Fig. 1 GPN-mediated reduction in lysotracker.

. Wild-type and NPC1 human fibroblasts were labeled with 200nM lysotracker red for 15min at room temperature. Lysosomes were destabilized with 300 M GPN and the rate of reduction in lysotracker fluorescence measured over time. A, confocal images over time of GPN mediated reduction in lysotracker fluorescence, NPC1 cells are brighter (caused by storage) but have similar rate of loss of fluorescence as wild-type cells. B, representative profile of loss of lysotracker fluorescence in a single NPC1 and wild-type fibroblast. C, quantitation of data shown in B, n=8 with minimum of 5 cells analysed per experiment. These data are consistent with there being no pH difference in NPC1 lysosomes.

Fig. 2 Determination of low affinity Rhod-dextran Kd at physiological and lysosomal pH. The Kd of rhod dextran was determined in vitro in a medium that mimics the intracellular environment (mM: 140 KCl, 10 NaCl, 1 MgCl2) at neutral or acidic pH using 10 mM Hepes (pH 7.2) or 10 mM acetate (pH 4.5) as buffers. Rhod dextran was used at 5 M and its fluorescence (excitation/emission, 544/590 nm) monitored at increasing free [Ca2+] generated by increasing the total added [Ca2+] (0-4 mM) in the presence of 5 mM Ca2+ chelator (the free [Ca2+] was calculated using Winmaxchelator 3.2, C. Patton, Stanford University; http://www.stanford.edu/~cpatton/). BAPTA and 5,5-dibromo-BAPTA were used as Ca2+ chelators since they exhibit good Ca2+ buffering with appropriate Ca2+ affinities over these pH ranges [1, 2]. To generate 1-50 mM free [Ca2+], Ca2+ chelators were omitted and the medium supplemented with these concentrations of Ca2+ alone. The data from each chelator were grouped according to pH, and plotted and fitted to a single binding site using GraphPad Prism 4 (San Diego, CA, USA). Each point represents the mean of two experiments.

Fig. 3 U18666A washout time-course. RAW macrophages were incubated with U18666A for 12h (T0) prior to U18666A washout. Cells were followed for 0-48h. A, Cells (control, red point and U18666A treated, black points) were incubated with 5 M Calcium Green 1-AM and Fura Red-AM as described (materials and methods), lysosomal calcium was released using GPN, all measurements were performed on a Zeiss LSM 510 confocal, n=4. Panel A shows recovery over time of reduced lysosomal calcium content caused by U18666A (normalised at 12h post washout). B, cells were either fixed at each timepoint for filipin staining (upper panels, false coloured in red) or live stained with BODIPY-LacCer (BOD-LC), n=4. Correction in trafficking with BOD-LC is observed at 12-24h, cholesterol levels start to normalise after 48h. These data are consistent with cholesterol and GSL storage being downstream of a primary lesion in lysosomal calcium homeostasis caused by sphingosine accumulation.

Fig. 4 Tetramethylrhodamine labeled sphingosine (TMR-sphingosine) is localised to lysosomes and induces an NPC1 phenotype. TMR-sphingosine (Echelon Biosciences Inc., Salt Lake City, Utah, USA) a sphingosine drivative with natural stereochemistry with the TMR group attached to the hydrocarbon tail. A, TMRsphingosine (2 M, 10min) induces a reduction in LE/Lys calcium content assessed by GPN induced lysis. B, TMR-sphingosine (2 M, 10min) co-localises with lysotracker green. C, TMR-sphingosine (2 M, 20h) induces classical NPC1 cellular phenotypes including mistrafficking of BODIPY-lactosylceramide (BODIPY-LacCer), cholesterol (filipin) and sphingomyelin (lysenin) storage (n=3, scale bar = 5 m). These data indicate that fluorescently labeled sphingosine behaves in an identical manner to endogenous sphingosine and following 10 minutes incubation all sphingosine is trapped within the acidic late endosome/lysosome compartments.

Fig. 5 Thapsigargin elevates LE/Lys calcium in wild-type but not NPC1 human fibroblasts. Wild-type (WT) and NPC1 human fibroblasts were loaded with 0.25mg/ml low affinity Rhod-dextran and 0.1mg/ml Alexa Fluor 488 dextran for 12h followed by 12h chase at 37C. Cells were incubated with 1 M thapsigargin for 15min and immediately visualised on a Zeiss LSM 510 confocal. The acidic store of WT cells double their calcium concentration following thapsigargin treatment whereas NPC1 acidic stores remain unchanged (n=4 with a minimum of 20 cells analysed per coverslip). These data are consistent with a defect in calcium store filling in NPC1.

Fig. 6 Chelation of cytosolic calcium elevation inhibits normalisation of NPC1 fibroblasts. Incubation of NPC1 cells with curcumin (30 M, 1h), to elevate cytosolic calcium, normalises cholesterol levels (filipin). Chelation of basal cytosolic calcium with BAPTA-AM (50 M, 1h) does not alter the cholesterol storage phenotype of NPC1 cells. Chelation of curcumin induced cytosolic calcium elevation with BAPTA-AM prevented normalisation of cholesterol storage in NPC1 cells (n=3, scale bar = 5 m). These data are consistent with the primary beneficial effect of curcumin in NPC1 being via its action on SERCA to elevate cytosolic calcium.

SUPPLEMENTAL REFERENCES 1. 2. Tsien, R.Y. (1980). New calcium indicators and buffers with high selectivity against magnesium and protons: design, synthesis, and properties of prototype structures. Biochemistry 19, 2396-2404. Harrison, S.M., and Bers, D.M. (1987). The effect of temperature and ionic strength on the apparent Ca-affinity of EGTA and the analogous Ca-chelators BAPTA and dibromo-BAPTA. Biochimica et biophysica acta 925, 133-143.