Background

Phosphoinositides, such as PI(4)P and PI(3)P, are phosphorylated derivatives of phosphatidylinositol, a phospholipid found in the cytoplasmic leaflet of eukaryotic membranes. The inositol ring of the phosphatidylinositol headgroup can be phosphorylated by specific kinases, and dephosphorylated by phosphatases, in position 3, 4, and 5 for a total of seven different phosphoinositides [1]. Phosphoinositides are key signaling lipids that regulate organelle membrane identity and dynamics. They notably play major roles in orchestrating various endolysosomal functions and dynamic events such as tubulation and fission, which are required for the reformation of lysosomes at the end of the autophagic and endocytic processes [2–5].

One elegant way to investigate the role of phosphoinositides in organelle biology is to modulate their levels at a given membrane. This can be done by specifically targeting phosphoinositide kinases or phosphatases to an organelle of interest. Targeting can be achieved by fusing the catalytic domain of kinases or phosphatases to a membrane-targeting domain (typically a transmembrane domain) of a resident membrane protein of the desired organelle. This allows for the constitutive depletion of the phosphoinositides upon the expression of the protein. However, to allow for temporal and acute regulation of the phosphoinositides species, the catalytic domain can be targeted using a chemically induced dimerization system. Here, one part of the dimer is anchored at a membrane of interest by fusing it to a transmembrane protein/targeting sequence, and the other is coupled to a cytosolic version of a phosphatase or kinase. Upon the addition of chemical inducers of dimerization, the binding affinity between the two-part dimerization system increases drastically, promoting the recruitment of the cytosolic part to the membrane-anchored part of the dimer. As a result, the recruited kinases and phosphatases actively phosphorylate and dephosphorylate the phosphoinositide substrate, respectively, depleting them from the organellar membrane.

In addition, the level of specific phosphoinositides at membranes in cells can be evaluated by the ectopic expression of genetically encoded phosphoinositides biosensors coupled to a fluorescent protein, such as GFP or mCherry. Over the last 20 years, high-affinity binding domains for specific phosphoinositides have been identified and used as biosensors to detect different phosphoinositide species [6]. Here, we present a protocol utilizing chemically induced dimerization systems to deplete specific phosphoinositides at endolysosomes and biosensors to detect their depletion. We describe how to use two different dimerization systems in cells, the GAI-GID1 and the FRB-FKBP, to deplete PI(4)P and PI(3)P, respectively, at Rab7-positive endolysosomes (hereafter referred to as endolysosomes). However, any of the two dimerization systems can be used to deplete the phosphoinositides on any other membrane-bound organelle of interest. In this protocol, we discuss the experimental setup to express these systems in mammalian cells and how to perform image acquisition and analysis to validate the depletion of PI(4)P or PI(3)P.

The FRB-FKBP dimerization system was invented in the 90’s [7] and relies on the dimerization of FRB and FKBP in the presence of rapamycin, a cell-permeant chemical most famous for its inhibitory effect on mTORC1 (mammalian target of rapamycin complex 1). The FRB and FKBP domains are relatively small (~10 kDa) and thus can be efficiently transfected and expressed, reducing the likeliness of impacting the function of the protein of interest fused to these domains. However, using rapamycin in the FRB-FKBP dimerization system may cause undesirable changes in certain mammalian biological processes such as autophagy and nutrient-sensing pathways, notably due to its effect on mTORC1. To circumvent this problem, rapalogs (rapamycin analogs) can be used to alleviate this cross-reactivity [8]. Another way to mitigate the complications of rapamycin is to use the plant-derived GA₃-AM-inducible GAI-GID1 dimerization system since there is no known mammalian target of GA₃-AM [9]. However, unlike the FRB-FKBP system, both GAI and GID1 are medium-sized protein domains (~59 and ~40 kDa, respectively) which may impede the efficient introduction or expression of proteins. Control experiments can then be performed to validate that tagging with dimerizing proteins does not impair the localization or the function of the protein of interest. Here, we present a protocol to target PI(4)P or PI(3)P phosphatases to endolysosomes and to validate phosphoinositides depletion using genetically encoded PI(4)P or PI(3)P biosensors.