Unrestricted Diffusion of Exogenous and Endogenous PIP₂ in Baby Hamster Kidney and Chinese Hamster Ovary Cell Plasmalemma

Abstract

We used two approaches to characterize the lateral mobility of phosphatidylinositol 4,5-bisphosphate (PIP₂) in the plasmalemma of baby hamster kidney and Chinese hamster ovary fibroblasts. First, nitrobenzoxadiazole-labeled C6-phosphatidylcholine and C16-PIP₂ were incorporated into plasma membrane “lawns” (∼20 × 30 μm) from these cells and into the outer monolayer of intact cells. Diffusion coefficients determined by fluorescence recovery after photobleaching were similar for the two lipids and were higher in lawns, ∼0.3 μm²/s, than on the cell surface, ∼0.1 μm²/s. For membrane lawns, the fractional recoveries (75–90%) were close to those expected from the fraction of total membrane bleached, and labeling by the probes was several times greater than for intact cells. Second, we analyzed cells expressing M1 muscarinic receptors and green fluorescent protein fused with PIP₂-binding pleckstrin-homology domains, Tubby domains or diacylglycerol (DAG)-binding C1 domains. On-cell gigaseal patches were formed with pipette tips >5 μm in diameter. When the agonist carbachol (0.3 mm) was applied either within or outside of the pipette, lipid signals crossed the pipette barrier rapidly in both directions and membrane blebbing occurred on both membrane sides. Accurate simulations of lipid gradients required diffusion coefficients >1 μm²/s. Exogenous DAG also crossed the pipette barrier rapidly. In summary, we found no evidence for restricted diffusion of signaling lipids in these cells. The lower mobility and incorporation of phospholipid at the extracellular leaflet may reflect a more ordered and condensed extracellular monolayer, as expected from previous studies.

Appendix

Simulations of PIP₃ and PIP₂ Diffusion

The most analyzed example of spatially organized phospholipid signaling is the PIP₃ gradient that develops from the leading to the trailing edge of migrating cells (Devreotes & Janetopoulos, 2003; Insall & Weiner, 2001). As underscored by simulations here, the existence of the PIP₃ gradients requires no restriction of lipid diffusion, and the time courses with which gradients develop are accounted for very well by phospholipid diffusion coefficients, as expected for simple bilayers with no restrictions. For the simulation illustrated in Figure 11, it is assumed that lipid kinases (PI3-kinases) are localized to 1% of total cell membrane at the left pole of the cell, while PIP₃ phosphatases (PTEN) are localized to 1% of the cell membrane at the right cell pole with enough activity to hydrolyze most of the cell’s PIP₃ within 2 min. With a diffusion coefficient expected for simple phospholipid bilayers, 1–4 μm²/s (Jacobson, 1983), the average PIP₃ molecule requires a good minute to traverse the cell length (20 μm) from its site of synthesis to its site of degradation. Gradients develop with apparent time constants of 30–60 s when lipid kinase activity is increased in a stepwise fashion, and these times are consistent with experimental data and previous simulations (Ma et al., 2004).

Fig. 11

Simulation of PIP₃ gradients in a cell assumed to be spherical and 20 μm in diameter with lipid kinases located to 1% of the cell area at one pole (PIP ₃ synthesis) and lipid phosphatases located to 1% of cell area at the opposite pole (PIP ₃ degradation). Degradation rate is 0.1/s within the degradation pole, and the diffusion coefficient is 2 μm²/s. The cell contains initially no PIP₃, and PIP₃ synthesis is initiated at time 0

As outlined in the “Introduction,” it is proposed that PIP₂ depletion during PLC activation can occur on a nanometer scale so as to mediate signals from specific receptors to specific ion channels (Cho et al., 2005a, 2005b). Figure 12 illustrates the drastic extent to which lipid mobility would have to be restricted to explain such highly localized signals. Figure 12a presents simulation of a 1-μ-diameter disk of membrane with 50,000 PIP₂ molecules (i.e., about 1% of total phospholipid in the cytoplasmic leaflet), as expected for the plasma membrane of most mammalian cells (e.g., Nasuhoglu et al., 2002). A single PLC is assumed to exist at the center of the membrane disk. Its activity is assumed to be 1,000 s⁻¹, the maximal rate that we can project from biochemical studies (Smrcka & Sternweis, 1993). When the PLC is activated, PIP₂ is depleted uniformly across the membrane disk. Similarly, PIP₂ gradients remain negligible when the activity is simulated in an infinite sheet of membrane (simulation not shown).

Fig. 12

Simulated PIP₂ gradients generated by a single PLC degrading PIP₂ at a rate of 1,000 s⁻¹. The PIP₂ density is assumed to be 50,000/μ² in steady state, equivalent to approximately 1% of the cytoplasmic monolayer; and the diffusion coefficient is 2 μm²/s. (a) Disk model. The single PLC begins to cleave PIP₂ at the tenth second. PIP₂ densities remain nearly homogeneous as PIP₂ is cleaved from the disk. (b) Infinite sheet model. Activity of a single PLC is initiated at 10 s and terminated at 150 s in an infinite sheet of membrane. Everywhere within the sheet, PIP₂ is assumed to be simultaneously and homogeneously degraded at 0.03/s and synthesized at 1,500 molecules s⁻¹ μm⁻². Results are given for the point of PIP₂ hydrolysis and for radii spaced at 0.5-μm intervals from the PLC activity. The approximately 6% depletion at the point of hydrolysis dissipates laterally with a space constant of 1 μm

In this same context, one may consider whether continuous metabolism of PIP₂ by phosphatases and lipid kinases might allow larger and/or more local gradients to develop with PLC activation. Figure 12b illustrates the PIP₂ gradients that are generated by the same punctate PLC activity in a membrane in which it is assumed that PIP₂ is generated and destroyed everywhere by lipid kinases and phosphatases so that the average PIP₂ lifetime is about 30 s. These assumptions are in fact required to reconstruct signals described subsequently for PH domains in intact cells. Activation of PLC with an activity of 1,000 s⁻¹ generates a gradient with a length constant of about 1 μm with these assumptions and with a peak magnitude of 6% of the background PIP₂. As pointed out by Cho et al. (2005a), for significant PIP₂ signals to occur between neighboring proteins, lipid diffusion must be restricted more than 100-fold, probably 1,000-fold.