Multi-Cell Line Analysis of Lysosomal Proteomes Reveals Unique Features and Novel Lysosomal Proteins

Abstract

Lysosomes, the main degradative organelles of mammalian cells, play a key role in the regulation of metabolism. It is becoming more and more apparent that they are highly active, diverse, and involved in a large variety of processes. The essential role of lysosomes is exemplified by the detrimental consequences of their malfunction, which can result in lysosomal storage disorders, neurodegenerative diseases, and cancer. Using lysosome enrichment and mass spectrometry, we investigated the lysosomal proteomes of HEK293, HeLa, HuH-7, SH-SY5Y, MEF, and NIH3T3 cells. We provide evidence on a large scale for cell type-specific differences of lysosomes, showing that levels of distinct lysosomal proteins are highly variable within one cell type, while expression of others is highly conserved across several cell lines. Using differentially stable isotope-labeled cells and bimodal distribution analysis, we furthermore identify a high confidence population of lysosomal proteins for each cell line. Multi-cell line correlation of these data reveals potential novel lysosomal proteins, and we confirm lysosomal localization for six candidates. All data are available via ProteomeXchange with identifier PXD020600.

Keywords: lysosomes; mass spectrometry; posterior probability analysis; proteomics; superparamagnetic iron oxide nanoparticles (SPIONs).

Graphical abstract

Fig. 1

Lysosomal stability and recovery vary across different cell types.A, workflow for lysosome enrichment and mass spectrometric analysis. B, activity of the lysosomal luminal enzyme β-hexosaminidase determined from postnuclear supernatant fractions of combined light/heavy SILAC cells. β-Hexosaminidase activity with/without the addition of Triton-X-100 relates to the total fraction of lysosomes contained in the sample and those that ruptured during cell lysis, respectively. C, lysosome recovery rates in input and eluate fractions of lysosome enrichment experiments after correction for the presence of background cells not receiving SPIONs determined by β-hexosaminidase activities. Shown are mean values ±SD, n = 4. LC-MS/MS, liquid chromatography–tandem mass spectrometry; MS, mass spectrometry; SILAC, stable isotope labeling by amino acids in cell culture; SPION, superparamagnetic iron oxide nanoparticle.

Fig. 2

Proteomic analysis of lysosome-enriched fractions identifies unique patterns of individual cell lines.A, average number of identified proteins detected for individual cell types. B, average number of identified putative lysosomal proteins detected for individual cell types. C, Pearson correlation values of log2 abundance ratios (SPIONs/control) for putative lysosomal proteins across individual cell lines. D, heatmap and unsupervised hierarchical clustering of log2 abundance ratios (SPIONs/control) for putative lysosomal proteins. Shown are mean values ±SD, n = 4. SPION, superparamagnetic iron oxide nanoparticle.

Fig. 3

Proteins with a putative lysosomal localization are identified with a higher reproducibility between individual cell lines. Dataset size as well as individual overlaps for distinct combinations are indicated. A, all proteins identified in the lysosome-enriched fractions of the individual cell types. B, putative lysosomal proteins identified in lysosome-enriched fractions of the individual cell lines.

Fig. 4

Known lysosomal proteins show both highly conserved and diverse expression levels between individual cell lines. For each protein, the median iBAQ values were determined and normalized to the median intensity of the same eight V-ATPase complex subunits in a replicate-wise manner. Proteins are either sorted based on their median intensity (scatter plot) or grouped in a cell line–wise manner (dotted box plot). A, proteins with known function as transporter, channel, or exchanger. B, members of, or proteins related to, mTORC1. C, proteins with a known function related to the hydrolysis of glycosidic bonds. Shown are log10 converted median-normalized iBAQ values for proteins detected in ≥3 replicates in each of ≥2 cell lines. iBAQ, intensity-based absolute quantification.

Fig. 5

Investigation of lysosomal proteins expression levels in whole cell lysate datasets of individual cell lines. For each protein, the median iBAQ values were determined and normalized to the median intensity of the same eight V-ATPase complex subunits utilized for the lysosome-enriched fractions in a replicate-wise manner. Proteins are either sorted based on their median intensity (scatter plot) or grouped in a cell line–wise manner (dotted box plot). A, proteins with known function as transporter, channel, or exchanger. B, members of, or proteins related to mTORC1. C, proteins with a known function related to the hydrolysis of glycosidic bonds. Shown are log10-converted median-normalized iBAQ values for proteins detected in three replicates in each of ≥2 cell lines. iBAQ, intensity-based absolute quantification.

Fig. 6

Bimodal distribution analysis of differentially SILAC-labeled populations of SPIONs-receiving and control cells identifies potential lysosomal proteins. Histograms indicate binned frequencies of log2-transformed normalized SILAC ratios across the datasets. Normal distributed populations were calculated using an expectation-maximization algorithm, and a p-value of ≤0.05 was applied as cutoff. Red lines indicate the background population showing a similar behavior between control cells and such receiving SPIONs. Green lines indicate proteins with a significant difference in their SILAC ratio for cells receiving SPIONs relative to the background population. SILAC, stable isotope labeling by amino acids in cell culture; SPION, superparamagnetic iron oxide nanoparticle.

Fig. 7

Frequency of identification across cell lines correlates with lysosomal localization and allows for identification of high confidence novel lysosomal proteins.A, gene ontology (GO) analysis of all proteins contained in the respective datasets (total proteins, TP) and proteins that were determined to be significantly overrepresented in SPIONs-receiving cells (based on bimodal distribution analysis, p-value ≤0.05). The percentage of proteins (relative to the respective dataset) is shown based on their assignment to significantly enriched GO terms (FDR <0.05, Fisher's test). B, correlation of identification frequency and GO term distribution for total proteins and such overrepresented in SPIONs-receiving cells. Shown values represent the percentage of proteins assigned to a respective category normalized to the value for considering presence in at least one cell line. C, distribution of putative lysosomal proteins depending on their identification frequency. D, distribution of proteins determined to be specifically enriched by SPIONs in bimodal distribution analyses excluding putative lysosomal proteins depending on their identification frequency.

Fig. 8

Investigation of subcellular localization of lysosomal candidate proteins shows their predominantly lysosomal localization. NIH3T3 cell lines stably transfected with N-/C-terminally HA-tagged constructs of TM7SF3, NDFIP2, SLC31A1, or SLC12A9, which were detected in the significantly enriched fractions (p ≤ 0.05) in at least five cell lines. A, colocalization with the lysosomal marker LAMP2 investigated by immunostaining. B, Pearson's and Manders' correlation coefficients for the colocalization of signals for HA and LAMP2 for individual cell lines. C, Western blot analysis of stable transfected and wildtype cells. Detection of construct with anti-HA; GAPDH serves as loading control.