Phosphatidylinositol Synthase and Diacylglycerol Platforms Bust a Move

Summary

Kim et al. (2011) challenge the dogma that phosphatidylinositol synthesis is restricted to the endoplasmic reticulum (ER) by showing that a mobile membrane compartment transports phosphatidylinositol synthase from the ER to numerous cellular compartments, including the plasma membrane. These findings significantly impact our view of phosphoinositide signaling in the cell.

Phosphatidylinositol is an essential phospholipid in eukaryotes, in part because of its requirement for production of both phosphoinositides and soluble inositol phosphates (Sasaki et al., 2009). These phosphorylated derivatives of phosphatidylinositol and inositol, respectively, are critical signaling molecules, and many intracellular systems receive regulatory input from them. Phosphatidylinositol biosynthesis is catalyzed by a single enzyme, phosphatidylinositol synthase (PIS), which utilizes inositol and cytidine-diphosphate (CDP)-diacylglycerol as substrates to produce phosphatidylinositol and cytidine-monophosphate. PIS is an integral membrane protein of the ER (Paulus and Kennedy, 1960; Antonsson et al., 1997).

PIS and phospholipases C participate in phosphoinositide-dependent growth factor-, G-protein-coupled receptor, and calcium signaling pathways. For purposes of this discussion, phospholipases C are active at the plasma membrane and hydrolyze the phosphoinositides derived from the phosphatidylinositol produced by PIS (Figure 1A). Thus, PIS resides in an organelle remote from sites of phospholipase C-mediated phosphoinositide signaling. It has been assumed that transfer-proteins mediate phosphatidylinositol trafficking between the ER and the plasma membrane to support phosphoinositide signaling (Michell, 1975). In this issue, Kim et al (2011) describe a mobile PIS-containing membrane compartment that supports generation of a phosphatidylinositol ‘signaling pool’ at the plasma membrane, and propose a new model for delivery of PIS to sites of phospholipase C action.

Figure 1.

The shifting landscape of phosphoinositide signaling. (A) Phosphatidylinositol (PtdIns) is phosphorylated at the plasma membrane by 4-OH kinase (1) and 4-phosphate 5-OH kinase (2) to generate PtdIns(4,5)P₂. Phospholipase C (PLC) cleaves PtdIns(4,5)P₂ to generate inositol-trisphosphate (IP₃) and diacylglycerol (DAG). (3) Diacylglycerol is converted to phosphatidic acid (PtdOH), and (4) phosphatidic acid is trafficked (via transfer proteins?) to ER where it is converted to CDP-DAG. PIS consumes CDP-DAG to generate a PtdIns pool (5) that is transported by phosphatidylinositol transfer proteins to plasma membrane. This cycle of phosphoinositide signaling and phosphatidylinositol synthesis is adapted from Michell (1975). (B) Kim et al (2011) report that phosphatidylinositol depletion mobilizes diacylglycerol via vesicular carriers (1) -- speculatively drawn as moving from plasma membrane to ER. A Sar1 GTPase-dependent cycle cooperates with PIS activity to generate a mobile platform loaded with CDP-diacylglycerol and phosphatidylinositol (2). The platform interacts with plasma membrane to which it delivers phosphatidylinositol (3). The exhausted PIS compartment recycles back to ER (4), perhaps serving as a vehicle for phosphatidic acid recycling to ER.

Phosphatidylinositol synthesis was linked with intracellular lipid trafficking by studies of phospholipase C-mediated phosphoinositide signaling at the plasma membrane (Michell, 1975). Those studies posed the question: how are phosphoinositides replenished at the plasma membrane in the face of robust consumption of these molecules by phospholipase C? The authors posited that stimulated phosphatidylinositol synthesis in ER generates a phosphatidylinositol pool earmarked for delivery to plasma membrane where it fuels phosphoinositide resynthesis (Figure 1A). How does the cell couple inositol phospholipid metabolic reactions executed in two discrete membranes? The model invoked a lipid trafficking cycle where soluble lipid-carrier proteins ferry either diacylglycerol or phosphatidic acid (produced by diacylglycerol kinases) from plasma membrane back to ER to fuel phosphatidylinositol synthesis. Phosphatidylinositol transfer proteins subsequently transport phosphatidylinositol from ER to plasma membrane (Figure 1A). Indeed, phosphatidylinositol transfer proteins are highly conserved and interpretations of their function borrow directly from this conjecture (Cockcroft and Carvou, 2007).

The veracity of the concept that phosphatidylinositol synthesis is restricted to ER has been questioned (Imai and Gershengorn, 1987; Kinney and Carman, 1990), primarily by biochemical evidence suggesting that ER and plasma membrane PIS activities exhibit distinct biochemical properties (Imai and Gershengorn, 1987). Moreover, there is a close temporal relationship between onset of agonist-stimulated phosphoinositide hydrolysis and phosphatidylinositol resynthesis, suggesting that the phosphatidylinositol ‘signaling’ pool is generated at the plasma membrane and not in ER. These sporadic challenges to the ‘ER only’ model failed to secure traction in the signaling field, however, and the ‘ER only’ concept for phosphatidylinositol synthesis enjoys dogma status.

Kim et al (2011) present new data on this issue by identifying functional pools of phosphatidylinositol in distinct endomembranes of mammalian cells. Their approach was to target bacterial phospholipase C (which hydrolyzes phosphatidylinositol but not phosphoinositides) to specific intracellular compartments for the purpose of inducing spatially-restricted phosphatidylinositol depletion. They monitored the consequences with a suite of imaging and metabolic labeling approaches. Unexpectedly, Kim et al. find phosphatidylinositol depletion evokes rapid deployment of PIS from ER via mobile vesicular compartments. These PIS-containing structures interact with other intracellular organelles, and define the phosphatidylinositol reservoirs that fuel plasma membrane phosphoinositide resynthesis. Moreover, Kim et al. observe that diacylglycerol is similarly packaged in mobile vesicular structures, but the diacylglycerol-containing vesicles appear to be physically distinct from the dynamic PIS-containing compartments. Although mobilization of both diacylglycerol and PIS might reflect an unusual response to a non-physiological stimulus, it seems more likely that phosphatidylinositol depletion simply increases the capacity of a physiological, low-level operation. Taken together, the data suggest that mammalian cells do not require transfer proteins to ferry newly synthesized phosphatidylinositol from ER to sites of phosphoinositide depletion. Rather, PIS incorporates into a mobile phosphatidylinositol-producing compartment. This portable platform distributes phosphatidylinositol throughout the cell for ‘on demand’ resupply of depleted intracellular membranes (Figure 1B).

The study of Kim et al (2011) raises many questions. What is the nature of the PIS-containing compartment? Is it a vesicle produced by an established vesicle-budding pathway (Kinney and Carman, 1990), or is it a tubulating sub-compartment of the ER? Is it a pre-existing structure whose loading with PIS is regulated, or is this a specialized compartment born of PIS recruitment into ER microdomains? Genesis of the PIS-containing structures requires a functional Sar GTPase cycle, which also regulates COPII-coated vesicle budding from ER (Jensen and Schekman, 2011). The relationship of the PIS-compartment to the COPII pathway for ER-derived vesicle budding remains unknown, however. Either way, the demonstration by Kim et al. that catalytically inactive PIS molecules are not recruited into these dynamic structures suggests that PIS activity contributes to formation of the mobile compartment. The mechanisms of PIS loading into this novel compartment will certainly be interesting. At issue is how cells sense the phosphatidylinositol deficit, and how sensing information is delivered to PIS and to a Sar GTPase-based machinery (Figure 1B)?

Are other protein and lipid cargos packaged into the mobile PIS-containing structures? Kim et al (2011) find neither of the two CDP-diacylglycerol synthase (CDS) isoforms in the PIS-containing compartment (Figure 1). Because the CDP-diacylglycerol produced by CDS1/2 is an essential substrate for phosphatidylinositol synthesis, it is not clear how any individual PIS-containing structure could support sustained phosphatidylinositol synthesis. While the imaging data do not exclude the possibility that a minor pool of CDS1/2 loads into the PIS-compartment, the data imply some mechanism must exist for active sorting of CDP-diacylglycerol into those structures (Figure 1B). In any event, it would seem that the phosphatidylinositol-synthetic capacity of any single PIS-compartment is limited.

How is phosphatidylinositol transferred from the PIS compartment to target membranes such as plasma membrane? Kim et al (2011) detect close apposition of the PIS-containing structures with plasma membrane using total internal reflection fluorescence microscopy, but fail to document direct fusion of the two membranes. ‘Kiss-and-run’ hemifusion is therefore an attractive mechanism for mobilizing phosphatidylinositol from the PIS-compartment to the plasma membrane (Figure 1B). Do phosphatidylinositol transfer proteins play a role in phospholipid-exchange reactions between the PIS-compartment and plasma membrane? Such a role appears unlikely given mounting evidence that phosphatidylinositol “transfer proteins” are in fact not transfer proteins, but scaffolds that facilitate efficient production of phosphoinositides (Bankaitis et al., 2010). Furthermore, mobilization from ER of a dedicated phosphatidylinositol-biosynthetic platform strikes directly at the heart of the principle assumptions upon which the lipid transfer protein concepts rest. As is so often the case, cells produce far more fascinating and intricate solutions to the problem than our simple minds imagine.