Rapid isolation of extracellular vesicles from cell culture and biological fluids using a synthetic peptide with specific affinity for heat shock proteins

Abstract

Recent studies indicate that extracellular vesicles are an important source material for many clinical applications, including minimally-invasive disease diagnosis. However, challenges for rapid and simple extracellular vesicle collection have hindered their application. We have developed and validated a novel class of peptides (which we named venceremin, or Vn) that exhibit nucleotide-independent specific affinity for canonical heat shock proteins. The Vn peptides were validated to specifically and efficiently capture HSP-containing extracellular vesicles from cell culture growth media, plasma, and urine by electron microscopy, atomic force microscopy, sequencing of nucleic acid cargo, proteomic profiling, immunoblotting, and nanoparticle tracking analysis. All of these analyses confirmed the material captured by the Vn peptides was comparable to those purified by the standard ultracentrifugation method. We show that the Vn peptides are a useful tool for the rapid isolation of extracellular vesicles using standard laboratory equipment. Moreover, the Vn peptides are adaptable to diverse platforms and therefore represent an excellent solution to the challenge of extracellular vesicle isolation for research and clinical applications.

Figure 1. Selection and validation of peptides with HSP-binding properties.

A. Peptide selection. Representative demonstration of peptide screening with recombinant HSPs using broad-range pH (3–10) isoelectric focusing (IEF) gels. Samples of 20 µg of the indicated peptides were applied at the anode and 2 µg of purified recombinant HSPs were applied at the cathode. The gradient of pH and electrophoretic directions are indicated on the left side of the gel. Abbreviations of recombinant HSP sources and horizontal lines are used to assist in sample identification in the distorted counter migrations affecting lanes 10–14. The complexes formed during counter migration are indicated at the right of the gel as “Vn-HSP complexes”. Estimates of the isoelectric focusing points of unbound gp96, HSP90 and HSP60 in the area of counter migrant distortion are indicated by red arrows (lane 10, 11 and 12). The yellow arrowhead at the bottom indicates unbound Vn peptide isoelectric focusing following counter migration against HSPs. Red arrowheads (lanes 5, 17 and 29) indicate the weakly staining salmon HSP70 (despite standardized dilution). The green arrowhead (middle of lane 11) indicates a complex with HSP90 resulting from Vn20 that has extended binding influence across preceding adjacent lanes. The green arrowhead at the base of the gel (lane 11) indicates the final focusing point of the errantly migrating Vn20. Abbreviations are as follows: Ec, E.coli dnaK (lanes 1, 14,26); Ad, A.davidanieli HSP70 (lanes 3, 15, 27); Mt, M.tuberculosis HSP70 (lanes 4, 16, 28); Sa, Chinook salmon HSP70 (lanes 5, 17, 29); Ra, rat HSP70 (lanes 6, 18, 30); Hu, human HSP70-1 (lanes 7, 19, 31); Bo, bovine HSP70-8 (8, 20, 32); gp78, hamster HSP70-5 (lane 9, 21, 33); gp96, canine GRP96 (lanes 10, 22, 34); h90, human HSP90 (lanes 11, 23, 35); h60, human HSP60 (lanes 12, 24, 36) and as blank lanes (lanes 13 and 25). B. Sequences of Vn96, Vn20 as well as Scrambled-Vn96 (Scr-Vn96) and their predicted 3D structures in aqueous solution using PEP-FOLD server . Red, blue, green, and black amino acid residues are acidic, basic, hydrophobic uncharged and other amino acid residues, respectively. Note that the Vn96 peptide favors a helical conformation. C. Validation of HSP binding by the Vn96 peptide via affinity pull-down of HSPs from total cell lysate. MCF-7 breast cancer cells were lysed and processed as described in experimental procedures. Streptavidin-coupled magnetic beads saturated with either biotinylated-Vn96 (b-Vn96) or biotinylated-scrambled sequence of Vn96 (b-Vn96-Scr) peptides were used to perform the pull-down assays. In the immunoblot, 1% volumes of total cell lysate were run as input proteins to compare with proteins bound by the Vn96 peptides. The heat shock proteins tested are indicated. The right lower panel shows both HSP27 (indicated as ‘>’), HSP10 (indicated as ‘<<’) and a non-specific band (indicated as ‘*’).

Figure 2. Characterization of extracellular materials precipitated by Vn96.

A. Vn96 peptides precipitate vesicular structures from conditioned cell culture media. The biotinylated-Vn96 (b-Vn96) precipitated materials from conditioned cell culture media previously incubated with the MDA-MB-231 breast cancer cell line were subjected to Proteinase K digestion. The transmission electron microscopy analysis were performed on the precipitated material from the b-Vn96 sample (left panel), proteinase K-digested b-Vn96 sample (middle panel), and the proteinase K-digested sample from b-Scr-Vn96 (right panel). The scale bars are 100 nm. B. Identification of exosome markers in the Vn96-purified EVs from conditioned cell culture media. 50 µg each of Vn96 peptide and Scr-Vn96 were incubated with 1 ml of conditioned cell culture media previously incubated with the breast cancer cell line MCF-7 at 4°C for overnight. Exosomes were also isolated from the same conditioned cell culture media by ultracentrifugation (UCF). The presence of HSP70, HSP90 and GAPDH were assessed by immunoblotting. C. CD63 immunoblot. 1 ml of pre-cleared MCF-7 conditioned cell culture media was incubated to precipitate EVs with indicated amount of peptides either overnight (O/N) at 4°C or 30 minutes at room temperature. Total cell lysate of MCF-7 (equivalent to 0.2×10⁶ cells) was used as a positive control and conditioned cell culture media alone (c. media) was used as negative control. SDS-PAGE was performed in non-reducing conditions for CD63 immunoblots as recommended by the supplier.

Figure 3. Visualization of Vn96 peptide-precipitated extracellular vesicles (EVs) from biological fluids.

Pre-cleared biological fluids (conditioned cell culture media, human plasma and human urine) were used as described in experimental procedures. The Vn96 peptide-precipitated materials were dispersed into solution using Proteinase K digestion prior to microscopic analysis. A. Transmission electron microscopy images of Vn96-precipitated material from the indicated samples (conditioned cell culture media and diluted human plasma). The scale bars are 100 nm. B. Atomic force microscopy (phase) image of the Vn96 peptide-precipitated (Proteinase K digested) EVs from human urine. A differential size distribution pattern is observed between EVs from urine of normal and prostate cancer subjects (equal scale). The enlarged area from the prostate cancer image was used to measure width and thickness of two individual EVs (right panel, i and ii) in nanometers (nm) are shown in the bottom panel. C. Pre- and post-digital rectal exam with prostatic massage (DRE) urine samples were collected from consenting donors. EVs were isolated in parallel from equal volumes of urine using Vn96 and ultracentrifugation (UCF) methods; immunoblot analyses were performed using antibodies against the proteins indicated. Representative results for pre- and post-DRE urine samples from a donor are shown. Mbr Frac = Membrane fraction of prostate adenocarcinoma cell line LNCap and MWM = Protein molecular weight markers.

Figure 4. Comparative miRNA-seq data for Vn96- and UCF-purified EVs from conditioned cell culture media.

A. Scatter plot comparing normalized expression profiles of miRNAs contained in EVs isolated from the indicated conditioned cell culture media using either ultracentrifugation or the Vn96 peptide. For example, MCF7_UCF and MCF7_VN96 indicate that EVs were purified from conditioned cell culture media previously incubated with MCF-7 cells by ultracentrifugation and the Vn96 peptide, respectively. High Pearson correlations between ultracentrifugation and Vn96 peptide methods of EV purification from the same sample validate Vn96 as an EV purification tool. B. Scatter plot comparing normalized expression profiles of miRNAs contained in EVs isolated from MCF-7 versus MDA-MB-231 conditioned cell culture media using the same purification method. C. Venn diagram of miRNAs contained in EVs isolated from MCF-7 conditioned cell culture media using different methods (ultracentrifugation, Vn96 peptide and a commercially-available exosome purification kit). Less than 10% differences were observed in the miRNA populations between the ultracentrifugation and Vn96 peptide methods, and the commercial kit and Vn96 peptide methods (left panel), but a wider variation in miRNA populations was observed in EVs from different cell lines (right panel).