• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of mbcLink to Publisher's site
Mol Biol Cell. Sep 1, 2009; 20(17): 3792–3800.
PMCID: PMC2735478

Basolateral Internalization of GPI-anchored Proteins Occurs via a Clathrin-independent Flotillin-dependent Pathway in Polarized Hepatic Cells

Keith E. Mostov, Monitoring Editor

Abstract

In polarized hepatocytes, the predominant route for apical resident proteins to reach the apical bile canalicular membrane is transcytosis. Apical proteins are first sorted to the basolateral membrane from which they are internalized and transported to the opposite surface. We have noted previously that transmembrane proteins and GPI-anchored proteins reach the apical bile canaliculi at very different rates. Here, we investigated whether these differences may be explained by the use of distinct endocytic mechanisms. We show that endocytosis of both classes of proteins at the basolateral membrane of polarized hepatic cells is dynamin dependent. However, internalization of transmembrane proteins is clathrin mediated, whereas endocytosis of GPI-anchored proteins does not require clathrin. Further analysis of basolateral endocytosis of GPI-anchored proteins showed that caveolin, as well as the small GTPase cdc42 were dispensable. Alternatively, internalized GPI-anchored proteins colocalized with flotillin-2–positive vesicles, and down-expression of flotillin-2 inhibited endocytosis of GPI-anchored proteins. These results show that basolateral endocytosis of GPI-anchored proteins in hepatic cells occurs via a clathrin-independent flotillin-dependent pathway. The use of distinct endocytic pathways may explain, at least in part, the different rates of transcytosis between transmembrane and GPI-anchored proteins.

INTRODUCTION

The plasma membrane of epithelial cells is divided into two domains, apical and basolateral, separated by tight junctions. Each domain has a specific protein and lipid composition, which is correlated with specialized functions. Sorting mechanisms must regulate trafficking of molecules to the appropriate membrane domain. These have been shown to operate along both the biosynthetic and endocytic pathways (reviewed in Rodriguez-Boulan et al., 2005 blue right-pointing triangle). Newly synthesized proteins are transported to polarized surfaces either by a direct or an indirect route. In the direct pathway, membrane proteins are primarily sorted in the trans-Golgi Network (TGN) and targeted directly to the apical or basolateral domain (Rodriguez-Boulan and Powell, 1992 blue right-pointing triangle). In the indirect pathway, however, TGN-derived carriers are first sent to the basolateral surface, from where they are selectively internalized and subsequently transcytosed to the opposite apical surface (reviewed in Mostov et al., 2000 blue right-pointing triangle, Rodriguez-Boulan et al., 2005 blue right-pointing triangle). In hepatocytes, unlike most simple polarized epithelial cells, the predominant route to the apical bile canalicular (BC) membrane is indirect (reviewed in Aït-Slimane and Hoekstra 2002 blue right-pointing triangle; Tuma and Hubbard, 2003 blue right-pointing triangle).

We have shown previously that, in polarized hepatic cells, both GPI-anchored proteins (APs) and single transmembrane (TMD) proteins use the transcytotic pathway. However, transcytosis of GPI-APs requires their integration into detergent-resistant microdomains called rafts, whereas TMD proteins exploit a raft-independent transcytotic pathway (Aït-Slimane et al., 2003 blue right-pointing triangle). In this study, we noted that transcytotic transport of TMD proteins was considerably more efficient than that of GPI-APs. A similar observation has been made in vivo regarding transcytosis of the GPI-AP 5′ nucleotidase (Schell et al., 1992 blue right-pointing triangle). Whether these variations reflect differences in transport kinetics or use of separate endocytic/transcytotic pathways has not been investigated. It has been reported that newly synthesized apical resident TMD proteins and the transcytosing polymeric immunoglobulin receptor (pIgA-R) follow the same transcytotic pathway (Hemery et al., 1996 blue right-pointing triangle, Ihrke et al., 1998 blue right-pointing triangle), but it is not clear whether GPI-APs enter this pathway. Moreover, except for the pIgA-R, which is internalized in clathrin-coated pits, the mechanisms that mediate basolateral internalization of apical resident proteins in hepatocytes are not known. The pIgA-R is endocytosed with tyrosine-based signals, which seem to be conventional clathrin-mediated internalization signals (Okamoto et al., 1992 blue right-pointing triangle). An important difference between the pIgA-R and BC resident proteins is that most of the latter are GPI-APs or TMD proteins with short cytoplasmic tails and no obvious internalization motif. It is generally assumed that in hepatocytes, TMD proteins are internalized via a clathrin-dependent pathway, like the pIgA-R. However, there are no functional studies demonstrating a role for clathrin in basolateral endocytosis of apical resident membrane proteins.

The endocytic pathway of GPI-APs has been extensively investigated in various nonpolarized cell types, but there is still no consensus concerning the mechanisms involved. Several studies performed in Cos-7 and Chinese hamster ovary (CHO) cells have suggested that internalization of GPI-APs is dynamin- and clathrin-independent (reviewed in Chatterjee and Mayor, 2001 blue right-pointing triangle; Mayor and Riezman, 2004 blue right-pointing triangle; Perret et al., 2005 blue right-pointing triangle). Other studies in Cos-7 and MA104 cells found GPI-APs in caveolin (Cav)-positive endocytic structures termed caveolae (Anderson et al., 1992 blue right-pointing triangle; Nichols, 2002 blue right-pointing triangle). However, these observations have been challenged in both cell types (Parton et al., 1994 blue right-pointing triangle; Fujimoto, 1996 blue right-pointing triangle; Fivaz et al., 2002 blue right-pointing triangle; Sabharanjak et al., 2002 blue right-pointing triangle). The group of Mayor proposed that GPI-APs are internalized in specialized compartments called GPI-anchored protein-enriched early-endosomal compartments via cdc42- and ARF1-regulated pathway (Sabharanjak et al., 2002 blue right-pointing triangle; Kumari and Mayor, 2008 blue right-pointing triangle). Finally a new pathway involving the scaffolding proteins flotillin-1 and flotillin-2 has been described for the endocytosis of GPI-APs in Cos-7 and CHO cells (Glebov et al., 2006 blue right-pointing triangle; Frick et al., 2007 blue right-pointing triangle). Thus far, the mechanisms that mediate the basolateral internalization of apical resident proteins in polarized hepatic cells have not been investigated.

We have investigated the mechanisms involved in the endocytic/transcytotic transport of apical resident TMD proteins and GPI-APs in polarized hepatic cells. We have compared the basolateral internalization of two GPI-APs, CD59 endogenously expressed, and GPI-green fluorescent protein (GFP) stably transfected in HepG2 cells, with that of two endogenous TMD proteins, dipeptidyl-peptidase IV (DPPIV) and aminopeptidase N (APN). Our data show that in polarized hepatic cells, GPI-APs and TMD proteins exploit distinct mechanisms for their endocytic/transcytotic transport. We demonstrate that internalization of TMD proteins is dynamin and clathrin dependent, whereas internalization of GPI-APs is dynamin dependent but clathrin independent. Furthermore, we found that flotillin-2 RNA interference (RNAi) inhibited basolateral endocytosis of GPI-APs without affecting internalization of TMD proteins. We propose that flotillin-2 is one determinant of the clathrin-independent endocytic pathway of GPI-APs in polarized hepatic cells.

MATERIALS AND METHODS

Reagents and Antibodies

Culture media, Alexa Fluor 488 (AF488)-labeled secondary antibodies, and Alexa Fluor 594-labeled transferrin (AF594-Tf), cholera toxin B subunit (AF594-CTX), and Topro-3 were purchased from Invitrogen (Cergy-Pontoise, France). Fluorescent secondary antibodies were from Jackson ImmunoResearch Laboratories (Montluçon, France). Peroxidase-conjugated secondary antibodies were from Rockland Immunochemicals (Gilbertville, PA). The enhanced chemiluminescence (ECL) detection kit was from GE Healthcare France (Orsay, France). The small-interfering RNA (siRNA)-clathrin heavy chain duplex was purchased from Eurogentec (Liège, Belgium). The siRNA–flotillin-2 duplex was obtained from Santa Cruz Biotechnology (Santa Cruz, CA), and control siRNA was obtained from Dharmacon RNA Technologies (Lafayette, CO). All other reagents were obtained from Sigma (St. Quentin-Fallavier, France) unless otherwise noted. The polyclonal antibodies (pAbs) anti-actin and anti-clathrin heavy chain were from Santa Cruz Biotechnology. The pAbs anti-flotillin-2 and anti-caveolin-1 were from BD Biosciences (Le-Pont-de-Claix, France). The monoclonal antibody (mAb) anti-CD59 was from Abcam (Cambridge, United Kingdom). The mAb anti-GFP was from Roche Diagnostics (Meylan, France). The pAb against GFP was raised in the laboratory by Jean-Louis Delaunay (UMRS 938, CDR Saint-Antoine, Paris, France). The pAb against APN was provided by Prof. Ann Hubbard (John Hopkins University, Baltimore, MD). The mAb directed against human DPPIV (HBB3/775) was provided by Prof. Hans-Peter Hauri (Biozentrum der Universität Basel, Basel, Switzerland).

Cell Culture

HepG2 cells were grown at 37°C in DMEM supplemented with 10% heat-inactivated (56°C; 30 min) fetal bovine serum, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin, under a 5% CO2, air atmosphere. For microscopy and biochemistry experiments, HepG2 cells were grown onto glass coverslips and 60-mm dishes, respectively.

DNA Constructs, Transfection, and Clonal Selection

The GPI-GFP construct produced by fusion of the folate receptor GPI-anchoring sequence with GFP was a kind gift from Dr. S. Mayor (National Centre for Biological Sciences, Bangalore, India). Flot2-GFP was a kind gift from Dr. R. Tikkanen (University Clinic, Frankfurt-am-Main, Germany). The Cav-1-GFP was from Dr. A. Le Bivic (Institut de Biologie du Developpement de Marseille, Faculté des Sciences de Luminy, Marseille, France). GFP constructs of cdc42 N17, rac N17, and rhoA N19 were obtained from Dr. S. Etienne-Manneville (Institut Pasteur, Paris, France). GFP construct of Eps15 DIIIΔ2 (control) and Eps15 EΔ95/295 (dominant-negative Eps15) were from Dr. A. Benmerah (Institut Cochin, Paris, France). GFP construct of dynamin 2 (control) and dynamin 2K44 (dominant-negative Dyn) were from Dr. M. McNiven (Mayo Clinic, Rochester, MN). HepG2 cells were transfected using nucleofector II (Amaxa, Cologne, Germany) using program 22 (solution V) according to the manufacturer's instructions. After 48 h, selection was started by addition of 1 mg/ml Geneticin (G418, Invitrogen, Cergy-Pontoise, France). Stable clones were isolated after 3 wk by using cloning cylinders. Positive cells were screened by GFP fluorescence.

Indirect Immunofluorescence and Confocal Microscopy

Cells were washed with phosphate-buffered saline with 0.5 mM CaCl2 and 1 mM MgCl2 (PBS+), fixed with 4% paraformaldehyde for 1 min at 4°C, and permeabilized in methanol for 10 min at 4°C. After blocking in 1% PBS/bovine serum albumin (BSA), cells were incubated for 1 h at room temperature with primary antibodies. After three washes in PBS, cells were incubated for 1 h with fluorescently labeled secondary antibodies. After three washes in PBS, cells were incubated for 10 min with RNAse A and then for 1 min with Topro-3 to stain nuclei. Confocal imaging was acquired with a TCS SP2 laser-scanning spectral system (Leica, Wetzlar, Germany) attached to a Leica DMR inverted microscope. Optical sections were recorded with a 63/1.4 immersion objective. Laser scanning confocal images were collected, and analyzed using the online Scan Ware software. Figure compilation was accomplished using Adobe Photoshop 5.5 and Adobe Illustrator 10 (Adobe Systems, Mountain View, CA).

Internalization Assays

Internalization assays were performed in two different ways according to a previously published protocol (Aït-Slimane et al., 2003 blue right-pointing triangle). In the first method, HepG2 cells were washed three times with HEPES-buffered (20 mM, pH 7.0) serum-free medium (HSFM). Cell surface antigens were labeled at 4°C for 30 min with specific primary antibodies diluted in HSFM/0.2% BSA. After surface labeling, cells were extensively washed with HSFM/0.2% BSA, placed in prewarmed complete medium, and incubated at 37°C for the indicated times. In the second method, HepG2 cells were continuously labeled with anti-CD59, DPPIV, or APN antibodies at 37°C for the indicated times. Noninternalized antibody–antigen complexes were removed by acid washing (200 mM glycine and 150 mM NaCl, pH 2.5) before fixation, permeabilization, and staining with fluorescently labeled secondary antibodies. For CTX and Tf labeling, HepG2 cells were incubated at 37°C for the indicated times with 10 μg/ml AF594-CTX or 30 μg/ml AF594-Tf. Excess CTX or Tf at the cell surface was removed by acid stripping. Fluorescence was examined by confocal scanning microscopy.

Analysis of Colocalization and Quantification of Total Fluorescence

To quantify the level of colocalization, 20–30 cells per experimental condition were randomly selected on the same coverslip among those that showed a well resolved pattern for the anti-CD59- or anti-DPPIV–labeled structures. The laser power and the levels for detector amplification were optimized for each channel before starting acquisition of the images. The ImageJ (National Institutes of Health, Bethesda, MD) 1.41 measure colocalization function was used to calculate the percentage of colocalization between the different probes. To quantify the amount of internalized cargo, 20–30 control and siRNA or dominant-negative mutant expressing cells per coverslip were randomly selected and imaged using the confocal microscope. All the images were taken with identical acquisition parameters. The amount of internalized cargo in treated cells was measured using ImageJ 1.41 and expressed as a percentage of that internalized in control cells.

Western Blot

Transfected cells were washed with PBS+ and lysed on ice for 30 min in TNE buffer (20 mM Tris-HCl, 150 mM NaCl, and 1 mM EDTA, pH 7.4) containing 1% (wt/vol) Triton X-100 in the presence of a protease inhibitor cocktail (Roche Diagnostics, Basel, Switzerland). Protein content was determined using the detergent-compatible bicinchoninic acid reagent (Interchim, Montluçon, France), with BSA as the standard, according to the manufacturer's instructions. Equal amounts of protein were directly processed for SDS-polyacrylamide gel electrophoresis on 7.5% polyacrylamide gels. Immunoblotting was performed using the appropriate primary antibodies followed by horseradish peroxidase-conjugated mouse-specific secondary antibody. Development of peroxidase activity was performed with the ECL detection kit (GE Healthcare France).

RESULTS

GPI-anchored Proteins and Single Transmembrane Domain Proteins Are Transcytosed with Different Efficiencies

We compared the kinetic of transcytosis of two GPI-APs, CD59 and GPI-GFP, to that of two TMD proteins, DPPIV and APN. CD59, DPPIV, and APN are endogenously expressed in HepG2 cells. The fusion protein GPI-GFP was stably expressed after transfection. Transcytosis was monitored by continuously labeling the basolateral pool of selected apical proteins for 20 min at 37°C. The cells were fixed, permeabilized, and the trafficked antibody–antigen complexes visualized with secondary antibodies. Figure 1A shows that, after 20 min, the GPI-APs were concentrated in small punctate structures particularly visible in the area directly underneath the basolateral membrane, whereas no apical staining was detected. By contrast, the TMD proteins were present mainly at the apical plasma membrane, and in the subapical area. To directly compare the kinetics of internalization of GPI-APs and TMD proteins, we followed the trafficking of CD59 and DPPIV simultaneously. HepG2 cells were incubated for 15 min at 0°C with antibodies to CD59 and DPPIV to allow their binding to the basolateral surface. Trafficking was initiated by raising the temperature to 37°C. After 20 min, the cells were processed as in Figure 1A. As shown in Figure 1B, both CD59 and DPPIV became internalized as reflected by the increase of the intracellular punctate staining. However, staining for CD59 remained largely at the periphery of the cells, whereas staining for DPPIV was observed deeper in the cells and at the level of the membrane (see the merged picture). These results show that GPI-APs and TMD proteins have different kinetics of internalization at the basolateral membrane of polarized hepatic cells.

Figure 1.
GPI-APs and TMD proteins are internalized and transcytosed with different kinetics in HepG2 cells. (A) Endogenous DPPIV, APN, CD59, and stably expressed GPI-GFP present at the BL membrane were continuously labeled with specific antibodies for 20 min at ...

GPI-anchored Proteins and Single Transmembrane Domain Proteins Are Internalized in Distinct Vesicles

In previous studies, we and others have shown that, in polarized hepatic cells, the transcytotic transport of GPI-APs occurs through a raft-mediated pathway, whereas TMD proteins do not exploit such transport mechanism (Aït-Slimane et al., 2003 blue right-pointing triangle; Nyasae et al., 2003 blue right-pointing triangle). This observation suggests that GPI-APs and TMD proteins may use different internalization mechanisms at the basolateral membrane. To test this hypothesis, we performed cotrafficking experiments using antibodies to transcytotic proteins, and we searched for colocalization with markers of known endocytic pathways. As shown in Figure 2A, nearly perfect colocalization was observed between DPPIV and Tf, a well established marker for the clathrin pathway, whereas only few vesicles costained with CD59 and Tf. Quantification of the results showed that >80% of DPPIV but only ~20% of CD59 colocalized with Tf. Similar to DPPIV, APN colocalized with Tf, whereas GPI-GFP behave like CD59 (Figure 2B). We also compared internalization of GPI-APs and TMD proteins with CTX which, depending on cell type, can be internalized via clathrin-dependent and/or -independent mechanisms (Torgersen et al., 2001 blue right-pointing triangle; Sandvig and van Deurs, 2002 blue right-pointing triangle; Nichols, 2003 blue right-pointing triangle). In HepG 2 cells, basolateral (BL) internalized CTX did not colocalize with Tf or with the early endosome marker EEA1 (data not shown), suggesting that BL uptake of CTX occurs via a clathrin-independent pathway. It is interesting that the GPI-AP CD59 was internalized in structures positive for CTX, whereas no colocalization was detected between DPPIV and CTX (Figure 2C). GPI-GFP also colocalized with CTX but APN did not, as shown in the quantitative analysis (Figure 2D). These results show that TMD proteins are likely to be internalized via clathrin-coated vesicles, whereas GPI-APs are excluded from Tf-positive structures and internalized in CTX-positive vesicles.

Figure 2.
GPI-APs and TMD proteins are internalized in distinct vesicles. HepG2 cells were allowed to internalize AF594-labeled Tf (A and B) or AF594-labeled CTX (C and D) with antibodies to DPPIV or CD59 for 5 min at 37°C. At the end of the 37°C ...

Basolateral Internalization of GPI-APs Is Eps15 Independent, Dynamin Dependent

The observation that GPI-APs were excluded from Tf-containing vesicles suggested that their basolateral endocytic/transcytotic transport occurs through a clathrin-independent pathway. To ascertain that clathrin is not involved in the basolateral internalization of GPI-APs, we transfected HepG2 cells with a dominant-negative isoform of Eps15 (EΔ95/295), which interferes with clathrin-coated pit assembly (Benmerah et al., 1999 blue right-pointing triangle). Figure 3A shows that the basolateral internalization of CD59 and CTX was not obviously impaired by expression of the Eps15 mutant. On the contrary, Tf internalization, a positive control of clathrin function was dramatically reduced, as well as DPPIV internalization. These morphological observations were quantified by measuring the relative fluorescence intensity present in control cells and cells expressing the Eps15 mutant (Figure 3C). Internalization of Tf and DPPIV was reduced from ~90 to 10%, whereas internalization of CD59 and CTX were minimally affected, shifting from ~90 to 75%.

Figure 3.
Effect of Eps15 and dynamin dominant-negative mutants on the BL endocytosis of GPI-APs and TMD proteins. HepG2 cells were transfected with Eps15 DIIIΔ2 (Eps15 CT), Eps15 EΔ95/295 (Eps15 MT), Dyn2WT, or Dyn2K44, all tagged with GFP. Three ...

Dynamin has been shown to regulate many vesiculation events at the plasma membrane, both clathrin dependent and clathrin independent (McNiven et al., 2000 blue right-pointing triangle). To investigate the role of dynamin in the endocytic process of GPI-APs, we transiently transfected HepG2 cells either with the wild-type dynamin (DynWT) or with the mutant defective in GTP binding and hydrolysis (DynK44A). Figure 3B shows that Tf, DPPIV, CTX, and CD59 were efficiently internalized in DynWT cells. In contrast, the internalization of all markers was considerably reduced in cells overexpressing the GTP-deficient mutant DynK44A. Quantitative analysis of internalized fluorescence (Figure 3D) confirmed that endocytosis of Tf and DPPIV was almost abolished in DynK44A cells and that internalization of CTX and CD59 was also strongly reduced, dropping to <20%. Together, these results show that internalization of GPI-APs is mediated by a clathrin-independent, dynamin-dependent mechanism.

Effect of Clathrin Heavy Chain Depletion on the Basolateral Internalization of GPI-APs

To obtain further support for a clathrin-independent mechanism of GPI-APs internalization in polarized hepatic cells, we inhibited expression of the clathrin heavy chain in HepG2 cells. We used a siRNA duplex that targets segment 3311-3333 of the clathrin heavy chain (Hinrichsen et al., 2003 blue right-pointing triangle). Control cells were transfected with a scrambled siRNA. To quantitate the reduction of clathrin heavy chain expression, transiently transfected HepG2 cells were lysed in SDS-sample buffer and analyzed by immunoblotting with mAb directed against the clathrin heavy chain. Three days after transfection, the signal from the clathrin heavy chain was strongly reduced (Figure 4A). We chose this time point to study the effect of clathrin depletion on the endocytic/transcytotic transport of GPI-APs and TMD proteins using the antibody trafficking assay. Transiently transfected HepG2 cells were incubated for 15 min at 37°C with antibodies to transcytosed proteins. The cells were fixed and permeabilized, and the trafficked antibody–antigen complexes were visualized with secondary antibodies. Cells were subsequently stained with a mAb directed against the clathrin heavy chain. Immunofluorescence showed that typical dot-like staining of clathrin was strongly reduced in many but not all cells (Figure 4B), thus making it possible to observe both clathrin-expressing and clathrin-depleted cells within the same field. Cells that did not express clathrin did not internalize Tf or DPPIV (Figure 4B). By contrast clathrin-depleted cells still internalized CTX as well as CD59. The results of intracellular fluorescence quantification (Figure 4C) gave values very similar to those obtained with the Eps15 mutant (Figure 3C). These results provide additional support for a clathrin-independent way of internalization of GPI-APs at the BL surface of polarized hepatic cells.

Figure 4.
Basolateral internalization of TMD proteins but not GPI-APs is clathrin-mediated. (A) Western blot analysis of clathrin heavy chain (HC) depletion in HepG2 cells. Lysates were prepared 72 h after transfection with clathrin siRNA or control siRNA. Actin ...

The Small GTPase cdc42 Is Not Involved in the Basolateral Endocytosis of GPI-APs in Polarized Hepatic Cells

The Rho family of small GTPases has been involved in several endocytic pathways, including pinocytosis and transcytosis (reviewed in Ellis and Mellor, 2000 blue right-pointing triangle). Among them, cdc42 has been involved in the internalization of GPI-APs (Sabharanjak et al., 2002 blue right-pointing triangle; Chadda et al., 2007 blue right-pointing triangle; Kumari and Mayor, 2008 blue right-pointing triangle). To examine whether cdc42 was involved in the regulation of clathrin-independent endocytosis of GPI-APs, we analyzed the internalization of both CD59 and CTX in HepG2 cells expressing a dominant-negative isoform of cdc42 (cdc42N17). We analyzed in parallel the internalization of Tf and DPPIV. In agreement with the previously published data showing that cdc42 was involved in the clathrin-mediated endocytosis of IgA in MDCK cells (Rojas et al., 2001 blue right-pointing triangle), we found that clathrin-mediated endocytosis of Tf and DPPIV was inhibited in cells expressing the dominant-negative isoform of cdc42 (Figure 5A). In contrast, the cdc42 mutant did not prevent endocytosis of CTX and CD59 (Figure 5A). These results were confirmed by the quantitative analysis of internalized fluorescence (Figure 5B), suggesting that in polarized hepatic cells cdc42 is not involved in the BL internalization of GPI-APs.

Figure 5.
The Rho family GTPase cdc42 is not involved in the basolateral internalization of GPI-APs. (A) HepG2 cells were transfected with cdc42N17 tagged with GFP. After transfection, cells were incubated for 5 min at 37°C with AF594-labeled Tf or AF594-labeled ...

Caveolin-1 Is Not Involved in the Basolateral Endocytosis of GPI-APs in Polarized Hepatic Cells

Because HepG2 cells do not express caveolin-1 endogenously (Fujimoto et al., 2000 blue right-pointing triangle; Fu et al., 2004 blue right-pointing triangle; Figure 6A), the possibility was raised that this protein may be required for efficient endocytosis of GPI-APs at the BL surface of HepG2 cells. To test this hypothesis, we stably transfected HepG2 cells with a cDNA encoding the caveolin-1-GFP fusion protein (Cav-1-GFP). Expression of Cav-1-GFP was first analyzed by Western blot using specific anti-caveolin-1 and anti-GFP antibodies. A single band of 48 kDa was detected in HepG2/Cav-1-GFP cells (Figure 6A). No expression of caveolin-1 protein was detected in nontransfected HepG2 cells (Figure 6A). The pattern of distribution of Cav-1-GFP in HepG2 cells was analyzed by confocal microscopy. As shown in Figure 6B, transfected cells displayed Cav-1-GFP as a punctate pattern at both the BC surface and the BL membrane, in addition to significant intracellular staining. To determine whether caveolin-1 is involved in the BL internalization of GPI-APs, we performed the antibody-trafficking assay. Cav-1-GFP–expressing cells were incubated continuously with antibodies to CD59. After 15 min at 37°C, the cells were fixed, permeabilized, and incubated with fluorescently labeled secondary antibodies. As shown in Figure 6C, there was no significant colocalization between Cav-GFP and CD59 as evidenced by the presence of distinct green and red vesicles throughout the cytoplasm. As expected, similar experiments comparing the internalization of Tf or DPPIV with caveolin did not show any colocalization (Figure 6C). These results suggest that basolateral internalization of GPI-APs does not occur via caveolae, even in caveolin-expressing HepG2 cells.

Figure 6.
Basolateral internalization of GPI-APs and TMD proteins occurs independently of caveolin-1. (A) Western blot of nontransfected (lane 1) and caveolin-1–GFP-transfected HepG2 cells by using anti-caveolin-1 antibody (lane 2) and anti-GFP antibody ...

Involvement of Flotillin in the Basolateral Internalization of GPI-APs in Polarized Hepatic Cells

Recent reports have described a new clathrin- and caveolin-independent endocytic pathway in mammalian cells. The two closely related proteins flotillin-1 and flotillin-2 serve as markers for this pathway (Glebov et al., 2006 blue right-pointing triangle; Frick et al., 2007 blue right-pointing triangle; Payne et al., 2007 blue right-pointing triangle). Hepatic cells express flotillin-2 but not flotillin-1 (our unpublished data; Volonté et al., 1999 blue right-pointing triangle). To test the possible involvement of flotillin-2 in the clathrin- and caveolin-independent internalization of GPI-APs, we first checked whether internalized CD59 colocalized with flotillin-2, in flotillin-2-GFP–expressing cells. Cells were incubated with antibodies to CD59 or DPPIV for 30 min at 0°C. Trafficking was initiated by raising the temperature to 37°C. After 30 min, the cells were fixed, permeabilized, and incubated with fluorescently labeled secondary antibodies. As shown in Figure 7A, after 30 min at 37°C, abundant puncta containing both flotillin-2-GFP and CD59 were observed at the periphery of the cells. In contrast, DPPIV, which is excluded from these structures, was observed deeper in the cells and at the BC membrane, where it colocalized with the steady-state localized flotillin2-GFP. Moreover, after a short period of internalization (2 min), CTX but not Tf nearly 100% colocalized with flotillin-2-GFP within the same vesicles (data not shown). To test whether flotillin-2 is required for the BL internalization of GPI-APs in HepG2 cells, we used siRNA to knock down the expression level of flotillin-2. The knockdown efficiency was tested by immunoblotting as well as by immunofluorescence. As shown in Figure 7B, 72 h after transfection, the flotillin-2 signal was strongly reduced, whereas the clathrin signal, used as an internal control had not changed. Compared with control cells in which flotillin-2 fluorescence was present at the BC membrane, at the BL surface and to a lesser extent in intracellular structures, cells transfected with flotillin-2 siRNA did not show detectable signal for flotillin-2 (Figure 7C). We analyzed the effect of flotillin-2 knockdown on the BL internalization of DPPIV and CD59. The efficiency of entry was determined using the assay described above for the effect of clathrin knockdown. Tf and CTX were used as markers for the clathrin-dependent and clathrin-independent pathways, respectively. Compared with control cells, cells transfected with siRNA against flotillin-2 showed ~70% decrease of CTX and CD59 internalization, whereas DPPIV and Tf entry were not inhibited by flotillin-2 knockdown (Figure 7D). To further determine whether flotillin-2 is required for the endocytic-transcytotic transport in polarized hepatic cells, we monitored the antibody-trafficking assay in control and siRNA treated. Cells were processed as in Figure 7A, except that the cells were incubated at 37°C for 60 min and were subsequently stained with a mAb directed against flotillin-2. As shown in Figure 7E, in control cells CD59 was largely detected in punctate structures visible in the cytoplasm, in the subapical area and at the BC membrane, whereas in siRNA-treated cells, CD59 remained largely at the BL surface and no BC staining was detected (Figure 7E). By contrast, a prominent and homogeneous distribution of the TMD protein DPPIV at the BC surface was apparent in both control and siRNA-treated cells (Figure 7E). Together, these experiments suggest that flotillin-2 plays a role in the basolateral internalization of GPI-APs in polarized hepatic cells.

Figure 7.
Basolateral internalization of GPI-APs involves flotillin-2. (A) Antibodies to CD59 and DPPIV were bound simultaneously to the basolateral surface of flotillin-2–GFP-expressing cells at 0°C. After unbound antibody was removed, the cells ...

DISCUSSION

In previous studies, we have made the observation that, in polarized hepatic cells, the transcytotic efflux of apical resident GPI-APs and TMD proteins occurs with different kinetics, suggesting that distinct classes of transcytosing proteins may use separate internalization mechanisms. In this study, we took endogenous CD59 and transfected GPI-GFP as representatives of apical GPI-APs in HepG2 cells, and the endogenous DPPIV and APN as apical TMD proteins. We show that TMD proteins are internalized through a clathrin-dependant mechanism, whereas internalization of GPI-APs is clathrin-independent but requires both dynamin and flotillin.

Internalization of TMD Proteins but Not GPI-APs Is Clathrin Mediated

Clathrin-dependent endocytosis is by far the better understood mechanism for internalization. Clathrin assembly can be driven by a variety of sorting motifs present in membrane receptors that can thus mediate their own endocytosis (Conner and Schmid, 2003 blue right-pointing triangle). The mechanism of internalization of GPI-APs or TMD proteins that do not bear internalization motifs remains elusive. Apical resident proteins like DPPIV and APN have very short cytoplasmic tails with only six and eight amino acids, respectively. It is therefore unlikely that their cytoplasmic tails contain a signal for incorporation into coated pits. Nevertheless, we found that BL endocytosis of APN and DPPIV was strongly inhibited by clathrin depletion or by expression of Eps15 dominant-negative mutants. In contrast, BL endocytosis of GPI-APs was little affected by clathrin depletion. These observations indicate that APN and DPPIV, but not GPI-APs, are able to enter coated pits. These differences may be linked to their affinity for distinct membrane microdomains. Indeed, GPI-APs preferentially partition into cholesterol-enriched membrane microdomains, called rafts (Aït-Slimane et al., 2003 blue right-pointing triangle; Nyasae et al., 2003 blue right-pointing triangle). Experiments using fluorescence resonance energy transfer have shown that cholesterol-dependent clusters of CTX bound to GM1 and GPI-APs are excluded from clathrin-coated pits (Nichols, 2003 blue right-pointing triangle). TMD proteins, which do not cluster in rafts may enter coated pits by diffusion and make opportunity of the efficient clathrin-mediated endocytic pathway to reach endosomal sorting compartments and achieve transcytosis, whereas GPI-APs would not be able to enter coated pits readily.

The Non-Clathrin Endocytic Pathway of GPI-APs

That GPI-APs are preferentially localized in rafts makes caveolae a good candidate for their endocytosis. Like rafts, these small invaginations are enriched in cholesterol and sphingolipids and can be isolated as detergent-resistant membranes (reviewed in Parton and Simons, 2007 blue right-pointing triangle). They have been implicated in endocytosis of CTX and GPI-APs (Kurzchalia and Parton, 1999 blue right-pointing triangle). However, formation of caveolae depends on the presence of caveolin, and not all cells do express caveolin. In HepG2 cells, we did not observe significant colocalization of GPI-APs and caveolae after induction of caveolae by caveolin overexpression. This makes it unlikely that caveolae could represent an endocytic pathway for GPI-APs in hepatic cells.

Several other raft-dependent entry mechanisms have been described that may be used by GPI-APs and/or raft-associated proteins (Mayor and Riezman, 2004 blue right-pointing triangle; Mayor and Pagano, 2007 blue right-pointing triangle). The raft-associated IL-2 receptor is endocytosed via a dynamin and RhoA-dependent mechanism in lymphocytes (Lamaze et al., 2001 blue right-pointing triangle). In contrast, Sabharanjak et al. (2002) blue right-pointing triangle found that internalization of GPI-APs was independent of dynamin and RhoA and was regulated by another small GTPase, cdc42, in CHO and Cos-7 cells. In our study, the entry of GPI-APs was dynamin dependent but did not seem to involve cdc42. Furthermore, we did not observe any effect of dominant-negative mutants of the Rho and Rac GTPases on the BL internalization of GPI-APs (our unpublished data). These different results suggest that endocytosis of GPI-APs and raft-associated proteins may be differentially regulated in different cell types or that more than one mechanism is involved in endocytosis of raft-associated proteins.

Involvement of Flotillin in Basolateral Endocytosis of GPI-APs

Flotillin-1 and flotillin-2, also known as reggie-2 and reggie-1, have been identified as plasma membrane-associated proteins that cocluster with GPI-APs in noncaveolar raft membrane microdomains (Lang et al., 1998 blue right-pointing triangle; Stuermer et al., 2001 blue right-pointing triangle). Flotillin-1 was recently implicated in the clathrin- and caveolin-independent endocytosis of GPI-APs and CTX in Cos-7 cells (Glebov et al., 2006 blue right-pointing triangle); and of proteoglycan-binding ligands in HeLa cells (Payne et al., 2007 blue right-pointing triangle). Moreover, a recent report suggested that flotillin-1 and flotillin-2 coassemble to generate membrane microdomains, and this coassembly was necessary to induce membrane invagination and vesicle budding (Frick et al., 2007 blue right-pointing triangle). In HepG2 cells, we observed a large colocalization between internalized CD59 and flotillin-2-GFP but not between internalized DPPIV and flotillin-2-GFP. Moreover, immediately after entry, CTX but not Tf, nearly 100% colocalized with flotillin-2-GFP within the same vesicles (data not shown). BL internalization of CD59 and CTX were strongly inhibited after down-expression of flotillin-2, proving that in polarized hepatic cells, the BL uptake of GPI-APs occurs via a flotillin-dependent pathway. It has been reported that the flotillin-dependent internalization of CTX and GPI-APs is dynamin independent (Glebov et al., 2006 blue right-pointing triangle), whereas another study showed that the entry of proteoglycan-binding ligands is dynamin and flotillin dependent (Payne et al., 2007 blue right-pointing triangle). Here, we showed that BL endocytosis of GPI-APs and CTX requires both flotillin and dynamin. These differences suggest that flotillin may play a role in several endocytic pathways, both dynamin dependent and independent.

It is generally considered that clathrin-dependent endocytosis is a very fast process (t1/2 < 10 s), and that clathrin-independent endocytosis proceeds more slowly (t1/2 > 1 min), with caveolae being the slowest (t1/2 > 20 min) (Conner and Schmid, 2003 blue right-pointing triangle). Our results are in agreement with flotillin-dependent endocytosis being slower than the clathrin-dependent mechanism. In these conditions, BL internalization of GPI-APs would be less efficient that BL internalization of TMD proteins. These differences may explain, to a large part, the different kinetics of transcytosis of these two classes of membrane proteins in HepG2 cells. One major question for the future is whether the clathrin-dependent pathway of TMD proteins and the flotillin-dependent pathway of GPI-APs converge in endosomes and whether they use common transcytotic compartments.

ACKNOWLEDGMENTS

We thank André Le Bivic, Satyajit Mayor, Ritva Tikkanen, Sandrine Etienne-Manneville, Alexandre Benmerah, and Mark McNiven for the generous gift of cDNAs; Ann. L. Hubbard and Hans-Peter Hauri for providing antibodies; Philippe Fontanges and Romain Morichon (Institut National de la Santé et de la Recherche Médicale IFR65) for help with the confocal microscopy; and Jean-Louis Delaunay for the critical reading of the manuscript.

Glossary

Abbreviations used:

AP
anchored protein
APN
aminopeptidase N
BC
bile canalicular
BL
basolateral
CTX
cholera toxin
DPPIV
dipeptidylpeptidase IV
mAb
monoclonal antibody
pAb
polyclonal antibody
Tf
transferrin
TMD
transmembrane domain.

Footnotes

This article was published online ahead of print in MBC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E09-04-0275) on July 15, 2009.

REFERENCES

  • Aït-Slimane T., Hoekstra D. Sphingolipid trafficking and protein sorting in epithelial cells. FEBS Lett. 2002;529:54–59. [PubMed]
  • Aït-Slimane T., Trugnan G., Van IJzendoorn S. C., Hoekstra D. Raft-mediated trafficking of apical resident proteins occurs in both direct and transcytotic pathways in polarized hepatic cells: role of distinct lipid microdomains. Mol. Biol. Cell. 2003;14:611–624. [PMC free article] [PubMed]
  • Anderson R. G., Kamen B. A., Rothberg K. G., Lacey S. W. Potocytosis: sequestration and transport of small molecules by caveolae. Science. 1992;255:410–411. [PubMed]
  • Benmerah A., Bayrou M., Cerf-Bensussan N., Dautry-Varsat A. Inhibition of clathrin-coated pit assembly by an Eps15 mutant. J. Cell Sci. 1999;112:1303–1311. [PubMed]
  • Chadda R., Howes M. T., Plowman S. J., Hancock J. F., Parton R. G., Mayor S. Cholesterol-sensitive Cdc42 activation regulates actin polymerization for endocytosis via the GEEC pathway. Traffic. 2007;8:702–717. [PubMed]
  • Chatterjee S., Mayor S. The GPI-anchor and protein sorting. Cell. Mol. Life Sci. 2001;58:1969–1987. [PubMed]
  • Conner S. D., Schmid S. L. Regulated portals of entry into the cell. Nature. 2003;422:37–44. [PubMed]
  • Ellis S., Mellor H. Regulation of endocytic traffic by rho family GTPases. Trends Cell Biol. 2000;10:85–88. [PubMed]
  • Fivaz M., Vilbois F., Thurnheer S., Pasquali C., Abrami L., Bickel P. E., Parton R. G., van der Goot F. G. Differential sorting and fate of endocytosed GPI-anchored proteins. EMBO J. 2002;21:3989–4000. [PMC free article] [PubMed]
  • Frick M., Bright N. A., Riento K., Bray A., Merrified C., Nichols B. J. Coassembly of flotillins induces formation of membrane microdomains, membrane curvature, and vesicle budding. Curr. Biol. 2007;17:1151–1156. [PubMed]
  • Fujimoto T. GPI-anchored proteins, glycosphingolipids, and sphingomyelin are sequestered to caveolae only after crosslinking. J. Histochem. Cytochem. 1996;44:929–941. [PubMed]
  • Fujimoto T., Kogo H., Nomura R., Une T. Isoforms of caveolin-1 and caveolar structure. J. Cell Sci. 2000;113:3509–3517. [PubMed]
  • Fu Y., Hoang A., Escher G., Parton R. G., Krozowski Z., Sviridov D. Expression of caveolin-1 enhances cholesterol efflux in hepatic cells. J. Biol. Chem. 2004;279:14140–14146. [PubMed]
  • Glebov O. O., Bright N. A., Nichols B. J. Flotillin-1 defines a clathrin-independent endocytic pathway in mammalian cells. Nat. Cell Biol. 2006;8:46–54. [PubMed]
  • Hemery I., Durand-Schneider A. M., Feldmann G., Vaerman J. P., Maurice M. The transcytotic pathway of an apical plasma membrane protein (B10) in hepatocytes is similar to that of IgA and occurs via a tubular pericentriolar compartment. J. Cell Sci. 1996;109:1215–1227. [PubMed]
  • Hinrichsen L., Harborth J., Andrees L., Weber K., Ungewickell E. J. Effect of clathrin heavy chain- and alpha-adaptin-specific small inhibitory RNAs on endocytic accessory proteins and receptor trafficking in HeLa cells. J. Biol. Chem. 2003;278:45160–45170. [PubMed]
  • Ihrke G., Martin G. V., Shanks M. R., Schrader M., Schroer T. A., Hubbard A. L. Apical plasma membrane proteins and endolyn-78 travel through a subapical compartment in polarized WIF-B hepatocytes. J. Cell Biol. 1998;141:115–133. [PMC free article] [PubMed]
  • Kumari S., Mayor S. ARF1 is directly involved in dynamin-independent endocytosis. Nat. Cell Biol. 2008;10:30–41. [PubMed]
  • Kurzchalia T. V., Parton R. G. Membrane microdomains and caveolae. Curr. Opin. Cell Biol. 1999;11:424–431. [PubMed]
  • Lamaze C., Dujeancourt A., Baba T., Lo C. G., Benmerah A., Dautry-Varsat A. Interleukin 2 receptors and detergent-resistant membrane domains define a clathrin-independent endocytic pathway. Mol. Cell. 2001;7:661–671. [PubMed]
  • Lang D. M., Lommel S., Jung M., Ankerhold R., Petrausch B., Laessing U., Wiechers M. F., Plattner H., Stuermer C. A. Identification of reggie-1 and reggie-2 as plasma membrane-associated proteins which cocluster with activated GPI-anchored cell adhesion molecules in non-caveolar micropatches in neurons. J. Neurobiol. 1998;37:502–523. [PubMed]
  • Mayor S., Pagano R. E. Pathways of clathrin-independent endocytosis. Nat. Rev. Mol. Cell Biol. 2007;8:603–612. [PubMed]
  • Mayor S., Riezman H. Sorting GPI-anchored proteins. Nat. Rev. Mol. Cell Biol. 2004;5:110–120. [PubMed]
  • McNiven M. A., Cao H., Pitts K. R., Yoon Y. The dynamin family of mechanoenzymes: pinching in new places. Trends Biochem. Sci. 2000;25:115–120. [PubMed]
  • Mostov K. E., Verges M., Altschuler Y. Membrane traffic in polarized epithelial cells. Curr. Opin. Cell Biol. 2000;12:483–490. [PubMed]
  • Nichols B. J. GM1-containing lipid rafts are depleted within clathrin-coated pits. Curr. Biol. 2003;13:686–690. [PubMed]
  • Nichols B. J. A distinct class of endosome mediates clathrin-independent endocytosis to the Golgi complex. Nat. Cell Biol. 2002;4:374–378. [PubMed]
  • Nyasae L. K., Hubbard A. L., Tuma P. L. Transcytotic efflux from early endosomes is dependent on cholesterol and glycosphingolipids in polarized hepatic cells. Mol. Biol. Cell. 2003;14:2689–2705. [PMC free article] [PubMed]
  • Okamoto C. T., Shia S. P., Bird C., Mostov K. E., Roth M. G. The cytoplasmic domain of the polymeric immunoglobulin receptor contains two internalization signals that are distinct from its basolateral sorting signal. J. Biol. Chem. 1992;267:9925–9932. [PubMed]
  • Parton R. G., Joggerst B., Simons K. Regulated internalization of caveolae. J. Cell Biol. 1994;127:1199–1215. [PMC free article] [PubMed]
  • Parton R. G., Simons K. The multiple faces of caveolae. Nat. Rev. Mol. Cell Biol. 2007;8:185–194. [PubMed]
  • Payne C. K., Jones S. A., Chen C., Zhuang X. Internalization and trafficking of cell surface proteoglycans and proteoglycan-binding ligands. Traffic. 2007;8:389–401. [PMC free article] [PubMed]
  • Perret E., Lakkaraju A., Deborde S., Schreiner R., Rodriguez-Boulan E. Evolving endosomes: how many varieties and why? Curr. Opin. Cell Biol. 2005;17:423–434. [PubMed]
  • Rodriguez-Boulan E., Kreitzer G., Müsch A. Organization of vesicular trafficking in epithelia. Nat. Rev. Mol. Cell Biol. 2005;6:233–247. [PubMed]
  • Rodriguez-Boulan E., Powell S. K. Polarity of epithelial and neuronal cells. Annu. Rev. Cell Biol. 1992;8:395–427. [PubMed]
  • Rojas R., Ruiz W. G., Leung S. M., Jou T. S., Apodaca G. Cdc42-dependent modulation of tight junctions and membrane protein traffic in polarized Madin-Darby canine kidney cells. Mol. Biol. Cell. 2001;12:2257–2274. [PMC free article] [PubMed]
  • Sabharanjak S., Sharma P., Parton R. G., Mayor S. GPI-anchored proteins are delivered to recycling endosomes via a distinct cdc42-regulated, clathrin-independent pinocytic pathway. Dev. Cell. 2002;2:411–423. [PubMed]
  • Sandvig K., van Deurs B. Transport of protein toxins into cells: pathways used by ricin, cholera toxin and Shiga toxin. FEBS Lett. 2002;529:49–53. [PubMed]
  • Schell M. J., Maurice M., Stieger B., Hubbard A. L. 5′nucleotidase is sorted to the apical domain of hepatocytes via an indirect route. J. Cell Biol. 1992;119:1173–1182. [PMC free article] [PubMed]
  • Stuermer C. A., Lang D. M., Kirsch F., Wiechers M., Deininger S. O., Plattner H. Glycosylphosphatidyl inositol-anchored proteins and fyn kinase assemble in noncaveolar plasma membrane microdomains defined by reggie-1 and -2. Mol. Biol. Cell. 2001;12:3031–3045. [PMC free article] [PubMed]
  • Torgersen M. L., Skretting G., van Deurs B., Sandvig K. Internalization of cholera toxin by different endocytic mechanisms. J. Cell Sci. 2001;114:3737–3747. [PubMed]
  • Tuma P. L., Hubbard A. L. Transcytosis: crossing cellular barriers. Physiol. Rev. 2003;83:871–932. [PubMed]
  • Volonté D., Galbiati F., Li S., Nishiyama K., Okamoto T., Lisanti M. P. Flotillins/cavatellins are differentially expressed in cells and tissues and form a hetero-oligomeric complex with caveolins in vivo. Characterization and epitope-mapping of a novel flotillin-1 monoclonal antibody probe. J. Biol. Chem. 1999;274:12702–12709. [PubMed]

Articles from Molecular Biology of the Cell are provided here courtesy of American Society for Cell Biology
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...