Acute depletion of plasma membrane phospholipids—dissecting the roles of PtdIns(4)P and PtdIns(4,5)P2

Autor: Natali Fili, Banafshé Larijani, Nirmal Jethwa
Rok vydání: 2012
Předmět:
Zdroj: Journal of Chemical Biology. 5:137-139
ISSN: 1864-6166
1864-6158
Popis: The study of signalling pathways and metabolic networks has long been assisted by tools which manipulate proteins: ectopic gene expression; gene knockouts or specific mutations; knockdown of expression levels by RNAi; or reductions in activity by chemical inhibitors. However, the relationship between a protein and its metabolite products (and the eventual biological effect) is complex. Biological membranes are typically composed of phospholipids which are synthesised, interconverted and recycled by a network of lipid metabolic enzymes [13]. These enzymes display a relative promiscuity for their substrates—both in terms of the headgroup (for example, MTM1 can dephosphorylate the 3-phosphate of PtdIns(3)P and PtdIns(3,5)P2) and the exact molecular species (the length, degree of saturation and combination of the fatty acyl chains). As a result, the potential pool of phospholipid species is incredibly diverse. As with any robust network, the overexpression or knockdown of an individual node results in an altered flux through other pathways to return the system to its ‘ideal’ state—such a global approach would interfere with multiple pathways and therefore may not have a predictable effect. A set of tools based on inducible protein dimerisation has been invaluable in dissecting the roles of various phospholipids in a number of biological contexts. Inducible dimerisation tools have two important advantages when studying phospholipid metabolism: (a) temporal control and (b) spatial control over lipid depletion. The tools typically consist of two parts—an anchor tied down to a defined membrane compartment and an active soluble enzyme, both tagged with protein domains which can dimerise in response to the binding of a drug [2] or light-induced conformational changes [8]. Upon dimerisation, the enzyme is recruited to the anchor, resulting in the acute depletion of a phospholipid from a specific compartment. The tool used by Hammond et al. [6] is based on two endogenous mammalian protein domains—FKBP (FK506 binding protein) and FRB (a fragment of the PI3K homolog FRAP/mTOR). The dimerisation of these domains was described by Chen et al. [2] and characterised structurally by Choi et al. [3] and has been used since to induce receptor dimerisation [11], to force Akt translocation to the plasma membrane [9], to control translation [12] and protein splicing [10], and to alter glycosylation [4]. It was commercialised by Ariad Pharmaceuticals and is now available from Clontech as the iDimerize Inducible Heterodimer System. It was first used to alter phosphoinositide metabolism by Fili et al. [5], who showed that acute depletion of PtdIns(3)P and PtdIns(3,5)P2 from early endosomes leads to a failure of this compartment to mature, eventually causing tubularisation of the endosomal network, i.e. an altered lipid composition can be sufficient to induce gross morphological changes in a membrane compartment. Hammond et al. [6] have used the rapalogue-induced dimerisation tool to investigate the roles of PtdIns(4)P and PtdIns(4,5)P2 at the plasma membrane. Using wild-type and mutant tandem chimeras of the phosphatase domains from the inositol polyphosphate 5-phosphatase E (which converts PtdIns(4,5)P2 to PtdIns(4)P) and the Sac1 phosphatase (which dephosphorylates PtdIns(4)P to PtdIns), they were able to independently manipulate the levels of these two lipids. Depletion of PtdIns(4)P does not have an effect on the steady state level of PtdIns(4,5)P2 as observed by PH-PLCd1 localisation, on PtdIns(4,5)P2-dependent processes like clathrin-mediated endocytosis or on the recovery of the PtdIns(4,5)P2 pool following heavy phospholipase C activation. This was a surprising result, given that PtdIns(4,5)P2 is replenished by 5-phosphorylation of PtdIns(4)P. It is possible that the PtdIns(4,5)P2 needed for prolonged signalling is provided by trafficking from other compartments or else it is directly produced as needed by a highly efficient PtdIns-5-kinase activity. The other possibility, not discussed here, is an increased flux through PtdIns(5)P and PtdIns-4 kinase activity to produce PtdIns(4,5)P2. However, without a pulse-chase experiment, it is not possible to know how the PtdIns(4,5)P2 is being replenished. The authors then observed the localisation of a series of short peptides taken from plasma membrane binding proteins, showing that both PtdIns(4)P and PtdIns(4,5)P2 were required for membrane localisation (unless an alternative targeting method existed, like the K-Ras prenylated tail). They also observed differential dependencies of selected ion channels for the phosphoinositides—TRPM8 requires PtdIns(4,5)P2 while TRPV1 is functional with either PtdIns(4)P or PtdIns(4,5)P2. The authors speculate that both PtdIns(4)P and PtdIns(4,5)P2 contribute to the negative charge of the plasma membrane inner leaflet; the PtdIns(4)P is acting as a general pool of polyvalent anions while PtdIns(4,5)P2 is undergoing rapid turnover or being masked by binding proteins. However, given that proteins associated with the membrane probably have both a general electrostatic requirement and a specific lipid requirement, it would also be informative to couple Sac1 phosphatase with a PtdIns 3-kinase or a 5-kinase—would the recruitment of membrane binding proteins be restored by a pool of phosphoinositides with the same charge but different structure? On a related note, a control experiment to manipulate the membrane charge by another method would definitively reveal the electrostatic contribution to protein binding—for example, by depleting negatively charged PtdSer from the plasma membrane (which should have a similar effect to depleting PtdIns(4)P) or by generating PtdSer in another compartment like the endoplasmic reticulum (whose large internal surface area should hopefully compete with the plasma membrane for electrostatic binding). Confirmation of the change in plasma membrane charge is also omitted—either by in vivo measurements of the membrane potential after lipid depletion or by modelling. Hammond et al. [6] have effectively shown that they have developed a rapamycin-inducible dimerisation tool which is able to successfully deplete PtdIns(4)P and/or PtdIns(4,5)P2 from the plasma membrane. The strength of the tool is that it allows the specific, acute depletion of lipid species from a defined compartment. However, the readout of lipid depletion is often based qualitatively on the binding of a PH domain. These are either transfected into cells, which creates problems with masking and sequestration [1], or used as recombinant probes, where the fixation conditions can profoundly alter the identity and localisation of the phospholipids [7]. Here, the authors have quantified the fluorescence intensity of PH domains, after confirming by lipid mass spectrometry that the changes in fluorescence correspond to changes in the mass of PtdIns(4)P and PtdIns(4,5)P2; this approach strikes a good balance between the most rigorous method of quantification and ease of data acquisition and analysis. The future of this work involves a number of difficult questions: how are the different pools of phosphoinositides created, maintained, trafficked and metabolised? How are the biophysical properties and functions of a membrane defined by small phosphoinositide populations? Certainly new tools to precisely manipulate these lipids spatially and temporally will help; interdisciplinary research will also be vital to probe these questions in different contexts and to gain a deeper understanding of the turnover of phospholipids in different compartments and the relationship between membrane morphology, dynamics and downstream signalling.
Databáze: OpenAIRE
Externí odkaz: