Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 8 (1), 1663043
eCollection

Systematic Characterization of Extracellular Vesicle Sorting Domains and Quantification at the Single Molecule - Single Vesicle Level by Fluorescence Correlation Spectroscopy and Single Particle Imaging

Affiliations

Systematic Characterization of Extracellular Vesicle Sorting Domains and Quantification at the Single Molecule - Single Vesicle Level by Fluorescence Correlation Spectroscopy and Single Particle Imaging

Giulia Corso et al. J Extracell Vesicles.

Abstract

Extracellular vesicles (EV) convey biological information by transmitting macromolecules between cells and tissues and are of great promise as pharmaceutical nanocarriers, and as therapeutic per se. Strategies for customizing the EV surface and cargo are being developed to enable their tracking, visualization, loading with pharmaceutical agents and decoration of the surface with tissue targeting ligands. While much progress has been made in the engineering of EVs, an exhaustive comparative analysis of the most commonly exploited EV-associated proteins, as well as a quantification at the molecular level are lacking. Here, we selected 12 EV-related proteins based on MS-proteomics data for comparative quantification of their EV engineering potential. All proteins were expressed with fluorescent protein (FP) tags in EV-producing cells; both parent cells as well as the recovered vesicles were characterized biochemically and biophysically. Using Fluorescence Correlation Spectroscopy (FCS) we quantified the number of FP-tagged molecules per vesicle. We observed different loading efficiencies and specificities for the different proteins into EVs. For the candidates showing the highest loading efficiency in terms of engineering, the molecular levels in the vesicles did not exceed ca 40-60 fluorescent proteins per vesicle upon transient overexpression in the cells. Some of the GFP-tagged EV reporters showed quenched fluorescence and were either non-vesicular, despite co-purification with EVs, or comprised a significant fraction of truncated GFP. The co-expression of each target protein with CD63 was further quantified by widefield and confocal imaging of single vesicles after double transfection of parent cells. In summary, we provide a quantitative comparison for the most commonly used sorting proteins for bioengineering of EVs and introduce a set of biophysical techniques for straightforward quantitative and qualitative characterization of fluorescent EVs to link single vesicle analysis with single molecule quantification.

Keywords: CD63; EV double labelling; EV engineering; EV protein sorting domains; Extracellular vesicles; exosomes; fluorescence correlation spectroscopy; single vesicle imaging.

Figures

Figure 1.
Figure 1.
Single molecule – single vesicle characterization of CD63-GFP-tagged extracellular vesicles. (a) Direct sucrose gradient fractionation of conditioned medium from CD63-GFP transfected HEK293T cells without prior ultracentrifugation. X-axis: measured sucrose concentrations in collected fractions. GFP fluorescent particles of ca 50–120 nm diameter (measured by FCS) co-fractionate with the EV marker Alix, peaking at ca 32–36% sucrose. (b) Following UF on a 100 kDa MWCO membrane, enriched medium was loaded onto a Superdex200 column for size exclusion chromatography, with continuous UV detection (blue line). The total GFP fluorescence count rate was measured in each of the 60 fractions by FCS (upper panel, green line). Western blots for different exosomal markers were performed after pooling of 4 fractions each into samples A-L as indicated, omitting one fraction in between pools. CD63-GFP fusion protein bands (multiple glycosylated isoforms) versus truncated GFP bands detected on an anti-GFP western blot are indicated by arrows and asterisks, respectively. (c) Cryo-EM analysis of UF-SEC isolated CD63-GFP versus native EVs (representative images). Scaling bar: 100 nm. (d) Left panel, LC-MS proteomics of native versus CD63-GFP HEK293T EVs. Spectral counts for a subset of proteins from the list in Table S1 are shown. Right panel, Venn diagram comparing the top 500 proteins, ranked by total spectral count (left), or all proteins detected at a < 1% false discovery rate (FDR) (right). (e) Individual fractions from size exclusion chromatography of HEK293T CD63-GFP EVs in Figure 1(b) were analysed by FCS with (red data points) and without (blue data points) vesicle disruption by the detergent NP40s. The translational diffusion time (τdiff, upper panel left y-axis and conversion into hydrodynamic diameter, right y-axis), molecular brightness (CPP, second panel) as well as absolute number of fluorescent molecules within the confocal volume (N, third panel) are depicted across the fractions. The ratio of freely diffusing GFP fluorescent particles before and after NP40s treatment (third panel) yields the average number of CD63-GFP molecules per particle (bottom panel). Frames indicate fractions comprising the two main fluorescent populations: the first indicates a heterogenous population comprising mostly GFP-tagged EVs, whereas the second one consists of monomeric GFP. (f) Disruption of EVs (stored at −80°C) with different mechanical and biochemical treatments based on the fraction of intact vesicles versus free GFP measured by FCS. (g) Time course of CD63-GFP vesicle release. The concentration of CD63-GFP positive vesicles in UF enriched samples from HEK293T cells was determined by FCS, the total number of vesicles between 70 and 140 nm in size was determined by NTA. Time points represent hours of conditioning, starting 5 h after transfection.
Figure 2.
Figure 2.
Overexpression of different GFP-tagged EV-markers in HEK293T parental cells. (a) LC-MS/MS proteomic analysis of HEK293T-EVs. The overall protein rank (X-axis) is plotted against the protein area expressed in base-10 log scale (Y-axis). Proteins which have been selected for the screening are marked in green. (b) Schematic representation of EV biogenesis with a zoomed-in vesicle depicting the proteins selected as loading moieties and their theoretical vesicular localization. (c) Western blot on HEK293T cells illustrating the molecular weight (kDa) of each GFP-fused protein (upper panels) and the expression of several EV markers across the samples (lower panels). β-actin was used as loading control.
Figure 3.
Figure 3.
Loading of different GFP-tagged EV markers into UF-qEV purified EVs. (a) Western blot of HEK293T-derived EVs illustrating the molecular weight (kDa) of each GFP-fusion protein (upper panels) and the expression of several EV markers across the samples (lower panels). (b) NTA analysis of isolated EVs showing the total particle concentration (black) and the concentration of GFP positive particles (green) in each sample (N = 3). (c) Table representing the percentage of GFP positive vesicles over total particle count (N = 3). (d) Single vesicle imaging using confocal fluorescence microscopy after spotting EVs onto coverslips (scale bar in the left-bottom corner is 20 µm, all overview images are shown to scale). Individually contrast-enhanced zooms are shown for illustration in the right-bottom corner (scale bar in the upper-right corner is 5 µm).
Figure 4.
Figure 4.
Fluorescence correlation spectroscopy of different GFP-tagged EVs. UF-qEV isolated vesicles, carrying different EV proteins fused to GFP were analysed by FCS without (red data points) and with vesicular disruption by NP40s (blue data points). (a) The diffusion time τ (µs), (b) the molecular brightness CPP (kHz) and (c) the number of particles is shown for each GFP-tagged EV protein. (d) The percentage of freely diffusing GFP fluorescent particles was calculated based on a 2-component FCS analysis using the τdiff of free versus vesicular GFP. (e) Schematic overview of the vesicular disruption by NP40s treatment and the relative observed parameters. (f) Average number of GFP molecules per vesicle based on increase in particle numbers and decrease in molecular brightness upon vesicle disruption by NP40s. Bar graphs indicate representative biological replicates (individual biological replicates are shown in Supplementary Figure 8(b)). Error bars represent standard deviations from at least four technical replicates.
Figure 5.
Figure 5.
Single vesicle quantification of GFP-tagged reporters in CD63 positive vesicles. (a) HEK293T cells were transfected with either CD63-GFP or CD63-mCherry, conditioned media were mixed and EVs were co-isolated by different protocols to assess single vesicle integrity, aggregation or fusion. CD63-GFP-mCherry transfected cells were used as control. (b) 1 µl of EV sample was spotted onto glass slides, covered with 0.17 mm coverslips to generate a thin EV layer, and imaged by high NA confocal or widefield fluorescence imaging. (c) EVs from co-isolated media were detected as light diffraction-limited spots in either green or red channels with negligible occurrence of double-positive vesicles, confirming detection and integrity of single vesicles after UF-SEC (left panel). UC isolation resulted in a small fraction of double-positive vesicles (middle panel). Double labelled control EVs are shown in the right panel. The isolated EVs were analysed by confocal microscopy and the colocalization was quantified based on overlap of the point spread functions in the two fluorescent channels to derive the number of GFP (green, line pattern), mCherry (red) and GFP/mCherry (yellow) double positive vesicles. (d) CD63-GFP/mCherry double labelled control EVs imaged by high NA widefield fluorescence imaging, confirming a high degree of co-localization. The EV point spread function (PSF, bottom left) in comparison to the PSF of 100 nm Tetraspec beads (bottom right) confirms detection as light diffraction-limited spots. (e) Co-localization of different GFP-tagged EV proteins with CD63-mCherry EVs by single vesicle widefield imaging. Images are shown after deconvolution and as maximum intensity projections (MIPs from 5 µm z-stacks; red: CD63-mCherry; green: GFP. scale bar 2 µm). Co-localization quantification as indicated in the image. (f) Percentage of CD63-mCherry (red), GFP (green) and double-positive EVs (yellow) for the different GFP constructs (technical replicates shown as individual bars).
Figure 5.
Figure 5.
(Continued).

Similar articles

See all similar articles

Cited by 1 article

References

    1. Del Conde I, Shrimpton CN, Thiagarajan P, et al. Tissue-factor-bearing microvesicles arise from lipid rafts and fuse with activated platelets to initiate coagulation. Blood. 2005;106:1604–22. - PubMed
    1. Segura E, et al. ICAM-1 on exosomes from mature dendritic cells is critical for efficient naive T-cell priming. Blood. 2005;106:216–223. - PubMed
    1. Viaud S, et al. Dendritic cell-derived exosomes promote natural killer cell activation and proliferation: a role for NKG2D ligands and IL-15Ra. PLoS One. 2009;4:e4942. - PMC - PubMed
    1. Lai RC, et al. Exosome secreted by MSC reduces myocardial ischemia/reperfusion injury. Stem Cell Res. 2010;4:214. - PubMed
    1. Camussi G, et al. Exosome/microvesicle-mediated epigenetic reprogramming of cells. Am J Cancer Res. 2011;1:98–110. - PMC - PubMed

Grant support

This work was supported by the Stiftelsen för Strategisk Forskning [SSF-IRC, FormulaEX];Vetenskapsrådet [VR-Med], as well as by Land Salzburg/IWB/EFRE 2014-2020 P1812596 „EV-TT“ and Land Salzburg/WISS 2025 20102-F1900731-KZP „EV-TT - Bpro“.

LinkOut - more resources

Feedback