Enterophilin-1, a New Partner of Sorting Nexin 1, Decreases Cell Surface Epidermal Growth Factor Receptor*

We previously described enterophilin-1 (Ent-1), a new intestinal protein bearing an extended leucine zipper and a B30.2 domain. Ent-1 expression is associated with growth arrest and enterocyte differentiation. To investigate the importance of Ent-1 in the differentiation, we performed a yeast two-hybrid screening. We identified sorting nexin 1 (SNX1) as a novel partner of Ent-1 and confirmed the specificity of interaction by co-immuno-precipitation experiments in mammalian cells. SNX1 is associated with endosomal membranes and triggers the endosome-to-lysosome pathway of epidermal growth factor receptor (EGFR). We observe by immunofluorescence microscopy that Ent-1 and SNX1 are co-localized on vesicular and tubulovesicular structures, which are different from early endosome antigen 1-containing endosomes. By gel filtration chromatography, we show that Ent-1 and SNX1 co-eluted in macromolecular complexes containing part of EGFR. Furthermore, overexpressed Ent-1 decreases cell surface EGFR. Ent-1 and SNX1 co-overexpression strongly extends

The intestinal epithelium undergoes continuous and rapid renewal, with proliferation of the multipotent stem cells limited to the crypts of Lieberkü hn. At the top of the crypt, cells lose their proliferative ability and complete differentiation during a highly organized migration along the crypt-villus axis (1)(2)(3). Two events are important to intestinal epithelial cell differentiation: the transition from stem cells to committed proliferative cells and the mechanisms responsible for the irreversible loss of proliferative potential as the committed cells start to differentiate. However, little is known regarding how these two events take place.
The regulation of epithelial cell growth and functional differentiation is susceptible to various influences along the cryptvillus axis, including growth factor-derived signals. The overall importance of growth factors in intestinal epithelial cell renewal is underscored by the fact that the normal intestinal development is severely perturbed in EGFR 1 knock-out mice (4). Interestingly, EGF binding in situ along the crypt-villus axis is higher in crypt than in villus enterocytes (5). These data correlated with studies performed with colon adenocarcinoma Caco-2 cells indicating that the expression of cell surface EGFR is dramatically decreased in well differentiated Caco-2 cells (6). Even if the cell growth arrest is well known to be associated with the differentiation process, the link between these two phenomena remains a confused point in the study of epithelial cell differentiation.
We have recently described enterophilins as a new family of intestinal proteins, with a carboxyl-terminal B30.2 domain and an extended leucine zipper in their amino-terminal part. Three members were identified: enterophilin-1 (Ent-1), enterophilin-2 (Ent-2), and a short form of Enterophilin-2 (Ent-2S). Ent-1 contains up to 45 regular heptad repeats and corresponds to a 65-kDa protein. Interestingly, Ent-1 was mostly expressed in the mid-crypt-villus axis, when cells stop their proliferation to start the differentiation process. In human intestinal epithelial carcinoma Caco-2 cells, Ent-1 ortholog expression pattern was positively correlated to growth arrest and terminal differentiation program. In addition, transfection of HT-29 cells with Ent-1 full-length cDNA inhibited cell growth and promoted an increase in alkaline phosphatase activity, an intestinal differentiation marker. Taken together, these results suggested a close relationship between Ent-1 expression and the enterocyte differentiation program (7).
To investigate the role of Ent-1 in growth arrest associated with enterocytic differentiation, we performed a yeast twohybrid screen to identify proteins interacting with Ent-1. We report herein that Ent-1 interacts with sorting nexin 1 (SNX1). The sorting nexins are an emerging family of proteins (for a review see Ref. 8), characterized by the presence of a phox homology domain (9,10). They are involved in the intracellular trafficking of several membrane receptors (11)(12)(13). SNX1, the best studied member of this family, recognizes the lysosomal targeting sequence code in EGFR. Moreover, SNX1 overexpres-sion decreases the amount of EGFR on the cell surface as a result of enhancing trafficking in the endosome-to-lysosome pathway (11). Furthermore, recent data demonstrate that mice lacking both SNX1 and SNX2 display alterations in proper cellular trafficking (14). We confirm the association of Ent-1 with SNX1 by biochemical and immunofluorescence experiments in mammalian cells. Furthermore, Ent-1 causes a significant diminution of cell surface EGFR. In this context, the identification of SNX1 as an Ent-1 partner and their cooperative role in cell surface EGFR decrease provide more evidence about the involvement of Ent-1 in the inhibition of cell proliferation associated with the enterocytic differentiation process.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid Screening-Full-length Ent-1 cDNA, the "leucine zipper" region, and the B30.2 domain were amplified by PCR using oligonucleotides containing NcoI and SalI restriction sites, and PCR products were inserted in pAS2-1 (Clontech Laboratories) at the corresponding sites. First, full-length Ent-1 containing vector pAS2-1 was used as a bait to screen an epithelial HeLa cell cDNA library constructed in pGAD GH vector at EcoRI/XhoI restriction sites (Clontech Laboratories). The yeast host strain CG-1945 (Clontech Laboratories) was transformed with the bait and library plasmids. Positive transformants were selected on medium lacking leucine, tryptophan, and histidine, and a colony lift filter assay was performed to test ␤-galactosidase activity. Harboring positive library clones were retransformed into strain SFY526 (Clontech Laboratories) with full-length Ent-1 bait and tested for expression of LacZ reporter gene to confirm the specificity of the interaction. cDNA inserts of true positive clones were sequenced and submitted to the data base. To pinpoint the interacting domain of Ent-1, the leucine zipper region or the B30.2 domain constructions were used in a one to one interaction test with positive library plasmids in the yeast strain SFY526 and tested for positive interaction as described above.
The cells were grown in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum (Invitrogen) and 100 g/ml penicillin/streptomycin (Invitrogen) in a humidified atmosphere containing 5% CO 2 . Nonessential amino acids (0.1 mM; Invitrogen) were added to the culture medium of Caco-2 cells. All cells were transfected with the different plasmids using FuGENE TM 6 (Roche Applied Science), except Caco-2 cells, which were transfected with a cationic lipid (LipofectAMINE; Invitrogen) according to the manufacturer's protocols.
Immunoprecipitation Experiments-Caco-2 cells were co-transfected with both pcDNA3/Myc-His-Ent-1 and pcDNA3/FLAG-SNX1. The experiments were performed 48 h after transfection, and all of the procedures were done at 4°C. The cells were washed three times with phosphate-buffered saline (PBS) and lysed for 15 min with lysis buffer (20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 10% glycerol) containing protease and phosphatase inhibitors (1 mM phenylmethanesulfonyl fluoride, 10 g/ml leupeptin, 1 g/ml aprotinin, 20 mM NaF, 2 mM Na 3 VO 4 ). Insoluble debris were removed by centrifugation at 14,000 ϫ g for 15 min, and the supernatant was subjected to a preclearing with 30 l of protein G-Sepharose for 30 min. Ent-1 was immunoprecipitated for 2 h with 5 g of monoclonal anti-Myc antibody, followed by 50 l of protein G-Sepharose for 1 h. After washing three times in buffer (20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.1% Nonidet P-40, 10% glycerol) containing protease and phosphatase inhibitors, the bound proteins were eluted by boiling the beads in Laemmli buffer (15). The precipitated proteins were then submitted to SDS-polyacrylamide gel electrophoresis and detected by immunoblotting.
Confluence-induced Differentiation of Caco-2 Cells-Caco-2 cells were seeded at 18,000 cells/cm 2 and grown in 60-mm dishes in the complete medium. The medium was changed every 2 days. The cells were washed twice in PBS and scraped at 4°C in PBS containing protease and phosphatase inhibitors (1 mM phenylmethanesulfonyl fluoride, 10 g/ml leupeptin, 1 g/ml aprotinin, 20 mM NaF, 2 mM Na 3 VO 4 ) at various times up to 23 days after plating. The samples were sonicated, and the protein concentration was determined according to the method of Bradford (16). Equal amounts of proteins were subjected to SDS-polyacrylamide gel electrophoresis. Then Ent-1 and EGFR expression were analyzed by immunoblotting.
Immunoblotting-The protein samples were submitted to SDS-polyacrylamide gel electrophoresis, transferred onto nitrocellulose membrane, and subjected to immunoblotting according to standard protocols (17). Rabbit polyclonal anti-Ent-1 peptide antibody was obtained from Eurogentec as previously described (7). Monoclonal antibody against c-Myc epitope (9E10) and polyclonal antibody against epidermal growth factor receptor (1005) were from Santa Cruz Biotechnology. Monoclonal antibody against flag (M2) was purchased from Sigma, and monoclonal antibody against GFP was from Roche Applied Bioscience. Revelation was done with horseradish peroxidase-conjugated anti-mouse IgG or anti-rabbit IgG antibody (Promega) and ECL (Amersham Biosciences) detection.
Immunofluorescence Microscopy-COS-7 cells were grown on sterile glass coverslips and transfected with pEGFP-C2-Ent-1 or pEGFP-C1-SNX1 or co-transfected with both pcDNA3/Myc-His-Ent-1 and pEGFP-C1-SNX1. For EGF stimulation, the cells were serum-starved for 24 h. The cells were incubated with 100 ng/ml EGF for 1 h at 4°C, rinsed twice in PBS, and finally incubated with warmed medium at 37°C for 10, 15, or 30 min. The cells were washed three times with ice-cold PBS, fixed for 15 min with 3% paraformaldehyde, permeabilized for 2 min with 0.2% Triton X-100, and saturated for 30 min with 0.2% gelatin. The cells were then incubated 60 min with the primary antibody (anti-EEA1 from Transduction Laboratories or anti-c-Myc (9E10) from Santa Cruz Biotechnology), and an immunostaining was performed with TRITC-conjugated anti-mouse IgG secondary antibody (Southern Biotechnology Associates). The coverslips were examined with a Zeiss Axioskop microscope or with a confocal microscope (Zeiss, LSM 510, Axiovert 100).
Analysis of Cell Surface EGFR Density by Flow Cytometry-COS-7 cells were transfected with pEGFP-C2-Ent-1, pEGFP-C1-SNX1, or empty vector and grown for 72 h. Then cells were collected by treatment with 2 mM EDTA in PBS and fixed in 1% formaldehyde for 30 min, and nonspecific sites were saturated with 1% bovine serum albumin for 30 min. Cell surface EGFR was detected by incubation with the primary antibody (clone LA1; Euromedex) for 60 min, followed by incubation with a R-phycoerythrin (RPE)-conjugated goat anti-mouse IgG secondary antibody (DAKO) for 60 min. The cells were then analyzed by flow cytometry (Beckman Coulter XL 4C). Expression of the GFP-fused proteins was monitored by fluorescence measurement, and the surface binding of the anti-EGFR antibody was measured specifically on the transfected cell population. To evaluate the effect of both Ent-1 and SNX1 on plasma membrane EGFR, we co-transfected COS-7 cells with pDsRed2-N1-Ent-1 and pEGFP-C1-SNX1 or both empty vectors. Cell surface EGFR was similarly immunolabeled, except for the secondary antibody: RPE-Cy5-conjugated rabbit anti-mouse IgG antibody (DAKO). The percentage of cell surface EGFR was measured by RPE (red) or RPE-Cy5 (deep red) fluorescence mean. The results were analyzed by paired t test and were considered significantly different (*, p Ͻ 0.005; **, p Ͻ 0.001) from controls (GFP-transfected cells for single transfection or GFP-and DsRed-co-transfected cells for double transfection).

Enterophilin-1 Interacts with Sorting Nexins 1 and 2-To
elucidate the role of Ent-1 in enterocyte differentiation, we performed a two-hybrid screening of epithelial HeLa cell library using full-length Ent-1 as a bait. Twenty-four positive clones were rescued and confirmed by retransformation into a second yeast host strain SFY526. The positive clones from the second screening were sequenced; nine clones encoded sorting nexins SNXs, a family of proteins playing an important role in protein trafficking among various organelles. Sequence analysis indicated that three clones encoded SNX1 and six encoded SNX2 as depicted by protein alignments presented in Fig. 1 (A  and B, respectively). These data also showed that the clones contained only a part of the coiled coil regions and the EGFR binding domain of SNX1, indicating that these minimal regions are sufficient for binding to Ent-1.
To pinpoint the binding domains of Ent-1 with its identified partners, we used the two-hybrid one to one interaction assay with the leucine zipper part or the B30.2 domain as a bait. The results showed that the B30.2 domain was unable to bind SNX1 and SNX2. Unfortunately, the leucine zipper part of Ent-1 was toxic for the two yeast strains. Accordingly, we suggested that the leucine zipper domain and/or full-length Ent-1 was necessary to mediate the binding with the sorting nexin proteins.
We previously demonstrated that Ent-1 overexpression inhibited cell proliferation. Thus, we decided to focus our investigations on SNX1 because it is the best characterized members of SNX family, and it has clearly been involved in the decrease of cell surface EGFR by enhancing trafficking in the endosometo-lysosome pathway. To further confirm the association of Ent-1 with SNX1, we performed co-immunoprecipitation experiments in intestinal mammalian cells. Myc-tagged Ent-1 and FLAG-tagged SNX1 were co-expressed in Caco-2 cells, and Ent-1 was subjected to immunoprecipitation with an anti-Myc antibody. Immunoprecipitated complexes were analyzed by immunoblotting, and data revealed the presence of both Ent-1 and SNX1, confirming the interaction between these two proteins in intestinal cells (Fig. 2).
Enterophilin-1 Co-localizes with SNX1 on Vesicular and Tubulovesicular Structures-Identification of SNXs as Ent-1 partners led us to focus on the cytoplasmic localization of Ent-1 by fluorescence microscopy. COS-7 cells were co-transfected with COOH-terminally tagged Ent-1 (pcDNA3/Myc-His-Ent-1) and NH 2 -terminally tagged SNX1 (pEGFP-C1-SNX1). Both Ent-1 (Fig. 3, top panel) and SNX1 (Fig. 3, middle panel) displayed a vesicular and tubulovesicular signal. We observed that part of both staining patterns perfectly overlapped (Fig. 3, bottom  panel, inset), indicating that within the cell, Ent-1 and SNX1 co-localized in similar structures. However, we noticed that Ent-1 was also located in cellular areas lacking SNX1 staining. Similarly, SNX1 was found in cytoplasmic pools that did not contain Ent-1. Thus, Ent-1 and SNX1 showed substantial but not complete co-localization.
Furthermore, we confirmed that NH 2 -terminally tagged Ent-1 displayed a cytoplasmic distribution similar to COOHterminally tagged Ent-1 in different epithelial cell lines. Indeed, we also observed vesicular and tubulovesicular structures with pEGFP-C2-Ent-1 in COS-7 cells, as well as in Madin-Darby canine kidney, HeLa, and Caco-2 cells (data not shown).
Ent-1-containing Structures Are Different from EEA1-positive Early Endosomes-Because SNX1 has been reported to be substantially associated to early endosomal membranes (19,20), we investigated whether Ent-1-containing vesicles corresponded to early endosomes. COS-7 cells were transfected with pEGFP-C2-Ent-1 and the early endosomal specific marker EEA1 was labeled with a specific antibody after various times of EGF stimulation (100 ng/ml) as indicated under "Experimental Procedures". Confocal microscopy analysis revealed that Ent-1 was still present on vesicular and tubulovesicular structures in cells stimulated for 10, 15, or 30 min with EGF (Fig. 4). However, no merge was observed with EEA1 staining at any time of stimulation, suggesting that Ent-1-containing vesicles were different from the previously defined EEA1-positive early endosomes (Fig. 4). By contrast, SNX1 partially overlapped with EEA1 signal under the same conditions of EGF stimula- FIG. 1. Schematic representation of sorting nexins 1 and 2. A, the structure of human SNX1 includes the phox homology domain (PX, residues 141-268), predicted coiled-coil domains H1 (residues 313-388) and H2 (residues 444 -472), and the EGFR binding domain (residues 465-522) as previously described (11). B, the structure of human SNX2 includes the phox homology domain (PX, residues 144 -270) and predicted coiled-coil domains H1 (residues 315-336), H2 (residues 362-390) and H3 (residues 441-479). The Ent-1-interacting clones isolated from the yeast two-hybrid screen were sequenced, and their predicted amino acid sequences were aligned below the domain structure of SNX1 (A) and SNX2 (B) analyzed with the SMART program.

FIG. 2. Ent-1 co-immunoprecipitates with SNX1 in Caco-2 cells.
Caco-2 cells were co-transfected with both pcDNA3/Myc-His-Ent-1 and pcDNA3/FLAG-SNX1 and grown for 48 h, and lysates were subjected to an Ent-1 immunoprecipitation (IP) with a monoclonal anti-Myc antibody (ϩ). To verify the specificity of the interaction, a control was performed without antibody (Ϫ). Immunoprecipitates were then submitted to SDS-polyacrylamide gel electrophoresis, and SNX1 and Ent-1 proteins were detected by immunoblotting with an anti-FLAG antibody and an anti-Myc antibody, respectively. tion. Nevertheless, as previously documented (19,21), a large proportion of SNX1 vesicles was clearly EEA1-negative (Fig. 5) and could correspond to Ent-1 merging area. Our immunofluorescence data demonstrated that Ent-1 and SNX1 co-localized in tubular and vesicular structures that were different from EEA1-positive early endosomes.
Enterophilin-1 and SNX1 Coexist in Macromolecular Complexes Containing EGFR in the Cell-Sorting nexins are supposed to act as a multimeric protein complex named the retromer complex by analogy with their yeast orthologs (22). Thus, we checked for the presence of Ent-1 in such macromolecular complexes by gel filtration chromatography. COS-7 cells were cotransfected with pEGFP-C2-Ent-1 and pcDNA3/FLAG-SNX1. After stimulation with 100 ng/ml EGF for 10 min to stimulate EGFR endocytosis, the cytosolic fractions were prepared and loaded onto the column. The different fractions were analyzed for the presence of GFP-Ent-1, FLAG-SNX1, and endogenous EGFR by Western blotting. An important signal for Ent-1 was obtained from fractions 26 -30, corresponding to large complexes of 670 -310 kDa, with a major elution in ϳ435 kDa (Fig.  6). Conversely, no signal was detected in the fractions corresponding to the size of monomeric Ent-1 (92 kDa for the GFP-tagged Ent-1), suggesting that Ent-1 does not exist as monomers in the cell. Interestingly, SNX1 displayed an elution peak in fractions 26 -30 of ϳ310 -670 kDa that perfectly matched with the Ent-1-enriched complexes. However, as already reported (18,22), SNX1 also presented a wide elution profile because of its property to self-assembly or to associate with other proteins. Thus, SNX1 was recovered in a second peak of higher size macromolecular structures (fractions 20 -22, ϳ1700 kDa). Furthermore, SNX1 did not appear in the monomeric form (ϳ60 kDa for the FLAG-tagged SNX1). Endogenous EGFR presented an elution profile in fractions 22-28, overlapping the Ent-1 and SNX1 peaks in fraction 26 corresponding to ϳ670-kDa heteromeric protein complexes. As with Ent-1 and SNX1, EGFR was not found as monomeric form. We can notice that a part of the three proteins was observed in the void volume (fractions 1-15) that could correspond to very large molecular mass complexes excluded from the column. Alternatively, this could be the result of nonspecific aggregation. Nevertheless, gel filtration data demonstrated a convincing overlap between Ent-1 and SNX1 elution profiles and strongly suggested the existence of Ent-1/SNX1/EGFR multimeric complexes involved in EGF-induced EGFR endocytosis.
Enterophilin-1 Enhances the Decrease of Cell Surface EGF Receptors-Gel filtration data led us to check for the effects of Ent-1 on EGFR cell surface pool. COS-7 cells were transfected with pEGFP-C2-Ent-1, pEGFP-C1-SNX1, or empty vector for 72 h, and cell surface EGFR was immunolabeled. By flow cytometry, cell surface EGFR expression was specifically monitored in the GFP-or GFP-Ent-1-transfected cells as indicated under "Experimental Procedures." Our results showed that Ent-1 induced a significant decrease of 35% of plasma mem- Enterophilin-1, SNX1, and EGFR Degradation brane EGFR, compared with the GFP-transfected cell population used as control (Fig. 7A). A similar decrease was obtained with cells overexpressing SNX1 in our experimental conditions (Fig. 7A), in agreement with previously published data (11). These results indicated that overexpressed Ent-1 was able to promote the down-regulation of cell surface EGFR.
To evaluate the cooperation between Ent-1 and SNX1 in the regulation of EGFR endocytosis, we co-transfected COS-7 cells with pDsRed2-N1-Ent-1 and pEGFP-C1-SNX1. Cell surface EGFR was immunolabeled and identically analyzed on doubletransfected cell population. We demonstrated that Ent-1 and SNX1 displayed a synergetic effect on cell surface EGFR re-moval because a 65% decrease of plasma membrane EGFR was quantified in double-transfected cells (Fig. 7B). Taken together, these data suggested a cooperative effect of Ent-1 and SNX1 on cell surface EGFR removal during endocytosis.
Enterophilin-1 Expression Correlates with the Decrease of EGFR during the Differentiation of Caco-2 Cells-To investigate the physiological relevance of the effects of Ent-1 overexpression on cell surface EGFR, we analyzed the expression patterns of both proteins in relation to enterocyte differentiation. Human colon carcinoma Caco-2 cell line is a well characterized model for the study of intestinal differentiation. We thus performed the confluence-induced differentiation of FIG. 6. Ent-1 co-elutes with part of SNX1 and EGFR in macromolecular complexes. COS-7 cells were co-transfected with pEGFP-C2-Ent-1 and pcDNA3/FLAG-SNX1, grown for 24 h, serum-starved during additional 24 h, and stimulated with 100 ng/ml EGF for 10 min. Cytosolic extract was obtained as described under "Experimental Procedures" and was loaded onto a Superose 6 HR 10/30 column. The fractions were collected, and an identical volume of fraction was subjected to immunoblotting analysis. Ent-1 was detected with monoclonal anti-GFP antibody, SNX1 with monoclonal anti-FLAG antibody, and EGFR with polyclonal anti-EGFR antibody. Vo represents the void volume (fractions 1-15).

Enterophilin-1, SNX1, and EGFR Degradation
Caco-2 cells. As previously shown (7), the cells reached confluence after 7 days, and the human ortholog of Ent-1, identified as a 65-kDa protein, displayed an increased expression beginning at day 9, sustained up to 23 days in culture. Interestingly, Western blot analysis showed a marked decrease of EGFR expression beginning at day 14, when the human ortholog of Ent-1 was highly expressed (Fig. 8). At day 2, we observed a weak expression of EGFR that could be expected knowing that Caco-2 cells start to proliferate after a lag time of 2 days after plating (7,23). The functional differentiation was followed by appearance of alkaline phosphatase activity, typically used as intestinal differentiation marker (data not shown). These re-sults demonstrated a correlation between the Ent-1 expression pattern and the decrease of EGFR expression during intestinal epithelial differentiation.

DISCUSSION
In this study, we sought partners of Ent-1 by a two-hybrid screening of an epithelial HeLa cell cDNA library. We demonstrated that Ent-1 interacted with sorting nexins 1 and 2. We first focused our investigations on the interaction of Ent-1 with SNX1, because SNX1 is the best studied mammalian member of the sorting nexin family. The association between Ent-1 and SNX1 was confirmed by immunoprecipitation experiments in epithelial mammalian cells. We also co-localized Ent-1 and SNX1 in vesicular and tubulovesicular structures. These results were consistent with recent studies that described sorting nexins in such vesicular and tubulovesicular compartments (19,24). As suggested by Kurten et al. (24), these tubulovesicular structures are not always detectable and may therefore represent transient entities. Ent-1-containing structures were clearly different from early endosomal marker EEA1 staining after EGF-induced EGFR endocytosis, whereas a certain proportion of SNX1 vesicles overlapped with EEA1-positive early endosomes. Such partial co-localization is consistent with recent data showing that SNX1-containing vesicles were part of the early endocytic compartment but were partially distinct from previously defined EEA1-containing endosomes and recycling transferrin receptor-containing endosomes (19,20,24). Further studies will attempt to more precisely define such EEA1-negative vesicles containing both Ent-1 and SNX1. SNX1 and SNX2 are the mammalian orthologs of the yeast vacuolar protein sorting Vps5p (25). Vps5p is a subunit of a large multimeric complex, termed the retromer complex, involved in retrograde transport of proteins from endosomes to the trans-Golgi network (26). In mammalian cells, homodimerization and heterodimerization of SNXs have been reported (22,24), and it has been proposed that complex formation between SNXs may be necessary for organizing functional units for receptor sorting and degradation. We showed a perfect co-elution of Ent-1 and SNX1 in 310 -670-kDa macromolecular complexes. Part of EGFR was also detected in the same complexes. We also demonstrated that neither Ent-1 and SNX1 nor EGFR eluted in their respective monomeric form. These data suggested that the three proteins interacted with each other or with other proteins in ϳ435-670-kDa macromolecular complexes. This was consistent with the work of Chin et al. (18), which demonstrated the elution of SNX1 in similar size complexes and the lack of monomeric-SNX1 form. Additionally, SNX1, but not Ent-1, was detected in higher molecular mass complexes (above 1700 kDa). Because a certain proportion of Caco-2 cells were grown in 60-mm dishes and harvested at various times up to 23 days after plating. Equal amounts of proteins were subjected to SDS-polyacrylamide gel electrophoresis. Ent-1 was detected with rabbit polyclonal anti-Ent-1 peptide antibody, and EGFR was detected with polyclonal anti-EGFR antibody.
Enterophilin-1, SNX1, and EGFR Degradation SNX1 and Ent-1 co-localized by immunofluorescence, these results emphasized that both proteins could interact with each other in the same multimeric complexes. Ent-1 could be a component of macromolecular complexes involved in endocytosis of EGFR. However, EGFR endocytosis is a very dynamic process, involving continuously remodeled complexes and structures. Thus, the presence of EGFR in Ent-1/SNX1-enriched complexes could be transient, strictly depending on well defined macromolecular complex formation along the EGFR endosome-to-lysosome pathway.
SNXs have been shown to modulate endocytosis of a variety of receptors (for a review, see Ref. 8), including EGFR (11), protease-activated receptor-1 (13), or low density lipoprotein receptors (27). Overexpressed SNX1 was clearly involved in the decrease of EGFR on the cell surface as a result of enhancing the rate of constitutive and ligand-induced degradation (11). We then investigated the role of Ent-1 on EGFR degradation. Our results showed that overexpression of Ent-1 significantly decreased the cell surface EGFR. Interestingly, EGFR degradation was strongly extended when Ent-1 and SNX1 were co-expressed. These results highlighted the synergetic effect of both proteins and were in favor of a role of Ent-1 in cell surface EGFR removal by endocytosis through its interaction with SNX1. Ent-1 could thus be considered as a new partner of SNX1, such as ACK2 (activated Cdc42-associated kinase 2), which promotes EGFR degradation through its interaction with SNX9 (28) or Hrs (hepatocyte growth factor-regulated tyrosine kinase substrate) that inhibits EGFR endocytosis through its binding to SNX1 (18). It will be interesting now to define whether Ent-1 could regulate other types of cell surface receptors as described for SNXs.
EGFR is an important mediator of intestinal epithelial cell proliferation. In fact, undifferentiated crypt cells display a higher EGF binding activity than differentiated villus enterocytes (5), which is correlated with the decrease of EGFR surface expression along the crypt-villus axis (6). Our data indicate that Ent-1 expression during spontaneous differentiation of Caco-2 cells correlates with the decrease of EGFR expression. Among several intestinal epithelial cell lines, Caco-2 cells have proven to be the most useful in vitro models. Indeed, Caco-2 cells are unique in their ability to initiate spontaneous differentiation on reaching confluence under normal culture conditions and undergo a maturation process closely resembling the functional differentiation found in normal intestine (23). Furthermore, we recently published that the onset of Ent-1 expression corresponded to cell growth arrest, preceding functional differentiation. Additionally, we reported a decrease of proliferation in Ent-1-transfected HT-29 cells (7). In this context, Ent-1-mediated cell surface EGFR removal adds to our understanding of the regulation of growth arrest in intestinal epithelium. This is directly related with the reported decrease of EGFR at the mid-villus axis (5).
To summarize, we demonstrated that Ent-1 was a new partner of SNX1. According to our two-hybrid results, we hypothesized an interaction mediated by the coiled-coil structures present in both the COOH-terminal region of SNXs and the NH 2 -terminal leucine zipper part of Ent-1. Works in progress in our laboratory presently aim at defining the exact SNX1binding regions on Ent-1 by generating Ent-1 mutants defective in binding SNX1. Additionally, Ent-1 was localized with SNX1 on vesicular and tubulovesicular structures, which were different from EEA1-positive early endosomes. Interestingly, Ent-1 co-eluted with SNX1 in macromolecular complexes containing part of EGFR and induced cell surface EGFR removal. Moreover, Ent-1 and SNX1 displayed a cooperative effect on EGFR degradation, strongly increasing plasma membrane EGFR removal. Further studies will bring new insights to precise exact molecular mechanisms by which Ent-1 and SNX1 regulate receptor vesicular trafficking. These data provide new evidence about Ent-1 involvement in down-regulation of mitogenic signal leading to cell growth arrest and enterocyte differentiation.