Reproduced with permission from EMBO Journal.

Over the years, many investigators have set out to look for a protein regulator of their favourite protein and instead found a phosphoinositide. Most labs have therefore developed at least a peripheral interest in this family of lipids. So they will be happy to read in the EMBO Journal that the laboratories of Harald Stenmark and Rob Parton have devised a new probe to study the cellular localization of phosphatidylinositol-3-phosphate (PtdIns(3)P).

The products of phosphatidylinositol-3-OH kinases function in processes as diverse as signal transduction, cytoskeletal organization and apoptosis. PtdIns(3)P is particularly interesting for membrane traffic aficionados, as it regulates transport along the endocytic pathway in all species where this has been studied. One of its activities is to recruit proteins that contain PtdIns(3)P-binding FYVE finger domains to membranes. But to which membranes?

Gillooly et al. reasoned that if PtdIns(3)P binds FYVE domains, then FYVE domains should bind PtdIns(3)P. They built a probe (2XFYVE) consisting of two FYVE domains from Hrs, a protein that acts in the endocytic pathway. A series of control experiments showed that 2XFYVE binds selectively to PtdIns(3)P, both in vivo and in vitro. The probe effectively competes with endogenous proteins for PtdIns(3)P binding when transfected into cells, and can be used for immunofluorescence as well as for immunoelectron microscopy studies.

The next step was to use 2XFYVE to localize PtdIns(3)P in the cell. The FYVE finger protein EEA1 is known to localize exclusively to early endosomes, where it is involved in membrane fusion. So it was clear from the start that there must be substantial amounts of PtdIns(3)P in the membrane of early endosomes. This was confirmed in this study — 2XFYVE, shown in red in the picture, colocalized extensively with EEA1, shown in green.

More surprisingly, immunoelectron microscopy using 2XFYVE revealed that PtdIns(3)P is also present on internal membranes of multivesicular late endosomes. This observation led the authors to speculate that the PtdIns(3)P-containing intralumenal vesicles arise from invagination of the endosomal membrane. This could sequester PtdIns(3)P away from the surface, stopping it from recruiting cytoplasmic proteins such as EEA1 to late endosomes. The origin of the convoluted morphology of multivesicular endosomes is still mysterious, and 2XFYVE might prove a useful tool to study this process.

This probe is not the first on the market. The pleckstrin homology (PH) domain of phospholipase Cδ1 and the PH domains of ARNO and Bruton's tyrosine kinase have been used to detect PtdIns(4,5)P2 and PtdIns(3,4,5)P3, respectively. But these probes cannot be used for electron microscopy, and neither of them shows the exquisite specificity for its target that 2XFYVE seems to have.