Photocleavable lipid nanoprobes for the size selective enrichment of vesicles

Abstract

The present invention provides lipid-based probes and uses thereof for studying vesicles. The probes are suitable for isolating vesicles from a heterogeneous population with high purity and yield. The lipid-based probes also are applicable to study the native state and intrinsic structures of isolated vesicles. The probes are readily amenable to a variety of purification and isolation techniques.

Description

[0001] TITLE OF THE INVENTION

[0002] Photocleavable Lipid Nanoprobes for the Size Selective Enrichment of Vesicles

[0003] CROSS REFERENCE TO RELATED APPLICATIONS

[0004] The present application claims priority to U.S. Provisional Patent Application No. 63 / 556,168, filed February 21, 2024, the disclosure of which is incorporated by reference herein in its entirety.

[0005] STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[0006] This invention was made with government support under Grant Number P30DK045735 awarded by the National Institute of Health. The government has certain rights in the invention.

[0007] BACKGROUND OF THE INVENTION

[0008] Extracellular vesicles (EVs), ranging from exosomes, micro vesicles, and ectosomes to location specific variants such as mitochondrial EVs, encapsulate diverse molecular components from their parent cell, such as proteins and microRNA, to fulfill specific physiological functions (Zhang, Q. et al. 2019. 27, 940-954. e6; Zhang, Q. et al. 2021. Nat. Cell Biol. 23, 1240-1254; Mathieu, M., et al. 2019. Nat. Cell Biol. 21, 9-17; van Niel, G. et al., 2018, Nat. Rev. Mol. Cell Biol. 19, 213-228; Jeppesen, D. K. et al. 2019. Cell 177, 428-445.el8; Clancy, J. W. et al., 2021, Nat. Cell Biol. 23, 1217-1219; Mathieu, M. et al. 2021, Nat. Commun. 12, 4389). It is widely accepted that exosome subpopulations, each with distinct biophysical and biochemical traits, uniquely affect recipient cells, making enriched separation methods for these subpopulations critically important. However, their inherent heterogeneity, coupled with the coexistence of nonvesicular extracellular nanoparticles (NVEPs) like lipoproteins (Kashkanova, A. D. et al. 2023. vesicles 12, 12348), protein complexes, exomeres (Zhang, Q. et al. 2019. Cell Rep. 27, 940-954. e6), and supermeres (Zhang, Q. et al. 2021. Nat. Cell Biol. 23, 1240-1254; Clancy, J. W. et al., 2021, Nat. Cell Biol. 23, 1217-1219), poses substantial challenges to their isolation and thorough characterization (Jeppesen, D. K. et al. 2019. Cell 177; Kowal, J. et al. 2016. Proc. Natl. Acad. Sci. 113, E968-E977). Ultracentrifugation and ultrafiltration methods often result in contamination and variability in sample preparation. While antibody-coupled magnetic beads provide a clean separation of exosomes, they fail to offer crucial size information often requiring additional differential / gradient ultracentrifugation, necessitating bulky equipment. Moreover, antibodybased affinity approaches tend to isolate only specific EV subpopulations based on these surface proteins, missing the inherent heterogeneity of the EV population. Thus, the field still faces challenges in managing yield and sorting heterogeneous populations while ensuring EVs' purity and structural integrity. To fully harness EVs' diagnostic and therapeutic promise (Cheng, L. & Hill, A. F. 2022, Nat. Rev. Drug Discov. 21, 379-399), developing a unified isolation technique that addresses the challenges of high-purity, efficient size-separation of EV subpopulations while retaining their innate heterogeneity and structure and function intact, yield, and scalability for large-scale applications is crucial.

[0009] Recent advancements in vesicle research have spotlighted the transformative capabilities of lipid rafts (Liu, H.-Y. et al. 2021. Adv. Mater. 33, 2008493) and lipid-based nanoprobes (LNs) (Pan, W.-L. et al. 2022. J. Extracell, vesicles 11, e!2281; Marvar, J. et al. 2023. Adv. Mater. Technol. 8, 2201937; Wan, Y. et al. 2017. Nat. Biomed. Eng. 1, 1-11; Liu, X. et al. 2021. Nano Lett. 21, 8817-8823). Instead of solely relying on surface-specific protein markers, these nanoprobes use strong hydrophobic interactions with EV membranes to capture and isolate all EV populations with high purity and without NVEPs. Consequently, LN-based isolations emerge as a promising alternative for cost-effective EV isolation for high purity and yield, offering a competitive edge over conventional strategies. While previous DNA-based lipid affinity-based capture techniques have achieved the isolation of EVs in their native form (Zhang, K. et al. 2019. ACS Sens. 4, 1245-1251), but it requires mandatory and expensive DNase enzyme digestion before obtaining them. To truly understand the role of EV heterogeneity and unlock EVs' diagnostic and therapeutic potential, a straightforward, cost-effective, and reliable method that can isolate native EVs and sort them by size-specific populations is essential.

[0010] Further, lipid-based nanoparticles, encompassing EVs, cell membrane vesicles (CMVs), and biomimetic artificial liposomes, have emerged as transformative tools in drug delivery, gene therapy, and diagnostics. These versatile nanocarriers can encapsulate and deliver a wide range of therapeutic payloads, including small molecules, proteins, and nucleic acids, offering unprecedented potential for addressing complex diseases such as cancer, genetic disorders, and neurodegenerative conditions. Their ability to overcome biological barriers, such as the blood-brain barrier, and enable precise therapeutic targeting has positioned them at the forefront of next-generation therapeutic development.

[0011] EVs, nano-sized vesicles secreted by nearly all cell types, naturally mediate intercellular communication through the transfer of bioactive molecules. This inherent function, coupled with their biocompatibility, low immunogenicity, and ability to traverse physiological barriers, makes them particularly attractive as delivery vehicles. Moreover, their surfaces can be engineered with targeting ligands to enhance specificity and delivery efficiency. Similarly, CMVs, derived from cell membranes, inherit the antigenic and functional profiles of their source cells, conferring unique biomimetic properties that include prolonged circulation, immune evasion, and precise targeting. For example, RBC-derived vesicles leverage surface proteins such as CD47 to evade immune clearance, while tumor derived vesicles exploit homotypic targeting to enhance specificity in cancer therapy. Artificial liposomes further complement these natural systems by offering unparalleled design versatility, allowing precise tuning of lipid composition, size, charge, and drug release profiles to maximize therapeutic efficacy.

[0012] Despite their transformative potential, the clinical translation of EVs, CMVs, and liposomes is hindered by critical challenges in isolation, purification, and characterization, especially at scales required for clinical applications. EVs are typically harvested from complex biological samples, such as cell culture supernatants or bodily fluids, which are rich in contaminants like protein aggregates, lipoproteins, and cellular debris. Similarly, the synthesis of liposomes often produces heterogeneous populations with residual solvents or unencapsulated byproducts. These complexities necessitate advanced purification techniques that preserve nanoparticle integrity while achieving high yields.

[0013] Traditional isolation methods, including ultracentrifugation, density gradient centrifugation, size exclusion chromatography, and tangential flow filtration, are often labor- intensive, time-consuming, and associated with low yields. Furthermore, the high shear forces involved in techniques such as ultracentrifugation can damage vesicle structure, compromising their functionality. These methods also struggle with orientation-specific separation — a significant limitation for CMVs, where right-side-out vesicles (RSOs) retain the native protein orientation critical for biological interactions, while inside-out vesicles (ISOs) may result in immune clearance or off-target effects. The inability to efficiently separate vesicles based on size and orientation presents a bottleneck in the development and translation of lipid nanoparticlebased therapeutics.

[0014] Conventional strategies are limited and have challenges related to yield, purity, scalability, and the preservation of the intrinsic structure of EVs. Thus, there is a need in the art for improved methods for enriching, purifying, and isolating extracellular vesicles from biological samples. This invention satisfies this unmet need.

[0015] SUMMARY OF THE INVENTION

[0016] In some embodiments, the invention provides a lipid nanoprobe (LN) comprising a lipid insert, a cleavable linker, a biocompatible polymer chain, and an affinity tag, wherein the cleavable linker connects the lipid insert and a first end of the biocompatible polymer chain, and wherein the affinity tag is connected to a second end of the biocompatible polymer chain. In some embodiments, the lipid insert is a sterol or a phospholipid. In some embodiments, the phospholipid is phosphoethanolamine. In some embodiments, the phosphoethanolamine is 1,2- distearoyl-sn-glycero-3-phosphoethanolamine (DSPE).

[0017] In some embodiments, the cleavable linker is a photocleavable linker. In some embodiments, the photocleavable linker is an1O2-cleavable linker. In some embodiments, the1O2-cleavable linker is (Z)-2, 2 '-(ethene- 1,2 diylbis(sulfanediyl)diethanamine) (BSDA).

[0018] In some embodiments, the biocompatible polymer chain is poly(ethylene glycol) (PEG).

[0019] In some embodiments, the average molecular weight of PEG is 1-10 KDa. In some embodiments, the average molecular weight of PEG is about 5 KDa.

[0020] In some embodiments, the affinity tag is selected from the group consisting of: biotin, D- biotin, desthiobiotin, avidin, streptavidin, and neutravidin. In some embodiments, the affinity tag is desthiobiotin.

[0021] In some embodiments, the invention provides a method for isolating vesicles from a sample, the method comprising: a) incubating the sample with a lipid nanoprobe (LN) to form LN-tagged vesicles , wherein the LN comprises a lipid insert, a linker, a biocompatible polymer chain, and an affinity tag, wherein the linker connects the lipid insert and a first end of the biocompatible polymer chain, and wherein the affinity tag is connected to a second end of the biocompatible polymer chain; b) capturing the LN-tagged vesicles by incubating the LN-tagged vesicles with a bead, wherein the bead comprises a binding partner for the affinity tag of the lipid nanoprobe; c) isolating the beads; d) releasing the LN-tagged vesicles from the beads; and e) collecting the LN-tagged vesicles .

[0022] In some embodiments, the beads are high-capacity agarose beads (HCABs).

[0023] In some embodiments, the step of isolating the beads further comprises a step of washing the beads one or more times with a sterile buffer, wherein the sterile buffer is water, PBS, HEPES, or Tris.

[0024] In some embodiments, the linker of the lipid nanoprobe is a photocleavable linker.

[0025] In some embodiments, the LN-tagged vesicles are released by exposing the isolated beads to Near-infrared (NIR). In some embodiments, the LN-tagged vesicles are released by contacting the isolated beads with an elution buffer comprising a competitive ligand for the binding partner.

[0026] In some embodiments, the concentration of the competitive ligand in the elution buffer is gradated to preferentially release a specific size of LN-tagged vesicle. In some embodiments, vesicles having a diameter of about 50 nm are preferentially released with an elution buffer having a concentration of the competitive ligand of about 0.1-10 mM.

[0027] In some embodiments, the method further comprises a step of passing the biological sample through a filter before step a). In some embodiments, the filter is a 0.2 pm or 0.4 pm filter.

[0028] In some embodiments, the isolated vesicles maintain their native state.

[0029] In some embodiments, the invention provides a kit comprising the lipid nanoprobe. In some embodiments, the kit further comprises one or more selected from the group consisting of HCABs, a filter, a wash buffer, and an elution buffer. In some embodiments, the HCABs are functionalized with a binding partner for the affinity tag of the lipid nanoprobe.

[0030] BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The following detailed description of embodiments of the invention will be better understood when read in conjunction with the appended drawings. It should be understood that the invention is not limited to the precise arrangements and instrumentalities of the embodiments shown in the drawings. Figure 1 A through Figure ID depicts a representative photosensitive lipid nanoprobe (LN) for isolating high purity vesicles . Figure 1A depicts representative chemical structures of the photosensitive LN: DSPE-NIR-PEG 0.1K-DB with the NIR-induced singlet oxygenXO2 -based cleavage mechanism of (Z)-2,2'-(ethene-l,2-diylbis(sulfanediyl)di ethanamine (BSDA) cleavable linker. (Bottom) The non-photosensitive analogue: DSPE-PEG_n-DB. Figure IB is a schematic depiction of a tri-phase protocol that includes lipid-affinity-based enrichment, purification, and isolation steps, designed to ensure high purity and yield in vesicle isolation. Figure 1C depicts representative data demonstrating de novo docking analysis of PEG- desthiobiotin(DB) affinity tag (left; designed using Avogadro (Hanwell, M. D. et al. 2012. J. Cheminformatics 4, 17)) and competitive binding of D-biotin (right) in the [3-barrel Avidin binding site (PDB ID: 2AVI). Figure ID depicts representative data demonstrating a SEM image of avidin-functionalized high-capacity agarose beads (HCAB) with captured Insulin cell derived- sEV. Scale Bar 1pm. Inset Left: arrows indicate captured sEV. Scale Bar 200 nm. Inset Right: Negative stain TEM image of eluted sEVs. Scale Bar 200 nm. For immunogold TEM experiments anti-CD81 antibodies was conjugated with 4-nm colloidal Gold nanoparticles (1:200) (Jackson ImmunoResearch Laboratories, West Grove, PA).

[0032] Figure 2A through Figure 2G depict representative data demonstrating optimization of HCAB and LN for liposome isolation. Figure 2A depicts representative temporal dependence asses using the colocalization efficiency of FITC and MEMGlow647™ for 60 nm and 140 nm vesicles with 5% LN. Figure 2B depicts representative data demonstrating the use of 5%, 3%, and 1% of LNs (mol / mol% in 3 mM lipid concentration) self-insertion in 60 nm and 140 nm vesicles. Colocalization indicates insertion of the LN into the vesicle. Figure 2C depicts representative data demonstrating the trapping efficiency analyses on neutravidin and monomeric avidin beads 200 pL slurry) with 60 + 140 nm liposome mixture containing DSPE and cholesterol LNs. Mean ± S.E.M (n = 4). Yellow highlight indicates the combination with the highest trapping efficiency. Two-way ANOVA multiple comparisons using Tukey’s method *p < 0.05, ****p < 0.0001. Figure 2D depicts representative data demonstrating the release efficiencies of 60 + 140 nm liposomes mixture isolated with 5% of 2 non-photosensitive probes (CHOL-PEG 5K-DB, DSPE-PEG 5K-DB) eluted over 4 elution steps (El-4) and the photosensitive probe (DSPE-NIR-PEG 0.1K-DB) eluted with a total 10-minute exposure of 850 nm NIR-light at 5 W / cm2. Mean ± S.E.M (n = 4). Figure 2E depicts a representative schematic of the NBD dequenching lipid mixing assay designed to elucidate bulk membrane fusion dynamics. v-SNARE liposomes containing NBD fluorophores (green) quenched by their proximity to Rhodamine fluorophores (pink), are reconstituted with v-SNARE (VAMP2) proteins. t-SNARE liposomes are reconstituted t-SNARE proteins but are devoid of any fluorescent dyes. The superfluous proteins in control v-SNARE and t-SNARE proteoliposomes were eliminated using the traditional ultracentrifugation, whereas for the LN (DSPE-NIR- PEG 0.1K-DB) purified v-SNARE proteoliposomes during the purification phase (refer to Figure IB) and was eluted with 2.5 mM D-biotin. LN was cleaved off one sample of LN purified v-SNARE proteoliposomes using (850 nm at 5 W / cm2) while the other sample contained uncleaved LNs. To block protein-to-protein interactions, negative control experiments incorporated the solubilized cytosolic domain of VAMP2 (CDV). Upon the act of membrane fusion, the spatial confinement of NBD fluorophores and Rhodamine diminishes, leading to the dequenching of NBD, resulting in a fluorescence resurgence. Figure 2F depicts representative data demonstrating a NBD dequenching lipid mixing assay showing normalized NBD fluorescence intensity for the control, cleaved and uncleaved LN of v-SNARE proteoliposomes with (w / ) and without (w / o) CDV. The terminal NDB fluorescence (averaged over the final 10 minutes) difference between samples w / and w / o CDV are listed on the left. Figure 2G depicts representative data demonstrating the bulk fusion rate for the first 30 minutes (linear region) estimated with a linear regression (R2 > 0.95). Secondary axis shows the change in bulk fusion rate between sample with and without CDV.

[0033] Figure 3A through Figure 31 depict representative data demonstrating sEV isolation and characterization from multiple cell types. Figure 3A and Figure 3B depict representative TEM Images of INS-sEV derived from INS cell culture media and Mcp-sEV sourced from primary BMDM cell culture milieu. Scale Bars 100 nm. Left and right insets: spotlight smaller and larger sEV profiles. Scale Bars 20 nm and 50 nm, respectively. Figure 3C and Figure 3D depict representative data demonstrating the NTA-derived size distribution and particle concentration charts, cross-validated with standalone DLS assessments for INS-sEV and Mcp-sEV. The red demarcation underscores the typical sEV size range. Mean ± S.E.M (n = 3). Figure 3E and Figure 3F depict representative data demonstrating dSTORM visualization spotlighting tetraspanin surface markers, namely CD9 and CD81, on the isolated INS-sEV and Mcp-sEV. Scale Bar, 5 pm. The dSTORM evaluation was grounded on the identification of sEVs presenting in excess of 15 localizations of either or both tetraspanin antibodies. Insets: magnify regions representing 16 localizations for CD9 and 21 for CD81. Scale Bar 1 pm. Figure 3G depicts representative images elucidating the diverse tetraspanin expression profiles — singlepositive (CD9+CD81- and CD9-CD81+) and double-positive (CD9+CD81+) configurations. Each line delineates the localization expanse of the highlighted sEVs. Scale Bar lOOnm Figure 3H and Figure 31 depict representative data demonstrating localization length profiles of the identified sEVs with tetraspanin markers, drawn for INS-sEV and Mcp-sEV. Insets: Pie charts showing the abundance of CD9+ (yellow), CD81+ (pink), and the double-positive CD9+CD81+ (orange) against the entirety of detected sEVs. INS-sEV (n = 2174, sEV analyzed), Mcp-sEV (n = 1709, sEV analyzed).

[0034] Figure 4A through Figure 4F depict representative data demonstrating size selective sorting of vesicles dependent on D-Biotin concentration. Figure 4A depicts representative data demonstrating the summary of mean diameters of liposomes with and without LNs of varying PEG lengths. The mean diameter was calculated by fitting a gaussian function on average of all DLS spectra, (n > 3). Figure 4B depicts representative data demonstrating the analysis of release efficiency for 60 nm and 140 nm liposomes integrated with photosensitive probe (DSPE-NIR-PEG 0.1K-DB) across a D-biotin concentration gradient spanning from 0.001 mM to 100 mM. Efficiency metrics are derived from the foundational extruded samples originating from a lipid concentration of 3mM. Mean ± S.D. (n = 3 for each concentration tier). Figure 4C depicts representative data demonstrating the DLS spectra visualizing size differentiation of liposomes contingent on D-biotin concentrations. Scale bar 25% intensity in vertical axis for each spectra, (n = 4). Figure 4D depicts representative data demonstrating NTA analysis of INS-sEV and Mcp-sEV size separation across select D-biotin concentrations: O.OOlmM, 0.1 mM, lOmM, and lOOmM. (n = 3 per concentration). Graphical representations delineate yield in relation to the mean vesicle size, plotted in descending order of increasing D- biotin concentration. Figure 4E depicts a representative summary plot demonstrating a negative correlation between D-biotin concentration and the diameter of the eluted vesicles, encompassing liposomes, INS-sEV, and Mcp-sEV. Each data point represents the global mean diameter calculated over 3 replicas of each concentration with a Lorentzian (Cauchy) fit that best correlate to the peak center. The error bars are the S.D. of the fits. All R2 > 0.9, with the sole exception of O.OOlmM Mcp-sEV. Figure 4F depicts representative data demonstrating terminal NBD fluorescence (last 10 minutes) of lipid mixing assay wit v-SNARE proteoliposomes, juxtaposing model synaptic vesicles proteoliposomes (SV) initially (1) eluted with 2.5 mM D-biotin and model Large Dense Core vesicles (LDCV) proteoliposomes eluted with (2) NIR light using the same HCABs; model liposomes at 1 :200 VAMP2 / lipid composition. Mean ± S.D. for hydrodynamic diameter (n = 2). All relevant DLS spectra can be found in Figure 15.

[0035] Figure 5A through Figure 5F depict representative data demonstrating the use of the photosensitive nanoprobe to enrich EV subpopulation. Figure 5A depicts representative data demonstrating hydrodynamic diameter of EVs sorted via the sequential dual elution strategy. Initially with a 2.5mM D-biotin elution followed by the 10 min explore to NIR- light while in suspension in the buffer. Mean DLS measurements are represented by the horizontal line, (n = 5). Unpaired t-test **** p < 0.0001. Figure 5B depicts representative data demonstrating Western blot analysis for exosomal markers (CD63, CD9, CD81) and synaptic protein Sytl in D- biotin and NIR-eluted samples. Figure 5C depicts representative data demonstrating the RNA- seq-derived pie chart illustrating small-RNA read percentages: miRNA, micro-RNA; tRNA; transfer RNA; snoRNA, small nucleolar; rRNA, ribosomal RNA; ncRNA, non-coding RNA; mRNA, messenger RNA. Figure 5D depicts representative two-color TIRF images (inverse color map) of EVs of Sytl and dual labelled CD63 and CD9. Green and red circles portray the EV area derived custom Mathematica code based on optimized thresholding of puncta (Figure 16cC. Merged profiles (Overlap) demonstrate overlapping puncta for Sytl and CD63 / CD9. Scale Bar 500 nm. Figure 5E depicts a representative Venn diagram depicting identified vesicles in each channel, including colocalized vesicles, for Control (C) and Alzheimer's model (AD) (induced by 100 mM Streptozotocin) samples. Circle sizes indicate EV counts; 1400 vesicles analyzed (n = 2). Figure 5F depicts a representative truncated violin plot of integrated intensity for Syt 1 - associated EVs. A647 corresponds to Alexa Fluor® 647 secondary antibody which was used to identify the integrated fluorescent of a single fluorescent unit. The center solid line indicates the median, dotted line indicates the 25th and 75th percentiles. The median of each population and the median difference between control Vs. Drug for sEVs (D-biotin elution) and larger EVs (NIR elution) are listed. Total of 1400 EVs were analyzed. Two-way ANOVA multiple comparisons using Tukey’s method *p < 0.05, ***p < 0.001.

[0036] Figure 6A through Figure 6C depict representative data demonstrating the monomeric avidin functionalized high-capacity agarose bead characterization. Figure 6A depicts representative SEM images of monomeric avidin beads, before (top) and after (bottom) INS-sEV capture. Scale Bar 10 pm. Figure 6B depicts representative data demonstrating the size distribution of the agarose beads conjugated with monomeric avidin beads. Figure 6C depicts representative data demonstrating the FTIR spectra of avidin conjugated beads and avidin beads with bound lipid nanoprobes. The characteristic amide I (-C=O stretching) and amide II (-C-N stretch and -C-N-H deformation) bands centered at 1640 and 1540 cm respectively, confirmed the avidin functionalization on the agarose beads in both cases. The prominent peak at 1066 cm'1(Varga, Z., et al. 2013. Eur. Polym. J. 49, 2415-2421) characteristic of R-O-P-O-R’ stretching), validated the DSPE-NIR-PEG 0.1K-DB nanoprobe binding, corroborating a previous study (. An additional peak at 2940 cm'1, associated to -CH2- stretching of the stearoyl backbone further supports this.

[0037] Figure 7A through Figure 7D depict representative data demonstrating the integration of LN into liposomes and micelle characterization. Figure 7A depicts representative single vesicle fluorescence images demonstrating colocalization of DSPE-PEG 5K-FITC labeled LNs with ATTO-647N-DOPE labeled liposomes. The merged images identify three distinct particle populations: liposomes with LN, liposomes without LN, and LN-aggregated micelles. Scale Bar 500 nm. Figure 7B depicts representative data demonstrating the analysis of mean pixel intensity across the 488 nm and 647 nm channels for all particles, with background signals minimized by thresholding. The analysis validated three distinct populations (n > 3,000 particles analyzed). Figure 7C depicts representative data demonstrating that LN micelle size distribution was verified using DLS and confirmed by TEM. Inset: A TEM image of a representative micelle. Scale Bar 10 nm. Figure 7D depicts representative data demonstrating that colocalization efficiency reveals the percentage of micelles as a function of varying LN quantities added (1, 3, and 5 mol / mol%). Lower colocalization values indicate a higher number of micelles present in the sample (n = 8).

[0038] Figure 8A through Figure 8G depict representative data demonstrating ab initio docking analysis of affinity tags and their experimental validation. Figure 8A depicts representative data demonstrating the binding affinities (kcal / mol) of 19 modes for two affinity tag variants, desthiobiotin and biotin, along with D-biotin used for competitive elution. Figure 8B depicts the representative ^-barrel avidin binding pocket with bound D-biotin molecule from PDB ID: 2AVI. Figure 8C depicts the representative docking position of mode 1 for D-biotin sourced from PubChem. Figure 8D and Figure 8E depict the representative docking position of mode 1 for PEG-biotin (Figure 8D) and PEG-desthiobiotin, generated using Avogadro software (Hanwell, M. D. et al. 2012. J. Cheminformatics 4, 17) (Figure 8E). Figure 8F depicts a representative SEM image contrasting bare HCAB (left) with HCAB incubated with 3 mM vesicles with lipid nanoprobe (LN) (right), Scale Bar 10 pm. A 200 pL slurry of HCAB shows an average coverage of 80% when incubated with 3 mM of lipids. Figure 8G depicts representative data demonstrating the experimental validation of the release efficiencies of liposomes (a mixture of 60 and 140 nm) captured with LN using biotin and desthiobiotin affinity tags. Desthiobiotin-PEG shows an improvement in release efficiencies by approximately 20%.

[0039] Figure 9A through Figure 9F depict representative data demonstrating the optimization of HCAB and LN for enhanced liposome capture. Figure 9A depicts representative data demonstrating the trapping efficiency of 60 nm liposomes (3mM total lipid concentration) using varying volumes of monomeric avidin-functionalized HCAB, expressed in slurry volume. Given that 4% beaded agarose is provided in a 50% slurry, 100 pL of settled resin equates to 200 pL of slurry. Figure 9B depicts representative data demonstrating the comparison of trapping efficiency for a 60 + 140 nm liposome mixture using new versus old HCAB, regenerated with a regeneration buffer (0.1M glycine, pH 2.8). The old HCAB was tested across four cycles (n = 4). Figure 9C depicts representative data demonstrating the release quantity of a 60 + 140 nm liposome mixture during purification shown as cumulative loss during six wash steps versus the first elution using 2.5mM D-biotin in Buffer A. Figure 9D depicts the representative visualization of 600 pL of slurry (300 pL resin) within 0.8 mL centrifuge columns equipped with 30 pm polyethylene filters: trapped liposomes are indicated by Alexa-647 dye (appearing blue) (top) after six wash steps and (bottom) after six biotin elution steps. Figure 9E depicts representative data demonstrating the trapping and release efficiency for 60 and 140 nm liposomes (3 mM total lipid concentration) integrated with 1, 3, or 5 [mol / mol%] DSPE- PEG 5K-DB. Figure 9F depicts representative data demonstrating the release efficiencies for CHOL-PEG 2K-DB and CHOL-PEG 5K-DB, assessing the impact of PEG spacer length.

[0040] Figure 10A through Figure 10E depict representative data demonstrating the evaluation of photosensitive LN performance. Figure 10A depicts the determination of the representative NIR cleavage efficiency initiated by incubating a liposome-containing sample with DSPE-NIR-FITC and DSPE-PEG 5K-DB. Post incubation, the mixture was introduced into a microfluidic chip functionalized with neutravidin, facilitating the specific capture and subsequent purification of EVs with Buffer A. Figure 10B depicts the subsequent to immobilization, the representative liposomes were subjected to graded NIR light exposures. Quantitative assessment of cleavage efficiency was derived from the average fluorescence intensity metrics pre- and post-exposure. (Inset): TIRF microscopy image exemplifying the immobilized EVs, which underpins the single vesicle fluorescence assessments. Scale Bar 500 nm. (Analysis is based on n > 90 vesicles per each time point). Figure 10C depicts a representative alternative cleavage strategy, direct exposure to UV light using a 365 nm lamp was employed, with intensity variations ranging from l-5mW / cm2(Yamahira, S. et al. 2022. J. Am. Chem. Soc. 144, 13154-13162). Figure 10D depicts representative data demonstrating the release efficiency of sequential exposure intervals, utilizing both CHOL-UV-PEG O.IK-B under UV light and DSPE-NIR-PEG 0.1K-DB under NIR light, both at an irradiance of 5mW / cm2. Representative aggregate data depicting the total release following a sustained 30-minute exposure period. Figure 10E depicts representative data demonstrating the NBD dequenching lipid mixing assay showing normalized fluorescence intensity traces of reconstituted proteoliposomes purified and isolated with CHOL-UV-PEG O.IK-B and DSPE-NIR- PEG_0.1K-DB using 10 min exposure to UV and NIR light. CDV negative controls (w / CDV) were performed to block protein interactions.

[0041] Figure 11A through Figure 11C depict the representative assessment of content retention and encapsulation efficiency post-isolation. Figure 11A depicts the representative visualization of single vesicle fluorescence, showcasing liposomes loaded with dequenched sulforhodamine B (SRB). Scale Bar 500 nm. Figure 1 IB depicts the SRB incorporation into the representative liposomes, providing insight into the labeling process for colocalization study. Figure 11C depicts representative data demonstrating the quantification of colocalization efficiency for liposomes measuring 60 nm and 140 nm in diameter, demonstrating size-selective labeling accuracy. Analysis of the integrated intensity values of SRB within liposomes, compared pre-isolation and post-isolation. This comparative measurement serves as a proxy for assessing the encapsulation efficiency and potential leakage of vesicle contents during the isolation protocol.

[0042] Figure 12A through Figure 12E depict representative data demonstrating the comparative analysis of sEV isolation techniques from INS-cell cultures. Figure 12A depicts representative data demonstrating that consistent sample volumes of 50 mL culture media were harvested from INS-cell cultures at approximately 50% confluence to standardize comparisons across all isolation methodologies. Figure 12B depicts representative negative stain TEM images showcasing sEVs isolated via this DSPE-NIR-PEG 0.1K-DB lipid affinity-based technique, illustrating the method's efficiency and specificity. Figure 12C depicts representative images of sEVs isolated using Total Exosome Isolation Reagent from cell culture media, displaying a polymer-matrix-associated foggy background. This method necessitates subsequent Size Exclusion Chromatography for polymer removal, potentially diminishing yield. Figure 12D depicts representative TEM images of sEVs obtained through a traditional ultracentrifugation protocol (Chhoy, P., et al. 2021. STAR Protoc. 2, 100303), noted for a noisy background and decreased yield. All Scale Bars 100 nm. Figure 12E depicts representative data demonstrating that NTA was utilized to quantify the concentration of particles within the 50-200 nm size range, (n = 3). Notably, *no NTA reading was achievable for the ultracentrifugation sample as the concentration fell below the instrument's sensitivity threshold. A tentative concentration four times lower than comparative methods was estimated based on TEM field of view analysis.

[0043] Figure 13 depicts representative data demonstrating the visualization of INS-cell uptake and internalization of sEVs isolated by different methods. Representative fluorescence image showing INS -cells having taken up and internalized the incubated sEVs, isolated using this DSPE-NIR-PEG_0.8K-DB LN (Left) and commercially available exosome isolation kit by thermos (Right). 15 mL of filtered (0.22 pm filter) culture media was incubated for 1-hour with 100 pm of DSPE-PEG 5K-FITC, fluorescently labelled LN variant. Following which the isolation was performed. In the case of DSPE-NIR-PEG 0.1K-DB, 150 pm of the LN was incubated together with 100 pm of DSPE-PEG 5K-FITC during the incubation step. Post isolation the sEVs were incubated with INS-cells for 2 hours at 37°C, washed, fixed with 2% PFA for 15 mins, washed again, stained with DAPI and imaged by a confocal microscope. The arrow on the merged image indicates the internalized sEVs in both cases. Scale Bar 5 pm.

[0044] Figure 14A through Figure 14E depict representative data demonstrating SMLM d-STORM characterization of sEV heterogeneity from BMDM and INS-cell cultures. Figure 14A and Figure 14B depict representative data demonstrating the localization length distribution for single-positive (CD9+CD81- and CD9-CD81+) and double-positive (CD9+CD81+) sEVs from Mcp-sEV and b) INS-sEV populations, with INS-sEVs further analyzed for double-negative (CD9-CD81-) sEVs using MEMGlow560™ dye. Figure 14C and Figure 14E depict representative data demonstrating the localization circularity data for Mcp-sEV and INS-sEV, respectively, revealing morphological attributes of the isolated vesicles. The dSTORM assessment required sEVs to exhibit more than >15 localizations of tetraspanin markers, ensuring accurate exosome detection. Figure 14E depicts representative data demonstrating the confirmation of exosomal tetraspanin proteins visualized in INS-sEV samples with (bottom) and without (top) anti-CD81 4-nm colloidal gold nanoparticles. Scale Bar 30 nm.

[0045] Figure 15A through Figure 15H depict representative data demonstrating the DLS spectra analysis of liposomes with and without LN modifications. Figure 15A depicts the representative DLS spectra of 60 nm liposomes extruded through 50 nm filters. Figure 15B depicts the representative DLS spectra of 140 nm liposomes extruded from 200 nm filters. Figure 15C depicts the representative DLS spectra of a 50% (v / v) mixture of 60 nm and 140 nm liposomes with no probe. Figure 5D depicts the representative DLS spectra of 60 nm liposomes incorporated with various non-photosensitive lipid nanoprobe (LN) variants. Figure 15E depicts the representative DLS spectra of 140 nm liposomes with all non-photosensitive LN variants. Figure 15F depicts the representative DLS spectra of 60 nm liposomes with 5% (mol / mol% in 3 mM total lipid concentration) DSPE-NIR-PEG 0.1K-DB, shown pre- (top) and post- (bottom) cleavage. Figure 15G depicts the representative DLS spectra of 140 nm liposomes with 5% DSPE-NIR-PEG 0.1K-DB, shown pre- (top) and post- (bottom) cleavage. (All liposome measurements n > 3). Figure 15H depicts the representative control v-SNARE proteoliposomes isolated with 5% DSPE-NIR-PEG_0.1K-DB: initial release of smaller synaptic vesicle (SV) model proteoliposomes was achieved with 2.5 mM D-biotin before PEG-DB cleavage (Left). Subsequently, larger dense-core vesicle (LDCV) model proteoliposomes were released using a sequential NIR-based elution on the same HCAB beads (Right). (All proteoliposome measurements n = 2).

[0046] Figure 16A through Figure 16C depict representative data demonstrating the sizebased isolation of N2A-derived EVs and extended characterization. Figure 16A depicts a representative schematic demonstrating the 50 most abundant miRNAs are depicted based on reads per million (RPM), with a color gradient from red to blue indicating most to least abundant, respectively, for sEVs isolated by biotin elution and larger EVs by NIR elusion. Figure 16B depicts representative data demonstrating the single vesicle immunofluorescence analysis demonstrates the integrated intensity for vesicles dual-labeled with CD63 and CD9 (CD63+CD9- , CD63-CD9+, and CD63+CD9+). The solid line in the center represents the median, and the dotted lines indicate the 25th and 75th percentiles. Two-way ANOVA with multiple comparisons using Tukey’s method indicated no significant difference between the exosome populations. Figure 16C depicts a flowchart outlining the representative single vesicle fluorescence analysis pipeline performed using custom Mathematica code. The process involves: (i) pre-processing the raw channel image with the 'BrightnessEqualize' function, followed by binarization using a custom threshold based on the channel used (647 / 488 / 532 nm), exposure time (100 ms), and laser intensity; (ii) using the binarized image to identify vesicles as non-overlapping bounding disks between 5 and 50 pixels in size to exclude noise and aggregates, utilizing the 'MorphologicalComponents' function to generate the final mask as a matrix; (iii) applying the mask to determine the component size (in pixels) and to extract the integrated intensity from the raw channel image, with the 'ComponentMeasuremenf function calculating mean intensity by dividing integrated intensity by component size. For colocalization, masks were multiplied to identify overlapping regions, and step-iii was repeated with the new overlap mask.

[0047] Figure 17 depicts a representative synthetic scheme for DSPE-NIR-PEG O.IK- DB.

[0048] Figure 18 depicts a representative synthetic scheme for DSPE-NIR-FITC.

[0049] Figure 19 depicts a representative synthetic scheme for CHOL-UV-PEG O.IK-B Figure 20 depicts a representativerH NMR Spectrum of Compound 2. Figure 21 depicts a representative13C NMR Spectrum of Compound 2.

[0050] Figure 22 depicts a representative NMR Spectrum of Compound 3 Figure 23 depicts a representative13C NMR Spectrum of Compound 3. Figure 24 depicts a representative NMR Spectrum of Compound 5. Figure 25 depicts a representative13C NMR Spectrum of Compound 5. Figure 26 depicts a representative NMR Spectrum of Compound 7 (DSPE- NIR-PEG 0.1K-DB).

[0051] Figure 27 depicts a representative13C NMR Spectrum of Compound 7 (DSPE- NIR-PEG 0.1K-DB).

[0052] Figure 28 depicts a representative NMR Spectrum of Compound 17 (DSPE-

[0053] NIR-FITC). Figure 29 depicts a representative NMR Spectrum of Compound 10.

[0054] Figure 30 depicts a representative13C NMR Spectrum of Compound 10. Figure 31 depicts a representativerH NMR Spectrum of Compound 11. Figure 32 depicts a representative13C NMR Spectrum of Compound 11. Figure 33 depicts a representative NMR Spectrum of Compound 12. Figure 34 depicts a representative13C NMR Spectrum of Compound 12.

[0055] Figure 35 depicts a representativerH NMR Spectrum of Compound 14 (CHOL- UV-PEG O.1K-B).

[0056] Figure 36 depicts a representative13C NMR Spectrum of Compound 14 (CHOL- UV-PEG O.1K-B).

[0057] Figure 37A through Figure 37C depict comparative analysis of the purity of plasma-derived exosomes using two different isolation techniques: Invitrogen-plasma isolation kit (Figure 37A), VESIQ-PLN based isolation (Figure 37B), and a bar chart summarizing the purity percentages (Figure 37C), highlighting the superior purity achieved with the VESIQ-PLN method.

[0058] Figure 38A through Figure 38G depict the Programmable Light-Driven (PLD) platform for rapid isolation of extracellular vesicles (evs) for large-scale applications. Figure 38A depicts a schematic of the PLD components: a photocleavable lipid nanoprobe (PLN) and a positively charged polymer (PLL). PLN comprises a hydrophobic lipid insert for EV membrane integration, a UV-sensitive linker (red), a spacer, and an affinity tag (grey circle, representing biotin or other affinity molecules), enabling dual-mode isolation. PLL facilitates electrostatic interactions with negatively charged EV membranes. Figure 38B depicts PLD complex formation through optimized mixing of PLL and PLN at a 25:75 ratio. Figure 38C depicts a representative PLD-mediated EV isolation workflow. The PLD complex binds EVs via electrostatic and hydrophobic interactions. Subsequent UV irradiation cleaves the PLN linker, releasing intact EVs while the non-vesicular contaminants are removed. Figure 38D depicts how nanoparticle tracking analysis (NTA) demonstrates comparable size distributions of EVs isolated via UV- or biotin-mediated elution. Figure 38E depicts quantitative analysis of EV capture efficiency demonstrates significantly higher EV capture within 15 and 30 minutes using the PLD platform compared to controls lacking PLN, highlighting rapid and efficient isolation. Figure 38F depicts fluorescence microscopy images of EVs labeled with MemGlow-488 membrane dye showing successful isolation after 15 mins of PLD platform incubation. Figure 38G depicts a representative cryo-transmission electron microscopy (cryo-TEM) image of an isolated EV, revealing intact, spherical morphology (Scale bar: 50 nm).

[0059] Figure 39A through Figure 39G depicts rapid isolation and functional characterization of biomimetic EVs fabricated via a bottom -up approach. Figure 39A depicts a schematic illustrating the bottom-up assembly of biomimetic EVs, incorporating SNARE proteins, using defined molecular components. Figure 39B depicts a representative NTA profile of isolated biomimetic EVs. Figure 39C depicts a fluorescence microscopy image of isolated vesicles, confirming successful isolation and purification (Scale bar: 5 pm). Figure 39D depicts a representative cryo-transmission electron microscopy (cryo-TEM) image of a single isolated biomimetic EV, showing an intact, unilamellar vesicle structure, indicating structural integrity post-isolation (Scale bar: 50 nm). Figure 39E depicts a size profile analysis of isolated vesicles using dynamic light scattering method at different UV irradiation durations (3, 10, and 20 minutes), demonstrating consistent size distribution (average diameters: 63±22 nm, 63±18 nm, and 65±22 nm, respectively) and confirming that the photocleavable linkers do not alter EV size or stability. Figure 39F depicts a schematic of a lipid fusion assay designed to assess the protein functionality of the isolated vesicles. Donor (V) vesicles containing both RHOD-PE and NBD- PE exhibit FRET, resulting in a fluorescence "OFF" state. UV-triggered cleavage of the linker on the vesicle surface allows for rapid purification via magnetic separation using the PLD platform. Subsequent mixing with target (T) vesicles allows for monitoring of v-SNARE protein functionality post-fusion, restoring NBD-PE fluorescence ("ON" state). Figure 39G depicts the quantification of membrane fusion efficiency using the FRET assay. UV irradiation for 1 minute promotes efficient fusion, demonstrating that the isolated vesicles retain their biological activity. Error bars represent standard deviation (n=3).

[0060] Figure 40A through Figure 40K depict the isolation and characterization of Red Blood Cell (RBC)-Derived Plasma Membrane nanoVesicles (PMnVs) with defined orientations and sizes using PLN probes. Figure 40A depicts a schematic overview of the generation of PMnVs with inside-out (ISO) and right-side-out (RSO) orientations. Microvesicles (2-6 pm) are generated from isolated RBCs via membrane blebbing. Nanovesicles are produced through sequential extrusion of the microvesicles. Figure 40B depicts a representative fluorescence microscopy image of RBC membrane-derived microvesicles labeled with an NBD-PE membrane dye. Length (Figure 40C) and size (Figure 40D) profde analysis of isolated microvesicles. Dithionite incubation destroys the NBD-PE fluorescence of the outer leaflet only. Sodium dithionite at a concentration of 20 mM (solid circles) or buffer (open circles) was added after 5 min at 37 °C to a solution of vesicles containing NBD-PE. The complete loss of the signal after the addition of dithionite shows that all the fluorescence is present in the outer leaflet (Figure 40E) but after extrusion the ratio becomes 43% and 57% indicating ISO and RSO distribution (Figure 40F). Figure 40G depicts a schematic of the strategy for purifying RSO nanovesicles as a drug delivery vector using the PLN probe. Figure 40H depicts a DLS analysis of purified RSO nanovesicles, showing a size distribution of around 100 nm. Figure 401 depicts fluorescence microscopy images of PMnV containing NBD-PE fluorescent membrane dye (Scale bar: 5 pm). Figure 40J depicts the quantification of purified RSO nanovesicles using a dithionite quenching assay, revealing an enrichment of RSO vesicles to 83%. Figure 40K depicts fluorescence microscopy images of purified RSO nanovesicles carrying VAMP2-GFP protein expressed in macrophage cell culture demonstrating the RSO orientation, retention of cargo, and potential for delivery to cells (Scale bar: 20 pm).

[0061] Figure 41A through 41F depicts the light-controlled size-based separation of vesicles using Photocl eavable Lipid Nanoprobes (PLNs). Figure 41 A depicts a schematic illustration of PLN integration into the membranes of vesicles, either synthetically assembled (top) or isolated EVs (bottom). Subsequent UV irradiation triggers PLN cleavage, with the extent of cleavage dependent on light intensity, exposure time, and the number of PLNs per vesicle. This differential cleavage allows for size-based separation. Figure 41B depicts representative Nanoparticle Tracking Analysis (NTA) profiles of vesicles after UV irradiation for 1.5 minutes (top) and 3 minutes (bottom), demonstrating a shift in the size distribution as a function of exposure time. Figure 41C depicts a quantification of vesicle size (diameter) from NTA data after 1.5 minutes and 3 minutes of UV irradiation. Smaller vesicles, with fewer integrated PLNs, undergo more rapid changes in size due to faster PLN cleavage. Error bars indicate standard deviation. Figure 4 ID depicts a bar graph comparing the area under the curve (AUC) of the NTA size distribution profiles after different UV irradiation times, representing the relative abundance of vesicles released. Changes in AUC reflect the light-induced separation. Figure 4 IE depicts dot blot analysis of exosomal markers (CD9, CD63, CD81, and Alix) in vesicles after different durations of UV irradiation (1.5, 3, and 10 minutes in PBS), confirming the exosomal origin of the vesicles and showing that the markers are largely unaffected by the light exposure under these conditions. Figure 41F depicts the quantification of biomimetic EVs size after 1 minute and 3 minutes of UV exposure, demonstrating that smaller vesicles are released in response to shorter UV exposure due to their lower PLN density, while larger vesicles require longer exposure for comparable changes. Statistical significance was determined using a t-test.

[0062] DETAILED DESCRIPTION

[0063] The present invention relates to compositions and methods for isolating vesicles. In some embodiments, the invention provides lipid nanoprobes (LNs) and methods of isolating extracellular vesicles using LNs or compositions comprising LNs. In some embodiments, the LN is used to isolate vesicles from biological samples. In certain embodiments, the LN comprises a lipid insert, an optional linker which can further be optionally a cleavable linker, a polymer- based tether, and an affinity tag. The cleavable linker can function as a bridge between the lipid insert and polymer tether.

[0064] In another aspect, this invention relates to methods of capturing and isolating vesicles using a LN of the invention. In some embodiments, the method comprises contacting the LN with a sample comprising vesicles. The lipid insert of the LN is configured in such a way that it embeds into the membrane of the vesicles and tethers the captured vesicles to a substrate surface through the affinity tag. In some embodiments, the substrate surface is an agarose bead. Once bound to the substrate surface the captured vesicles are purified through filtration and various washing steps. After purification, the vesicles are released from the substrate surface. In some embodiments, the vesicles are released from the substrate surface by competitive elution or cleavage of a cleavable linker in the LN.

[0065] In some embodiments, the vesicles are isolated from a sample by contacting a sample comprising vesicles with an LN immobilized on a substrate surface. In some embodiments, the LN is covalently bound to a substrate surface. In some embodiments, the LN is non-covalently bound to the substrate through the interaction of an affinity tag (e.g., a ligand) conjugated to the LN and a binding partner (e.g., a receptor) conjugated to a surface substrate.

[0066] In some embodiments, the LN is incubated with a sample containing vesicles to form LN-tagged vesicles. The LN-tagged vesicles are captured by incubation with a surface substrate. Tn some embodiments, the substrate surface comprises a binding partner that binds to the affinity tag conjugated the LN. The LN-tagged vesicles bound to the surface substrate are purified using filtration and / or a series of washing steps. In certain embodiments, the vesicles are released from the beads by cleaving a cleavable linker in the LN. In some embodiments, the linker is cleaved by exposing the beads to near-infrared (NIR) light. In certain embodiments, the vesicles are released from the beads by incubating the beads with an elution buffer comprising a competitive ligand that binds to the binding partner on the substrate surface. In some embodiments, the concentration of the competitive ligand in the elution is graduated to preferentially release a specific size of LN-tagged vesicles. In some embodiments, the isolated vesicles maintain their native state.

[0067] Definitions

[0068] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Although any methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, the preferred methods and materials are described.

[0069] As used herein, each of the following terms has the meaning associated with it in this section.

[0070] The articles “a” and “an” are used herein to refer to one or to more than one (z.e., to at least one) of the grammatical object of the article. By way of example, “an element” means one element or more than one element.

[0071] “About” as used herein when referring to a measurable value such as an amount, a temporal duration, and the like, is meant to encompass variations of ±20%, ±10%, ±5%, ±1%, or ±0. 1% from the specified value, as such variations are appropriate to perform the disclosed methods.

[0072] The term “assessing” includes any form of measurement, and also includes determining if an element is present or not. The terms “determining,” “measuring,” “evaluating,” “assessing” and “assaying” are used interchangeably and may include quantitative and / or qualitative determinations. Assessing may be relative or absolute. “Assessing binding” includes determining the amount of binding, and / or determining whether binding has occurred (i .e., whether binding is present or absent). “Assessing activity” includes determining the amount of activity, and / or determining whether an activity has occurred (i.e., whether an activity is present or absent).

[0073] As used herein, the term “bind” or “binding” refers to the specific association or other specific interaction between two molecular species, such as, but not limited to, proteinprotein, lipid-lipid, lipid-protein, and lipid-membrane interactions, for example, the specific association between proteins and their receptors and their ligands, enzymes and their substrates, etc. Such binding may be specific or non-specific, and can involve various noncovalent interactions, such as including hydrogen bonding, metal coordination, hydrophobic forces, van der Waals forces, pi-pi interactions, and / or electrostatic effects.

[0074] “Contacting” refers to a process in which two or more molecules or two or more components of the same molecule or different molecules are brought into physical proximity such that they are able to undergo an interaction. Molecules or components thereof may be contacted by combining two or more different components containing molecules, for example by mixing two or more solution components, preparing a solution comprising two or more molecules such as target, candidate or competitive binding reference molecules, and / or combining two or more flowing components. Alternatively, molecules or components thereof may be contacted combining a fluid component with molecules immobilized on or in a cell or on or in a substrate, such as a polymer bead, a membrane, a polymeric glass substrate or substrate surface derivatized to provide immobilization of target molecules, candidate molecules, competitive binding reference molecules or any combination of these. Molecules or components thereof may be contacted by selectively adjusting solution conditions such as, the composition of the solution, ion strength, pH or temperature. Molecules or components thereof may be contacted in a static vessel, such as a microwell of a microarray system, or a flow-through system, such as a microfluidic or nanofluidic system. Molecules or components thereof may be contacted in or on a variety of cells, media, liquids, solutions, colloids, suspensions, emulsions, gels, solids, membrane surfaces, glass surfaces, polymer surfaces, vesicle samples, bilayer samples, micelle samples and other types of cellular models or any combination of these.

[0075] “Isolated” means altered or removed from the natural state. For example, a vesicle naturally present in a living animal is not “isolated,” but the same vesicle or membrane partially or completely separated from the coexisting materials of its natural state is “isolated.” An isolated vesicle or membrane can exist in substantially purified form, or can exist in a non-native environment such as, for example, a host cell.

[0076] As used herein, the terms “peptide,” “polypeptide,” and “protein” are used interchangeably, and refer to a compound comprised of amino acid residues covalently linked by peptide bonds. A protein or peptide must contain at least two amino acids, and no limitation is placed on the maximum numb. er of amino acids that can comprise a protein’s or peptide’s sequence. Polypeptides include any peptide or protein comprising two or more amino acids joined to each other by peptide bonds. As used herein, the term refers to both short chains, which also commonly are referred to in the art as peptides, oligopeptides and oligomers, for example, and to longer chains, which generally are referred to in the art as proteins, of which there are many types. “Polypeptides” include, for example, biologically active fragments, substantially homologous polypeptides, oligopeptides, homodimers, heterodimers, variants of polypeptides, modified polypeptides, derivatives, analogs, fusion proteins, among others. The polypeptides include natural peptides, recombinant peptides, synthetic peptides, or a combination thereof.

[0077] A “variant” of a lipid nanoprobe can include lipid nanoprobes with a cleavable or non-cleavable linker, additional polymer chains or extension of polymer chains, or affinity or fluorescence tags for tagging vesicles.

[0078] A “ligand” is a protein, molecule, or compound that binds to the receiving protein or molecule.

[0079] An “tag” is a protein, peptide, polypeptide, nucleic acid, or small molecule which is attached to another protein, lipid, nucleic acid, or small molecule. For example, a tag may be an affinity tag which interacts covalently or non-covalently with a binding partner (e.g., biotin and streptavidin, lysophosphatidic acid and lysophosphatidic acid receptor 1 (LPAR1), an aptamer and a glycan, an azide and an alkyne, etc.). Other examples of tags include fluorescent tags (e.g., fluorescein, green fluorescent protein (GFP), etc.), radioactive tags (e.g.,3H,32P, etc.), stable isotope tags (e.g.,2H,13C, etc.), and any other tag known in the art.

[0080] A “binding partner” or “capture molecule” is a protein, peptide, polypeptide, nucleic acid, or small molecule that specifically associates or interacts with another molecular species, such as, but not limited to, proteins, peptides, polypeptides, lipids, or metabolites. For example, the specific association between proteins and their receptors and their ligands, enzymes and their substrates, etc. A “competitive ligand” is a protein, molecule, or compound that binds to a ligand and inhibits or attenuates the interaction of other proteins, molecules, or compounds with that ligand. The competitive ligand can have an increased affinity to the ligand compared to the other interactor.

[0081] A “vesicle” is a lipid bilayer structure. The vesicle can be naturally occurring vesicles within or outside of the cell or synthesized. The vesicle can comprise molecules such as, but not limited to, lipids, proteins, and nucleic acids.

[0082] Throughout this disclosure, various aspects of the invention can be presented in a range format. It should be understood that the description in range format is merely for convenience and brevity and should not be construed as an inflexible limitation on the scope of the invention. Accordingly, the description of a range should be considered to have specifically disclosed all the possible sub-ranges as well as individual numerical values within that range. For example, description of a range such as from 1 to 6 should be considered to have specifically disclosed sub-ranges such as from 1 to 3, from 1 to 4, from 1 to 5, from 2 to 4, from 2 to 6, from 3 to 6 etc., as well as individual numbers within that range, for example, 1, 2, 2.7, 3, 4, 5, 5.3, and 6. This applies regardless of the breadth of the range.

[0083] Description

[0084] The invention is based, in part, on the development of a lipid nanoprobe (LN). In certain embodiments, the lipid nanoprobe is used to isolate vesicles. The LN comprises three domains: a lipid insert, a polymer-based tether, and a tag, and an optional linker domain.

[0085] In some embodiments, a lipid insert is conjugated to a biocompatible polymer. In some embodiments, the biocompatible polymer is further conjugated to a tag.

[0086] In some embodiments, a lipid insert is conjugated to a linker domain. In some embodiments, the linker domain is further conjugated to biocompatible polymer. In some embodiments, the biocompatible polymer is further conjugated to a tag. In some embodiments, the invention is related to compositions comprising an LN. In some embodiments, the compositions are capable of interacting with lipid membranes. In some embodiments, the compositions interact with or embed themselves into the lipid membranes of vesicles. In some embodiments, the compositions are used to isolate membrane-bound vesicles from a sample. In some embodiments, the compositions are used to isolate extracellular vesicles. Exemplary vesicles include, but are not limited to, exosomes, micro vesicles, ectosomes, exomeres, supermeres, synaptic vesicles , secretory vesicles , and large dense core vesicles . In some embodiments, the isolated membranes comprise lipoproteins, protein complexes, and proteoliposomes.

[0087] In some embodiments, the isolated vesicle is a small vesicle, that has a diameter of about 50-200 nanometer (nm). In some embodiments, the diameter of the isolated vesicles can range in size from 20, 30, 40, 50, 60, 70, 80, 90, 100, 150, 200, 250, 300, 350, 400, 450, and 500 nanometers (nm). In some embodiments, the isolated vesicle maintains its native state and intrinsic structure. In some embodiments, the isolated vesicle maintains its integrity and functionality.

[0088] Compositions

[0089] In one aspect, the disclosure provides a lipid nanoprobe (LN). In some embodiments the LN comprises a lipid insert, a polymer tether, and a tag. In some embodiments, the connectivity of the LN is lipid insert-polymer tether-tag. In some embodiments, the LN further comprises a linker. In some embodiments, the connectivity of the LN is lipid insert- polymer tether-linker-tag. In some embodiments, the connectivity of the LN is lipid insert-linker- polymer tether-tag.

[0090] In some embodiments, an LN of the invention comprises a lipid insert domain that embeds or integrates into the membrane of vesicles. In some embodiments, the lipid insert is a triglyceride, a fatty acid, a phospholipid, a sphingolipid, a steroid, a glycolipid, a derivative thereof, or a combination thereof. Examples of suitable fatty acids include, but are not limited to, lauric acid, myristic acid, palmitic acid, stearic acid, arachidic acid, myristoleic acid, palmitoleic acid, sapienic acid, oleic acid, elaidic acid, vaccenic acid, linoleic acid, linoelaidic acid, linolenic acid, arachidonic acid, eicosapentaenoic acid, docosahexaenoic acid, and derivatives thereof. Examples of suitable phospholipids include, but are not limited to, phosphatidic acids, phosphorylethanolamines, phosphatidylcholines, phosphatidylserines, phosphoinositides, and phosphosphingolipids, glycophospholipids, and derivatives thereof. Examples of suitable sphingolipids include, but are not limited to, sphingosine, ceramides, cerebrosides, and gangliosides. Examples of suitable steroids include, but are not limited to, cholestanes, cholanes, pregnanes, androstanes, estranes, and derivatives thereof. In some embodiments, the lipid insert is a cholestane. In some embodiments, the cholestane is cholesterol.

[0091] In some embodiments, the lipid insert is a phospholipid. In some embodiments, the phospholipid is a phosphorylethanolamine. In some embodiments, the phosphorylethanolamine is a stearoyl phosphorylethanolamine. In some embodiments, the stearoyl phosphorylethanolamine is l,2-distearoyl-sn-glycero-3-phosphoethanolamine (DSPE).

[0092] In some embodiments, an LN of the present invention can integrate into vesicle membranes with high efficiency. In some embodiments, the insertion efficiency (how well the LN incorporates itself into the lipid bilayer) is greater than about 30%, greater than about 40%, greater than about 50%, greater than about 60%, greater than about 70%, greater than about 80%, or about greater than 90%.

[0093] In some embodiments, the LN of the invention comprises a linker domain that connects the lipid insert to a polymer tether and tag. In some embodiments, the linker is a non- cleavable linker, wherein the linker is not susceptible to cleavage under biological conditions. Exemplary non-cleavable linkers include any known non-cleavable linker known in the art, such as, but not limited to, peptidyl linkers, polymeric linkers, maleimidocaproyl linkers, and 4- maleimidomethyl cyclohexane- 1 -carboxylate linkers. In some embodiments, the linker is a cleavable linker. In some embodiments, the cleavable linker is an enzymatically cleavable linker. In some embodiments, the cleavable linker is a photocleavable linker, such as, but not limited to, nitrophenyl ether linkers, 4-bromomethyl-3 -nitrobenzoyl linkers, 3-cyanovinylcarbazole linkers, and bromo-coumarin linkers. In some embodiments, the photocleavable linker is cleaved by exposure to ultraviolent (UV) or near-infrared (NIR) light. In some embodiments, the photocleavable linker is an1O2-cleavable linker. In some embodiments, the1O2-cleavable linker is a dialkoxy anthracene linker, bis(sulfanediyl) linker, or derivative thereof. In some embodiments, the1O2-cleavable linker is a (Z)-2,2’-(ethene-l,2,-diyl- bis(sulfanediyl))diethanamide (BSD A) linker.

[0094] In some embodiments, an LN of the invention comprises a biocompatible polymer (a polymer tether domain) that facilitates the binding of the LN with a substrate surface. In some embodiments, the biocompatible polymer comprises one or more elected from the group consisting of: DNA, peptides, polyethylene glycol) (PEG), poly(lactic acid) (PLA), poly(lactic- co-glycolic acid) (PLGA), polypropylene oxide) (PPO), poly(caprolactone) (PCL), poly(y-L- glutamic acid) (PGA), poly(L-phenylalanine ethyl ester) (PAE), poly(L-Lysine) (PLL), poly(L- histidine) (PLH), poly(ethylene amine) (PEI), and derivatives thereof. In some embodiments, the polymer tether comprises PEG. In some embodiments, the average molecular weight of the PEG tether is 1 KDa, 2 KDa, 3 KDa, 4 KDa, 5 KDa, 6 KDa, 7 KDa, 8 KDa, 9 KDa, or 10 KDa. In some embodiments, the PEG tether has an average molecular weight of about 5 KDa.

[0095] In some embodiments, the LN of the invention comprises a tag domain. In some embodiments, the tag is an affinity tag, a fluorescent tag, a radioactive tag, or a stable isotope tag. Examples of suitable affinity tags include, but are not limited to, biotin, desthiobiotin, D- biotin, streptavidin, avidin, NeutrA vidin®, streptavidin-binding protein (SBP), Strep-tag, albumin, AU1 epitope, AU5 epitope, bacteriophage T7 epitope (T7-tag), bacteriophage V5 epitope (V5-tag), bluetongue virus tag (B-tag), calmodulin, calmodulin binding protein, cellulose, cellulose binding protein, chitin, E2 epitope, FLAG epitope, human influenza hemagglutinin (HA), histidine affinity tag (HAT), HSV epitope, KT3 epitope, Myc epitope, PDZ domain, PDZ ligand, polyarginine (Arg-tag), polyaspartate (Asp-tag), polycysteine (Cys-tag), polyhistidine (His-tag), polyphenylalanine (Phe-tag), Sl-tag, VSV-G, an aptamer, a glycan, an azide, and an alkyne. In some embodiments, the LN comprises a fluorescence tag. Examples of fluorescence tags include, but are not limited to, green fluorescent protein (GFP), fluorescein, rhodamine, cyan fluorescent protein (CFP), mCherry, and red fluorescent protein. In some embodiments, the fluorescent tag is a synthetic fluorophore, such as, but not limited to, Cy5 / 7, ATTO, AlexaFluoro, and Rhod-B. In some embodiments, the fluorescence tag is a photoconvertible fluorescent protein. In some embodiments, the fluorescent tag is a photoswitchable fluorescent protein, such as, but not limited to, Dronpa. In some embodiments, the fluorescent tag is used in combination with DNA-based point accumulation for imaging in nanoscale topography (DNA PAINT) or microscopy, such as, but not limited to direct stochastic optical reconstruction microscopy (d-Storm). In some embodiments, the affinity tag is biotin, D- biotin, or desthiobiotin. In some embodiments, the biotin, D-biotin, or desthiobiotin can bind with avidin, streptavidin, or NeutrA vidin® on a surface substrate. Other suitable tag systems include nucleic acid hybridization, aptamer-target binding, protein-protein interaction, and other molecular systems with specific high binding affinity.

[0096] Methods In some embodiments, the present invention provides methods of isolating vesicles with high purity and efficiency. In some embodiments, the method comprises the steps of: a) incubating the sample with an LN of the present invention; b) capturing the LN-tagged vesicles by incubating the LN-tagged vesicles with a bead, wherein the bead comprises a binding partner for the affinity tag of the lipid nanoprobe; c) isolating the beads; d) releasing the LN- tagged vesicles from the beads; and e) collecting the LN-tagged vesicles.

[0097] In some embodiments, the method yields high purity LN-tagged vesicles. In some embodiments, the purity of LN-tagged vesicles is about 50%, about 60%, about 70%, about 80%, about 90%, and about 100%.

[0098] In some embodiments, before step a), the method comprises a step of filtering the sample. Filters used for purifying vesicles include, but are not limited to, a 0.1 micron filter, a 0.2 micron filter, a 0.3 micron filter, a 0.4 micron filter, and a 0.5 micron filter.

[0099] In some embodiments, step a) comprises incubating the sample with an LN of the present invention. Examples of samples comprising vesicles include, but are not limited to, body fluids (e.g., blood, urine, plasma, cerebrospinal fluid (CSF), saliva, etc.) from a subject (e.g., human, non-human mammal, reptile, amphibian, bird, fish, etc.), tissue samples from a subject, and cell culture medium (e.g., primary tissue, immortalized cell culture, extracellular vesicle culture, etc.). In some embodiments, the volume of the sample containing vesicles is between about 1 mL and about IL. In some embodiments, the volume is about 1 mL, about 10 mL, about 25 mL, about 50 mL, about 75 mL, or about 100 mL. In some embodiments, the sample comprising the vesicles is incubated with an LN between about 1 minute and about 2 days. In some embodiments, the incubation time is about 1 hour, 2 hours, 4 hours, 6 hours, 8 hours, 10 hours, 12 hours, 14 hours, 16 hours, 18 hours, 20 hours, 22 hours, or 24 hours. In some embodiments, the LN concentration added to the sample is between about 0.1 mol% to about 20 mol%. In some embodiments, the concentration is about 0.5 mol%, about 1 mol%, about 2 mol%, about 3 mol%, about 4 mol%, about 5 mol%, about 6 mol%, about 7 mol%, about 8 mol%, about 9 mol%, and about 10 mol%. In some embodiments, the amount of LN added to the sample is between about 10 pg and about 10 mg. In some embodiments, the amount of LN added is about 25 pg, about 50 pg, about 75 pg, about 100 pg, about 125 pg, about 150 pg, about 200 pg, about 400 pg, about 600 pg, about 800 pg, about 1 mg, about 2 mg, about 4 mg, about 5 mg, about 6 mg, about 7 mg, about 8 mg, about 9 mg, or about 10 mg. In some embodiments, step b) comprises capturing the LN-tagged vesicles by incubating the LN-tagged vesicles with a substrate surface comprising a binding partner that binds to a tag of the LN. In some embodiments, the substrate is a bead including, but not limited to, agarose beads, polystyrene beads, polyethylene glycol (PEG) beads, dextran beads, magnetic beads, glass beads, plastic beads, silica beads, metal beads, ceramic beads, and any combination thereof. In some embodiments, the agarose bead is a high-capacity agarose bead (HCAB). In some embodiments, the diameter of the bead is between about 1 pm and about 100 pm. In some embodiments, the diameter of the bead is about 1 pm, about 5 pm, about 10 pm, about 15 pm, about 20 pm, about 25 pm, about 30 pm, about 35 pm, about 40 pm, about 50 pm, about 60 pm, about 70 pm, about 80 pm, about 90 pm, and about 100 pm. In some embodiments the substrate surface comprises a binding partner that binds to the tag of an LN. Examples of binding partners include, but are not limited to, avidin, monomeric avidin, streptavidin, or NeutrAvidin®. In some embodiments, the LN-tagged vesicles are incubated with a surface substrate for between about 5 minutes and about 24 hours. In some embodiments, the incubation time is about 5 minutes, about 15 minutes, about 20 minutes, about 30 minutes, about 40 minutes, about one hour, about 1.5 hours, about 2 hours, about 3 hours, about 4 hours, about 5 hours, about 6 hours, 8 about hours, about 10 hours, about 12 hours, about 14 hours, about 16 hours, about 18 hours, about 20 hours, about 22 hours, or about 24 hours. In some embodiments, the LN-tagged vesicles are incubated with a surface substrate with stirring, with shaking, with rocking, or with a combination thereof. In some embodiments, the incubation is undisturbed by stirring, shaking, and / or rocking.

[0100] In some embodiments, non-specific binding and superfluous protein reconstitutions are removed from LN-tagged vesicles bound to a surface substrate (e.g., a bead) through various isolation and purification steps c). In some embodiments, step c) comprises isolating beads and bound LN-tagged vesicles. Examples of isolation methods include, but are not limited to, filtration, centrifugation, precipitation, magnetic isolation, and decanting. In some embodiments, beads are isolated by filtering the sample containing LN-tagged vesicles bound to beads. In some embodiments, the filter is a 10 pm filter, 20 pm filter, 30 pm filter, 40 pm filter, or 50 pm filter. In some embodiments, the beads are isolated by centrifuging the sample containing LN-tagged vesicles bound to beads and removing the supernatant. In some embodiments, the sample is centrifuged at about 100 * g, about 150 * g, about 200 x g, or about 300 x g in some embodiments, the sample is centrifuged for about 10 seconds, about 20 seconds, about 30 seconds, about 45 seconds, about 60 seconds, about 2 minutes, about 3 minutes, about 4 minutes, or about 5 minutes. In some embodiments, captured LN-tagged vesicles bound to beads are subjected to one or more washes, or rinses, to remove unbound impurities. In some embodiments, the beads are subjected to one wash cycle, two wash cycles, three wash cycles, four wash cycles, five wash cycles, or six wash cycles. Examples of suitable wash buffers include, but are not limited to, sterile water, saline, phosphate buffered saline (PBS), tris(hydroxymethyl)aminomethane (Tris), 4-(2 -hydroxy ethyl)-l -piperazineethanesulfonic acid (HEPES), and piperazine-N,N'-bis(2-ethanesulfonic acid (PIPES). Other additional components of the wash buffers that may be useful include salts, reducing agents, and detergents. Additional enrichment and purification steps can include the use of antibodies, size-based filtration, or other forces such as electrostatic force, di electrophoretic force, gravity, centrifugal force, and any combination thereof.

[0101] In certain embodiments, step d) comprises releasing the captured LN-tagged vesicles from the substrate surface (e.g., a bead). In some embodiments, the captured LN-tagged vesicles are released from a substrate surface by contacting the LN-tagged vesicles and bound substrate with a competitive ligand for the binding partner on the substrate. A competitive ligand may be the same as the tag in the LN or may be a different ligand which also binds to the binding partner of the substrate surface. As an example, LN-tagged vesicles comprising a desthiobiotin tag bound to a substrate surface by an avidin binding partner are released by contacting the substrate with a competitive ligand, such as D-biotin. In some embodiments, the captured LN- tagged vesicles are incubated in an elution buffer comprising a competitive ligand. In some embodiments, the concentration of the competitive binding partner is between about 0.001 pM and about 10 mM. In some embodiments, the concentration is about 0.001 pM, about 0.05 pM, about 0.1 pM, about 0.2 pM, about 0.3 pM, about 0.4 pM, about 0.5 pM, about 0.6 pM, about 0.7 pM, about 0.8 pM, about 0.9 pM, about 1 pM, about 1.5 pM, about 2 pM, about 3 pM, about 4 pM, about 5 pM, about 6 pM, about 7 pM, about 8 pM, about 9 pM, about 10 pM, about 20 pM, about 30 pM, about 40 pM, about 50 pM, about 60 pM, about 70 pM, about 80 pM, about 90 pM, about 100 pM, about 200 pM, about 300 pM, about 400 pM, about 500 pM, about 600 pM, about 700 pM, about 800 pM, about 900 pM, about 1 mM, about 2 mM, about 3 mM, about 4 mM, about 5 mM, about 6 mM, about 7 mM, about 8 mM, about 9 mM, or about 10 mM. In some embodiments, the concentration of competitive ligand in the elution buffer is gradated to preferentially release a specific size of LN-tagged vesicle.

[0102] In certain embodiments, the captured LN-tagged vesicles are released from the substrate surface by cleaving a linker in the LN. In some embodiments, the linker is a photocleavable linker and the LN-tagged vesicles are released by exposure of the LN-tagged vesicles bound to a surface substrate by exposure to light. In some embodiments, the light source is ultraviolet (UV) light, visible light, infrared light (IR), or near-IR light (NIR). In some embodiments, the LN-tagged vesicles are exposed to a light source for between about 1 minute and about 30 minutes. In some embodiments, the LN-tagged vesicles bound to a surface substrate are exposed to a light source for about 1 minute, about 2 minutes, about 3 minutes, about 4 minutes, about 5 minutes, about 6 minutes, about 7 minutes, 8 minutes, about 9 minutes, about 10 minutes, about 12 minutes, about 14 minutes, about 16 minutes, about 18 minutes, about 20 minutes, about 25 minutes, or about 30 minutes. In some embodiments, the LN-tagged vesicles are exposed to light with a wavelength between about 200 nm and about 1,000 nm. In some embodiments, the wavelength is about 200 nm, about 250 nm, about 300 nm, about 350 nm, about 400 nm, about 450 nm, about 500 nm, about 550 nm, about 600 nm, about 650 nm, about 700 nm, about 750 nm, about 800 nm, about 850 nm, about 900 nm, about 950 nm, or about 1,000 nanometers (nm). In some embodiments, the intensity of the light is between about 0.5 mW / cm2and about 25 mW / cm2. In some embodiments, the intensity of light is about 0.5 mW / cm2, about 0.75 mW / cm2, about 1 mW / cm2, about 1.25 mW / cm2, about 1.5 mW / cm2, about 1.75 mW / cm2, about 2 mW / cm2, about 2.5 mW / cm2, about 3 mW / cm2, about 3.5 mW / cm2, about 4 mW / cm2, about 4.5 mW / cm2, about 5 mW / cm2, about 5.5 mW / cm2, about 6 mW / cm2, about 6.5 mW / cm2, about 7 mW / cm2, about 7.5 mW / cm2, about 8 mW / cm2, about 9 mW / cm2, about 10 mW / cm2, about 15 mW / cm2, about 20 mW / cm2, or about 25 mW / cm2. In some embodiments, the LN-tagged vesicles bound to a surface substrate are exposed to one or more cycles of light. In some embodiments, the LN-tagged vesicles bound to a surface substrate are exposed to one cycle, two cycles, three cycles, or four cycles of light exposure. In some embodiments, releasing the captured LN-tagged vesicles from the substrate surface comprises a combination of competitive elution and light-induced elution.

[0103] In some embodiments, step e) comprises collecting the isolated, purified vesicles. In some embodiments, the isolated vesicles are collected by centrifugation. In some embodiments isolated vesicles are centrifuged at between about 10 * g and about 200,000 *g. In some embodiments, the EVs are centrifuged at about 10 x g, about 50 x g, about 100 x g, about 150 x g, about 200 x g, about 300 x g, about 400 x g:about 500 x g;about 600 x g?about 700 x g, about 800 x g, about 900 x g, about 1,000 x g, about 2,000 x g, about 2,000 x g, about 4,000 x g, about 5,000 x g, about 6,000 x g?about 7,000 x g, about 8,000 x g, about 9,000 x g;about 10,000 x g, about 20,000 x g, about 30,000 x g, about 40,000 x g, about 50,000 x g, about 60,000 x g, about 70,000 x g, about 80,000 x g, about 90,000 x g, about 100,000 x g, or about 200,000 x g. in some embodiments, the isolated LN-tagged vesicles are centrifuged for between about 1 second and about 5 hours. In some embodiments, the LN-tagged vesicles are centrifuged for about 1 second, about 2 seconds, about 3 seconds, about 4 seconds, about 5 seconds, about 6 seconds, about 7 seconds, about 8 seconds, about 9 seconds, about 10 seconds, about 20 seconds, about 30 seconds, about 45 seconds, about 60 seconds, about 2 minutes, about 3 minutes, about 4 minutes, about 5 minutes, about 6 minutes, about 7 minutes, about 8 minutes, about 9 minutes, about 10 minutes, about 15 minutes, about 20 minutes, about 30 minutes, about 40 minutes, about 50 minutes, about 60 minutes, about 2 hours, about 3 hours, about 4 hours, or about 5 hours. In addition, other methods useful for purifying and collecting vesicles can include differential ultracentrifugation, size-exclusion chromatography, vesicle precipitation with polymer matrices, ultrafdtration, density gradient ultracentrifugation, immuno-affinity capture, and any combination thereof.

[0104] In some embodiments, the self-insertion efficiency (how well the LN incorporates itself into the lipid bilayer) of the LN is about is greater than about 30%, such as greater than about 40%, about 50%, about 60%, about 70%, and even greater than about 80% and about 95%.

[0105] In some embodiments, the trapping efficiency (how well the vesicles are captured by the substrate surface; bulk fluorescence measurements based on total vesicles present in the sample) of substrate surface is greater than about 30%, such as greater than about 40%, about 50%, about 60%, about 70%, and even greater than about 80% and about 95%.

[0106] In some embodiments, the release efficiency (how well the vesicles are released from the substrate surface; bulk fluorescence measurements based on total) of methods described herein are greater than about 30%, such as greater than about 40%, about 50%, about 60%, about 70%, and even greater than about 80% and about 95%. In some embodiments, a LN of the present invention is first tagged with beads (like a fishing technique) and then incubated with the sample.

[0107] In certain embodiments, the isolated vesicles can be analyzed for their contents such as lipids, proteins, and nucleic acids using typical procedures and protocols. For example, after isolation the vesicles can be released from the substrate surface and the lipid, protein, and nucleic acid content from the isolated and released vesicles can then be extracted and analyzed. For example, protein can be extracted from isolated EVs and analyzed via western blot analysis. In addition, RNA can be extracted from isolated vesicles and analyzed via RNA Seq. In some embodiments, morphology of the isolated vesicles can be characterized using DNA-based point accumulation for imaging in nanoscale topography or microscopy, such as, but not limited to direct stochastic optical reconstruction microscopy.

[0108] In certain embodiments, the isolated vesicles are nano plasma membrane vesicles (NPMVs). In some embodiments, the isolated vesicles can exist in an inside-out (ISO) orientation, wherein the interacting proteins on the plasma membrane face inward, or a rightside-out (RSO) orientation, wherein the interacting proteins on the plasma membrane face outward. In some embodiments, the sample comprises a mixture of vesicles in an ISO and RSO orientation and the method is selective in isolating vesicles of the RSO orientation or of the ISO orientation.

[0109] Benchtop Kit for Wide Accessibility

[0110] In certain embodiments, the disclosure provides kits for isolating vesicles. In some embodiments, the kit comprises an LN of the present invention. In some embodiments, the kit further comprises one or more selected from the group consisting of: agarose beads, wash buffers, elution buffers, filters, columns, tubes, and instructive material. In some embodiments, the kit comprises a positive control sample and / or a negative control sample.

[0111] EXPERIMENTAL EXAMPLES

[0112] The invention is further described in detail by reference to the following experimental examples. These examples are provided for purposes of illustration only, and are not intended to be limiting unless otherwise specified. Thus, the invention should in no way be construed as being limited to the following examples, but rather, should be construed to encompass any and all variations which become evident as a result of the teaching provided herein.

[0113] Without further description, it is believed that one of ordinary skill in the art can, using the preceding description and the following illustrative examples, make and utilize the present invention and practice the claimed methods. The following working examples therefore are not to be construed as limiting in any way the remainder of the disclosure.

[0114] Example 1 : Photocleavable Lipid Nanoprobes for the Size Selective Enrichment of Vesicles: Isolating Extracellular Vesicle Subpopulations

[0115] The biggest challenge in the current separation methods for lipid bilayer- encapsulated vesicles, such as exosomes, secretory, and synthetic vesicles, lies in the absence of a unified approach that seamlessly delivers high-purity, efficiency, yield, and scalability for large-scale applications. Addressing this gap, this study developed an innovative approach leveraging light-responsive lipid nanoprobes (LNs), explicitly engineered for the size-specific isolation and enrichment of vesicle subpopulations, particularly focusing on the sub-200 nm in size. This unique design facilitates seamless integration with various vesicle types and extracellular vesicles (EVs) and enhances precision in size-specific vesicle isolation. These LNs are synthesized with a photosensitive moiety and affinity tags, allowing for the controlled release and isolation of vesicles under precise light modulation and competitive elution. The nanoprobes undergo cleavage upon light activation, disengaging chemical linkers and presenting the vesicles in their native form. Notably, this method is tailored to meet large-scale isolation requirements, ensuring reliable scalability for industrial and clinical applications. This method opens new avenues in exploring, analyzing, and utilizing vesicles and EV subpopulations in various biomedical applications, including diagnostics, therapeutic delivery, and biomarker discovery.

[0116] Development of Photosensitive Lipid-affinity-based Approach for Native Vesicle Isolation

[0117] Our vesicle isolation method relies on two principal components: the lipid nanoprobe (LN) and the associated capture unit. The LN design features a range of constructs, including those with a photocleavable linker - DSPE-NIR-PEG O.1K-DB (Figure 1 A - Top) and their non-photocleavable counterparts - DSPE-PEG_n-DB (Figure 1 A - Bottom). All LN constructs used in the study can be found in Table 1. Structurally, the nanoprobe incorporates four main components: a lipid insert which ensures a robust interaction with the lipid bilayers, a photosensitive linker for regulated release upon near-infrared (NIR) light exposure, a PEG-based tether for the strategic spatial separation of the vesicles, and an biotin-derivative based tag for affinity capture and retention also enabling a secondary mode of release via competitive elusion with Biotin containing buffer.

[0118] Table 1: Lipid Nanoprobes used in Vesicle Isolation and Vesicle Labeling. aLNs were synthesized inhouse and the synthesis process and characterization are described below.bLNs were purchased from NANOCs

[0119] As a capture unit, high-capacity monomeric avidin- functionalized agarose beads (HCABs) were employed (Figure 6A). The HCABs, with an average diameter of approximately 25 pm (Figure 6B), present a pronounced difference in size, being approximately five times larger than previously reported magnetic counterparts (Wan, Y. et al. 2017. Nat. Biomed. Eng. 1, 1-11; Zhang, K. et al. 2019. ACS Sens. 4, 1245-1251; Shao, H. et al. 2015. Nat. Commun. 6, 6999). This dimensional advantage effectively minimizes crosstalk between larger vesicles and multiple beads. Yet, their size remains optimally compact, ensuring an enhanced surface-to-volume ratio compared to a flat capture surface, amplifying both vesicle capture efficiency and cost-effectiveness. To facilitate vesicle capture, LNs were incubated in a vesicle-containing sample. Once the LNs incorporate themselves into the vesicle membranes, they're captured by HCABs (Figure IB - D). Alternatively, the LNs can also be first tagged with beads (like a fishing technique) and then incubated with the sample. The capture of the LNs on the Beads were verified by identifying the presence of DSPE using the characteristic of R-O-P-O-R’ stretching (Varga, Z , Mihaly, et al. 2013. Eur. Polym. J. 49, 2415-2421) (Figure 6C). The PEG component, bridging the affinity tag and lipid insert, acts dual fold, preventing non-specific vesicle attachment to the HCABs and functioning as an adaptive spacer. After vesicle capture and enrichment (Figure IB & D), their isolation was accomplished in one of two ways (Figure IB). The first was introducing a competitive elution process with D-biotin, effectively displacing the LN-bound biotin derivative (Figure 1C). The alternative method involved a photolytic approach wherein light-induced cleavage was initiated, specifically targeting the PEG-affinity tag subunit.

[0120] Our method enables precise vesicle isolation and unlike other lipid- Nanoprobes (Wan, Y. et al. 2017. Nat. Biomed. Eng. 1, 1-11), offers a choice between chemical or light-based separation based on the application's needs. More importantly, by the ability to cleave off the functional unit, this technique bridges a significant gap by enabling the isolation of native vesicles directly from complex biological environments.

[0121] Optimization and Characterization of LNs for Optimal Vesicle Isolation

[0122] Validation of LN Self-insertion Efficiency and Affinity Tag Efficacy To investigate the temporal dependence of this approach, a fluorescent alternative of the DSPE-based LN was synthesized, replacing the desthiobiotin affinity tag with a fluorescein moiety (DSPE-NIR-FITC and DSPE-PEG 5K-FITC). In the initial studies, a 5% (which will refer to mol / mol%) probe was incubated with liposomes of two sizes (60 nm and 140 nm) labeled with ATTO 647N-DOPE fluorescent lipid. The colocalization efficiency, serving as a proxy for self-insertion efficacy, increased notably within an hour, showing a nearly 70% rise in colocalization for the 140 nm liposomes (Figure 2A). Conversely, the 60 nm liposomes exhibited reduced colocalization, possibly due to photobleaching effects, combined by the limited probe incorporation due to the vesicle size. However, for native EVs, longer incubation times, preferably overnight, are recommended for optimal EV enrichment based on the sample volume and HCAB resin quantity. Experimentation with different LN concentrations (1%, 3%, and 5%) confirmed the expected inverse relationship: lower probe concentrations resulted in fewer integrated probes (Figure 2B). Interestingly, the unincorporated fluorescent components (Figure 7A-B) formed a small subset of aggregated LNs which subsequent analyses identified as micelles with a size range <10nm (Figure 7C). Previous studies reported similar sized micelles, with DSPE-PEG 2K based LNs (Sun, F. et al. 2018. Mol. Pharm. 15, 3786-3795). Similarly, these micelles were only visible in LNs with a larger PEG spacer (DSPE-PEG 5K-FITC) predominantly at lipid concentration exceeding 3% of 3 mM lipids (Figure 7D), indicating a critical micelle concentration < 90 pM, a threshold occasionally reached in large sample volumes.

[0123] In efforts to improve on the traditional biotin-based affinity tag, the performance of biotin with its lower-affinity counterpart desthiobiotin was compared. While biotin derivatives have been used in LNs for vesicle sequestration (Wan, Y. et al. 2017. Nat. Biomed. Eng. 1, 1-11), their release methods have not been extensively studied. The initial Ab initio docking simulations indicated that adding a PEG segment reduced its affinity within the - barrel binding pocket (Figure 1C & 8A-E). The incorporation of desthiobiotin further reduced this affinity (Figure 8A-E). Even though in both cases the LN incorporated vesicles were able to be captured efficiently (Figure 8F), this lower affinity enabled the competitive displacement of these bound vesicles with solubilized D-biotin elution buffer. True to the simulated binding affinities, experiments validated that desthiobiotin based affinity tag enhanced competitive elution, facilitating larger target recovery with gentle elution conditions (Figure 8G).

[0124] Optimization of Vesicle Capture and Release Next, the efficacy of DSPE (l,2-distearoyl-sn-glycero-3- phosphoethanolamine) and cholesterol as potential lipid tags was examined, while also examining their synergistic performance with HCABs functionalized with monomeric avidin and neutravidin, commonly used low binding streptavidin derivatives. Results showed that the monomeric avidin functionalized HCABs consistently seemed to have better vesicle trapping efficiencies compared to neutravidin HCABs (Figure 2C). Despite its documented lower biotinbinding affinity (Kohanski, R. A. & Daniel Lane, M. 1990. Methods in Enzymology. Academic Press, vol. 184 194-200), it is believed that the compact structure of monomeric avidin offers enhanced accessibility. Thus, this cost-effective variant is also less susceptible to nonspecific interactions and aggregation, resulting in a higher grafting densities with a larger biotin binding capacity (Janolino, V. G. et al. 1996. Appl. Biochem. Biotechnol. 56, 1-7). Strikingly, the synergy between monomeric avidin and DSPE yielded a trapping efficiency of nearly 75% for a 60 and 140 nm liposome mixture (3 mM total lipid cone.), a significant (>20%) improvement compared to other combinations. It was possible to further enhance the efficiencies to ~ 95% with larger quantities of HCAB (Figure 9A). Even though cholesterol based probes have been noted to incorporate better in cell membrane bilayers and EVs (Wan, Y. et al. 2018. Cancer Res. 78, 798-808; Wan, Y. et al. 2019. Lab. Chip 19, 2346-2355), their binding to the HCABs seemed compromised. Similar to this observation, a recent study reported that at higher concentrations cholesterol based LNs with a biotin moiety had lower avidin binding due to steric hindrances possible due to the formation of cholesterol lipid domains after bilayer incorporation (Vabbilisetty, P. et al. 2018. ACS Omega 3, 1589-1599). Furthermore, the reusability of the regenerated HCAB was evaluated and results showed that these beads remained functional for up to four cycles, with only a 10-30% loss in trapping efficiency across repeated uses (Figure 9B).

[0125] During the purification phase, the stability of captured liposomes was monitored and results showed that there was a cumulative loss of - 5% across six wash cycles, indicating successful vesicle retention during this process (Figure 9C). In the elution phase, the two methodologies: competitive elution and light-activated elution (Figure 2D) were compared. Both methods can be used interchangeably to achieve high-purity vesicles. However, the light cleavage removes any residual PEG and desthiobiotin moiety from the linker and provide a near native vesicle (Figure IB).

[0126] For competitive elution, initially, a buffer containing 2.5 mM of D-biotin was used. Cholesterol-based LN integrated liposomes had a slightly lower release efficiency compared to DSPE-based LNs integrated liposomes. Results showed that smaller liposomes had a larger release efficiency. A cumulative of 6 elutions was required to achieve >50% elution for smaller liposomes (Figure 9D & E). In fact, near 100% release efficiency was achieved with 3% LNs for 60 nm vesicles (Figure 9E). Results showed that increasing the PEG spacer length correlated with an increase in the release efficiency per elution (Figure 9F). As discussed before, this would come at the cost of the co-isolation of LN micelles in concentrated liposome samples (Figure 7B). Light-activated elution was possible via a photosensitive cleavage mechanism using a102-cleavable linker, (Z)-2,2'-(ethene-l,2 diylbis(sulfanediyl)diethanamine (BSDA) (Li, J. et al. 2019. J. Am. Chem. Soc. 141, 4073-4079), which bridges between PEG chains and DSPE (Figure 1 A). Typically, common photosensitive linkers use reactive oxygen species cleavage which require ultraviolet (UV) or NIR light in combination with a photosensitizer to generate reactive oxygen species (ROS) (Zhou, H. et al. 2023. Nano Lett. 23, 3661-3668; Duosiken, D. et al. 2022. J. Am. Chem. Soc. 144, 2455-2459). These linkers often disrupt vesicle carriers and are used in pharmaceutical drug delivery applications (Lu, M. et al. 2022. Adv. Mater. 34, 2204765; Qin, Y. et al. 2019; J. Am. Chem. Soc. 141, 8943-8950; Zhou, H. et al. 2023. Nano Lett. 23, 3661-3668). It is also known that high-intensity near-infrared (NIR) and UV light can induce in situ ROS (Qin, Y. et al. 2019. J. Am. Chem. Soc. 141, 8943-8950; Golovynska, I., Golovynskyi, S. & Qu, J. 2023. Photochem. Photobiol. 99, 106-119; Jockusch, S. et al. 2008. Photochem. Photobiol. Sci. 7, 235-239; Kochevar, I. E. 2004. Sci. STKE. pe7- pe7). By directly applying non-invasive light, ROS was generated at a level sufficient to cleave the sulfur double bond to release the vesicle (Figure 2D). To quantify1O2-cleavable linker cleavage per vesicle experiments were carried out to measure single vesicle fluorescence analysis with an LN fluorescent variant (DSPE-NIR-FITC) that was incorporated together with this nonphotosensitive variant (DSPE-PEG 5K-DB), which allowed for the capture of the vesicles in a neutravidin coated slide (Figure 10A). Results showed -40% cleavage for a total of 10 min exposure to 850 nm (5mW / cm2) of NIR light (Figure 10B). This led to the release -20% of the trapped vesicles in the 1stelution (Figure 2D). Experiments then compared it with a reported UV- responsive isoform (Yamahira, S. et al. 2022. J. Am. Chem. Soc. 144, 13154-13162), using direct UV cleavage (Figure 10C). Results showed a progressively increasing temporal dependence compared to the NIR probe, but it resulted in a -8% decrease in total release efficiency for a total of 10 min exposure (Figure 10D).

[0127] Vesicle Integrity and Protein Functionality

[0128] To ensure that this isolation technique preserves protein functionality, a lipid mixing assay was conducted, a pivotal test for verifying the biological activity of proteins within isolated vesicles. In this assay, liposomes were reconstituted with soluble N- ethylmaleimide sensitive factor attachment protein receptors (SNAREs), which play a central role in mediating membrane fusion — a fundamental process in various cellular functions. SNAREs consist of two types: v-SNAREs located on vesicle membranes and t-SNAREs on target membranes. They form four-helix bundles that bring these membranes together, facilitating fusion. This experiment involved reconstituting the neuronal / exocytotic v-SNARE vesicle-associated membrane protein VAMP2 into liposomes (final lipid to protein ratio « 200: 1). These liposomes also contained a fluorescence resonance energy transfer (FRET) pair of dye-labeled lipids, NBD-DOPE and Rhodamine-DOPE. In this context, the NBD signal serves as a readout for lipid-mixing kinetics. When fusion occurs, the spacing between the FRET dye pair increases, allowing the NBD signal's changes to be tracked. Importantly, it's worth noting that experiments subjected only the control proteoliposomes to the standard ultracentrifugation for excess protein removal after reconstitution. For the other samples, liposomes were incubated with DSPE-NIR-PEG 0.1K-DB and isolated, where the excess proteins were removed during the purification stage. Part of the isolated sample was then exposed to light to cleave the functional PEG-DB component in LN-cleaved sample (Figure 2E).

[0129] To provide a comprehensive assessment, the lipid mixing assay results were compared with and without the presence of soluble cytosolic domain v-SNAREs (CDV), which acted as a competitive inhibitor. Experiments were then used to evaluate if this isolation technique altered the surface protein interactions between t-SNAREs and v-SNAREs. The bulk fusion rate of the first 30 minutes was evaluated by fitting a linear regression. Firstly, when LN were present on the liposome surface, the change in the rate of fusion with and without CDV decreased significantly (Figure 2G). Secondly, by cleaving the LN, the results approached the control levels and also accentuated the effect of CDV on blocking protein interactions. This could also be seen by the terminal NBD fluorescence difference, where the cleaved LN purified v-SNARE showed the largest difference (Figure 2F). These results not only strongly support the use of this isolation technique in such cases but also emphasize the need for LN cleavage for the preservation of surface protein functionality imperative to retain the biological activity of isolated vesicles. In a parallel investigation, the performance of this UV-responsive isoform was assessed. Notably, results showed that prolonged UV exposure (for more than 10 minutes) rendered vesicles non-functional, ultimately compromising protein activity (Figure 10E).

[0130] Additionally, to ensure that this technique effectively retained vesicle contents without leakage, 60 and 140 nm liposomes loaded with sulforhodamine B (SRB) dye were synthesized and used as model cargo. Using single-vesicle fluorescence analysis, experiments assessed colocalization efficiency and SRB integrated intensity in colocalized vesicles before and after isolation (Figure 11). This analysis helped demonstrate content preservation showing that the SRB dye, serving as a representative marker for vesicle cargo, remained intact in the colocalized vesicles throughout the isolation process.

[0131] Small Extracellular Vesicle (sEV) Isolation

[0132] Our attention then turned to the isolation of sEVs using the photosensitive LNs. To enhance the purity of the sEV samples, an additional 0.22 pm filtration step was introduced, a departure from conventional liposome and proteoliposome isolation protocols. This step was essential to exclude any cellular debris and dead cells. Importantly, this protocol eliminated the need for ultracentrifugation, streamlining the entire isolation process.

[0133] Experiments were conducted with two different cell cultures: a primary macrophage (Mcp) cell line and an insulin (INS) beta-cell line. This dual approach served two purposes. Firstly, it allowed the adaptability of this method to be assessed across diverse cellular environments by comparing primary and conventional cellular models. Secondly, by comparing Mcp with INS beta cells, experiments demonstrated the technique's robustness, given their distinct patterns of sEV secretion.

[0134] The isolation of small extracellular vesicles (sEVs) was successful in both cell types (Figure 3). To assess the purity of the eluted sEV samples, negative stain TEM imaging was conducted (Figure 3A & B). For comparison, INS-sEV samples were isolated from a 50 mL culture medium at -50% confluence of INS cells (Figure 12A) using a well-established ultracentrifugation protocol (Chhoy, P. et al. 2021. STAR Protoc. 2, 100303) and a commonly available polymer precipitation-based commercial kit (Total exosome isolation kit, Thermo). This method showcased a substantial advantage in terms of sample purity (Figure 12B-D). The polymer precipitation method exhibited notable contamination, possibly due to traces of the PEG-based polymer matrix (Figure 9B), necessitating an additional size exclusion chromatography (SEC) step (Figure 9B). However, the precipitation method yielded a 9-fold increase in particle concentration between 60-140 nm. It's important to note that polymer precipitation-based isolation strategies are not membrane-specific, resulting in the co-isolation of non-vesicular entities, NVEP including larger lipoproteins, which contribute to the overall particle count (Kashkanova, A. D. et al. 2023. J. Extracell, vesicles 12, 12348; Van Deun, J. et al. 2020. Adv. Biosyst. 4, 1900310). When compared to ultracentrifugation, the samples not only exhibited higher purity but also demonstrated a 4-fold enrichment (Figure 12E). These findings underscore the versatility and effectiveness of this nanoprobe-enhanced protocol. To assess cellular uptake and functionality, the isolated INS-sEV samples were tagged with DSPE- PEG 5K-FITC as an EV tracker and subsequently exposed them to INS cells. Thew EVs displayed clear evidence of cellular internalization for both commercial kits and this proposed method (Figure 13).

[0135] The size of vesicles was then analyzed using dynamic light scattering (DLS) and nanoparticle tracking analysis (NTA) to profile sEVs derived from both cell lines (Figure 3C & D). The consistent vesicle size across replicates highlighted the reproducibility of this method. Importantly, more than 90% of these vesicles fell within the 50-200 nm range, meeting the criteria for sEVs. Starting with a 50 ml volume, approximately ~107particles per milliliter was obtained for both samples, with a predominant diameter around 100 nm. It is noteworthy that the yield from Mcp-sEVs was three times less than that from INS-sEVs, possibly due to phagocytic activity inherent to macrophages, potentially leading to EV reuptake explaining the difference in yield. Further analysis was conducted using exosomal markers CD9 and CD81, with dSTORM single molecule localization microscopy (SMLM) (Figure 3E & F). The LN's dual purpose, aiding in fixation during SMLM sample preparation, allowed biotinylated sEVs to adhere directly to avi din-functionalized slides for imaging analysis (Figure 10A).

[0136] This analysis confirmed the trend of a reduced exosomal count in Mcp-sEV samples, aligning with the previously observed differences in yield between the two cell types. To assess vesicle size, these experiments leveraged the capability of this method to identify localization lengths specifically for CD9 and CD81 carrying exosomes (Figure 3G-I). Intriguingly, these exosomal localization lengths exhibited a smaller peak diameter compared to the DLS and NTA results. However, it's noteworthy that recent studies have highlighted that DLS and NTA techniques can sometimes provide slightly larger diameter values and exhibit broader size distributions (Kashkanova, A. D. et al. 2022. Nat. Methods 19, 586-593). Additionally, it's important to clarify that the DLS and NTA results encompass the entire sEV population within the sample, not just exosomes. Nonetheless, despite this distinction, the shape of the distribution patterns in the localization length data closely resembled those observed in DLS and NTA measurements. This alignment further strengthens the robustness of these findings.

[0137] Using this approach, the size distribution contribution of single positive (CD9+CD81‘ and CD9'CD81+) and double-positive (CD9+CD81+) exosomes within the exosome cohort were assessed. Detailed size distribution analysis revealed that smaller vesicles were predominantly CD81+in macrophage samples (Figure 14A), a distinction not evident in INS- sEVs (Figure 14B). Furthermore, INS-sEVs were additionally labeled using the membrane- associated dye MEMGlow560™, revealing a distinct sEV subpopulation within the expected size range that lacked both tetraspanins (Figure 1 IB). This observation suggests potential limitations in solely relying on affinity-based isolation strategies, even though other exosome tetraspanin markers like CD63 were not tested. Interesting, this double-negative (CD9 CD81 ) sEVs exhibited a peak >150 nm, beyond the size cutoff for exosomes size range (Dixson, A. C. et al. 2023. | Nature Reviews Molecular Cell Biology. Nat. Rev. Mol. Cell Biol. 24(7):454-476). Circularity studies conducted on the isolated population showed a mean circularity > 0.7 for all isolated sEVs, indicating a retention of morphology (Saftics, A. et al. 2023. J. Extracell, vesicles 12, 12346) (Figure 14C & D). As a validation of this isolation technique's ability to capture exosomes within the lower size range, negative staining with anti-CD81 4 nm colloidal Au-Tag was conducted. This analysis revealed a representative exosome measuring approximately 30 nm (Figure 14E), showcasing the capability to isolate smaller exosomes while ensuring their separation from lipoproteins and other non-vesicular extracellular particles more abundant in that size range.

[0138] Size-selective Vesicle Enrichment.

[0139] To comprehend the rationale behind selectively isolating smaller vesicles, experiments were conducted using these standard liposomes with hydrodynamic diameters of 60 nm and 140 nm. As part of this investigation, analysis began with constructing a standard curve representing the integration of 5% LN constructs with different PEG lengths (as listed in Table 1). Intriguingly, despite the increment in PEG length, results showed that the hydrodynamic diameter exhibited a saturation effect. Specifically, for the smaller 60 nm liposomes, the diameter increased by approximately +10 nm, while the larger 140 nm liposomes exhibited an increase of roughly + 20 nm (Figure 4A). To further corroborate these findings and confirm the cleavage of the PEG-DB tether, experiments utilized the photosensitive DSPE-NIR-PEG O.1K- DB probe. Following light cleavage, results demonstrated that this technique effectively removed the PEG components from the liposomes, thereby restoring them to their standard sizes.

[0140] Next, experimented were used to evaluate the release efficiency of these liposomes across a gradient of D-biotin concentrations, spanning from 0.1 mM to 100 mM (Figure 4A). As expected, a concentration-dependent trend was observed. At lower D-biotin concentrations, fewer vesicles were released, whereas this release efficiency increased progressively with rising D-biotin concentrations. However, the amount of smaller 60 nm liposomes released was larger, and at the highest concentration tested (100 mM), results showed an almost 10% higher release compared to 140 nm liposomes. This was attributed to the inherent differences in the number of LNs anchored to the beads. Smaller vesicles have a significantly lower number of LNs attached to them, a 6-fold fewer, as illustrated in Table 2. This difference in probe density among vesicles of varying sizes plays a pivotal role in determining their release efficiency. able 2: Number of lipid nanoprobes per liposome based on its size.

[0141] To test if varying the concentration could selectively enrich the release liposomes based on sized, small and large liposomes were mixed together. As the concentration increased, results show that the liposomes released had decreasing mean hydrodynamic diameter (Figure 4C). Despite the broader peaks associated with NT A, as reported in previous studies (Kashkanova, A. D. 2022. Nat. Methods 19, 586-593), results consistently showed this trend in both INS-sEV and Mip-sEV samples (Figure 4D), with the mean diameter decreasing as the D- biotin concentration increased (Figure 4E). However, at lower D-biotin concentrations, the amount of sEVs released decreased. For Mcp-sEVs, even though the size at the mode was determined, most sizes were below the instrument’s threshold of 1x107parti cles / mL.

[0142] These findings highlight the potential of gentle D-biotin elution LNs for selective sEV isolation. This novel concentration-dependent vesicle amplification mechanism not only provides a versatile platform for researchers to tailor vesicle size enrichment according to their specific needs but also has broad applications in the liposome domain. In cases where the loss of vesicles during elution could be an issue, increasing the lipid concentration could compensate for this loss. This adaptable approach opens new avenues for researchers to explore and manipulate vesicle populations effectively.

[0143] To illustrate the versatility of this approach, experiments were conducted to differentiate and sort model proteoliposomes resembling synaptic vesicles (SVs) and large dense core vesicles (LDCVs) from a control proteoliposome sample. This was achieved through a sequential elution process, starting with 2.5 mM D-biotin elution to release the smaller vesicles, followed by NIR elution to release the remaining larger vesicles . Remarkably, this was achieved without the need for conventional ultracentrifugation processes.

[0144] This lipid mixing assay was then employed to investigate lipid mixing kinetics as a function of vesicle size (Figure 4B). Smaller liposomes exhibited faster fusion, with the most to least fusogenic proteoliposomes showing a decreasing terminal final NBD fluorescence. This observation aligns with a previous study that sorted liposomes using a DNA- brick assisted ultracentrifugation approach (Yang, Y. et al. 2021. Nat. Chem. 13, 335-342). While this study focused on the minimal fusion machinery (SNAREs) to demonstrate the concept, this platform can, in principle, be adapted to model more complex physiological conditions. For instance, it can be extended to study the interplay between membrane curvature and molecular factors governing vesicular fusion, such as Sytl or Muncl8, underpinning vesicle dynamics in neurons and other cellular processes.

[0145] Isolating Extracellular Vesicle Subpopulations

[0146] As demonstrated before, the distinctive design of this photosensitive LNs was utilized to harness dual elution methods. To further validate this idea on EV samples, neuroblastoma N2A cell line derived EVs was subjected to this protocol. By employing a 0.4pm filter for the cell culture media to increase the EV cutoff to ensure a larger size diversity of EVs. Once the EVs were captured and purified, a D-biotin concentration of 2.5 mM, the system preferentially released sEVs exhibiting a mean hydrodynamic diameter spanning 100-150 nm (Figure 5A). A secondary NIR exposure was conducted to recover the larger EVs retained on the HCABs. As expected, the photocleavage process eluted vesicles of larger sizes, with hydrodynamic diameters between 220-350nm.

[0147] Western blot analyses further supported these findings by confirming the presence of canonical exosomal tetraspanin markers — CD63, CD9, and CD81 — in both eluted samples (Figure 5B). However, a thorough examination of the sRNA cargo revealed notable disparities in their relative abundances (Figure 5C). Specifically, the sEVs obtained via biotin elution showed a twofold increase in mature micro-RNA (miRNA) levels. In contrast, the larger EVs obtained from the NIR elution exhibited a fivefold increase in ribosomal RNA (rRNA). This difference suggests the intriguing possibility of distinct cellular origination and cargo enrichment, depending on the vesicle size. It is well-documented that rRNA is a major constituent of microvesicle, presumably due to their closer association with cellular compartments where ribosomal machinery is abundant (Miranda, K. C. et al. 2010. Kidney Int. 78, 191-199; Cheng, L. & Hill, A. F. 2022. Nat. Rev. Drug Discov. 21, 379-399). Conversely, mature miRNAs, which are frequently associated with post-transcriptional regulation, are a common presence in sEVs, suggesting their involvement in intercellular communication. In the EV samples, two of the most abundant miRNAs were noteworthy (Figure 16A). MiRNA 335 : 5p has been linked to cancer and is known as a cancer-secreted exosomal miRNA (Sun, X. et al. 2021. Mol. Ther. Nucleic Acids 24, 164-174; Chen, D. et al. 2020. Front. Oncol. 10). Meanwhile, miRNA 16:5p has shown promise as a potential diagnostic marker for Alzheimer's disease (Liu, S. et al. 2022. Pharmacother. 148, 112681).

[0148] Moving beyond tetraspanins, these assays also delved into the synaptic protein Sytl, which has recently been proposed as an Alzheimer's Disease (AD) biomarker. The literature presents a dichotomy — some studies postulate elevated Sytl levels (Ohrfelt, A. et al. 2016. Alzheimers Res. Ther. 8, 41), especially in cerebrospinal fluid samples, while others argue the opposite (Jia, L. et al. 2021. Alzheimers Dement. 17, 49-60; Goetzl, E. J. et al. 2016. FASEB J. 30, 4141-4148). A recent research study has successfully isolated exosomes from CSF, suggesting a consortium of synaptic proteins, including SYT1, as potential biomarkers (Jia, L. et al. 2021. Alzheimers Dement. 17, 49-60). However, many studies label their vesicles as "exosomes" without comprehensively analyzing the canonical tetraspanin markers, potentially misleading the term. To illustrate this point, a study associated with an Alzheimer's Disease (AD) model induced by the neurotoxicity of streptozotocin on N2A cells was utilized (Chen, L. et al. 2020. J. Cell. Mol. Med. 24, 10458-10467; Park, J. et al. 2020 Front. Cell. Neurosci. 14; Kadhim, H. J. et al. 2022. Sci. Rep. 12, 21878; Biswas, J. et al. 2016. Mol. Neurobiol. 53, 2794- 2806). Drawing inspiration from this AD model, N2A cells were exposed to 100 pM STZ for a duration of 48 hours. Subsequently, sEVs and larger EVs were isolated. Imaging revealed the widespread presence of tetraspanins, CD9 and CD63, in conjunction with Sytl (Figure 5D). Remarkably, although the Western blot indicated the presence of all proteins, minimal colocalization of 1-2% delineated two distinct EV pools: one abundant with SYT1 and the other predominantly marked by the exosome-associated canonical markers, CD9 and CD63 (Figure 5E). After STZ administration, there was a noticeable increase in SYT1 -associated EVs. These SYTl-rich EVs appeared to lack the exosomal markers under investigation, raising questions about the broad usage of the term 'exosome.'

[0149] Furthermore, integrated intensity evaluations revealed an SYT1 overexpression in both EV categories (Figure 5F). This was not the case for EVs containing CD9 and CD63 (Figure 16B). Notably, sEVs exhibited a median increase that was twice as pronounced, corresponding to a 4-fold increase in the fluorescence of the Alexa Fluor® 647- conjugated secondary antibody. It's important to note that even with monoclonal SYT1 primary antibodies, multiple secondary antibodies could bind to each primary SYT1 antibody. This still results in at least a 4-fold increase in SYT1 representation within sEVs compared to their larger counterparts. These findings emphasize the urgent need to separate sEVs based on size, as it shed slight on subtle yet significant differences in cargo and surface proteins, which might play a crucial role in disease diagnostics.

[0150] Interestingly, in a direct comparison, experiments were performed with this photosensitive lipid nanoprobe (PLN) method to isolate exosomes from AD plasma samples and benchmarked against the Invitrogen plasma exosome isolation kit. This approach not only maintained high yield but also achieved superior purity with no contamination of molecular debris (Figure 37). This high purity level is critical as it reduces the risks of false positives and false negatives, thus enhancing the accuracy of biomarker-based diagnostics.

[0151] As the field of vesicle research increasingly emphasizes size-based classifications, this photosensitive LN-based isolation method emerges as a precise, scalable, and potentially universally applicable approach for isolating and enriching EVs and synthetic vesicles. Its implications extend across biological research, diagnostics, and therapeutics. In the realm of EV research, where conventional sEV isolation methods often rely on ultracentrifugation, this innovative 3-step technique showcases superior effectiveness. This study has verified the purity of this EV isolation following the guidelines set forth by the international society of extracellular vesicles (ISEV) (Thery, C. et al. 2018. J. Extracell, vesicles 7, 1535750). This method not only outperforms traditional approaches but also maintains the integrity of the vesicles while offering adaptability for various applications. This method's capacity to isolate sEVs from diverse biological sources, including primary macrophage, N2A, and insulin beta-cell lines, underscores its broad applicability. Furthermore, this research demonstrates its potential for precise size-based vesicle categorization. By distinguishing between small and large EVs derived from the N2A neuroblastoma cell line, this study provides unique RNA profiles and protein markers, emphasizing the importance of specific classifications. This innovative technique, with potential enhancements in the future, holds promising implications for both basic research and commercial applications, including drug-delivery systems and disease-specific EV isolation.

[0152] In conclusion, this pioneering work showcases transformative insights, indicating a possible paradigm shift in vesicle isolation and size separation (Shah, S. et al. 2020. Adv. Drug Deliv. Rev. 154-155, 102-122; Amati, A. M. et al. 2020. Biochem. Soc. Trans. 48, 1473-1492). As these results delve further into this domain, gaining a profound understanding of vesicle subsets becomes crucial, particularly in the diagnostics and therapeutic sectors. The future of this field calls for a heightened and discerning approach, emphasizing the importance of sophisticated yet straightforward isolation strategies. This research serves as a foundational step not only in the realm of EV studies using such probes but also as an initial phase in developing user-friendly kits for vesicle isolation.

[0153] The materials and methods employed in these experiments are now described. Materials and Methods

[0154] Non-photosensitive chemistry probes were obtained from NANOCs.

[0155] Photosensitive LNs were synthesized purified in-house, see Supplementary Note 2 for synthesis and characterization details. All lipids were purchased from Avanti Polar Lipids (Table 3). All LNs, lipids and buffer compositions are listed in Table 1, 3, and 4. Table 3: Synthetic Lipids used in Liposomes and Proteoliposomes.

[0156] Table 4: Lipid Compositions (Top) and Buffer Ingredients (Bottom). Numeric values refer to molar percentages and ratios. aDLS characterization of liposomes and proteoliposomes can be found in Figure 15.

[0157] Liposome and proteoliposome preparation.

[0158] Solvent evaporation and lipid rehydration.

[0159] To prepare liposomes containing 3 mM of total lipids of specific composition (Table 4), appropriate volumes of lipid stocks (dissolved in chloroform) were mixed in a borosilicate glass disposable culture tube. The mixture was blown dry under N2 gas for at least 30 min. The resulting lipid film at the bottom of the tube was further dried for Ih in a desiccator under vacuum. Unless noted otherwise, 1 mL of Buffer A (Table 4) was added to the tube and agitated for 30 min. To test for the leakage, 10 mM sulforhodamine B (SRB) (dissolved in Buffer A) was used instead for rehydration. The glass tubes were always wrapped with aluminum foil to reduce photobleaching of the fluorescent labels.

[0160] Proteoliposome Reconstitution. v-SNARE (VAMP2) and t-SNARE used in this study were purified as described in the Supporting information section 1 .4. They were reconstituted into liposomes using rapid detergent (1% w / v Octyl-beta-Glucoside) dilution and dialysis method as described previously (Ramakrishnan, S. et al. 2020. eLife 9, e54506; Ramakrishnan, S. et al. 2019. FEBS Lett. 593, 154-162; Bera, M. et al. 2023. Proc. Natl. Acad. Sci. 120, e2311484120). In short, lipids (lipid composition in Table 4) were dried down to a film in a glass test tube using nitrogen gas and then stored under a vacuum for Ih to remove any residual chloroform. Lipids were resuspended in protein (final protein / lipid ratio for proteoliposomes are shown Table 4.) and buffer A up to a final volume of 100 pL and then passed through a detergent removal spin column (Pierce) following the manufacturer’s protocol. Volume was increased to 150 pL and dialyzed against 25 mM HEPES, 120 mM KC1, and 1 mM DTT overnight to remove any residual detergent. The next day, 150 pL of 80% OptiPrep™ (lodixanol) (in the same buffer) was mixed with the sample and then loaded into a 0.8-mL tube (Beckman Coulter), overlaid with 250 pL of 30% OptiPrep™, followed by 50 pL of HEPES buffer. Tubes were centrifuged in an SW55 rotor at 48,000 rpm for 4 h. Proteoliposomes were collected from the gradient near the buffer / 30% OptiPrep™ interface. The reconstitution efficiency for both SNAREs have been previously validated to be nearly identical (50%-60%).

[0161] Lipid mixing assay.

[0162] FRET-based lipid mixing experiments were conducted on a plate reader (Molecular Devices). 5 pL of v-SNARE liposomes labeled with a pair of FRET dyes (donor: NBD-DOPE, acceptor: Rhodamine-DOPE) and 45 pL of unlabeled t-SNARE liposomes (Table 4), with a total lipid concentration of 3 mM were separately heated to 37 °C and transferred to a Pierce™ 96-Well Polystyrene Plates. NBD fluorescence was monitored at emission / excitation of -535 / 460 nm every 1 min for 1 hr and 40 mins. The fluorescence signal was normalized using Norm. NBD fluorescence [% of Imax] = 100%x(It-I min) / (Imax_Imin ), where Imin was the lowest observed NBD fluorescence intensity and Imax was the maximum NBD fluorescence intensity observed after addition of detergent (2.5% [w / v] n-dodecylmaltoside [DM]).

[0163] Protein Production and Purification for Lipid Mixing Assays The v and t-SNAREs were expressed and purified as described previously1 4Briefly, proteins were expressed in E. coli strain Rosetta 2 (DE3) (Novagen, Madison, WI) by inducing bacteria with 0.5 mM IPTG for 4 h at 37 °C. Cells were pelleted and lysed in buffer containing 400 mM KC1, 25 mM HEPES, 4% TritonX-100 (v / v), 10% glycerol (v / v), pH 7.4 supplemented with 0.2 mM Tris (2-carboxyethyl) phosphine hydrochloride (TCEP, ThermoFisher, Waltham, MA), and 1 mM phenylmethylsulfonyl fluoride (PMSF, SigmaAldrich). After running samples through the cell disruptor (Avestin, Ottawa, Canada), the lysate was clarified using a 45 Ti rotor (Beckman Coulter, Atlanta, GA) at 35k rpm for 30 min at 4 °C. The supernatant was then incubated with HisPur NiNTA beads (ThermoFisher, Waltham, MA) with constant agitation overnight at 4 °C. The resin was washed with lysis buffer containing 1% octyl glucoside (Chem-Impex, Wood Dale, IL). Protein was eluted with 350 mM Imidazole and the concentration was determined using a Bradford Assay (BioRAD, Hercules, CA).

[0164] Tissue culture.

[0165] Bone marrow -derived macrophage extraction & primary culture.

[0166] Primary bone marrow-derived macrophage (BMDM) cells were obtained through the bone marrow of the hind limbs in C57BL / 6J mouse (JAX™) wild type (WT) mice. WT mice were first euthanized through CO2 overdose for 15 min; mice are sequentially euthanized with bone extraction immediately following the euthanasia procedure of one mouse before the euthanasia of the next. Incisions near the groin and ankle area were made to remove the skin and tissue surrounding the thighs and the hind limbs were dislocated from the ball-and- socket hip joint. Muscle was removed carefully upon the separation of the femur, tibia and fibula, and hind foot. The femur and tibia and fibula bones were immediately placed in ice cold Gibco Iscove’s Modified Dulbecco’s Medium (IMDM) (ThermoFisher) media to keep moist. Under the tissue culture hood, bones were dipped in 70% EtOH 5X then H2O 5X to sterilize and wash off excess non-bone contaminants. In a petri dish, small segments of the two ends of the femur and tibia and fibula were sectioned off with a razor blade to reveal bone marrow that was flushed out with cold IMDM through a needle and syringe. Aggregated bone marrow was separated and further passed through 40pm mesh syringe filters into a 15 mL tube. Upon obtaining a pellet through one centrifuge spin down (300 * g for 5 min), 1 mL of Ack Lysing buffer (Lonza) was added for 30 seconds to lyse red blood cells. lOmL IMDM was used to neutralize the buffer and the sample was centrifuged to obtain the final pellet containing bone marrow-derived cells. 5 million cells were plated in one 10 cm non-TC treated dish with lOmL IMDM, 10% ultracentrifuged exosome-free FBS, 1% Penicillin-Streptomycin, and 20 ng / mL recombinant macrophage colony-stimulating factor (M-CSF). Media with 10 ng / mL M-CSF was changed every 2-3 days with the macrophages fully differentiated and ready for experimentation on day 7. Culture media was extracted at -50% confluence.

[0167] INS-1 rat insulinoma culture.

[0168] The clonal beta-cell line 832 / 3, derived from the parental INS-1 rat insulinoma cells, was purchased from Millipore Sigma and cultured in complete Roswell Park Memorial Institute (RPMI) medium (Sigma Aldrich) supplemented with 10% ultracentrifuged exosome-free FBS (Sigma Aldrich), 10 mM HEPES (Sigma Aldrich), 2 mM L-Glutamine (Sigma Aldrich), 1 mM sodium pyruvate (Sigma Aldrich) and 0.05 mM beta-mercaptoethanol (Sigma Aldrich) in 5% CO2 at 37 °C. Media was changed every 2-3 days with the macrophages fully differentiated and ready for experimentation on the day 7. Culture media was extracted at -50% confluence (Figure 12A).

[0169] Neuroblastoma N2A culture.

[0170] Neuro-2a (N2a) CCL-131™ cell line was purchased from ATCC and cultured with Dulbecco's Modified Eagle's Medium (DMEM) (Sigma Aldrich) supplemented with 10% exosome-free FBS and 100 pg mL1penicillin / streptomycin (Gibco) in 5% CO2 at 37 °C. Media was changed every 2-3 days with the cells fully differentiated and ready for experimentation on day 7. For the AD model the cells were treated with varied concentrations of 100 pm STZ for 48 h. Culture media was extracted at -50% confluence.

[0171] Liposomes, proteoliposome and EV isolation.

[0172] If not mentioned otherwise, in most experiments 200 pL of avidin (monomeric / neutravidin) - HCAB slurry (Thermo Scientific) was pipetted into 0.8 mL centrifuge columns with 30 pm polyethylene filters (Thermo Scientific) (Figure 10D). For larger volumes 5 mL and 10 mL centrifuge columns were used (Thermo Scientific). Given that 4% beaded agarose is provided in a 50% slurry, 100 pL of settled resin equates to 200 pL of slurry. Buffer exchange was performed on the HCAB with Buffer A for liposome and proteoliposome isolation or PBS for EV isolation, through a series of 5 - 200 pL washes based on slurry volume. In parallel, 5% of LN (150 pM for 3 mM lipid concentration) was incubated for Ih with liposomes and proteoliposomes if not mentioned otherwise. In some cases, 1 and 3 mol / mol% were tested for the micelle characterization study. For EV isolation, 100 pg of LN was added into 50 mL of culture media and incubated for Ih. The buffer exchanged HCAB were then incubated with LN incorporated vesicle sample and shaken for 30 minutes. In cases of EV isolation with large 50 mL sample volumes the HCABs were left overnight to maximize vesicle capture and enrichment. Following incubation, the sample was centrifuged (150 x g for 20 sec) to collect the post-capture sample which would be used to determine the liposome trapping efficiency of the HCABs. The HCAB were prevented from drying and washed 5-10 times using the 200 pL (based on slurry volume) of relevant buffer during the purification stage based on the sample. EV samples were washed with up to 2 mL of PBS. For the elution step, to concentrate the sample, 100 pL (1 / 2 slurry volume) of 2.5 mM D-biotin in the buffer was pipetted into HCAB and agitated for a 2- minute incubation period. For sequential D-biotin elution the process was repeated. For the D- biotin concentration experiments, elution buffer - buffer A or PBS with desired D-biotin concentration was used. For light-based elution, the purified HCAB was incubated with 100 pL (1 / 2 slurry volume) of buffer A or PBS without D-biotin and the column was exposed to either 840 or 365 nm lamp at 5mW / cm2for NIR and UV, respectively. The sample was exposed to light for 10 min, if not mentioned otherwise. The solution was then centrifuged (200 x g for 20 sec) to elute isolated vesicles . For reuse, the HCAB were washed with regeneration buffer containing 0.1M glycine at pH 2.8 (Thermo Scientific) and centrifuged (200 * g for 15 sec) a total of 5 times, then exchanged with buffer 5 times. Note that all centrifugation steps were performed using a benchtop unit to speed up the process, however, with larger sample volumes centrifuges can be completely avoided if needed.

[0173] Liposome trapping and release quantification assays.

[0174] The trapping and release were evaluated using bulk fluorescence measurements obtained using a plate reader (Molecular Devices). The liposome solution was diluted from 3mM to 50 pM (60X) to allow for clearer downstream fluorescence analysis. To retain quantitative relevance all flow throughs including the intermediate washes and eluted samples were diluted 60X. The trapping efficiency [% of Max] = 100% x((linitial-lpost.capture- Ibuffer) / (Iinitiai-Ibuffer)), where the linitiai is the initial fluorescence intensity of the liposomes. Buffer is the intensity of the buffer and Ipost-capture is the intensity of the sample collected post-capture. For the release efficiency [% of Trapped] = 100%x((Ieiute-Ibuffer) / (Iinitiai-Ipost-caPture-Ibuffer)) and release efficiency [% of Max] = 1 00%x((Ieiute-IbufFer) / (Iinitiai-IbufFer)), where Liute is the intensity of the eluted sample. The fluorescence intensity readings were taken over 5 minutes at 20 sec intervals and averaged to account for photobleaching effects.

[0175] Liposomes, proteoliposome and EV size characterization.

[0176] Dynamic light scattering (DLS).

[0177] DLS was performed using Wyatt DynaPro Nanostar model WDPN-06. 25 pL of each elution was pipetted into a disposable cuvette. Peak radius cutoffs were at 0.5 nm and 10,000 nm. 10 acquisitions were collected at 25°C for each sample with one acquisition captured every 10 seconds. Batch DLS with regularization processing was utilized to average repeat experiments to calculate average hydrodynamic diameter within each sample. This allowed for the visualization of the sizes of single vesicles and the ability to determine the presence of any aggregated vesicles as outliers. Samples were not diluted in this study. The raw data was obtained and fitted with GraphPad Prism.

[0178] Nanoparticle Tracking Analysis (NTA).

[0179] NTA was performed on a ZetaView® x30, Next Generation Nanoparticle Tracking Analyzer (Particle Metrix GmbH), equipped with a 405 nm laser and a CMOS camera. Data analysis was performed on the ZetaView® software (version 8.04.04), applying a bin class width of 5 nm, a minimum brightness of 25, a minimum area of 5, a maximum area of 1,000, and a trace length of 15. First, the device was first calibrated with polystyrene beads (Sigma- Aldrich) diluted by 1 :500,000 (v / v) at a concentration of -100 particles per frame. For the measurements, the shutter was set to 150, the sensitivity to 85, and the frame rate to 30. EV samples were diluted 1 :10 (v / v) in PBS for the measurements. The raw data was plotted and fitted with GraphPad Prism.

[0180] DLS and NTA size data analysis.

[0181] For both DLS and NTA the experimental replicates (3-5) were first averaged and fitted using either a Gaussian function G(x) or a Lorentzian L(x) based on the presence or absence of trailing end of the distribution and were verified with R-squared value. where ‘a’ is the amplitude or peak intensity of the curve, ‘b’ is the mean diameter where the peak occurs, and ‘c’ is the standard deviation which controls the width of the bell curve. where a is the amplitude or peak intensity of the curve, ‘b’ is the center of the, and c is the standard deviation which controls the width of the bell curve. From this equation, the poly dispersity index, a measure of size heterogeneity in the sample, was also calculated by squaring the ratio of standard deviation to mean size (-) .

[0182] Single-molecule localization microscopy (SMLM).

[0183] SMLM of EVs were obtained using a temperature-controlled Nanoimager

[0184] S Mark II microscope from ONI (Oxford Nanoimaging, Oxford, UK) equipped with a lOOx, 1.4NA oil immersion objective, an XYZ closed-loop piezo 736 stage, and 405 nm / 150 mW, 473 nm / 1 W, 560 nm / 1 W, 640 nm / 1 W lasers, as well as dual / triple emission channels split at 640 / and 555 nm. For sample preparation a manufacturers microfluidic slide coated with avidin was used. The sEVs were prelabelled with antibody for CD9 and CD81 as mentioned in supporting information section 1.2. The samples were incubated in the channels for 20 mins a washed with PBS before imaging. Two-channel (647 and 488 nm) dSTORM data (5,000 frames per channel) or three channels (2,000 frames per channel) (647, 555 and 488 nm) were acquired sequentially at 30 Hertz in total internal reflection fluorescence (TIRF) mode. Two color was used for Mq>- sEVs and three color was used for INS-sEVs with an additional label using the membrane- associated dye MEMGlow560™. Before each imaging session, beads slide calibration was performed to align fluorescent channels, achieving a channel mapping precision smaller than 12 nm. Single-molecule data was filtered using NimOS (Version 1.18.3, ONI) based on the point spread function shape, photon count and localization precision to minimize background noise and remove low-precision and non-specific co-localization. All pictures were analyzed using algorithms developed by ONI via their CODI website platform. For the SMLM experiments anti- CD9 antibody was conjugated with Alexa Fluor 488 dye and anti-CD81 antibody was conjugated with Alexa Fluor 647 dye (Thermo) using the Apex Antibody Labelling Kit (Invitrogen) according to the manufacturer’s protocol.

[0185] Single vesicle fluorescence microscopy.

[0186] A custom TIRF microscopy, previously described (Bera, M. et al. 2023. Proc. Natl. Acad. Sci. 120, e2311484120; Kalyana Sundaram, R. V. et al. 2022. Small 18, 2205567). Nikon inverted microscope equipped with three laser lines (488, 532, and 633 nm), a Photometries DV2 dual view, and an Andor EMCCD digital camera was used to was used for single vesicle fluorescence imaging. Two fluorophore channels were used based on the experiment; FITC and MEMGlow647™ were performed in the 488 and 647 channel, SRB and Atto-647N-PE were performed in the 555 and 647 channel; CD9 / CD63 labelled with Alexa Fluor 488 and Sytl labeled with Alexa Fluor 647 were performed in 488 and 647 channel. For all measurements, images were obtained at 512 * 512 pixels with a dynamic range of 16-bit grayscale. The exposure time was set to 150 ms in both channels. Images were converted to TIFF using FIJI (Schindelin, J. et al. 2012. Nat. Methods 9, 676-682) and analyzed with custom Wolfram Mathematica algorithm. The flowchart in figure 16C describes the data analysis pipeline. All colocalization efficiencies were computed using the identified vesicles in each channel. For the TIRF single vesicle immunofluorescence experiments anti-CD9 and anti-CD63 antibodies were conjugated with Alexa Fluor 488 dye, while the anti-Sytl antibody was conjugates with Alexa Fluor 647 dye.

[0187] Scanning-electron microscopy (SEM).

[0188] Samples were fixed in a 2% EMS-quality paraformaldehyde aqueous solution and washed twice with lx PBS or Buffer A, based on the EV or liposome sample, respectively. The agarose beads with vesicles were used within 7 days of cell culture media extraction. All samples were dehydrated in a series of increasing ethanol concentrations (10, 20, 30, 50,70, 90,100%), following which, they were transferred onto conductive double sided carbon tape fixed on a SEM stage and dried. Prior to imaging the samples were thoroughly blow dried under a fume hood. A Hitachi SU8230 SEM was used for all the imaging under low beam energies (1.0-3.0kV). aIncorporate 1% w / v Octyl-beta-Glucoside (OG) detergent in Buffer A for membrane protein solubilization during v-SNARE and t-SNARE reconstitution for proteoliposomes.

[0189] Negative stain electron microscopy.

[0190] For preparing the negative stain EM grids of sEVs and Au-labelled sEVs (CD81 was labelled - see Supporting information), 5 pL of the isolated samples (in PBS) were directly applied to a glow discharged Carbon Type-B, 400 mesh, Copper grid (Ted Pella). After 1 -minute incubation on the grid, the samples were quickly washed with 5 pL of 2% (w / v) uranyl formate solution, followed by staining with another 5 pL of 2% uranyl formate for 1 minute. The grids were imaged using an 80 kV JEOL JEM-1400Plus microscope equipped with a bottommount 4 k x 3 k charge-coupled device camera (Advanced Microscopy Technologies).

[0191] RNA isolation and sequencing.

[0192] RNA isolation from EVs.

[0193] The Total exosome RNA and protein isolation kit (Invitrogen) was utilized for recovery of RNA from both NIR and biotin eluted N2A EV. Two hundred microliter of each sample (brought up to volume with PBS if necessary) was combined with 205 pL of 2* denaturing solution, vortexed to lyse, and then incubated on ice for 5 min. After incubation, 410 pL of Acid-Phenol: Chloroform was added to the mixture and vortexed for 30-60 s to mix. Samples were then centrifuged for 5 min at 10,000 x g at room temperature to separate the mixture into aqueous and organic phases. Once centrifugation was complete, the aqueous (upper) phase was carefully removed without disturbing the lower phase or the interphase and transferred to a fresh tube. One point twenty-five volumes of 100% EtOH was added to the aqueous phase for each sample then vortexed to mix. About 700 pL of volume was placed onto spin column in a collection tube then spun at 10,000 * g for 15 s to move the sample through the filter cartridge. Samples were then washed once with 700 pL Wash Solution 1 x and 2 x with 500 pL wash solution 2 / 3 (centrifuged at 10,000 x g for 15 s for each wash). After washing, filter was dried by spinning for an additional 1 min at 10,000 x g. The filter cartridge was transferred into a fresh collection tube and 50 pL of preheated (95 °C) nuclease-free water was applied to the center of the filter. Samples were centrifuged for 30 s at 10,000xg to recover the RNA, then a second 50 pL volume of preheated (95 °C) nuclease-free water was applied to the center of the filter and centrifuged for 30 s at 10,000xg. After the second spin, the eluate containing the RNA was collected and stored at -20 °C. miRNA Seq quality control.

[0194] RNA quality was determined by estimating the A260 / A280 and A260 / A230 ratios by nanodrop. RNA was run on the Agilent Bioanalyzer gel or Agilent Fragment Analyzer. miRNA Seq library prep.

[0195] Library preparation was performed using the Qiagen QIAseq miRNA

[0196] Library Kit. 5uL of RNA eluate was taken for each sample. Adapters were ligated sequentially to the 3’ and 5’ ends of miRNAs prior to cDNA synthesis with UMI assignment, cDNA cleanup, amplification, and final library cleanup. The library was run on the Agilent Tapestation and size selection was performed if there was excess adapter dimer. The library was then quantified by qRT-PCR using a commercially available kit (KAPA Biosystems).

[0197] Flow cell preparation and sequencing.

[0198] Sample concentrations were normalized to 1.2 nM and loaded onto an

[0199] Illumina NovaSeq flow cell at a concentration that yields 5-10 million passing filter clusters per sample. Samples were sequenced using lOObp paired-end sequencing on an Illumina NovaSeq according to Illumina protocols. The lObp unique dual index was read during additional sequencing reads that automatically follow the completion of read 1. Data generated during sequencing runs were simultaneously transferred to the YCGA high-performance computing cluster. A positive control (prepared bacteriophage Phi X library) provided by Illumina was spiked into every lane at a concentration of 0.3% to monitor sequencing quality in real time.

[0200] Data analysis and storage.

[0201] Signal intensities were converted to individual base calls during a run using the system's Real Time Analysis (RTA) software. Base calls were transferred from the machine's dedicated personal computer to the Yale High Performance Computing cluster via a 1 Gigabit network mount for downstream analysis. Primary analysis - sample de-multiplexing and alignment to the human genome - was performed using Illumina's CASAVA 1.8.2 software suite. The data was returned if the sample error rate was less than 2% and the distribution of reads per sample in a lane was within reasonable tolerance.

[0202] Western Blot.

[0203] EV samples were lyzed with electrophoresis (Laemmli) sample buffer and boiled the mixture at 95-100 °C for 5 min. Proteins were separated by SDS / PAGE method, and performed western blots based on the Biorad western blot protocol. The blots were stained with antibodies depending on the protein analyzed and incubated overnight. The antibody used and the dilution can be found in Table 5. The blot was developed with ECL western blot substrate. Table 5: Primary Antibodies used to Characterize Isolated sEV.

[0204] Statistics and reproducibility.

[0205] Data were represented as the means ± SEM or mean ± SD, mentioned in all appropriate cases. Unpaired Student’s t test was used to analyze data with only two sets. Two- way analysis of variance (ANOVA) was performed to determine whether there was a significant difference between more than two datasets, following Tukey’s method, using GraphPad Prism 6.0. Group differences at the level ofP < 0.05 were considered statistically significant. Asterisk (*) represented p < 0.05; double asterisk (**) represented p < 0.01; triple asterisk (***) represented p < 0.001; quadruple asterisk (****) represented p < 0.0001.

[0206] Attenuated Total Reflectance - Fourier Transform Infrared (ATR- FTIR) Spectroscopy

[0207] The ATR-FTIR spectra were acquired using a Cary 600 series FTIR (Agilent Technologies, Inc., USA) spectrometer using a macro-ATR accessory with a Zn / Se crystal. The spectra were measured in a range from 4,000 to 500 cm164 interferograms were collected at a nominal resolution of 2 cm1and averaged for both the background and samples. The beads samples were dried, and the measurements were done at room temperature, immediately after mounting the sample.

[0208] Thin Layer Chromatography (TLC) and Column Chromatography

[0209] Analytical thin-layer chromatography (TLC) was performed using SILICYCLE® Inc. glass-backed silica gel hard layer with 20 x 20 cm size (F254, 250 pm thickness) and developed plates were visualized using a UV lamp and / or stained with Iodine (I2), potassium permanganate (KMnCh), p-anisaldehyde (CH3OC6H4CHO), phosphomolybdic acid (i2MoO3.H3PO4.xH2O) or ninhydrin. Normal phase flash column chromatography was conducted using either silica gel 60 A (32-63 microns) or an automated Biotage® Isol era™ One flash purification system equipped with a 10 g SNAP Ultra (HP Sphere, 25 pm silica) cartridge. Whichever column chromatography was performed, the desired fractions (confirmed by TLC or UV) were collected and concentrated under reduced pressure to obtain the product.

[0210] Nuclear Magnetic Resonance (NMR)

[0211] All NMR data were acquired at ambient temperature, unless otherwise indicated. NMR solvents, chloroform-D (CDCh) and methanol-D4 (CD3OD) were purchased from Cambridge Isotopes Laboratories, Inc. and used as received. CD3OD ampules were used immediately upon opening. NMR spectra were processed with MestReNova software (v. 10.0.2) using the baseline and phasing correction features. Multiplicities and coupling constants were calculated using the multiplet analysis feature with automated and / or manual intervention, as necessary. ’H NMR spectra were obtained on Agilent 400 MHz, 500 MHz, or 600 MHz spectrometers. Proton chemical shifts (8) are reported in ppm and referenced to residual solvent peaks for CDCh (8 7.26 ppm) and CD3OD (84.87 ppm). Proton data are reported as chemical shift, multiplicity (noted as singlet (s), doublet (d), triplet (t), quartet (q), pentet (p), heptet (hept), multiplet (m), broad singlet (bs), doublet of doublets (dd), doublet of doublet of doublets (ddd), doublet of doublet of triplets (ddt), doublet of triplets (dt), doublet of triplet of triplets (dtt), etc.) coupling constants [Hz], and integration.13C NMR spectra were obtained on Agilent 400(101) MHz, 500 (126) MHz, or 600 (150) MHz spectrometers with full proton decoupling. Carbon chemical shifts (8) are reported in ppm and referenced to residual solvent peaks for CDCh (8 77.16 ppm) and CD3OD (8 49.00 ppm) with multiplicity and coupling constants [Hz] indicated when present.

[0212] Mass Spectrometry

[0213] High-resolution mass spectrometry (HRMS) was conducted by the Chemical and Biophysical Instrumentation Center in the chemistry department at Yale University, on a Waters Xevo Q-TOF high-resolution Mass Spectrometry using ESI.

[0214] Synthesis Chemicals for lipid Nanoprobe Synthesis

[0215] Room temperature is considered as 20-25 °C. All chemical reactions were carried out under normal conditions without exclusion of air or moisture, unless otherwise stated. All commercially available reagents and solvents were obtained from common suppliers [Ambeed, TCI Chemicals, Thermo Scientific Chemicals, Acros organics, MilliporeSigma, Nanocs, BroadPharm] and used without further purification unless otherwise reported. Dichloromethane (CH2CI2), chloroform (CHCI3), A, A -di methyl form am ide (DMF), tetrahydrofuran (THF), acetonitrile (MeCN), and ethanol (CH3CH2OH) were dried over alumina and stored in molecular sieves. Triethylamine (Et3N) and A / A-Diisopropylethylamine (z-PnNEt) were distilled over calcium hydride (CaPE) under a nitrogen atmosphere prior to use. Deionized water was used for reactions and extraction mediums. HPLC grade solvents were used for all other chromatography.

[0216] Synthetic Scheme for DSPE-NIR-PEG 0.1 K-DB (Figure 17)

[0217] Compound 2. cis-l,2-dichloroethylene (Compound 1, 113 pL, 1.5 mmol), 2-(Boc- amino)ethanethiol (558 pL 3.3 mmol), and sodium hydroxide (132 g, 3.3 mmol) were added to ethanol (10 mL) in an oven-dried round bottom flask (100 mL). The mixture was stirred at 55 °C for 24 h, cooled it to room temperature, and the solvent was removed under reduced pressure. The residue was resuspended in chloroform (20 mL) and water (20 mL), extract the aqueous layer with chloroform (3 x 20 mL), washed the combined organic extracts with brine (20 mL), and dried over anhydrous sodium sulfate. The solvent was then removed under reduced pressure and the crude residue was purified by automated normal phase chromatography (Biotage®, SNAP Ultra 10 g; dichloromethane / acetonitrile 90 / 10) to obtain a white powder (493 mg, 87%). 'H NMR (600 MHz, Chloroform-D, 8): 6.12 (s, 2H), 3.34 (d, J= 6.7 Hz, 4H), 2.85 (t, J= 6.6 Hz, 4H), 1.44 (s, 18H).13C NMR (151 MHZ, Chloroform-D, 8): 155.9, 124.4, 79.7, 40.7, 34.5, 28.5. ESLMS ( / z . [M-H]’ calculated for C16H29N2O4S2, 377.5450; found, 377.5449.

[0218] Compound 3. Compound 2 (450 mg, 1.2 mmol) and trifluoroacetic acid (1 mL, 13 mmol) were added to di chloromethane (4 mL) in an oven-dried round bottom flask (25 mL). The mixture was stirred at room temperature for 2 h and the solvent was removed under reduced pressure. The residue was resuspended in dichloromethane (5 mL) and lowered the flask temperature to 0 °C before adding triethylamine (4 mL) to it. The mixture was stirred at room temperature for 0.5 h. The solvent was then removed under reduced pressure and the crude residue was purified by automated normal phase chromatography (Biotage®, SNAP Ultra 10 g; di chloromethane / methanol 20 / 80) to obtain a brown liquid (493 mg, 95%). 'H NMR (600 MHz, Methanol-D4, 8): 6.30 (s, 2H), 3.14 (t, J = 7.1 Hz, 4H), 3.00 (t, J= 7.1 Hz, 4H).13C NMR (151 MHz, Methanol-D4, 8): 125.0, 40.6, 31.7. ESI-MS (m / z) [M-H]' calculated for C6H13N2S2, 177.3020; found, 177.3018.

[0219] Compound 5. DSPE-NHS (100 mg, 0.11 mmol), Compound 3 (20 mg, 0.11 mmol), and triethylamine (139 pL, 1 mmol) were added to dichloromethane (4 m ) in an oven-dried round bottom flask (25 mL). The mixture was stirred at room temperature for 24 h and monitored with TLC. The solvent was removed under reduced pressure and the crude residue was purified by automated normal phase chromatography (Biotage®, SNAP Ultra 10 g; di chloromethane / methanol 20 / 80) to obtain a white powder (92 mg, 83%).1H NMR (600 MHz, Methanol-D4, 8): 7.98 (s, 1H), 6.31 (d, J= 1.4 Hz, 1H), 6.18 (m, 1H), 5.35 (td, J= 17.0, 16.0, 9.0 Hz, 2H), 4.50-1.00 (m, 87H), 0.90 (t, J= 6.7 Hz, 6H).13C NMR (151 MHz, Methanol-D4, 8): 180.1, 163.2, 163.0, 159.9, 130.8, 127.4, 125.0, 123.2, 121.1, 119.2, 118.1, 117.2, 115.3, 59.4,

[0220] 58.3, 54.8, 47.9, 42.0, 40.9, 40.6, 40.5, 37.0, 35.3, 35.2, 34.8, 33.4, 33.1, 31.8, 31.7, 31.2, 30.8,

[0221] 30.8, 30.8, 30.6, 30.5, 30.5, 30.3, 30.2, 29.1, 28.1, 26.7, 26.2, 26.1, 23.7, 23.7, 18.4, 14.7, 14.5,

[0222] 14.4, 14.2, 9.2. ESI-MS ( z): [M-H]’ calculated for C51H97N3O10PS2, 1007.3495; found, 1007.3472.

[0223] Compound 7. Compound 5 (50 mg, 50 pmol), Desthiobiotin-PEG3-NHS ester (Compound 6, 25 mg, 50 pmol), and triethylamine (1 pL) were added to dichloromethane (3 mL) in an oven-dried round bottom flask (25 mL). The mixture was stirred at room temperature for 24 h and monitored with TLC. The solvent was removed under reduced pressure and the crude residue was purified by automated normal phase chromatography (Biotage®, SNAP Ultra 10 g; dichloromethane / methanol 20 / 80) to obtain a white powder (62 mg, 88%). 'l l NMR (600 MHz, Methanol-D4, 8): 5.85 (s, 1H), 5.62 (d, J= 9.5 Hz, 1H), 5.35 (dt, J= 21.8, 4.5 Hz, 2H), 4.70 - 1.02 (m, 121H), 0.89 (t, J= 6.7 Hz, 6H). 'H NMR (151 MHz, Methanol-D4, 8): 180.1, 176.3, 174.9, 173.3, 172.9, 172.7, 166.2, 159.8, 130.9, 130.8, 129.0, 125.9, 124.0, 102.2,

[0224] 99.2, 84.0, 71.6, 71.5, 71.5, 71.4, 71.3, 71.3, 71.2, 71.2, 70.6, 68.5, 68.2, 67.6, 62.1, 59.4, 57.4,

[0225] 54.8, 52.7, 49.8, 49.7, 40.3, 36.9, 35.7, 34.7, 33.1, 30.8, 30.8, 30.8, 30.7, 30.6, 30.5, 30.5, 30.3,

[0226] 30.3, 30.2, 30.2, 30.2, 29.1, 27.1, 26.8, 26.7, 26.5, 26.4, 26.3, 26.1, 26.0, 23.7, 23.7, 15.6, 15.6, 14.5, 14.5, 14.4, 14.4, 13.3. [M-H]’ calculated for C70H130N6O16PS2, 1406.9223; found, 1406.9211.

[0227] Synthetic Scheme for DSPE-N1R-FITC (Figure 18)

[0228] Compound 8. Compound 5 (10 mg, 10 pmol), Fluorescein NHS Ester (5 mg, 10 pmol), and tri ethylamine (10 pL) were added to dimethyl formamide (2 mL) in an oven-dried round bottom flask (25 mL). The mixture was stirred at room temperature for 24 h and monitored with TLC. The solvent was removed under reduced pressure and the crude residue was purified by automated normal phase chromatography (Biotage®, SNAP Ultra 10 g; di chloromethane / methanol 20 / 80) to obtain a white powder (9 mg, 65%).1H NMR (400 MHz, MethanoLDUCDCh-Di, 5): 8.65 (s, 2H), 8.27 (m, 3H), 8.08 (s, 1H), 7.43 (d, J= 8.1 Hz, 2H), 0.99 (s, 6H). [M-H]’ calculated for C72H109N3O16PS2, 1367.7548; found, 1367.6500.

[0229] Synthesis ofCHOL-UV-PEG 0.1K-B (Figure 19)

[0230] Compound 10. Cholesterol (Compound 9, 1.20 g, 3.10 mmol), N,N'- disuccinimidyl carbonate (2.38 g, 9.30 mmol), and triethylamine (10 mL) were added to chloroform (25 mL) in an oven-dried round bottom flask (100 mL). The mixture was refluxed for 24 h, cooled it to room temperature, and the solvent was removed under reduced pressure. The residue was resuspended in chloroform (25 mL) and water (25 mL), extract the aqueous layer with chloroform (3 x 25 mL), washed the combined organic extracts with brine (25 mL), and dried over anhydrous sodium sulfate. The solvent was then removed under reduced pressure and the crude residue was purified by automated normal phase chromatography (Biotage®, SNAP Ultra 10 g; dichloromethane / acetonitrile 90 / 10) to obtain a white powder (1.4, 88%). 'H NMR (600 MHz, Chloroform-D, 8): 5.42 (m, 1H), 4.60 (tt, J= 11.0, 5.9 Hz, 1H), 2.83 (s, 4H), 2.48 (d, J= 7.3 Hz, 2H), 2.01 - 0.82 (m, 38H), 0.68 (s, 3H).13C NMR (151 MHz, Chloroform-D, 8): 168.9, 151.0, 138.7, 123.9, 82.4, 56.8, 56.3, 50.1, 42.5, 39.8, 39.7, 37.7, 36.8, 36.6, 36.3, 35.9, 32.0, 31.9, 29.9, 28.4, 28.2, 27.5, 25.6, 24.8, 24.4, 24.0, 23.0, 22.7, 21.2, 19.4, 18.9, 12.0. ESIMS (m / z) [M-H]’ calculated for C32H48NO5, 526.7460; found, 526.7455.

[0231] Compound 11. N-Boc-ethylenediamine (32 pL, 0.2 mmol), N-Boc- ethylenediamine (106 mg, 0.2 mmol), and triethylamine (139 pL, 1 mmol) were added to dichloromethane (4 mL) in an oven-dried round bottom flask (25 mL). The mixture was stirred at room temperature for 24 h and monitored with TLC. The solvent was removed under reduced pressure and the crude residue was purified by automated normal phase chromatography (Biotage®, SNAP Ultra 10 g; dichloromethane / methanol 20 / 80) to obtain a white powder (103 mg, 90%). 'H NMR (600 MHz, Chloroform-D, 8): 5.33 (d, J= 4.9 Hz, 1H), 5.29 - 4.89 (m, 2H), 4.45 (m, 1H), 3.84 - 0.77 (m, 53H), 0.65 (s, 3H).13C NMR (151 MHz, Chloroform-D, 8): 177.7, 156.5, 139.9, 122.6, 79.5, 74.5, 56.8, 56.2, 50.1, 42.4, 41.3, 40.8, 39.8, 39.6, 38.6, 37.1, 36.6,

[0232] 36.3, 35.9, 32.0, 31.9, 28.5, 28.4, 28.3, 28.3, 28.2, 28.1, 24.4, 23.9, 22.9, 22.6, 21.1, 19.4, 18.8, 11.9. ESI-MS (w / z): [M-H]’ calculated for C35H59N2O4, 571.8649; found, 571.8633.

[0233] Compound 12. Compound 11 (100 mg, 0.17 mmol) and trifluoroacetic acid (1 mL, 13 mmol) were added to di chloromethane (4 mL) in an oven-dried round bottom flask (25 mL). The mixture was stirred at room temperature for 2 h and the solvent was removed under reduced pressure. The residue was resuspended in dichloromethane (5 mL) and lowered the flask temperature to 0 °C before adding triethylamine (4 mL) to it. The mixture was stirred at room temperature for 0.5 h. The solvent was then removed under reduced pressure and the crude residue was purified by automated normal phase chromatography (Biotage®, SNAP Ultra 10 g; di chloromethane / methanol 20 / 80) to obtain a brown liquid (76 mg, 94%).1H NMR (400 MHz, Chloroform-D, 8): 5.46 - 5.39 (m, 1H), 4.81 (dp, J= 16.0, 4.9 Hz, 1H), 3.97 (s, 1H), 3.44 (s, 3H), 3.20 - 3.10 (m, 3H), 2.68 - 0.37 (m, 43H).13C NMR (101 MHz, Chloroform-D, 8): 138.6, 124.0, 78.9, 56.8, 56.3, 50.1, 42.4, 39.8, 39.7, 37.6, 36.9, 36.7, 36.3, 36.0, 32.0, 31.9, 28.4, 28.2,

[0234] 27.4, 24.4, 24.0, 24.0, 23.0, 22.9, 22.9, 22.7, 21.2, 19.4, 18.9, 12.0. ESI-MS (wCz): [M-H]’ calculated for C30H51N2O2, 417.7459; found, 417.7449.

[0235] Compound 14. Compound 12 (11 mg, 24 pmol), PC biotin-PEG3-NHS ester (Compound 13, 20 mg, 24 pmol), and tri ethylamine (1 pL) were added to di chloromethane (5 mL) in an oven-dried round bottom flask (25 mL). The mixture was stirred at room temperature for 24 h and monitored with TLC. The solvent was removed under reduced pressure and the crude residue was purified by automated normal phase chromatography (Biotage®, SNAP Ultra 10 g; dichloromethane / methanol 20 / 80) to obtain a white powder (25 mg, 88%). 'H NMR (600 MHz, Chloroform-D, 8): 7.53 (s, 1H), 7.00 (d, J= 32.2 Hz, 1H), 6.30 (s, 3H), 5.51 - 0.41 (m, 94H).13C NMR (151 MHz, Chloroform-D, 8): 175.2, 164.8, 162.0, 154.9, 154.2, 147.6, 147.2, 141.1, 139.9, 139.7, 134.3, 122.7, 119.5, 117.4, 115.5, 109.1, 108.2, 70.0, 68.7, 56.8, 56.5, 56.3, 55.7, 55.2, 50.3, 45.2, 42.4, 39.6, 39.2, 36.7, 36.3, 35.9, 35.7, 33.4, 32.0, 31.8, 31.6, 29.8,

[0236] 29.5, 28.2, 28.1, 28.1, 25.7, 25.6, 25.0, 24.0, 23.0, 22.9, 22.9, 22.8, 22.7, 22.4, 22.2, 21.2, 20.4, 19.9, 19.4, 18.9, 18.4, 14.3, 13.5, 12.0. ESI-MS [M-H]’ calculated for C62H98N7O14S, 1197.4790; found, 1197.4788

[0237] Example 2: Programmable Light-Driven Platform for Rapid Isolation and Fractionation of Extracellular, Cell Membrane, and Biomimetic Lipid Vesicles for Large Scale Clinical Applications

[0238] The rapid development of lipid-based nanoparticles, including extracellular vesicles (EVs), cell membrane vesicles (CMVs), and biomimetic artificial nanovesicle formulations, has opened transformative avenues in drug delivery, gene therapy, and diagnostics. These nanoparticles exhibit a unique capacity to encapsulate and deliver a wide range of therapeutic payloads, from small molecules to nucleic acids, offering immense potential for treating diseases such as cancer and neurodegenerative disorders. Despite their potential, current isolation and purification methods remain cumbersome, inefficient, and prone to structural and functional degradation of these fragile carriers, significantly hindering their scalability for clinical applications. The present novel material platform overcomes these challenges and provides a high-throughput, programmable, light-driven approach for rapid (< 3 minutes) and scalable isolation of lipid vesicles. This innovative approach introduces size-dependent incorporation of light-sensitive probes into vesicle membranes, enabling the selective release and fractionation of vesicles with unprecedented speed and purity. Additionally, the present example demonstrates the co-existence of right-side-out (RSO) and inside-out (ISO) vesicles within CMVs and, for the first time, the selective isolation of RSOs with desired cargo in a single step. These RSOs, owing to their orientation and integrity, represent superior candidates for drug delivery applications. The ability to rapidly isolate and enrich vesicle subpopulations with distinct biological and therapeutic properties could unlock new potentials in precision medicine and advanced therapeutic development.

[0239] Recent advancements have explored lipid-based probes that integrate into vesicle membranes through hydrophobic interactions, facilitating selective isolation and the removal of non-vesicular contaminants. However, current probe-based methods often rely on affinity tags for capture and elution, introducing additional steps and hindering scalability. The use of photocleavable reactive groups can be used to address these limitations by achieving tag-free vesicle purification in a near-native state, critical for downstream applications like mass spectrometry.

[0240] Building on this foundation, the present example relates to a novel, light-driven platform for the rapid (<3 minutes) isolation and size-dependent fractionation of lipid nanoparticles, including EVs and liposomes. This approach employs specially designed, lightsensitive probes that selectively incorporate into vesicle membranes in a size-dependent manner. Smaller vesicles, with higher curvature, incorporate fewer probes, while larger vesicles incorporate more, enabling precise photolysis-triggered release and fractionation. Notably, this platform enriches RSO vesicles, addressing the critical need for orientation-specific separation.

[0241] This innovative system offers a scalable, efficient, and versatile solution to the challenges of lipid nanoparticle purification, preserving vesicle integrity and functionality. By addressing key challenges of this field, this light-activated approach paves the way for the development of safe, effective, and scalable gene therapies and other advanced therapeutic modalities. The subsequent sections detail the methodology, results, and implications of this transformative technology, underscoring its potential to redefine nanomedicine.

[0242] Development of a programmable light-sensitive system for rapid isolation of near-native vesicles.

[0243] The present PLD platform was developed to overcome the limitations of traditional extracellular vesicle (EV) isolation methods, focusing on rapidity, scalability, and structural preservation. The platform consists of two key components: a photocleavable lipid nanoprobe (PLN) and a positively charged polymer (PLL). The PLN incorporates into EV membranes via its hydrophobic lipid insert and features a UV-sensitive photocleavable linker, a spacer, and an affinity tag, enabling dual isolation mechanisms. The PLL complements this system by forming electrostatic interactions with negatively charged EV membranes, enhancing capture efficiency (Figure 38A). To optimize EV capture and ensure compatibility with downstream UV-mediated release, we tested varying PLL:PLN ratios and determined that a 25:75 ratio was optimal (Figure 38B). This balance allowed for effective EV binding without compromising vesicle stability, a critical factor for maintaining functional integrity during isolation. A streamlined workflow for EV isolation was developed using the PLD platform, combining vesicle capture, UV-mediated release, and removal of non-vesicular contaminants (Figure 38C). In this process, the PLD complex binds EVs through a combination of electrostatic and hydrophobic interactions. A short UV exposure cleaves the PLN linker, selectively releasing intact EVs while contaminants remain bound to the platform. This UV-triggered mechanism eliminates the need for labor-intensive elution steps common in affinity-based techniques, drastically reducing isolation time to less than 15 minutes. By avoiding harsh centrifugation or chemical treatments, this workflow preserves the integrity of isolated vesicles, making it suitable for both small- and large-scale applications.

[0244] To validate the platform's performance, the size distribution of isolated EVs was analyzed using nanoparticle tracking analysis (NTA). The results showed that vesicles isolated using UV-mediated elution had size distributions comparable to those obtained with biotin- mediated elution, with a predominant peak around 100 nm (Figure 38D). This demonstrates that the present method isolates vesicles without introducing structural deformation. Additionally, EV capture efficiency was quantified over time. Within 15 and 30 minutes, the present platform captured significantly higher amounts of EVs compared to controls lacking PLN (Figure 38E). These results highlight the rapidity and efficiency of the PLD platform, enabling high-throughput isolation without sacrificing quality.

[0245] To ensure the structural integrity of the isolated EVs, fluorescence microscopy and cryo-transmission electron microscopy (cryo-TEM) analyses were conducted. Fluorescence imaging with MemGlow-488 labeling confirmed the successful capture and release of intact EVs after 15 minutes of incubation (Figure 38F). Cryo-TEM imaging revealed that the isolated EVs retained their characteristic spherical morphology and intact lipid bilayer structure (Figure 38G). Compared to conventional isolation techniques, such as ultracentrifugation and size exclusion chromatography, our UV-triggered approach achieves rapid, high-purity isolation while maintaining vesicle morphology and functionality. Furthermore, the ability to scale the platform for larger sample volumes represents a significant advantage for clinical and industrial applications. These findings demonstrate that our PLD platform preserves the delicate structure of EVs, making them suitable for sensitive downstream applications, including drug delivery and biomarker discovery. PLN Probes preserve the vesicle morphology and membrane protein functionality

[0246] The performance of the PLD platform was tested by evaluating the structural and functional integrity of biomimetic extracellular vesicles (bEVs) isolated using this approach. Biomimetic EVs were prepared by mixing natural EVs with synthetic lipids, followed by extrusion through 50 nm and 200 nm membranes. This approach facilitated the incorporation of synthetic lipids into the EV structure, enabling precise control over vesicle composition. Size analysis using dynamic light scattering (DLS) and nanoparticle tracking analysis (NTA) demonstrated effective control over vesicle size. DLS measurements revealed average diameters of 84 ± 12 nm and 146 ± 22 nm for vesicles extruded through 50 nm and 200 nm membranes, respectively. Similarly, NTA showed mean diameters of 96 nm and 149 nm, validating the size distribution consistency across techniques. When the vesicle populations from the 50 nm and 200 nm extrusions were combined, DLS reported a mean size of 136 ± 25 nm, while NTA measured a mean size of 130 nm, confirming the generation of multimodal vesicle populations with controlled size distributions.

[0247] To simulate large-scale conditions, the combined vesicle populations were diluted in 50 mL of buffer and incubated with the PLD platform to facilitate interaction with the probes. After 15 minutes of incubation, the beads were collected and exposed to UV light for 90 seconds to induce photocleavage. This step released the bound vesicles from the beads without requiring external chemicals or elution agents, such as D-biotin. DLS analysis of the released vesicles after 1.5 minutes of UV irradiation showed a uniform size distribution with a mean diameter of 100 nm, confirming that the vesicle morphology was preserved throughout the process (Figure 39B). This rapid binding, stabilization, and release of vesicles within a short timeframe underscores the efficiency of the PLD platform for large-scale isolation workflows.

[0248] Fluorescence microscopy further confirmed successful isolation by revealing clear, bright puncta corresponding to vesicles. To evaluate the lipid bilayer integrity, detergent was added to the vesicle samples. The detergent dissolved the lipid bilayer, confirming the vesicular structure and susceptibility of the isolated EVs to membrane disruption (Fig. SI.2). These findings validate the efficiency of the PLD platform in recovering intact and structurally sound vesicles. To assess whether the UV probes used in the PLD platform caused vesicle aggregation, we incubated the biomimetic EV samples with UV probes for varying durations (3, 10, and 20 minutes) and performed dynamic light scattering (DLS) analysis. The size distributions remained consistent across all conditions, with average diameters of 63 ± 22 nm, 63 ± 18 nm, and 65 ± 22 nm, respectively (Figure 39E). These results demonstrate that the UV probes do not induce vesicle aggregation or compromise vesicle stability, further highlighting the robustness of the PLD platform for vesicle processing.

[0249] To test the functional integrity of the isolated vesicles, a lipid fusion assay was conducted to measure v-SNARE-mediated membrane fusion, a critical function of biomimetic EVs (Figure 39F). Donor bEVs (labeled as V) were prepared containing V-SNARE proteins labeled with RHOD-PE and NBD-PE fluorophores to establish a FRET pair, resulting in a fluorescence "OFF" state. After UV-triggered release, these donor vesicles were mixed with target vesicles (T) to assess fusion. Successful membrane fusion disrupted the FRET pair, restoring NBD-PE fluorescence to the "ON" state, which served as a functional readout of v- SNARE activity. Our quantitative analysis revealed that vesicles exposed to 1 minute of UV irradiation exhibited fusion efficiency similar to density gradient centrifugation, confirming that the PLD platform preserves membrane protein functionality during isolation (Figure 39G). However, extending UV exposure to 2 minutes led to reduced fusion efficiency, likely due to overexposure compromising protein activity.

[0250] These results collectively demonstrate that the PLD platform isolates bEVs rapidly and efficiently while maintaining their structural and functional integrity. The platform’s ability to produce biologically active vesicles suitable for functional assays highlights its potential for diverse applications in nanomedicine, including drug delivery, gene therapy, and diagnostics. By combining speed, scalability, and functionality, the PLD platform offers a transformative solution to longstanding challenges in EV isolation and purification, paving the way for its adoption in both research and clinical applications.

[0251] Isolation and Orientation-Specific Enrichment of Nano Plasma Membrane Vesicles (NPMVs) Using PLN Probes

[0252] Nano plasma membrane vesicles (NPMVs) have gained significant attention for their applications in drug delivery, biosensing, and in vivo imaging. However, isolating NPMVs with a correct orientation that mimics the natural geometry of cellular membranes remains a challenge. Right-side-out (RSO) vesicles, in which the glycans and interacting proteins on the plasma membrane face outward, are critical for achieving effective target-cell adhesion and minimizing immune clearance. In contrast, inside-out (ISO) vesicles can result in adverse effects, such as off-target interactions and rapid immune clearance. Given their abundance and simplicity, red blood cells (RBCs) serve as an ideal source for generating NPMVs.

[0253] The preparation of NPMVs involves a top-down approach, typically starting with the generation of cell membrane vesicles (CMVs) via hypotonic treatment, homogenization, and high-speed centrifugation. Sequential extrusion of CMVs through membranes with defined pore sizes produces NPMVs of approximately 100 nm in diameter. Due to the random wrapping of membranes during extrusion, NPMVs inherently exhibit mixed orientations. RSO vesicles are preferred as delivery platforms because their external membrane orientation maintains functional interactions with target cells, whereas ISO vesicles pose limitations. Addressing the orientation of coreless NPMVs is, therefore, critical for ensuring their efficacy as therapeutic carriers.

[0254] Extrusion process exhibit mixed-orientations in NPMVs lipid asymmetry.

[0255] PLN probes were employed to investigate and regulate the orientation of NPMVs. Amphiphilic PLN probes were designed to anchor selectively onto the outer membrane of RBCs, leveraging their integration into the outer phospholipid bilayer before extrusion. The workflow is illustrated in Figure 40 A. The results demonstrated that NPMVs extruded from RBC membranes inherently exhibit mixed orientations due to the random wrapping of membranes during the extrusion process.

[0256] Fluorescence microscopy confirmed the successful labeling of CMVs with NBD- PE dye, as evidenced by bright circular signals (Figure 40B). Length and size profiles of the vesicles, measured via size characterization techniques, showed well-defined peaks corresponding to the expected uniformity of the membrane vesicles (Figure 40C, D). The dithionite quenching assay demonstrated that NBD-PE fluorescence was selectively quenched in the outer leaflet of the vesicles, confirming the preservation of lipid asymmetry (Figure 40E). Following extrusion, the vesicle population was found to consist of 57% RSO vesicles and 43% ISO vesicles (Figure 40F). These findings confirm the mixed orientation distribution of extruded NPMVs, with a slight preference for RSO vesicles due to the random wrapping principle. PLN selectively isolates RSOs from mixed populations.

[0257] To address the challenge of isolating RSO vesicles from mixed populations, the method was developed by leveraging functionalized PLN probes. These probes were modified with DSPE lipids on one end for membrane integration and biotin affinity tags on the other end for bead-based separation. The CMVs were extruded through defined pore-size membranes, and the resulting NPMVs were incubated with monomeric avidin beads to enable selective isolation of biotinylated RSO vesicles (Figure 40G). The moderate affinity of monomeric avidin for biotin enabled efficient washing of non-biotinylated ISO vesicles, while bound RSO vesicles were eluted using D-biotin in phosphate-buffered saline (PBS).

[0258] DLS analysis of the enriched RSO vesicles revealed a uniform size distribution with an average diameter of approximately 100 nm, consistent with expectations (Figure 40H). Fluorescence microscopy confirmed the successful enrichment of NBD-PE-labeled vesicles with defined orientation, as evidenced by homogeneous fluorescent signals (Figure 401) Quantification using a dithionite quenching assay demonstrated that the RSO vesicle population was enriched to 83% following the isolation process (Figure 40J). This represented a significant improvement compared to the initial mixed population and underscored the efficiency of the PLN probe-enabled isolation workflow. Further functional validation was performed using RSO vesicles carrying VAMP2-GFP protein, which confirmed the retention of membrane protein functionality. Fluorescence microscopy showed intact RSO vesicles with retained protein expression, validating their potential for delivering cargo to target cells (Figure 40K). The high purity and functional integrity of the enriched RSO vesicles demonstrate the suitability of this platform for therapeutic and diagnostic applications.

[0259] These results highlight the capability of PLN probes to selectively isolate RSO vesicles from mixed NPMV populations. By preserving structural integrity and membrane orientation, this approach offers a robust platform for generating high-purity RSO vesicles suitable for drug delivery, biosensing, and other biomedical applications. The selective enrichment of vesicles with the correct orientation and functionality underscores the transformative potential of the PLN probe-enabled platform for targeted therapeutic delivery and precision medicine. Size-Specific isolation, sorting of lipid vesicles using PLN Probes

[0260] Rapid Size-Specific Isolation and Enrichment of Lipid Vesicles

[0261] Next, the efficacy of PLNs to achieve size-specific separation of vesicles using light was tested. The approach was applied to both EVs and bEVs, as outlined in the schematic workflow (Figure 41A). PLNs were integrated into the lipid membranes of the vesicles, and UV irradiation was used to cleave the probes in a size-dependent manner. Factors such as light intensity, exposure time, and the density of PLNs per vesicle membrane were tested to observe their influence on the separation process. Vesicle size distributions were analyzed after UV exposure using DLS. The results demonstrated a distinct size-dependent release of vesicles as a function of UV irradiation time (Figure 41B). After 1.5 minutes of exposure, smaller vesicles dominated the population, while prolonged exposure of 3 minutes resulted in the release of larger vesicles. Quantification of vesicle sizes (Figure 41C) confirmed that smaller vesicles, with fewer PLNs incorporated into their membranes, responded more rapidly to UV exposure. In contrast, larger vesicles required longer exposure times to achieve similar separation. The area under the curve (AUC) of DLS size distribution profiles was measured to estimate the relative abundance of vesicles released at different UV exposure times (Figure 4 ID). The analysis revealed that number of particles released depends on the irradiation times. These findings validated the capability of the PLN probes to achieve light-controlled, size-specific separation of vesicles.

[0262] Dot blot analysis of canonical exosomal markers (CD9, CD63, CD81, and Alix) were performed across different UV exposure durations (1.5, 3, and 10 minutes). The tests showed that these markers were preserved across all experimental conditions (Figure 4 IE), indicating that the light exposure and probe cleavage process did not compromise vesicle integrity or protein composition.

[0263] To test whether the size separation would work for all types of vesicles, mechanism underlying the size-specific separation by quantifying vesicle diameters after 1 and 3 minutes of UV exposure (Figure 4 IF). The results showed that smaller vesicles, due to their lower PLN density, responded more rapidly to UV irradiation, resulting in faster cleavage and release. In contrast, larger vesicles required extended UV exposure for comparable PLN cleavage. Statistical analysis revealed a significant difference in the size distributions between the two exposure durations, highlighting the precision and robustness of the PLN platform for size-specific vesicle separation. The present example introduces a groundbreaking material platform for the rapid isolation of EVs in their native state, achieving high purity at a large scale in under 15 minutes. This platform addresses critical challenges in producing clinical-grade EVs for large-scale therapeutic applications. It is also demonstrated that PLNs allow precise, light-controlled separation of EVs, NPMVs, and bEVs with size and orientation specificity. The present approach preserves vesicle integrity and membrane protein functionality, validated through multiple advanced techniques. PLNs enable size- and orientation-dependent separation without the need for complex centrifugation or chromatography, achieving size-specific isolation in less than 3 minutes. Notably, the selective enrichment of right-side-out (RSO) vesicles is a critical advancement, ensuring preserved membrane orientation for effective target-cell interactions. This platform is fast, scalable, and adaptable, offering a clear advantage over traditional methods. It eliminates the need for chemical elution agents, minimizes contamination risks, and preserves the bioactive properties of vesicles, making it ideal for a wide range of applications in nanomedicine. With its simplicity, precision, and speed, this platform has the potential to transform the approach to liquid biopsies, drug delivery, and the development of next-generation therapies.

[0264] The materials and methods employed in these experiments are now described. Materials and Methods

[0265] Non-photosensitive chemistry probes were obtained from NANOCs. Photosensitive LNs were synthesized purified in-house, see Supplementary Note 2 for synthesis and characterization details. All lipids were purchased from Avanti Polar Lipids.

[0266] Lipid mixing assay.

[0267] FRET-based lipid mixing experiments were conducted on a plate reader (Molecular Devices). 5 pL of v-SNARE liposomes labeled with a pair of FRET dyes (donor: NBD-DOPE, acceptor: Rhodamine-DOPE) and 45 pL of unlabeled t-SNARE liposomes, with a total lipid concentration of 3 mM were separately heated to 37 °C and transferred to a Pierce™ 96-Well Polystyrene Plates. NBD fluorescence was monitored at emission / excitation of -535 / 460 nm every 1 min for 1 hr and 40 mins. The fluorescence signal was normalized using Norm. NBD fluorescence [% of Imax] = 1 00%x (It-I min) / (Imax“Imin ), where I min was the lowest observed NBD fluorescence intensity and Imax was the maximum NBD fluorescence intensity observed after addition of detergent (2.5% [w / v] n-dodecylmaltoside [DM]).

[0268] INS-1 rat insulinoma culture.

[0269] The clonal beta-cell line 832 / 3, derived from the parental INS-1 rat insulinoma cells, was purchased from Millipore Sigma and cultured in complete Roswell Park Memorial Institute (RPMI) medium (Sigma Aldrich) supplemented with 10% ultracentrifuged exosome- free FBS (Sigma Aldrich), 10 mM HEPES (Sigma Aldrich), 2 mM L-Glutamine (Sigma Aldrich), 1 mM sodium pyruvate (Sigma Aldrich) and 0.05 mM beta-mercaptoethanol (Sigma Aldrich) in 5% CO2 at 37 °C. Media was changed every 2-3 days with the macrophages fully differentiated and ready for experimentation on the day 7. Culture media was extracted at -50% confluence (Fig. S7a).

[0270] EV isolation.

[0271] If not mentioned otherwise, in most experiments 200 pL of avidin (monomeric / neutravidin) - HCAB slurry (Thermo Scientific) was pipetted into 0.8 m centrifuge columns with 30 pm polyethylene filters (Thermo Scientific) (Fig. S5d). For larger volumes 5 mb and 10 mL centrifuge columns were used (Thermo Scientific). Given that 4% beaded agarose is provided in a 50% slurry, 100 pL of settled resin equates to 200 pL of slurry. Buffer exchange was performed on the HCAB with Buffer A for liposome and proteoliposome isolation or PBS for EV isolation, through a series of 5 - 200 pL washes based on slurry volume. In parallel, 5% of LN (150 pM for 3 mM lipid concentration) was incubated for Ih with liposomes and proteoliposomes if not mentioned otherwise. In some cases, 1 and 3 mol / mol% were tested for the micelle characterization study. For EV isolation, 100 pg of LN was added into 50 mL of culture media and incubated for Ih. The buffer exchanged HCAB were then incubated with LN incorporated vesicle sample and shaken for 30 minutes. In cases of EV isolation with large 50 mL sample volumes the HCABs were left overnight to maximize vesicle capture and enrichment. Following incubation, the sample was centrifuged (150 g x 20 sec) to collect the post-capture sample which would be used to determine the liposome trapping efficiency of the HCABs. The HCAB were prevented from drying and washed 5-10 times using the 200 pL (based on slurry

[0272] 1 volume) of relevant buffer during the purification stage based on the sample. EV samples were washed with upto 2 mL of PBS. For the elution step, to concentrate the sample, 100 pL (1 / 2 slurry volume) of 2.5 mM D-biotin in the buffer was pipetted into HCAB and agitated for a 2- minute incubation period. For sequential D-biotin elution the process was repeated. For the D- biotin concentration experiments, elution buffer - buffer A or PBS with desired D-biotin concentration was used. For light-based elution, the purified HCAB was incubated with 100 pL (1 / 2 slurry volume) of buffer A or PBS without D-biotin and the column was exposed to either 840 or 365 nm lamp at 5mW / cm2for NIR and UV, respectively. The sample was exposed to light for 10 min, if not mentioned otherwise. The solution was then centrifuged (200 g x 20 sec) to elute isolated vesicles. For reuse, the HCAB were washed with regeneration buffer containing 0. IM glycine at pH 2.8 (Thermo Scientific) and centrifuged (200 g x 15 sec) a total of 5 times, then exchanged with buffer 5 times. Note that all centrifugation steps were performed using a benchtop unit to speed up the process, however, with larger sample volumes centrifuges can be completely avoided if needed. This process has been summarized in Figure 38B.

[0273] Liposomes, proteoliposome and EV size characterization.

[0274] Dynamic light scattering (DLS).

[0275] DLS was performed using Wyatt DynaPro Nanostar model WDPN-06. 25 pL of each elution was pipetted into a disposable cuvette. Peak radius cutoffs were at 0.5 nm and 10,000 nm. 10 acquisitions were collected at 25°C for each sample with one acquisition captured every 10 seconds. Batch DLS with regularization processing was utilized to average repeat experiments to calculate average hydrodynamic diameter within each sample. This allowed us to visualize the sizes of single vesicles and determine the presence of any aggregated vesicles as outliers. Samples were not diluted in this study. The raw data was obtained and fitted with GraphPad Prism.

[0276] Nanoparticle Tracking Analysis (NTA).

[0277] NTA was performed on a ZetaView® x30, Next Generation Nanoparticle Tracking Analyzer (Particle Metrix GmbH), equipped with a 405 nm laser and a CMOS camera. Data analysis was performed on the ZetaView software (version 8.04.04), applying a bin class width of 5 nm, a minimum brightness of 25, a minimum area of 5, a maximum area of 1000, and a trace length of 15. First, the device was first calibrated with polystyrene beads (Sigma- Aldrich) diluted by 1 :500,000 (v / v) at a concentration of -100 particles per frame. For the measurements, the shutter was set to 150, the sensitivity to 85, and the frame rate to 30. EV samples were diluted 1 : 10 (v / v) in PBS for the measurements. The raw data was plotted and fitted with GraphPad Prism.

[0278] DLS and NTA size data analysis.

[0279] For both DLS and NTA the experimental replicates (3-5) were first averaged and fitted using either a Gaussian function G(x) or a Lorentzian L(x) based on the presence or absence of trailing end of the distribution and were verified with R-squared value.

[0280] _0 5(L— L)2

[0281] G(x) = aevc , where ‘a’ is the amplitude or peak intensity of the curve, ‘b’ is the mean diameter where the peak occurs, and ‘c’ is the standard deviation which controls the width of the bell curve. where a is the amplitude or peak intensity of the curve, ‘b’ is the center of the, and c is the standard deviation which controls the width of the bell curve. From this equation, we also calculated the poly dispersity index, a measure of size heterogeneity in the sample, by squaring

[0282] C the ratio of standard deviation to mean size (-) . b

[0283] Single vesicle fluorescence microscopy.

[0284] A custom TIRF microscopy, previously described (Bera, M. et al., 2023, Proc.

[0285] Natl. Acad. Sci., 120, e2311484120; Kalyana Sundaram, R. V. et al., 2022, Small, 18, 2205567)

[0286] - Nikon inverted microscope equipped with three laser lines (488, 532, and 633 nm), a Photometries DV2 dual view, and an Andor EMCCD digital camera was used to was used for single vesicle fluorescence imaging. Two fluorophore channels were used based on the experiment; FITC and MEMGlow647™ were performed in the 488 and 647 channel, SRB and Atto-647N-PE were performed in the 555 and 647 channel; CD9 / CD63 labelled with Alexa Fluor 488 and Sytl labeled with Alexa Fluor 647 were performed in 488 and 647 channel. For all measurements, images were obtained at 512 * 512 pixels with a dynamic range of 16-bit grayscale. The exposure time was set to 150 ms in both channels. Images were converted to TIFF using FIJI61and analyzed with custom Wolfram Mathematica algorithm. All colocalization efficiencies were computed using the identified vesicles in each channel.

[0287] Negative stain electron microscopy.

[0288] For preparing the negative stain EM grids of sEVs and Au-labelled sEVs (CD81 was labelled - see Supporting information), 5 pL of the isolated samples (in PBS) were directly applied to a glow discharged Carbon Type-B, 400 mesh, Copper grid (Ted Pella). After 1-minute incubation on the grid, the samples were quickly washed with 5 pL of 2% (w / v) uranyl formate solution, followed by staining with another 5 pL of 2% uranyl formate for 1 minute. The grids were imaged using an 80 kV JEOL JEM-1400Plus microscope equipped with a bottom-mount 4 k x 3 k charge-coupled device camera (Advanced Microscopy Technologies).

[0289] Western Blot.

[0290] EV samples were lyzed with electrophoresis (Laemmli) sample buffer and boiled the mixture at 95-100 °C for 5 min. Proteins were separated by SDS / PAGE method, and performed western blots based on the Biorad western blot protocol. The blots were stained with antibody depending on the protein analyzed and incubated overnight. The blot was developed with ECL western blot substrate.

[0291] Statistics and reproducibility.

[0292] Data were represented as the means ± SEM or mean ± SD, mentioned in all appropriate cases. Unpaired Student’s t test was used to analyze data with only two sets. Two- way analysis of variance (ANOVA) was performed to determine whether there was a significant difference between more than two datasets, following Tukey’s method, using GraphPad Prism 6.0. Group differences at the level ofP < 0.05 were considered statistically significant. Asterisk (*) represented p < 0.05; double asterisk (**) represented p < 0.01; triple asterisk (***) represented p < 0.001; quadruple asterisk (****) represented p < 0.0001.

[0293] The disclosures of each and every patent, patent application, and publication cited herein are hereby incorporated herein by reference in their entirety. While this invention has been disclosed with reference to specific embodiments, it is apparent that other embodiments and variations of this invention may be devised by others skilled in the art without departing from the true spirit and scope of the invention. The appended claims are intended to be construed to include all such embodiments and equivalent variations.

Claims

CLAIMSWhat is claimed is:

1. A lipid nanoprobe (LN) comprising: a lipid insert; a cleavable linker; a biocompatible polymer chain; and an affinity tag, wherein the cleavable linker connects the lipid insert and a first end of the biocompatible polymer chain, and wherein the affinity tag is connected to a second end of the biocompatible polymer chain.

2. The lipid nanoprobe of claim 1, wherein the lipid insert is a sterol or a phospholipid.

3. The lipid nanoprobe of claim 2, wherein the phospholipid is phosphoethanolamine.

4. The lipid nanoprobe of claim 3, wherein the phosphoethanolamine is 1,2- distearoyl-sn-glycero-3-phosphoethanolamine (DSPE).

5. The lipid nanoprobe of claim 1, wherein the cleavable linker is a photocleavable linker.

6. The lipid nanoprobe of claim 5, wherein the photocleavable linker is an 'Ch- cleavable linker.

7. The lipid nanoprobe of claim 6, wherein the 'Ch-cleavable linker is (Z)-2,2'- (ethene-l,2-diylbis(sulfanediyl)di ethanamine) (BSD A).

8. The lipid nanoprobe of claim 1, wherein the biocompatible polymer chain is poly(ethylene glycol) (PEG).

9. The lipid nanoprobe of claim 8, wherein the average molecular weight of PEG is 1-10 KDa.

10. The lipid nanoprobe of claim 9, wherein the average molecular weight of PEG is about 5 KDa.

11. The lipid nanoprobe of claim 1, wherein the affinity tag is selected from the group consisting of: biotin, D-biotin, desthiobiotin, avidin, streptavidin, and neutravidin.

12. The lipid nanoprobe of claim 11, wherein the affinity tag is desthiobiotin.

13. A method for isolating vesicles from a sample, the method comprising: a) incubating the sample with a lipid nanoprobe (LN) to form LN-tagged vesicles, wherein the LN comprises a lipid insert, a linker, a biocompatible polymer chain, and an affinity tag, wherein the linker connects the lipid insert and a first end of the biocompatible polymer chain, and wherein the affinity tag is connected to a second end of the biocompatible polymer chain; b) capturing the LN-tagged vesicles by incubating the LN-tagged vesicles with a bead, wherein the bead comprises a binding partner for the affinity tag of the lipid nanoprobe; c) isolating the beads; d) releasing the LN-tagged vesicles from the beads; and e) collecting the LN-tagged vesicles.

14. The method of claim 13, wherein the beads of step b) are high-capacity agarose beads (HCABs).

15. The method of claim 13, wherein step c) further comprises a step of washing the beads one or more times with a sterile buffer, wherein the sterile buffer is water, PBS, HEPES, or Tris.

16. The method of claim 13, wherein the linker of the lipid nanoprobe is a photocl eavable linker.

17. The method of claim 16, wherein step d) comprises releasing the LN-tagged vesicles by exposing the isolated beads to Near-infrared (NIR).

18. The method of claim 13, wherein step d) comprises releasing the LN-tagged vesicles by contacting the isolated beads with an elution buffer comprising a competitive ligand for the binding partner.

19. The method of claim 18, wherein the concentration of the competitive ligand in the elution buffer is gradated to preferentially release a specific size of LN-tagged vesicle.

20. The method of claim 19, wherein smaller vesicles are preferentially released as the concentration of the competitive ligand in the elution buffer increases, wherein the smaller vesicles have a diameter of about 50 nm.

21. The method of claim 13, wherein the method further comprises a step of passing the biological sample through a filter before step a).

22. The method of claim 22, wherein the filter is a 0.2 pm or 0.4 pm filter.

23. The method of claim 13, wherein the isolated vesicles from step e) maintain their native state.

24. A kit comprising the lipid nanoprobe of claim 1.

25. The kit of claim 25, wherein the kit further comprises one or more selected from the group consisting of: HCABs, a filter, a wash buffer, and an elution buffer.

26. The kit of claim 26, wherein the HCABs are functionalized with a binding partner for the affinity tag of the lipid nanoprobe.

Images