FIGURE 3. GSL synthesis depends on PtdIns4P, and not on membrane trafficking at the GC.

a–c, GSL synthesis: labelling with [³H]sphingosine for 60 min (a) or 120 min (b, c); VSVG transport: labelling with [³⁵S]methionine for 10 min, chase for 60 min; transport of VSVG quantified as fraction of VSVG acquiring resistance to endoglycosidase H (Endo H). a, Upper panel: cells mock-treated (control) or treated with 10 M ONO-RS-082 or 200 M dicoumarol 30 min before and during pulse–chase. Open columns, GSL synthesis; filled columns, VSVG transport. Lower panel: representative autoradiography of Endo-H-induced shift in VSVG mobility on SDS–PAGE. b, Cells were mock-treated (control), knocked down for Bet3 (Bet3 KD), or exposed to 50 g ml^-1 cycloheximide (CHX) both 60 min before and during [³H]sphingosine pulse. Data are expressed as percentages of relative controls. White columns, GSL synthesis; grey columns, SM synthesis; black columns, VSVG transport. c, Cells were mock-treated (white columns), knocked down for PI(4)KII (light grey columns) or PI(4)KIII (dark grey columns), or treated with 1 M PAO (black columns). d, e, GSL synthesis controls TGN-to-PM membrane trafficking. d, FAPP2 KD cells (filled columns) were transfected with siRNA-resistant forms of wild-type FAPP2 (FAPP2 WT) or FAPP2-W407A or FAPP2-R18L. Open column, mock-transfected cells. TGN-to-PM transport of VSVG was quantified by counting cells showing PM staining for VSVG. e, Cells were mock-treated, knocked down for FAPP2 or GCS, or treated with 75 M d-PDMP for 1 h. For release of [³⁵S]glycosaminoglycans, see Methods. Data are shown as means  s.d. for three independent experiments, each in duplicate (n = 3). Asterisks indicate statistical significance.