Biophysical Features of Outer Membrane Vesicles (OMVs) from Pathogenic Escherichia coli: Methodological Implications for Reproducible OMV Characterization

Abstract

Background/Objectives: Bacterial outer membrane vesicles (OMVs) play a role in bacterial communication, virulence, antimicrobial resistance, and host–pathogen interaction. OMV isolation is a key step for studying these particles’ functions; nevertheless, isolation procedures can greatly influence the yield, purity, and structural integrity of OMVs, thereby affecting downstream biological analyses and functional interpretation. Methods: In this study, we compared the efficacy of two OMV isolation techniques, differential ultracentrifugation (dUC) and size-exclusion chromatography (SEC), in separating and concentrating vesicles produced by two Escherichia coli strains belonging to uropathogenic (UPEC) and Shiga toxin-producing (STEC) pathotypes. The isolated OMVs were characterized using a multi-analytical approach including transmission and scanning electron microscopy (TEM, SEM), nanoparticle tracking analysis (NTA), dynamic light scattering (DLS), ζ-potential measurement, and protein quantification to assess the purity of the preparations. Results: Samples obtained by dUC exhibited higher total protein content, broader particle size distributions, and more pronounced contamination by non-vesicular material. In contrast, SEC yielded morphologically homogeneous and structurally well-preserved vesicles, higher particle-to-protein ratios, and lower total protein content, reflecting reduced co-isolation of protein aggregates. NTA and DLS analyses revealed polydisperse populations in samples obtained with both isolation methods, with DLS measurements highlighting the contribution of larger or transient aggregates. ζ-potential values were close to neutrality for all samples, consistent with limited electrostatic repulsion and with the aggregation tendencies observed in some preparations. Conclusions: This study describes features of OMV produced by two relevant E. coli strains considering two isolation strategies which exert method- and strain-dependent effects on vesicle properties, including size distribution and surface charge, and emphasizes the trade-offs between yield, purity, and vesicle integrity.

1. Introduction

Outer membrane vesicles (OMVs) are nanosized bilayered particles produced by Gram-negative bacteria, increasingly recognized as mediators of bacterial communication, virulence, and antimicrobial resistance. They transport cargo such as toxins, adhesins, lipopolysaccharides, nucleic acids, and other virulence-associated molecules to other bacterial cells or host tissues, influencing colonization, immune modulation, horizontal gene transfer, and tissue damage [1,2].

Among clinically relevant pathogens, uropathogenic Escherichia coli (UPEC) is a leading cause of urinary tract infections, while Shiga toxin-producing E. coli (STEC) is associated with severe gastrointestinal disease and complications such as Haemolytic Uraemic Syndrome (HUS) [3,4,5,6]. Cross-pathotype strains, displaying virulence features of both extraintestinal pathogenic E. coli and STEC, have been identified as the cause of HUS associated with septicemia [7]. In both pathotypes, OMVs have emerged as critical vectors of virulence, making them promising targets for diagnostics, vaccines, and therapeutic interventions [8,9]. Recent studies have expanded knowledge on OMV cargo diversity and host–pathogen interactions. For instance, OMVs from STEC O157:H7 have been shown to carry multiple virulence factors, including Stx2a, CdtV, hemolysin, and flagellin, which can be internalized by intestinal epithelial and endothelial cells, leading to cellular injury [10]. Other studies have demonstrated that OMVs can translocate across the intestinal epithelial barrier through both paracellular and transcellular pathways, particularly under inflammatory conditions [11]. OMVs can also transport small RNAs. In a recent work, several small RNAs could be identified within OMVs derived from E. coli O26:H11 and O80:H2 strains carrying hlyF, suggesting potential roles in stress adaptation, Shiga toxin regulation, and interbacterial communication [12].

As this evidence highlights the multiple roles of OMVs in bacterial physiology and disease, the need for methodological standardization has become increasingly evident. The broader extracellular vesicle (EV) research community benefits from the Minimal Information for Studies of Extracellular Vesicles (MISEV) guidelines (MISEV2014, MISEV2018, and most recently MISEV2023), which recommend transparent reporting and standardized experimental workflows for EV isolation, characterization, and functional assays [13,14,15]. MISEV2023 update definitions, quality control criteria, and recommendations for EV uptake assays and in vivo studies, emphasizing the need for rigorous distinction between vesicular and non-vesicular particles [15].

Regarding bacterial extracellular vesicles (BEVs), a recent meta-analysis emphasizes a lack of methodological transparency and standardization, showing low “EV-metric” scores and inconsistent reporting of biophysical and biochemical parameters [16]. Consequently, the reproducibility and interpretability of OMV-based biological findings remain limited. Importantly, several studies have provided standardized protocols and methodological guidance specifically for bacterial OMV isolation and characterization, highlighting the importance of reproducibility and comparative evaluation of techniques. For example, Ellis and Kuehn (2010) comprehensively reviewed OMV composition and immunological relevance; Chutkan et al. (2013) and Klimentová and Stulík (2015) described quantitative and qualitative OMV isolation procedures; and Dauros Singorenko et al. (2017) compared different prokaryotic vesicle isolation methods, illustrating method-dependent differences in vesicle purity and yield [17,18,19,20]. These studies underscore the value of methodological rigor while emphasizing that no single approach universally addresses all experimental needs.

Among available isolation methods, differential ultracentrifugation (dUC) and size-exclusion chromatography (SEC) remain the most employed approaches. dUC is often regarded as the traditional standard, but it is time-consuming, prone to vesicle morphological alteration, aggregation, and often results in a higher co-isolation of non-vesicular material (e.g., protein contaminants like hemolysin, flagellin, GroEL) [13,14,15,21,22]. Conversely, SEC provides a faster, more reproducible, and scalable alternative, enabling superior removal of protein and LPS contaminants [22,23,24]. Despite the supposed advantages of the SEC, comparative evaluations of these two OMV isolation methods and their performance remain limited.

Here, we present an integrated physicochemical comparison of particle samples isolated by dUC and SEC from UPEC and STEC strains, with the aim of providing additional information on vesicle yield, purity, structural integrity, and reproducibility of biological replicate preparations, in line with the principles of MISEV2023. Our findings provide methodological insights to guide the selection of OMV isolation approaches and promote standardized and comparable workflows in BEV studies.

2. Results

2.1. Microscopic Characterization of Vesicle Preparations

Transmission and scanning electron microscopy (TEM and SEM) analyses confirmed the isolation of vesicular particles from STEC and UPEC strains using both dUC and SEC methods (Figure 1). Fraction 1 (F1) and Fraction 2 (F2) correspond to the pellets obtained from the first and second ultracentrifugation steps, respectively, as described in Materials and Methods. The purity and composition of the preparations varied between methods and strains. In dUC fractions from STEC, F1 contained abundant non-vesicular material, including protein aggregates and pili-like structures (Figure 1A,B), whereas F2 was enriched in vesicles with minimal contamination (Figure 1C,D). Samples obtained from the UPEC strain by dUC showed greater contamination, with detectable pili-like structures in both F1 and F2 fractions (Figure 1G–J). SEC preparations yielded a higher apparent vesicle density for both strains compared to dUC (Figure 1E,F,K,L). Residual contaminants, particularly pili-like structures, were still detectable in all samples, with UPEC preparations being again the most affected. In all samples, it was possible to observe a dense proteinaceous matrix associated with vesicles, suggesting the co-isolation of extracellular or matrix-bound proteins. In UPEC samples, this matrix appeared more abundant and compact, consistent with previous reports describing the high extracellular protein content and biofilm-associated material typical of UPEC strains [25,26].

Based on the dUC-F2 vesicular features by electron microscopy, such fractions were selected for subsequent comparisons with SEC samples.

2.2. Nanoparticle Tracking Analysis of OMVs

Nanoparticle tracking analysis (NTA) was used to assess the size distribution and concentration of OMVs isolated with dUC-F2 and SEC. Representative NTA profiles and statistics are reported in Figure 2 and

Table 1. Summary of NTA-derived OMV characteristics. Values represent the mean ± standard error of the mean (SeM) for Mean size, Mode size, D10/D50/D90 percentiles, and particle concentration. Span was calculated for each biological replicate as span = (D90 − D10)/D50 and reported as mean ± standard deviation (SD).

Sample	STEC dUC-F2 (nm ± SeM)	STEC SEC (nm ± SeM)	UPEC dUC-F2 (nm ± SeM)	UPEC SEC (nm ± SeM)
Mean	164.8 ± 2.0	120.5 ± 1.4	168.0 ± 2.9	171.0 ± 1.00
Mode	142.1 ± 7.4	95.5 ± 4.1	149.5 ± 7.8	160.1 ± 5.4
SD	50.8 ± 1.2	44.5 ± 1.7	53.0 ± 4.3	48.5 ± 1.5
D10	110.4 ± 1.8	78.7 ± 1.3	117.2 ± 3.3	119.7 ± 1.5
D50	155.6 ± 2.4	108.7 ± 1.3	156.0 ± 2.4	165.2 ± 1.2
D90	231.3 ± 3.9	176.4 ± 4.6	235.4 ± 7.8	225.7 ± 2.5
Span ± SD	0.78 ± 0.10	0.90 ± 0.11	0.74 ± 0.05	0.64 ± 0.09
Concentration (Particles/mL)	(5.11 ± 0.70) × 1010	(6.81 ± 0.37) × 1010	(1.54 ± 0.10) × 1010	(8.46 ± 0.22) × 1010

, respectively, while the complete datasets of all biological replicates are reported in Supplementary Figure S1.

As shown in Table 1, for the STEC strain, the average particle concentration was very similar between the two methods, resulting in (5.11 ± 0.70) × 10¹⁰ particles/mL in the dUC-F2 fraction and (6.81 ± 0.37) × 10¹⁰ particles/mL in the SEC preparation. Statistical analyses did not reveal a significant difference in particle concentration between the two methods (p = 0.307), indicating similar concentration yields for STEC-derived OMVs (Supplementary Figure S4). Whereas the vesicle modal diameter—calculated over the three biological replicates—was 142.1 ± 7.4 nm and 95.5 ± 4.1 nm, respectively. Unpaired t-tests confirmed that the differences in modal diameter were statistically significant (p = 0.0134), whereas F-tests indicated no significant difference in variance (p = 0.56), suggesting that the observed size reduction in SEC samples is consistent across replicates (Supplementary Figure S4). The smaller size observed in SEC samples is consistent with the higher resolving capacity of SEC over dUC (Figure 2; Table 1; Supplementary Figure S1).

For the UPEC strain, the F2 fraction yielded an average of (1.54 ± 0.10) × 10¹⁰ particles/mL, while the SEC-derived sample showed a concentration of (8.46 ± 0.22) × 10¹⁰ particles/mL. Although SEC-derived samples showed higher mean particle concentrations, this difference did not reach statistical significance (p = 0.0778), and no significant difference in variance was observed between methods (p = 0.122). The modal diameters were 149.5 ± 7.8 nm for dUC-F2 and 160.1 ± 5.4 nm for SEC, respectively (Figure 2; Supplementary Figure S1). Statistical comparison by unpaired t-test revealed no significant difference in modal diameter between the methods (p = 0.212), and F-test confirmed similar variance (p = 0.254) (Supplementary Figure S4). In addition to the main peak, NTA profiles showed secondary populations of larger diameter but lower abundance in all preparations, ranging from ~300 to ~700 nm, likely reflecting minor subpopulations or transient aggregates that contribute to the measured size distribution without affecting the modal diameter (Figure 2; Table 1; Supplementary Figures S1).

Overall, both isolation methods yielded vesicle populations within the expected nanometric range for E. coli-derived OMVs [27]. dUC-F2 fractions showed a broader particle size distribution (mean span values ± SD = 0.78 ± 0.10 for STEC and 0.74 ± 0.05 for UPEC), while SEC samples were more uniform for UPEC (0.64 ± 0.09) and slightly more variable for STEC (0.90 ± 0.11), consistent with higher polydispersity or the presence of transient aggregates (Table 1).

Representative NTA recordings for STEC dUC-F2 and UPEC SEC OMVs are available as Supplementary Videos S1 and S2, illustrating vesicle Brownian motion and dispersion patterns. Corresponding still frames are shown in Supplementary Figures S2 and S3.

2.3. Physicochemical Characterization of OMVs by Dynamic Light Scattering (DLS) and Zeta-Potential (ζ)

DLS analysis showed that vesicle size was influenced by both the bacterial strain and the isolation procedure employed. As shown in Supplementary Table S1, the STEC SEC sample yielded a smaller hydrodynamic diameter (265 nm) than the STEC dUC-F2 sample (733 nm). Conversely, for UPEC samples, dUC-F2 resulted in smaller vesicles with an average hydrodynamic diameter of 219 nm, compared to 535 nm obtained with SEC. Statistical analysis of hydrodynamic diameter measurements confirmed that STEC dUC-F2 vesicles were significantly larger than STEC SEC and UPEC dUC-F2 vesicles (p < 0.05, p < 0.01). Moreover, as shown in Figure 3, DLS did not identify a single, uniform population in terms of vesicle size, and the high PDI values observed for most samples indicated polydispersity (Supplementary Table S1). Nevertheless, when considering the average measurements from different isolates of each bacterial strain, DLS consistently revealed a population of vesicles around 100 nm, in agreement with NTA and microscopic observations.

ζ-potential measurements were also performed, and all samples exhibited values close to zero (Figure 3 and Supplementary Table S1). These near-neutral ζ-potential values were insufficient to ensure repulsive interactions in some preparations; thus, the large aggregates observed in the size measurements may have been caused by the tendency of some vesicles to undergo coalescence.

2.4. Protein Quantification of OMV Preparations

Protein content and purity ratio of OMV preparations (dUC-F2 and SEC-derived) from STEC and UPEC strains were assessed using Qubit measurements in combination with NTA particle counts, according to the Webber and Clayton equation, as described in Materials and Methods [28]. This metric was used as a standardized proxy for OMV purity, allowing comparison between methods. Mean values, SD, and calculated particle-to-protein (purity) ratios for each sample are reported in

Table 2. Protein concentration (μg/mL) ± SD, particle concentration (particles/mL) ± SeM, and purity ratio ± SD for OMVs fractions derived from STEC and UPEC strains and isolated with dUC and SEC.

Strain and Fraction	Qubit Protein Concentration (μg/mL) ± SD	NTA Concentration (Particles/mL) ± SeM	Purity Ratio ± SD
STEC dUC_F2	1186 ± 569	(5.11 ± 0.70) × 1010	(4.31 ± 2.07) × 107
STEC SEC	399 ± 252	(6.81 ± 0.37) × 1010	(1.71 ± 1.08) × 108
UPEC dUC_F2	963 ± 527	(1.54 ± 0.10) × 1010	(1.60 ± 0.88) × 107
UPEC SEC	367 ± 216	(8.46 ± 0.22) × 1010	(2.30 ± 1.36) × 108

.

Statistical analysis of independent biological replicates did not reveal significant differences in mean particle-to-protein ratios between dUC-F2 and SEC-derived preparations for either STEC or UPEC strains (unpaired t-test, p > 0.05). The mean particle-to-protein ratios tended to be higher in SEC-derived samples, consistent with the trend toward improved OMV purity. However, variance analysis demonstrated a significantly lower dispersion of particle-to-protein ratios in SEC-derived samples compared to dUC-F2 preparations (F-test, p < 0.01 for both strains), indicating improved batch-to-batch consistency and methodological reproducibility.

3. Discussion

The present study provides an integrated biophysical characterization of the outer membrane vesicles (OMVs) produced by two unrelated pathogenic E. coli (STEC and UPEC strains) and obtained through two widely used isolation methods, dUC and SEC. By combining complementary analytical approaches, our work highlighted how both methodological and biological factors may influence the structural integrity, purity, and size of OMV preparations, thereby affecting quantitative and qualitative parameters commonly used to define OMV preparations, leading to practical considerations for downstream applications.

Electron microscopy confirmed that both dUC-F2 and SEC-derived preparations contained vesicular structures with strain-dependent differences in homogeneity. Despite the presence of impurities, SEC-derived vesicles appeared more uniform in shape and better preserved structurally compared to those isolated by dUC. These qualitative differences are indicative of variable degrees of co-isolated non-vesicular material and are consistent with higher estimated purity observed for SEC-derived samples.

NTA and DLS analyses provided complementary information when applied to the analysis of the OMV preparations obtained with the two isolation methods. Differences in vesicle size were more pronounced for STEC than UPEC, as observed in both NTA and DLS measurements, highlighting the strain-dependent variability of OMV dimensions and confirming that methodological choices can shape biophysical readouts. The apparent discrepancies observed between the measurements should be interpreted considering the intrinsic principles of the two techniques. Although both NTA and DLS provide sphere-equivalent hydrodynamic diameters, NTA reports particle number-based modal diameters by tracking the diffusion of individual particles with minimal intensity weighting, whereas DLS measures the collective motion of a large number of particles simultaneously and, being intensity-weighted, emphasizes the contribution of larger or aggregated particles [29,30]. Therefore, the larger average diameters obtained by DLS may reflect the transient aggregation of vesicles and the contribution of the surrounding proteinaceous material. This effect was particularly evident in UPEC-derived samples, and it is consistent with their higher PDI values. Moreover, the minor populations of larger vesicles detected by NTA (up to 761 nm) are composed of particles with low abundance and do not significantly affect the modal diameter. As regards particle concentration data, no statistically significant differences were observed between dUC-F2 and SEC-derived preparations for either strain. Nevertheless, SEC samples consistently displayed higher mean particle concentrations, particularly for UPEC, indicating a trend toward increased vesicle enrichment without a concomitant increase in variability.

Combining these parameters makes it possible to evaluate OMV preparations more thoroughly. SEC offers clear advantages in terms of processing time, reproducibility, improved batch-to-batch consistency and structural preservation, allowing also a better separation of OMV-enriched from protein-enriched fractions. In contrast, dUC, although requiring long processing time and facilitating the co-isolation of proteins and aggregated material, remains suitable when the goal is to recover broader-sized vesicle populations together with released soluble components [31].

As expected, dUC-F2 captured more heterogeneous populations, spanning different sizes and aggregation states, while SEC provided better size resolution at the potential cost of partially losing larger or more complex vesicular subpopulations [25].

In addition to methodological factors, strain-specific characteristics had a significant impact on vesicle properties. Differences between STEC and UPEC strains in membrane composition, LPS structure, and the regulatory pathways controlling stress responses and vesicle release likely underline the observed variability in vesicle size, polydispersity, aggregation propensity, and the relative performance of the two methods [32,33]. These intrinsic features explain why SEC and dUC produced distinct outcomes depending on the pathotype and highlight the importance of evaluating OMV isolation strategies in a strain-dependent manner.

ζ-potential analysis further characterized the surface charge of vesicles, indicating relatively low absolute values in all preparations and therefore limited electrostatic repulsion among particles. It should be noted that ζ-potential primarily reflects the vesicle surface charge; however, loosely associated protein aggregates or extracellular matrix components may also carry charge and contribute to transient aggregation, without being individually resolved by ζ-potential measurements. Accordingly, ζ-potential values should be interpreted with size distribution and PDI data to better contextualize aggregation phenomena. These results are consistent with the broader DLS profiles observed in some preparations and with the tendency of OMVs to form transient aggregates in aqueous suspension, depending on ionic strength and surface composition.

Protein quantification confirmed lower total protein content and higher particle-to-protein ratios in the SEC-derived fraction, indicating improved purity and reduced co-isolation of soluble proteins. Since the Qubit fluorimeter estimates total protein but cannot distinguish vesicular from soluble protein content, its measurements should be interpreted in combination with particle counts (NTA) to accurately evaluate purity [34].

Overall, both the amount and the purity of vesicle preparations should be interpreted in relation to the intended downstream application. Within this comparative framework, particle concentration primarily reflects isolation yield, whereas size distribution, polydispersity, and ζ-potential inform on vesicle heterogeneity and aggregation state or tendency, which may be relevant for interpreting OMV behavior in biological or ecological contexts. Furthermore, as suggested by previous studies, additional detection techniques might be used to further characterize any extracellular contaminants that may be desirable [16].

This study provides parameters and practical orientations towards the selection of OMV isolation and characterization approaches, contributing to a more consistent comparison of commonly used methodologies. By extending recent observations that OMV isolation procedures can influence vesicle composition and biological readouts, we provide evidence that these effects may also have an impact on OMV biophysical properties in a strain-dependent manner, highlighting the need for integrated physicochemical characterization prior to functional interpretation [35].

Our results are consistent with literature evidence indicating that SEC generally provides better outcomes when purity, structural integrity, and reproducibility are prioritized, such as in cellular or immunological assays evaluating vesicle or protein–host interaction. These findings support the importance of the integration of multiple complementary analytical methods for accurate OMV characterization, but also for the development of standardized workflows adaptable to diverse bacterial strains.

Future studies should be carried out to expand this comparative framework to include functional assays, inter-laboratory validations, and emerging analytical platforms, such as high-resolution flow cytometry, which offer more quantitative and discriminative OMV profiling.

4. Materials and Methods

4.1. Bacterial Strain Selection and Growth Conditions

Two unrelated pathogenic E. coli strains were selected for the study: UPEC strain LC2, serotype O77:H18, isolated from the urine of a woman suffering from recurrent cystitis [36], and STEC strain ED1374, serotype O80:H2, a food isolate from cow milk [12]. Briefly, single colonies (n = 6 for each replicate) were grown overnight at 37 °C in 600 mL of LB broth (Sigma-Aldrich/Merck KGaA, Darmstadt, Germany) in continuous shaking at 120 rpm. The cultures were then subjected to centrifugation at 9800× g for 10 min at 4 °C to remove bacterial cells. The collected supernatant was filtered with a 0.22 μm polyethersulfone (PES) vacuum filter (MilliporeSigma, Burlington, MA, USA) to remove any residual bacteria. The resulting filtration was divided into two 300 mL aliquots to be processed by dUC and SEC, respectively.

4.2. Purification of Bacterial Outer Membrane Vesicles

Three independent culture supernatants of each strain were used for the isolation of OMVs produced by UPEC LC2 and STEC ED1374, processed throughout three independent extractions for each method to produce three consistent biological replicates for each method and for each strain. The experimental design adopted for assessing the yield and purity of the OMVs obtained by dUC and SEC is described in Figure 4.

The dUC protocol for OMVs isolation was used following the procedure described by Rueter and Bielaszewska [37] with some modifications and optimization as described in our previous work [12]. After filtration (0.22 μm), the culture supernatants (~300 mL) were subjected to two sequential ultracentrifugation steps at 100,000 and 200,000× g for two hours each at 4 °C, using Optima XPN ultracentrifuge (Beckman Coulter, Brea, CA, USA) equipped with a Type 45 Ti fixed-angle rotor and 70 mL polycarbonate tubes. The resulting pellets from each step were collected separately, referred to as Fraction 1 (F1) and Fraction 2 (F2), respectively. Each pellet was gently resuspended in a total volume of 900 µL of sterile filtered PBS (pH 7.4; Sigma-Aldrich/Merck KGaA, Darmstadt, Germany) for further analyses.

As for the SEC method, the filtered supernatants (~300 mL) were first concentrated using Amicon Ultra-15 centrifugal filters with a 100 kDa molecular weight cut-off (Sigma-Aldrich/Merck KGaA, Darmstadt, Germany) by centrifuging at 3000–5000× g until the entire supernatant volume was processed and a final volume of 1 mL was obtained. The concentrate was then subjected to size-exclusion chromatography using qEVoriginal 70 nm Gen2 columns (Izon Science Europe Ltd., Oxford, UK), following the manufacturer’s instructions. Based on the column calibration profile, four fractions (numbers 2–5), which corresponded to the vesicle-enriched eluate, were pooled, yielding a total volume of 2.8 mL per preparation.

A sterility assay was performed by plating 10 μL of each OMV preparation, obtained using both isolation methods, onto LB agar plates (Sigma-Aldrich/Merck KGaA, Darmstadt, Germany), then incubated at 37 °C for 48h. All OMV fractions were stored at −80 °C in low-protein-binding microcentrifuge tubes (Thermo Fisher Scientific, Waltham, MA, USA), and freeze–thaw cycles were avoided prior to downstream analyses.

4.3. Scanning and Transmission Electron Microscopy

Morphological characterization of freshly prepared OMV samples was performed by Scanning and Transmission Electron Microscopy (SEM and TEM). For SEM analysis, 50 μL of each OMV suspension was left to adhere to poly-L-lysine-coated round glass coverslips (10 mm, SIAL, Rome, Italy) for five hours at room temperature. According to Barbieri et al. (2025), samples were initially fixed with 2.5% glutaraldehyde (Sigma-Aldrich/Merck KGaA, Darmstadt, Germany) in sodium cacodylate buffer (Sigma-Aldrich/Merck KGaA, Darmstadt, Germany), then post-fixed with 1% osmium tetroxide (Sigma-Aldrich/Merck KGaA, Darmstadt, Germany) [12]. Subsequently, samples were dehydrated by immersion in a graded ethanol (starting from 30% to 100%, Sigma-Aldrich/Merck KGaA, Darmstadt, Germany) followed by hexamethyldisilazane (HMDS, Sigma-Aldrich/Merck KGaA, Darmstadt, Germany) treatment and air-dried for two hours. The dried coverslips were then mounted on aluminum stubs, sputter-coated with a 15 nm chromium layer (Quorum TurboQ ES, Laughton, East Sussex, UK), and analyzed using GeminiSEM 450 (ZEISS, Carl Zeiss AG, Oberkochen, Germany). For TEM analysis, 10 μL of each OMV preparation was deposited onto carbon-coated TEM grids and incubated at room temperature for five minutes. The excess liquid was removed, and the grids were air-dried. A 1:1 solution of 4% ammonium molybdate (pH 6.8) and 2% fosfotungstic acid (AGAR Scientific, Rotherham, UK) was used as a negative stain. The samples were visualized using a PHILIPS EM208 TEM (Eindhoven, The Netherlands) equipped with a Megaview II SIS digital camera (Olympus, Tokyo, Japan).

4.4. Nanoparticles Tracking Analysis

Nanoparticles Tracking Analysis (NTA) was used to measure particle size and concentration of the OMV preparations. In detail, each OMV fraction was analyzed with the Nanosight NTA NS300 instrument, software version 3.4 (Malvern Panalytical, Malvern, Worcestershire, UK) using a Blue488 laser (Cobolt, Gothenburg, Sweden) with the following parameters: 15 camera level, 5 detection threshold, room temperature, water viscosity, 5 video-acquisitions of 60 s. Samples were diluted in sterile PBS (Sigma-Aldrich/Merck KGaA, Darmstadt, Germany) to fall within the optimal particle concentration range recommended by the manufacturer. Mean size (nm), mode size (nm), D10, D50, D90 percentiles, and concentration (particles/mL) were tabulated using for each parameter the mean of five reads and plotted as particle size versus number of particles per mL. The overall size distribution and concentration values were reported as mean ± standard error (SeM), calculated across the three biological replicates per sample. The span of the size distribution was calculated for each replicate using the formula (D90-D10)/D50 and reported as mean ± standard deviation (SD).

4.5. Dynamic Light Scattering and ζ-Potential

The size distribution and ζ-potential were also determined using a Malvern NanoZetaSizer (Malvern Instruments, Worcestershire, UK), equipped with a 5 mW He-Ne laser (λ = 632.8 nm). For Dynamic Light Scattering (DLS) measurements, the scattered light intensity was recorded at a 90° detection angle. The normalized intensity autocorrelation functions were computed using a logarithmic digital correlator and analyzed by the cumulant method to determine the hydrodynamic diameter and the Polydispersity Index (PDI). For ζ-potential determination, the electrophoretic mobility (μ) of the vesicles was measured by dielectrophoretic light scattering (DELS) using the same instrument. The ζ-potential was then calculated from the measured electrophoretic mobility values according to the Smoluchowski equation (ζ = µη/ε), where η and ε represent the viscosity and the permittivity of the solvent, respectively. For the hydrodynamic diameter, PDI, and ζ-potential results, the average values of the mean and the standard deviation (SD) of the three biological replicates per sample were calculated.

4.6. Evaluation of Purity of OMV Preparations

The total protein content of the OMV preparations was quantified with a Qubit 4 fluorometer using the Qubit Protein Assay kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. EVs’ sample purity was estimated by means of the Webber and Clayton’s equation, which defines the particle-to-protein ratio (R) as the number of particles per microgram of protein [28]:

$$\mathrm{R}={\displaystyle \frac{\mathrm{N}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{ }\mathrm{o}\mathrm{f}\mathrm{ }\mathrm{p}\mathrm{a}\mathrm{r}\mathrm{t}\mathrm{i}\mathrm{c}\mathrm{l}\mathrm{e}\mathrm{s}}{\mathrm{A}\mathrm{m}\mathrm{o}\mathrm{u}\mathrm{n}\mathrm{t}\mathrm{ }\mathrm{o}\mathrm{f}\mathrm{ }\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{t}\mathrm{e}\mathrm{i}\mathrm{n}}}$$

For protein concentration results, the average values of the mean and the standard deviation (SD) of the triplicates were calculated.

4.7. Statistical Analyses

Statistical analyses were performed using GraphPad Prism software (version 8.0, San Diego, CA, USA). For comparisons of NTA-derived modal diameters and particle concentrations between isolation methods within each strain, unpaired two-tailed Student’s t-tests were applied. Equality of variances was assessed using F-tests.

For DLS, PDI, and ζ-Potential measurements, statistical analyses were carried out using a two-way ANOVA, followed by Tukey’s post hoc multiple comparisons test to evaluate the effects of strain and isolation method.

Data are presented as mean ± standard deviation (SD), unless otherwise stated. Differences were considered statistically significant at p < 0.05, with p < 0.01 indicating stronger statistical significance.

5. Conclusions

This study provides useful information about OMV from pathogenic E. coli by highlighting how dUC and SEC affect vesicle yield, purity, and structural integrity. Further investigations are needed to deepen the understanding of E. coli OMVs. The integration of biochemical and compositional analyses (e.g., SDS-PAGE, proteomics, lipid or LPS quantification) represents complementary levels of OMV characterization and may be essential for studies aimed at defining vesicle preparations and composition and/or their biological activity.