COG lobe B sub-complex interacts with v-SNARE GS15 on Golgi vesicles

Our data indicate that COG sub-complexes exist as separate entities on Golgi membranes/vesicle-like structures. This data is in good agreement with previous observations that t-SNARE STX5 and SM protein Sly1p interact with lobe A subunit COG4^18,19 and that exogenously expressed COG sub-complexes preferentially interact with different vesicle-tethering factors⁶. The vesicle-like localization of the lobe B sub-complex suggests that lobe B subunits might interact directly or indirectly with vesicle-associated proteins involved in tethering/fusion, such as v-SNAREs. The major intra-Golgi v-SNARE is GS15/Bet1L^20,21 which is the partner of STX5. Previous yeast two hybrid screens failed to identify COG as a partner for GS15^17,22. However, GFP IP from cells that stably express GFP-GS15 revealed strong interaction with the endogenous lobe B subunit COG8 (Fig. 4A). Moreover, in vivo co-expression experiments showed that lobe B expressed in HEK293T cells preferentially binds to GFP-GS15 (Fig. 4B,C). This interaction is likely to be direct, as evidenced by purified GS15-GST very efficiently interacting with purified COG in the in vitro setting (Fig. 4D), with preferential binding to lobe B sub-complex compared to the full COG complex. Importantly, removal of the SNARE domain (aa 21–86) completely abolished this interaction, indicating that lobe B preferentially interacts with the SNARE domain of v-SNARE GS15. GS15 is present on both Golgi rims and Golgi-derived vesicles²⁰; therefore we sought to determine the membrane compartment where lobe B interacts with GS15. Membranes from cells co-expressing GFP-GS15 and myc-tagged lobe B subunits were separated by differential centrifugation into P15 and P100 fractions (Fig. 4E). Golgi resident protein GPP130²³ was found primarily associated with P15 fraction while GFP-GS15 was present in both P15 and P100 fractions (Fig. 4F, Supplemental Figure 4B), indicating that the P15 fraction is enriched in Golgi membranes while P100 contains Golgi-derived vesicles. Pull-down of GFP-associated proteins revealed that lobe B preferentially interacts with GS15 in the vesicle (P100) fraction (Fig. 4E,F). Vesicle-associated lobe B-GS15 interaction appeared to be strong (up to 12% pull-down) and highly specific for the v-SNARE GS15 with no significant interaction detected between lobe B and either Golgi tethering factor p115 (data not shown) or endosomal v-SNARE Vamp3 (Supplemental Figure 4B). In contrast to lobe B-GS15 vesicle association, lobe A subunit COG4 was shown to interact with Golgi t-SNARE STX5¹⁷ while COG2 directly binds to Golgi coiled-coil tethering factor p115²⁴. P115 was originally found to be active in the in vitro intra-Golgi trafficking assay²⁵ along with the COG complex^4,26, indicating that these studies likely represent functional protein-protein interactions.

Open in a separate window

Figure 4

COG lobe B sub-complex preferentially interacts with GS15.

(A) GFP-GS15 interacts with endogenous COG8. HEK293T cells were transiently transfected with GFP-GS15 or GFP then immunoprecipitated. IPs were probed by western blot with antibodies against COG8, STX5, and GFP. (B) GFP-GS15 preferentially interacts with lobe B in vivo. HEK293T cells were transiently transfected with COG-Myc multi-expression plasmids (lobe A expression, lobe B expression, or the whole COG complex). 24h after COG-myc transfection cells were transfected with GFP-GS15. 24h post transfection cellular lysates were immunoprecipitated. IPs were probed by western blot with antibodies against Myc and GFP. (C) Co-IP efficiency values were generated by calculating the relative density of the co-IP Myc signal relative to the density of the input Myc signal and the recovered GFP signal in four independent experiments. (D) COG-GS15 in vitro binding assay. Glutathione beads were incubated with recombinant GST, GST-tagged full-length (FL) GS15 cytoplasmic domain (aa 1–86), GS15 with deletion of N-terminal domain (Δ1–20), and deletion of SNARE domain (Δ21–86). Bound proteins were eluted from beads, and probed by western blot with Myc antibodies. IP efficiency values were generated by normalizing the density of the IP Myc signal to the density of the whole COG/Lobe B input Myc signal. (E) Lobe B interacts with GS15 on vesicle membranes. HEK293T cells were transiently transfected with lobe B and GFP-GS15. 48h after transfection, cells were collected and fractionated into large membrane (P15), small membrane (P100), and cytosol (S100) by differential centrifugation. The membrane pellets were solubilized with Triton-X100 and then incubated with GBP-beads. IPs were probed by western blot with anti-Myc, anti-GFP, and anti-GPP130. (F) Co-IP efficiency values were calculated by dividing co-IP by input in four independent experiments. (G) HeLa cells stably expressing GFP-GS15 were stained for COG3 and COG8 and imaged using the Zeiss ELYRA S1. Line plots are shown measuring the relative value of signal intensity (y-axis) over the distance measured in pixels (x-axis). White arrows in the merged image indicate GS15 and COG8 co-localizing on vesicle-like structures. Bar, 10μm (inset 1μm).

In order to complement the biochemical data, we employed superresolution microscopy (Fig. 4G and Supplemental Figure 4C). In HeLa cells stably expressing GFP-GS15, a sub-population of lobe B subunit COG8 (P=0.59±0.09, n=15), but not lobe A subunit COG3 (P=0.29±0.04, n=15), was detected on GS15-positive vesicle-like structures (white arrowheads on the overlay insert). In summary, these results indicate that lobe B of the COG complex preferably interacts with v-SNARE GS15, and furthermore that this interaction is likely occurring on vesicle membranes.

Cell culture

HeLa and HEK293T cells (ATCC) were cultured in DMEM/F-12 medium (Thermo Scientific) supplemented with 15mM HEPES, 2.5mM L-glutamine and 10% FBS (Atlas Biologicals) at 37°C and 5% CO₂ in a 90% humidified incubator. YFP-COG3/CFP-COG6 stable cells were generated by transfection of CFP-COG6 plasmids into HeLa cells stably expressing YFP-COG3³⁷ followed by selection with Zeocin and single colony isolation.

HEK293T COG1 and COG8 stable knockouts were generated using CRISPR^48,49,50. gRNA sequences were provided by Horizon Discovery, COG8 Guide ID:123842 GGTGGAGGATGAAGGGCTCC, COG1 Guide ID: 129696 TTTCGAGACGCATGGAGCGG. HEK293T cells were transfected with plasmid containing gRNA with Cas9-dasherGFP (Horizon Discovery). Eight days after transfection, COG depleted cells were identified by increased cell surface binding of GNL-647 and single-cell sorted based on high 647 fluorescence intensity using the FACSARIA (BD Biosciences). Single cell colonies were then expanded. Cells were verified as nulls by WB and sequencing. To create the COG1/COG8 double KO cell line the COG1 KO cells obtained in the first round of selection were transfected again with COG8 gRNA/Cas9 construct, single sorted and verified by WB and sequencing.

Size exclusion chromatography

Size-exclusion chromatography was performed using Superose 6 10/300GL column with the AKTA purifier FPLC (GE Healthcare). The column was equilibrated in column buffer (125mM Potassium Acetate, 25mM HEPES, pH 7.2). Ferritin from equine spleen (Sigma, Type I, saline solution), alcohol dehydrogenase (Worthington Biochemical Corporation), and bovine serum albumin (RPI) diluted with column buffer with 0.1% Tween 20 to 10mg/mL were used for column calibration. The void volume (Vo) at ~7.5mL was used with the elution volume (Ve) to generate a protein standard curve to calculate the size of eluted proteins.

Sample preparation: HeLa cells were washed with DPBS and collected by scraping in column buffer with Halt protease inhibitor cocktail (Thermo Scientific) and 1mM PMSF. Collected cells were lysed by vortexing with 0.5mm glass beads (BioSpec) on ice. Total cell lysates were centrifuged at 660xg for 5min at 4°C to isolate the post-nuclear supernatant (PNS). PNS was separated into supernatant (cytosol) and pellet (membrane) fractions by ultracentrifugation at 100,000xg for 1hour at 4°C in a Beckman Optima ultracentrifuge using TLA 55 rotor. Supernatant (cytosol fraction) was concentrated with a NanoSep 30K centrifugal device for 20min at 20,000×g at 4°C (~10 fold concentration). Membrane pellet (membrane fraction) was suspended in 1%Triton X-100 in DPBS with protease inhibitors, spun down at 20,000xg at 4°C, and solubilized membranes collected. Triton X-100 (final concentration 1%) was added to cytosol fraction as well. Cytosol fraction and Triton X-100 soluble membrane fraction were loaded on column equilibrated with column buffer plus 0.1% Tween 20 at 0.4mL/min, 0.5mL fractions were collected after 0.35 CV elution. Fractions were concentrated by TCA precipitation (trypsin inhibitor added to final concentration of 0.01μg/μL, 5μL 0.02% sodium deoxycholate (Sigma) added and samples incubated for 15min on ice. 168.5μL of 24% TCA added and samples incubated on ice 1hour. Precipitated proteins were collected by centrifugation at 20,000xg for 20min at 4°C and resuspended in 45μl sample buffer.

Native IP of endogenous COG subunits from intracellular membranes

HeLa cells were collected on ice in DPBS containing Halt protease inhibitor cocktail (Thermo Scientific), and lysed by vortexing with 0.5mm glass beads (BioSpec). Total cell lysates were centrifuged at 660×g for 5min at 4°C to isolate the post-nuclear supernatant (PNS). PNS was separated into supernatant (cytosol) and pellet (membrane) fractions by ultracentrifugation at 100,000×g for 1hour at 4°C in a Beckman Optima ultracentrifuge using TLA 55 rotor. Membrane pellet (P100) was then resuspended in 1ml of DPBS containing 1% Triton X-100 and Halt protease inhibitor cocktail, and incubated for 30min on ice. Triton X-100 (final concentration 1%) was added to cytosol fraction as well. 30μl of protein A-Sepharose CL-4B beads was added to both supernatant and pellet fractions and incubated with gentle mixing by inversion for 1h on ice and finally clarified by centrifugation at 20,000×g for 10min to produce S20 (for cytosolic proteins) and P20 (for detergent-soluble membrane proteins) fractions. S20 and P20 were mixed with 3μg of affinity purified anti-hCOG3, or anti-hCOG6 antibodies and incubated overnight on ice. Samples were clarified by centrifugation as described above. 60μl of Protein G-Agarose beads slurry (50%) was added to each tube and incubated for 2h with gentle inversion on the rotator in a cold room. Beads were then washed four times with TBST. Samples were eluted in 2X sample buffer, and heated for 5min at 95°C to elute the bound proteins.

siRNA-induced knockdown

Transfection was performed using RNAiMAX siRNA Transfection Reagent (Life Technologies), following a protocol recommended by Life Technologies and cells were analyzed 72–96hours after transfection. (siRNA Human COG1-4: GUAGCGGCCUCUCCAUGAA).

Plasmid preparation and transfection

Mammalian expression constructs were generated using standard molecular biology techniques or obtained as generous gifts. See Table 1 for complete list of plasmids. Plasmids were isolated from bacterial cells using the QIAprep Spin Miniprep or Midiprep Kits (Qiagen). Plasmid transfections were performed using Lipofectamine 2000 (Life Technologies).

Fractionation of mechanically disrupted cells

HEK293T cells were collected by gentle resuspension and pelleted by centrifugation at 500xg for 2min. Cell pellet was resuspended in cold homogenization buffer (25mM HEPES-KOH pH 7.2, 150mM KCl, 2mM EDTA, 5μL/mL Halt protease inhibitor cocktail (Thermo Scientific), and 1mM PMSF). Cells were lysed by passage through a 27G needle 30 times, or until 90% of cells were broken. Cell lysate was centrifuged at differential speeds as follows: the whole cell lysate was centrifuged 500×g for 5min at 4°C to remove unbroken cells and nuclei. Remaining supernatant (S0.5) was centrifuged at 15,000×g for 10min at 4°C to pellet large membranes including Golgi (P15). S15 was ultracentrifuged at 100,000×g for 1hour at 4°C in a Beckman Optima ultracentrifuge using TLA 55 rotor to pellet remaining membranes including vesicles. Recovered cell pellets were washed with homogenization buffer and resuspended in 0.5mL of lysis buffer (50mM Tris, 150mM NaCl, pH 7.4) containing 1% Triton and 5μL/mL Halt protease inhibitor cocktail (Thermo Scientific), and 1mM PMSF and incubated on ice for 30min. Cell lysates were cleared by centrifugation at 20,000xg and recovered supernatants used for subsequent analysis.

Immunofluorescence microscopy

Cells were grown on 12mm glass coverslips (#1, 0.17mm thickness) one day before transfection. After transfection cells were fixed and stained as described previously⁹. In short, cells were fixed in 4% paraformaldehyde (16% stock solution; Electron Microscopy Sciences). Cells were then treated with 0.1% Triton X-100 for one minute. After that the cells were treated with 50mM ammonium chloride for 5min. Cells were washed with DPBS and blocked twice for 10min with 1% BSA, 0.1% saponin in DPBS. Cells then were incubated for 1hour with primary antibody diluted in antibody buffer (1% cold fish gelatin, 0.1% saponin in DPBS) at room temperature. Cells were washed four times with DPBS and incubated for 30min with fluorescently tagged secondary antibody in antibody buffer at room temperature. After that, coverslips were washed four times with DPBS, rinsed with ddH₂O, and mounted on glass microscope slides using Prolong^® Gold antifade reagent (Life Technologies). Cells were imaged with the 63X oil 1.4 numerical aperture (NA) objective of a LSM510 Zeiss Laser inverted microscope outfitted with confocal optics and four lines of excitation lasers (405, 488, 561, and 642nm) under the control of Zen 2009 software (Zeiss). The pixel size was 93nm. The “RGB profiler” plug-in of ImageJ (http://rsbweb.nih.gov/ij) was used to generate line plots for individual channels. The “Coloc_2” plug-in was utilized to calculate Pearson coefficient for co-localization. At least three independent experiments were performed to calculate both mean and standard deviation values.

Live Cell Microscopy

HeLa cells that stably express YFP-COG3 and CFP-COG6 were grown on glass-bottom culture plates. Live cells were imaged using a laser-scanning confocal microscope (LSM 510 Zeiss Laser inverted microscope outfitted with confocal optics) equipped with an environmental control system (Live cell system, BioVision Technologies) set to 37°C and 5% CO₂. Cells were cultured in phenol red–free DMEM/F12 (Sigma-Aldrich) with 1% FBS. Brefeldin A (BFA) (final concentration of 0.25μg/mL) in 10% FBS in DMEM/F-12 medium without antibiotics was added to cells in glass-bottom culture plates. Cells were imaged after 15min incubation.

Super-resolution Microscopy

Cells grown on coverslips were fixed and labeled with affinity-purified antibodies to COG3 and COG8 as detailed above. Cells co-expressing fluorescent protein tagged COG subunits were initially transfected with siRNAs to corresponding subunit. 48h later cells were co-transfected with DNA plasmids expressing siRNA-resistant tagged COG subunits. 24h later cells were fixed and analyzed. 3D-SIM samples were labeled with mouse monoclonal anti-COG3 and rabbit polyclonal affinity purified anti-COG8 primary and anti-mouse Cy3b and anti-rabbit-Alexa488 secondary antibodies. STED samples were labeled with mouse monoclonal anti-COG3 and rabbit polyclonal affinity purified anti-COG8 primary and anti-mouse ATTO488 and anti-rabbit-Alexa647 secondary antibodies.

The superresolution 3D-SIM images were acquired using an Elyra PS.1 (Zeiss, Jena, Germany) equipped with an oil Plan-Apochromat objective lens (100×, NA 1.46), EMCCD camera, and four lines of excitation lasers (405, 488, 561, and 642nm) under the control of Zen 2012 software (Zeiss). The chromatic shift aberration of the system was calibrated and corrected by TetraSpeck beads (110nm in diameter; Invitrogen). The pixel size was 40nm. The z-step of image stacks was 80nm. The intensity line profile of the Golgi region imaged by 3D-SIM and Pearson’s coefficient of colocalization was generated using ImageJ (National Institutes of Health, Bethesda, MD). Pearson’s coefficient was calculated using individual optical slices from the entire Golgi area of 10–18 cells.

Two-color STED images were acquired on the Leica TCS SP8 STED 3x microscope essentially as in Pellett et al.⁵¹. Two-color STED time series were obtained sequentially line by line by alternating the excitation lasers and the STED pulsed and gated lasers tuned to 775 and 592nm for both Alexa 647 and ATTO488. A 100x/1.4 NA oil immersion objective lens (HC PL APO SC2, Leica Microsystems) was used for imaging. The pixel size was 20nm.

Transmission Electron Microscopy

Cells grown on collagen-coated sapphire discs were high pressure frozen using a Leica EM PACT2 high pressure freezer with rapid transfer system. Samples were transferred under liquid nitrogen to cryovials containing 2% OsO₄/0.1% glutaraldehyde/1%H₂O in acetone. Samples were freeze substituted in a Leica EM AFS2 freeze substitution and low temperature embedding system under the following schedule: −90°C for 40h, warm 3°C/h to −60°C, −60°C for 8h, warm 3°C/h to −30°C, −30°C for 8h, warm 3°C/h to 0°C. Following freeze substitution, samples were rinsed three times (10 minutes each) in acetone followed by tannic acid at 4°C (1% tannic acid with 1% H2O in acetone) for 1h. Samples were rinsed three times in acetone followed by 1h osmium wash (1% OsO₄/1% H₂O solution in acetone) at 4°C. Samples were rinsed three times in acetone and embedded in Araldite 502/Embed 812 resin (EMS).

Ultrathin sections were imaged at 80kV on a FEI Technai G2 TF20 transmission electron microscope and images were acquired with a FEI Eagle 4kX USB Digital Camera

GFP binding protein (GBP) Co-immunoprecipitation

DNA encoding GFP binding protein (GBP)⁵² (synthesized by Genscript) was inserted into pET24B vector. Recombinant His6x-tagged GBP was purified on Talon resin (Clontech), and then conjugated to high density glyoxal agarose beads (Agarose Bead Technologies). GBP beads were used for IP similar to agarose protein G beads. Cells were collected by trypsinization (HeLa) or by gentle resuspension (HEK293T), and lysed in 50mM Tris, pH 7.4, 150mM NaCl; 1% Triton X-100 supplemented with Halt protease inhibitor mixture (Thermo Scientific) and 1mM PMSF for 1hour on ice.

Cell lysate was spun at 20,000×g and supernatant was added to 20–25μL of 50% GBP-bead suspension. Bead-lysate mix was incubated for 2h at room temperature while rotating. Unbound material was removed and the beads were washed three times in 0.05% Triton X-100 in PBS and eluted in 2X sample buffer to a total volume of 50μL.

Galanthus nivalis lectin labeling

Cells grown on 12mm glass coverslips (#1, 0.17mm thickness) were washed 3 times with DPBS and then fixed for 15min with 1% PFA (16% aqueous stock solution; Electron Microscopy Sciences). After fixation, cells were incubated with 1% BSA in DPBS for 10min. Cells were then incubated with Galanthus Nivalus lectin (GNL) labeled with AlexaFluor 647 (Invitrogen) diluted 1:500 in 0.1% BSA in PBS for 30min. Cells were washed 5 times with DPBS and then fixed again with 4% PFA for 15min. After second fixation cells were either mounted on glass microscope slides using Prolong^® Gold antifade reagent (Life Technologies) or were stained with antibodies following the IF protocol detailed above.

Flow analysis using Galanthus nivalis lectin

COG1 KO, COG8 KO and COG1/COG8 double KO cells were grown in 3 wells of a 6 well plate along with wild-type HEK293T cells. 1day after plating cells were transfected with Sialytransferase-RFP (a gift from J. Rothman lab, Yale University) and pCOG1-HA-GS-COG8-3Myc and/or hTEV in pLIK (a gift from the Puhl Lab, NIH) using Lipofectamine 2000 (Life Technologies). 48–72hours after transfection cells were washed with DPBS, then resuspended in 0.1% BSA in DPBS then moved to a tube and placed on ice. After centrifugation at 800xg for 3minutes supernatant was removed and cells were resuspended in 0.1% BSA in DPBS with Galanthus Nivalus lectin (GNL) labeled with AlexaFluor 647 (Life Technologies) diluted 1:500. After incubation cells were placed on ice in the dark for 30minutes and resuspended in 0.1% BSA in DPBS with DAPI to distinguish live/dead cell populations. Cells were then analyzed using the LSRFortessa Flow Cytometer (BD Biosciences) or Attune NxT Acoustic Focusing cytometer (Life Technologies). Cells were gated based on shape, singlets, and DAPI signal. Transfected cells were also gated using ST-RFP signal to track transfection. Processing of data was done using NovoExpress software (ACEA Biosciences). The same protocol was followed for the COG1-GFP-GS-COG8-3myc construct, with the exception of gating for GFP instead of co-transfected ST-RFP.

Isolation of whole COG and lobe B complex from HEK cells

HEK293T cells grown in five 10-cm dishes were transfected with MultiLabel vectors with Lipofectamine 2000 reagent (Life Technologies) as previously described⁶. (lobe B expression: GFP-hCOG5+hCOG6/8-3Myc+hCOG7-3Myc; whole COG expression: GFP-mCOG3+hCOG1/2/4-3Myc+hCOG6/8-3Myc+hCOG5/7-3Myc). Twenty four hours after transfection, cells were treated with 5mM sodium butyrate. Cells were harvested 40h after COG transfection and incubated in hypotonic lysis buffer (20mM HEPES ,pH 7.5) supplemented with 1mM PMSF and HALT protease inhibitor cocktail (Thermo Scientific) on ice for 10min. Cells were further disrupted by passing through a 24-gauge needle 20 times. Then cell lysate was supplemented with 150mM potassium acetate. After removing unbroken cells and nuclei by centrifugation at 2000xg for 10min at 4°C, 1% Octyl β-D-glucopyranoside was added into the supernatant which was centrifuged again at 20,000xg for 10min at 4°C.

The final cell supernatant (1ml) was loaded onto 4ml Sephacryl S-300 HR gel filtration (GF) column which was pre-equilibrated with GF buffer (20mM HEPES, pH 7.5, 150mM potassium acetate, 10% glycerol) supplemented with 0.1% Octyl β-D-glucopyranoside to remove small protein contaminants. Fractions enriched with COG complex were collected and kept on ice.

Recombinant His6x-tagged GBP was purified with TALON resin (Clontech) following a standard protocol. Fractions with highest concentrations of GBP were pooled together and extensively dialyzed against the dialysis buffer (25mM HEPES-KOH, pH 7.4, 25mM KCl, 2.5mM magnesium acetate). TALON beads were incubated with purified recombinant GBP for 30min at room temperature with rotating. Unbound proteins were removed by washing. Then Sephacryl S-300 HR GF fractions containing COG complex or lobe B were incubated with TALON-GBP beads for 30min at room temperature. TALON-GBP beads were washed with GF buffer supplemented with 10mM imidazole and then eluted with GF buffer supplemented with 350mM imidazole. Fractions with COG complex were extensively dialyzed against the GBP dialysis buffer supplemented with 10% glycerol and 1mM DTT at 4°C overnight by using Micro Float-a-Lyzer (Spectrum labs, MWCO: 100 kD). After dialysis, proteins were frozen in liquid nitrogen and stored at −80°C.

Preparation of recombinant proteins

cDNAs encoding full-length human GS15 cytoplasmic domain (aa 1–86), cytoplasmic GS15 with deletion of N-terminal domain (Δ1–20), and deletion of SNARE domain (Δ21–86) were subcloned into pET28a (EMD Biosciences) with His6X tag fused at N-terminus and GST tag fused at C-terminus of the expressed protein. GS15 proteins were expressed in E. coli BL21(DE3) cells. pGEX4T-1 (GE Healthcare) was utilized to express recombinant GST protein in E. coli XL-10 Gold (Stratagene). All bacterial cultures were grown at 37°C with shaking until they reached mid-log (OD₆₀₀=0.5–0.6). Expression of proteins was induced by the addition of 0.2mM isopropyl-1-thio-β-D-galactopyranoside (IPTG), and then the culture was grown over night at 22°C. The cells were collected by centrifugation at 6000g for 30min at 4°C and lysed in lysis buffer (50mM Tris, 0.3M NaCl, pH 8.0, 40μg/ml lysozyme, 1mM PMSF). His6X-tag conjugated GS15 proteins were purified by TALON metal affinity resin (Clontech) following a standard protocol. GST was purified by glutathione Sepharose 4B beads following a standard protocol. Fractions with highest concentrations of proteins were pooled together and extensively dialyzed against the buffer (25mM HEPES, pH 7.4, 150mM KCl, 1mM DTT, 0.05% Tween-20). After dialysis, proteins were frozen in liquid nitrogen and stored at −80°C.

COG-GS15 binding assay

Glutathione Sepharose 4B beads were incubated with purified recombinant GST, full-length (FL) of GS15 cytoplasmic domain (aa 1–86), cytoplasmic GS15 with deletion of N-terminal domain (Δ1–20), and deletion of SNARE domain (Δ21–86) in the buffer (20mM HEPES, pH 7.4, 150mM KCl, 1mM DTT, 0.05% Tween-20) for 2hours at 4°C with rotating. Unbound proteins were removed by extensive washing. Then whole COG complex or lobe B isolated from HEK cells were incubated with beads in the buffer (25mM HEPES-KOH, pH 7.4, 25mM KCl, 2.5mM magnesium acetate, 1mM DTT, 0.05% Tween-20) for 1hour at room temperature with rotating. After washing, the remaining beads-bound material was eluted with 2X SDS-PAGE sample buffer with heating for 10min at 70°C and resolved with SDS–PAGE and WB.

SDS–PAGE and Western blotting (WB)

SDS–PAGE and WB were performed as described previously¹². Blots were incubated with primary rabbit polyclonal antibodies and mouse monoclonal antibodies, followed by incubation with appropriate secondary antibodies conjugated with IRDye 680 or IRDye 800 dyes (LI-COR). Blots were scanned and analyzed with an Odyssey IR imaging system (LI-COR).

Multiple COG expression assay for co-IP

HEK293T cells were transfected with multi-expression vectors as previously described⁶ for 36h using Lipofectamine 2000 (Life Technologies) reagent. (lobe A expression: hCOG1/2/4-3Myc+mCOG3-3Myc; lobe B expression: hCOG6/8-3Myc+hCOG5/7-3Myc; lobe A+lobe B; hCOG1/2/4-3Myc+mCOG3-3Myc+hCOG6/8-3Myc+hCOG5/7-3Myc). After 18h of COG-Myc expression cells were transfected again with GFP expression plasmids using Lipofectamine 2000 (Life Technologies) reagent. Cells were harvested 36h after initial COG transfection and processed for IP.

We are thankful to M. Davidson, F. Hughson, A. Linstedt, J. Rothman, E. Sztul, D. Ungar and others who provided reagents and critical reading of the manuscript. We also would like to thank Richard Smith for technical support. This work was supported by the National Institute of Health (R01GM083144, U54 GM105814).