How can lipids cross the Rubicon?

Untitled Document


An everyday question in eukaryotic cell biology is how water-soluble molecules cross the membrane barriers of the elaborate, membranous architecture of the cell.

Membrane penetration depends on lipid solubility of molecules, but hydrophilic molecules, ions and even water itself require specific proteins called channels or transporters to help them cross biological membranes.

Interestingly, we ask the reverse question less frequently: How do lipid molecules cross the aqueous barrier that separates the membranes? The importance of this question, however, is increasingly recognized in light of recent developments in the field of nonvesicular lipid transfer.

It has been well documented that membranes of distinct organelles have unique lipid compositions. Most of the cells’ structural lipids are synthesized in the endoplasmic reticulum or are taken up from the cell exterior via specific receptors by endocytosis, and they need to reach their ultimate destinations.

Given the very intense vesicular trafficking between organelles, this was assumed to be the primary mechanism by which lipids move from one membrane compartment to another. However, several observations suggested that vesicular trafficking might not be the sole mechanism of lipid transfer and that, in fact, cells have to ensure that vesicular trafficking does not alter the lipid profiles of the various organelles.

One of the first indications of the requirement for a soluble protein for lipid transport was the discovery of the StAR protein that is needed for cholesterol to cross the intermembrane space of mitochondria . Other examples include the inability of cholesterol to be recycled from the lysosomes in Niemann-Pick disease type C, or the requirement for the CERT protein for the transport of ceramide from the endoplasmic reticulum to the Golgi. Other lines of research showed that soluble Sec14 proteins are needed for yeast to support transport or utilization of phosphatidylinositol and phosphatidylcholine.

A unique relationship exists between some of the soluble lipid-transfer proteins that carry out nonvesicular lipid transfer of phosphoinositides. The discovery that the presence of a phosphatidylinositol 4-phosphate (PI4P)-binding pleckstrin homology domain is a common feature of several proteins involved in lipid transport initially highlighted this relationship. Subsequent studies showed that pleckstrin homology domain interaction with PI4P helps to dock these lipid-transfer proteins to their target membranes, which happened to be the Golgi for OSBPs, CERT and FAPP2.

While this view is still valid, further research showed that some OSBP proteins (OSBP in mammals and Osh4 in yeast) can transport PI4P back from the Golgi to the endoplasmic reticulum, where the PI4P phosphatase Sac1 converts PI4P back to phosphatidylinositol. These findings suggest that PI4P phosphorylation in the Golgi and dephosphorylation in the endoplasmic reticulum maintain a PI4P gradient that is required for the efficient transport of cholesterol from the endoplasmic reticulum to the Golgi.

While this PI4P-gradient-energized transport phenomenon was described in the Golgi, some data suggest that it might be a more general principle governing lipid fluxes between other membranes. For example, in yeast, the Osh3 protein was found to enable the endoplasmic reticulum-bound Sac1 enzyme to dephosphorylate PI4P made by the plasma-membrane-localized STT4 phosphatidylinositol 4-kinase in endoplasmic reticulum-plasma membrane contact zones. These data can be interpreted as demonstrating the ability of Sac1 to act on plasma-membrane-localized PI4P in trans when Osh3 is present or as indicating that the Osh3 brings PI4P back to the endoplasmic reticulum to be acted upon by Sac1. We speculated that the latter case would be more consistent with the model suggested by the studies on cholesterol transport .

More studies will be needed to explore the universality of these models. However, it is already notable that, in either case, endoplasmic reticulum-synthesized phosphatidylinositol (the precursor of PI4P) has to reach the membrane where it is converted to PI4P (at the Golgi, plasma membrane or endosomes).

It long has been postulated that soluble transport proteins – PITPa, PITPβ or Nir2 (and Sec14 in yeast) – can either help transport phosphatidylinositol from the endoplasmic reticulum to their target membranes or help the various PI 4-kinases utilize phosphatidylinositol in their respective locations .

Schematic cartoon of a cell showing the contact zones formed between membranes of various organelles (red circles). Sterol transport from the endoplasmic reticulumto the Golgi was shown to be mediated by the OSBP protein, driven by a PtdIns4P gradient between the Golgi and the ER. OSBP transports PtdIns4P in the reversedirection in exchange for cholesterol. Hypothetical model of a PtdIns4P-driven lipid exchange at the ER-PM contact zones. Yeast studies showed that the ER-localizedSac1 enzyme can hydrolyze PtdIns4P generated in the plasma membrane by the PI4KA ortholog Stt4p in the presence of the OSH3 protein. The mammalian lipid transferprotein (LTPx) is still to be identified. It is possible that a different PI transfer protein (PITP) facilitates the transfer of PtdIns between the ER and the partner membranein the different contact zones.  

The small phosphatidylinositol transfer protein , PITPb, works at the Golgi, while the larger Nir2 protein was found both at the Golgi and in endoplasmic reticulum-plasma membrane contact zones in stimulated cells.

Deciphering the integration of these pathways will allow us better to understand how cells maintain their unique membrane compositions and how defects in nonvesicular lipid transfer contribute to human disease. 

Tamas Balla Tamas Balla ( is is a senior investigator at the National Institutes of Health’s Eunice Kennedy Shriver National Institute of Child Health and Human Development.