Optimizing of Culture Conditions for the Isolation and Characterization of Extracellular Vesicles Derived from Lactobacillus crispatus PMC201

1. Introduction

According to the World Health Organization (WHO) and the U.S. Food and Administration (FDA), probiotics are microorganisms that are beneficial to the health of the host when consumed in sufficient quantities [1] and food ingredients that can increase the intake of beneficial microorganisms in the human body when consumed [2]. Probiotics are microorganisms that are beneficial to the host. They are commonly found in the digestive tract, where they live in symbiosis with humans. They are subjects of much microbiome research. Probiotics are known to benefit human health and play a particularly significant role in supporting good health [3]. They hold significant potential in mitigating conditions associated with dysbiosis, including allergic disorders, gastrointestinal diseases, metabolic ailments, oncological conditions, and cardiovascular pathologies [4]. Probiotics produce a wide range of bioactive compounds including vitamins, short-chain fatty acids (SCFAs), amino acids, peptides, enzymes, and exopolysaccharides (EPS) that can promote beneficial health effects for the host [5]. Particularly, lactic acid bacteria (LAB) celebrated for their ability to produce Gram-positive lactate have been categorically presented as probiotic frontrunners [6]. Probiotics play a significant role in enhancing host defenses against infections through diverse mechanisms [7]. They reinforce the epithelial barrier, inhibit pathogen adhesion, and competitively modulate colonization of pathogenic microbes. Additionally, probiotics can produce antimicrobial agents, modify toxins, and boost host immune responses through both nonspecific and specific immune stimulation [8]. However, understanding specific pathways and regulatory molecules responsible for beneficial effects of probiotics remains challenging [9]. Significant progress has been made in chemistry, biology, and membrane vesicle research over the past decade. These advancements have enhanced the efficiency and targeted delivery of probiotics, expanding their utility in treating various diseases [10]. Recent investigations have emphasized the significance of EVs derived from probiotics and indicated their significant role in enhancing probiotic efficacy [11,12]. EVs play essential roles in interactions between bacteria and hosts, transporting signals associated with immune responses and modulating diverse signaling pathways [13]. Protein compositions of bacteria-derived extracellular vesicles (BEVs) vary depending on specific attributes such as taxonomic categorizations (genus, species, lineage) and symbiotic or pathogenic characteristics. BEVs appear as spherical or nano-sized vesicles, with diameters ranging from 20 nm to 400 nm. They transport abundant microRNAs, mRNAs, proteins, and assorted bioactive factors [14,15]. BEVs primarily activate immune cells. In microbial communities, bacterial cells engage in complex interactions, wherein EVs play significant roles in both cooperative and competitive strategies [16]. Several studies have reported that bacterial EVs can mediate resistance through various mechanisms, including horizontal gene transfer [17], representation of extracellular and intracellular antibiotics by EVs, and presence of enzymes for antibiotic resistance associated with EVs. BEVs can bind/capture antibiotics in the extracellular compartment, thereby protecting microbial communities and providing bacteria with defense against various antibiotics. EVs represent a set of microbial-associated molecular patterns (MAMPs) recognized by specific receptors expressed in epithelial and immune cells of the host [18]. These pattern recognition receptors (PRRs) are considered vital components of innate immunity as they can detect intestinal microbes and mediate appropriate immune responses [19]. Toll-like receptors (TLRs) primarily located on the cell surface or within endosomal membranes can initiate cellular signaling pathways that regulate immune and defensive responses when they bind to MAMPs such as polysaccharides, lipopolysaccharides (LPS), and lipoproteins found on surfaces of EVs. These attributes suggest the involvement of vesicles in the interplay between infection and the immune system [20,21]. Previous studies have reported that two Lactobacillus strains, L. crispatus and L. gasseri, isolated from vaginal microbiota of healthy women, can release EVs to protect human cells and tissues from HIV-1 infection. These EVs contain various bacterial metabolites and proteins associated with anti-HIV-1 effects [22]. Moreover, EVs can directly deliver therapeutic substances, supporting targeted action on specific molecules or cells. This can enhance treatment efficiency and precision compared to using microbial strains alone as therapeutics. Previous studies have confirmed the effectiveness of Lactobacillus crispatus PMC201 in reducing M. tuberculosis H37Rv and extensively drug-resistant (XDR) M. tuberculosis in co-culture conditions [23]. This study was conducted to find optimal conditions for studying probiotic derived EVs. In addition, the corresponding proteome was confirmed. For L. crispatus PMC201, EVs were isolated using food-grade medium [24], a self-developed medium rather than commercial media. Proteomic analysis was then performed using liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) [25,26]. In L. crispatus PMC201, a total of 64 proteins were identified, predominantly composed of cell wall constituents.

2. Materials and Methods

2.1. Optimized Culture Media for Isolating of EVs

Food-grade culture medium, a self-made medium rather than a commercial one, was prepared by adding 10 g of Soy Peptone A2 SC, 10 g of yeast extract, 20 g of glucose, 1 ml of Tween 80, 0.1 g of magnesium sulfate, and 5 g of sodium chloride to 1 L of water

2.2. Isolation of Bacterial EVs

L. crispatus PMC201 was cultured in a food-grade medium at 37°C on a shaking incubator (Biofree, Seoul, KOR) for 20 hours. Subsequently, the optical density (OD) of the probiotic culture was confirmed to be in the range of 1.0–1.2 using a portable VIS spectrophotometer (DR 1900, Hatch) at 600 nm. L. crispatus PMC201 was pelleted by centrifugation twice at 3515 × g for 20 minutes at 4°C. The culture supernatant of L. crispatus PMC201 was filtered through a 0.22 μm membrane (Merck Millipore, Burlington, MA, USA) and concentrated 10-fold using a 100-kDa membrane filter (Pall, Port Washington, NY, USA). It was further concentrated 10-fold with an Amicon Ultra Centrifugal Filter (100 kDa MWCO). EVs derived from L. crispatus PMC201 were purified from culture supernatants concentrated 100-fold. For the ultracentrifugation (UC) method, the concentrated conditioned medium was centrifuged at 150,000 g for 2 hours at 4°C using an ultracentrifuge Optima MAX-XP (Beckman Coulter, Fullerton, CA, USA) with an MLA-55 rotor. Pelleted EVs were resuspended in PBS. Size exclusion chromatography (SEC) method obtains fractions using a qEVoriginal/35 nm Gen 2 column (Izon Science, Christchurch, NZ). Protein content was quantified using a Micro BCA assay (Thermo Scientific, 23235).

2.3. Bio-TEM Imaging of EVs

To visualize EVs samples, transmission electron microscopy (TEM) was utilized. For negative staining, EVs samples were affixed to a carbon-coated 200-mesh copper grid and then stained with 2% (w/v) uranyl acetate. Imaging of EVs was performed using a 120 kV TEM (Bio-TEM, Tecnai G2 Spiri TWIN, FEI). Diameters of EVs were determined using ImageJ software, with measurements obtained at 200 nm through a CCD camera.

2.4. NTA Measurements of EVs

Size distribution analysis and concentration measurements were performed using the nanoparticle tracking analysis (NTA) and Nano Sight NS300 system with a sCMOS camera, Green488 laser, and NTA software Version 3.4 (Malvern Instruments, Malvern, UK). Before analysis, samples were vortexed and diluted to 1 mL with PBS buffer. Measurements were performed at 25°C. Samples were recorded three times for 60 s with a camera level of 16 and the threshold parameter set at 5 to obtain concentrations in the range of 10⁷–10⁹ particles mL−1 (corresponding to 20–100 particles per frame).

2.5. Methanol/Chloroform Precipitation

Proteins were precipitated using a methanol/chloroform mixture to remove lipids. A 100 µg protein sample was mixed with methanol (260 µL) and vortexed. After centrifugation at 9,000 x g for 10 seconds, chloroform (65 µL) was added and the mixture was vortexed and centrifuged again. After adding HPLC-grade water (195 µL), the solution was centrifuged at 16,000 x g for 2 minutes to pellet proteins. Methanol (195 µL) was then added to the protein pellet, vortexed, and centrifuged at 16,000 x g for 3 minutes. The supernatant was removed and the protein pellet was air-dried before resuspension in 30 µL of 2X dye for further analysis.

2.6. SDS-PAGE and Coomassie Blue Staining

SDS-PAGE was conducted to separate proteins based on molecular weight. Equal numbers of purified EVs derived from L. crispatus PMC201 were analyzed using denaturing polyacrylamide gel electrophoresis with a linear concentration gradient of polymer (4–15% Tris-glycine gel). Following electrophoresis at 150 V for 70 min, the gel was then stained with Gel Code Blue Stain Reagent (Thermo, USA), destained, and imaged using a Bio-Rad ChemiDoc MP Imaging System

2.7. Proteomic Analysis of EVs Derived from L. crispatus PMC 201

All analyses were performed in triplicate to ensure reliability. A total of 200 µg of EVs underwent methanol/chloroform precipitation according to the established protocol (PMID: 33420990). Precipitated proteins were subsequently subjected to in-gel digestion by trypsin following procedures outlined in a previous study (PMID: 34200450). Resulting tryptic peptides were dissolved in 20 μL of buffer A (0.1% formic acid in water) and 5 μL was injected onto a trap column, Acclaim PepMap C18 nano Viper 100 (75 μm × 2 cm, 3 μm) (Thermo Fisher Scientific) with a flow rate of 5 μL/min using 95% buffer A for 4 min for LC-MS/MS analysis. This analysis was performed using an Orbitrap Eclipse™ Tribrid™ mass spectrometer (Thermo Fisher Scientific, San Jose, USA) coupled with an Ultimate™ 3000 UHPLC system (Thermo Fisher Scientific, USA). Peptides were separated on an analytical column, PepMap RSLC C18 ES803A (75 μm × 50 cm, 2 μm) (Thermo Fisher Scientific), using a 180 min gradient with 5-90% of solvent B (0.1% formic acid in acetonitrile) at a flow rate of 300 mL/min. The mass spectrometer was operated in a data-dependent Top 20 scans mode, switching between MS and MS2. Peptides were analyzed with the following parameters: 10 ppm of mass accuracy, 1850 V of ion spray voltage, 275°C of capillary temperature, m/z 375-1575 resolution for full scans, and 120,000 higher-energy collisional dissociation activation scans with 35% normalized collision energy. The quadrupole isolation window was set to 1.4 Da. MS/MS spectra were detected using an Orbitrap instrument with a resolution of 30,000. MS raw files were analyzed using the MaxQuant program (version 2.2.0.0) (Max Planck Institute, Germany) against the L. crispatus PMC201 protein database (release 2023_04) from UniProt. A tolerance of 20 ppm was applied for precursor ions and fragment ions. Trypsin digestion allowed for up to two potential missed cleavages. Modifications included fixed carbamidomethylation of cysteine (57 Da) and variable deamidation of asparagine and glutamine (1 Da), methionine oxidation (16 Da), and N-terminal acetylation (42 Da). All identified proteins underwent statistical processing using the Scaffold program (version 5.3.0). Proteins meeting a protein threshold of ≥ 0.95 and containing a minimum of 2 peptides with a protein threshold of ≥ 0.95 were selected for further analysis. Protein quantitation was conducted using normalized weighted spectra in the Scaffold program.

2.8. Integrated Bioinformatics Analysis of protein Sequences from Proteomic Data

Protein bioinformatics analysis was conducted using a subset of data obtained through the Scaffold program (version 5.3.0) for proteomic analysis. Identified protein sequences were analyzed through protein-protein interaction network analysis using String (version 12.0). Protein network visualization used Cytoscape (version 3.8.2). Protein location prediction used DeepLoc2. Subsequently, sequences were mapped to Gene Ontology (AmiGO 2) to verify their biological and functional properties. Furthermore, enriched pathways associated with identified proteins were evaluated using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Data were visualized using R programming. Data were visualized using R programming.

2.9. Statistical Analysis

NTA data used GraphPad Prism(version.10) software to conduct the statistical analysis, specifically using the t-tests (p-value < 0.05 or 0.01). Statistical significance was considered for P-values (P) of ≤0.05. Protein pathways data values were based on intensity values derived from network analyses (PPI enrichment p-value <1.0$\times$10^-16, FDR ≤ 0.05).

3. Results

3.1. Isolation of EVs Derived from L. crispatus PMC201

EVs were isolated by removing bacterial pellets and cell fragments, EVs were filtered and concentrated using a 100 kDa cutoff to prevent the presence of EVs with molecular weights < 100 kDa in the supernatant. Finally, EVs were pelletized through ultracentrifugation to remove supernatants and any remaining soluble proteins to isolate EVs (Figure 1A). Subsequently, we identified proteins secreted by these EVs. To isolate EVs, SEC and UC methods were employed. The presence of EVs in the isolated sample was confirmed through TEM and NTA. Both SEC and UC verified that L. crispatus PMC201 secreted EVs.

3.2. Comparison of EVs Size and Concentration

3.2.1. Comparison of EVs between Culture Media

EVs isolated from L. crispatus PMC201 were obtained using the SEC method under two culture conditions using MRS media or food-grade media. SEC yielded fractions 7 and 8, which were typically associated with EVs. To quantify nanoparticles in EVs obtained from both culture media, NTA was employed (Figure 2, Table 2). Although the size of L. crispatus PMC201 was confirmed to be larger in MRS media than in food-grade media, the particle number was measured to be higher for L. crispatus PMC201 grown in food-grade media. Since this was a result obtained by diluting 10 times, when total particles were checked, it was confirmed that EVs obtained by culturing L. crispatus PMC201 in food-grade media secreted 2.5 times more particles. Sizes of EVs obtained from both media were in the range of 30 nm to 200 nm. However, since food-grade media clearly secreted a large amount of EVs at this concentration, it is efficient to isolated EVs by culturing L. crispatus PMC201 in food-grade media (Table 1) [PMC 201- MRS: (Mean: 136+/-7.1 nm, Mode: 107.6+/-8.7 nm, SD: 62.3+/-12.6 nm, Concentration (particles/mL): 1.06$\times$10⁷+/-9.60$\times$10⁶), PMC 201-FGM: (Mean: 119.8+/-5.3 nm, mode: 88.05+/- 6.7 nm, SD: 58.6+/-12.1 nm, concentration (particles/mL): 2.31$\times$10⁸+/ -1.02$\times$10⁷)]

3.2.2. Comparison of EVs According to Isolation Methods

EVs isolated from L. crispatus PMC201 was obtained using SEC and UC methods (Figure 1A). Since previous results showed that particle secretion was more prominent in food-grade media than in MRS, EVs were isolated from strains grown in food-grade media for both SEC and UC. Fractions 1-12 of SEC were identified, yielding fractions 7 and 8, which were generally associated with EVs (Figure 3B). To quantify nanoparticles in EVs obtained from both concentration methods, NTA was employed (Figure 3, Table 2). For PMC 201, the number of nanoparticles was similar with both SEC and UC. However, the size was significantly larger in UC than in SEC [PMC 201 - SEC: (Mean: 119.8+/-5.3 nm, Mode: 88.05+/-6.7 nm, SD: 58.6+/-12.1 nm, Concentration (particles/mL): 2.31$\times$10⁸+/-1.02$\times$10⁷), PMC 201 - UC: (Mean: 166.9+/-13.5 nm, Mode: 110.1+/-7.0 nm, SD: 89.9+/-4.1 nm, Concentration (particles/mL): 3.50$\times$10⁸+/-2.54$\times$10⁷)](Figure 3A). As a result of comparing EVs isolated from L. crispatus PMC201 separated in SEC and UC, both size and particle number were observed to be higher in UC than in SEC (Figs. 3C, D). In the case of particle number, those in UC secreted 1.5 times more particles than those in SEC. This suggests that it is more efficient to isolate EVs using UC than using SEC.

3.3. EVs Derived from L. crispatus PMC201 Confirmed by TEM

Shapes of EVs isolated from L. crispatus PMC201 were confirmed by TEM. Imaging confirmed the presence of EVs derived from L. crispatus PMC201 within the size range of 100-200 nm (Figure 4). Previously, measuring the size using NTA (Figure 3) revealed that nanoparticles of all samples had sizes in a range of 100 nm to 200 nm. Samples isolated by UC showed a spectrum of cup-shaped extracellular vesicles of diverse sizes and shapes (Figure 4A). In contrast, samples isolated with SEC had fewer impurities, although they had a smaller number of EVs compared to those isolated with UC (Figure 4B). Specifically, EVs derived from L. crispatus PMC 201 exhibited a relatively high purity with comparatively uniform shape and size distribution.

3.4. Proteome Profiling of EVs Derived from L. crispatus PMC201

3.4.1. LC-MS/MS Analysis

To identify proteins secreted by EVs derived from L. crispatus PMC201, isolated EV samples (n = 3) were digested in a solution and analyzed using LC-MS/MS with an Orbitrap Eclipse TM Tribrid TM mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA) coupled with an Ultimate TM 3000 UHPLC system (Thermo Fisher Scientific, USA). All identified proteins were statistically processed using the Scaffold program (version 5.3.0) (Figure 1B). Only proteins with a protein threshold ≥ 0.95 and one minimum peptide of two were selected for further analyses. Under these conditions, 64 proteins were identified in EVs derived from L. crispatus PMC201 (Figure 5A). In EVs derived from L. crispatus PMC201, S-layer protein showed the highest abundance at 301 (Figure 5B). This proteomic study identified proteins in EVs derived from L. crispatus PMC201 (Table 3).

3.4.2. Confirmation of Protein Sizes through Coomassie Blue Staining

Through Coomassie blue staining, we assessed protein sizes, revealing a range of sizes with several bands exhibiting distinct visibility. Notably, for EVs derived from L. crispatus PMC201, the most prominent band fell within the 42-49 kDa range (Figure 5F). LC/MS/MS analysis of proteomics data revealed that a 47 kDa S-layer protein was the most prominent band for EVs derived from L. crispatus PMC201 in terms of abundance (Figure 5C, D).

3.5. Prediction of Protein Pathways and Interaction of EVs Derived from L. crispatus PMC201

3.5.1. Cellular Components in Protein from EVs Derived from L. crispatus PMC201

Analysis of EVs derived from L. crispatus PMC201 identified distinct cellular components, including the extracellular region, peptidoglycan-based cell wall, cell wall, and S-layer with strength (Log10(observed / expected)) values ranging from 0.91 to 1.47, indicating moderate to high enrichment effects. Corresponding false discovery rates ranged from 0.0014 to 0.0145 (Table 4).

3.5.2. Analysis of Biological Processes and KEGG Pathways in Proteins from EVs Derived from L. crispatus PMC201

Gene Ontology (GO) analysis was performed to profile protein compositions of EVs and identify key biological processes associated with their cargo (Figure 6A, Table 5). EVs derived from L. crispatus PMC201 proteins were enriched in metabolic processes, emphasizing the significant role of EVs in mediating metabolic activity. EVs derived from L. crispatus PMC201 proteins showed significantly higher counts and strengths in glycolytic process (9 of 11, strength: 1.38) and gluconeogenesis (4 of 7, strength: 1.23). They also showed higher counts and strengths in the pyruvate metabolic process (10 of 12, strength: 1.39). Proteins extracted from EVs derived from L. crispatus PMC201 displayed pathways associated with carbohydrate metabolism, amino acid biosynthesis, and microbial metabolism. EVs derived from L. crispatus PMC201 proteins also showed a higher prevalence of pathways related to biosynthesis of primary and secondary metabolites, such as biosynthesis of secondary metabolites (20 of 100, strength: 0.71) and carbon metabolism (17 of 38, strength: 1.06) (Figure 6B, Table 6).

3.5.3. Investigation into Subcellular Localization of Proteins EVs Derived from L. crispatus PMC201

A substantial fraction of proteins within EVs derived from L. crispatus PMC201, amounting to 46.59%, predominantly localized to the extracellular compartment, with an additional 19.09% exhibiting a distinct presence within the cytoplasm. These results demonstrate the significant participation of EVs derived from L. crispatus PMC201 in extracellular matrix dynamics, underscoring their significance in intercellular signaling and tissue homeostasis (Figure 6C).

3.5.4. Protein Interactions within EVs Derived from L. crispatus PMC201

Interactions among proteins identified in EVs derived from L. crispatus PMC201 revealed a network primarily centered around large ribosomal subunit protein, glyceraldehyde-3-phosphate dehydrogenase(gap), enolase (eno), glucose-6-phosphate isomerase (pgi), phosphoglycerate kinase (pgk), and L-lactate dehydrogenase (ldh) (Figure 7A). These proteins, when considered as the core, are associated with essential biological processes such as pyruvate metabolic process, glycolytic process, generation of precursor metabolites and energy, and organic acid metabolic process.

4. Discussion

EVs derived from L. crispatus PMC201 strain and the proteome they possess were investigated in this study. While EVs derived from probiotic bacteria are significant, proteomic studies on isolated EVs are also being conducted. However, there is a shortage of reports regarding functional characteristics and mechanisms of action of probiotic EVs and proteomes [27]. The function of EVs varies not only depending on the type and origin of bacteria, but also on the quantity of EVs themselves and conditions of bacteria and their culture methods [28]. Therefore, we used food-grade media [24] instead of conventional MRS media. The secretion of EVs particles was more effectively achieved with the food-grade media than with the MRS media. EVs derived from Lactobacillus spp. were associated with immune responses. Treatment with EVs derived from L. plantarum can inhibit the secretion of major inflammatory cytokines and increase the secretion of the anti-inflammatory cytokine IL-10 without causing tissue damage [29]. Additionally, attention has been paid to EVs because they have been found to help drug delivery across the blood-brain barrier (BBB) using both passive and active targeting strategies, promote gene therapy for targeted tissues, and deliver chemotherapy drugs to tumors [30]. These characteristics suggest the possibility of targeting the lungs, which are involved in tuberculosis, for therapeutic purposes. M.tuberculosis is recognized by the C-type lectin pattern recognition receptor (PRR) DC-SIGN (dendritic cell-specific intercellular adhesion molecule 3-capture nonintegrin) [31,32]. DC-SIGN is a transmembrane protein with a C-type calcium-dependent lectin, which can capture antigens for processing and presentation of a viral and bacterial nature. Viruses (such as HIV-1, Ebola, Juniper, and dengue fever) and important bacteria such as M. tuberculosis have previously been shown to be recognized by lectins [33,34,35]. Among many Lactobacillus species reported as probiotics, several possess S-layer proteins that exhibit an antiviral activity against herpesvirus-2 [36]. The S layer of Lactobacillus acidophilus has been demonstrated to possess a protective antiviral effect against Junín virus, an important human pathogen, by blocking virus entry into host cells through the DC-SIGN receptor [37]. Based on these findings, it is conceivable that the S-layer protein of L. crispatus PMC201 might influence the interaction between M.tuberculosis and DC-SIGN, potentially resulting in an anti-tuberculosis effect. When efficacies of specific proteins derived from bacterial EVs are examined, it has been found that glycerol-3-phosphate dehydrogenase (GAPDH) secreted from L. reuteri BBC3 vesicles can reduce intestinal inflammation damage induced by lipopolysaccharides (LPS) [38]. Lipoteichoic acid of L. paracasei D3-5 cell walls can enhance mucin (Muc2) expression and reduce intestinal inflammation by regulating the TLR-2/p38-MAPK/NF-kB pathway [39]. Additionally, GAPDH secreted from L. plantarum EVs has been found to have anti-inflammatory effects against Staphylococcus aureus infection and ability to bind to human colonic mucus [40]. The function of each protein within EVs can be noted and the mechanism of action can be understood through studies on interactions of found proteins. GAPDH, one of the most abundant proteins in L. crispatus PMC201, has previously been proven to be present in L. plantarum EVs with anti-inflammatory effects [40]. It might have an efficacy in treating tuberculosis. Proteins of EVs derived from L. crispatus PMC201 included S-layer, GADPH, and various other proteins such as peptidoglycan-binding protein, N-acetyl muramidase, beta-lactamase family protein, and enolase. Studies on the mechanisms of action and interactions of EVs or proteins from lactic acid bacteria strains are still lacking. Their effects as tuberculosis therapeutics have not been revealed yet. If we look at the efficacy and mechanism of action of each protein, we can predict that we will be able to find a protein that has potential as an adjuvant for tuberculosis treatment. Therefore, further studies are needed to identify characteristics of proteins, find proteins with immunomodulatory mechanisms related to tuberculosis treatment, and identify protein combinations that can induce synergistic effects.

5. Conclusions

This is a follow-up study of Lactobacillus crispatus PMC201, a strain with anti-tuberculosis activity. EVs derived from this strain were confirmed. In addition, experiments were conducted on medium compositions and isolation methods to effectively isolate EVs. This paper optimized culture conditions for characterization and isolation of EVs derived from L. crispatus PMC201, showing yield differences when L. crispatus PMC201 was cultured in commercial media MRS and food-grade media. Food-grade media secreted about 2.5 times more EVs, allowing for more efficient EV isolation. In addition, since the particle number was higher when EVs were isolated by UC than by SEC method, proteomic analysis was performed using food-grade media and UC method. A total of 64 proteins were identified in EVs derived from L. crispatus PMC 201, with s-layer proteins showing the highest abundance. Identified proteins were mainly located in the extracellular space. The analysis confirmed biological processes and protein-protein interactions of proteins identified in EVs derived from L. crispatus PMC 201. In conclusion, our study confirmed that strain L. crispatus PMC201 with anti-tuberculosis efficacy secreted EVs and found a medium that could obtain a high yield of its EVs. We further found that this method was effective for purifying EVs from probiotics, isolating EVs from anti-TB strains, and optimizing culture conditions for therapeutic development. This method could also produce therapeutics for other important bacterial pathogens.