Subunit Structure of a Mammalian ER/Golgi SNARE Complex
2000; Elsevier BV; Volume: 275; Issue: 50 Linguagem: Inglês
10.1074/jbc.m007684200
ISSN1083-351X
AutoresDalu Xu, Ashwini P. Joglekar, Antionette L. Williams, Jesse Hay,
Tópico(s)Calcium signaling and nucleotide metabolism
ResumoSNAP receptor (SNARE) complexes bridge opposing membranes to promote membrane fusion within the secretory and endosomal pathways. Because only the exocytic SNARE complexes have been characterized in detail, the structural features shared by SNARE complexes from different fusion steps are not known. We now describe the subunit structure, assembly, and regulation of a quaternary SNARE complex, which appears to mediate an early step in endoplasmic reticulum (ER) to Golgi transport. Purified recombinant syntaxin 5, membrin, and rbet1, three Q-SNAREs, assemble cooperatively to create a high affinity binding site for sec22b, an R-SNARE. The syntaxin 5 amino-terminal domain potently inhibits SNARE complex assembly. The ER/Golgi quaternary complex is remarkably similar to the synaptic complex, suggesting that a common pattern is followed at all transport steps, where three Q-helices assemble to form a high affinity binding site for a fourth R-helix on an opposing membrane. Interestingly, although sec22b binds to the combination of syntaxin 5, membrin, and rbet1, it can only bind if it is present while the others assemble; sec22b cannot bind to a pre-assembled ternary complex of syntaxin 5, membrin, and rbet1. Finally, we demonstrate that the quaternary complex containing sec22b is not an in vitroentity only, but is a bona fide species in living cells. SNAP receptor (SNARE) complexes bridge opposing membranes to promote membrane fusion within the secretory and endosomal pathways. Because only the exocytic SNARE complexes have been characterized in detail, the structural features shared by SNARE complexes from different fusion steps are not known. We now describe the subunit structure, assembly, and regulation of a quaternary SNARE complex, which appears to mediate an early step in endoplasmic reticulum (ER) to Golgi transport. Purified recombinant syntaxin 5, membrin, and rbet1, three Q-SNAREs, assemble cooperatively to create a high affinity binding site for sec22b, an R-SNARE. The syntaxin 5 amino-terminal domain potently inhibits SNARE complex assembly. The ER/Golgi quaternary complex is remarkably similar to the synaptic complex, suggesting that a common pattern is followed at all transport steps, where three Q-helices assemble to form a high affinity binding site for a fourth R-helix on an opposing membrane. Interestingly, although sec22b binds to the combination of syntaxin 5, membrin, and rbet1, it can only bind if it is present while the others assemble; sec22b cannot bind to a pre-assembled ternary complex of syntaxin 5, membrin, and rbet1. Finally, we demonstrate that the quaternary complex containing sec22b is not an in vitroentity only, but is a bona fide species in living cells. solubleN-ethylmaleimide-sensitive factor attachment protein receptor Golgi SNARE of 28 kDa glutathioneS-transferase hemagglutinin normal rat kidney synaptosomal-associated protein of 25 kDa SNARE containing glutamine in the zero-layer position SNARE containing arginine in the zero-layer position vesicle-associated membrane protein vesicular tubular cluster dithiothreitol endoplasmic reticulum bovine serum albumin polyacrylamide gel electrophoresis Complexes of soluble N-ethylmaleimide-sensitive factor attachment protein receptors (SNAREs)1 on opposing membranes are required for membrane fusion within the secretory pathway (1Rothman J.E. Nature. 1994; 372: 55-63Crossref PubMed Scopus (2011) Google Scholar, 2Hay J.C. Scheller R.H. Curr. Opin. Cell Biol. 1997; 9: 505-512Crossref PubMed Scopus (254) Google Scholar). The atomic structure of a single SNARE complex that is required for synaptic vesicle fusion has been solved and was shown to consist of a highly twisted parallel bundle of four amphipathic helices, one from the vesicle protein VAMP, one from the plasma membrane protein syntaxin 1A and two from the plasma membrane protein SNAP-25 (3Sutton R.B. Fasshauer D. Jahn R. Brunger A.T. Nature. 1998; 395: 347-353Crossref PubMed Scopus (1932) Google Scholar, 4Poirier M.A. Xiao W. Macosko J.C. Chan C. Shin Y.K. Bennett M.K. Nat. Struct. Biol. 1998; 5: 765-769Crossref PubMed Scopus (419) Google Scholar). Although the contacting surfaces of the helices are mostly hydrophobic, a layer of residues in the center of the bundle, referred to as the "zero layer," are comprised of glutamines for the plasma membrane SNAREs (Q-SNAREs) and arginine for the vesicle SNARE (an R-SNARE). Most SNARE proteins appear to possess either a glutamine or arginine at that deduced position, although the functional importance of these residues are not known (5Fasshauer D. Sutton R.B. Brunger A.T. Jahn R. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 15781-15786Crossref PubMed Scopus (756) Google Scholar). Although the precise role of SNAREs in the specific docking and fusion of transport vesicles remains controversial, different types of experiments indicate that SNARE complexes are the core machinery for intracellular membrane fusion. For example, purified synaptic SNAREs catalyze the rapid and efficient fusion of synthetic vesicles (6Parlati F. Weber T. McNew J.A. Westermann B. Sollner T.H. Rothman J.E. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 12565-12570Crossref PubMed Scopus (217) Google Scholar, 7Nickel W. Weber T. McNew J.A. Parlati F. Sollner T.H. Rothman J.E. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 12571-12576Crossref PubMed Scopus (160) Google Scholar). Our understanding of SNARE function is limited by a lack of comparisons between SNARE complexes from different fusion events. Without structural knowledge of several SNARE complexes, we cannot understand the functional distinction between Q- and R-SNAREs, nor the general significance of the synaptic four-helix bundle. Although many SNAREs have been discovered and implicated in membrane fusion events, it is only the exocytic complex for which even a subunit structure is known. SNARE complexes outside the synapse have been studied primarily using immunoprecipitation experiments in cellular detergent extracts. Five SNAREs involved in yeast vacuolar fusion were shown to coprecipitate, implying the existence of a pentameric cis-SNARE complex (8Ungermann C. von Mollard G.F. Jensen O.N. Margolis N. Stevens T.H. Wickner W. J. Cell Biol. 1999; 145: 1435-1442Crossref PubMed Scopus (134) Google Scholar). However, because at least one of these proteins, Ykt6p, has been shown to operate in multiple transport steps (8Ungermann C. von Mollard G.F. Jensen O.N. Margolis N. Stevens T.H. Wickner W. J. Cell Biol. 1999; 145: 1435-1442Crossref PubMed Scopus (134) Google Scholar, 9McNew J.A. Sogaard M. Lampen N.M. Machida S. Ye R.R. Lacomis L. Tempst P. Rothman J.E. Sollner T.H. J. Biol. Chem. 1997; 272: 17776-17783Abstract Full Text Full Text PDF PubMed Scopus (178) Google Scholar), it is possible that the five associated SNAREs are present in two or more distinct complexes with overlapping members. Likewise, in ER/Golgi transport, yeast Sed5p (and mammalian syntaxin 5) coprecipitated with an array of other SNAREs (10Sogaard M. Tani K. Ye R.R. Geromanos S. Tempst P. Kirchhausen T. Rothman J.E. Sollner T. Cell. 1994; 78: 937-948Abstract Full Text PDF PubMed Scopus (443) Google Scholar, 11Hay J.C. Chao D.S. Kuo C.S. Scheller R.H. Cell. 1997; 89: 149-158Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar), including Gos1p, Bos1p, Sec22p, Ykt6p, and Bet1p (in mammals syntaxin 5 coimmunoprecipitated GOS-28, membrin, sec22b, and rbet1). Further immunoprecipitations in mammals demonstrated that syntaxin 5 and GOS-28 appeared to be in separate complexes from syntaxin 5 and sec22b, membrin, and rbet1 (12Hay J.C. Klumperman J. Oorschot V. Steegmaier M. Kuo C.S. Scheller R.H. J. Cell Biol. 1998; 141: 1489-1502Crossref PubMed Scopus (140) Google Scholar). The latter group probably functions in membrane fusion between ER-derived vesicles and vesicular tubular clusters (VTCs) or between homotypically fusing ER-derived vesicles (13Rowe T. Dascher C. Bannykh S. Plutner H. Balch W.E. Science. 1998; 279: 696-700Crossref PubMed Scopus (98) Google Scholar). Syntaxin 5, rbet1, membrin, and sec22b have all been localized by immunoelectron microscopy to VTCs (12Hay J.C. Klumperman J. Oorschot V. Steegmaier M. Kuo C.S. Scheller R.H. J. Cell Biol. 1998; 141: 1489-1502Crossref PubMed Scopus (140) Google Scholar), and all four of their yeast counterparts exist on ER-derived vesicles generated in vitro (14Cao X. Barlowe C. J. Cell Biol. 2000; 149: 55-66Crossref PubMed Scopus (118) Google Scholar, 15Rexach M.F. Latterich M. Schekman R.W. J. Cell Biol. 1994; 126: 1133-1148Crossref PubMed Scopus (94) Google Scholar, 16Lian J.P. Ferro-Novick S. Cell. 1993; 73: 735-745Abstract Full Text PDF PubMed Scopus (121) Google Scholar). These four ER/Golgi SNAREs could either form two or more overlapping ternary complexes or a single quaternary complex containing all four members. Immunoprecipitation experiments could not distinguish between these possibilities, whereas functional results from yeast mutants suggested that members of this group of proteins functioned in at least two distinct fusion complexes (17Spang A. Schekman R. J. Cell Biol. 1998; 143: 589-599Crossref PubMed Scopus (109) Google Scholar). The precise compositions of ER/Golgi SNARE complexes were not clearly demonstrated using purified yeast ER/Golgi SNAREs, because Sed5p appeared to engage in several potentiated ternary complexes including Sed5p·Sec22p·Bos1p and Sed5p·Bet1p·Bos1p (18Stone S. Sacher M. Mao Y. Carr C. Lyons P. Quinn A.M. Ferro-Novick S. Mol. Biol. Cell. 1997; 8: 1175-1181Crossref PubMed Scopus (43) Google Scholar, 19Sacher M. Stone S. Ferro-Novick S. J. Biol. Chem. 1997; 272: 17134-17138Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar). Which one of these ternary complexes, if any, was a complete SNARE complex? The simple possibility that all four proteins form a single quaternary complex has not been experimentally addressed. We used purified mammalian ER/Golgi SNAREs to precisely define the higher-order SNARE complexes they can form in vitro. Here we document the subunit composition, assembly, and regulation of a quaternary SNARE complex that likely mediates the first fusion event in the secretory pathway. The complex demonstrates a remarkable similarity to the synaptic fusion complex, establishing a pattern to be repeated at other transport steps. Syntaxin 5, membrin, and rbet1, three Q-SNAREs, assemble to create a high affinity binding site for sec22b, an R-SNARE. This assembly pattern is consistent with sec22b opposing the other three proteins on opposite membranes, as VAMP opposes SNAP-25 and syntaxin 1 in the synapse. The resulting quaternary complex has a stoichiometry of 1:1:1:1 (Q:Q:Q:R, as in the synapse), perhaps arranged in a four-helix bundle as in the synapse, but where the Q-helices of membrin and rbet1 (one from each) play the role of the two SNAP-25 helices. In addition, we have explored the kinetic assembly mechanism of the ER/Golgi complex and its negative regulation by the syntaxin 5 amino-terminal domain. Monoclonal and affinity purified polyclonal anti-SNARE antibodies were described previously (11Hay J.C. Chao D.S. Kuo C.S. Scheller R.H. Cell. 1997; 89: 149-158Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar, 12Hay J.C. Klumperman J. Oorschot V. Steegmaier M. Kuo C.S. Scheller R.H. J. Cell Biol. 1998; 141: 1489-1502Crossref PubMed Scopus (140) Google Scholar). Anti-VAMP2 polyclonal antibody was obtained from Stressgen (Victoria, British Columbia, Canada). GST fusion proteins were expressed in Escherichia coli using vector pGEX-KG (20Guan K.L. Dixon J.E. Anal. Biochem. 1991; 192: 262-267Crossref PubMed Scopus (1641) Google Scholar), whereas hexahistidine-tagged sec22b employed pQE-9 (Qiagen). See the Fig. 1 legend for the specific residues of each protein expressed. Bacteria were grown in Luria Broth at 37 °C to an A600 of 0.4–0.6, when they were either shifted to 15 °C (for GST-syntaxin 5 constructs) for 30 min prior to protein induction at 15 °C for 2–3 h, or induced immediately at 37 °C (all other constructs) for 2–3 h. GST fusion protein production was induced with 0.1 mmisopropyl-1-thio-β-d-galactopyranoside; His6-sec22b with 1 mmisopropyl-1-thio-β-d-galactopyranoside. After harvesting the cultures, bacteria were resuspended in French Press buffer (for all constructs except sec22b: 50 mm Tris, pH 8.0, 0.1m NaCl, 1 mm EDTA, 0.05% Tween 20, 1 mm DTT, 2 μg/ml leupeptin, 4 μg/ml aprotinin, 1 μg/ml pepstatin A, and 1 mm phenylmethylsulfonyl fluoride; for sec22b, the same buffer lacking DTT) at 20 ml/liter of culture, French Pressed twice, and centrifuged at 20,000 × g for 20 min, and the supernatant (S1) was recovered. For His6-sec22b, GST-rbet1, and GST-syntaxin 5 (residues 251–333), the S1 was immediately centrifuged at 100,000 ×g for 45 min, and the supernatant was collected. For GST-syntaxin 5 (residues 55–333), the S1 was adjusted to 0.35% sodium sarkosyl, mixed gently for 30 min, and then supplemented to 1% Triton X-100 and 10% glycerol. After centrifugation at 100,000 ×g for 45 min, the supernatant was collected. For GST-membrin, the S1 was centrifuged at 100,000 × g for 45 min, the supernatant was discarded and the pellet was homogenized in the original volume of French Press buffer. This was treated with sarkosyl, Triton X-100, and glycerol and centrifuged as above. The final 100,000 × g supernatant was retained. For GST fusion proteins, the supernatant from the final 100,000 ×g spin was loaded onto a glutathione-Sepharose column (Amersham Pharmacia Biotech), washed extensively with phosphate-buffered saline, and eluted with 50 mm Tris, pH 8.0, 20 mm glutathione containing 0.1% Triton X-100. For His6-sec22b, the final supernatant was loaded onto a column of Ni2+-nitrilotriacetic acid-agarose (Qiagen); washed with 50 mm Tris, pH 8.0, 0.3 m NaCl, 0.05% Tween 20; washed again with the same buffer plus 0.025m imidazole; and eluted with a 0.025–0.25 mgradient of imidazole in the same buffer. From this point on, manipulations varied for each protein. Glutathione-purified GST-rbet1 was cleaved in solution with an optimal concentration (determined on each occasion) of human thrombin (Sigma) for 1 h at room temperature and then passed over a Q-Sepharose column equilibrated with 20 mm Tris, pH 8.0, 1 mm EGTA, 1 mm DTT, to remove GST. Rbet1 in the flow-through was then concentrated on a YM-3 membrane (Amicon) to ∼100 μg/ml, dialyzed for ≥4 h into Buffer A (20 mmHepes, pH 7.2, 0.15 m KCl, 2 mm EDTA, 5% glycerol) then supplemented with 1 mm DTT, 2 μg/ml leupeptin, 4 μg/ml aprotinin, 1 μg/ml pepstatin A, and stored at −80 °C. Glutathione-purified syntaxin 5 (residues 55–333) was cleaved in solution with an optimal concentration of human thrombin for 1 h at room temperature, then either dialyzed into Buffer A or in some cases further purified by velocity sedimentation on 5–30% glycerol gradients run in Buffer A plus 0.2% Triton X-100. The 5-ml gradients were overlaid with 0.5 ml sample and spun at 201,000 ×g (at rav) for 24 h in a Beckman MLS-50 rotor and the purest fractions were supplemented with protease inhibitors (see above) and stored at −80 °C. Glutathione-purified GST-membrin was first desalted into 20 mm Tris, pH 7.6, 0.1m KCl, 1 mm EGTA on Sephadex G-25 (Amersham Pharmacia Biotech), cleaved in solution with thrombin, passed back over glutathione-Sepharose to deplete GST and uncleaved fusion protein, and then purified on velocity gradients as described above. Ni2+-NTA-agarose-purified sec22b was separated from a major degradation product by two consecutive runs on a 100-ml Superose 12 column (Amersham Pharmacia Biotech) run in 20 mm Tris, pH 7.6, 0.1 m KCl, 1 mm EGTA. The purified His6-sec22b was then dialyzed into Buffer A, supplemented with protease inhibitors as above, and stored at −80 °C. All binding incubations were conducted in Buffer A (see above) containing 0.1% Triton X-100. A typical bead-binding reaction consisted of 10 μl of 50% glutathione-Sepharose beads containing 124 pmol (620 nm) of immobilized protein and 80–620 pmol (400 nm to 3.1 μm) of soluble binding partners in a final volume of 200 μl. After 1 h at 4 °C with constant agitation, beads were washed three times with Buffer A plus 0.1% BSA prior to solubilization of the bound proteins in SDS-PAGE sample buffer. All Western blots that were quantitated utilized 125I-labeled secondary antibodies and standards quantified relative to BSA. For solution binding assays, which were incubated for varying time periods at 4 °C (see each figure legend), a 300-μl reaction typically contained ∼600 pmol (2 μm) of each protein, 250 μl of which was injected onto a 24-ml Superdex 200 gel filtration column (Amersham Pharmacia Biotech) run in Buffer A containing 0.1% Triton X-100 and 30 μg/ml BSA. Individual column fractions were either analyzed directly by SDS-PAGE and Western blotting or were precipitated with acetone prior to gel and Western analysis. For the experiment in Fig. 8, A and B, typical ternary or quaternary binding incubations were carried out for 1 h, followed by gel filtration as usual. To the purified ternary complex (fraction 17) we added twice the normal concentration of sec22b and incubated for an additional hour. This fraction along with the corresponding purified quaternary complex fraction were then mock-immunodepleted using BSA-blocked 1:1 protein A: protein G beads, and the supernatant was collected and centrifuged at 100,000 ×g for 30 min. Supernatant from the 100,000 ×g spin was immunoprecipitated using BSA-blocked protein A/G beads for a control or the same beads containing covalently cross-linked (using dimethylpimelimidate) monoclonal 4E11 anti-rbet1 antibodies. The beads were then washed three times with Buffer A containing 0.1% Triton X-100, and noncovalently bound proteins were solubilized for SDS-PAGE as described below in the immunoprecipitation section.Figure 4Formation of a highly cooperative ternary complex composed of rbet1, membrin, and syntaxin 5. Purified soluble SNARE proteins were mixed in solution for 1 h in the combinations indicated (right), and gel-filtered on Superdex 200. Individual column fractions were immunoblotted for rbet1 (top four panels), membrin (fifth and sixth panels) and syntaxin 5 (bottom two panels). The lower band in the seventh and eighth panels is a degradation product of syntaxin 5 (see Fig. 1). Elution positions of native thyroglobulin (669 kDa), catalase (232 kDa), BSA (67 kDa), and chymotrypsinogen A (25 kDa) are indicated with arrows above the fraction numbers.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 3Syntaxin 5 and rbet1 potentiate each other in binding to bead-immobilized GST-membrin. Soluble syntaxin 5 (400 nm in A, 2 μm in B) and rbet1 (3.1 μm in A, 1.5 μm inB) were mixed with glutathione beads preloaded with either GST or GST-membrin (620 nm). After a 1-h incubation at 4 °C, beads were washed three times and analyzed by quantitative immunoblotting for syntaxin 5 (A) or rbet1 (B). Ponceau staining of blots indicated that equal amounts of GST or GST-membrin were loaded for each condition (not shown). At the doses used in part A, a maximum of 6.6% of the added syntaxin 5 bound to the beads with a molar ratio of 0.044 (syntaxin 5/GST-membrin). In part B, 14.3% of added rbet1 bound with a molar ratio of 0.36 (rbet1/GST-membrin). All values represent the mean of two duplicates ± S.E.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 1Purified recombinant ER/Golgi SNAREs.Proteins were expressed as GST fusions (membrin, rbet1, and syntaxin 5) or a hexahistidine-tagged construct (His6-sec22b) inE. coli and purified by glutathione- or nickel-affinity chromatography. Thrombin cleavage of fusion proteins and subsequent purification steps are described under "Experimental Procedures." Approximately 200 ng of each protein was loaded in each gel lane. The His6-sec22b and rbet1 constructs encoded essentially the entire cytoplasmic domains of the proteins (2–196 and 1–95, respectively). Rbet1 appears as a doublet of bands that behave nearly identically in binding assays; the lower band is a carboxyl-terminal degradation/truncation product (amino acids 1 to ∼90). Syntaxin 5 encoded the entire cytoplasmic domain of the smaller 34-kDa syntaxin 5 isoform (residues 55–333 of the entire syntaxin 5 open reading frame, Ref. 24Hui N. Nakamura N. Sonnichsen B. Shima D.T. Nilsson T. Warren G. Mol. Biol. Cell. 1997; 8: 1777-1787Crossref PubMed Scopus (81) Google Scholar). The upper syntaxin 5 band contains residues 55–333, the second band is an amino-terminal degradation product (amino acids 70–333), and the third band is a carboxyl-terminal degradation product (amino acids 55–251) produced during thrombin digestion (see the legend to Fig. 7 for details). The upper of the two syntaxin 5 degradation products was active in some binding reactions, the lower was not. The membrin construct encoded essentially the entire protein including the transmembrane domain (residues 2–212). These four purified protein constructs were used throughout our studies except where specifically stated otherwise.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 8The isolated ternary complex does not bind sec22b. Purified soluble SNAREs were mixed and incubated in solution to form the ternary complex lacking sec22b (as in Fig. 4) or the complete quaternary complex (as in Fig. 6), and the respective complexes were purified by gel filtration. The isolated ternary complex was then incubated for 2 h at 4 °C with purified sec22b at twice the concentration normally used for quaternary complex assembly. Both complexes were then immunoprecipitated with anti-rbet1 monoclonal antibodies (4E11) that do not affect the formation of SNARE complexes.A, compositions of the samples prior to immunoprecipitation, analyzed by immunoblotting. B, immunoprecipitated pellets analyzed by immunoblotting. Control represents the ternary mixture immunoprecipitated with beads lacking antibody 4E11. Immunoblotting of the supernatant remaining after immunoprecipitation revealed that all of the rbet1 had been depleted in both of the samples precipitated with immune beads (not shown). C, purified soluble SNAREs were mixed and incubated in solution to form the ternary complex lacking sec22b. After 2.5 h at 4 °C, sec22b was added to the ternary mixture (sequential assembly) and a new binding reaction was begun containing identical amounts of all four proteins (simultaneous assembly). After 1 h at 4 °C, both binding mixtures were gel filtered and individual column fractions were immunoblotted for sec22b. A, B, andC utilized only the membrane-proximal helix of syntaxin 5 (amino acids 251–333) rather than the full cytoplasmic domain.View Large Image Figure ViewerDownload Hi-res image Download (PPT) E. coli strain NM522 containing the protein construct was streaked on minimal M9 plates and grown 48 h at 37 °C. Fresh colonies were then used without storage to innoculate 100 ml of minimal medium (1× M9 salts, 0.2% dextrose, 0.1 mm CaCl2, 2.0 mmMgCl2, 50 μg/ml ampicillin, and 10 μg/ml thiamine hydrochloride) and grown overnight to high density. One ml of this culture was then used to innoculate fresh 100-ml cultures in the above medium containing 0.5 mCi of [U-14C]glucose (ICN cat. no. 1104701, 316 mCi/mmol). Protein induction and purification were essentially as described above. Solution-binding reactions were similar to nonradioactive reactions (above). After gel filtration as usual, fractions 17 and 18 were pooled and either directly acetone precipitated or immunoprecipitated with monoclonal 4E11 anti-rbet1 antibody covalently cross-linked (using dimethylpimelimidate) to a 1:1 mixture of protein A- and protein G-Sepharose (Amersham Pharmacia Biotech). Samples were solubilized in SDS-PAGE sample buffer and electrophoresed using high Tris/urea gels (21Schlenstedt G. Gudmundsson G.H. Boman H.G. Zimmermann R. J. Biol. Chem. 1990; 265: 13960-13968Abstract Full Text PDF PubMed Google Scholar) as described. Gels were Coomassie Blue-stained and dried, and bands were first quantitated with a laser densitometer and then exposed to Kodak Biomax-MS x-ray film with a Kodak Biomax Transcreen-LE intensifying screen. Radioactive bands were then excised, minced, and incubated overnight at 50 °C with 1 ml of TS-2 tissue solubilizer (Research Products International) and counted in 10 ml 3a20 toluene-based counting mixture (Research Products International). To calculate molar stoichiometries using Coomassie Blue staining, stain intensities were divided by the predicted protein molecular weight and expressed relative to syntaxin 5. To calculate stoichiometries using radioactivity, 14C counts in each band were divided by the predicted number of carbon atoms in each protein construct (calculated from the sequence), and the resulting quotient was expressed relative to that of syntaxin 5. NRK cells were transfected as before (22Hay J.C. Hirling H. Scheller R.H. J. Biol. Chem. 1996; 271: 5671-5679Abstract Full Text Full Text PDF PubMed Scopus (77) Google Scholar) with a HA-syntaxin 5 full-length construct in pCMV4 (23Bennett M.K. Garcia-Arraras J.E. Elferink L.A. Peterson K. Fleming A.M. Hazuka C.D. Scheller R.H. Cell. 1993; 74: 863-873Abstract Full Text PDF PubMed Scopus (591) Google Scholar) or Myc-tagged full-length sec22b in pCDNA3 (11Hay J.C. Chao D.S. Kuo C.S. Scheller R.H. Cell. 1997; 89: 149-158Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar). Forty-eight hours after transfection, cells were depleted of intracellular ATP by rinsing them twice with glucose-free buffer (25 mm Hepes, pH 7.3, 70 mm sucrose, 130 mm NaCl, 4.8 mm KCl, 1.3 mmCaCl2, 1.2 mm MgSO4), and they were incubated for 10 min at 37 °C with the same buffer containing 0.2 mm iodoacetic acid (sodium salt) and 0.1 μmantimycin A. Cells were then solubilized in KCl buffer (20 mm Hepes, pH 7.2, 0.1 m KCl, 2 mmEGTA, 2 mm EDTA) containing 1% Triton X-100, 1 mm DTT, 2 μg/ml leupeptin, 4 μg/ml aprotinin, 1 μg/ml pepstatin A, and 1 mm phenylmethylsulfonyl fluoride and centrifuged at 100,000 × g for 30 min. The clarified supernatant was subjected to immunoprecipitation using 1:1 protein A: protein G beads loaded with anti-Myc (9E10 from our tissue culture medium) antibodies covalently cross-linked with dimethylpimelimidate. 0.4 ml of extract was incubated with 10 μl of 50% antibody beads for 2 h. The beads were washed three times with KCl buffer containing 0.1% Triton X-100, and noncovalently bound proteins were solubilized by incubation for 5 min at room temperature with nonreducing SDS-PAGE sample buffer. After pelleting the beads, the bead-free supernatant was adjusted to ∼5% β-mercaptoethanol and subjected to SDS-PAGE and Western blotting. The immunoprecipitations with wild-type NRK and PC12 cells shown in Fig. 10, B and C followed essentially the same protocol except that we used rabbit preimmune IgG or affinity-purified anti-syntaxin 5 antibody covalently cross-linked to protein A-Sepharose. We expressed in bacteria the following protein constructs: 1) GST fused to the full cytoplasmic domain of the shorter isoform of rat syntaxin 5 (amino acids 55–333 where amino acid 1 would be the N terminus of the longer syntaxin 5 isoform, Ref. 24Hui N. Nakamura N. Sonnichsen B. Shima D.T. Nilsson T. Warren G. Mol. Biol. Cell. 1997; 8: 1777-1787Crossref PubMed Scopus (81) Google Scholar), 2) GST fused to full-length rat membrin (amino acids 2–212), 3) hexahistidine-tagged mouse sec22b cytoplasmic domain (amino acids 2–196), and 4) GST fused to rat bet1 cytoplasmic domain (amino acids 1–95). Although we were unable to express the cytoplasmic domain of membrin in bacteria, we found that the full-length protein expressed well and was partially soluble. In addition, we found that partially soluble GST-membrin and GST-syntaxin 5 fusion proteins could be completely solubilized using the ionic detergent sarkosyl (25Frangioni J.V. Neel B.G. Anal. Biochem. 1993; 210: 179-187Crossref PubMed Scopus (833) Google Scholar), which was then replaced with Triton X-100 during the purification on glutathione-Sepharose (see "Experimental Procedures" for details). The resulting GST-syntaxin 5 and GST-membrin were fully soluble and functional in binding reactions. All expressed proteins were first purified by glutathione-Sepharose or nickel-affinity chromatography, the GST moiety was cleaved from GST fusion proteins, and the proteins were further purified as detailed under "Experimental Procedures." Fig. 1 shows a Coomassie Blue-stained gel of the purified proteins used for most of the studies. Although all of the four VTC-localized ER/Golgi SNAREs coprecipitated in detergent extracts (12Hay J.C. Klumperman J. Oorschot V. Steegmaier M. Kuo C.S. Scheller R.H. J. Cell Biol. 1998; 141: 1489-1502Crossref PubMed Scopus (140) Google Scholar), we wanted to characterize the sets of direct protein interactions responsible for their association. We therefore immobilized GST-rbet1, GST-membrin, and GST-syntaxin 5 onto glutathione beads and incubated them separately with soluble rbet1, membrin, syntaxin 5, or sec22b. After washing with
Referência(s)