Proteomics analysis of CDVs

Purified CDVs and EVs together with their parental cells, were analyzed for their proteome profiles. Cell pellets (50 μg) were solubilized in 8 M urea (GE Healthcare, #17-1319-01) and 100 mM Tris–HCl (Invitrogen, #AM9855-G). For CDVs and EVs, a total protein of 50 µg of each sample was precipitated with chilled acetone (Sigma-Aldrich, #179124-500ML) at − 20 °C overnight and solubilized using the same buffer as described for cell pellets. The sample of each group was reduced with 10 mM dl-Dithiothreitol (Sigma-Aldrich, #D0632-10G) at 56 °C for 30 min and alkylated with 25 mM iodoacetamide (Sigma-Aldrich, #I1149-25G) at 25 °C for 30 min. The sample was diluted with 100 mM Tris–HCl to decrease the urea concentration to 1 M, pH 8. The protein mixture was then digested by Trypsin/Lys-C Mix (Promega, #V5073) at 37 °C for 16 h in a ratio of 1:50 (enzyme: protein, w/w). The resultant peptides were cleaned using an OASIS SPE cartridge (Waters, USA), dried in Savant SpeedVac (Thermo Fisher Scientific, USA), and resuspended in 0.1% formic acid (Thermo Fischer Scientific, #LS118-4) in water at 0.5 µg/µL. Digested peptides were analyzed using Q-Exactive plus mass spectrometry with an EASY nLC 1000 liquid chromatography system (Thermo Fisher Scientific, USA). The mass spectrometer was operated in data-dependent mode with a full scan (m/z 350–2000) followed by MS/MS for the top 20 precursor ions in each cycle and the data-dependent neutral loss method. The acquired MS/MS spectra were subjected to searches against the Uniprot-Human database (Jun 2018; 73099 sequences) using SEQUEST software in Proteome Discoverer v2.4 (Thermo Fisher Scientific, version 2.4). Two missed trypsin cleavages were allowed, and the peptide mass tolerances for MS/MS and MS were set to 0.02 Da and 10 ppm, respectively. Other parameters used for the SEQUEST searches included the fixed modification of carbamidomethylation at cysteine (+ 57.021 Da) and the variable modification of oxidation at methionine (+ 15.995 Da). The abundance of each peptide was measured as the number of peptide spectrum match (PSM) events. Data was presented in the form of abundance for each protein.

Proteins identified from all three different replicates of each group were filtered first based on the requirement of at least two peptides per protein. Only proteins with measurable abundance values were included in the analyses. Then, the differentially expressed proteins were identified using the criteria of fold change > 2 and p-value < 0.05 in relative protein abundance values to cells. The most abundant membrane protein markers were selected from the top 5% of the CDV proteome dataset (by abundance) and > seven-fold of the relative abundance compared to the parental cells. Principal component analysis (PCA) and Venn diagram were performed using Proteome Discoverer v2.4.