No abstract

The mechanism by which newly synthesized sterols are transported from their site of synthesis, the endoplasmic reticulum (ER), to the sterol-enriched plasma membrane (PM) is not fully understood. Studies in mammalian cells suggest that newly synthesized cholesterol is transported to the PM in Golgi-bypassing vesicles and/or via a nonvesicular process. Using the yeast Saccharomyces cerevisiae as a model system, we now rule out an essential role for known vesicular transport pathways in transporting the major yeast sterol, ergosterol, from its site of synthesis to the PM. We use a cyclodextrin-based sterol capture assay to show that transport of newly synthesized ergosterol to the PM is unaltered in cells defective in Sec18p, a protein required for almost all intracellular vesicular trafficking events; we also show that transport is not blocked in cells that are defective in formation of transport vesicles at the ER or in vesicle fusion with the PM. Our data suggest instead that transport occurs by equilibration (t(1/2) approximately 10-15 min) of ER and PM ergosterol pools via a bidirectional, nonvesicular process that is saturated in wild-type exponentially growing yeast. To reconcile an equilibration process with the high ergosterol concentration of the PM relative to ER, we note that a large fraction of PM ergosterol is found condensed with sphingolipids in membrane rafts that coexist with free sterol. We propose that the concentration of free sterol is similar in the PM and ER and that only free (nonraft) sterol molecules have access to a nonvesicular transport pathway that connects the two organelles. This is the first description of biosynthetic sterol transport in yeast.

Newly synthesized proteins and lipids are transported across the Golgi complex via different mechanisms whose respective roles are not completely clear. We previously identified a non-vesicular intra-Golgi transport pathway for glucosylceramide (GlcCer)--the common precursor of the different series of glycosphingolipids-that is operated by the cytosolic GlcCer-transfer protein FAPP2 (also known as PLEKHA8) (ref. 1). However, the molecular determinants of the FAPP2-mediated transfer of GlcCer from the cis-Golgi to the trans-Golgi network, as well as the physiological relevance of maintaining two parallel transport pathways of GlcCer--vesicular and non-vesicular--through the Golgi, remain poorly defined. Here, using mouse and cell models, we clarify the molecular mechanisms underlying the intra-Golgi vectorial transfer of GlcCer by FAPP2 and show that GlcCer is channelled by vesicular and non-vesicular transport to two topologically distinct glycosylation tracks in the Golgi cisternae and the trans-Golgi network, respectively. Our results indicate that the transport modality across the Golgi complex is a key determinant for the glycosylation pattern of a cargo and establish a new paradigm for the branching of the glycosphingolipid synthetic pathway.

We have prepared acyl chain-defined D-erythro-sphingomyelins and D-erythro-dihydrosphingomyelins and compared their properties in monolayer and bilayer membranes. Surface pressure/molecular area isotherms of D-erythro-N-16:0-sphingomyelin (16:0-SM) and D-erythro-N-16:0-dihydrosphingomyelin (16:0-DHSM) show very similar packing properties, except that the expanded-to-condensed phase transition (crystallization) occurs at a lower surface pressure for 16:0-DHSM. The measured surface potential was generally about 100 mV less for 16:0-DHSM monolayers compared to 16:0-SM monolayers. The condensed domains (crystals) that formed in 16:0-SM monolayers as a function of compression displayed star-shaped morphology when viewed under an epifluorescence microscope. 16:0-DHSM monolayers did not form similar crystals upon compression. 16:0-DHSM was degraded much faster by sphingomyelinase from Staphylococcus aureus than 16:0-SM (10-fold difference in enzyme activity needed for comparable hydrolytic rate). Cholesterol desorption from 16:0-DHSM to cyclodextrin was slightly slower (approximately 20%) than the rate measured from 16:0-SM monolayers (at 60 mol % cholesterol). The bilayer melting temperature of 16:0-DHSM was 47.7 degrees C (DeltaH 8.3 kcal/mol) whereas it was 41.2 degrees C for 16:0-SM (DeltaH 8.1 kcal/mol). Cholesterol/16:0-DHSM bilayers (15 mol % sterol) had more condensed domains than comparable 16:0-SM bilayers, as evidenced from the quenching resistance of DPH in DHSM membranes. We conclude that cholesterol interacts more favorably with 16:0-DHSM and that the membranes are more condensed than comparable 16:0-SM-containing membranes.

We recently showed that transport of ergosterol from the ER (endoplasmic reticulum) to the sterol-enriched PM (plasma membrane) in yeast occurs by a non-vesicular (Sec18p-independent) mechanism that results in the equilibration of sterol pools in the two organelles [Baumann, Sullivan, Ohvo-Rekilä, Simonot, Pottekat, Klaassen, Beh and Menon (2005) Biochemistry 44, 5816-5826]. To explore how this occurs, we tested the role of proteins that might act as sterol transporters. We chose to study oxysterol-binding protein homologues (Osh proteins), a family of seven proteins in yeast, all of which contain a putative sterol-binding pocket. Recent structural analyses of one of the Osh proteins [Im, Raychaudhuri, Prinz and Hurley (2005) Nature (London) 437, 154-158] suggested a possible transport cycle in which Osh proteins could act to equilibrate ER and PM pools of sterol. Our results indicate that the transport of newly synthesized ergosterol from the ER to the PM in an OSH deletion mutant lacking all seven Osh proteins is slowed only 5-fold relative to the isogenic wild-type strain. Our results suggest that the Osh proteins are not sterol transporters themselves, but affect sterol transport in vivo indirectly by affecting the ability of the PM to sequester sterols.