Acute manipulation of Golgi phosphoinositides to assess their importance in cellular trafficking and signaling

Abstract

Phosphoinositides are essential lipid regulators of trafficking and signaling pathways of all eukaryotic cells. Phosphatidylinositol 4-phosphate (PtdIns4P) is an intermediate in the synthesis of several important phosphoinositide species but also serves as a regulatory molecule in its own right. Phosphatidylinositol 4-kinases are most abundant in the Golgi but are also found in the plasma membrane and in endocytic compartments. To investigate the role of Golgi PtdIns4P in orchestrating trafficking events, we used a unique drug-inducible molecular approach to rapidly deplete PtdIns4P from Golgi membranes by a recruitable Sac1 phosphatase enzyme. The utility of the system was shown by the rapid loss of Golgi localization of PH domains known to bind PtdIns4P after Sac1 recruitment to the Golgi. Acute PtdIns4P depletion prevented the exit of cargo from the Golgi destined to both the plasma membrane and the late endosomes and led to the loss of some but not all clathrin adaptors from the Golgi membrane. Rapid PtdIns4P depletion in the Golgi also impaired but did not eliminate the replenishment of the plasma membrane PtdIns(4,5)P(2) during phospholipase C activation revealing a hitherto unrecognized contribution of Golgi PtdIns4P to this process. This unique approach will allow further studies on the role of phosphoinositides in endocytic compartments that have evaded detection using the conventional long-term manipulations of inositide kinase and phosphatase activities.

Fig. 1.

Rapamycin-induced elimination of PtdIns4P from Golgi membranes. (A) Principle of rapamycin-induced modification of membrane lipid composition. (B) Distribution of the Tgn38-FRB-CFP and cytosolic mRFP-FKBP-fused Sac1 phosphatase expressed in COS-7 cells. (C) Recruitment of the cytosolic Sac1 phosphatase to the Golgi decreases the localization of the PtdIns4P reporter, GFP-OSH1-PH domain. The decrease in the OSH1-PH domain intensity over the Golgi parallels the increase in the mRFP-FKBP12-Sac1 signal over the same area.

Fig. 2.

The effect of PtdIns4P elimination on VSVg trafficking in COS-7 cells. Cells were transfected with Tgn38-FRB (without a fluorescent tag) together with either the cytosolic mRFP-FKBP only (A) or the cytosolic mRFP-FKBP-Sac1 phosphatase (B and C) and the VSVg protein fused to photoactivable GFP. Cells were kept at 39.5 °C overnight and shifted to 20 °C the next day for 1 h. This allowed the refolded VSVg to enter and accumulate in the Golgi. Cells were then mounted on the microscope's heated stage (35 °C) and treated immediately with rapamycin to induce recruitment of the cytosolic constructs to the Golgi (A and C). Cells that showed Golgi recruitment (within 5–10 min of rapamycin) were selected and the Golgi area was photoactivated using a 405-nm laser line. The intensity change in the photoactivated VSVg was then followed in the GFP channel. VSVg trafficking was also recorded in cells expressing the Sac1 phosphatase but without recruitment to the Golgi (B). Representative images are shown at 0 min (after photoactivation) and at 50 min. The bar diagram (D) shows the summary calculated from seven cells expressing FKBP only or the recruitable Sac1 without recruitment, and from six cells in the case of recruited Sac1 phosphatase (means ± SEM).

Fig. 3.

The effect of PtdIns4P elimination on mannose 6-phosphate receptor trafficking in COS-7 cells. Cells were transfected with the CI-M6PR fused to photoactivable-GFP (A Left) either alone or together with the Golgi-targeted Tgn38-FRB (nonfluorescent) and the cytosolic mRFP-FKBP-only (A Middle) or the cytosolic mRFP-FKBP-Sac1 phosphatase (A Right). The next day the mounted live cells were preincubated for 10 min with rapamycin (100 nM) and the Golgi region of cells that showed recruitment was photoactivated. Time-lapse images were followed for 5 min with GFP excitation. (A) Representative images of each group of cells at the indicated time points. (B) Combined result of morphometric analysis of the recordings using MetaMorph software. The graphs show the number of objects around the Golgi that are above an arbitrary threshold normalized to the value found right after photoactivation (at 0 min). The graph illustrates the first 2.5 min of the recordings. (Means ± SEM of 12, 12, and 21 cells for the black, gray, and open symbol traces, respectively.)

Fig. 4.

Changes in the localization of Arf1, clathrin and the monomeric clathrin adaptor GGAs to the Golgi upon acute Sac1-phosphatase recruitment. Cells were transfected with the Golgi-targeted Tgn38-FRB-CFP (not displayed) together with the cytosolic mRFP-FKBP12-Sac1 and one of the following constructs: Arf1-GFP (A), clathrin light chain-GFP (B), thimidine kinase (TK) promoter driven GGA1-GFP (C), GGA2-GFP (D), and GGA3-GFP proteins (E). Pictures were taken before and after (5–8 min) rapamycin (100 nM). (F) Summary of intensity changes observed over the Golgi upon Sac1-phosphatase recruitment. Means ± SEM of number (n) of cells: Arf1 (n = 3), Golgi clathrin (n = 6), GGA1 (n = 6), GGA2 (n = 7), and GGA3 (n = 4). (Because the cytosolic background intensity increases after delocalization, a complete loss of Golgi localization in most constructs does not cause more than a 60% decrease in the initial intensity.)

Fig. 5.

The effects of PtdIns4P elimination on the resynthesis of the plasma membrane PtdIns(4,5)P₂ pools during phospholipase C activation. (A) Current model of how the PM is supplied with PtdIns(4,5)P₂ in mammalian cells. Because a large fraction of PtdIns4P is found in the Golgi (PI4P with red), some PtdIns4P may reach the PM via vesicular transport. (B) HEK-293 cells stably expressing AT1a receptors were cotransfected with the PtdIns(4,5)P₂ reporter PLCδ1PH-GFP, the Tgn38-FRB-CFP, and either the cytosolic mRFP-FKBP-only or the cytosolic mRFP-FKBP-Sac1-phosphatase constructs. Next day cells were examined on the stage of Zeiss LSM 510-META confocal microscope at room temperature. Rapamycin (100 nM) was added 10 min before stimulation with 100 nM AngII. The cytoplasmic fluorescence of the cells was calculated from ROIs drawn in the cytosol (outside the nucleus) from time-lapse images. Curves were normalized to their initial prestimulatory values and averaged either from cells showing Golgi recruitment or from cells without recruitment. Means ± SEM are shown from 60 to 80 cells recorded in multiple dishes from two independent experiments. The curves are plotted such that an increase in cytosolic PLCδ1PH-GFP fluorescence is shown as a downward change to better conceptualize that it reflects a drop in the plasma membrane PtdIns(4,5)P₂. Also note that in all of the cells expressing the Sac1 construct there are larger changes in PtdIns(4,5)P₂ than in those expressing the FKPB-only construct (compare black traces in the Upper and Lower panels), but a significantly slower recovery is observed only in the cells that showed Golgi recruitment of the Sac1 phosphatase (Lower trace).