GPI anchoring leads to sphingolipid-dependent retention of endocytosed proteins in the recycling endosomal compartment

Abstract

Glycosylphosphatidylinositol (GPI) anchoring is important for the function of several proteins in the context of their membrane trafficking pathways. We have shown previously that endocytosed GPI-anchored proteins (GPI-APs) are recycled to the plasma membrane three times more slowly than other membrane components. Recently, we found that GPI-APs are delivered to endocytic organelles, devoid of markers of the clathrin-mediated pathway, prior to their delivery to a common recycling endosomal compartment (REC). Here we show that the rate-limiting step in the recycling of GPI-APs is their slow exit from the REC; replacement of the GPI anchor with a transmembrane protein sequence abolishes retention in this compartment. Depletion of endogenous sphingolipid levels using sphingolipid synthesis inhibitors or in a sphingolipid-synthesis mutant cell line specifically enhances the rate of endocytic recycling of GPI-APs to that of other membrane components. We have shown previously that endocytic retention of GPI-APs is also relieved by cholesterol depletion. These findings strongly suggest that functional retention of GPI-APs in the REC occurs via their association with sphingolipid and cholesterol-enriched sorting platforms or 'rafts'.

Fig. 1. Trafficking kinetics of FR-GPI measured in endocytic compartments of CHO cells. FRαTb-1 cells were incubated with PLR to determine the extent of accumulation of endocytosed FR-GPI in the peripheral GEECs (arrows) and the pericentriolar REC (arrowheads) after a 20 (A) or 60 min (B) pulse of the label (see Materials and methods). (A) and (B) show gray-scale intensities of endocytosed PLR normalized to surface FR-GPI expression. The bar shows a gradient of gray-scale intensities with the lowest (black) at the bottom of the bar to the highest (white) at the top of the bar. (C and D) REC-localized Cy5-Tf images of the same cells in (A) and (B), obtained as described in Materials and methods. The REC is marked with a white boundary for two representative cells. Scale bar represents 5 µm. (E) Change in surface receptor-normalized fluorescence in intracellular compartments plotted against time of incubation for the total endosomal FR-GPI (open squares), FR-GPI in REC (open triangle) and FR-GPI in GEECs (filled circles). The lines represent the best fit for the data points using the kinetic equation described in Materials and methods. Note that GEECs reach a steady state of PLR labeling much more quickly than the REC or the total endosomal receptors. (F) Rate constant of exit (k_e) from the indicated compartments.

Fig. 2. Retention of FR-GPI in the REC requires the GPI anchor. FRαTb-1 (A) and FR-TMTb-1 (B) cells expressing FR-GPI and FR-TM, respectively, were labeled with PLR and Cy5-Tf for the indicated times and imaged using a wide-field microscope. Pseudo-colored merges of PLR (encoded green) and Cy5-Tf (encoded red) images are shown in the extreme right-hand panel (‘Merge’). Images at the two time points in (A) and (B) were selected for matching total TfR expression and the endosomal PLR fluorescence was normalized to surface receptor expression. Note the change in PLR fluorescence in the REC (arrowheads) compared with that in peripheral endosomes in (A) (arrows indicate GEECs) or (B) (arrows indicate sorting endosomes) between the 20 and 40 min time points. Scale bar represents 10 µm.

Fig. 3. Schematic of the sphingolipid biosynthesis pathway. The sites of action of the various sphingolipid inhibitors used in this study and the location of the sphingolipid synthesis block in the mutant cell line, LY-B, are indicated by italic text.

Fig. 4. Sphingolipid depletion increases the rate of recycling of FR-GPI. Fluorescence associated with internalized FR-GPI bound to PLF at indicated times was normalized to surface receptor levels and plotted against time in untreated (control; filled circles) and FB₁-treated (open circles) FRαTb-1 cells (A and B) and in FB₁-treated FRαTb-1 cells supplemented with exogenous N-hexanoyl-d-sphingosine (C₆-cer; open squares) and N-palmitoyl-dl-erythro-dihydrosphingosine (NPDS; open triangles). Fluorescence associated with internalized PLF bound to FR-GPI at the indicated times was normalized to surface receptor levels and plotted against time in the FR-GPI-expressing LY-B cell line (C) grown in sphingolipid-deficient (Sph–; open circles) or sphingolipid-sufficient medium (Sph+; filled circles). The solid lines in (A)–(C) are the best fit for the data points using the kinetic equation described in Materials and methods. (D) Rate constants of FR-GPI recycling (k_e) under various treatment conditions. Data are derived from two to six independent experiments for each rate constant determination.

Fig. 5. Rate of recycling of DAF in sphingolipid-depleted cells. (A) DAF-associated fluorescence intensity at the indicated times was normalized to surface receptors and plotted against time in untreated (control; filled circles) and FB₁-treated (FB₁; open circles) cells. The solid lines represent the best fit for the data points using the kinetic equation described in Materials and methods. (B) Rate constants for DAF recycling in treated and untreated cells.

Fig. 6. Endocytic pathway of FR-GPI and TfR in sphingolipid-depleted cells. FRαTb-1 cells were incubated with PLF and Alexa₅₆₈-Tf to steady state, and images of FR-GPI (A, C and E) and TfR (B, D and F) were collected using a wide-field microscope for untreated (A and B), FB₁-treated (C and D) and C₆-ceramide-replenished FB₁-treated cells (E and F). Note that the majority of the intracellular FR-GPI is extensively co-localized with TfR in the REC (arrows). Scale bar represents 10 µm.

Fig. 7. Sphingolipid depletion does not affect the rate of TfR recycling. (A) Internalized Cy3-Tf fluorescence bound to TfR (circles) or δ3–59TfR (triangles) at indicated times was normalized to surface receptor expression and plotted against time in untreated (control; filled symbols) and FB₁-treated (FB₁; open symbols) cells, in FRαTb-1 (TfR) and δ3–59TRVb (δ3–59TfR) cells, respectively. (B) Internalized Cy5-Tf fluorescence bound to TfR at the indicated times was normalized to surface receptor expression and plotted against time in LY-B cells grown in sphingolipid-deficient (open circle; Sph–) or sphingolipid-sufficient medium (filled circle; Sph+). The solid lines in (A) and (B) represent the best fit for the data points using the kinetic equation described in Materials and methods.