SEARCH

SEARCH BY CITATION

Keywords:

  • ergosterol;
  • lipid rafts;
  • plasma membrane;
  • secretory pathway;
  • sorting;
  • sphingolipids;
  • trans-Golgi network;
  • vesicles

Abstract

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

Previous work has showed that ergosterol and sphingolipids become sorted to secretory vesicles immunoisolated using a chimeric, artificial raft membrane protein as bait. In this study, we have extended this analysis to three populations of secretory vesicles isolated using natural yeast plasma membrane (PM) proteins: Pma1p, Mid2p and Gap1*p as baits. We compared the lipidomes of the immunoisolated vesicles with each other and with the lipidomes of the donor compartment, the trans-Golgi network, and the acceptor compartment, the PM, using a quantitative mass spectrometry approach that provided a complete lipid overview of the yeast late secretory pathway. We could show that vesicles captured with different baits carry the same cargo and have almost identical lipid compositions; being highly enriched in ergosterol and sphingolipids. This finding indicates that lipid raft sorting is a generic feature of vesicles carrying PM cargo and suggests a common lipid-based mechanism for their formation.

The trans-Golgi network (TGN) is a conserved organelle of eukaryotic cells that governs the trafficking and sorting of newly synthesized proteins and lipids destined to different organelles (1). From the TGN in the yeast Saccharomyces cerevisiae two pathways emanate, which provide the growing yeast plasma membrane (PM) with distinct membrane constituents. The two carriers can be separated based on their different densities and therefore are classified as light- and heavy-density secretory vesicles (LDSV and HDSV) (2–4). The two vesicle types transport different cargo proteins; LDSV are responsible for transporting integral PM proteins, e.g. Pma1p, Hxt2p and glycosyl-phosphatidylinositol (GPI)-anchored Gas1p, as well as a variety of cell wall proteins such as Bgl2p and Chs3p, whereas HDSV transport the periplasmic enzymes invertase and acidic phosphatase (2,4–7).

The generation of HDSV is clathrin (Chc1p) and dynamin (Vps1p) dependent and upon deletion of these genes, cargo is re-routed to LDSV (3). The same effect was observed when genes encoding proteins from the late endosomal pathway, like PEP12 (the t-SNARE of late endosomes) or VPS4 (AAA-ATPase involved in the multivesicular body sorting), were deleted (3,4). These and other observations suggest a partial overlap of the machineries involved in the formation of HDSV and the vacuolar sorting machinery.

Proteins transported to LDSV have different trafficking requirements, independent of clathrin and dynamin. Chs3p, Fus1p and the artificial lipid raft protein FusMidp require the exomer complex for their exit from the TGN, whereas Pma1p or Hxt2 do not (8–12). In addition, it was shown that the so-called ‘raft lipids'—in yeast, ergosterol and the sphingolipids inositolphosphoceramide (IPC), mannosyl-IPC (MIPC) and mannosyl-di-IPC (M(IP)2C)—facilitate proper cell surface trafficking of many PM and GPI-anchored proteins, e.g. Gas1p, tryptophan permease Tat2p, uracil permease Fur4p, PM ATPase Pma1p and also FusMidp (8,13–17). These and other yeast surface proteins become associated with sphingolipids and sterols in the Golgi apparatus (as reflected by acquired detergent resistance (18,19)) or, in the case of Gas1p, become associated with detergent-resistant membranes which are already present in the endoplasmic reticulum (ER) (13,20,21). The self-association of sphingolipids and sterols with proteins is responsible for lipid raft formation in vivo(19) and their connectivity is manifested in the capability of yeast lipids for phase separation (22).

The functional requirement of lipid rafts in cargo sorting at the TGN was previously demonstrated by genetics and biochemistry (8,20). Later by employing an immunoisolation technique, we purified yeast secretory vesicles carrying the lipid raft marker protein FusMidp. Using quantitative shotgun lipidomics, we showed that these vesicles were enriched in ergosterol and sphingolipids (23). This observation raised the question whether this finding was general, i.e. applicable also to other types of cargo proteins destined for the PM.

In this work, we isolated three additional secretory vesicle populations carrying different bait-cargoes: Pma1p; Mid2p, a cell wall integrity sensor and the donor of the transmembrane and intracellular domain for the FusMid chimeric model protein (24) and Gap1*p, the ubiquitylation mutant of Gap1p known to be associated with lipid rafts and delivered directly to the PM (21). We also isolated the TGN/endosome (TGN/E) system as the donor organelle and the PM as the acceptor compartment of the secretory vesicles. We show that the lipidomes of the distinct secretory vesicle populations were almost identical and specifically enriched in ‘raft lipids' as compared to the donor compartment, the TGN/E. These results unequivocally show the functional involvement of lipid rafts in sorting and delivery of membrane proteins from the TGN to the cell surface in yeast.

Results

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

Immunoisolation of yeast secretory vesicles bound for the PM

To isolate yeast secretory vesicles transporting different cargo proteins to the cell surface, we applied the immunoisolation method introduced in our previous work (23). The baits for the immunoisolation were PM proteins C-terminally tagged with a linker-separated, Tobacco Etch Virus (TEV) protease cleavable, ninefold myc epitope (LTLM9). We used the multispanning transmembrane ATPase Pma1p and amino acid transporter Gap1(K9K16)p (Gap1*p), and the single transmembrane domain protein Mid2p. To monitor subcellular localization during the purification procedure, Mid2p and Gap1*p were additionally fused with GFP yielding the constructs: Mid2GFPLTLM9 and Gap1*GFPLTLM9, respectively (GFP was omitted for Pma1p resulting in Pma1pLTLM9). These constructs were expressed in the temperature-sensitive exocyst mutant sec6-4(25). At the permissive temperature (24°C), secretory vesicles containing bait proteins were transported to the PM, whereas at the restrictive temperature (37°C) they accumulated intracellularly (Figure S1A,B). Similarly, the soluble HDSV cargo invertase (Suc2p) tagged with RFP (invertaseRFP) showed intracellular accumulation only at the restrictive temperature. This behavior was found for all bait proteins and was consistent with our previous data (23).

Cells that had accumulated secretory vesicles at the nonpermissive temperature were then disrupted mechanically and fractionated using sequential centrifugation steps, which yielded the input material for the isopycnic sucrose gradient centrifugation used for the prepurification of organelles. All immunoisolation baits showed similar density distribution as previously observed for FusMidp (23). Secretory vesicles were immunoisolated using the sucrose density gradient fraction with a refractive index of n = 1.407 (Figure S1C). Mouse anti-myc antibodies were added to the fraction, which was incubated with the self-made immunoadsorbent bed consisting of microgranular cellulose coupled to sheep anti-mouse antibody (23,26). At the same time, antibodies directed against the endosomal t-SNARE Pep12p were added to retain endosomes (immunodepletion) on the immunoadsorbent, greatly improving the purity of the isolated secretory vesicles (23). Bait-carrying secretory vesicles were specifically eluted using TEV protease, whereas Pep12p containing membranes remained bound. The observed shift in electrophoretic mobility between the bait protein in the input and eluate confirmed the changes in protein size caused by cleavage of the 9myc tag (Figure 1).

image

Figure 1. Immunoisolation of Gap1*p- (A), Mid2p- (B) and Pma1p-vesicles (C). Several organelle markers were monitored throughout the isolation of vesicles. Anti-GFP antibodies were used to detect the baits. For Pma1p-vesicles, anti-Pma1p antibodies were used (two sizes of Pma1p detected are 120 kDa, which corresponds to the uncleaved bait, and 100 kDa to its cleaved version as well as the endogenous Pma1p. Both versions were detected only prior to TEV elution). InvertaseRFP (InvRFP) was used as a specific marker for HDSV. Pep12p is the late endosome t-SNARE, Dpm1p is an ER protein. Bind, material bound to the immunoadsorbent; ELUATE, eluted material released from the immunoadsorbent after TEV protease cleavage; FT, flow through; Input, sucrose gradient fraction; Remaining, material not eluted from the immunoadsorbent by TEV protease; Wash, supernatant of the final wash.

Download figure to PowerPoint

To assess the purity of yielded vesicles, marker proteins for major contaminants were screened throughout the immunoisolation procedure (Figure 1). ER contamination, as indicated by Dpm1p, could not be detected in eluates, which also did not contain detectable levels of the endosomal marker Pep12p. InvertaseRFP was not present in the eluates; most of it being separated during gradient centrifugation. These findings show that similar to FusMidp-vesicles, Pma1p-, Mid2p- and Gap1*p-vesicles are biochemically distinct from invertaseRFP carriers (23).

Different vesicles show comparable cargo protein composition

Having isolated secretory vesicles carrying different bait-cargo, we compared their protein composition by western blotting a set of marker proteins (Figure 2). All eluates invariably contained significant amounts of endogenous Pma1p and Gas1p. In addition, endoglucanase Bgl2p, a non-membrane, soluble LDSV protein, was always detected in all four vesicle preparations (27). We noted differences in marker protein abundances between vesicle preparations, however they fluctuated between experiments and therefore we concluded that they are not significant and rather reflect experimental variations. To further test whether different cargo proteins co-populate the same fraction of vesicles, we co-expressed the FusMidRFP construct together with the respective immunoisolation bait: Pma1p, Mid2p or Gap1*p. After isolation we could detect the fully glycosylated, mature post-Golgi form (24) of FusMidRFP in secretory vesicles isolated with all baits, showing that this artificial cargo populates the same carriers as natural cargo proteins.

image

Figure 2. Cargo similarities in different bait vesicles. Vesicle pellets were analyzed by western blotting for the presence of different cargo markers. Pma1p is the major PM ATPase in yeast, Gas1p is the major GPI-anchored protein, Bgl2p is the cell wall endoglucanase and all are LDSV markers. FusMidRFP construct was co-expressed as a second marker concurrently to the vesicle baits and was detected with the anti-RFP antibodies.

Download figure to PowerPoint

All immunoisolated secretory vesicles are enriched in sphingolipids and ergosterol

To determine the lipid composition of the different secretory vesicle preparations, we performed a comparative lipidomics analysis of the immunoisolated material. We used a mass spectrometry (MS)-based quantitative shotgun lipidomics platform for absolute quantification (i.e. picomole or mole percentage) of ergosterol, sphingolipid, glycerophospholipid and glycerolipid species (28).

The lipidomes of secretory vesicles isolated using four different cargo baits were almost identical (Figure 3, Table S1). The main observation was the strong enrichment of sphingolipids and ergosterol in the vesicle lipidomes relative to their donor compartment. Ergosterol was enriched more than twofold compared with the TGN/E donor compartment, and with 31 mol%, it was also the most abundant lipid in the vesicle preparations. Total sphingolipids were enriched by a factor of ∼1.8, but the individual classes showed different degrees of enrichment: IPC and MIPC were enriched 1.9- and 2.5-fold, respectively, whereas M(IP)2C was enriched ∼1.6-fold. IPC and M(IP)2C were the two most abundant sphingolipid classes in the vesicle lipidomes, contributing on average 10.1 and 8.1 mol%, respectively.

image

Figure 3. Lipid class and category composition of FusMidp-, Mid2p-, Gap1*p-, Pma1p-vesicles, TCE, TGN/E compartments and the PM. The mole% of lipid class was calculated as the sum of the mole% of lipid species of the respective lipid class and of categories as the sum of constituting classes (inset). For vesicles, TGN/E and PM each bar represents n = 4 experiments and for TCE n = 16 experiments mean estimate ±SD. TCE, total cell extract; TGN/E, trans-Golgi network/endosome; SV, secretory vesicles; PM, plasma membrane; CL, cardiolipin; DAG, diacylglycerol; PA, phosphatidic acid; PS, phosphatidylserine; PE, phosphatidylethanolamine; PC, phosphatidylcholine; PI, phosphatidylinositol; Erg, ergosterol; Cer, ceramide; IPC, nositolphosphoceramide; MIPC, mannosyl-IPC; M(IP)2C, mannosyl-di-IPC. Inset: GL, glycerolipids (DAG); SP, sphingolipids (IPC, MIPC, M(IP)2C) and ceramides; GP, glycerophospholipids (PA, PS, PE, PC, PI); ST, sterols (Erg).

Download figure to PowerPoint

Lipid composition changes gradually along the late secretory pathway

To assess the differences in lipid composition of the major organelles within the late secretory pathway, we also isolated the late Golgi apparatus membranes, from which secretory vesicles are generated, and the PM, which is the final destination of all secretory traffic. Because in yeast all known TGN markers cycle continuously between late Golgi apparatus and endosomal compartments, we had to isolate a mixture of these organelles further referred to as TGN/E. For this task, we used the immunoisolation technique with the bait being wild-type Gap1p, which cycles within the TGN/E compartment under our experimental conditions (23,29,30). PMs were isolated according to the established protocols based on density gradient centrifugation (31), as described previously (23). The lipid composition of both organelles was analyzed in the same manner as the vesicles (Figure 3, Table S1). The results showed the expected lipid gradient along the secretory pathway with the lipid composition of secretory vesicles positioned between the donor TGN/E and acceptor PM. In the TGN/E membranes, sphingolipids and ergosterol combined accounted for ∼27 mol% of total lipids. In secretory vesicles, half of the membrane was constituted of sphingolipids and sterols (∼53 mol%) and these lipid categories peak at the PM, where they made up ∼82 mol% of total lipids.

Interestingly, at the PM only M(IP)2C and ceramides were significantly enriched, whereas the two other sphingolipid classes were either reduced (IPC) or maintained (MIPC) as compared to the secretory vesicles. Glycerophospholipids showed the opposite trend with decreasing amounts toward the PM. The only exception in this category was phosphatidic acid (PA), which was gradually enriched along the secretory pathway. Notable was also that cardiolipin (CL) was barely detectable in all preparations indicating minimal contamination with mitochondria (Figure 3).

Detailed analyses of the glycerophospholipid species revealed changes in their fatty acid (FA) composition between different organelles. The double bond (DB) index, which denotes the total number of DBs within a lipid molecule, decreased along the secretory pathway (Figure 4). This effect spans all glycerophospholipid classes, with the exception of phosphatidylcholine (PC) at the PM (Figure S3A). The ratios between doubly unsaturated glycerophospholipids (composed of two monounsaturated FAs; DB = 2) and monounsaturated (DB = 1) or fully saturated ones (DB = 0) were the lowest at the PM.

image

Figure 4. Glycerophospholipids DB index analyses. The mole% of given DB index for glycerophospholipids was calculated as the sum of the mole% of glycerophospholipid species with the respective DB. For the secretory vesicle (SV), bar is an average for all types of vesicles (n = 16 experiments mean estimate, 4 for each type of vesicles), TCE represents n = 16 and TGN/E and PM n = 4 experiments mean estimate; all ±SD.

Download figure to PowerPoint

Immunoisolated vesicles show enrichment of C42 and C46 sphingolipid species

Analyses of the individual species of sphingolipids in the vesicle fractions revealed that some species were selectively enriched compared to others (Figure 5). In the total cell extract (TCE), the most abundant sphingolipids carried 44 carbon atoms in total [this number is a sum of the length of a long chain base (LCB) and amide-linked fatty acid (FA) moiety], with C46 and C42 species much less abundant. This abundance profile was generally preserved in the secretory vesicles, but the ratios between 42, 44 and 46 carbon atoms-containing species were significantly different from the TCE. In secretory vesicles, the ratio of sphingolipid species with 46 to species with 44 carbon atoms was 0.29, whereas TCE showed a ratio of 0.12 (Figure 5B). Similarly, ratios of C42 to C44 sphingolipids differed between the vesicles and the TCE, with values of 0.06 and 0.02, respectively, showing that the vesicles were specifically enriched in C46 and C42 sphingolipid species. Both TGN/E and PM had lower ratios of 46 to 44 and 42 to 44 carbon atom-containing sphingolipids, as compared with secretory vesicles. Interestingly, the enrichment of C46 and C42 sphingolipids in vesicles was found for all sphingolipid classes, although it was most pronounced for MIPC. The ratio of C46 to C44 MIPC species in the vesicles amounted to 0.50, whereas the ratios for IPC and M(IP)2C species were 0.24. The effect of specific enrichment of 46 and 42 carbon atoms-containing sphingolipid species was not correlated with their hydroxylation; the majority of sphingolipids contained three hydroxyl groups and there was no significant difference in the degree of hydroxylation between secretory vesicles and the other organelles of the secretory pathway (Figure S3B).

image

Figure 5. Sphingolipid species analyses. A) The mole% of given sphingolipid length was calculated as the sum of the mole% of sphingolipid (IPC, MIPC and M(IP)2C) species of the respective length. For the secretory vesicle (SV), bar is an average for all types of vesicles (n = 16 experiments mean estimate, 4 for each type of vesicles), TCE represents n = 16, TGN/E and PM n = 4 experiments mean estimate; all ±SD. Forty carbon atoms species are not shown (on average in all samples <0.06% of total category). B) Carbon atoms sphingolipids (46 to 44) ratios in different organelles. Total sphingolipids (SLs) constitute IPC, MIPC and M(IP)2C. Structure of the most abundant M(IP)2C species—44:0;4, with the LCB and amide-linked FA moiety in red and blue.

Download figure to PowerPoint

Discussion

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

Different cargo vesicles belong to one uniform class

In our previous work, we isolated yeast secretory vesicles carrying the model lipid raft protein construct FusMidp (23). Using this artificial bait protein, we were able to show that cargo sorting at the TGN selectively enriches ergosterol and sphingolipid species in secretory vesicles. In this study, we show that the enrichment of ‘raft lipids' is a general characteristic for the formation of secretory vesicles carrying PM protein cargo.

To this end, we immunoisolated secretory vesicles carrying Pma1p, Mid2p and Gap1*p as natural PM cargoes and also included FusMidp-vesicles as an additional reference. Comparative western blotting for a set of marker proteins revealed a surprisingly similar protein composition for each of the isolated secretory vesicles. They did not show significant ER or endosomal impurities, but all contained endogenous Bgl2p, Gas1p and Pma1p (Figures 1 and 2). Moreover, FusMidp when co-expressed as cargo was also present in all the isolated vesicles regardless of the bait used for isolation (Figure 2). The observation that the different bait vesicles contain the same LDSV marker proteins strongly suggests that these carriers are actually representative of the same vesicle pool and are generated by a single sorting process at the TGN, a conclusion further supported by the similarities of their lipidomes.

Lipidomics of yeast late secretory pathway

Using quantitative shotgun lipidomics, we determined the lipid compositions of each vesicle fraction, including the previously characterized FusMidp-vesicles (Figure 3, Table S1). Our analysis revealed that all vesicle preparations have almost identical lipid compositions in terms of mole fractions of measured lipid classes and species. Comparing the lipid composition of the secretory vesicles with that of their donor compartment (TGN/E), ergosterol and sphingolipids are enriched in vesicular carriers, whereas glycerophospholipids are depleted. One exception is PA, the concentration of which increases in the vesicles.

In the lipid composition of the acceptor compartment, the PM, ergosterol and M(IP)2C contents are further increased over the secretory vesicle composition, whereas IPC level is reduced. The lipidomes of the TGN/E, vesicle and PM compartments are in agreement with the observed gradual enrichment of sphingolipids and ergosterol along the secretory pathway toward the PM (32–35). The PM exhibits the highest sphingolipid and ergosterol content (35–38), explaining its increased resistance to solubilization with detergents (13,39).

We noted that ergosterol and M(IP)2C are enriched to a much higher degree at the PM than in the secretory vesicles and correspondingly most of glycerophospholipids at the PM are reduced below the levels measured for the vesicles. This indicates that lipid composition of the PM is regulated by other processes as well, including sorting of lipids during endocytosis, metabolism and perhaps by exchange over direct organelle contact sites or lipid transfer proteins (40–43).

Along the secretory pathway, we also recorded an increase of saturated and monounsaturated over diunsaturated glycerophospholipids species (Figures 4 and S3A); the PM being highly enriched in glycerophospholipid species containing one saturated acyl chain. This result is in agreement with the previously reported enrichment of saturated phosphatidylserine (PS) and phosphatidylethanolamine (PE) species at the PM (38). Schneiter et al. (38) proposed three possible explanations for these lipidomic differences: lipid selection and/or remodeling in the Golgi apparatus, in situ acyl chains remodeling at the PM and selective removal of ‘non-fitting’ lipid species from the PM. Our data on secretory vesicle lipidomes, when compared to the donor TGN/E compartment, show that a major change occurs at the PM, which becomes significantly enriched in monounsaturated glycerophospholipid species, supporting remodeling or selective lipid removal only after PM delivery.

Additionally, monounsaturated compared to diunsaturated PC was shown to increase membrane order as reflected by the generalized polarization of membranes stained with Laurdan probes (44). In our previous study, we observed that the FusMidp-vesicle membranes are significantly more ordered when compared to TGN/E (23). However, because the DB index of glycerophospholipids of these two preparations is similar, we conclude that the observed increase of membrane order is solely because of to the higher sphingolipid and ergosterol content.

More detailed insight into lipid species composition revealed that the isolated vesicles are specifically enriched in sphingolipid species with 46 and 42 carbon atoms (Figure 5). Yeast sphingolipids contain a LCB with a polar head group and an amide-linked FA moiety (Figure 5C). Yeast LCBs can contain 16, 18 or 20 carbon atoms, whereas the FA moiety is usually 22–26 carbon atoms long (45,46). C44 species primarily comprise C18 LCB and C26 FA, whereas C46 species contain C20 LCB and C26 FA, and C42 species have C18 LCB and C24 FA (28). We noted that the C46 and C42 sphingolipid species enriched in isolated vesicles differ in length between LCB and FA chains by six carbons, in contrast to the depleted C44 species, with a length difference of eight carbon atoms. Therefore, the observed changes can be considered as a specific enrichment of sphingolipids with particular length difference between their LCB and FA chains. Whether the observed enrichment of particular sphingolipid species could play a role in vesicle formation remains to be seen. For instance, differences in acyl chain lengths of asymmetric lipid molecules could be involved in interdigitation of the lipid leaflets (47). While it is known that the length of the FA in the sphingolipid is important for correct trafficking of transmembrane proteins (8,17), the fine-tuning of sphingolipid species could also have functional implications.

The lipid composition of carriers from the second secretory pathway, HDSV, remains to be investigated. Because no transmembrane receptor for HDSV cargo that could serve as immunoisolation bait is known, we could not address this issue using our methods. We assume that HDSV generation does not depend on ergosterol and sphingolipid segregation because lipid mutants affecting these lipids were not identified in a genetic screen for vacuolar/endosomal sorting factors (48).

Model for carrier formation and cargo sorting at the TGN

The finding that vesicles carrying different cargo proteins to the PM have similar lipid compositions, i.e. being enriched in ergosterol and sphingolipids, and share the most abundant PM markers as cargoes suggests a common mechanism for the formation of secretory vesicles at the yeast TGN. Therefore, we propose a general model for cargo and lipid sorting in PM-destined carrier formation at the level of TGN.

In contrast to classical coat–adaptor mechanisms, carrier formation and budding in our model are driven mainly by lipid phase separation and domain-induced budding based on lipid raft clustering (23,49,50) and facilitated by peripheral membrane proteins such as BAR (BIN/amphiphysin/RVS) domain-containing proteins and the cytoskeleton (51). Cargo would be included in the carriers by partitioning into lipid rafts or by becoming recruited/clustered by additional factors, as in the case of Pma1p by Ast1p (20). Cargo could use dedicated mechanisms for specific recruitment, such as Chs3p and exomer, allowing cell cycle-controlled trafficking (12). Cargo not destined for the PM would be excluded from these carriers and instead recruited by positive sorting signals into clathrin-dependent (like HDSV) and AP-3 routes or for intracellular trafficking triggered by ubiquitylation (52). How the triggering of lipid raft clustering is regulated remains to be established in future studies. Here, we note that lipid rafts coalescence could be facilitated by cargo proteins with self-clustering potential with highly glycosylated GPI-anchored proteins as possible candidates. Also cytosolic proteins, such as Kes1p and Rvs161p previously identified in the screen for sorting factors (8), could further regulate the process. This lipid raft domain-induced budding model would be similar to that proposed for apical raft sorting in epithelial cells (53,54), or for Shiga toxin or virus-mediated raft endocytosis (55,56).

Materials and Methods

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

Yeast strains and media

Yeast strain used was KSY302, which carries the temperature-sensitive allele sec6-4 in leu2-3, 112; ura3-52 background (from P. Novick, Yale University School of Medicine). All media were synthetic, containing yeast nitrogen base (Difco), CSM-ura-leu (MP Biomedicals) with 2% raffinose (Sigma-Aldrich) and 2% galactose (Merck) for induction.

Plasmids

All plasmids were based on the p41x backbone, which is the yeast centromere plasmid, and all constructs were under galactose-inducible GALS promoter (57). p145 is the invertaseRFP-carrying plasmid (23) in p415GALS backbone and p147 is FusMidGFPLTLM9 (FusMidp bait) (23), p146 is Gap1(K9K16)GFPLTLM9 (Gap1*p bait based on Gap1(K9K16)p construct from B. André, ULB), p165 is Gap1GFPLTLM9 (TGN/E bait based on wild-type Gap1GFP construct from C. Kaiser, MIT), p167 is Pma1pLTLM9 (Pma1p bait omitting GFP) and p174 is Mid2GFPLTLM9 (Mid2p bait), all in p416GALS backbone.

Subcellular fractionation and immunoisolation

All organelles were purified from the same yeast background (KSY302) cultivated in the same conditions with baits and other constructs expressed as described. Purification of vesicles and TGN/E system was performed using self-made immunoadsorbent microgranular cellulose with the procedure described previously (23) with one additional 10-min 20 000×g centrifugation of eluted material to remove any residual traces of cellulose. PM was purified (31) with one additional sucrose gradient centrifugation as described (23).

Antibodies

Mouse anti-myc antibody (c-Myc [9E10], Santa Cruz Biotechnology, Inc.), mouse anti-GFP (Roche), rabbit anti-RFP (from M. Zerial, MPI-CBG), rabbit anti-Gas1p (from H. Riezman, University of Geneva), mouse anti-Pep12p (Invitrogen), mouse anti-Dpm1p (Invitrogen), rabbit anti-Pma1p (from. R. Serrano, University of Valencia) and rabbit anti-Bgl2p (from R. Schekman, University of California) were used in this study.

Microscopy

Microscopy was performed with an epi-fluorescent microscope (BX61; Olympus) with UPlanSApo 100× NA 1.40 oil immersion objective (Olympus). Pictures were acquired at room temperature in synthetic defined medium with a charge-coupled device camera (SPOT; Diagnostic Instruments, Inc.) using MetaMorph software (MDS Analytical Technologies).

Mass spectrometric lipid analysis

The lipid compositions of secretory vesicles, TGN/E, PM and TCEs were determined by quantitative shotgun lipidomic analysis as previously described (23,28). In short, samples were mixed with 30 µL of internal lipid standard mixture, providing a total spike of 24 pmol diacylglycerol (DAG) 17:0-17:0, 22 pmol PA 17:0-14:1, 41 pmol PE 17:0-14:1, 41 pmol PS 17:0-14:1, 42 pmol PC 17:0-14:1, 40 pmol PI 17:0-14:1, 14 pmol CL 15:0-15:0-15:0-16:1, 22 pmol ceramide 18:0;3/18:0;0, 37 pmol IPC 18:0;2/26:0;0, 36 pmol MIPC 18:0;2/26:0;0, 31 pmol M(IP)2C 18:0;2/26:0;0 and 57 pmol cholesterol-D7. Samples were subsequently subjected to two-step lipid extraction executed at 4°C (28). The lower organic phases were isolated and evaporated in a vacuum evaporator at 4°C. Lipid extracts were dissolved in 100 µL chloroform/methanol (1:2; v/v) and analyzed by quantitative MS using an LTQ Orbitrap XL (Thermo Fisher Scientific) equipped with a robotic nanoflow ion source TriVersa NanoMate (Advion Biosciences, Inc.). PA, PS, PE, PC, CL, PI, IPC, MIPC and M(IP)2C species were monitored by negative ion mode FT MS analysis, whereas DAG as well as ceramide species were monitored by positive ion mode FT MS analysis. Ergosterol and cholesterol-D7 were subjected to chemical sulfation and analyzed by negative ion mode FT MS analysis (58). Identification and quantification of molecular lipid species were performed using LipidXplorer software (59).

Lipid species were annotated according to their molecular composition. Glycerophospholipid and DAG species are annotated as: <lipid class><sum of carbon atoms in the two FAs>:<sum of DBs in the two FAs> (e.g. PI 34:1). Sphingolipid species are annotated as: <lipid class><sum of carbon atoms in the LCB and FA moiety>:<sum of DBs in the LCB and the FA moiety>;<sum of hydroxyl groups in the LCB and the FA moiety> (e.g. IPC 44:0;4).

Acknowledgments

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

We thank Maria Carvalho for the help with ergosterol quantification; Julio Sampaio, Andrej Shevchenko and Kai Schuhmann for the helpful advice concerning lipid MS and Howard Riezman, Randy Schekman, Ramon Serrano and Marino Zerial for sharing antibodies. This work was supported by DFG ‘Schwerpunktprogramm1175’ grant no: SI459/2-1, DFG ‘Transregio 83’ grant no: TRR83 TP02, ESF ‘LIPIDPROD’ grant no: SI459/3-1, BMBF ‘ForMaT’ grant no: 03FO1212, the Klaus Tschira Foundation and the Danish Council for Independent Research (09-072484, CSE) and Lundbeckfonden (R45-A4342, CSE).

References

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

Supporting Information

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

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Figure S1: Intracellular accumulation of the immunoisolation baits at the restrictive temperature (37°C) in sec6-4 cells (A and B) and the organelle marker profile of the sucrose gradient used for prepurification of secretory vesicles (C). A) Gap1&ast;GFPLTLM9 and (B) Mid2GFPLTLM9 immunoisolation baits and invertaseRFP (InvRFP) expressed for 45 min. At the permissive temperature (24°C), immunoisolation baits reached the PM, while InvRFP was secreted. At the restrictive temperature (37°C), both proteins accumulated intracellularly. Bars, 2 µm. C) The sucrose gradient was fractionated into 1 mL fractions and analyzed by western blot analysis for the immunoisolation baits FusMidGFPLTLM9 or Gap1&ast;GFPLTLM9, Mid2GFPLTLM9 with anti-GFP antibodies and Pma1LTLM9 with anti-Pma1p antibodies. InvRFP is a HDSV marker, Pep12p is an endosomal marker and Dpm1p is an ER marker. A fraction 7 was used for the immunoisolation of secretory vesicles. The graph shows the refractive index (n; solid line) as a function of fraction number. DIC, differential interference contrast.

Figure S2: Immunoisolation FusMidp-vesicles (A), TGN/E system (B) and PM (C). Several organelle markers were monitored throughout the isolation. A and B) Anti-GFP antibodies were used to detect the FusMidp and TGN/E baits. invertaseRFP (InvRFP) was used as a specific marker for HDSVs. Pep12p is the late endosome t-SNARE and Dpm1p is an ER protein. C) Pma1p is the major PM and Gas1p is the major GPI-anchored protein in yeast, both PM markers. Remaining antibodies descriptions as in (A) and (B). 1st gradient, output of the first gradient; Bind, material bound to the immunoadsorbent; ELUATE, eluted material released from the immunoadsorbent after TEV protease cleavage; FT, flow through; IN, input of the first gradient; Input, sucrose gradient fraction 7; PM, the final PM preparation; Remaining, material not eluted from the immunoadsorbent by TEV protease; Wash, supernatant of the last wash.

Figure S3: DB index for glycerophospholipid classes (A) and species composition of sphingolipid classes (B) in different organelles. A) The mole% of given DB index for class was calculated as the sum of the mole% of glycerophospholipid species with the respective DB. B) Species composition of sphingolipid classes. Species below 1% of total class are not plotted. For the secretory vesicle (SV), bar is an average for all types of vesicles (n = 16 experiments mean estimate, 4 for each type of vesicles), TCE represents n = 16 and TGN/E and PM n = 4 experiments mean estimate; all ±SD.

Table S1: Detailed species composition of FusMidp-, Mid2p-, Gap1&ast;p-, Pma1p-vesicles, TCE, TGN/E compartments and the PM. Quantitative lipidomic analysis of different secretory vesicles, TGN/E and PM allowed absolute quantification of 104 molecular lipid species. Lipid composition is in mol% to show the stoichiometric relationship between the lipid species. For vesicles, TGN/E and PM the value represents n = 4 experiments and for TCE n = 16. Erg, ergosterol; Cer, ceramide.

FilenameFormatSizeDescription
TRA_1221_sm_fs1.pdf2470KSupporting info item
TRA_1221_sm_fs2.pdf572KSupporting info item
TRA_1221_sm_fs3.pdf363KSupporting info item
TRA_1221_sm_ts1.xls86KSupporting 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.