Signaling-dependent immobilization of acylated proteins in the inner monolayer of the plasma membrane

Parts used

Key findings

• PM-GFP, an EGFP fused to the 10-amino-acid N-terminal myristoylation and palmitoylation sequence of Lyn, localizes mainly to the inner monolayer of the plasma membrane in RAW264.7 macrophages, with a variable amount of endomembrane signal that depends on expression level. In unstimulated cells at 37 °C, photobleached PM-GFP recovers almost completely (mobile fraction 1.10 ± 0.04, half-time 15 ± 2 s, n = 8), corresponding to a diffusion coefficient of 1.5 × 10^-10 cm²/s.
• Diacylated PM-GFP diffuses in the RAW264.7 macrophage plasma membrane at 1.5 × 10^-10 cm²/s, close to the 2.4 × 10^-10 cm²/s of NBD-phosphatidylserine and 2.2 × 10^-10 cm²/s of BODIPY-phosphatidylethanolamine, and nearly three orders of magnitude below the 9 × 10^-7 cm²/s reported for GFP in water. These values imply that acyl-chain interactions within the bilayer, not drag from the GFP moiety, set the mobility of a Lyn-anchored GFP.
• At the phagocytic cup formed around IgG-opsonized 8.3-µm beads on RAW264.7 macrophages, the mobile fraction of PM-GFP (EGFP carrying the Lyn N-terminal acylation sequence) falls to 0.36 ± 0.01 (n = 8), compared with 0.8–1 in unengaged membrane of the same cell. The fraction that remains mobile recovers at the same rate as PM-GFP in the bulk membrane.
• GFP-tH, a GFP fused to the C-terminal 9 amino acids of H-Ras (prenylated and doubly acylated, the complete H-Ras membrane-targeting domain), targets almost exclusively to the plasma membrane of RAW264.7 macrophages. In unengaged membrane it recovers rapidly after photobleaching (half-time 10 ± 0.5 s, mobile fraction 0.78 ± 0.01, n = 8), while at the phagocytic cup its mobile fraction falls to 0.37 ± 0.01 (n = 8).
• Two unrelated inner-leaflet acyl anchors, the Lyn N-terminal 10 amino acids (PM-GFP) and the H-Ras C-terminal 9 amino acids (GFP-tH), are both immobilized at the macrophage phagocytic cup. Because the H-Ras C terminus is not expected to bind Fc receptors, direct association of the Lyn sequence with the receptor complex is unlikely to explain the immobilization.
• GFP anchored to the outer monolayer through a GPI motif (GPI-GFP or GPI-YFP, made from the 26-amino-acid insulin signal sequence, the fluorescent protein and the 43-amino-acid GPI motif of decay-accelerating factor) is mobile in resting RAW264.7 macrophages (mobile fraction 0.81 ± 0.04, n = 8). Its mobile fraction and diffusion rate at the phagocytic cup are virtually identical to those in unengaged membrane, unlike the inner-leaflet probes PM-GFP and GFP-tH.
• GFP-tK, a GFP fused to the C-terminal 17 amino acids of K-Ras (a CAAX prenylation box plus a polycationic sequence that directs the probe to anionic lipids of the cytosolic face of the plasma membrane), is highly mobile in RAW264.7 macrophages in unengaged membrane (mobile fraction 1.13 ± 0.03, half-time 9 ± 1 s, n = 6). At the phagocytic cup it is only marginally slowed (mobile fraction 0.90 ± 0.03, half-time 7.8 ± 1.4 s, n = 6), in contrast to the diacylated PM-GFP and GFP-tH.
• Depleting cholesterol with 10 mM methyl-β-cyclodextrin for 30 min at 37 °C removes 50% of total cellular cholesterol from RAW264.7 macrophages and leaves the immobilization of PM-GFP at the phagocytic cup intact (mobile fraction 0.44 ± 0.02, n = 7). In unstimulated cells the same treatment alters the PM-GFP mobile fraction (1.27 ± 0.1) and half-time (23 ± 4 s) only marginally, and it does not impair ingestion of opsonized beads.
• The raft-associated transmembrane adaptor LAT, fused to GFP (LAT-GFP) and expressed in RAW264.7 macrophages, localizes to the plasma membrane and shows the same mobility at the phagocytic cup as elsewhere in the membrane. Together with the persistence of PM-GFP immobilization after cholesterol extraction and the unchanged mobility of GPI-YFP, this indicates that cholesterol-enriched rafts are not noticeably immobilized near engaged Fcγ receptors.
• Preventing actin polymerization with 2 µM cytochalasin D allows cup formation around IgG-opsonized beads but does not abolish the reduced mobility of PM-GFP at the cup, and does not change PM-GFP mobility in bulk membrane (mobile fraction 0.84). In sealed phagosomes of 3.1-µm beads, where actin has dissociated from the membrane, PM-GFP stays largely immobile (mobile fraction 0.26 ± 0.02) compared with the bulk plasma membrane (0.9 ± 0.02). Corralling by the actin cytoskeleton is therefore unlikely to cause the immobilization.
• Inhibiting Src-family kinases with 10 µM PP1 for 1 h prevents the immobilization of PM-GFP at the phagocytic cup of RAW264.7 macrophages (mobile fraction 0.76 ± 0.03 at the cup versus 0.96 ± 0.03 in unengaged membrane, n = 8) and leaves GPI-YFP mobility unaltered. PP1 does not prevent cup formation but virtually completely inhibits particle engulfment, so the immobilization of acylated inner-leaflet probes depends on tyrosine-kinase signaling and not on receptor clustering alone.
• Inhibiting phosphatidylinositol 3-kinase with 100 µM LY294002 for 30 min arrests phagosome development after cup formation and prevents accumulation of 3′-phosphorylated phosphoinositides at the cup, yet PM-GFP remains immobilized there (mobile fraction 0.46) and the exofacial GPI-YFP is unaffected. Generation of 3′-phosphorylated inositides is therefore not required to retain acylated molecules in the inner leaflet of the phagocytic cup.
• During phagocytosis of large beads by RAW264.7 macrophages, membrane remodeling begins before the cup seals, and up to 40% of PM-GFP or GPI-anchored YFP fluorescence is lost from the base of the cup within 3 min. Photobleaching experiments are therefore limited to the first 100 s after bead engagement, and because GPI-YFP, which is remodeled at a similar rate, and GFP-tK recover to a much greater extent, endocytosis does not account for the incomplete recovery of PM-GFP and GFP-tH.

Abstract

Phospholipids play a critical role in the recruitment and activation of several adaptors and effectors during phagocytosis. Changes in lipid metabolism during phagocytosis are restricted to the phagocytic cup, the area of the plasmalemma lining the target particle. It is unclear how specific lipids and lipid-associated molecules are prevented from diffusing away from the cup during the course of phagocytosis, a process that often requires several minutes. We studied the mobility of lipid-associated proteins at the phagocytic cup by measuring fluorescence recovery after photobleaching. Lipid-anchored (diacylated) fluorescent proteins were freely mobile in the unstimulated membrane, but their mobility was severely restricted at sites of phagocytosis. Only probes anchored to the inner monolayer displayed reduced mobility, whereas those attached to the outer monolayer were unaffected. The immobilization persisted after depletion of plasmalemmal cholesterol, ruling out a role of conventional "rafts." Corralling of the probes by the actin cytoskeleton was similarly discounted. Instead, the change in mobility required activation of tyrosine kinases. We suggest that signaling-dependent recruitment of adaptors and effectors with lipid binding domains generates an annulus of lipids with restricted mobility.