Fig 2: Mass spectrometry-based proteomic analysis of ASM-regulated DRM-associated proteins. (A) A schematic illustration of the proteomic procedure to analyze ASM-regulated DRM-associated proteins. (B) A partial list of proteins identified the ASM-sensitive DRM proteome. Proteins identified are cross-referenced to the palmitoyl-proteome database (SwissPalm) and 65 proteins are found to be known palmitoylated proteins (Table S3). Selected proteins from Table S3 are listed here. Semi-quantitative mass spectrometry data, fold of protein abundance change between control siLuciferase and siASM, as well as numbers of peptides identified in each condition, for the listed proteins are shown in the middle columns. The numbers of articles that reported the particular protein found in the palmitoyl-proteome is indicated in the last column. (C) A pie illustration to indicate the functional categories and the percentage of proteins in each category. Analysis is based on the identified palmitoylated proteins in the ASM-sensitive DRM proteome (Table S3). (D) Validation of the mass spectrometry-based quantitation by western blot analysis. Selected proteins from the list in panel B were subjected to western blot analysis using corresponding antibodies. Cells were treated with ASM siRNA or desipramine, or corresponding controls, and subjected by discontinuous sucrose gradient ultracentrifugation analysis. The pooled DRM fractions are analyzed as described in Fig. 1G. (E) The ABE assay to detect protein palmitoylation. Lysates from U373-MG cells were exposed to NEM to block all unmodified free thio groups (-SH) in the cysteine residues. The NEM-treated cell lysates were next exposed to hydroxylamine (HA), or left untreated. Both groups were then incubated with biotin-HPDP and subsequently pulled down by the streptavidin-agarose resins, followed by western blot analysis. Palmitoylated protein are enriched by HA treatment as compared to the no prior HA treatment group. Whole cell lysates (WCL) were used as control.

Fig 3: DRM-association and intracellular localization of the palmitoylation-defective SNAP23 and Lyn. (A) Mutations of specific cysteines in SNAP23 and Lyn greatly reduced their palmitoylation. Wild-type (wt) and cysteine to serine (CS) mutants of SNAP23 and LYN were expressed as GFP-SNAP23 and GFP-Lyn fusion proteins, with GFP tagged at the N-terminus of SNAP23 and C-terminus of Lyn, respectively. Expression constructs were stably transfected into U373-MG cells. The cell lysates were prepared and palmitoylation of the wild type and CS mutants were analyzed using the ABE assay as in Fig. 2E. While the exogenous CS mutant of SNAP23 and Lyn has greatly reduced palmitoylation levels, the palmitoylation status of the endogenous SNAP23 and Lyn are not affected, further serving as controls. (B) Cysteine mutants of SNAP23 and Lyn failed to associate with the DRM fractions. Wild-type and CS mutants of GFP-SNAP23 and GFP-Lyn proteins were stably transfected into U373-MG cells. The cells were lysed in Brij58-containing lysis buffer and fractionated in the discontinuous sucrose gradient ultracentrifugation, as in Fig. 1B, and the pooled fractions were examined by western blot analysis, as described in Fig. 1G. Only the wild-type but not the CS mutant proteins are found in the DRM fractions, while the CS mutant proteins are found mostly in the pooled soluble fractions. (C) The association of SNAP23 with the DRM fractions is sensitive to the ASM inhibitor desipramine. U373-MG cells were treated with the ASM inhibitor desipramine or vehicle control DMSO for 12 h. Lysates were prepared in the Brij58 lysis buffer and subjected to the sucrose discontinuous sucrose gradient ultracentrifugation as described in Fig. 1B. The fractions were analyzed by western blot analysis using anti-SNAP23 antibodies. (D) CS mutants of GFP-SNAP23 and GFP-Lyn failed to appear on the plasma membrane and were accumulated in the intracellular compartment. Wild-type and CS mutants of GFP-SNAP23 and GFP-Lyn constructs were transfected into U373-MG cells. The localization of the GFP-tagged ectopically expressed proteins on the plasma membrane and intracellular compartment was examined by fluorescence microscopy. Scale bars: 10 µm.

Fig 4: ASM acts at the plasma membrane to regulate the intracellular trafficking of palmitoylated proteins. (A) U373-MG cells were treated with 50 nM control and two independent ASM siRNAs for 48 h. Alternatively, cells were treated with 25 μM desipramine or DMSO for 12 h. The localization of ASM and its catalytic product, ceramide, was examined by specific anti-ASM and anti-ceramide antibody immunostaining. The plasma membrane staining of both ASM and ceramide (indicated by arrows) disappeared in cells after treatment with ASM siRNAs or with the ASM inhibitor desipramine. (B) The distribution of ganglioside GM1 on the plasma membrane is not affected by ASM inactivation. The experiments were conducted similarly as panel A except that cells were immunostained with an anti-GM1 antibody (immunostaining of GM1 on the plasma membrane is indicated by arrows). (C) Cells were transfected with siRNAs as in panel A for 48 h. Live cells were stained directly with anti-ASM antibodies at room temperature for 1 h. Cells were then fixed and processed for staining with by secondary antibodies and imaging. (D) Extracellular recombinant ASM enzyme can rescue the trafficking defects of SNAP23 in the ASM siRNA-treated cells. U373-MG cells were transfected with 50 nM control and ASM siRNAs for 48 h. In one set of the ASM siRNA-treated cells (right panel), purified recombinant active recombinant human ASM enzyme (4 μg/ml final) was added to the culture media for 2 h. Cells were then fixed, processed by immunostaining with anti-SNAP23 antibody, and examined by fluorescence microscopy. (E) For quantification the relative abundance of SNAP23 under all three conditions in panel C was measured (200 cells from each condition), as in Fig. 3C, and then normalized to control cells. Three independent experiments were measured to obtain the average of SNAP23 distribution on the plasma membrane (statistically significant, P<0.01, Student's t-test). (F) A model for the ASM function in regulation of SNAP23 and Lyn trafficking. ASM acts at the outer-leaflet of the plasma membrane (PM) to convert sphingomyelins (SM) to ceramides (CER). While CER are found in both leaflets of the plasma membrane, SM are mainly on the outer leaflet of the plasma membrane. Ceramides can retrogradely traffic back to Golgi (TGN). On the Golgi membrane, CER recruit the palmitoylated proteins such as SNAP23 and Lyn to facilitate the transport of these proteins out of the Golgi. SNAP23 and Lyn are carried by the ceramide-rich transport vesicles (possibly endosomes) en route to the plasma membrane, which eventually fused with the plasma membrane. On the plasma membrane, SNAP23 and Lyn are also associated with CER-rich lipid microdomains. When SNAP23 and Lyn are depalmitoylated, the depalmitoylated proteins become cytosolic and then associate with Golgi to be palmitoylated again, starting another cycle of trafficking. When ASM is inactivated, SNAP23 and Lyn are trapped in the Golgi, due to the lack of the CER on the Golgi membrane, which are essential for the transport of the palmitoylated proteins out of the Golgi. Scale bar: 10 μm.

Fig 5: Fractionation of the ASM-regulated membrane-associated proteins by discontinuous sucrose gradient ultracentrifugation. (A) A schematic workflow of the discontinuous sucrose gradient fractionation procedure. (B) The distribution of tyrosine kinases IGF-1R and Yes in the discontinuous sucrose gradient in control (DMSO) and desipramine (Desi, 25 µM, 12 h) treated U373-MG cells by anti-IGF-1R and anti-Yes antibody immunoblotting. Flotillin was used as a lipid raft marker. Fractions were collected from the top (fraction #1) to the bottom of the gradient (fraction #16). The distribution of IGF-1R or Yes was reduced in the DRM fractions (#1–4) after ASM inhibition. (C) Loss of ASM reduced the levels of Yes in the detergent resistant membrane fraction. U373-MG cells were transfected with 50 nM control (luciferase, siLuc) and ASM siRNAs (siASM) for 48 h and the cells were harvested in the Brij58-containing lysis buffer and fractionated and analyzed as in B. (D) The same as in B except U373-MG cells were treated control (PBS) or 10 mM MßCD for 1 h and the cells were harvested in the Brij58-containing lysis buffer, fractionated and analyzed as in B. (E) The reduction of ASM activities in the desipramine-treated cells or the ASM siRNAs-treated cells were determined using the sphingomyelinase assay, as compared with control cells. Data are means±s.d. (n=3) with three independent repeats. (F) Knockdown efficiency of two independent ASM siRNAs on the ASM protein levels in U373-MG cells. ASM was immunoprecipitated from the lysates and then analyzed by western blot by an anti-ASM antibodies. (G) Independent verification of the effects of ASM siRNAs, desipramine and MßCD on the distribution of Yes in discontinuous sucrose gradient. U373-MG cells were transfected with 50 nM control (siLuc) and two independent ASM siRNAs for 48 h or treated with 25 µM desipramine for 10 h or 10 mM MßCD for 1 h. The cells were harvested in the Brij58-containing lysis buffer and fractionated. The distribution of Yes was monitored in the pooled DRM fractions (fraction #1–4) or soluble fractions (fraction #13–16) by western blotting.