Sphingolipids containing very long-chain fatty acids regulate Ypt7 function during the tethering stage of vacuole fusion

Abstract

Sphingolipids are essential in membrane trafficking and cellular homeostasis. Here, we show that sphingolipids containing very long-chain fatty acids (VLCFAs) promote homotypic vacuolar fusion in Saccharomyces cerevisiae. The elongase Elo3 adds the last two carbons to VLCFAs that are incorporated into sphingolipids. Cells lacking Elo3 have fragmented vacuoles, which is also seen when WT cells are treated with the sphingolipid synthesis inhibitor Aureobasidin-A. Isolated elo3Δ vacuoles show acidification defects and increased membrane fluidity, and this correlates with deficient fusion. Fusion arrest occurs at the tethering stage as elo3Δ vacuoles fail to cluster efficiently in vitro. Unlike HOPS and fusogenic lipids, GFP-Ypt7 does not enrich at elo3Δ vertex microdomains, a hallmark of vacuole docking prior to fusion. Pulldown assays using bacterially expressed GST-Ypt7 showed that HOPS from elo3Δ vacuole extracts failed to bind GST-Ypt7 while HOPS from WT extracts interacted strongly with GST-Ypt7. Treatment of WT vacuoles with the fluidizing anesthetic dibucaine recapitulates the elo3Δ phenotype and shows increased membrane fluidity, mislocalized GFP-Ypt7, inhibited fusion, and attenuated acidification. Together these data suggest that sphingolipids contribute to Rab-mediated tethering and docking required for vacuole fusion.

Keywords: Elo3; HOPS; SNARE; Vps33; Ypt7.

Figure 1

Vacuole morphology is perturbed when sphingolipid synthesis is interrupted.A, WT and elo3Δ cells were grown overnight, back-diluted to OD₆₀₀ ∼0.7 in fresh YPD and grown for 2 h. WT cells were then treated with 0.125 μg/ml aureobasidin A (AbA) and incubated for 4 h. Cells were incubated with 5 μM FM4-64 to label vacuoles. Cell walls were stained calcofluor white. Images are representative of three separate trials. Scale bar: 5 μm. B, Vacuole fragmentation was quantified using data from A. Data was graphed as column scatter plots with error bars representing mean ± 95% CI. Significance was determined with one-way ANOVA for multiple comparisons [F (11, 249) = 186.5, ∗∗p < 0.0001]. (n = 3). Tukey’s multiple comparison test was performed for each cluster bin for individual p values. Ten fields containing ≥20 cells were counted per condition per experiment. Each data point represents a single field. ∗∗∗∗p < 0.0001; ns, not significant.

Figure 2

Vacuoles from elo3Δ cells are fusion impaired.A, time course of fusion for WT or elo3Δ vacuoles. Fusion activity was expressed as the percentage of WT activity. Significance was determined using one-way ANOVA for multiple comparisons for each time point between strains [F (5, 12) = 15.62; ∗∗∗∗p < 0.0001]. Tukey’s multiple comparison test was used for individual p values (n = 3). Error bars represent mean ± SE. ∗∗p < 0.01. ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. B, endpoint fusion (90 min) reactions were performed with combinations of WT and elo3Δ (Δ) vacuoles. Significance was determined using one-way ANOVA for multiple comparisons with WT/WT as a control [F (4, 9) = 10.07; ∗∗p = 0.0022]. Dunnett’s test for multiple comparison was used for individual p values. Error bars represent mean ± SE. (n = 3). ∗∗p < 0.01. C, measurement of vacuole diameters after fusion. WT and elo3Δ vacuoles were incubated for the indicated times at 27 °C and images were taken by fluorescence microscopy. Diameters of individual vacuoles in clusters were measured using ImageJ. Data was graphed in a scatter plot of the pooled data. Significance was determined using one-way ANOVA for multiple comparisons [F (7, 1288) = 53.22 p < 0.0001]. Tukey’s test for multiple comparisons was used for individual p values. The bars represent the median with upper and lower quartiles. (n = 3). ∗p < 0.05, ∗∗∗∗p < 0.0001. Scale bar: 5 μm. D, lipid mixing experiments were performed with WT or elo3Δ vacuoles. Fluorescence (ex = 544 nm, em = 590 nm) was measured every 60 s after the addition of ATP. Left, A representative run of the experiment. Right, Quantitation of three experiments calculated by measuring the difference in fluorescence between no ATP and ATP for WT and elo3Δ at the end of the assay. Error bars represent mean ± SE (n = 3). Data was analyzed using an unpaired two-tailed t test ∗∗p < 0.01. E, Western blots of WT and elo3Δ vacuoles that were incubated on ice or at 27 °C for 1 h. Vacuole extracts were resolved by SDS-PAGE and transferred to nitrocellulose for immunoblotting using antibodies against the specified proteins.

Figure 3

C26-SL are important for membrane fluidity and acidification.A, vacuoles were mixed with MC540 in PS buffer. MC540 fluorescence was measured (ex = 530 nm, em = 585 nm) and intensities were normalized to the WT set to 1. Data was analyzed using an unpaired two-tailed t test. Error bars represent mean ± SE (n = 3). ∗∗p < 0.01. B, WT vacuole fluidity was determined as described above in the presence or absence of 500 μM dibucaine in 0.5% ethanol. Here, changes in fluorescence were normalized to each strain that was set to 1. Changes in fluorescence was analyzed using one-way ANOVA for multiple comparisons with no treatment (0 μM dibucaine/0.5% ethanol) as a control [F (4, 14) = 14.87; ∗∗∗∗p < 0.0001]. Dunnett’s multiple comparison test was used for individual p values. Error bars are mean ± SE. (n = 3). ∗∗p < 0.01. C, vacuole acidification was determined by measuring changes in AO fluorescence at 520 nm. Fluorescence decreased as vacuoles were acidified. Reactions were incubated with or without ATP and fluorescence was measured every 20 s. The protonophore FCCP was added at 500 s to collapse the proton gradient. Fluorescence was normalized to initial intensities at the time of adding ATP and set to 1. D, quantitation of multiple experiments shown in C. Fluorescence values at 400 s (WT peak acidification) were averaged for each strain. Significance was measured using one-way ANOVA for multiple comparisons [F (3, 14) = 80.30, ∗∗∗∗p < 0.0001]. Tukey’s post hoc multiple comparisons test was used for individual p values. Error bars represent mean ± SE. (n > 4). ∗∗∗p < 0.001.

Figure 4

Vertex enrichment of ergosterol and PI3P does not require C26-SL.A, docking reactions using filipin to mark ergosterol. FM4-64 was added at the end of the reaction and docked vacuoles were visualized by fluorescence microscopy. Scale bars: 5 μm. B, quantitation of ratiometric fluorescence intensities for vertices (V) and outer edge (O) in panel A. The data points were pooled from multiple experiments where 15 to 20 clusters with ≥10 vacuoles per cluster were analyzed per experiment (n > 400 vertices for each strain; n > 200 outer edge measurements for each strain). Error bars represent geometric means ± geometric SD (n = 3). Significance was measured using one-way ANOVA for multiple comparisons [F (3, 452) = 24.71, ∗∗∗p < 0.0001]. Tukey’s post hoc multiple comparisons test was used for individual p values. ∗p < 0.05; ns, not significant. C, docking reactions of purified vacuoles were performed in the presence of Cy5-FYVE to mark PI3P. MDY-64 was added at the end of the reaction and docked vacuoles were visualized using fluorescence microscopy. Scale bars: 5 μm. D, Quantitation of ratiometric fluorescence intensities for Cy5-FYVE at V and O in panel C as described in A. Error bars represent geometric means ± geometric SD (n = 3). Significance was measured using one-way ANOVA for multiple comparisons [F (3, 291) = 6.152, ∗∗∗p = 0.0005]. Tukey’s post hoc multiple comparisons test was used for individual p values. ∗∗∗∗p < 0.0001; ∗∗∗p < 0.001 ns, not significant.

Figure 5

Vacuole tethering is defective in elo3Δ vacuoles.A, vacuoles were incubated for 20 min at 27 °C under docking conditions. Following incubation reaction mixtures were placed on ice and stained with MDY64. WT vacuoles were incubated with purified 2 μM GDI to inhibit the Ypt7-dependent tethering of vacuoles. Shown are two examples for each strain and condition. Scale bars: 5 μm. B, quantitation of vacuole clustering from A. Vacuole clusters (∼300) were counted for each condition. Significance was measured using one-way ANOVA for multiple comparisons [F (14, 295) = 11.75, p < 0.0001]. Tukey’s post hoc multiple comparisons test was used for individual p values. Error bars represent mean ± SE (n = 3). ∗∗p < 0.01; ∗∗∗p < 0.001; ns, not significant. C, co-isolation experiments performed with GST-Ypt7 immobilized on glutathione agarose resin and preloaded with GTP. 6X fusion reactions containing vacuoles from WT or elo3Δ were solubilized with TX100 and incubated separately with Ypt7-loaded beads. The amount of Vps33 and Vps18 that was bound to resin and found in the input sample were determined by Western blotting. PD, pulldown. Input lanes show to 10% of the total material added to GST-Ypt7 loaded beads. D, quantification and comparison of the GST-Ypt7 pull-down efficiency. Efficiency was calculated by dividing the area of Vps33 or Vps18 pulldown band by the area of the corresponding GST-Ypt7 band in each experiment as determined by ImageJ. Error bars represent mean ± SE (n = 4). Significance was measured using one-way ANOVA for multiple comparisons [F (5, 18) = 56.88 p < 0.0001]. Šidák’s test was used for pairwise comparisons between strains and individual p values. ∗∗∗∗p < 0.0001. ns, not significant. E, sensitivity of fusion to GDI. Fusion assays using WT and elo3Δ vacuoles were performed with a dose curve of GDI. Each strain was normalized to its own maximum fusion set to 1. Graphpad Prism was used to log transform, normalize and fit the data to a non-linear regression curve to yield IC₅₀ values of GDI for vacuole fusion. Error bars represent the mean ± SE. (n = 3).

Figure 6

Ypt7 is mislocalized on elo3Δ vacuoles.A, WT and elo3Δ vacuoles containing GFP-Ypt7 were incubated under docking conditions for 20 min at 30 °C. After incubation, the reactions were placed on ice and labeled with PSS-380. A large cluster of elo3Δ vacuoles is shown to see more vertices. Arrows point at representative vertices. Arrowheads point at representative outer edges. Scale bar: 5 μm. B, cumulative distribution plot depicting the percentile values of GFP-Ypt7/PSS-380 ratio for vertex (V) and outer edge (O) in WT vs elo3Δ vacuoles. Each curve is comprised of pooled data points from three experiments. Each experiment used 10 to 15 clusters with ≥10 vacuoles per cluster (n > 400 vertices for each strain; n > 200 outer edge measurements for each strain) for each condition. (n = 3). C, Quantitation of ratiometric GFP fluorescence at vertices (V) and outer edge (O) in B. The data points were pooled from multiple experiments and ≥15 to 20 clusters with at ≥10 vacuoles per cluster per experiment for each strain. Error bars represent geometric means ± geometric SD. (n = 3). Significance was measured using one-way ANOVA for multiple comparisons [F (3, 634) = 6.779, ∗∗∗p = 0.0002]. Tukey’s post hoc multiple comparisons test was used for individual p values. ∗∗∗∗p < 0.0001; ns, not significant. D, Western blot of WT and elo3Δ vacuoles. Antibody against Ypt7 was used to show levels of GFP-Ypt7 (∼49 kDa). Antibody against Vph1 was used as a loading control. GFP-Ypt7 levels on WT vacuoles treated with 250 μM dibucaine is also shown.

Figure 7

HOPS localization to vertex domains is not affected on elo3Δ vacuoles.A, WT and elo3Δ vacuoles containing Vps33-GFP were analyzed for vertex enrichment as described above. Scale bar: 5 μm. B, quantitation of ratiometric of Vps33-GFP fluorescence intensities for V and O in panel C. Data was collected as described in (B). Error bars represent geometric means ± geometric SD. (n = 3). Significance was measured using one-way ANOVA for multiple comparisons [F (3, 1148) = 24.71, ∗∗∗p < 0.0001]. Tukey’s post hoc test for multiple comparisons was used for individual p values. ∗∗∗∗p < 0.0001; ns, not significant. C, WT and elo3Δ vacuoles containing GFP-Vps39 were analyzed for vertex enrichment as described above. Scale bar: 5 μm. D, quantitation of ratiometric of GFP-Vps39 fluorescence intensities for V and O in panel C. Data was collected as described in (B). Error bars represent geometric means ± geometric SD. (n = 3). Significance was measured using one-way ANOVA for multiple comparisons [F (3, 1110) = 92.40) = ∗∗∗p < 0.0001]. Tukey’s post hoc test for multiple comparisons was used for individual p values. ∗∗∗∗p < 0.0001; ns, not significant.

Figure 8

Treating WT cells with Aureobasidin A mimics the altered GFP-Ypt7 localization seen in elo3Δ cells.A, wild type and elo3Δ yeast expressing GFP-Ypt7 were grown overnight in selective media and back diluted to ∼0.7 OD₆₀₀ the next day. Cells were transferred to YPD broth and incubated for 4 h at 30 °C while shaking. WT cells were also treated with AbA. Cells were then incubated with 10 μM MDY-64 for 15 min in the dark before visualization. Images are representative of three separate trials. Arrows represent complete rings of GFP-Ypt7 around a vacuole. Scale bars: 4 μm. B, quantitation of the number of vacuoles with complete rings of GFP-Ypt7 per cell. Ten 63x fields were counted for a total of 200 to 600 vacuoles per condition/strain for each experiment. The number of vacuoles with complete GFP-Ypt7 in each field was converted to a percentage of total amount of vacuoles. Bars represent mean ± 95% CI. Significance was measured using one-way ANOVA for multiple comparisons [F (2, 90) = 16.00, ∗∗∗∗p < 0.0001]. Tukey’s post hoc multiple comparisons test was used for individual p values. ∗∗∗p < 0.0001, ∗∗∗∗p < 0.0001; ns, not significant. C-D, WT and elo3Δ yeast expressing GFP-Vps39 (C), and Vps33-GFP (D) were grown as described in panel A overnight and diluted to ∼0.7 OD₆₀₀ the next day. Cells were transferred to YPD broth and incubated for 4 h at 30 °C while shaking. Cells were stained with FM4-64 and visualized by fluorescence microscopy. Images are representative of three separate trials. Scale bars: 4 μm.

Figure 9

Dibucainetreatment of WT reproduces the elo3Δ phenotype.A, docking reactions containing of WT vacuoles containing GFP-Ypt7 were performed in the with or without 250 μM Dibucaine. FM4-64 was added at the end of the reaction and docked vacuoles were visualized using fluorescence microscopy. Scale bar: 5 μm. B, cumulative distribution plot depicting the percentile values of GFP-Ypt7/FM4-64 ratio for vertex (V) and outer edge (O) and the effect of dibucaine. Each curve is comprised of pooled data points from three experiments. Each experiment used 10 to 15 clusters with ≥10 vacuoles per cluster (n > 400 vertices for each strain; n > 200 outer edge measurements for each strain) for each condition. C, geometric means ± geometric SD for the data in panel B. Significant differences were measured using one-way ANOVA for multiple comparisons [F (3, 1607) = 149.8, ∗∗∗∗p < 0.0001]. Tukey’s post hoc test was used for individual p values. ∗∗∗∗p < 0.0001; ns, not significant. D, fusion reactions were performed with WT vacuoles treated with a dosage curve of Dibucaine. IC₅₀ was determined using Prism. Error bars represent mean ± SE. (n = 3). E, gain of resistance vacuole fusion assays were performed in the presence of buffer, 1 mM NEM, 2 μM GDI or 500 μM dibucaine. Fusion reactions were incubated for 120 min at 27 °C. Reactions were treated with reagents or placed on ice at the indicated time points. Fusion was normalized to the buffer-treated reactions at each time point. F, calculated half-times of inhibition from data in E. Significant differences were measured using one-way ANOVA for multiple comparisons [F (3, 10) = 63.4, ∗∗∗∗p < 0.0001]. Tukey’s post hoc test was used for individual p values. Error bars represent mean ± SE. (n = 4). ∗p < 0.05, ∗∗∗∗p < 0.0001, ns, not significant. G, vacuole acidification was determined for WT in the presence of dibucaine at the indicated concentrations. AO fluorescence was measured at 520 nm every 20 s. A reaction lacking ATP was used as a negative control. FCCP was added at 600 s to collapse the proton gradient. Fluorescence was normalized to initial intensities at the time of adding ATP and set to 1. H, quantitation of three experiments shown in D. Fluorescence intensities were averaged for each condition at 400 s. Significant differences were measured using one-way ANOVA for multiple comparisons [F (9, 20) = 19.31, ∗∗∗∗p < 0.0001]. Tukey’s post hoc test was used for individual p values. Error bars represent mean ± SE. (n = 3). ∗∗∗∗p < 0.0001.