrated that PMA-elicited endocytosis of the transferrin receptor and Na⁺-K⁺ ATPase does not involve their PKC-dependent phosphorylation, because enhanced internalization persists despite the presence of PKC site mutations. The possible role of the PKC- isoform has not, to our knowledge, been heretofore addressed.

Membrane trafficking is a complex process that involves vesicle budding, specific transport, vesicle docking, and membrane fusion events. A bewildering number of proteins have been identified in recent years that may participate in or regulate these various steps. Understanding the specific role of PKC in plasma membrane remodeling and endocytosis is further complicated by the presence of multiple pathways rather than a single pathway for endocytosis. At present, at least five independent forms of endocytosis are recognized: a clathrin-dependent pathway, macropinocytosis, a caveolar pathway, a clathrin- and caveolin-independent pathway, and phagocytosis. The molecular mechanisms involved in these pathways are distinct (37) and appear to be independently regulated. Uptake of fluid-phase markers can be differentially affected by stimuli that regulate clathrin-dependent and -independent endocytosis. Interestingly, the five major endocytic pathways all share the common feature of involvement of the actin cytoskeleton, although agents that affect microfilament architecture have been shown to exert opposite effects on these processes in many instances. For example, in proximal tubule-derived opossum kidney cells, phorbol esters and cytochalasin D enhanced fluid-phase uptake of FITC-dextran but inhibited adsorptive endocytosis of albumin (18), and in Vero kidney cells, the endocytic uptake of the membrane probe ricin and the fluid-phase marker lucifer yellow could be selectively modulated by cytochalasin, phorbol esters, and epidermal growth factor without alteration of the clathrin-dependent uptake of transferrin receptor (38).

The plasma membrane is intimately associated with a dense actin filament cortex beneath the cytoplasmic leaflet of the lipid bilayer, and it is perhaps not surprising that cortical actin interacts with various pathways of endocytosis and exocytosis (18, 25, 40, 46). Disruption of cortical actin appears to inhibit receptor-dependent endocytic events, perhaps by altering the clustering of ligands within coated pits or by affecting the molecular mechanisms of detachment of newly formed membrane invaginations. However, cortical F-actin appears also to act as a physical barrier to membrane fusion and budding events, and disassembly of this layer by chemical manipulation of actin polymerization or by agonist-regulated pathways appears to locally destabilize the plasma membrane and thereby favor a generalized increase in both endocytic and exocytic processes. Phorbol esters exert profound effects on cell shape and membrane spreading, effects that have been linked to their ability to reorganize actin filaments, microtubules, and their associated proteins (2). The present experiments suggest a link between the ability of PKC to remodel F-actin and the ability to enhance basolateral fluid-phase endocytosis. PKC- has been proposed to affect membrane traffic in other cells by an actin-dependent mechanism. For example, activated PKC- has been shown to bind actin within nerve endings and has been proposed to thereby participate in the exocytic neurotransmitter release (35). Cytochalasin D, like PMA, enhanced basolateral uptake of FITC-dextran; moreover, the effects of PMA on endocytosis were blocked by two chemically dissimilar stabilizers of F-actin, phalloidin and jasplakinolide.

The present experiments further suggest a role for MARCKS or a MARCKS-like protein in PKC-regulated basolateral endocytosis. Although MARCKS is one of the major cellular targets of PKC, its precise physiological function remains to be clearly established. MARCKS appears to associate with plasma membranes either directly via its myristoylated tail or indirectly via an effector protein (1). Hartwig et al. (21) showed that nonphosphorylated MARCKS acts as an actin cross-linking protein. Phosphorylation of MARCKS by PKC at four serine residues in its effector domain inhibits its actin cross-linking activity and results in translocation of MARCKS from the membrane (presumably cytoskeletally associated) environment to the cytosol. A number of PKC isoforms, including PKC-, have been shown to be MARCKS kinases (45). Our data are consistent with a model in which PMA, acting via PKC-, induces MARCKS phosphorylation and releases it from the basolateral membrane; as a result, MARCKS-mediated actin cross-linking is disrupted, actin filament disassembly is promoted, and membrane invagination and endocytosis are favored. In support of this concept, we demonstrated that 1) PMA induces MARCKS translocation in T84 cells from membrane to cytosolic fraction with an appropriate time course that parallels endocytosis; 2) pharmacological blockade of PKC- inhibits MARCKS movement and PMA-elicited endocytosis; and 3) a myristoylated tet-Ser PSD peptide blocks MARCKS movement and endocytosis in response to PMA, but a tet-Ala PSD peptide blocks neither. Such data should be interpreted with caution, because a number of other potential PKC targets such as adducin and MARCKS-related protein (MRP, also called MacMARCKS or F52) also contain a MARCKS-like PSD (12) and thus could be functionally inhibited by these novel myristoylated PSD peptides. Agents that affect actin assembly were shown in glioma cells to induce MARCKS translocation and to enhance PKC-induced MARCKS translocation (15), effects that we confirm for intestinal epithelial cells in the present study.

In summary, the present study indicates that phorbol esters, acting via the novel Ca²⁺-independent -isoform of PKC, selectively enhance basolateral but not apical membrane endocytosis in T84 cells by an F-actin-dependent mechanism. A novel myristoylated PSD peptide inhibits both MARCKS translocation and PMA-stimulated basolateral endocytosis, suggesting that MARCKS (or a protein containing a MARCKS-like PSD) may represent a link between PKC-stimulated actin rearrangement and endocytosis. This phenomenon may account for the profound alteration of the transport characteristics of the basolateral membrane of various polarized epithelia in response to agents that affect PKC.