The Rab5 Effector EEA1 Interacts Directly with Syntaxin-6*

The fusion of transport vesicles with their cognate target membranes, an essential event in intracellular membrane trafficking, is regulated by SNARE proteins and Rab GTPases. Rab GTPases are thought to act prior to SNAREs in vesicle docking, but the exact biochemical relationship between the two classes of molecules is not known. We recently identified the early endosomal autoantigen EEA1 as an effector of Rab5 in endocytic membrane fusion. Here we demonstrate that EEA1 interacts directly and specifically with syntaxin-6, a SNARE implicated in trans-Golgi network to early endosome trafficking. The binding site for syntaxin-6 overlaps with that of Rab5-GTP at the C terminus of EEA1. Syntaxin-6 and EEA1 were found to colocalize extensively on early endosomes, although syntaxin-6 is present in the trans-Golgi network as well. Our results indicate that SNAREs can interact directly with Rab effectors, and suggest that EEA1 may participate intrans-Golgi network to endosome as well as in endocytic membrane traffic.

The delivery of transport vesicles to their correct destination membrane is ensured by a complex molecular machinery. The SNARE 1 and Rab protein families have been assigned a special role in vesicle targeting (1). SNARE proteins on vesicles form tight complexes with their complementary SNARE proteins on target membranes. This docks the vesicle to the target membrane, and membrane fusion ensues (2,3). Although a large number of SNARE molecules have been detected and localized to distinct intracellular membranes, SNARE pairing does not confer sufficient specificity to vesicle delivery because SNAREs interact rather promiscuously with each other (4). Small GTPases of the Rab family appear to add an additional layer of specificity. Like SNAREs, distinct Rab GTPases are localized to distinct membranes and control distinct membrane trafficking routes (5,6). One of their main functions appears to be the recruitment of proteins that function as tethers prior to SNARE complex formation (1). However, the molecular mechanism that couples Rab-mediated tethering to SNARE complex formation is not known.
Homotypic fusion between early endosomes can be readily reconstituted in vitro and provides a convenient system to examine the role of Rab and SNARE proteins (7)(8)(9). Such fusion requires the presence of Rab5 on both endosome membranes (10), as well as the Rab5 effector, EEA1 (11). The findings that EEA1 contains two spatially separate Rab5 binding sites, forms rod-shaped coiled-coil dimers, and is required prior to endosomal SNARE function have suggested that it may act as a tethering factor (11)(12)(13). Here we have investigated if EEA1 is able to interact with SNAREs on early endosomes.
Antibodies-A human autoimmune serum against EEA1 (16) and an affinity purified rabbit anti-EEA1 antibody (11) were used. Mouse monoclonal anti-Myc epitope antibody was from the 9E10 hybridoma (17). A mouse monoclonal anti-syntaxin-6 antibody was purchased from Transduction Laboratories. Horseradish peroxidase-conjugated goat antibodies against human, mouse, and rabbit IgG, fluorescein isothiocyanate-labeled goat antibodies against human IgG, and lissaminerhodamine-labeled goat antibodies against mouse IgG were purchased from Jackson Immunoresearch.
Yeast Two-hybrid Methods-The yeast reporter strain L40 (15) was co-transformed with the indicated pLexA and pGAD plasmids, and ␤-galactosidase activities of the transformants were determined as described previously (18).
Cells and Transfections-For transient overexpression studies, BHK-21 cells were infected for 1 h with T7 RNA polymerase recombinant vaccinia virus (vT7) and then transfected at 37°C with pGEM-1 plasmids containing the cDNA of interest, using DOTAP (Boehringer, Mannheim), as described previously (19). The cells were analyzed 6 h post-transfection.
Confocal Immunofluorescence Microscopy-Cells on coverslips were fixed with 3% paraformaldehyde, permeabilized with 0.05% Saponin (Sigma) and stained with primary antibodies followed by fluorescein isothiocyanate or lissamine-rhodamine-conjugated secondary antibodies, as described (14). The coverslips were examined with a Leica TCS NT confocal microscope equipped with a Kr/Ar laser.
Binding of Cytosolic and Membrane-associated EEA1 to GST-Syntaxin-6 -HeLa cells grown in 15-cm dishes were washed in ice-cold PBS, scraped, and homogenized in 400 l of homogenization buffer (HB) (20 mM Hepes, pH 7.2, 100 mM KCl, 2 mM MgCl 2 ), 1 mM DTT by passage through a 22 gauge needle six times. A post-nuclear supernatant was obtained by centrifugation for 5 min at 6000 rpm. The post-nuclear supernatant was centrifuged at 60,000 rpm for 30 min at 4°C in a Beckman TLA 100.2 rotor to obtain a cytosolic and a membrane fraction. The membrane fraction was solubilized in HB-1 mM DTT, 1% Triton X-100 (TX-100) containing a mixture of protease inhibitors without EDTA (Roche Molecular Biochemicals) for 45 min on ice before centrifugation at 60,000 rpm for 30 min at 4°C in a Beckman TLA 100.2 rotor. The cytosol and the soluble membrane fraction were then incubated with 5 g of recombinant GST, GST-syntaxin-6, GST-syntaxin-7, or GSTsyntaxin-16 proteins prebound to glutathione-Sepharose (Amersham Pharmacia Biotech) for 2 h at 4°C. The beads were subsequently washed four times with ice-cold HB, 0.5% TX-100. EEA1 associated with the beads was detected by SDS-polyacrylamide gel electrophoresis (PAGE), followed by immunoblotting, using a rabbit anti-EEA1 serum and the SuperSignal chemiluminescence kit from Pierce.
In Vitro Binding Studies with Recombinant Proteins-Aliquots (25 l) of glutathione-Sepharose beads (Amersham Pharmacia Biotech) were washed three times with HB before incubation with 0.1 nmol of GST or GST-syntaxin-6 for 1 h at room temperature. The beads were then washed three times with HB, 0.5% TX-100 before incubation with 1 g of recombinant EEA1 proteins (His 6 -EEA1, MBP-EEA1 1-209 , MBP-EEA1 1257-1411 , and MBP-EEA1 1257-1411 C1405S ) in HB, 0.5% TX-100 containing 5 mg/ml bovine serum albumin for 1 h at 4°C. Finally, the beads were washed four times with HB, 0.5% TX-100. Recombinant EEA1 proteins associated with the beads were detected by SDS-PAGE, followed by immunoblotting with a human anti-EEA1 serum and the SuperSignal chemiluminescence kit from Pierce. In some cases, MBP-EEA1 1257-1411 was incubated with Rab5-GDP or Rab5-GTP␥S prior to addition to the beads. His 6 -Rab5 (2 mM) (20) was preincubated with 10 mM GDP or GTP␥S in 20 mM Hepes, pH 7.2, 100 mM K-acetate, 0.5 mM MgCl 2 , 2 mM EDTA and 1 mM dithiothreitol for 30 min at 25°C. The MgCl 2 was then adjusted to 15 mM, MBP-EEA1 1257-1411 (50 nM) was added, and the incubation was continued for 30 min at 25°C. The mixture was then added to glutathione-Sepharose beads containing 0.1 nmol of GST or GST-syntaxin-6 and left at 4°C for 60 min. Finally, the beads were washed three times with the same buffer containing 15 mM MgCl 2 and 0.05% TX-100, and protein associated with the beads was detected by SDS-PAGE followed by immunoblotting with anti-MBP antibodies (11).
Co-immunoprecipitation of EEA1 and Syntaxin-6 -BHK-21 cells grown in 10-cm dishes were transfected as described above. 6 h post transfection, the cells were washes three time with ice-cold PBS and lysed in HB, 1% TX-100 containing a mixture of protease inhibitors without EDTA (Roche Molecular Biochemicals) for 20 min on ice. The lysate was centrifuged at 10,000 rpm for 10 min, and the supernatant was incubated with a human anti-EEA1 serum, with normal human serum, or with 20 l of anti-Myc-agarose beads (Santa Cruz Biotechnology) at 4°C for 15 h. Twenty l of protein G-agarose beads (Santa Cruz Biotechnology) were added to the lysate in the former cases, and incubation was continued for 1 h at 4°C. The beads were then washed three times with HB, 1% TX-100 and once with PBS. Precipitated proteins were detected by SDS-PAGE, followed by immunoblotting with anti-EEA1 serum or anti-Myc antibody.

EEA1
Interacts Specifically with Syntaxin-6 in the Yeast Twohybrid System-Because Rab GTPases appear to act upstream of SNARE proteins in membrane docking/fusion, we investigated the possibility that EEA1 may bind to a SNARE molecule as well as to Rab5-GTP. For this purpose, we cloned EEA1 into a yeast two-hybrid "bait" vector and the cytoplasmic domains of various endosomal/trans-Golgi network (TGN) syntaxins (SNAREs) into a "prey" vector. The bait and prey plasmids were cotransformed into a two-hybrid reporter yeast strain, which was subsequently assayed for activation of the reporter gene, lacZ. As shown in Table I, neither syntaxin-7, which is thought to regulate trafficking between endosomes and lysosomes (21, 22), syntaxin-11, which is thought to regulate trafficking between late endosomes and the TGN (23), nor syntaxin-16, which is found in the TGN FIG. 1. Biochemical interaction between EEA1 and syntaxin-6. a, GST, GST-syntaxin-6 (GST-Stx6), GST-syntaxin-7 (GST-Stx7), or GST-syntaxin-16 (GST-Stx16) immobilized on glutathione-Sepharose was incubated with cytosol or a membrane extract from HeLa cells to study their binding to EEA1. b, GST or GST-syntaxin-6 (GST-Stx6) immobilized on glutathione-Sepharose was incubated with His 6 -EEA1, MBP-EEA1 1257-1411 (CT), MBP-EEA1 1257-1411 C1405S , or MBP-EEA1 1-209 (NT). c, BHK-21 cells were transfected with EEA1 and/or Myc-syntaxin-6, as indicated. Proteins were immunoprecipitated (IP) using a human anti-EEA1 serum, normal human serum (nhs), or anti-Myc antibodies and detected by immunoblotting (IB) using anti-Myc antibodies or anti-EEA1 serum. d, GST-syntaxin-6 (lanes 1-3) or GST (lanes 4) were immobilized on glutathione-Sepharose and incubated with MBP-EEA1 1257-1411 that had been preincubated with buffer alone (lanes 1 and 4) or with His 6 -Rab5 loaded with GTP␥S (lanes 2) or GDP (lanes 3). The bound MBP-EEA1 1257-1411 was detected by immunoblotting with anti-MBP antibodies (11). The experiment was done in duplicate.  (14), were found to interact with EEA1 in the two-hybrid system. In contrast, syntaxin-6, which has been implicated in TGN-endosome trafficking (24), was found to interact strongly with EEA1, as indicated by the high ␤-galactosidase activity associated with the yeast transformants. The specific interaction between EEA1 and syntaxin-6 could also be demonstrated when EEA1 was cloned into the prey vector and the syntaxins were cloned into the bait vector, although these bait constructs resulted in some reporter gene activation by themselves (data not shown). We detected no interaction between syntaxin-6 and the GTPase-deficient Rab5 mutant, Rab5 Q79L , or with another Rab5 effector, Rabaptin-5 (18,20). Altogether, these results indicate that EEA1 and syntaxin-6 interact specifically with each other.
To identify the syntaxin-6 interacting domain of EEA1, we tested deletion mutants of EEA1 against syntaxin-6 in the two-hybrid system. While the N terminus of EEA1 showed no interaction, the C terminus (residues 1257-1411) was found to interact with syntaxin-6 ( Table I). We have previously identified residues 1277-1411 as the minimal endosomal binding domain of EEA1 (25), but this region showed no significant interaction with syntaxin-6, suggesting that the syntaxin-6 binding region is slightly larger than the minimal endosome binding region. Residues 1277-1348 constitute a minimal Rab5-binding domain that interacts with Rab5 Q79L (Table I) (11), whereas the very C terminus of EEA1 (residues 1325-1411) comprises a phosphatidylinositol 3-phosphate (PtdIns-(3)P) binding FYVE finger (11,12,26). Neither the minimal Rab5 binding region nor the FYVE finger alone interacted with syntaxin-6 (Table I). Interestingly, a mutation (C1405S) that abolishes PtdIns (3)P and endosome binding (25,26) led to a loss of syntaxin-6 binding as well. Taken together, these results suggest that the syntaxin-6 binding region of EEA1 encompasses both the Rab5 binding region and the FYVE finger.
In Vitro and in Vivo Biochemical Interactions between EEA1 and Syntaxin-6 -To study whether the interaction between EEA1 and syntaxin-6 can be detected biochemically, we prepared fusion proteins between GST and the cytoplasmic domains of syntaxin-6, syntaxin-7, and syntaxin-16. GST alone or the fusion proteins were immobilized on glutathione-Sepharose beads, which were incubated with cytosol and membrane extract from HeLa cells. After washing the beads, we analyzed the bound material by SDS-PAGE and Western blotting with anti-EEA1 antibodies. As shown in Fig. 1a, a significant portion of cytosolic and membrane-associated EEA1 bound to GST-syntaxin-6, whereas EEA1 did not associate with GST alone, GST-syntaxin-7, and GST-syntaxin-16. We also detected no interaction with GST-syntaxin13 (not shown). To test if the interaction between EEA1 and syntaxin-6 is direct, we performed a similar GST pull-down experiment using recombinant EEA1 instead of cytosol. Like cytosolic EEA1, the recombinant full-length EEA1 was found to bind specifically to GST-syntaxin-6 ( Fig. 1b). Likewise, the recombinant C terminus of EEA1 (as a fusion with maltose binding protein, MBP) was found to interact with syntaxin-6, whereas the C1405S mutant and the N terminus showed no interaction. These experiments support the data from the two-hybrid system and indicate that the C terminus of EEA1 interacts directly and specifically with syntaxin-6.
To investigate if EEA1 and syntaxin-6 can interact in vivo, we subjected cell lysates to immunoprecipitation with anti-EEA1 antibodies and studied by SDS-PAGE and Western blotting if syntaxin-6 was coimmunoprecipitated. As shown in Fig.  1c, in cells coexpressing Myc-epitope-tagged (17)  indicate that EEA1 interacts with syntaxin-6 in vivo as well as in vitro.
The assignment of binding sites for both Rab5 and syntaxin-6 to the C terminus of EEA1 led us to investigate if Rab5 and syntaxin-6 bind competitively to EEA1. For this purpose, we immobilized GST-syntaxin-6 on glutathione-Sepharose beads and studied the binding of MBP-EEA1 1257-1411 in the presence of Rab5 complexed with either GDP or the slowly hydrolyzable GTP analogue, GTP␥S (Fig. 1d). Under the experimental conditions used, we detected some unspecific binding of the EEA1 construct to GST alone (lanes 4), but the binding to GST-syntaxin-6 (lanes 1) was significantly higher. Interestingly, binding was reduced to background level in the presence of Rab5-GTP␥S (lanes 2) but remained unaltered in the presence of Rab5-GDP (lanes 3). This indicates that the binding sites for Rab5-GTP and syntaxin-6 at the C terminus of EEA1 physically overlap.
Syntaxin-6 Colocalizes with EEA1 on Early Endosomes-To study if EEA1 and syntaxin-6 colocalize in the cell, we doublestained BHK cells with antibodies recognizing the endogenous proteins. As found previously (16), EEA1 was detected on vesicular structures, representing early endosomes (Fig. 2a). Syntaxin-6, on the other hand, was abundant in a juxtanuclear region (Fig. 2b). However, syntaxin-6 was also detected on small vesicles, and it colocalized extensively with EEA1 on early endosomes (Fig. 2c, yellow). This localization is consistent with a previous electron microscopic study that showed syntaxin-6 present on TGN vesicles and endosomes as well as in the TGN (24). When the GTPase-deficient Q79L mutant of Rab5 is expressed, early endosomes expand because of increased fusion activity (20). To investigate if syntaxin-6 is recruited onto these enlarged endosomes, we transfected cells with Rab5 Q79L and stained them with anti-EEA1 and anti-syntaxin-6 antibodies. The large endosomes clearly contained EEA1 (Fig. 2d) as well as syntaxin-6 ( Fig. 2e), and the two proteins colocalized on these structures (Fig. 2f). When a panel of epitope-tagged syntaxins were coexpressed with Rab5 Q79L , only syntaxin-6, syntaxin-7, and syntaxin-13 were found to be recruited on the enlarged early endosomes, whereas syntaxin-2, syntaxin-4, syntaxin-5, syntaxin-11, and syntaxin-16 were not (data not shown). These results indicate that endogenous syntaxin-6 colocalizes with EEA1 on endosomes, and that syntaxin-6 represents a minority of syntaxins that can be found on these organelles. The colocalization between syntaxin-6 and EEA1 is consistent with the idea that these two proteins may functionally interact to regulate TGN-endosome trafficking.

DISCUSSION
This paper provides the first evidence for a direct interaction between a Rab effector and a SNARE molecule. Previously, a direct interaction between the Rab GTPase Ypt1p and the SNARE Sed5p has been detected in yeast. However, this interaction has so far not been demonstrated to be GTP-dependent (27). Another yeast protein, Vac1p, has recently been shown to interact directly with the Rab5 homologue Vps21p as well as with Vps45p, a member of the Sec1p family of putative SNARE regulators (28,29). Interestingly, Vac1p and Vps45p have previously been found in a complex with Pep12p (30), a possible yeast homologue of syntaxin-6 (31). Vac1p has been implicated in the docking of post-Golgi vesicles with early endosomes (28,30,32,33), and it shares regions of sequence similarity with EEA1, including a phosphatidylinositol 3-phosphate binding FYVE finger (25). This raises the possibilities that Vac1p may represent an EEA1 homologue in yeast (34) and that Vac1p/ Pep12p and EEA1/syntaxin-6 may regulate equivalent trafficking steps. The fact that syntaxin-6 interacts with a mammalian homologue of Vps45p (35) is consistent with this view.
We propose that Rab5 and EEA1, like their putative yeast homologues, Vps21p and Vac1p, may play a role in TGN to early endosome trafficking in addition to regulating endocytic membrane fusion. The interaction between syntaxin-6 and EEA1 could thus mediate the tethering of a post-Golgi vesicle to an early endosome. Our finding that Rab5-GTP and syntaxin-6 bind competitively to the membrane-proximal C terminus of EEA1 further suggests that syntaxin-6 binding may cause a dissociation between EEA1 and Rab5. We speculate that another (unknown) SNARE present on the endosome may then pair with syntaxin-6 to form a tight pre-fusion complex. It will now be interesting to study if Rab5 and EEA1 may regulate, for instance, the routing of lysosomal enzymes from the TGN to endosomes. Likewise, it will be important to study if syntaxin-6 may regulate homotypic early endosome fusion. The existence of multiple endosomal SNAREs raises the possibility that EEA1 may interact with other SNAREs to regulate this process.