SEARCH

SEARCH BY CITATION

Keywords:

  • COG;
  • glycosylation;
  • Golgi;
  • retrograde traffic;
  • tethering

Abstract

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Supplementary Material
  7. Acknowledgments
  8. References
  9. Supporting Information

Defects in conserved oligomeric Golgi (COG) complex result in multiple deficiencies in protein glycosylation. On the other hand, acute knock-down (KD) of Cog3p (COG3 KD) causes accumulation of intra-Golgi COG complex-dependent (CCD) vesicles. Here, we analyzed cellular phenotypes at different stages of COG3 KD to uncover the molecular link between COG function and glycosylation disorders. For the first time, we demonstrated that medial-Golgi enzymes are transiently relocated into CCD vesicles in COG3 KD cells. As a result, Golgi modifications of both plasma membrane (CD44) and lysosomal (Lamp2) glycoproteins are distorted. Localization of these proteins is not altered, indicating that the COG complex is not required for anterograde trafficking and accurate sorting. COG7 KD and double COG3/COG7 KD caused similar defects with respect to both Golgi traffic and glycosylation, suggesting that the entire COG complex orchestrates recycling of medial-Golgi-resident proteins. COG complex-dependent docking of isolated CCD vesicles was reconstituted in vitro, supporting their role as functional trafficking intermediates. Altogether, the data suggest that constantly cycling medial-Golgi enzymes are transported from distal compartments in CCD vesicles. Dysfunction of COG complex leads to separation of glycosyltransferases from anterograde cargo molecules passing along secretory pathway, thus affecting normal protein glycosylation.

The Golgi apparatus is a hub for membrane-trafficking pathways, organizing both anterograde and retrograde trafficking of molecules (1). It plays a key role in the intracellular trafficking, processing and secretion of glycoproteins, glycolipids and proteoglycans (2). Sequential modifications of glycoproteins by Golgi enzymes depend on the non-uniform distribution of different glycosylation enzymes within Golgi stack (3–5).

Anterograde vesicular transport and cisternal maturation are two alternative models of intra-Golgi transport (6). A fundamental difference between these two models is the differential movement of resident and cargo proteins. The anterograde vesicular transport model predicts that cargo molecules will move forward in transport vesicles, while resident proteins are specifically retained. The cisternal maturation model predicts that cargo molecules will move through the stack passively as the cisternae move forward, while resident proteins will be recycled by retrograde transport to establish differential concentrations across the stack. These two models are not mutually exclusive and may occur simultaneously (7). Inherent in the cisternal maturation model is the importance of recycling in localization of glycosyltransferases, SNAREs and other resident Golgi proteins.

Studies in yeast and mammalian cells have led to the identification of several multisubunit protein complexes that are thought to be involved in Golgi vesicle tethering and/or compartment function, including the conserved oligomeric Golgi (COG) (8,9), the TRAPP, (10) and the GARP (Vps51–54) (11) complexes (12). COG complex is preferentially localized to the cis/medial cisternae of the Golgi apparatus (13–15) and is involved in Golgi membrane traffic (16–20). Conserved oligomeric Golgi complex is a peripheral membrane hetero-oligomer which consists of eight subunits named COG1–8 (15–18,21–23). On the basis of yeast and mammalian genetic and biochemical studies (22–25) and on the results of the electron microscopy (15), COG subunits have been grouped into two lobes consisting of COG1–4 (essential Lobe A) and COG5–8 (non-essential Lobe B). Mutations in the subunits of the COG complex severely distress Golgi glycosylation machinery (17,20–22,26). In ΔCOG1 and ΔCOG2 CHO mutant cells processing of glycoproteins and glycolipids is defective and heterogeneous, resulting in substantial global alterations in cell surface glycoconjugates (27). A recently found mutation in human COG7 gene leads to the type II congenital disorder of glycosylation (CDG) (26). The heterogeneity of protein glycosylation defects suggests that mutations in COG complex affect the activity or compartmentalization of multiple Golgi enzymes without sizeable disruption of secretion and endocytosis. The activity of glycosylation enzymes depends on their proper intra-Golgi localization (28,29). Thus, COG may play a direct role in transport, retention and/or retrieval of components of Golgi-glycosylation machinery. Studies in yeast have identified a large number of COG-interacting genes encoding proteins implicated in Golgi trafficking (8,9,17). We have shown that COG complex directly interacts with Rab GTPase Ypt1p, intra-Golgi SNAREs, as well as with the COPI coat complex. In addition, electron microscopy revealed that cog2 and cog3 temperature-sensitive yeast mutants accumulate vesicles (30).

It has been recently demonstrated that the acute knock-down (KD) of the COG3 was accompanied by reduction in Cog1, Cog 2 and Cog 4 protein levels and resulted in the accumulation of COG complex-dependent (CCD) vesicles carrying Golgi v-SNARE molecules (19). Prolonged block in CCD vesicle tethering is accompanied by substantial fragmentation of the Golgi ribbon. Fragmented Golgi membranes maintain their juxtanuclear localization, cisternal organization and competence for anterograde protein trafficking to the plasma membrane. These findings let us hypothesize that COG complex acts as a tether which connects COPI vesicles with cis-Golgi membranes during retrograde intra-Golgi traffic. Additional evidence that COG plays a role in the retrograde vesicular transport of Golgi proteins, including glycosylation enzymes, came from surveying the steady-state levels of Golgi proteins in wild-type and COG-deficient mammalian cells (31). Seven Golgi membrane proteins, including processing enzyme α-1,3-1,6-mannosidase II (Mann II), were found to exhibit reduced steady state levels in both ΔCOG1 and ΔCOG2 CHO cells.

How can the COG complex determine localization of enzymes within the Golgi? One way to achieve this is to interact directly with cytoplasmic tails of resident Golgi proteins and retain them in an appropriate compartment. Another way is to control the basic structure of the Golgi or its lumenal environment (pH, ion concentrations) (21). The most likely hypothesis is that the COG complex is directly involved in tethering of intra-Golgi COPI vesicles which transport recycling enzymes to appropriate compartments. To test this hypothesis and determine molecular mechanism by which malfunctioning of the COG complex may generate defects in Golgi-glycosylation machinery, we investigated cellular phenotypes after both acute and prolonged COG3 and COG7 KD. We found that progression of COG3 KD was positively correlated with Golgi-glycosylation defects, and glycosylation of both lysosomal and plasma membrane proteins was severely altered after prolonged COG3 KD. Underglycosylated proteins were correctly delivered to their cellular locations, indicating that even a prolonged block in Cog3p function does not affect anterograde trafficking and protein sorting. We also discovered that as early as 3 days after COG3 KD medial-Golgi enzyme, β-1,2-N-acetylglucosaminyltransferase-1 (GlcNAcT1) becomes mislocalized into vesicular fraction. GlcNAcT1 containing vesicles were distinct from ER, endosomes and lysosomes and most likely identical to previously described CCD vesicles enriched with Golgi v-SNARE proteins (19). Similar relocation into CCD vesicles was found for another medial-Golgi enzyme, Mann II and to a lesser extent medial/trans-Golgi enzyme N-acetylgalactosaminyltransferase 2 (GalNAcT2) and trans-Golgi enzyme β-1,4galactosyltransferase (GalT). Prolonged COG3 KD finally leads to degradation of both GlcNAcT1 and Mann II. Similar effects were observed in COG7 KD and COG3/COG7 double KD HeLa cells.

On the basis of our findings, we conclude that alterations in the function of COG complex caused specific mislocalization of medial-Golgi enzymes and Golgi v-SNAREs into recycling intra-Golgi CCD vesicles. Efficient and COG-dependent docking of CCD vesicles to purified Golgi membranes was reconstituted in vitro, supporting functional status of these transport intermediates. Conserved oligomeric Golgi complex most likely functions in tethering of retrograde enzyme-carrying CCD vesicles to the proper Golgi compartment.

Results

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Supplementary Material
  7. Acknowledgments
  8. References
  9. Supporting Information

Prolonged COG3 knock-down leads to defects in protein glycosylation

Our previous studies demonstrated that acute depletion of Cog3p caused accumulation of non-tethered CCD vesicles and Golgi fragmentation but, at least initially, did not result in increased gel mobility of Golgi glycoprotein GPP130, arguing that protein glycosylation was not altered (19). This result was rather surprising, since severe Golgi-glycosylation defects were previously observed in both ΔCOG1 and ΔCOG2 CHO mutant cells (21), and similar glycosylation abnormalities were detected in yeast cog3-ts (sec34–2) mutant (17). We hypothesized that the observed phenotype of the acute COG3 KD might represent a primary Cog3p-dependent defect in membrane trafficking, leading to deficient Golgi glycosylation. To test out this hypothesis, we investigated whether the extended deficiency in Cog3p function would affect Golgi-glycosylation machinery (Figure 1A). First, we determined the steady-state glycosylation status of two endogenous membrane glycoproteins – plasma membrane-localized CD44 (32) and extensively glycosylated with 16 N-linked carbohydrate chains lysosomal resident protein Lamp2 (33). HeLa cells were treated with COG3 siRNA for 3, 6 and 9 days. Acute KD is referred to as 3 days after siRNA treatment; prolonged KD is 6 and 9 days after siRNA treatment. Cog3 protein has been efficiently depleted, and after 3 days of KD, its level comprised 10% of the initial one (Figure 1A, right panel, upper lane). We have previously shown that acute COG3 KD affects the stability of Lobe A but not Lobe B subunits of the COG complex (19). Similarly, there was destabilization of Cog4p and most probably whole Lobe A after prolonged COG3 KD. Cellular levels of Lobe B subunits were also slightly decreased (Figure S1, available online at http://www.blackwell-synergy.com). Prolonged COG3 KD or similar transfection with control (scrambled) siRNA did not affect viability of HeLa cells.

image

Figure 1. Prolonged COG3 knock down(KD) affects glycosylation and stability of membrane glycoproteins of secretory pathway.A) Steady-state cellular level and gel mobility of glycoproteins of secretory pathway. HeLa cells were mock (0) or COG3 siRNA treated for 3, 6 and 9 days. Total cell lysates were prepared at corresponding time points, subjected to SDS-PAGE and WB with the indicated antibodies. Right panel represents quantification of immunoblotted proteins. B) Increase in gel mobility of CD44 after COG3 KD is due to its defective glycoprotein modification. Cell lysates after 3, 6 and 9 days of COG3 KD were incubated with a PNGase F (500 U/reaction) as described in Materials and Methods. Samples were loaded on SDS-PAGE and WB with anti-CD44 antibody. C) Pulse-chase labeling of CD44 and Lamp2 glycoproteins. Cells were metabolically labeled with 35S-Methionine for 10 min and then chased for indicated time points. Cell lysates were subjected to IP with anti-CD44 and anti-Lamp2 antibodies. Precipitated proteins were loaded on SDS-PAGE and visualized by PhosphorImager. Star indicates a non-specific band.

In agreement with our previous observations, after 3 days of COG3 KD, all tested glycoproteins, i.e. GPP130, Lamp2 and CD44, did not show a visible change in their mobility on SDS-PAGE (Figure 1A, compare control (0) and 3 days lanes). In contrast, after 6 days of COG3 KD, the gel mobility of both Lamp2 and CD44 was altered, indicating the production of underglycosylated protein species. The glycosylation defects became more pronounced after 9 days of COG3 KD. Western blot analysis demonstrated that a decrease in molecular weight of both CD44 and Lamp2 positively correlated with the duration of COG3 KD (Figure 1A, left panel, compare molecular weights of CD44 and Lamp2 after 3 and 9 days of COG3 KD). Smearing of the bands corresponding to CD44 and Lamp2 may argue for heterogeneity in glycosylation, emerging after 6 and 9 days of COG3 KD. Interestingly, the bulk of the GPP130 did not change its gel mobility even after 9 days of COG3 KD. Instead, the level of GPP130 protein was dramatically decreased after prolonged COG3 KD (Figure 1A, right panel). A similar expression profile was observed for Lamp2. After 9 days of COG3 KD, the protein level of Lamp2 was reduced by more than 50% (Figure 1A, right panel). The latter result might be caused by the instability of underglycosylated Lamp2 protein (34).

To confirm our assumption that the increase in gel mobility is caused by defective glycosylation, both control and COG3 KD cell lysates were subjected to peptide:N-glycosidase F (PNGase F) treatment (Figure 1B). We found that deglycosylation of all forms of CD44 observed in COG3 KD cells produced a single 62 kDa polypeptide. A similar result was obtained for Lamp2, in which the molecular weight was reduced from approximately 110 to 82 kDa after PNGase F treatment (data not shown). This result confirmed our hypothesis that increased gel mobility of CD44 and Lamp2 after COG complex KD is entirely due to deficient glycosylation. After prolonged COG3 KD, underglycosylated glycoproteins were sensitive to Endo H treatment (Figure S2, available online at http://www.blackwell-synergy.com), indicating defects in early Golgi glycosylation.

To determine whether COG3 KD affects the kinetics of glycoprotein modifications in the Golgi, we performed a pulse-chase experiment (Figure 1C). Both the extent of glycosylation and its kinetics were almost indistinguishable for control and 3 days COG3 KD cells, with only minute smearing of CD44 band observed at the 120 min time point (Figure 1C, panel a). This smearing most likely represents the early onset of glycosylation defects. In contrast, a pronounced defect in glycosylation of CD44 and Lamp2 (Figure 1C, 9 KD lanes) was detected after 9 days of COG3 KD. Even after 2 h of chase, both CD44 and Lamp2 remained underglycosylated. Prolonged (9 days) growth on plates and/or continuous treatments with transfection reagent slowed down the kinetics of CD44 glycosylation in both COG3 siRNA and mock-transfected cells. Interestingly, in both plates, 120 min was a sufficient time to chase the bulk of CD44 to the Golgi glycosylated form, indicating that protein movement through the Golgi in COG3 KD cells was not sufficiently altered. We have previously observed a similar slight delay in plasma-membrane delivery of VSVG in COG3 KD cells (19).

The behavior of Lamp2 was more complex, with the accumulation of a whole array of partially glycosylated forms accumulated in COG3 KD cells after 120 min of chase. These findings allowed us to hypothesize that after prolonged COG3 KD, either glycoproteins are not targeted properly and thus can not encounter the Golgi glycosylation machinery or Golgi glycosylation machinery itself is mislocalized, thus not allowing Golgi enzymes to process proteins.

Underglycosylated glycoproteins are correctly localized in COG3 KD cells

To address the first question of possible impairment of anterograde trafficking of glycoproteins of the secretory pathway under conditions of COG3 KD, we determined localization of CD44 and Lamp2 after prolonged COG3 KD. As shown in Figure 2(A), after 9 days of COG3 KD, CD44 was mostly localized on the plasma membrane. A small portion of CD44 signal was detected on intracellular membranes that were distinct from GPP130-positive membranes and most likely represented recycling pool of CD44. Strikingly, staining of non-permeabilized cells demonstrated that the intensity of CD44 signal was similar for both control (100%) and 9 days of COG3 KD (94%) cells, indicating that anterograde trafficking and plasma membrane expression of CD44 was not compromised (Figure 2B). There was no staining for the lumenal Golgi protein GlcNAcT1 in non-permeabilized cells (Figure S3 available online at http://www.blackwell-synergy.com).

image

Figure 2. Underglycosylated CD44 and Lamp2 are correctly localized in control and COG3 knock down(KD) cells.A) Localization of CD44 is similar in control and COG3 KD cells. Control and 9 days COG3 KD cells were fixed, permeabilized and triple stained with anti-CD44 (red), GPP130 (green) and DAPI (blue). Arrows indicate plasma membrane. B) Plasma membrane delivery of CD44 is not compromised in COG3 KD cells. Control and 9 days COG3 KD cells were fixed and stained with anti-CD44 (green) without permeabilization. C) Localization of Lamp2 is similar in control and COG3 KD cells. Control and 9 days COG3 KD cells were fixed and triple stained with anti-Lamp2 (red), GPP130 (green) and DAPI (blue). D) Lamp2 is colocalized with lysosomes in control and COG3 KD cells. Control and 9 days of COG3 KD cells were treated with Cascade Blue Dextran as described in Materials and Methods and stained with anti-Lamp2 (green). Cascade Blue Dextran is visualized at 420 nm spectrum (blue). Arrows indicate lysosomes. All images were collected at equal signal gains using CARV microscopy. Bar, 10 µm.

Similarly, proper localization was found for the lysosomal protein Lamp2 (Figure 2C). Nine days after COG3 KD, Lamp2 was associated with scattered membrane compartments distinct from endosomes, ER and Golgi remnants (data not shown). To obtain additional evidence that the Lamp2-positive compartments are indeed lysosomal in nature, cells were treated with Cascade Blue Dextran, which is known to travel via the endocytic pathway finally accumulating in lysosomes (Figure 2D). Immunofluorescent assays demonstrated that in both control (80 ± 3% of colocalization) and 9 days (69 ± 4% of colocalization) of COG3 KD cells, endocytosed dextran was colocalized with Lamp2, indicating that underglycosylated Lamp2 was sorted properly to lysosomes in COG3-depleted cells.

Medial-Golgi enzymes are severely mislocalized in COG3 KD cells

Results from the experiments described above allowed us to conclude that in Cog3p deprived cells, underglycosylated proteins are delivered properly to both the plasma membrane and lysosomes. Most likely, on their way to final destination, they are transported through the fragmented Golgi. Thus, the anterograde protein flow is undistorted, and the trafficking itinerary is not modified in COG KD cells. Why do glycoproteins become underglycosylated? We hypothesized that Golgi-glycosylation machinery itself is mislocalized in COG3 KD cells. Both IF and subcellular fractionation were employed to determine localization of key components of the Golgi-glycosylation machinery. We have specifically focused our investigation on GlcNAcT1 localized at steady-state within the medial-Golgi cisternae (5). It has been shown that newly synthesized GlcNAcT1 is transported rapidly through the Golgi stack to the trans-Golgi network and then is recycled back to the cis-Golgi with a half-time of about 150 min, suggesting that this protein is continuously recycled from the late Golgi (29,35). Furthermore, GlcNAcT1 was detected in COPI-dependent transport intermediates which fused with the cis-cisternae in the in vitro assays (36,37). To account for any non-specific staining with primary or secondary antibodies in IF experiments, we used a mixed population (1:1) of HeLa cells stably expressing Golgi glycosyltransferases tagged either with myc or vsv epitope tag. It has been previously shown that in these cells, tagged enzymes were only slightly (two- to fourfold) overexpressed, and their localization and trafficking was indistinguishable from endogenous enzymes (5). We found that GlcNAcT1 is tightly associated with the Golgi ribbon in mock-transfected cells treated with scrambled siRNA and is rapidly redistributed into vesicular structures after COG3 KD (Figure 3A). A subpopulation of the enzyme colocalized with GM130 and most likely remained associated with fragmented Golgi mini-stacks. A similar vesicular distribution of GlcNAcT1 was observed in cells after 3 and 6 days of COG3 KD (Figure 3A, 3 and 6 days rows). More reticular GlcNAcT1-positive structures were found in cells after prolonged Cog3p depletion (Figure 3A, 9 days row). These reticular structures were similar to classical ER appearance, and indeed partial colocalization of Golgi enzyme with ER marker PDI was observed in cells after 9 days of COG3 KD (Figure S4, available online at http://www.blackwell-synergy.com). Localization of GlcNAcT1 in acutely depleted cells was distinct from that of ER, endosomal and lysosomal markers (Figure 3B), and most likely, indicated accumulation of the enzyme in CCD vesicles carrying v-SNAREs GS15 and GS28 and cis-Golgi glycoprotein GPP130 (19). On the glycerol velocity gradient, GPP130 was detected in the same fractions as GlcNAcT1. Indeed, separation of cell lysates on glycerol velocity gradient demonstrated that a significant fraction of GlcNAcT1 was associated with small vesicles (Figure 3C, fractions 3 and 4). CCD vesicles associated fraction of GlcNAcT1 was increased from 5% (control) to more than 50% (3 and 6 days COG3 KD). The vesicular fraction in 9 days COG3 KD cells decreased as compared to acutely depleted cells, arguing that the majority of GlcNAcT1 molecules were now associated with larger membrane structures like Golgi or ER (Figure 3C, lower panel, fraction 13). This biochemical result is in good agreement with the IF data. Both PDI and GM130 were exclusively found in soluble (fraction 1) and Golgi/ER (fraction 13) pools (19; data not shown).

image

Figure 3. COG3 knock down (KD) induces relocation of Golgi enzyme β-1,2-N-acetylglucosaminyltransferase-1(GlcNAcT1).A) Control (0), 3, 6 and 9 days after COG3 KD mixture of cells stably expressing GlcNAcT1-myc and Mann II-vsv were fixed and triple stained with anti-myc (red), Golgi marker anti-GM130 (green) and DAPI (blue). Arrows indicate conserved oligomeric Golgi complex-dependent (CCD) vesicles. Star indicates cell without GlcNAcT1-myc. B) 3 days COG3 KD cells as in (A) were fixed and triple stained with anti-myc (red), ER marker PDI (green), early endosomal marker EEA1 (green), lysosomal marker Lamp2 (green) and DAPI (blue). C) Glycerol velocity gradient fractions were immunoblotted with anti-myc (left panel) and quantified (right panel). Combined signal of fractions 1–13 was taken for 100%. Quantification of each lane represents an average of three independent experiments. Black, red, blue and green lines represent distribution of GlcNAcT1 in control, 3, 6 and 9 days of COG3 KD cells, respectively.

To test whether the localization of other Golgi enzymes is affected by COG3 KD, we used cells stably expressing vsv-tagged medial-Golgi enzyme Mann II and medial/trans-Golgi enzyme GalNAcT2 (5). IF analysis revealed that upon COG3 KD, both Mann II and, to a lesser extent, GalNAcT2 were redistributed into vesicular structures, supporting our hypothesis of their accumulation in CCD vesicles (Figure 4A). Indeed, there was significant increase in Mann II protein in the vesicle fraction isolated on glycerol velocity gradient (Figure 4B). Relative enrichment of Mann II in the vesicle fraction (approximately four- to fivefold) was similar to one observed for GPP130 and GlcNAcT1 (Figure 4B, compare GPP130, GlcNAcT1 and Mann II rows). Vesicular accumulation of GalNAcT2-vsv (approximately twofold) was at the level previously observed for GalNAcT2-GFP and GS28 (19). We concluded that COG3 KD causes specific accumulation of medial-Golgi enzymes in recycling intra-Golgi CCD vesicles. Another trans-Golgi enzyme, GalT, was found mostly on fragmented Golgi (Figure 4A, GalT row). This result is in good agreement with GalT-GFP localization in COG3 KD cells observed previously (19).

image

Figure 4. COG3 knock down(KD) induces relocation and partial degradation of medial-Golgi enzymes.A) Control and 3 days COG3 KD cells stably expressing GlcNAcT1-myc, Mann II-vsv and GalNAcT2-vsv were fixed and stained with anti-myc, anti-vsv, or anti-GalT antibodies. Images were collected using CARV microscopy. Bar, 10 µm. B) Control and COG3 KD vesicle fractions (combined glycerol gradient fractions 3 and 4) were immunoblotted with GPP130, anti-myc for GlcNAcT1, anti-vsv for Mann II, anti-vsv for GalNAcT2 and anti-PDI antibodies as indicated. C) Control, 3 and 9 days COG3 KD cell lysates (10 µg) were immunoblotted with anti-vsv (upper panel), anti-myc (lower panel), anti-human GS28 and anti-PDI antibodies. Star indicates putative degradation product of GlcNAcT1.

Since the severe redistribution of Golgi enzymes was observed after both acute and prolonged COG3 KD, we questioned how it would affect enzyme stability. Western blot analysis of total cellular homogenates revealed that prolonged, but not acute, COG3 KD resulted in a twofold decrease of Mann II cellular level (Figure 4C, Mann II row) and in accumulation of GlcNAcT1-degradation products (Figure 4C, GlcNAcT1 panel). We also noticed a slight increase in total cellular level of GlcNAcT1. Intracellular levels of the control protein PDI was found unchanged, while the level of another CCD vesicle protein GS28 (19) was also reduced by approximately 50% after prolonged COG3 KD (Figure 4C, GS28 row). These data correlate well with previously observed degradation of GS28 (GOS-28) in ΔCOG1 CHO cells (31).

COG7 KD leads to mislocalization of GlcNAcT1 and GS15

To test whether the entire COG complex is required for proper localization of Golgi glycosyltransferases, we used an siRNA strategy to knock-down Cog7p (subunit of COG Lobe B). It has been found recently that mutation in human COG7 gene led to secretion of underglycosylated proteins (26), a mutant phenotype similar to the one observed after prolonged COG3 KD. After 3 days of COG7 KD, the Cog7p was efficiently depleted (Figure 5B), and both GlcNAcT1 and GS15 became severely mislocalized (Figure 5A). This effect was specific since both GM130 (Figure 5A, panel GM130) and GalNAcT2 (Figure 5, panel GalNAcT2) remained associated with large perinuclear Golgi fragments. Transfection with control (scrambled) siRNA did not have any effect on localization of Golgi proteins (Figure 5A, top row). Similarly, to the prolonged COG3 KD, the SDS-PAGE mobility of lysosomal glycoprotein Lamp2 in COG7 KD cells was altered, indicating defects in Golgi glycosylation. In addition to obvious similarities between COG3 KD and COG7 KD mutant phenotypes, we noticed some unique characteristics of COG7 KD cells – glycosylation defects were manifested slightly earlier and GM130/GalNAcT2-containing Golgi membranes were less fragmented and often misshapen into ‘cotton ball-like’ structures. We also detected that the cellular level of Golgi-resident proteins GM130 and Syntaxin 5 was reduced in COG7 KD cells to less than 50% as compared to mock-treated or COG3KD cells. A Golgi-located short 37 kDa form of Syntaxin 5 was affected to a greater extent with only approximately 20% protein remained in COG7 KD cells. We have shown previously that yeast Syntaxin 5 homologue, Sed5p, interacted with COG complex directly (17), and the cellular level of Sed5p was severely reduced in double ΔCOG1COG6 mutant cells (23). Reduced level of GM130 and Syntaxin 5 in COG7 KD cells may reflect direct interaction between Cog7p and/or Lobe B with Golgi membrane-associated components of vesicle docking/fusion machinery. Observed differences between COG3 KD and COG7 KD phenotypes are likely due to diverse specialization of two lobes of the COG complex and will be examined in future studies.

image

Figure 5. COG7 knock down (KD) induces relocation of GlcNAcT1 and GS15.A) Mixture of cells expressing GlcNAcT1-myc and GalNAcT2-vsv were mock (control) or COG7 siRNA treated for 3 days. Fixed cells were stained with anti-myc or anti-vsv antibodies. Images were collected at equal signal gains, using CARV microscopy. Bar, 10 µm. B) Control, COG3, COG3 & COG7 and COG7 KD whole-cell lysates (10 µg protein/lane) were immunoblotted with anti-Cog3, anti-Cog4, anti-Cog6 and anti-Cog7 antibodies. C) Lysates as in (B) were immunoblotted with indicated antibodies, anti-myc for GlcNAcT1; PDI was used as a loading control. Star indicates putative degradation product of GlcNAcT1.

Simultaneous depletion of Cog3p and Cog7p affected glycosylation of Lamp2 and stability of GlcNAcT1 more severely as compared to individual COG subunit KD (Figure 5C) (COG3 & 7 lane), indicating faster progression of medial-Golgi enzyme-trafficking defects in a double Lobe A/Lobe B mutant. This result correlates well with the previously observed severe glycosylation defect in double ΔCOG1COG6 yeast mutant cells (23).

On the basis of the results described above, we concluded that any alterations in COG complex function in human cells cause specific mislocalization of preferentially medial-Golgi enzymes and Golgi v-SNARE molecules and their accumulation in normally transient CCD vesicles. Conserved oligomeric Golgi complex most likely functions in tethering of these transport intermediates to proper Golgi compartment. The cellular biosynthetic machinery is able to ‘ignore’ the initial defects in Golgi protein recycling, but continuous separation of glycosylation machinery from the secretory pathway affects Golgi modifications of secretory, plasma membrane, and lysosomal glycoproteins and could ultimately lead to congenital disorders of glycosylation (26).

CCD vesicles dock to Golgi in vitro in a Cog3p-dependent reaction

Results obtained from this and previous (19) works indicated that both medial-Golgi glycosyltransferases and intra-Golgi SNAREs transiently accumulated in CCD vesicles in COG3 KD cells. To test if these vesicles represent a functional intra-Golgi transport intermediate, we designed an in vitro system which measured vesicle docking/fusion with isolated rat liver Golgi (RLG). The system is similar to the yeast COPII vesicle tethering setup (38) and is based on sedimentation properties of isolated RLG (pelletable at 10 000 × g) and CCD vesicles obtained from COG3 KD HeLa cell lysate (not pelletable at 20 000 × g). Both GlcNAcT1-myc and GPP130 were used as vesicle markers. GS28 was used as Golgi marker. Monoclonal antibodies to ratGS28 and human GPP130 specifically recognized corresponding proteins in RLG and HeLa cell lysates (Figure S7, available online at http://www.blackwell-synergy.com). We have found that CCD vesicles are able to dock to isolated Golgi (Figure 6). The amount of sedimentable vesicle marker (up to 30% from total input) was proportional to the amount of added Golgi membranes and vesicle–Golgi association was resistant to 250 mm KCl wash, which normally strips vesicles from Golgi membranes (39), indicating tight association and/or complete fusion (Figure 6A). Acceptor Golgi membranes were sensitive to proteinase K pretreatment, indicating requirement for peripheral and/or transmembrane proteins. Indeed both COG3 and Golgi SNARE GS28 were completely destroyed by Proteinase K treatment (Figure 6B and data not shown). Most importantly, docking of CCD vesicles was sensitive (approximately 70% inhibition) to addition of anti-COG3 IgG but not to addition of pre-immune (control) IgG (Figure 6B, compare α-COG3 (+) and control IgG columns). We have concluded that CCD vesicles are functional intra-Golgi intermediates capable of docking to Golgi membranes in a COG complex-dependent reaction.

image

Figure 6. Conserved oligomeric Golgi complex-dependent(CCD) vesicles dock to Golgi membranes in vitro.A) Vesicle docking is proportional to amount of added Golgi and resistant to salt wash. Cells stably expressing GlcNacT1-myc were treated with COG3 siRNA. A vesicle fraction from COG3 KD cells (S20) was incubated with purified rat liver Golgi as described in Materials and Methods. After incubation with CCD vesicles, Golgi membranes were pelleted at 10 000 × g and washed with low or high-salt buffer. Golgi membranes were repelleted and analyzed for CCD vesicle marker GlcNAcT1 by WB with anti-myc antibodies. B) Vesicle docking is sensitive to protease treatment of acceptor Golgi membranes and is inhibited by anti-COG3 IgG. CCD vesicle marker GlcNAcT1 (anti-myc Ab), GPP130 (anti-human GPP130) and Golgi marker GS28 (anti-rat GS28 Ab) were detected by WB. Star indicates heavy chain of anti-Cog3p IgG bound to Golgi membrane.

Discussion

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Supplementary Material
  7. Acknowledgments
  8. References
  9. Supporting Information

The COG complex has been assigned a role of one of the key components in intracellular membrane trafficking (40). Majority of up-to-date data argue that the COG complex resides on cis/medial-Golgi and functions as a tether of retrograde intra-Golgi vesicles (15–17,31). Indeed, both yeast (30) and mammalian (19) cells deficient in a Cog3p subunit accumulate multiple CCD vesicles. These vesicles most likely are COPI coated and packed with recycling Golgi SNAREs (GS15, GS28) (41) and putative retrograde cargo receptors as GPP130 (42). Another well-studied feature of COG-deficient cells is significantly impaired modification of glycoconjugates. In both mammalian (ΔCOG1, ΔCOG2 and ΔCOG7) (21,26) and yeast (ΔCOG1, ΔCOG3, ΔCOG4 and ΔCOG6) (17,18,22,23) COG mutant cells, virtually all N- and O-linked Golgi-associated glycosylation reactions are impaired.

In this article, we examined the hypothesis that retrograde intra-Golgi trafficking of components of the glycosylation machinery is directed by the COG complex. We have shown that the duration of COG3 KD positively correlates with development of Golgi-glycosylation defects (Figure 1). Sorting and delivery of anterograde secretory cargo proteins (Figure 2) is not altered. For the first time, we demonstrated that depletion of both Lobe A (Cog3p) and Lobe B (Cog7p) of the COG complex severely affects localization of medial-Golgi enzymes GlcNAcT1 and Mann II, inducing their relocation into CCD vesicles. This finding agrees well with the observed mislocalization of yeast Golgi-glycosylation enzyme Och1p in a cog3 mutant (20).

CCD vesicles are likely to originate from either the trans-Golgi or TGN, since both GlcNAcT1 and Mann II are known to cycle through the Golgi stack to the trans-Golgi network and then back to the cis-Golgi (29,35). It has been shown recently that formation of COPI vesicles is linked to the assembly of the actin complex (43,44). The actin cytoskeleton is affected in yeast cog3 mutant cells, (13) and actin is shown to be coimmunoprecipitated with the mammalian COG complex (45). One attractive idea is that COG complex directs the movement of CCD vesicles along specialized intra-Golgi actin railways through communication with the actin cytoskeleton. In support to this notion, Cog3p (CG3248) was found to interact with the actin-binding protein Arp3 (CG7558) in a recent Drosophila two-hybrid screen (46). We also observed that CCD vesicles are positioned along the actin cables in COG3 KD cells (Figure S6).

Vesicles containing Golgi enzymes are likely to be retrograde in respect to direction of their trafficking since both endogenous (CD44 and Lamp2) (Figure 2) and model (GFP-VSVG) (19) anterograde secretory cargo molecules are not detected in CCD vesicles at the fluorescence microscopy level; after prolonged KD of COG complex, vesicles are partially consumed by the ER, depositing their content into the endoplasmic reticulum (Figure 3C and S4). Latter result is in agreement with previously observed partial relocation in ER of Mann II in ΔCOG1 and ΔCOG2 CHO mutant cells (31). We have previously demonstrated that both GS15 and GS28 are enriched in CCD vesicles (19). Most likely, these v-SNAREs form a functional fusion complex with t-SNARE Syntaxin 5, which itself is a COG-interacting protein (17). One plausible explanation for the eventual consumption of CCD vesicles by ER is based on the observation that the long form of Syntaxin 5 cycles between the Golgi and ER (47). We have found that the ER form of Syntaxin 5 is less sensitive to COG7 KD as compared to the Golgi form of t-SNARE (Figure 5C). Consequently, in COG7 KD cells, partial colocalization of GlcNAcT1 with ER markers was observed as early as 4 days after KD (Figure S5, available online at http://www.blackwell-synergy.com). Therefore, a relative increase in the ER localized form of Syntaxin 5 in COG KD cells could potentiate fusion of CCD vesicles with the ER membrane. Interestingly, in recently described mammalian cells depleted of Syntaxin 5, the COG-sensitive protein GS28 was found in dispersed vesicle structures (48).

The in vitro experiments (Figure 6) support the idea that CCD vesicles are functional intra-Golgi trafficking intermediates. Isolated CCD vesicles are capable of docking and, most probably, fusion with purified Golgi membranes, since as a result of vesicle–Golgi coincubation, vesicle cargo proteins GlcNAcT1 and GPP130 became associated with large membranes and this association is salt-resistant. Vesicle docking is dependent on Golgi peripheral and/or transmembrane proteins, since proteinase treatment of isolated Golgi virtually abolished vesicle docking. The efficient docking of vesicles requires a functional COG complex on the Golgi membrane, since both protease treatment and pretreatment with anti-Cog3p IgGs efficiently block vesicle–Golgi interaction. Latter result supports the model of Golgi-localized COG complex tethering intra-Golgi retrograde vesicles (Figure 7).

image

Figure 7. Model of the function of conserved oligomeric Golgi(COG) complex in trafficking of Golgi enzymes. The COG complex primarily resides on the medial-Golgi. It orchestrates tethering of constantly cycling retrograde COG complex-dependent (CCD) vesicles that bud from trans-Golgi (solid line). These Golgi intermediates carry resident Golgi proteins, including medial-Golgi glycosyltransferases. Lobe B of the COG complex might also associate with trans-Golgi and accept vesicles that retrieve trans-Golgi enzymes from trans-Golgi network and endosomal compartments (dashed line). During malfunction of COG complex, retrograde membrane intermediates accumulate in cytoplasm.

Both GPP130 and GlcNAcT1 containing vesicles behave similarly in the in vitro system, suggesting that the CCD vesicle population is homogeneous and that both proteins are being recycled using the same trafficking intermediates. Conversely, known trafficking itineraries of GPP130 and GlcNAcT1 are different. Medial-Golgi localized GlcNAcT1 cycles via the trans-Golgi and ER (29), while cis-Golgi-resident GPP130 visits the endosomal system and plasma membrane (49). Prolonged COG3 and COG7 depletion ultimately leads to partial relocalization of the GlcNAcT1 signal to the ER, while GPP130 is not detected in the ER (unpublished data). There was a reduction of GPP130 (but not GlcNAcT1) protein levels after prolonged COG KD (compare Figure 1A and Figure 4C). Additional biochemical and immuno-EM studies will be required to characterize CCD vesicles in details.

Comparison of COG3 and COG7 KD phenotypes indicates that depletion of either Lobe A or Lobe B subcomplexes primarily affects localization of GlcNAcT1 and GS15, indicating that the whole COG complex is required for proper recycling of cis/medium-Golgi-resident proteins. On the other hand, acute COG7 KD phenotype differs from COG3 KD. Glycosylation defects are manifested slightly earlier in COG7 KD cells (Figure 5C and data not shown), and Golgi membranes in those cells are often misshapen into cotton ball or sponge-like structures. In addition, acute double COG3/COG7 KD affected glycosylation of both Lamp2 and GlcNAcT1 more severely as compared to single depletion, indicating that double KD is affecting activity or compartmentalization of multiple Golgi enzymes. There is evidence that loss-of-function mutation in Lobe A subunits causes defects in early Golgi-glycosylation reactions (17,20,21), whereas a loss-of-function mutation in Lobe B subunits causes trans-Golgi-glycosylation defects (26). Taking into account that in mammalian cells Lobe B exists both as a part of the large COG complex and as a separate small subcomplex (15), we propose the model (Figure 7). Our model suggests that the whole Lobe A/Lobe B COG complex regulates efficient vesicle tethering to the cis/medial-Golgi membranes. The lobe B subcomplex could also specifically tether vesicles to trans-Golgi cisternae. Similar separation of functions was discovered recently for cis- and trans-Golgi-operating anterograde tethers TRAPP I and II (10). Proposed compartment-specific functions for COG lobes may reflect a common principle in the evolution of oligomeric complexes operating in membrane trafficking.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Supplementary Material
  7. Acknowledgments
  8. References
  9. Supporting Information

Reagents and antibodies

Reagents were as follows: PNGase F, EndoH (New England Biolabs, Beverly, MA, USA); Cascade Blue Dextran (Molecular Probes, Eugene, OR, USA); Protein G-Sepharose (Calbiochem, La Jolla, CA, USA); Protein A-Sepharose (Amersham Biosciences, Piscataway, NJ, USA). Antibodies used for Western Blotting (WB), immunofluorescence (IF), immunoprecipitation (IP) studies were obtained from commercial sources and as gifts from generous individual investigators or generated by us as indicated below. Antibodies (and their dilutions) were as follows: rabbit – both anti-myc (WB 1:5000, IF 1:3000) and anti-vsv (WB 1:2500, IF 1:400) from Bethyl Laboratories (Montgomery, TX, USA); anti-Cog3p (WB 1:1000) (16); anti-GPP130 (WB 1:1000, IF 1:2000; Covance Laboratories, Madison, WI, USA); anti-CD44 (WB 1:400; Santa Cruz Biotechnology, Santa Cruz, CA, USA); murine: anti-GM130 (IF 1:250); anti-human GS28 (WB 1:1000, IF 1:100) and anti-EEA1 (IF 1:250) (all from BD Biosciences, San Jose, CA, USA); anti-GPP130 (WB 1:100, gift from Adam Linstedt, CMU); anti-PDI (WB 1:5000, IF 1:200, IP 1:2000; Affinity BioReagents, Golden, CO, USA); anti-rat GS28 (WB 1:500; Stressgene, Victoria, BC, Canada); anti-GAPDH (WB 1:1000; Ambion, Austin, TX, USA); anti-CD44 (clone H4C4) and Lamp2 (clone H4B4) (WB 1:200, IF 1:100; Developmental Studies Hybridoma Bank, University of Iowa); and anti-GalT (IF 1:20, gift from Brian Storrie, UAMS)

Mammalian cell culture

Monolayer HeLa cells were cultured in DMEM/F-12 media supplemented with 15 mm HEPES, 2.5 mm l-glutamine, 5% FBS, 100 U/mL penicillin G, 100 µg/mL streptomycin, and 0.25 µg/mL amphotericin B. Cells were grown at 37 °C and 5% CO2 in a humidified chamber. HeLa cells stably expressing tagged Golgi apparatus proteins were maintained in the presence of 0.4 mg/ml G418 sulfate. HeLa cells stably expressing GlcNAcT1-myc, Mann II-vsv and GalNAcT2-vsv were obtained from B. Storrie's laboratory (UAMS). All cell culture media and sera were obtained from Invitrogen (Carlsbad, CA, USA).

RNA interference

Human COG3 was targeted with siRNA duplex as described previously (19) by Oligofectamine (Invitrogen, Carlsbad, CA, USA). To achieve 3, 6 and 9 days of COG3 KD, cells were transfected each 72-h period. Human COG7 was targeted with a mixture of two Stealth siRNA duplexes (target sense, 1-CCAAGCUCUC-CAGAACAUGCCCAAA; 2-CCUGAAAAUCCCUCUUUGCC-AAGUAU) (Invitrogen). Stealth siRNA duplex (target sense, CCAACCGACUUAAUGGCGCGGUAUU) was used as a mock control (Invitrogen). Cells were transfected with Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) twice each 24-h period according to protocol recommended by Invitrogen. For IF microscopy and WB, HeLa cells were grown in 35 mm dishes (on coverslips for IF) at 60% confluence and lyzed in 2% SDS after 3, 6 and 9 days of COG3 KD or 3 days of COG7 KD.

Radioactive labeling and immunoprecipitation

Pulse-chase experiments were performed using the 35S-Methionine (ICN Biomedicals, Aurora, OH, USA). HeLa cells grown in four 60-mm dishes were mock or siRNA treated for 3 and 9 days, washed twice with DPBS and starved in DMEM without methionine for 30 min. Pulsed by addition of 200 µCi/mL 35S-Methionine for 10 min and chased with complete growth medium containing 2 mm of cold methionine/cysteine mix for 0, 30, 60 and 120 min (at 37 °C). All further steps were performed at room temperature. Cells were washed twice with DPBS and lyzed in 1 mL of TES buffer [50 mm Tris-HCl, pH 7.4, 150 mm NaCl, 1 mm EDTA, 1% Triton X-100, 0.1% SDS and 0.2% sodium azide, protease inhibitor cocktail]. Lysates were incubated with 30 µL of protein A Sepharose CL-4B beads on the tube rotator for 1 h and centrifuged at 20 000 × g for 10 min. Supernatants were transferred to new tubes and incubated overnight with 2 µg of anti-CD44 antibodies (1:50) in a cold room. Samples were clarified by centrifugation as above, and 20 µL of Protein G-Agarose beads were incubated for 1 h and centrifuged at 110 × g for 1 min, washed four times with TBST. Beads were then transferred to a new tube, resuspended in 30 µL of ×1 sample buffer; concentrated supernatants were loaded on SDS-PAGE. Proteins were quantified using a Phosphoimager analysis. After IP with anti-CD44, supernatants of the cell lysates were subjected to IP with anti-Lamp2 (2 mkg).

Immunofluorescence microscopy

Cells grown on coverslips were processed at room temperature as described previously with some modifications (15). Cells were washed once with PBS and fixed by incubating for 10 min with 4% paraformaldehyde (Electron Microscopy Systems, Washington, PA, USA) in PBS, pH 7.4. The coverslips were then washed for 1 min with 0.1% TritonX-100 in PBS, incubated in 0.1% Na-borohydride in PBS for 5 min and washed with 50 mm NH4Cl in PBS for 5 min. Cells were then blocked in 0.1% saponin, 1% BSA in PBS for 15 min. Cells were then incubated for 30 min with primary antibodies and washed off extensively with 0.1% saponin in PBS. Secondary antibodies (Alexa®594 goat anti-rabbit IgG conjugate and Alexa®488 goat anti-mouse IgG conjugate (Molecular Probes) diluted 1:400 in 1% gelatin and 0.1% saponin in PBS were applied for 30 min, and then coverslips were extensively washed with PBS and water. The coverslips were mounted in Prolong Gold Antifade Reagent (Molecular Probes). IF microscopy was performed using an epifluorescence microscope (Axiovert 200; Carl Zeiss International, Thornwood, NY, USA) equipped with the CARV I and CARV II Confocal Imager modules (BioVision Technologies, Exton, PA, USA) with a Plan-Apochromat ×63 oil immersion lens (NA 1.4) at RT. The images were obtained as confocal stacked images and processed on Macintosh computers using IPLab 3.9.3 software (Scanalytics, Fairfax, VA, USA). During the processing stage, individual image channels were pseudocolored with RGB values corresponding to each of the fluorophore emission spectral profiles. Images were cropped using Adobe Photoshop 6.0 software.

Endocytosis of fluorescent dextran

HeLa cells stably expressing GlcNAcT1-myc were grown on coverslips in 6-well plates. Cells were mock or COG3 siRNA treated for 9 days. Cells were treated with Cascade Blue Dextran as previously described (50) with some modifications. Cell cultures were incubated with Cascade Blue Dextran at 0.6 mg/mL in culture medium for 12 h before performing IF assay. Cells were washed with PBS and incubated in fresh medium for 3 h; coverslips were fixed and stained with anti-Lamp2 and anti-myc antibodies.

Treatment with endoglycosydases

HeLa cells were grown in 6-well plates and transfected with COG3 siRNA for 3, 6 and 9 days. Cells were lyzed in 2% SDS and denatured for 15 min at 95 °C. To decrease SDS concentration to 0.5%, 9 µL of each lysate was diluted in 27 µL of water. One half was incubated with supplied buffer and the other with PNGase F (51) (1 µL, 500 units) or Endo H (0.1 µL, 100 units). Samples were incubated at 37 °C for 1 h, dissolved in ×6 sample buffer, subjected to SDS-PAGE and immunoblotted with anti-CD44 and anti-Lamp2 antibodies.

Glycerol velocity gradient

Gradient fractionation was prepared as described previously (19,52) with some modifications. Control and COG3 KD cells were analyzed in pairs simultaneously and under the same conditions. GlcNAcT1-myc were grown in one 6-cm plate, treated with COG3 siRNA, collected by trypsinization, pelleted (2000 × g for 2 min), washed once with PBS and STE buffer (250 mm sucrose, 10 mm triethylamine, pH 7.4, 1 mm EDTA) and homogenized by 20 passages through a 25-gauge needle in 0.6 mL of STE-S (STE buffer without sucrose) containing cocktail of protein inhibitors (Roche Molecular Biochemicals, Indianapolis, IN, USA). Efficiency of homogenization was determined by staining with Trypan Blue. Cell homogenates were centrifuged at 1000 × g for 2 min to obtain PNS. This PNS (0.6 mL) was layered on linear 10–30% (wt/vol) glycerol gradient (12 mL in 10 mm triethylamine, pH 7.4 and 1 mm EDTA on a 0.5 mL 80% sucrose cushion) and centrifuged at 280 000 × g for 60 min in a SW40 Ti rotor (Beckman Coulter, Miami, FL, USA). One milliliter fractions were collected from the top. All steps were performed at 4 °C. Fifty microliter of each fraction as well as an aliquot of PNS were combined with 6× sample buffer, loaded on SDS-PAGE and analyzed by WB. For the analysis of CCD vesicle pool, proteins from fractions 3–5 of glycerol velocity gradient were concentrated by TCA precipitation (53).

In vitro CCD vesicle docking assay

Acceptor RLG membranes were isolated as described previously (19). 20 000 × g supernatant (S20) from COG3 KD HeLa cells stably expressing GlcNAcT1-myc was used as a source of both donor CCD vesicles and cytosol. To prepare cell homogenates of 3 days, COG3 KD cells were grown in 10-cm plate, washed twice with PBS and once with 20 mm HEPES pH 7.4 buffer containing 250 mm sucrose. Sucrose buffer was removed, and cells were scraped from the dish in 1 mL of cold 20 mm HEPES pH 7.4 containing cocktail of protein inhibitors (Roche) and 1 mm DTT. Cells were homogenized on ice by 20 passages through a 25-gauge needle. Unbroken cells were removed by centrifugation at 1000 × g for 2 min to obtain PNS. After that, membranes were stabilized by addition of an equal volume of buffer containing 25 mm KCl and 2.5 mm MgOAc (final concentrations). Large membranes were subsequently removed by centrifugation at 10 000 × g for 10 min and by repeated centrifugation at 20 000 × g for 10 min to obtain S20. Standard vesicle docking reaction (100 µL total volume) contained 50 µL of S20, 0–10 µL of acceptor Golgi membranes (1.5 mg/mL), 10 µL of ATP-regeneration mixture in a 20 mm HEPES, pH 7.4, 25 mm KCl and 2.5 mm MgCl2, 1 mm DTT (buffer HKMD). Reaction was incubated for 30 min at 37 °C, cooled on ice, and then Golgi membranes were pelleted at 10 000 × g for 10 min and washed once with 100 µL of HKMD buffer or high salt buffer (HKMDS buffer with 250 mm KCl). Pellet was resuspended in 20 µL of SDS sample buffer; 10 µL of sample was loaded on SDS-PAGE and analyzed by WB. For Proteinase K treatment, Golgi membranes were incubated for 5 min at 37 °C with Proteinase K (0.25 µg/mL) (Sigma Chemical, St Louis, MO, USA). The reaction was stopped by addition of 1/10 of the volume of 10× cocktail of protein inhibitors, and the membranes were washed twice with HKMD buffer. For mock treatment, cocktail of protein inhibitors was added before Proteinase K treatment. For IgG interference experiment, 5 µL of affinity-purified anti-Cog3p IgGs (0.54 mg/mL) or preimmune IgGs in HKMD buffer were added to vesicle-docking reaction.

SDS-polyacrylamide gel electrophoresis and Western blotting

SDS-PAGE and WB were performed as described previously (17). A signal was detected using a chemiluminescence reagent kit (PerkinElmer Life Sciences, Boston, MA, USA) and quantified using ImageJ software (http://rsb.info.nih.gov/ij/).

Supplementary Material

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Supplementary Material
  7. Acknowledgments
  8. References
  9. Supporting Information

The following figures are available as part of the online article from http://www.blackwell-synergy.com

Figure S1: Stability of COG subunits is altered after 9 days of COG3 KD. Control and 9 days of COG3 KD cell lysates (10 μg of total protein each) were immunoblotted with antibodies against COG complex subunits. GAPDH was used as a loading control.

Figure S2: Lamp2 becomes EndoH sensitive after 6 and 9 days of COG3 KD. Control (0), 3, 6 and 9 days of COG3 KD total cell lysates were treated with EndoH as described in Materials and Methods. Samples were immunoblotted with anti-Lamp2 antibodies. Compare lanes 6 and 9 in control and EndoH panels.

Figure S3: CD44 is primarily localized on the plasma membrane in permiabilized and non-permeabilized cells after 9 days of COG3 KD. Non-permiabilized (upper row) and permeabilized (lower row) cells after 9 days of COG3 KD were triple stained with anti-CD44 (green), anti-myc for GlcNAcT1 (red) and DAPI (blue). There is no GlcNAcT1 signal in non-permeabilized cells. Images were collected at equal signal gains using CARV II microscopy. Bar 10 μm.

Figure S4: GlcNAcT1 is partially colocalized with ER marker PDI after 9 days of COG3 KD. Cells after 9 days of COG3 KD were fixed and stained with anti GlcNAcT1-myc (red), ER marker PDI (green) and nuclei marker DAPI (blue). Partial colocalization of GlcNAcT1 and PDI signals is evident in perinuclear region (merged image). Images were collected at equal signal gains, using CARV II microscopy. Bar 10 μm.

Figure S5: GlcNAcT1 is partially colocalized with ER marker PDI after 3 days of COG7 KD. Cells after 3 days of COG7 KD were stained with anti-myc (red), anti-PDI (green) and DAPI (blue). Notice partial colocalization of GlcNAcT1 and PDI signals. Arrows indicate ER nuclear ring. Images were collected at equal signal gains using CARV microscopy. Bar 10 μm.

Figure S6: CCD vesicles are positioned along actin cables in COG3 KD cells. Cells after 3 days of COG3 KD were stained with anti GlcNAcT1-myc (green), phalloidin-Alexa 594 (red) and DAPI (blue). Notice that CCD vesicles (arrowheads) are positioned along actin cables (arrows) in the merged image. Images were collected using CARV microscopy. Bar 10 μm.

Figure S7: Antibodies recognized specific antigens in RLG and HeLa S20 fractions. Aliquots of RLG and COG3 KD HeLa S20 (approximately 10 μg each) were separated on 10% SDS-PAGE and immunoblotted with antibodies as indicated.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Supplementary Material
  7. Acknowledgments
  8. References
  9. Supporting Information

We are very grateful to O. Pavliv for excellent technical assistance. We also thank M. Jennings, F. Hughson, A. Linstedt, B. Storrie, E. Sztul, D. Ungar and others who provided reagents and critical reading of the manuscript. Supported by grants from the National Science Foundation (MCB-0234822) and the DOD (DAMD17-03-1-0243).

References

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Supplementary Material
  7. Acknowledgments
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Supplementary Material
  7. Acknowledgments
  8. References
  9. Supporting Information

Figure S1. Stability of COG subunits is altered after 9 days of COG3 KD. Control and 9 days of COG3 KD cell lysates (10 μg of total protein each) were immunoblotted with antibodies against COG complex subunits. GAPDH was used as a loading control.

Figure S2. Lamp2 becomes EndoH sensitive after 6 and 9 days of COG3 KD. Control (0), 3, 6 and 9 days of COG3 KD total cell lysates were treated with EndoH as described in Materials and Methods. Samples were immunoblotted with anti-Lamp2 antibodies. Compare lanes 6 and 9 in control and EndoH panels.

Figure S3. CD44 is primarily localized on the plasma membrane in permiabilized and nonpermiabilized cells after 9 days of COG3 KD. Non-permiabilized (upper row) and permiabilized (lower row) cells after 9 days of COG3 KD were triple stained with anti-CD44 (green), anti-myc for GlcNacT1 (red) and DAPI (blue). There is no GlcNacT1 signal in non-permiabilized cells. Images were collected at equal signal gains, using CARV II microscopy. Bar 10μm.

Figure S4. GlcNacT1 is partially co-localized with ER marker PDI after 9 days of COG3 KD. Cells after 9 days of COG3 KD were fixed and stained with anti- GlcNacT1-myc (red), ER marker PDI (green) and nuclei marker DAPI (blue). Partial co-localization of GlcNacT1 and PDI signals is evident in perinuclear region (merged image). Images were collected at equal signal gains, using CARV II microscopy. Bar 10μm.

Figure S5. GlcNacT1 is partially co-localized with ER marker PDI after 3 days of COG7 KD. Cells after 3 days of COG7 KD were stained with anti-myc (red), anti-PDI (green) and DAPI (blue). Notice partial co-localization of GlcNacT1 and PDI signals. Arrows indicate ER nuclear ring. Images were collected at equal signal gains, using CARV microscopy. Bar 10μm.

Figure S6. CCD vesicles are positioned along actin cables in COG3 KD cells. Cells after 3 days of COG3 KD were stained with anti- GlcNacT1-myc (green), phalloidin Alexa 594 (red) and DAPI (blue). Notice, that CCD vesicles (arrowheads) are positioned along actin cables (arrows) in the merged image. Images were collected using CARV microscopy. Bar 10μm.

Figure S7. Antibodies recognized specific antigens in RLG and HeLa S20 fractions. Aliquots of RLG and COG3 KD HeLa S20 (~ 10 μg each) were separated on 10% SDSPAGE and immunoblotted with antibodies as indicated.

FilenameFormatSizeDescription
tra_440_sm_7_2a_figures.pdf1653KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.