Heterogeneity of late endosome/lysosomes shown by multiplexed DNA-PAINT imaging

Abstract

Late endosomes/lysosomes (LELs) are crucial for numerous physiological processes and their dysfunction is linked to many diseases. Proteomic analyses have identified hundreds of LEL proteins; however, whether these proteins are uniformly present on each LEL, or if there are cell-type-dependent LEL subpopulations with unique protein compositions is unclear. We employed quantitative, multiplexed DNA-PAINT super-resolution imaging to examine the distribution of seven key LEL proteins (LAMP1, LAMP2, CD63, Cathepsin D, TMEM192, NPC1, and LAMTOR4). While LAMP1, LAMP2, and Cathepsin D were abundant across LELs, marking a common population, most analyzed proteins were associated with specific LEL subpopulations. Our multiplexed imaging approach identified up to eight different LEL subpopulations based on their unique membrane protein composition. Additionally, our analysis of the spatial relationships between these subpopulations and mitochondria revealed a cell-type-specific tendency for NPC1-positive LELs to be closely positioned to mitochondria. Our approach will be broadly applicable to determining organelle heterogeneity with single organelle resolution in many biological contexts.

Figure 1.

Quantitative DNA-PAINT pipeline for characterizing LEL membrane proteins. Schematic shows LELs with target proteins imaged in this study. Target proteins (e.g., LAMP1 and LAMP2) are labeled with primary and secondary antibodies for DNA-PAINT imaging. Step 1 shows the data acquisition step in which a dual-color DNA-PAINT image is acquired (in this case, LAMP1 and LAMP2). Subsequently, in Step 2, LAMP1 is segmented to be used as the reference channel, and the LAMP2 raw localizations (target channel) are overlaid onto the segmented LELs. During colocalization (Step 3), the target localization density is measured in the region of the segmented LELs (colocalization region), as well as a local background region surrounding the LEL. If the density within the colocalization region is significantly (>3 standard deviations) enriched over the background density that LEL is determined to have both target and reference proteins. Several quantitative outputs are obtained as a result of this pipeline. Cell scale bars = 10 µm. Inset scale bars = 500 nm.

Figure S1.

Overexpression of target proteins confirms antibody (AB) specificity. (A–F) Overexpression of tagged constructs of (A) LAMP1-Flag, (B) LAMP2-BFP, (C) CD63-HA, (D) TMEM192-3xHA, (E) NPC1-Flag, and (F) LAMTOR4-Flag were used to verify antibody specificity in HeLa cells. Widefield images showing the co-staining with antibodies to the target protein (magenta) and the overexpression tag (yellow) colocalized for all proteins listed. Cell scale bars, 10 µm. Inset scale bars, 1 µm.

Figure 2.

Dual-color DNA-PAINT in HeLa cells identifies heterogeneity in the enrichment of canonical LEL proteins. (A–F) Representative DNA-PAINT images of LAMP1 or LAMP2 reference channel (magenta) and target protein channel (yellow) in HeLa cells for (A) LAMP2, (B) CD63, (C) Cathepsin D, (D) TMEM192, (E) NPC1, and (F) LAMTOR4. Arrows indicate LELs positive for both LAMP1/LAMP2 reference and target protein, arrowheads indicate LELs positive for LAMP1/LAMP2 reference and negative for the target protein, and the dotted circle indicates which LEL arrow or arrowhead refers to. Cell scale bars, 10 µm. Inset scale bars, 1 µm.

Figure 3.

Dual-color DNA-PAINT in ARPE-19 cells identifies heterogeneity in the enrichment of canonical LEL proteins. (A–F) Representative DNA-PAINT images of LAMP1 or LAMP2 reference channel (magenta) and target protein channel (yellow) in ARPE-19 cells for (A) LAMP2, (B) CD63, (C) Cathepsin D, (D) TMEM192, (E) NPC1, and (F) LAMTOR4. Arrows indicate LELs positive for both LAMP1/LAMP2 reference and target protein, arrowheads indicate LELs positive for LAMP1/LAMP2 reference and negative for target protein, and the dotted circle indicates which LEL arrow or arrowhead refers to. Cell scale bars, 10 µm. Inset scale bars, 1 µm.

Figure S2.

Controls demonstrate optimal fixation and permeabilization conditions as well as the robustness of DNA-PAINT imaging. (A) Representative DNA-PAINT image of glyoxal-fixed HeLa cells stained for LAMP1 and LAMP2 shows significant disruption of LAMP2 staining. (B and C) Representative DNA-PAINT image of HeLa cells fixed with warm 4% PFA and permeabilized with (B) 0.2% Triton X-100 or (C) 0.1% saponin. (D) Quantitative analysis of protein density shows a significant reduction with 0.2% Triton X-100. Mann Whitney U-Test was performed to compare LAMP2 protein density in saponin-treated cells (N = 3 biological replicates, n = 16 cells, replotted from Fig. 4) with Triton X-100 treated cells (N = 1 biological replicate, n = 6 cells) (*** P = 0.0002). (E and F) Plots show the percent area of an LEL calculated from the reference image (LAMP2 for Cathepsin D, LAMP1 for all other targets) covered by the localizations from the target protein in HeLa and ARPE-19 cells. (E) shows LAMP1 calculated as the reference channel while (F) shows all other targets calculated as the target channel on a LAMP1/LAMP2 reference channel. Colored lines show traces from 500 randomly selected LELs per target. Solid black line indicates average over all traces for that target. As LAMP1 was used as the reference channel, these traces always reached 100% coverage. Other targets cover a variable percent of LEL area, and importantly individual traces approach a plateau by 25,000 frames for all targets, indicating majority of target protein localizations have been captured by this imaging time. (G and H) (G) Representative dual-color DNA-PAINT images in HeLa cells and (H) quantification in HeLa and ARPE-19 cells of LAMP1 and EEA1 shows distinct organelle populations labeled by each marker, indicating that LAMP1 does not label EEA1-positive early endosomes. N = 3 biological replicates per cell type, n = 14 HeLa cells, n = 14 ARPE-19 cells. Cell scale bars, 10 µm. Inset scale bars, 1 µm. (I and J) LAMP1 and LAMP2 levels on individual LELs in (I) HeLa and (J) ARPE-19 cells show minimal correlation. Pearson’s correlation coefficients (R²) were 0.138 and 0.263 for HeLa and ARPE-19 cells, respectively. (K) LAMP1 localization density is consistent across multiple distinct dual-color experiments in HeLa cells. Kruskal–Wallis test was performed on median LAMP1 density per cell across experiments where LAMP1 was used as the reference channel, P = 0.5342, not significant.

Figure 4.

Dual-color DNA-PAINT identifies LEL subpopulations with variable protein makeup that persist across epithelial cell types. (A and B) Box and whisker plots showing the percent colocalization of target proteins with (A) LAMP1 used as reference or (B) LAMP2 used as a reference in HeLa (magenta), HeLa LAMP1-GFP overexpressing (green), or ARPE-19 cells (cyan). All targets were imaged in three independent biological replicates. (A) For LAMP2 (n = 16 HeLa cells, n = 13 HeLa LAMP1-GFP overexpressing cells, n = 11 ARPE-19 cells), CD63 (n = 14 HeLa cells, n = 13 ARPE-19 cells), TMEM192 (n = 13 HeLa cells, n = 14 ARPE-19 cells), NPC1 (n = 16 HeLa cells, n = 15 HeLa LAMP1-GFP overexpressing cells, n = 14 ARPE-19 cells), and LAMTOR4 (n = 16 HeLa cells, n = 14 ARPE-19 cells). CBC refers to the target protein colocalization as analyzed using a coordinate-based colocalization analysis. (B) For Cathepsin D (n = 15 HeLa cells, n = 13 ARPE-19 cells) and LAMP1 (n = 16 HeLa cells, n = 13 HeLa LAMP1-GFP overexpressing cells, n = 11 ARPE-19 cells). Plot line color indicates cell type, black circles indicate individual cells. Mann–Whitney U test was performed to compare percent colocalization of LAMP2 with LAMP1 (n = 16 HeLa cells, n = 13 HeLa LAMP1-GFP overexpressing cells) (**** P < 0.0001), NPC1 with LAMP1 (n = 16 HeLa cells, n = 13 HeLa LAMP1-GFP overexpressing cells) (P = 0.3527, no significance), and LAMP1 with LAMP2 (n = 16 HeLa cells, n = 12 HeLa LAMP1-GFP overexpressing cells) (**** P < 0.0001) in HeLa versus HeLa LAMP1-GFP overexpressing cells. Vertical dashed lines separate different target proteins. (C and D) Violin plots of target protein density for (C) high-density and (D) low-density targets. Plot line color indicates cell type, black line indicates median target density on LELs positive for a given target. Black circles, squares, and diamonds represent individual cells from three different biological replicates. Mann–Whitney U test was performed to compare median target densities in HeLa versus HeLa LAMP1-GFP overexpressing cells for LAMP2 (n = 16 HeLa cells, n = 13 HeLa LAMP1-GFP overexpressing cells) (* P = 0.0172) and NPC1 (n = 16 HeLa cells, n = 13 HeLa LAMP1-GFP overexpressing cells) (P = 0.0683). Vertical dashed lines separate different target proteins. (E–G) Pattern analysis of target proteins on individual LEL membranes. (E) Representative zoomed-in DNA-PAINT image of an individual LEL as defined by LAMP1 or LAMP1-GFP localizations (magenta, first panel), raw localizations of target protein on the LEL (yellow, second panel), nanoclusters of the target channel segmented using DBSCAN and pseudocolored to indicate distinct nanoclusters (third panel), and these nanoclusters overlaid on LEL reference channel alphashape bounding area calculated from the LAMP1 or LAMP1-GFP localizations (fourth panel). Scale bar, 100 nm. (F) Violin plots show nanocluster density (number of localizations of the target protein per unit area of the nanocluster). Plot line color indicates cell type, black line indicates median nanocluster density on LELs positive for a given target. Black circles, squares, and diamonds represent individual cells from three different biological replicates. Vertical dashed lines separate different target proteins. Mann-Whitney U test was performed to compare median nanocluster densities in HeLa cells between NPC1 (n = 16 cells) and LAMTOR4 (n = 16 cells) (**** P < 0.0001) and for NPC1 in HeLa (n = 16 cells) versus HeLa LAMP1-GFP overexpressing cells (n = 13 cells) (**** P < 0.0001). (G) Violin plots show nanocluster diameter. Vertical dashed lines separate different target proteins. Mann–Whitney U test was performed to compare median nanocluster diameters in HeLa cells between NPC1 (n = 16 cells) and LAMTOR4 (n = 16 cells) (**** P < 0.0001) and for NPC1 in HeLa (n = 16 cells) versus HeLa LAMP1-GFP overexpressing cells (n = 13 cells) (* P < 0.0251). (H) FRC analysis (left y-axis) and localization precision measurements (right y-axis) for NPC1 and LAMTOR4 target protein images. Circles indicate individual fields of view. Solid line indicates data plotted on the left versus right y-axis.

Figure S3.

Cathepsin D, TMEM192, and NPC1 antibodies are robust and specific. (A) Representative DNA-PAINT image of LAMP2 and Cathepsin D when the Cathepsin D protein is knocked down via siRNA in HeLa cells shows a significant reduction in Cathepsin D signal on LAMP2 with respect to wildtype conditions. Cell scale bars, 10 µm. Inset scale bars, 1 µm. (B) Representative DNA-PAINT image of LAMP2 and TMEM192 in HeLa cells imaged using an alternative TMEM192 antibody shows a staining pattern consistent with results from the main antibody used. Cell scale bars, 10 µm. Inset scale bars, 1 µm. (C) Representative DNA-PAINT image of LAMP1 and TMEM192 when the TMEM192 protein is knocked down via siRNA in HeLa cells shows a significant reduction in TMEM192 signal on LAMP1 with respect to wildtype conditions. Cell scale bars, 10 µm. Inset scale bars, 1 µm. (D) Representative DNA-PAINT image of LAMP1 and TMEM192 in TMEM192-3xHA overexpressing (OE) HeLa cells shows near-complete overlap with LAMP1. Cell scale bars, 10 µm. Inset scale bars, 1 µm. (E) Representative DNA-PAINT image of LAMP1 and NPC1 in NPC1-null HeLa cells shows minimal overall signal and overlap with LAMP1 with respect to wildtype conditions. Cell scale bars, 10 µm. Inset scale bars, 1 µm. (F–H) WB analyses of (F) Cathepsin D, (G) TMEM192, and (H) NPC1 levels in HeLa cells. (F) Cathepsin D antibody recognizes both uncleaved pro-Cathepsin D and cleaved Cathepsin D protein products, as expected. Cathepsin D siRNA knockdown in HeLa cells shows a reduction in Cathepsin D signal. (G) TMEM192 siRNA knockdown in HeLa cells shows a reduction in TMEM192 signal. (H) NPC1-null HeLa cells show loss of NPC1 signal. (I) Quantification of colocalization of Cathepsin D, TMEM192, and NPC1 in HeLa cells. Plot line color indicates cell type, black circles indicate individual cells, solid black line indicates data plotted on left Y-axis (LAMP2 reference) versus right y-axis (LAMP1 reference). Vertical dashed lines separate different target proteins. Mann–Whitney U-tests were performed to compare Cathepsin D siRNa knockdown HeLa cells (N = 1 biological replicate, n = 5 cells) with HeLa cells (N = 3 biological replicates, n = 15 cells, replotted from Fig. 4) (*** P = 0.0005), HeLa cells imaged with TMEM192 main antibody (N = 3 biological replicates, n = 13 cells, replotted from Fig. 4), or alternative antibody (N = 1 biological replicate, n = 6 cells) (P = 0.6388, no significance), TMEM192 siRNA knockdown HeLa cells (N = 1 biological replicate, n = 4 cells) with HeLa cells (N = 3 biological replicates, n = 13 cells, replotted from Fig. 4) (** P = 0.0034), TMEM192-3xHA overexpressing cells (N = 1 biological replicate, n = 7 cells) with wildtype HeLa cells (N = 3 biological replicates, n = 13 cells, replotted from Fig. 4) (* P = 0.0297), and NPC1-null HeLa cells (N = 1 biological replicate, n = 6 cells) with HeLa cells (N = 3 biological replicates, n = 16 cells, replotted from Fig. 4) (**** P < 0.0001). (J) Protein density of Cathepsin D, TMEM192, and NPC1 in HeLa cells. Plot line color indicates cell type, black circles indicate individual cell medians, solid black line indicates data plotted on the left y-axis (LAMP2 reference) versus the right y-axis (LAMP1 reference). Vertical dashed lines separate different target proteins. Mann–Whitney U-tests were performed to compare: Cathepsin D siRNa knockdown HeLa cells (N = 1 biological replicate, n = 5 cells) with HeLa cells (N = 3 biological replicates, n = 15 cells, replotted from Fig. 4) (** P = 0.0015), TMEM192 siRNA knockdown HeLa cells (N = 1 biological replicate, n = 4 cells) with HeLa cells (N = 3 biological replicates, n = 13 cells, replotted from Fig. 4) (** P = 0.0034), and TMEM192-3xHA overexpressing cells (N = 1 biological replicate, n = 7 cells) with HeLa cells (N = 3 biological replicates, n = 13 cells, replotted from Fig. 4) (* P = 0.0236), and NPC1-null HeLa cells (N = 1 biological replicate, n = 6 cells) with HeLa cells (N = 3 biological replicates, n = 16 cells, replotted from Fig. 4) (**** P < 0.0001). Source data are available for this figure: SourceData FS3.

Figure S4.

Controls show the robustness of object-based colocalization analysis for low-density targets, and LAMP1-GFP overexpression (OE) in HeLa cells induces changes in LELs. (A) Box and whisker plots of percent colocalization of target proteins on LELs in HeLa cells imaged with varying primary antibody (AB) concentrations. The solid line indicates data plotted on the left y-axis (LAMP1 reference) versus the right y-axis (LAMP2 reference). Mann–Whitney U tests were performed to compare LAMP2 1:100 (standard dilution, N = 3 biological replicates, n = 16 cells, replotted from Fig. 4) with LAMP2 1:2,000 (20-fold lower concentration, N = 1 biological replicate, n = 7 cells) (**** P < 0.0001), NPC1 1:100 (standard dilution, N = 3 biological replicates, n = 16 cells, replotted from Fig. 4) with NPC1 1:50 (twofold higher concentration, N = 1 biological replicate, n = 7 cells) (P = 0.5245, no significance), TMEM192 1:25 (standard dilution, N = 3 biological replicates, n = 13 cells) with TMEM192 1:12.5 (twofold higher concentration, N = 1 biological replicate, n = 6 cells) (P = 0.21, no significance), and TMEM192 1:12.5 (twofold higher concentration, N = 1 biological replicate, n = 6 cells) with TMEM192 imaged with an alternative antibody (N = 1 biological replicate, n = 6 cells, replotted from Fig. S3) (P = 0.4848, no significance). (B) Violin plots of target protein density on LELs in HeLa cells imaged with varying primary antibody concentrations. Mann–Whitney U tests were performed to compare LAMP2 1:100 (standard dilution, N = 3 biological replicates, n = 16 cells, replotted from Fig. 4) with LAMP2 1:2,000 (20-fold lower concentration, N = 1 biological replicate, n = 7 cells) (**** P < 0.0001), NPC1 1:100 (standard dilution, N = 3 biological replicates, n = 16 cells, replotted from Fig. 4) with NPC1 1:50 (twofold higher concentration, N = 1 biological replicate, n = 7 cells) (P = 0.249, no significance), and TMEM192 1:25 (standard dilution, N = 3 biological replicates, n = 13 cells) with TMEM192 1:12.5 (twofold higher concentration, N = 1 biological replicate, n = 6 cells) (P = 0.5789, no significance). (C) Colocalization percentage versus median protein density per cell plots for NPC1 and LAMP2 imaged on LAMP1 reference LELs in HeLa cells. Black circles represent individual cells. Red boxes indicate protein density values between 3,500 and 6,500 localizations/area where both NPC1 and LAMP2 targets have similar median densities but LAMP2 shows much higher percent colocalization with LAMP1. (D) Colocalization percentage versus median density per cell plots for TMEM192 in HeLa cells and TMEM192 in TMEM192-3xHA overexpressing HeLa cells imaged on LAMP1 reference LELs. Black circles represent individual cells. Red boxes indicate protein density values between 1,000–4,000 localizations/area where both WT and TMEM192 overexpressing cells have similar median protein densities but TMEM192 shows much higher percent colocalization with LAMP1 in overexpressing cells. (E) Box and whisker plots of colocalization percentage of target proteins with LAMP1 analyzed with two different colocalization thresholds. NPC1 and TMEM192 both show minimal differences in colocalization when analyzed using the standard cutoff threshold for colocalization of three standard deviations above background density when compared to the less-stringent two standard deviation above background density threshold. Mann-Whitney U tests were performed to compare percent colocalization with LAMP1 for NPC1 using three standard deviation versus two standard deviation filter (N = 3 biological replicates, n = 16 cells) (P = 0.0849, no significance) and TMEM192 using three standard deviation versus two standard deviation filter (N = 3 biological replicates, n = 13 cells) (P = 0.2642, no significance). (F and G) Representative DNA-PAINT images of (F) LAMP2 and (G) NPC1 with LAMP1 in HeLa cells overexpressing LAMP1-GFP. Arrows indicate LELs positive for both LAMP1-GFP and target protein, arrowheads indicate LELs positive for LAMP1 and negative for target protein, dotted line indicates which LEL arrow or arrowhead refers to. Cell scale bars, 10 µm. Inset scale bars, 1 µm. (H) Comparison of LEL area in wild type versus LAMP1-GFP overexpressing HeLa cells. Black circles indicate individual cell medians. Mann-Whitney U-test was performed to compare cell medians of HeLa cells (n = 75 cells) with LAMP1-GFP overexpressing HeLa cells (n = 26 cells) (* P = 0.0197).

Figure 5.

Drug-based perturbations impact LEL subpopulations. (A–D) Representative DNA-PAINT images of LAMP1 reference channel (magenta) and target protein channel (yellow) in (A and B) HeLa cells treated for 16 h with either 15 μM EN6 or (C and D) 100 nm BafA1. Arrows indicate LELs positive for both LAMP1 and target protein, arrowheads indicate LELs positive for LAMP1 and negative for target protein, dotted circle indicates which LEL arrow or arrowhead refers to. Cell scale bars, 10 µm. Inset scale bars, 1 µm. (E) Box and whisker plots showing the percent colocalization of target proteins with LAMP1 reference in HeLa (re-plotted from Fig. 4), EN6-treated HeLa, or BafA1-treated HeLa cells. All targets were imaged in three or four independent biological replicates. Plot line color indicates cell type, black circles indicate individual cells. Cells were clustered using a standard minimum area filter unless otherwise specified. Mann–Whitney U test was performed to compare the percent colocalization of NPC1 in untreated HeLa with NPC1 in EN6-treated HeLa (n = 16 HeLa cells, n = 14 EN6-treated HeLa cells) (**** P < 0.0001), NPC1 in untreated HeLa with NPC1 in BafA1-treated HeLa (n = 16 HeLa cells, n = 21 BafA1-treated HeLa cells) (*** P = 0.001), NPC1 in BafA1-treated HeLa with untreated HeLa using 350-nm-diameter filter (n = 21 BafA1-treated HeLa cells, n = 16 HeLa cells) (P = 0.058), LAMTOR4 in untreated HeLa with LAMTOR4 in EN6-treated HeLa (n = 16 HeLa cells, n = 13 EN6-treated HeLa cells) (*** P = 0.0005), LAMTOR4 in untreated HeLa with LAMTOR4 in BafA1-treated HeLa (n = 16 HeLa cells, n = 12 BafA1-treated HeLa cells) (**** P < 0.0001), and LAMTOR4 in BafA1-treated HeLa cells with untreated HeLa using 350-nm-diameter filter (n = 12 BafA1-treated HeLa cells, n = 16 HeLa cells) (**** P < 0.0001). (F) Violin plots of target protein density on reference LELs in HeLa (re-plotted from Fig. 4), EN6-treated HeLa, BafA1-treated HeLa cells, or HeLa cells with LELs segmented using 350-nm-diameter filter. Plot line color indicates cell type, black line indicates median target density on LELs positive for a given target. Black circles represent individual cells from three or four independent biological replicates. Mann-Whitney U test was performed to compare median target densities in untreated HeLa with NPC1 in EN6-treated HeLa (n = 16 HeLa cells, n = 14 EN6-treated HeLa cells) (P = 0.061), NPC1 in untreated HeLa with NPC1 in BafA1-treated HeLa (n = 16 HeLa cells, n = 21 BafA1-treated HeLa cells) (P = 0.8204, no significance), NPC1 in BafA1-treated HeLa with untreated HeLa using 350-nm-diameter filter (n = 21 BafA1-treated HeLa cells, n = 16 HeLa cells) (P = 0.8678, no significance), LAMTOR4 in untreated HeLa with LAMTOR4 in EN6-treated HeLa (n = 16 HeLa cells, n = 13 EN6-treated HeLa cells) (**** P < 0.0001), LAMTOR4 in untreated HeLa with LAMTOR4 in BafA1-treated HeLa (n = 16 HeLa cells, n = 12 BafA1-treated HeLa cells) (*** P = 0.0004), and LAMTOR4 in BafA1-treated HeLa cells with untreated HeLa using 350-nm-diameter filter (n = 12 BafA1-treated HeLa cells, n = 16 HeLa cells) (**** P < 0.0001). (G) Violin plots of median LEL area in untreated HeLa cells, BafA1-treated HeLa cells, or HeLa cells with LELs segmented using a 350-nm-diameter filter. All cells imaged with LAMP1 and NPC1 as well LAMP1 and LAMTOR4 were combined for each condition. Mann–Whitney U test was performed to compare median LEL areas per cell in untreated HeLa with BafA1-treated HeLa (n = 32 HeLa cells, n = 33 BafA1-treated HeLa cells) (**** P < 0.0001), and BafA1-treated HeLa with untreated HeLa using 350-nm-diameter filter (n = 32 HeLa cells, n = 33 BafA1-treated HeLa cells) (**** P < 0.0001). (H) Box and whisker plots showing the total NPC1-positive LEL area per cell in untreated HeLa cells, BafA1-treated HeLa cells, or HeLa cells with LELs segmented using a 350-nm-diameter filter. Mann–Whitney U test was performed to compare the total NPC1-positive LEL area per cell in untreated HeLa with BafA1-treated HeLa (n = 16 HeLa cells, n = 21 BafA1-treated HeLa cells) (* P = 0.0346), and BafA1-treated HeLa with untreated HeLa using 350-nm-diameter filter (n = 21 BafA1-treated HeLa cells, n = 16 HeLa cells) (** P = 0.0064). (I) Box and whisker plots showing the total NPC1-negative LEL area per cell in untreated HeLa cells, BafA1-treated HeLa cells, or HeLa cells with LELs segmented using 350-nm-diameter filter. Mann-Whitney U test was performed to compare NPC1-negative LEL area per cell in untreated HeLa with BafA1-treated HeLa (n = 16 HeLa cells, n = 21 BafA1-treated HeLa cells) (* P = 0.0162), and BafA1-treated HeLa with untreated HeLa using 350-nm-diameter filter (n = 21 BafA1-treated HeLa cells, n = 16 HeLa cells) (P = 0.089).

Figure S5.

Bulk analysis of drug-treated HeLa cells, examining LEL subpopulations with additional cellular context in ARPE-19 cells, and multiplexed imaging controls. (A and B) Representative WBs of (A) NPC1 and (B) LAMTOR4 in untreated, EN6-treated, or BafA1-treated conditions in HeLa cells. EN6 treatment leads to an increase in both NPC1 and LAMTOR4. BafA1 treatment leads to an increase in NPC1 but a decrease in LAMTOR4. (C–G) Cumulative density function plots of LEL distance from the nucleus from dual-color DNA-PAINT imaging experiments in ARPE-19 cells for subpopulations containing LAMP1 or LAMP2 only or LAMP1/LAMP2 and target protein: (C) CD63, (D) Cathepsin D, (E) TMEM192, (F) NPC1, and (G) LAMTOR4. Distance to the nucleus was normalized per cell to the maximum distance from the nucleus of an LEL in that cell, with values closest to zero indicating greatest proximity to the nucleus. Line indicates median with standard deviation between biological replicates. Kolmogorov–Smirnov tests were performed on the mean distributions from three independent biological replicates: for CD63 n = 13 cells, P = 0.193, no significance; for Cathepsin D n = 13 cells, P = 0.794, no significance; for TMEM192 n = 14 cells, P = 0.261, no significance; for NPC1 n = 14 cells, P = 0.140, no significance; for LAMTOR4 n = 14 cells, P = 0.193, no significance. (H–J) Representative three-color DNA-PAINT image of LAMP1, NPC1, and mitochondria (TOM20) in ARPE-19 cells. (H and I) (H) Raw images and (I) post-processed image showing a spatial map of LELs with or without NPC1 in relation to mitochondria. (J) Quantification shows combined subpopulations of seven cells from three biological replicates. Cell scale bars, 10 µm. Inset scale bars, 1 µm. (K) Acetylated tubulin and mitochondria (TOM20) in HeLa cells imaged using two rabbit primary antibodies pre-labeled with secondary rabbit nanobodies, showing minimal cross-talk between the two targets. Cell scale bars, 10 µm. Inset scale bars, 1 µm. (L) Localizations corresponding to Tetraspeck beads imaged in a multicolor DNA-PAINT acquisition were used for image registration. Localizations are shown before and after post-processing alignment. Scale bar, 100 nm, Source data are available for this figure: SourceData FS5.

Figure 6.

Spatial analysis reveals the location of LEL subsets with respect to the nucleus, mitochondria, and the TGN. (A–E) Cumulative density function plots of LEL distance from the nucleus calculated from dual-color DNA-PAINT images in HeLa cells for subpopulations containing LAMP1 or LAMP2 only or LAMP1/LAMP2 and target protein: (A) CD63, (B) Cathepsin D, (C) TMEM192, (D) NPC1, and (E) LAMTOR4. Distance to the nucleus was normalized per cell to the maximum distance from the nucleus of an LEL in that cell, with values closest to zero indicating the greatest proximity to the nucleus. The line indicates median with standard deviation between biological replicates. Kolmogorov–Smirnov tests were performed on the mean distributions from three independent biological replicates: for CD63, n = 14 cells, P = 0.261, no significance; for Cathepsin D, n = 15 cells, P = 0.193, no significance; for TMEM192, n = 13 cells, P = 0.794, no significance; for NPC1, n = 16 cells, P = 0.261, no significance; for LAMTOR4, n = 16 cells, P = 0.069, no significance. (F–H) Three-color DNA-PAINT image of LAMP1, NPC1, and mitochondria (TOM20) in HeLa cells. (F and G) (F) Raw images and (G) post-processed image showing a spatial map of LELs with or without NPC1 in relation to mitochondria. (H) Quantification shows combined subpopulations of five cells from three biological replicates. Cell scale bars, 10 µm. Inset scale bars, 1 µm. (I–L) Three-color DNA-PAINT image of LAMP1, NPC1, and TGN (P230) in HeLa cells. (I and J) (I) Raw images and (J) post-processed images showing a spatial map of LELs with or without NPC1 in relation to TGN. Cell scale bars, 10 µm. Inset scale bars, 1 µm. (K and L) Quantification of five cells from three independent biological replicates shows (K) percent of NPC1-positive LELs colocalized with the TGN compared to LELs independent from the TGN and (L) distance of each LEL subtype to the TGN normalized to the maximum distance in each cell. Mann-Whitney U test was used to compare percent colocalization with NPC1 for LELs on TGN with LELs independent from TGN (n = 5 HeLa cells) (P = 0.4206, no significance).

Figure 7.

Higher-order multiplexing reveals molecularly distinct LEL subsets. (A–C) Three-color DNA-PAINT image of LAMP1, LAMTOR4, and NPC1 in HeLa cells reveals four distinct subpopulations. (A and B) (A) Raw images and (B) post-processed image showing a spatial map of these four subsets. (C) Quantification shows combined subpopulations of four cells from three biological replicates. Cell scale bars, 10 µm. Inset scale bars, 1 µm. (D–F) Four-color DNA-PAINT image of LAMP1, CD63, LAMTOR4, and NPC1 in ARPE-19 cells reveals eight distinct subpopulations. (D and E) (D) Raw images and (E) post-processed images showing a spatial map of these eight subsets. (F) Quantification shows combined subpopulations of four cells from three biological replicates. Cell scale bars, 10 µm. Inset scale bars, 1 µm.