Targeting of glycosyl-phosphatidylinositol (GPI)-anchored proteins (GPI-APs) in polarized epithelial cells depends on their association with detergent-resistant membrane microdomains called rafts. In MDCK cells, GPI-APs associate with rafts in the trans-Golgi network and are directly delivered to the apical membrane. It has been shown that oligomerization is required for their stabilization in rafts and their apical targeting. In hepatocytes, GPI-APs are first delivered to the basolateral membrane and secondarily reach the apical membrane by transcytosis. We investigated whether oligomerization is required for raft association and apical sorting of GPI-APs in polarized HepG2 cells, and at which step of the pathway oligomerization occurs. Model proteins were wild-type GFP-GPI and a double cysteine GFP-GPI mutant, in which GFP dimerization was impaired. Unlike wild-type GFP-GPI, which was efficiently endocytosed and transcytosed to the apical surface, the double cysteine mutant was basolaterally internalized, but massively accumulated in early endosomes, and reached the bile canaliculi with delayed kinetics. The double cysteine mutant was less resistant to Triton X-100 extraction, and formed fewer high molecular weight complexes. We conclude from these results that, in hepatocytes, oligomerization plays a key role in targeting GPI-APs to the apical membrane, by increasing their affinity for rafts and allowing their transcytosis.

The plasma membrane (PM) of polarized epithelial cells is divided into two domains, apical and basolateral, separated by tight junctions. Each domain performs distinct functions and therefore has a unique protein and lipid composition. Generation and maintenance of the PM polarized organization is achieved through the combination of sorting signal and sorting machineries that allow PM proteins to be transported to the PM domain where they will be functional (Mostov et al., 2003).

Basolateral sorting signals are short cytoplasmic amino acid sequences often characterized by dileucine or tyrosine motifs (Matter and Mellman, 1994), which are recognized by clathrin adaptors (Sugimoto et al., 2002; Simmen et al., 2002). Apical sorting signals are more diverse. Several determinants have been identified, including membrane anchoring, N-and O-glycosylation, and unrelated proteinaceous motifs located in the cytoplasmic, transmembrane or extracellular domains (Rodriguez-Boulan et al., 2005). Glycosyl-phosphatidylinositol (GPI) anchor was the first apical sorting signal to be identified. Addition of a GPI anchor to a secretory protein targeted the chimeric protein to the apical membrane of MDCK cells (Lisanti et al., 1989). Likewise, substitution of the transmembrane domain of a basolateral PM protein by a GPI anchor redirected the protein to the apical membrane (Brown et al., 1989). Sorting of GPI-anchored proteins (GPI-APs) is thought to be mediated by affinity of the GPI anchor with sphingolipid- and cholesterol-rich microdomains called rafts (Simons and Ikonen, 1997). Indeed, GPI-APs associate with detergent-resistant membranes (DRMs) during their passage through the Golgi apparatus where sorting is thought to occur (Brown and Rose, 1992). However, the role of the GPI anchor as apical determinant has been questioned because certain GPI-APs are sorted to the basolateral domain (Zurzolo et al., 1993; McGwire et al., 1999; Sarnataro et al., 2002). It has been suggested that GPI-APs need additional features to be sorted apically. For instance, addition of N-glycosylation sites was necessary to allow a chimeric GPI-AP to be apically sorted in MDCK cells (Benting et al., 1999). Conversely, removal of the N-glycosylation sites of a natural GPI-AP resulted in the protein being targeted predominantly to the basolateral surface (Pang et al., 2004). However, recent results did not corroborate the direct role of glycosylation in apical targeting of GPI-APs (Catino et al., 2008). The role of glycosylation may be indirect by allowing apical proteins to cluster in rafts (Rodriguez-Boulan and Gonzalez, 1999; Schuck and Simons, 2004). Independently of glycosylation, oligomerization by itself was recognized as a key feature for apical sorting of GPI-APs. Paladino et al. showed that apical but not basolateral GPI-APs formed oligomers, and impairment of oligomerization led to their mis-sorting in an unpolarized fashion (Paladino et al., 2004; Paladino et al., 2007). Like glycosylation, oligomerization may increase raft affinity of GPI-APs, promoting their stabilization in apical carriers.

Most studies of GPI-APs targeting have been carried out in simple epithelial cells, especially MDCK cells. In these cells, it is generally assumed that GPI-APs are targeted to the apical PM directly from the trans-Golgi network (Lisanti et al., 1990; Arreaza and Brown, 1995; Paladino et al., 2006; Hua et al., 2006). By contrast, apical trafficking of transmembrane and GPI-APs in hepatocytes generally occurs through transcytosis (Bartles et al., 1987; Schell et al., 1992; Maurice et al., 1994; Bastaki et al., 2002), with the exception of multi-spanning transmembrane proteins, which are transported via the direct route (Kipp and Arias, 2000). We have previously demonstrated that GPI-APs are sorted to the apical PM of HepG2 cells by transcytosis and that this process is dependent on their association with Triton X-100 DRMs (Aït-Slimane et al., 2003). Moreover, we have shown that flotillin-2 controls the first step of GPI-AP transcytosis, i.e., basolateral endocytosis (Aït-Slimane et al., 2009). Here we investigated whether oligomerization plays a role in the apical sorting of GPI-AP in hepatocytes, and at which step, i.e. Golgi-to-basolateral or basolateral-to-apical membranes, or both. For these studies, we used a GFP-GPI construct in which green fluorescent protein (GFP) is fused to the GPI anchor attachment signal from the folate receptor and a double cysteine GFP-GPI mutant (S49/71), in which the two cysteines, Cys 49 and Cys 71 involved in GFP oligomerization were mutated. These constructs were stably expressed in HepG2 cells. We found that the double cysteine mutant accumulated in intracellular structures and its basolateral-to-apical transcytosis was less efficient compared to its wild-type counterpart. We show that the S49/71 mutant was not able to oligomerize. We also found that the mutant was much less DRM-associated. Together our results demonstrate that, in hepatocytes, oligomerization and association with lipid rafts are required for efficient basolateral-to-apical transcytosis of GPI-APs.

Subcellular localization of wild-type GFP-GPI and GFP-GPI S49/71 in HepG2 cells

In order to investigate the role of oligomerization in the sorting of GPI-AP in hepatocytes, we used the reported protein GFP fused to the GPI anchor attachment signal of the folate receptor and the double cysteine mutant S49/71. These two cysteines are involved in GFP oligomerization through disulfide bond formation (Jain et al., 2001). Both constructs contained a c-myc tag at the NH2 terminus allowing their detection using an anti-c-myc antibody. Wild-type GFP-GPI and the mutant were stably transfected in HepG2 cells. Wild-type GFP-GPI was principally expressed at the bile canalicular membrane with some basolateral staining (Fig. 1A), as already reported for endogenous GPI-APs in polarized hepatocellular cell lines (Aït-Slimane et al., 2003; Nyasae et al., 2003). The mutant showed a different pattern of localization. Not only the bile canalicular membrane was labeled, but there was also a strong labeling of intracellular vesicles (Fig. 1A). This result suggested that the S49/71 mutant had a trafficking defect. In order to more precisely determine whether the intracellular vesicles corresponded to a block in the exocytotic (Golgi-to-basolateral) or transcytotic (basolateral-to-apical) pathway, the localization of the mutant was compared to those of giantin and TGN46, which are markers of the Golgi apparatus and trans-Golgi network respectively. As shown in Fig. 1B, no significant colocalization of the S49/71 mutant was detected with either one of the two markers. This result suggests that the exocytic transport of the mutant was not defective, and that the transcytotic transport was more likely impaired.

Fig. 1.

The double cysteine mutant is retained in intracellular vesicles. (A) HepG2 cells expressing GFP-glycosyl-phosphatidylinositol (GPI) and the double cysteine GFP-GPI mutant (S49/71) were grown on coverslips. GFP-GPI-expressing cells were fixed and directly examined for the fluorescence of GFP. Cells expressing the mutant were fixed, permeabilized and stained with an antibody against c-myc followed by Alexa-Fluor 488-conjugated secondary antibody. Confocal microscopy shows that at steady state the GFP-GPI preferentially localizes at the bile canalicular (BC) surface (asterisk), whereas the double cysteine mutant is detected at the BC surface and in intracellular vesicles. Nuclei are stained with Topro-3. (B) HepG2 cells expressing the double cysteine GFP-GPI mutant (S49/71) were grown on coverslips, fixed, permeabilized and stained with an antibody against c-myc followed by anti-giantin or anti-TGN46 antibodies and then incubated with Alexa-Fluor 594-conjugated secondary antibody and Alexa-Fluor 488-conjugated secondary antibody. Merged images show that at steady state no significant colocalization was detected between the double cysteine mutant and either marker. BCs are indicated by asterisks. Scale bars: 10 µm.

Fig. 1.

The double cysteine mutant is retained in intracellular vesicles. (A) HepG2 cells expressing GFP-glycosyl-phosphatidylinositol (GPI) and the double cysteine GFP-GPI mutant (S49/71) were grown on coverslips. GFP-GPI-expressing cells were fixed and directly examined for the fluorescence of GFP. Cells expressing the mutant were fixed, permeabilized and stained with an antibody against c-myc followed by Alexa-Fluor 488-conjugated secondary antibody. Confocal microscopy shows that at steady state the GFP-GPI preferentially localizes at the bile canalicular (BC) surface (asterisk), whereas the double cysteine mutant is detected at the BC surface and in intracellular vesicles. Nuclei are stained with Topro-3. (B) HepG2 cells expressing the double cysteine GFP-GPI mutant (S49/71) were grown on coverslips, fixed, permeabilized and stained with an antibody against c-myc followed by anti-giantin or anti-TGN46 antibodies and then incubated with Alexa-Fluor 594-conjugated secondary antibody and Alexa-Fluor 488-conjugated secondary antibody. Merged images show that at steady state no significant colocalization was detected between the double cysteine mutant and either marker. BCs are indicated by asterisks. Scale bars: 10 µm.

The transcytotic rate of the S49/71 mutant is slower

In order to determine whether accumulation of the S49/71 mutant in intracellular vesicles was due to a transcytosis defect, we compared the kinetic of transcytosis of wild-type and mutant GFP-GPIs. We followed in parallel the trafficking of CD59, a GPI-AP endogenously expressed in HepG2 cells. HepG2 cells were incubated for 30 min at 0°C with antibodies against c-myc or CD59 to allow their binding to the surface. Trafficking was initiated by raising the temperature to 37°C. After various chase periods, the cells were fixed, permeabilized and incubated with fluorescently labeled secondary antibodies. In the case of the mutant that does not fluoresce, the bile canaliculi were visualized by staining with an antibody against dipeptidyl-peptidase IV (DPPIV). As shown in Fig. 2, after incubation on ice, the antigen-antibody complexes were exclusively detected at the basolateral membrane. After 60 min at 37°C, some wild-type GFP-GPI had already completed transcytosis as indicated by the detection of internalized antibodies complexed with antigens both in intracellular vesicles and at the membrane of bile canaliculi. Subsequently, labeling of the bile canalicular membranes increased while the basolateral staining decreased. The endogenous GPI-AP, CD59 behaved very similarly, with transcytosis being completed within about 3 h. After 6 h, a prominent and homogeneous distribution of both antibody-conjugated proteins at the bile canalicular surface was apparent. In contrast, in the case of the double cysteine mutant, the antibody–antigen complexes did internalize, but they accumulated intracellularly and the bile canalicular membranes began to be labeled only after 3 h. After 6 h, the bile canalicular membranes were labeled, but intracellular vesicles were still detected (Fig. 2H). These results show that mutation of the cysteines involved in GFP-GPI dimerization delays transport of the mutant from the basolateral to the apical surface, and suggest that oligomerization plays a role in the transcytosis efficiency of this GPI-anchored protein in HepG2 cells.

Fig. 2.

The double cysteine GFP-GPI mutant (S49/71) is internalized and transcytosed with slower kinetics. Non-transfected HepG2 cells and HepG2 cells expressing GFP-GPI or the double cysteine GFP-GPI mutant (S49/71) were incubated for 30 min at 0°C with anti-CD59 and anti-c-myc antibodies, respectively. Immediately (A, A′ and E) or after incubation at 37°C for 60 min (B, B′ and F), 180 min (C, C′ and G) or 360 min (D, D′ and H), cells were fixed and permeabilized. Primary antibodies were visualized with secondary Alexa-Fluor 594-conjugated antibody. Panels E′, F′, G′ and H′ show the steady-state distribution of DPPIV stained with anti-DPPIV and Alexa-Fluor 488-conjugated secondary antibody. Nuclei are stained with Topro-3. BCs are marked by asterisks. Scale bars: 10 µm.

Fig. 2.

The double cysteine GFP-GPI mutant (S49/71) is internalized and transcytosed with slower kinetics. Non-transfected HepG2 cells and HepG2 cells expressing GFP-GPI or the double cysteine GFP-GPI mutant (S49/71) were incubated for 30 min at 0°C with anti-CD59 and anti-c-myc antibodies, respectively. Immediately (A, A′ and E) or after incubation at 37°C for 60 min (B, B′ and F), 180 min (C, C′ and G) or 360 min (D, D′ and H), cells were fixed and permeabilized. Primary antibodies were visualized with secondary Alexa-Fluor 594-conjugated antibody. Panels E′, F′, G′ and H′ show the steady-state distribution of DPPIV stained with anti-DPPIV and Alexa-Fluor 488-conjugated secondary antibody. Nuclei are stained with Topro-3. BCs are marked by asterisks. Scale bars: 10 µm.

The double cysteine mutant forms fewer high molecular weight complexes

Mutation of the cysteines was designed in order to prevent oligomerization of GFP-GPI. To verify that the mutant was indeed defective in oligomerization, we studied its ability to form high molecular weight (HMW) complexes compared to wild-type GFP-GPI and CD59. We used velocity gradients, in which proteins sediment according to their molecular weight (Paladino et al., 2007). Flotillin-2, a 47 kDa protein, which forms HMW complexes (Neumann-Giesen et al., 2004), was taken as a positive control. We found that similar to CD59, wild-type GFP-GPI was essentially detected in fractions corresponding to complexes of high molecular weights. By contrast, the double cysteine mutant was mainly detected in the 30 kDa range (fraction 3), as expected for the GFP monomer, with only a small fraction being recovered as HMW complexes (fractions 5–7) (Fig. 3).

Fig. 3.

The double cysteine mutant forms fewer high molecular weight complexes. HepG2 cells expressing GFP-GPI or the double cysteine GFP-GPI mutant (S49/71) were lysed in buffer containing 0.4% SDS and 0.2% Triton X-100 and run through a non-linear 5–30% sucrose gradient. Fractions of 500 µl were collected from the top to the bottom of the gradient. Proteins were immunoprecipitated with the polyclonal anti-c-myc antibody, run on SDS-PAGE and revealed by western blotting using the monoclonal anti-GFP. CD59 was detected by western blotting using the polyclonal anti-CD59 antibody. Flotillin-2 was used as a control protein, which forms high molecular weight (HMW) complexes. The molecular weight of the monomeric forms of each protein is indicated. The position on the gradient of molecular weight markers is indicated on top of the panel.

Fig. 3.

The double cysteine mutant forms fewer high molecular weight complexes. HepG2 cells expressing GFP-GPI or the double cysteine GFP-GPI mutant (S49/71) were lysed in buffer containing 0.4% SDS and 0.2% Triton X-100 and run through a non-linear 5–30% sucrose gradient. Fractions of 500 µl were collected from the top to the bottom of the gradient. Proteins were immunoprecipitated with the polyclonal anti-c-myc antibody, run on SDS-PAGE and revealed by western blotting using the monoclonal anti-GFP. CD59 was detected by western blotting using the polyclonal anti-CD59 antibody. Flotillin-2 was used as a control protein, which forms high molecular weight (HMW) complexes. The molecular weight of the monomeric forms of each protein is indicated. The position on the gradient of molecular weight markers is indicated on top of the panel.

The double cysteine mutant accumulates in early endosomes

In a previous study, we have shown that, in HepG2 cells, GPI-APs are internalized in flotillin-2-containing endosomes, whereas TMD proteins are internalized in Early Endosome Antigen 1 (EEA1)-containing endosomes (Aït-Slimane et al., 2009). In order to test whether the double cysteine mutant accumulated in flotillin-2-positive structures, we searched for colocalization of the S49/71 mutant with flotillin-2. As shown in Fig. 4A, the double cysteine mutant did not colocalize with flotillin-2, whereas a large colocalization was observed between the wild-type GFP-GPI and flotillin-2. We therefore searched for colocalization with EEA1. As shown in Fig. 4B, abundant puncta containing both the double cysteine mutant and EEA1 were observed at the periphery of the cells, indicating that the S49/71 accumulated in early endosomes.

Fig. 4.

The double cysteine mutant accumulates in early endosomes. (A) HepG2 cells expressing the double cysteine mutant (S49/71) (top panel) or wild-type GFP-GPI (bottom panel) were grown on coverslips, fixed, permeabilized and stained with an antibody against c-myc followed by anti-flotillin-2 (Flot2) antibody and incubated with Alexa-Fluor 594-, 488- or 647-conjugated secondary antibodies. Because wild-type GFP-GPI is not significantly detected intracellularly at steady state, HepG2 cells expressing wild-type GFP-GPI were incubated for 60 min at 37°C with antibody to c-myc. At the end of the 37°C incubation, antibody–antigen complexes remaining at the basolateral surface were removed by acid washing. Internalized GFP-GPI was visualized by Alexa-Fluor 594-conjugated secondary antibody (shown in green pseudocolor) and flotillin-2 was visualized by Alexa-Fluor 647-conjugated secondary antibody (shown in red pseudocolor). (B) HepG2 cells expressing the double cysteine GFP-GPI mutant (S49/71) were grown on coverslips, fixed, permeabilized and stained with an antibody against c-myc followed by anti-Early Endosome Antigen 1 (EEA1) antibody and then incubated with Alexa-Fluor 594-conjugated secondary antibody and Alexa-Fluor 488-conjugated secondary antibody. Nuclei are stained with Topro-3. BCs are marked by asterisks. Scale bars: 10 µm.

Fig. 4.

The double cysteine mutant accumulates in early endosomes. (A) HepG2 cells expressing the double cysteine mutant (S49/71) (top panel) or wild-type GFP-GPI (bottom panel) were grown on coverslips, fixed, permeabilized and stained with an antibody against c-myc followed by anti-flotillin-2 (Flot2) antibody and incubated with Alexa-Fluor 594-, 488- or 647-conjugated secondary antibodies. Because wild-type GFP-GPI is not significantly detected intracellularly at steady state, HepG2 cells expressing wild-type GFP-GPI were incubated for 60 min at 37°C with antibody to c-myc. At the end of the 37°C incubation, antibody–antigen complexes remaining at the basolateral surface were removed by acid washing. Internalized GFP-GPI was visualized by Alexa-Fluor 594-conjugated secondary antibody (shown in green pseudocolor) and flotillin-2 was visualized by Alexa-Fluor 647-conjugated secondary antibody (shown in red pseudocolor). (B) HepG2 cells expressing the double cysteine GFP-GPI mutant (S49/71) were grown on coverslips, fixed, permeabilized and stained with an antibody against c-myc followed by anti-Early Endosome Antigen 1 (EEA1) antibody and then incubated with Alexa-Fluor 594-conjugated secondary antibody and Alexa-Fluor 488-conjugated secondary antibody. Nuclei are stained with Topro-3. BCs are marked by asterisks. Scale bars: 10 µm.

The double cysteine mutant and DPPIV share the same transcytotic pathway

To further characterize the compartment in which the mutant accumulated, we performed co-trafficking experiment using antibodies to DPPIV and CD59, a GPI-anchored protein endogenously expressed in HepG2 cells, and we searched for colocalization with the S49/71 mutant. S49/71-expressing HepG2 cells were continuously labeled either with anti-DPPIV and anti-c-myc antibodies or with anti-CD59 and anti-c-myc antibodies for 60 min at 37°C. The cells were fixed, permeabilized and incubated with fluorescently labeled secondary antibodies. We observed a large colocalization between internalized S49/71 and DPPIV but not between internalized S49/71 and CD59 (Fig. 5A). Quantification of the results showed that 80% of DPPIV but only ∼30% of CD59 colocalized with the double cysteine mutant (Fig. 5B). These results suggest that impairment of GFP-GPI oligomerization leads to its sorting into the pathway used by TMD proteins.

Fig. 5.

(A) The double cysteine mutant and dipeptidyl-peptidase IV (DPPIV) share the same transcytotic pathway. HepG2 cells expressing the double cysteine GFP-GPI mutant (S49/71) were allowed to internalize DPPIV and anti-c-myc or CD59 and anti-c-myc antibodies continuously for 60 min at 37°C. At the end of the 37°C incubation, antibody–antigen complexes remaining at the basolateral surface were removed by acid washing. Cells were fixed, permeabilized and internalized antibodies were visualized with Alexa-Fluor 488- and 594-conjugated secondary antibodies. (B) The colocalization of the double cysteine mutant with DPPIV and CD59 was quantified using the ImageJ 1.41 measure colocalization function on multiple confocal sections of at least 30 cells in three independent experiments. BCs are marked by asterisks. Scale bar: 10 µm.

Fig. 5.

(A) The double cysteine mutant and dipeptidyl-peptidase IV (DPPIV) share the same transcytotic pathway. HepG2 cells expressing the double cysteine GFP-GPI mutant (S49/71) were allowed to internalize DPPIV and anti-c-myc or CD59 and anti-c-myc antibodies continuously for 60 min at 37°C. At the end of the 37°C incubation, antibody–antigen complexes remaining at the basolateral surface were removed by acid washing. Cells were fixed, permeabilized and internalized antibodies were visualized with Alexa-Fluor 488- and 594-conjugated secondary antibodies. (B) The colocalization of the double cysteine mutant with DPPIV and CD59 was quantified using the ImageJ 1.41 measure colocalization function on multiple confocal sections of at least 30 cells in three independent experiments. BCs are marked by asterisks. Scale bar: 10 µm.

Dynamin is required for many endocytic processes, especially the clathrin-dependent endocytosis (Conner and Schmid, 2003). Therefore, we investigated the effect of disruption of dynamin function on the internalization of the S49/71 mutant, by using dynasore, a cell permeable small molecule shown to be a potent inhibitor of dynamin function (Macia et al., 2006). Treatment of HepG2 cells by dynasore significantly impaired internalization of the S49/71 mutant similarly to that of transferring (Tf), a well established marker of the clathrin pathway, and to that of DPPIV, which has been shown to be internalized via a clathrin-dependent pathway in hepatic cells (Aït-Slimane et al., 2009). In contrast to Tf and DPPIV, the internalization of GFP-GPI was not obviously impaired by dynasore (Fig. 6A). These morphological observations were quantified by measuring the fluorescence intensity in cells treated with dynasore relative to that in control cells. Internalization of Tf, DPPIV and S49/71 was reduced from ∼95% to 20%, whereas internalization of GFP-GPI was minimally affected, shifting from ∼90% to 65% (Fig. 6B). These results provide additional support for a clathrin-dependent way of internalization of the double cysteine mutant.

Fig. 6.

Dynasore blocks the internalization of the double cysteine mutant. (A) HepG2 cells expressing the S49/71 mutant were treated with 80 µM dynasore for 15 min, followed by internalization assays for DPPIV, S49/71 and GFP-GPI for 45 min at 37°C and compared with control dimethyl sulfoxide-treated cells. For transferring (Tf) internalization assay, cells were treated with 80 µM dynasore for 30 min and then allowed to internalize Alexa-Fluor 594-labeled Tf for 5 min at 37°C. Dynasore was maintained during the internalization processes. At the end of the 37°C incubation, Tf and antibody–antigen complexes remaining at the basolateral surface were removed by acid washing. Cells were fixed, permeabilized and internalized antibodies were visualized with Alexa-Fluor 594-conjugated secondary antibody. (B) The amount of internalized Tf, DPPIV, S49/71 and GFP-GPI was measured in dynasore-treated cells and expressed as a percentage of that internalized in control cells. Filled bars represent control cells and open bars represent cells treated with dynasore. Quantification of three independent experiments is shown. BCs are marked by asterisks. Scale bar: 10 µm.

Fig. 6.

Dynasore blocks the internalization of the double cysteine mutant. (A) HepG2 cells expressing the S49/71 mutant were treated with 80 µM dynasore for 15 min, followed by internalization assays for DPPIV, S49/71 and GFP-GPI for 45 min at 37°C and compared with control dimethyl sulfoxide-treated cells. For transferring (Tf) internalization assay, cells were treated with 80 µM dynasore for 30 min and then allowed to internalize Alexa-Fluor 594-labeled Tf for 5 min at 37°C. Dynasore was maintained during the internalization processes. At the end of the 37°C incubation, Tf and antibody–antigen complexes remaining at the basolateral surface were removed by acid washing. Cells were fixed, permeabilized and internalized antibodies were visualized with Alexa-Fluor 594-conjugated secondary antibody. (B) The amount of internalized Tf, DPPIV, S49/71 and GFP-GPI was measured in dynasore-treated cells and expressed as a percentage of that internalized in control cells. Filled bars represent control cells and open bars represent cells treated with dynasore. Quantification of three independent experiments is shown. BCs are marked by asterisks. Scale bar: 10 µm.

Both wild-type and S49/71 mutant are incorporated into DRMs but with different efficiencies

GPI-APs associate with rafts and this association is thought to contribute to their apical sorting. We have already shown that, in HepG2 cells, indirect apical sorting of GPI-APs occurs via Triton X-100 resistant microdomains (Aït-Slimane et al., 2003). In order to determine whether mutation of the two cysteines affected the ability of GFP-GPI to associate with rafts, we compared the detergent-resistance of wild-type and mutant GFP-GPIs. Cells were lysed in Triton X-100 at 4°C and DRMs were prepared as described (Aït-Slimane et al., 2003). Wild-type GFP-GPI was mainly recovered in the floating DRM fractions [3–5], which also contained flotillin-2, an established marker for DRM fractions (Fig. 7A). By contrast, the double cysteine mutant was much less associated with DRM fractions [3–5], the majority of the protein being recovered in the soluble fractions [7–12]. These results demonstrate that impairment of GFP-GPI oligomerization affects its ability to associate with DRMs despite the presence of the GPI anchor.

Fig. 7.

GFP-GPI and GFP-GPI S49/71 are incorporated into detergent-resistant membranes with different efficiencies. (A) HepG2 cells expressing GFP-GPI or the double cysteine GFP-GPI mutant (S49/71) were lysed in TNE/TX-100 buffer (see Materials and Methods) at 4°C and run through a 5–40% sucrose gradient. Fractions of 1 ml were collected from top to bottom after centrifugation to equilibrium. Proteins were immunoprecipitated with an antibody against c-myc, run on SDS-PAGE and revealed by western blotting using the monoclonal anti-GFP antibody. Flotillin-2 was analyzed in parallel as a marker to indicate DRM fractions. (B) HepG2 cells expressing GFP-GPI or the double cysteine GFP-GPI mutant (S49/71) were pulse chased with [35S]methionine for 10 min, followed by incubation in chase medium for the indicated times. After extraction in TNE/TX-100 buffer at 4°C, both the soluble (S; supernatant) and insoluble (I; pellet) fractions were collected by centrifugation, immunoprecipitated with the polyclonal anti-c-myc antibody and analyzed by SDS-PAGE. (C) Quantification of three independent experiments. Filled bars represent soluble proteins and open bars represent insoluble proteins.

Fig. 7.

GFP-GPI and GFP-GPI S49/71 are incorporated into detergent-resistant membranes with different efficiencies. (A) HepG2 cells expressing GFP-GPI or the double cysteine GFP-GPI mutant (S49/71) were lysed in TNE/TX-100 buffer (see Materials and Methods) at 4°C and run through a 5–40% sucrose gradient. Fractions of 1 ml were collected from top to bottom after centrifugation to equilibrium. Proteins were immunoprecipitated with an antibody against c-myc, run on SDS-PAGE and revealed by western blotting using the monoclonal anti-GFP antibody. Flotillin-2 was analyzed in parallel as a marker to indicate DRM fractions. (B) HepG2 cells expressing GFP-GPI or the double cysteine GFP-GPI mutant (S49/71) were pulse chased with [35S]methionine for 10 min, followed by incubation in chase medium for the indicated times. After extraction in TNE/TX-100 buffer at 4°C, both the soluble (S; supernatant) and insoluble (I; pellet) fractions were collected by centrifugation, immunoprecipitated with the polyclonal anti-c-myc antibody and analyzed by SDS-PAGE. (C) Quantification of three independent experiments. Filled bars represent soluble proteins and open bars represent insoluble proteins.

The double cysteine mutant is slowly incorporated into Triton X-100-insoluble membranes

Association of GPI-APs with DRMs is thought to occur along the biosynthetic pathway, in the trans-Golgi network (Brown and Rose, 1992). To determine at which step DRM association of GPI-GFP S49/71 was affected, we performed pulse/chase experiments. Proteins were pulse-labeled with [35S]methionine/cysteine for 10 min, and the detergent resistance of the wild-type and mutant GFP-GPIs was analyzed after different periods of chase. As shown in Fig. 7B,C, immediately after the pulse, both proteins were almost totally soluble. Wild-type GFP-GPI became progressively insoluble, 65% being associated with DRMs after 1.5 h, and 90% after 5 h. In contrast, GFP-GPI S49/71 was only weakly detected in the insoluble fractions after 1.5 h, and only 50% was in DRMs after 5 h. These results show that the rate of DRM association of GFP-GPI is greatly influenced by its ability to oligomerize.

In this study, we investigated whether oligomerization is required for apical sorting of GPI-APs in hepatocytes, and at which step of the pathway oligomerization occurs. Using the GPI-GFP chimera, which oligomerizes and a mutant, in which the two cysteines involved in GFP oligomerization were mutated, we show that oligomerization is required for raft association and efficient basolateral to apical transcytosis of GFP-GPI in hepatocytes. Previous studies in MDCK cells have shown that oligomerization of GPI-APs occurs in the TGN and that impairment of oligomerization leads to mis-sorting of GPI-APs to the basolateral plasma membrane (Paladino et al., 2004; Paladino et al., 2007). In HepG2 cells, in which apical resident proteins are first delivered to the basolateral membrane, the double cysteine GFP-GPI mutant (S49/71), which does not oligomerize accumulated intracellularly. The step which was impaired was not between the Golgi and the basolateral membrane because the S49/71 mutant did not colocalize with Golgi markers. Rather, it is transcytosis between the basolateral to the apical membrane which was dramatically delayed, suggesting a sorting defect along the transcytotic pathway. The difference between MDCK cells and HepG2 cells is in accordance with the fact that in MDCK cells the TGN is the major site for sorting of apical membrane proteins, whereas in hepatocytes, sorting occurs in endosomes after proteins have reached the basolateral plasma membrane (Wang and Boyer, 2004).

The main mechanism for apical delivery of GPI-APs is their clustering into glycosphingolipid/cholesterol-enriched microdomains, called rafts (Simons and Ikonen, 1997; Schuck and Simons, 2004). The model established in MDCK cells proposes that rafts are formed in the TGN and act as sorting platforms for inclusion of cargo proteins destined to the apical plasma membrane. In hepatic cells, we and others have previously shown that apical sorting of GPI-APs is also raft-dependent, although association with DRMs occurs with variable kinetics (Aït-Slimane et al., 2003; Nyasae et al., 2003). Herein, we observed that wild-type GFP-GPI became insoluble after 90 min of chase, i.e. much later than at the TGN as already observed for 5′-nucleotidase in Wif-B cells (Nyasae et al., 2003). These observations are in keeping with the fact that in hepatic cells, traffic from TGN to basolateral membrane does not depend on rafts while basolateral-to-apical transport is dependent on raft association (Nyasae et al., 2003). The S49/71 mutant was much less associated with Triton X-100 resistant membranes and became Triton X-100 insoluble with a delayed kinetics, compared to the wild-type GFP-GPI. Although in MDCK cells the S49/71 mutant was still largely associated with DRMs (Paladino et al., 2004), reduced ability to dimerize may lower affinity to DRMs. Indeed, it has been shown that oligomeric GPI-APs have a higher affinity for rafts than their monomers (Cunningham et al., 2003; Fivaz et al., 2002). Our results strongly suggest that transcytosis of GFP-GPI is directly linked to its ability to oligomerize and associate with rafts. As already proposed by Paladino et al. (Paladino et al., 2004), the role of oligomerization would be to increase raft affinity of GPI-APs. Strong association with rafts would promote their apical sorting, either along the direct route in MDCK cells (Paladino et al., 2004), or along the transcytotic pathway in hepatocytes, as observed here for GFP-GPI.

Intriguingly, accumulation of the oligomerization-defective S49/71 mutant occurred in an EEA1-positive compartment which is used by TMD proteins but not by GPI-APs. Thus, we have shown that transcytosis of GFP-GPI as well as other GPI-APs in hepatocytes does not proceed through EEA1-positive endosomes but involves a flotillin-2-positive pathway (Aït-Slimane et al., 2009). The internalization of the S49/71 mutant in vesicles containing the transcytosing TMD protein DPPIV, but not in vesicles containing the transcytosing endogenous GPI-AP CD59, suggests that the S49/71 mutant would not be able to enter flotillin-2-positive endosomes, and is mis-sorted to the clathrin-mediated endocytic/transcytotic pathway.

Nevertheless, inclusion into EEA1-positive endosomes cannot explain the slow transcytosis of the S49/71 mutant since DPPIV and other TMD proteins are efficiently transcytosed from this compartment. The molecular machinery of transcytosis is not well known. A number of molecules controlling vesicle budding and fusion have been identified (Tuma and Hubbard, 2003). Specific proteins like the raft-associated MAL (Ramnarayanan and Tuma, 2011) and multimeric protein complex, the retromer (Vergés et al., 2004) have been involved. However, the exact mechanism for cargo selection and sorting is still elusive. Rafts seem to play a decisive role in this process, as raft depletion specifically inhibits basolateral-to-apical transcytosis (Nyasae et al., 2003). On the other hand, a transcytotic signal in the cytoplasmic tail of the polymeric IgA receptor has been identified, suggesting that transcytosis of certain TMD proteins may be signal-mediated (Luton et al., 2009). Transcytosis of GPI-APs, which do not have a cytoplasmic domain, may depend solely on raft association. Therefore, the GFP-GPI mutant, which is weakly associated with rafts will not enter the transcytotic pathway and will remain in the endosome. In CHO cells, it has been observed that the fate of GPI-APs differs depending on their association with lipid domains, and that GPI-APs poorly associated with DRMs accumulate in recycling endosomes (Fivaz et al., 2002). The mutant GFP-GPI may nevertheless transcytose by default, but this would be a slow process as only a minor fraction of internalized membranes undergoes transcytosis. Therefore, oligomerization, by promoting raft association plays a key role in transcytosis of GPI-APs in hepatic cells.

Reagents and antibodies

Alexa Fluor 488 and Alexa Fluor 594 secondary antibodies, cell culture media, Topro-3 and the rabbit polyclonal anti-giantin were purchased from Invitrogen (Cergy-Pontoise, France). [35S]methionine/cysteine (‘in vitro cell lab mix’), Protein A-Sepharose, nitrocellulose and the ECL detection Kit were from GE Healthcare France (Orsay, France). The mouse monoclonal anti-GFP was from Roche Diagnostics (Meylan, France). Monoclonal and polyclonal anti-CD59 antibodies and monoclonal anti-TGN46 were from Abcam (Cambridge, UK). Monoclonal and polyclonal anti-c-myc antibodies and goat polyclonal anti-EEA1 were purchased from Santa Cruz Biotechnology (Santa-Cruz, CA). Monoclonal anti-flotillin-2 was purchased from BD Bioscience (Benton and Dickinson France, Le Pont-de-Claix, France). Peroxidase-conjugated secondary antibodies were from Rockland Immunochemicals (Gilbertville, PA). Secondary antibodies anti-mouse or rabbit IgG conjugated with FITC or Cy3 were purchased from Jackson Immunoresearch (Suffolk, England). Monoclonal anti-DPPIV and polyclonal anti-GFP were produced in the laboratory. All other reagents were obtained from Sigma (St-Quentin-Fallavier, France).

Cell culture and transfection

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. HepG2 cells were transfected with cDNAs encoding GFP-GPI and GFP-GPI S49/71 constructs using the Nucleofector system (Amaxa, Cologne, Germany). The GFP-GPI construct was obtained from S. Mayor (National Center for Biological Sciences, Bangalore, India) and the GFP-GPI S49/71 construct from Chiara Zurzulo (Institut Pasteur, Paris, France). Both constructs were produced by fusion of the folate receptor GPI-anchoring sequence with GFP and contained a c-myc tag at the N-terminus. Transfected cells were selected by geneticin sulfate (G418). Stable clones were isolated using cloning cylinders and screened by direct GFP fluorescence or immunofluorescence.

Indirect immunofluorescence and confocal microscopy

HepG2 cells grown on glass coverslips for 3 days were washed with PBS+ (phosphate-buffered saline with 0.5 mM CaCl2 and 1 mM MgCl2), fixed with 4% paraformaldehyde (PFA) for 1 min at 4°C and subsequently permeabilized with methanol for 10 min at 4°C. After blocking in PBS+/1%BSA, cells were incubated for 1 h at room temperature with primary antibodies, then for 1 h with fluorescently-labeled secondary antibodies. After three washes with PBS, cells were incubated with RNase A and then with Topro-3 to stain the nuclei. Confocal imaging was acquired with a Leica TCS SP2 Laser Scanning Spectral system 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 on-line ‘Scan Ware’ software. Images were processed using Image J and Photoshop software. Figure compilation was accomplished using Adobe Photoshop 5.5 and Adobe Illustrator 10.

Transcytosis assay

Stably transfected 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. Non-internalized antibody–antigen complexes were removed by acid washing (200 mM glycine, 150 mM NaCl, pH 2.5) before fixation, permeabilization and staining with fluorescently labeled secondary antibodies. Fluorescence was examined by confocal laser scanning microscopy.

Detergent resistant membrane (DRM) preparation and analysis

DRM preparation was performed according to previously published protocol (Aït-Slimane et al., 2003). Cells grown to confluence were rinsed with PBS and lysed for 30 min in TNE/TX-100 buffer (20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100 and protease inhibitors) on ice. Lysates were scraped from the dishes, and homogenized by passing through a 22-gauge needle. Extracts were brought to 40% sucrose and placed at the bottom of 5–35% sucrose gradients. Gradients were centrifuged at 38,000 rpm for 18 h at 4°C in a Beckman SW 41 rotor. Fractions of 1 ml were harvested from the top of the gradient. Proteins were immunoprecipitated with the polyclonal anti-c-Myc. After washing, samples were analyzed by SDS-PAGE and western blotting using the monoclonal anti-GFP.

Metabolic labeling, detergent extraction and immunoprecipitation

Cells were pre-incubated for 2×30 min at 37°C in methionine/cysteine-free medium. They were pulse-labeled with 300 µCi/ml [35S]methionine/cysteine at 37°C, and chased in DMEM containing excess cysteine and methionine (2 mM and 1 mM, respectively) for the indicated periods. At each time point, cells were washed with PBS+ on ice, and lysed in 1 ml of TNE/TX-100 buffer for 30 min on ice. Lysates were centrifuged at 13,000 rpm for 2 min at 4°C. Supernatants representing the soluble material were collected. Pellets were solubilized in 100 µl solubilization buffer (50 mM Tris-HCl, pH 8.8, 5 mM EDTA, 1% SDS). Both soluble and insoluble fractions were adjusted to 0.1% SDS and then immunoprecipitated at 4°C with the polyclonal anti-c-Myc. Immunoprecipitates were analysed by SDS-PAGE. Gels were dried and submitted to fluorography. The fluorograms were scanned and the bands were quantified using Scion Image software.

Velocity gradients

The procedure was essentially performed as described (Paladino et al., 2007). HepG2 cells were grown to confluence in 100-mm dishes, washed in PBS+ and lysed in lysis buffer (20 mM Tris, pH 7.4, 100 mM NaCl, 0.4% SDS and 0.2% Triton X-100) for 30 min on ice. Cells were scraped from dishes, sheared through a 26-g needle. Lysates were centrifuged at 9,000 rpm for 10 min to pellet nuclei. The cleared lysates were layered onto 5–30% sucrose gradients, and ultracentrifugation was performed at 45,000 rpm for 16 h using a Beckman SW55 rotor. Fractions of 500 µl were harvested from the top of the gradient. Proteins were immunoprecipitated using specific antibodies. High-molecular-weight (HMW) complexes were analyzed by SDS-PAGE and western blot.

We thank Chiara Zurzolo and Satyajit Mayor for the generous gift of cDNAs; Philippe Fontanges and Romain Morichon (INSERM IFR65) for imaging with the confocal microscope; and Chantal Housset for the critical reading of the manuscript. We also thank Odile Colard for help in the initial phase of this work.

Author contributions

T.A.-S. and M.M. conceived and designed the experiments; R.G., T.A.-S. and J.-L.D. performed the experiments; R.G. and T.A.-S. analyzed the data; T.A.-S. and M.M. wrote the paper.

Funding

This research was funded by Institut National de la Santé et de la Recherche Médicale (INSERM) and by Université Pierre et Marie Curie (UPMC).

Aït-Slimane
T. A.
,
Trugnan
G.
,
Van IJzendoorn
S. C. D.
,
Hoekstra
D.
(
2003
).
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
14
,
611
624
.
Aït-Slimane
T.
,
Galmes
R.
,
Trugnan
G.
,
Maurice
M.
(
2009
).
Basolateral internalization of GPI-anchored proteins occurs via a clathrin-independent flotillin-dependent pathway in polarized hepatic cells.
Mol. Biol. Cell
20
,
3792
3800
.
Arreaza
G.
,
Brown
D. A.
(
1995
).
Sorting and intracellular trafficking of a glycosylphosphatidylinositol-anchored protein and two hybrid transmembrane proteins with the same ectodomain in Madin-Darby canine kidney epithelial cells.
J. Biol. Chem.
270
,
23641
23647
.
Bartles
J. R.
,
Feracci
H. M.
,
Stieger
B.
,
Hubbard
A. L.
(
1987
).
Biogenesis of the rat hepatocyte plasma membrane in vivo: comparison of the pathways taken by apical and basolateral proteins using subcellular fractionation.
J. Cell Biol.
105
,
1241
1251
.
Bastaki
M.
,
Braiterman
L. T.
,
Johns
D. C.
,
Chen
Y. H.
,
Hubbard
A. L.
(
2002
).
Absence of direct delivery for single transmembrane apical proteins or their ‘Secretory’ forms in polarized hepatic cells.
Mol. Biol. Cell
13
,
225
237
.
Benting
J. H.
,
Rietveld
A. G.
,
Simons
K.
(
1999
).
N-Glycans mediate the apical sorting of a GPI-anchored, raft-associated protein in Madin-Darby canine kidney cells.
J. Cell Biol.
146
,
313
320
.
Brown
D. A.
,
Rose
J. K.
(
1992
).
Sorting of GPI-anchored proteins to glycolipid-enriched membrane subdomains during transport to the apical cell surface.
Cell
68
,
533
544
.
Brown
D. A.
,
Crise
B.
,
Rose
J. K.
(
1989
).
Mechanism of membrane anchoring affects polarized expression of two proteins in MDCK cells.
Science
245
,
1499
1501
.
Catino
M. A.
,
Paladino
S.
,
Tivodar
S.
,
Pocard
T.
,
Zurzolo
C.
(
2008
).
N- and O-glycans are not directly involved in the oligomerization and apical sorting of GPI proteins.
Traffic
9
,
2141
2150
.
Conner
S. D.
,
Schmid
S. L.
(
2003
).
Regulated portals of entry into the cell.
Nature
422
,
37
44
.
Cunningham
O.
,
Andolfo
A.
,
Santovito
M. L.
,
Iuzzolino
L.
,
Blasi
F.
,
Sidenius
N.
(
2003
).
Dimerization controls the lipid raft partitioning of uPAR/CD87 and regulates its biological functions.
EMBO J.
22
,
5994
6003
.
Fivaz
M.
,
Vilbois
F.
,
Thurnheer
S.
,
Pasquali
C.
,
Abrami
L.
,
Bickel
P. E.
,
Parton
R. G.
,
van der Goot
F. G.
(
2002
).
Differential sorting and fate of endocytosed GPI-anchored proteins.
EMBO J.
21
,
3989
4000
.
Hua
W.
,
Sheff
D.
,
Toomre
D.
,
Mellman
I.
(
2006
).
Vectorial insertion of apical and basolateral membrane proteins in polarized epithelial cells revealed by quantitative 3D live cell imaging.
J. Cell Biol.
172
,
1035
1044
.
Jain
R. K.
,
Joyce
P. B. M.
,
Molinete
M.
,
Halban
P. A.
,
Gorr
S. U.
(
2001
).
Oligomerization of green fluorescent protein in the secretory pathway of endocrine cells.
Biochem. J.
360
,
645
649
.
Kipp
H.
,
Arias
I. M.
(
2000
).
Newly synthesized canalicular ABC transporters are directly targeted from the Golgi to the hepatocyte apical domain in rat liver.
J. Biol. Chem.
275
,
15917
15925
.
Lisanti
M. P.
,
Caras
I. W.
,
Davitz
M. A.
,
Rodriguez-Boulan
E.
(
1989
).
A glycophospholipid membrane anchor acts as an apical targeting signal in polarized epithelial cells.
J. Cell Biol.
109
,
2145
2156
.
Lisanti
M. P.
,
Le Bivic
A.
,
Saltiel
A. R.
,
Rodriguez-Boulan
E.
(
1990
).
Preferred apical distribution of glycosyl-phosphatidylinositol (GPI) anchored proteins: a highly conserved feature of the polarized epithelial cell phenotype.
J. Membr. Biol.
113
,
155
167
.
Luton
F.
,
Hexham
M. J.
,
Zhang
M.
,
Mostov
K. E.
(
2009
).
Identification of a cytoplasmic signal for apical transcytosis.
Traffic
10
,
1128
1142
.
Macia
E.
,
Ehrlich
M.
,
Massol
R.
,
Boucrot
E.
,
Brunner
C.
,
Kirchhausen
T.
(
2006
).
Dynasore, a cell-permeable inhibitor of dynamin.
Dev. Cell
10
,
839
850
.
Matter
K.
,
Mellman
I.
(
1994
).
Mechanisms of cell polarity: sorting and transport in epithelial cells.
Curr. Opin. Cell Biol.
6
,
545
554
.
Maurice
M.
,
Schell
M. J.
,
Lardeux
B.
,
Hubbard
A. L.
(
1994
).
Biosynthesis and intracellular transport of a bile canalicular plasma membrane protein: studies in vivo and in the perfused rat liver.
Hepatology
19
,
648
655
.
McGwire
G. B.
,
Becker
R. P.
,
Skidgel
R. A.
(
1999
).
Carboxypeptidase M, a glycosylphosphatidylinositol-anchored protein, is localized on both the apical and basolateral domains of polarized Madin-Darby canine kidney cells.
J. Biol. Chem.
274
,
31632
31640
.
Mostov
K.
,
Su
T.
,
ter Beest
M.
(
2003
).
Polarized epithelial membrane traffic: conservation and plasticity.
Nat. Cell Biol.
5
,
287
293
.
Neumann-Giesen
C.
,
Falkenbach
B.
,
Beicht
P.
,
Claasen
S.
,
Lüers
G.
,
Stuermer
C. A. O.
,
Herzog
V.
,
Tikkanen
R.
(
2004
).
Membrane and raft association of reggie-1/flotillin-2: role of myristoylation, palmitoylation and oligomerization and induction of filopodia by overexpression.
Biochem. J.
378
,
509
518
.
Nyasae
L. K.
,
Hubbard
A. L.
,
Tuma
P. L.
(
2003
).
Transcytotic efflux from early endosomes is dependent on cholesterol and glycosphingolipids in polarized hepatic cells.
Mol. Biol. Cell
14
,
2689
2705
.
Paladino
S.
,
Sarnataro
D.
,
Pillich
R.
,
Tivodar
S.
,
Nitsch
L.
,
Zurzolo
C.
(
2004
).
Protein oligomerization modulates raft partitioning and apical sorting of GPI-anchored proteins.
J. Cell Biol.
167
,
699
709
.
Paladino
S.
,
Pocard
T.
,
Catino
M. A.
,
Zurzolo
C.
(
2006
).
GPI-anchored proteins are directly targeted to the apical surface in fully polarized MDCK cells.
J. Cell Biol.
172
,
1023
1034
.
Paladino
S.
,
Sarnataro
D.
,
Tivodar
S.
,
Zurzolo
C.
(
2007
).
Oligomerization is a specific requirement for apical sorting of glycosyl-phosphatidylinositol-anchored proteins but not for non-raft-associated apical proteins.
Traffic
8
,
251
258
.
Pang
S.
,
Urquhart
P.
,
Hooper
N. M.
(
2004
).
N-glycans, not the GPI anchor, mediate the apical targeting of a naturally glycosylated, GPI-anchored protein in polarised epithelial cells.
J. Cell Sci.
117
,
5079
5086
.
Ramnarayanan
S. P.
,
Tuma
P. L.
(
2011
).
MAL, but not MAL2, expression promotes the formation of cholesterol-dependent membrane domains that recruit apical proteins.
Biochem. J.
439
,
497
504
.
Rodriguez-Boulan
E.
,
Gonzalez
A.
(
1999
).
Glycans in post-Golgi apical targeting: sorting signals or structural props?
Trends Cell Biol.
9
,
291
294
.
Rodriguez-Boulan
E.
,
Kreitzer
G.
,
Müsch
A.
(
2005
).
Organization of vesicular trafficking in epithelia.
Nat. Rev. Mol. Cell Biol.
6
,
233
247
.
Sarnataro
D.
,
Paladino
S.
,
Campana
V.
,
Grassi
J.
,
Nitsch
L.
,
Zurzolo
C.
(
2002
).
PrPC is sorted to the basolateral membrane of epithelial cells independently of its association with rafts.
Traffic
3
,
810
821
.
Schell
M. J.
,
Maurice
M.
,
Stieger
B.
,
Hubbard
A. L.
(
1992
).
5′nucleotidase is sorted to the apical domain of hepatocytes via an indirect route.
J. Cell Biol.
119
,
1173
1182
.
Schuck
S.
,
Simons
K.
(
2004
).
Polarized sorting in epithelial cells: raft clustering and the biogenesis of the apical membrane.
J. Cell Sci.
117
,
5955
5964
.
Simmen
T.
,
Höning
S.
,
Icking
A.
,
Tikkanen
R.
,
Hunziker
W.
(
2002
).
AP-4 binds basolateral signals and participates in basolateral sorting in epithelial MDCK cells.
Nat. Cell Biol.
4
,
154
159
.
Simons
K.
,
Ikonen
E.
(
1997
).
Functional rafts in cell membranes.
Nature
387
,
569
572
.
Sugimoto
H.
,
Sugahara
M.
,
Fölsch
H.
,
Koide
Y.
,
Nakatsu
F.
,
Tanaka
N.
,
Nishimura
T.
,
Furukawa
M.
,
Mullins
C.
,
Nakamura
N.
 et al. (
2002
).
Differential recognition of tyrosine-based basolateral signals by AP-1B subunit mu1B in polarized epithelial cells.
Mol. Biol. Cell
13
,
2374
2382
.
Tuma
P. L.
,
Hubbard
A. L.
(
2003
).
Transcytosis: crossing cellular barriers.
Physiol. Rev.
83
,
871
932
.
Vergés
M.
,
Luton
F.
,
Gruber
C.
,
Tiemann
F.
,
Reinders
L. G.
,
Huang
L.
,
Burlingame
A. L.
,
Haft
C. R.
,
Mostov
K. E.
(
2004
).
The mammalian retromer regulates transcytosis of the polymeric immunoglobulin receptor.
Nat. Cell Biol.
6
,
763
769
.
Wang
L.
,
Boyer
J. L.
(
2004
).
The maintenance and generation of membrane polarity in hepatocytes.
Hepatology
39
,
892
899
.
Zurzolo
C.
,
Lisanti
M. P.
,
Caras
I. W.
,
Nitsch
L.
,
Rodriguez-Boulan
E.
(
1993
).
Glycosylphosphatidylinositol-anchored proteins are preferentially targeted to the basolateral surface in Fischer rat thyroid epithelial cells.
J. Cell Biol.
121
,
1031
1039
.