Results

Antimicrobial Peptides

Introduction

Antimicrobial resistance (AMR) is a growing global health crisis, with antibiotic-resistant infections causing over 1.2 million deaths annually [1]. Pseudomonas aeruginosa, a gram-negative opportunistic pathogen, is particularly concerning due to its intrinsic resistance to multiple antibiotics and its ability to rapidly acquire new resistance mechanisms.

Antimicrobial peptides (AMPs) have emerged as a promising alternative to traditional antibiotics. These naturally occurring molecules are part of the innate immune system of many organisms and exhibit broad-spectrum antimicrobial activity [2,3]. Unlike conventional antibiotics that target specific cellular processes, AMPs primarily act by disrupting bacterial cell membranes, making it more difficult for bacteria to develop resistance [4].

With CAPTURE, we focus on two specific AMPs: Sushi S1 and CONGA-Q7. Sushi S1, derived from horseshoe crab, has shown potent activity against gram-negative bacteria [3]. CONGA-Q7, a synthetic AMP designed using machine learning algorithms, demonstrates broad-spectrum antimicrobial activity [5]. Our goal is to design a plasmid specific for Pseudomonas that can express these AMPs effectively. By engineering bacteria to produce their own AMPs, we aim to create a self-sustaining antimicrobial system that could potentially overcome the limitations of external AMP application, such as rapid degradation and high production costs [6].

To develop an effective AMP-based system against P. aeruginosa, we first needed to characterize and optimize individual components of our design. We began by using E. coli as a model organism due to its rapid growth and ease of genetic manipulation. This approach allowed us to fine-tune our system before transitioning to the more challenging P. aeruginosa.

Our initial experiments focused on designing and testing an inducible expression system for AMPs using the pET22 plasmid, which allows controlled expression via IPTG induction. Learn more on our Plasmid Design Page.

Signal Peptide

Key findings:

• The extracellular signal peptide HSTII fused to Sushi has the highest growth inhibition on E. coli BL21(DE3)
• pET-Plasmid is not suitable as a negative control for the induced expression experiments

Aim:

Our model AMP Sushi is known to disrupt bacterial membranes when applied externally. The first critical question we sought to answer was the optimal cellular location for AMP expression: Would the AMPs be effective when expressed inside the bacterial cell, or would they need to be secreted to the extracellular space? To address this, we designed three constructs with various signal peptides:

• Sushi without a signal peptide for intracellular expression (BBa_K5057004)
• Sushi fused to the pelB signal peptide for periplasmic localization, and (BBa_K5057013)
• Sushi linked to the Heat Stable Toxin II (HSTII) signal peptide for extracellular secretion (BBa_K5057007).

The pelB signal peptide was chosen to target the periplasm, as gram-negative bacteria have two membranes, and we wanted to explore if periplasmic localization would affect Sushi’s membrane-disrupting ability. The HSTII signal peptide was selected to direct Sushi to the extracellular space, mimicking its natural mode of action.

Experimental Setup:

We tested the signal peptide sequences HSTII and pelB linked to Sushi and Sushi without a signal peptide in the pET-22(b)+-Vector in E. coli BL21(DE3). As a control, uninduced cultures and E. coli BL21(DE3) containing the empty pET-22b(+) vector (pET-Plasmid) were used. For each construct, we inoculated 100 mL of LB medium supplemented with ampicillin (LB-Amp) using overnight cultures. Cultures were grown at 37°C with shaking at 200 rpm until they reached mid-log phase. At an OD₆₀₀ (optical density at 600 nm) of 0.6, protein expression was induced by adding 0.5 mM IPTG (Isopropyl-β-D-thiogalactopyranoside) and cultures were incubated at 30°C with continuous shaking at 200 rpm. To monitor bacterial growth, we measured OD₆₀₀ hourly for six hours post-induction. Simultaneously, we assessed cell viability using the drop plate method. At each time point, cultures were serially diluted (from 10^-1 to 10^-5) and spotted on LB-Amp plates. After an overnight incubation at 37°C, cell viability was determined.

Results:

Our experiments revealed significant differences in bacterial growth inhibition among the various Sushi constructs and controls (Figure 2). In all cases, induced cultures exhibited higher growth inhibition compared to uninduced cultures.

Growth Inhibition in Liquid Cultures:

When comparing the effect of Sushi at different cellular localizations, we observed the strongest growth inhibition with HSTII-Sushi (secreted), followed by pelB-Sushi (periplasm) and Sushi (intracellular). Notably, cultures expressing HSTII-Sushi did not surpass an OD₆₀₀ of 0.9 post-induction. Unexpectedly, the control with the empty pET-22b(+) vector also exhibited growth inhibition compared to that of pelB-Sushi and Sushi without signal peptide. This observation suggests potential metabolic burden or toxicity associated with the plasmid induction, independent of Sushi expression.

Viability Assessment via Drop Count Assay:

The growth inhibition could not only be observed in liquid cultures but also for the drop count assay performed 4 hours post-induction (Figure 3). The assay revealed distinct differences in colony-forming ability among the constructs: For pET-pelB-Sushi colonies were observed up to the dilution of 10^-3, indicating moderate growth inhibition. Colonies for pET-Sushi were only visible at 10^-1 dilution, suggesting stronger growth inhibition compared to pelB-Sushi. Despite showing the lowest OD₆₀₀ in liquid culture, pET-HSTII-Sushi, unexpectedly produced more colonies than pET-Sushi. The empty pET Plasmid induced control culture showed similar colony formation to pET-HSTII-Sushi.

Conclusion:

Our experiments revealed complex dynamics in the antimicrobial activity of Sushi peptide with different localization signals. the extracellular signal peptide HSTII, when fused to Sushi, achieved the lowest survival rate of E. coli BL21(DE3) in liquid culture. This finding aligns with previous studies indicating that Sushi exhibits enhanced antimicrobial activity when applied externally to bacterial cells [3].

Based on these results, we decided on AMPs fused to HSTII in the next experiments, while other cellular localizations were provisionally left aside.

An unexpected observation was the growth inhibition in E. coli cultures induced with the empty pET plasmid. We hypothesize that this effect may be due to the expression of an unknown protein or metabolic burden associated with plasmid induction. This finding underscores the importance of carefully selecting appropriate controls in expression studies. To address this issue, we have replaced the empty pET plasmid with pET-mCherry as our negative control.

HSTII-mCherry Induction Conditions

Key findings:

• The optimal conditions for induction of pET-HSTII-mCherry (BBa_K5057009) in E. coli BL21(DE3) with pRARE2 LysS plasmid: 18°C, 0.1 mM IPTG, 24 hours
• Fluorescence detected in supernatant suggested successful export of mCherry (BBa_K5057003), mediated by the HSTII signal peptide (BBa_K5057002)

Aim:

As genes expressed in trans can have different effects on bacteria, we wanted to optimize the expression conditions of HSTII-mCherry regulated by the T7-promoter in E. coli. We aimed to determine the optimal IPTG concentration and temperature for induction.

Experimental setup:

Initial tests with E. coli BL21(DE3) showed high basal expression of HSTII-mCherry without induction (data not shown), making it difficult to properly characterize. Therefore, we chose to continue with E. coli BL21(DE3) containing a pRARE2 LysS plasmid. This plasmid encodes a lysozyme which binds to the T7 polymerase leading to inhibited basal expression, helping us with the characterization of HSTII-mCherry.

E. coli BL21(DE3) pRARE2 LysS cells were transformed with pET-HSTII-mCherry. The cultures were incubated in 700 mL MH- medium containing ampicillin and chloramphenicol (MH-Amp-Cm) at 37°C to an OD₆₀₀ of 0.6. Four different concentrations of IPTG (0.1 mM, 0.5 mM, 0.7 mM, 1.0 mM) were tested at three different temperatures (18°C, 25°C, 37°C) in 20 mL of culture. As a control, an uninduced culture was used.

Expression of mCherry was determined by fluorescence measurements of pellets and supernatants. After 4, 12 or 24 hours, depending on the temperature, 3x 1 mL per culture was pelleted and diluted to an OD₆₀₀ of 1.0. The pellets were washed two times with phosphate-buffered saline (PBS). The supernatants (SN) were directly used for fluorescent measurement. Fluorescent measurements were carried out as technical duplicates using a plate reader at an excitation of 570 nm.

See Notebook Protocol

Results:

Induced samples consistently showed higher fluorescence compared to uninduced samples, indicating successful induction in most conditions. The low fluorescence in uninduced cultures indicates minimal basal expression from the T7 promoter. However, since the difference between uninduced and induced cultures is notable, therefore IPTG can be used for inducible expression reliably. All supernatants showed a fluorescence intensity which is approximately five-fold lower compared to pellets, which is due to the fact that the proteins are less concentrated in the supernatant in comparison to the pellet.

Temperature had a significant impact on induction efficacy. The fluorescent signal was at its lowest at 37°C, suggesting minimal or no induction of mCherry expression at this temperature. For cultures at 30°C, no significant difference between different IPTG concentrations were observed. The fluorescent signal in the pellet was comparable to the uninduced sample, while the signal of the induced supernatant was slightly increased.

The overall highest fluorescence in both pellet and supernatant compared to uninduced control was observed for 18°C samples after 24 hours. Induction with 0.1 mM IPTG yielded higher fluorescent signals compared to other concentrations.

Conclusion:

The results demonstrate successful optimization of HSTII-mCherry expression in E. coli BL21(DE3) pRARE2 LysS cells. The optimal conditions for the expression of HSTII-mCherry were determined with 0.1 mM IPTG at 18°C for 24 hours.

The fluorescence detected in the supernatant, particularly at 18°C and 30°C, suggests successful export of mCherry, mediated by the HSTII signal peptide. With these optimized conditions we can further characterize our AMPs fused to HSTII.

HSTII Function

Key findings:

• HSTII successfully transports mCherry to extracellular localization

Aim:

We wanted to verify the functionality of HSTII as an extracellular signal sequence with mCherry as a reporter protein.

Experimental Setup:

E. coli BL21(DE3) with pRARE2 LysS plasmid were transformed with the pET-HSTII-mCherry and the pET-mCherry, respectively. The cultures were diluted to an OD₆₀₀ of 0.1 and incubated in 30 mL MH-medium with ampicillin and chloramphenicol at 37°C to an OD₆₀₀ of 0.6. The cultures were induced with 0.1 mM IPTG. After 24 hours at 18°C and 200 rpm continuous shaking, 2x 1 mL of the culture was pelleted and diluted to an OD₆₀₀ of 1.0. The supernatants (SN) were stored for fluorescent measurement. The pellets were washed two times with phosphate-buffered saline (PBS). The pellets and the supernatants were used for fluorescent measurement using a plate reader with an excitation of 570 nm and an emission of 610 nm.

Results:

The pellets of cultures expressing mCherry or HSTII-mCherry showed the comparable level of fluorescence (Figure 7). Supernatants from HSTII-mCherry cultures exhibited higher fluorescence compared to those from mCherry cultures. The overall fluorescence in supernatants was lower than in pellets for both constructs, consistent with lower protein concentration in the extracellular medium.

Conclusion:

The follow-up comparative study between HSTII-mCherry and mCherry alone provided strong evidence for the functionality of the HSTII signal peptide. The higher fluorescence signal in the supernatant of HSTII-mCherry cultures compared to mCherry cultures suggests successful export of the fusion protein mediated by the HSTII signal peptide. The comparable fluorescence levels in the pellets indicate that the presence of the HSTII signal does not significantly affect overall mCherry expression or folding.

These results validate our approach of using HSTII as a signal peptide for extracellular localization of our target proteins. With the optimized conditions we can further characterize our AMPs fused to HSTII.

The overall fluorescence of the supernatants is lower in contrast to the pellets due to the amount of cells used for measurement.

Evaluation of the Antimicrobial Activity of Synthetic Sushi and CONGA-Q7

Key findings:

• Synthetic Sushi showed no detectable bactericidal effect on E. coli BL21(DE3) in liquid culture or on agar plates.
• Positive controls (kanamycin and ampicillin) demonstrated expected growth inhibition
• D-CONGA-Q7 (BBa_K5057005), another AMP tested, showed some inhibition of E. coli and P. fluorescens growth
• L-CONGA-Q7 showed partial growth inhibition of E. coli but no detectable effect on P. fluorescens at 40 µM

Aim:

To evaluate the antimicrobial activity of synthetic Sushi peptide against E. coli BL21(DE3), we compare its efficacy to established antibiotics and another AMP. We chose the synthetic AMP called d-amino acid CONsensus with Glycine Absent (D-CONGA-Q7) developed by William C Wimley [5]. In literature, CONGA-Q7 has been only synthesized and tested using D-amino acids. We aim to evaluate the expression of the L-amino acid version of CONGA-Q7 (read more on our Plasmid Design). Therefore we examined D-CONGA-Q7 as well as L-CONGA-Q7 against E. coli BL21(DE3) and P. fluorescens DSM 50090. This experiment is crucial for validating our approach of using Sushi as an antimicrobial agent in CAPTURE.

Experimental Setup:

Liquid Culture Assay - Sushi

An overnight culture of E. coli BL21(DE3) was grown in MH-medium to an OD₆₀₀ of 0.3. 500 µL of bacterial culture were incubated with Sushi (20 µM) at 37°C and 800 rpm shaking for 2 hours. As a positive control, kanamycin (40 µg/mL) was used. As a negative control untreated E. coli BL21(DE3) were used. Bacterial viability was determined by drop count assay.

Liquid Culture Assay - CONGA-Q7

500 µL of bacterial culture were incubated with D- or L-CONGA-Q7 (40 µM) at 37 °C or 30°C respectively and 800 rpm shaking for 3 hours. As a positive control kanamycin was used at 40 µg/mL for E. coli and 80 µg/mL for P. fluorescens. As a negative control 0,025% acetic acid, the solvent of D- and L-CONGA-Q7, and bacteria without treatment were used. Every hour a 7x 10-fold serial dilution was plated on MH-Agar using the Drop-Count method.

Agar Plate Assay

E. coli BL21(DE3) were grown overnight in MH-medium. 50 µL of the cultures were plated on a MH-plate. Drops of 10 µL Sushi, ampicillin (100 µg/ml, positive control) or water (negative control) were applied. D-CONGA-Q7 was included for comparison. The plate was incubated at 37°C overnight.

Results:

Liquid culture assay - Sushi

The control of non-treated E. coli BL21(DE3) showed normal growth. The positive control kanamycin showed total growth inhibition of E. coli. The incubation with synthesized Sushi showed no detectable bactericidal effect.

Liquid culture assay - CONGA-Q7

Untreated E. coli exhibited normal growth patterns throughout the experiment. Acetic acid (0.025%, solvent control) had no significant impact on bacterial growth compared to untreated controls. kanamycin treatment demonstrated complete growth inhibition, with the exception of minimal colony formation in the first three dilutions at the 2-hour time point.

D-CONGA-Q7 treatment resulted in complete elimination of viable bacteria from the first hour onwards. No colonies were observed at any dilution or timepoint, indicating potent bactericidal activity.

L-CONGA-Q7 treatment exhibited a less pronounced but still significant antimicrobial effect. Reduced colony-forming units compared to untreated controls, indicating partial growth inhibition. The effect was notably less potent than that of D-CONGA-Q7.

CONGA-Q7 was also tested on P. fluorescens DSM 50090. Here treatment with D-CONGA-Q7 also led to no bacterial growth on the plates for all three hours. L-CONGA-Q7 had no visible effect on the growth of P. fluorescens. The positive control kanamycin worked since it effectively inhibited all bacterial growth. The negative controls with acetic acid and without treatment showed similar colony forming units.

Agar plate assay

On the agar plate assay, the synthetic Sushi produced no visible growth inhibition zone, while ampicillin created a clear inhibition zone around the applied drop. D-CONGA-Q7 demonstrated some growth inhibition, but less than ampicillin.

Conclusion:

Contrary to our expectations based on literature [3], synthetic Sushi did not exhibit antimicrobial activity against E. coli BL21(DE3) in either liquid or agar plate assays. This lack of activity was observed despite using concentrations similar to those reported in previous studies. Several factors could explain these unexpected results:

• Potential degradation of synthetic peptide during shipping or storage.
• Possible issues with peptide solubility.
• Strain-specific resistance of E. coli BL21(DE3).

The positive results obtained with kanamycin, ampicillin, and D-CONGA-Q7 confirm the validity of our experimental setup and suggest that the lack of activity is specific to our synthetic Sushi.

Our results demonstrate the potent antimicrobial activity of synthetic D-CONGA-Q7 against E. coli BL21(DE3) and P. fluorescens DSM 50090, with efficiency comparable to kanamycin positive control. Synthetic D-CONGA-Q7 treatment showed a negative effect on growth of E. coli, comparable to that of the kanamycin positive control. Although L-CONGA-Q7 had a reduced effect on bacterial growth, it is a promising candidate for expression. L-CONGA-Q7 showed no visible growth inhibition on P. fluorescens, indicating the need for further investigations. We will therefore test higher concentrations of L-CONGA-Q7 on P. fluorescens.

In the next step we want to proceed with our original plan to express Sushi peptides in bacteria, which could provide a more reliable and effective approach for antimicrobial activity. By bypassing the need for external peptide application, we could overcome potential issues related to peptide stability, delivery, or bacterial uptake.

Expression and Activity of HSTII-Sushi in E. coli

Key findings:

• HSTII-Sushi at 18°C in E. coli BL21(DE3) with pRARE2 LysS at 18°C does not inhibit bacterial growth
• HSTII-Sushi expression at 30°C effectively inhibits growth of E. coli BL21(DE3) for 6 hours post-induction
• IPTG concentrations between 0.1 mM and 1.0 mM show similar effects on HSTII-Sushi expression and bacterial growth inhibition at 30°C
• E. coli BL21(DE3) expressing HSTII-Sushi recovers growth after 24 hours

Aim:

To evaluate the optimal expression conditions of HSTII-Sushi and its antimicrobial activity compared to HSTII-mCherry in E. coli. We aimed to determine the most effective temperature and IPTG concentration for HSTII-Sushi expression and bacterial growth inhibition.

HSTII-Sushi Expression at 18°C:

Based on our previous experiments of HSTII-mCherry expression we decided to use these conditions (18°C, 0.1 mM IPTG) as a starting point to evaluate the expression of HSTII-Sushi in E. coli BL21(DE3) pRARE LysS cells.

Experimental Setup:

Overnight cultures of pET-HSTII-Sushi and pET-HSTII-mCherry were pelleted, diluted to an OD₆₀₀ of 0.1 and then grown in MH-medium with ampicillin and chloramphenicol to an OD₆₀₀ of 0.6 at 37°C. Cultures were induced with 0.1 mM IPTG and incubated at 18°C, 200 rpm. OD₆₀₀ was measured hourly for eight hours.

Results:

All cultures (induced and uninduced) showed similar growth patterns over 8 hours (Figure 15). There was no notable effect on E. coli growth upon induction of HSTII-Sushi.

Conclusion:

The conditions that worked best for HSTII-mCherry (18°C, 0.1 mM IPTG) were not effective for HSTII-Sushi expression or activity in E. coli BL21(DE3) pRARE2 LysS cells. With Western-Blot analysis we did not detect any Sushi peptide (data not shown). We concluded that under these conditions, Sushi is either not expressed or in too low concentrations. This led us to explore alternative expression conditions.

HSTII-Sushi Expression at 30°C

Following our initial experiments at 18°C, which failed to demonstrate the expected growth inhibition of E. coli BL21(DE3) expressing HSTII-Sushi, we hypothesized that the optimal expression conditions might differ between HSTII-mCherry and HSTII-Sushi. This led us to investigate the effect of higher temperatures on HSTII-Sushi expression and activity. In preliminary studies, we observed significant growth inhibition of E. coli BL21(DE3) cells expressing HSTII-Sushi when cultured at 30°C in LB medium (Figure 16). To maintain consistency with our other experiments and to rule out media-specific effects, we decided to repeat this experiment using MH medium, which we had used in our previous experiments.

Experimental Setup:

We used the same experimental set up as before, but changed the induction condition to 0.5 mM IPTG at 30°C using E. coli BL21(DE3) cells in MH medium. Additionally we performed the experiment with pET-Sushi-6xHis and took 1 mL samples of the culture after 1,2,4,6,8 and 24 hours post-induction. After centrifugation, the pellets were used to perform a Western-Blot using anti-His-antibodies.

Results:

Induced pET-HSTII-Sushi cultures showed a significant growth inhibition, while uninduced HSTII-Sushi and mCherry cultures achieved an OD₆₀₀ of 6.18. Induced mCherry cultures showed slower but continuous growth, in contrast to induced HSTII-Sushi cultures, which stagnated at OD₆₀₀ of approx. 0.7 at 8 hours post-induction.

Western blot analysis was performed to detect the expression of HSTII-Sushi-6xHis (BBa_K5057012) containing a C-terminal 6xHis-tag for identification. Detectable bands at approximately 5 kDa were observed at 1, 2, 4 and slightly 8 hours post-induction (Figure 18). Notably, no bands were visible in samples collected at 6 and 24 hours post-induction as well as for the uninduced culture with HSTII-Sushi-6xHis. The protein is potentially degraded over time or forms complexes. The control HSTII-mCherry showed detectable bands approx. the size of the expected 30 kDa.

Conclusion:

This experiment demonstrates that HSTII-Sushi expression in E. coli BL21(DE3) at 30°C effectively inhibits bacterial growth. The expression of HSTII-Sushi-6xHis was detected on a Western Blot.

Optimization of IPTG Concentration for HSTII-Sushi Expression

From our previous experiments we have learned that optimal expression conditions are protein-specific and cannot be universally applied across different proteins. By systematically testing a range of IPTG concentrations while maintaining a constant temperature of 30°C, we seek to optimize the expression conditions for HSTII-Sushi .

Experimental Setup:

Overnight cultures of pET-HSTII-Sushi and pET-HSTII-mCherry were pelleted and then grown in MH-Amp to an OD₆₀₀ of 0.6. pET-HSTI-Sushi was induced with 0.1 mM, 0.5 mM, 0.7 mM and 1.0 mM IPTG. pET-HSTII-mCherry (control) was induced with 0.5 mM in a culture volume of 50 mL. For both constructs an uninduced culture was grown at 30°C. OD₆₀₀ was measured every hour.

Results:

All induced HSTII-Sushi cultures showed growth inhibition within the first hour post-induction (Figure 20). In the first six hours the cultures showed no growth above an OD₆₀₀ of 0.9. After six hours, the cultures began to recover. After 24 hours, the induced HSTII-Sushi cultures reached similar OD₆₀₀ values as the controls.

To assess the impact of IPTG concentration on HSTII-Sushi expression and bacterial growth inhibition, we focused our analysis on the first eight hours post-induction. During this period, all induced cultures showed significant growth inhibition compared to the control cultures (Figure 21). Notably, we observed no substantial differences in growth patterns among the various IPTG concentrations tested (0.1 mM, 0.5 mM, 0.7 mM, and 1.0 mM). This suggests that within the range of concentrations examined, the induction of HSTII-Sushi expression and its subsequent antimicrobial effect are not highly sensitive to IPTG concentration.

Conclusion:

Our experiments demonstrate that HSTII-Sushi expression in E. coli BL21(DE3) at 30°C effectively inhibits bacterial growth. The antimicrobial effect is most pronounced in the first 6 hours post-induction, indicating a potent but time-limited activity of the expressed HSTII-Sushi. The recovery of bacterial growth after 24 hours could be due to plasmid loss following ampicillin degradation. The similar results observed across different IPTG concentrations (0.1-1.0 mM) indicate that HSTII-Sushi expression and activity are not highly sensitive to inducer concentration within this range. This robustness allows flexibility in experimental design and potential applications.

Overall Conclusion:

Our series of experiments demonstrate that the expression and activity of HSTII-Sushi in E. coli is highly dependent on temperature and strain characteristics.

Comparison AMPs: Sushi, CONGA-Q7, Imitate

Key findings:

• Expressed CONGA-Q7 demonstrates potent antimicrobial activity comparable to Sushi against E. coli BL21(DE3)
• Imitate (BBa_K5057006), designed as a non-functional control, unexpectedly inhibited the growth of E. coli similar to Sushi and CONGA-Q7
• Sushi and Imitate were successfully detected on a Western Blot

Aim:

After successfully characterizing Sushi, the next step was to compare the antimicrobial efficacy of expressed Sushi with our other AMP CONGA-Q7. In order to work efficiently the minimal inhibitory concentration of the AMP needs to be very low, therefore it is important to screen several peptides. Additionally, we aim to assess the activity of Imitate, a designed peptide intended to be non-functional.

Experimental Setup:

Overnight cultures of pET-HSTII-Sushi, pET-HSTII-Imitate (BBa_K5057015), pET-HSTII-CONGA-Q7 (BBa_K5057008) and pET-HSTII-mCherry (control) were pelleted and then grown in MH-Amp to an OD₆₀₀ of 0.6. Cultures were induced with 0.5 mM IPTG and incubated at 30°C, 200 rpm. OD₆₀₀ was measured hourly for seven hours. Samples (1 mL) were collected every hour and after 24 hours for Western Blot analysis.

Results:

Cultures expressing HSTII-Sushi exhibited consistent growth inhibition after induction as observed in previous experiments (Figure 19). HSTII-CONGA-Q7 showed similar growth inhibition to Sushi for the first hours and started to regrow after five hours to an OD₆₀₀ of 2.0 at 7 hours.

Unexpectedly, HSTII-Imitate, designed as a non-functional control, displayed growth inhibition comparable to both Sushi and CONGA-Q7. All induced cultures maintained OD₆₀₀ below two for 7 hours post-induction.

Uninduced cultures and mCherry controls showed normal growth.

In a direct comparison of cultures from HSTII-Sushi, HSTII-CONGA-Q7 and HSTII-Imitate, no significant difference was observed (Figure 20). Sushi, CONGA-Q7 and Imitate showed growth inhibition in comparison to uninduced cultures.

Western blot analysis was performed to detect the expression of HSTII-Sushi-6xHis and HSTII-Imitate-6xHis , both containing a C-terminal 6xHis-tag for identification. For HSTII-Sushi-6xHis, detectable bands at approximately 5 kDa were observed at 2, 3 and slightly 4 hours post-induction (Figure 25). Notably, no bands were visible in samples collected at 8 and 23 hours post-induction, suggesting potential protein degradation or complex formation over time.

Distinct bands, similar in size to HSTII-Sushi, were detected for HSTII-Imitate at 2 and 4 hours post-induction. This confirms successful expression of the designed peptide and indicates a comparable molecular weight to HSTII-Sushi.

HSTII-mCherry was only detectable in the pellet at 23 hours after induction, not in the supernatant (Figure 26).

Conclusion:

HSTII-CONGA-Q7 demonstrated bactericidal effects comparable to HSTII-Sushi against E. coli BL21(DE3). This result establishes HSTII-CONGA-Q7 as a promising alternative candidate for our antimicrobial peptide expression system.

Contrary to our intentions, HSTII-Imitate, designed as a non-functional control, exhibited significant growth inhibition of E. coli BL21(DE3), similar to both Sushi and CONGA-Q7. This unexpected outcome suggests that the antimicrobial activity of these peptides may be more dependent on their physicochemical properties rather than their specific amino acid sequences.

Effect AMPs in Supernatant

Key findings:

• Supernatants from E. coli BL21(DE3) expressing HSTII-Sushi, -CONGA-Q7, -Imitate and -Sushi-6xHis showed no detectable growth inhibition against plated E. coli BL21(DE3) or P. fluorescens DSM 50900

Aim:

To determine whether the expressed AMPs exert their effects solely on the producing bacteria or if secreted AMPs in the supernatant can inhibit growth of surrounding bacteria.

Experimental Setup:

Overnight cultures of E. coli BL21(DE3) and P. fluorescens DSM 50090 were plated on MH-Agar and incubated for three hours at 37°C for E. coli or 30°C for P. fluorescens. Overnight cultures of E. coli BL21(DE3) expressing pET-HSTII-Sushi, pET-HSTII-CONGA-Q7 and pET-HSTII-Imitate, respectively, were grown in MH-Amp to an OD₆₀₀ of 0.6 at 30°C and then induced with 0.5 mM IPTG. After four hours of incubation at 30°C and 200 rpm, 1 mL of the culture was pelleted. 5 µL of the supernatant was applied on plates with E. coli or P. fluorescens. As a negative control 5 µL of MH-medium was applied to the plates, as a positive control 5 µL kanamycin (40 µg/ml) was used. The plates were incubated at 37°C (E. coli) and 30°C (P. fluorescens) for 8 hours.

Results:

No visible growth inhibition zones was observed on E. coli BL21(DE3) or P. fluorescens DSM 50900 plates for any of the tested AMPs (HSTII-Sushi, -CONGA-Q7, -Imitate, and -Sushi-6xHis).

Conclusion:

The supernatants from E. coli BL21(DE3) expressing various HSTII-fused AMPs (Sushi, CONGA-Q7, Imitate, and Sushi-6xHis) did not demonstrate detectable growth inhibition against plated E. coli BL21(DE3) or P. fluorescens DSM 50900. This suggests that the concentration of secreted AMPs in the supernatant is insufficient to inhibit bacterial growth under these conditions. The AMPs may also not be effectively secreted into the supernatant. Moreover, AMPs may be unstable in the extracellular environment, leading to fast degradation.

The effectiveness of kanamycin in inhibiting growth confirms the validity of our experimental setup.

These results indicate that the antimicrobial activity of our expressed peptides may be primarily intracellular or membrane-associated, rather than secreted. Further investigation into the localization and concentration of these AMPs will be necessary to fully understand their mode of action and potential applications in our CAPTURE system.

Effect of Purified AMPs from Supernatant

Key findings:

• HSTII -Sushi-6xHis and HSTII-Imitate-6xHis peptides purified from 1 L culture of E. coli BL21(DE3) via affinity chromatography showed no growth inhibition against plated E. coli BL21(DE3) or P. fluorescens DSM 50900

Aim:

This experiment was performed to evaluate the antimicrobial activity of purified and concentrated AMPs against surrounding bacteria, addressing the possibility that the previous lack of activity in supernatants was due to insufficient peptide concentration.

Experimental Setup:

20 µL overnight Cultures of E. coli BL21(DE3) and P. fluorescens DSM 50090 were plated on MH-Agar and incubated for three hours at 37°C for E. coli and 30°C for P. fluorescens. In parallel overnight cultures of E. coli BL21(DE3) containing the plasmids pET-HSTII-Sushi-6xHis and pET-HSTII-Imitate-6xHis were grown in 1 L MH-Amp to an OD₆₀₀ of 0.3 at 25°C and then induced with 1 mM IPTG. After 18 hours of incubation at 25°C and 200 rpm, the cultures were pelleted, resuspended in 30 mL buffer (50 mM Tris, 800 mM NaCl, 20 mM Imidazole, 10% glycerol), lysed by sonication for 10 min at 80% power, pelleted for 1 hour. Clarified lysate purified by affinity chromatography. 500 µL of 1 mL eluate were additionally concentrated for 2 hours. 5 µL of the purified peptide was applied on plates with E. coli or P. fluorescens. As a negative control 5 µL of MH-media were applied to the plates, as a positive control 5 µL kanamycin (40 µg/ml). The plates were incubated at 37°C (E. coli) and 30°C (P. fluorescens) for 8 hours.

Results:

Neither the purified HSTII-Sushi-6xHis nor HSTII-Imitate-6xHis peptides produced visible growth inhibition zones on E. coli BL21(DE3) or P. fluorescens DSM 50900 plates (Figure 29). This lack of activity was observed for both the initial purified peptides and their concentrated forms. In contrast, the positive control (kanamycin) effectively inhibited growth of both bacterial strains, confirming the validity of the assay.

Conclusion:

The purified peptides did not show growth inhibition of plated E. coli BL21(DE3) or P. fluorescens DSM 50900, while kanamycin effectively inhibited bacterial growth. This result, combined with our previous findings using unpurified supernatants, suggests that:

• The expressed peptides may lack intrinsic antimicrobial activity in their current form
• The peptides may require specific conditions or cofactors not present in our assay to show antimicrobial effects
• The concentration or stability of the purified peptides may still be insufficient for observable activity

Cytotoxicity

Key findings:

• Preliminary results indicate minimal cytotoxic effect of AMPs on A549 lung epithelial cells.
• All tested AMPs (Sushi, D-GONGA, L-GONGA) showed comparable or lower cytotoxicity than the negative control.

Aim:

To quantify and evaluate the cytotoxic effect of synthesized AMPs (Sushi S1, D-CONGA-Q7, L-CONGA-Q7) on A549 lung epithelial cells using SYTOX Green assay.

Experimental Setup:

The experimental procedure was adapted from Ghimire et al. 2023 [5] to assess AMP-induced membrane damage in A549 cells. The human cell line A549 (alveolar basal epithelial cells) were grown in T-25 flasks in Dulbecco’s modified Eagle’s Medium (DMEM) with 10% fetal bovine serum (FBS) at 37°C, 5% CO₂ see cell culture protocol.

The day before the cytotoxicity experiment, 10⁴ cells/well were plated in a 96-well tissue-culture plate (see Figure 30).

On the day of the assay, the media was replaced with AMP samples prepared in Opti-MEM. Opti-MEM alone and 100% water were used as negative and positive control, respectively. The AMP solvent varies, therefore, samples with dimethyl sulfoxide (DMSO), water and 0.025% acetic acid were used as solvent controls. Cells were treated with the highest concentration of AMP implemented in the experiments of PBS washing. The samples were prepared according to the scheme in figure 30.

The cells are treated with the compound and 0.5 µM SYTOX Green was added at the same time. An endpoint measurement after 60 min incubation at 37°C is performed with an excitation wavelength of 483 nm and an emission wavelength of 523 nm. The experiment was performed in technical duplicates.

The two values of each sample were averaged and the error bars represent the technical error.

Results:

SYTOX Green specifically stains cells with compromised membrane integrity. By quantifying the fluorescence, we can assess the extent of membrane damage caused by our synthesized AMPs (Figure 31). neither Sushi S1, L-CONGA-Q7, D-CONGA-Q7 or our AMP control, Imitate, have any cytotoxic effect. The positive control, 100 µL H₂O, exhibited the highest fluorescence intensity, indicating maximum cell death.

Cells treated with AMPs exhibited a lower or similar fluorescence intensity when compared with the negative control. D-CONGA-Q7 demonstrated the lowest cytotoxicity among the tested AMPs. Although the values of some samples demonstrated a lower fluorescence intensity compared to the negative control, this could be due to an error in the measurement of the plate reader.

Conclusion:

Our preliminary findings suggest that the synthesized AMPs have a minimal cytotoxic effect on A549 lung epithelial cells. D-CONGA-Q7 exhibits the lowest cytotoxicity compared to L-CONGA-Q7. However, additional replicates are necessary to confirm the results and to minimize technical variabilities.

Next, we plan to investigate dose-dependent cytotoxicity by testing a range of AMP concentrations to gain insights into optimal treatment ranges. Extending the duration of testing will enable the assessment of long-term cytotoxic effects, which are crucial for understanding the safety of AMPs. Finally, the use of complementary assays such as MTT will confirm the SYTOX Green results and provide a more comprehensive view of the effects of AMP on cell viability and membrane integrity.

Lipid-based Nanocarriers

Production of Giant Unilamellar Vesicles

Key findings:

• Giant unilamellar vesicles (GUVs) with diameters up to 100 µm can be produced using a method called polyvinyl-alcohol (PVA) swelling.
• GUVs can be visualized by fluorescence microscopy using the lipophilic fluorescent dye DiO or the fluorescent lipid Atto 647N DOPE.

Aim:

With these initial experiments, we were exploring a method called PVA swelling to produce giant unilamellar vesicles (GUVs).

Experimental Setup:

GUVs were produced using two different gels: polyvinyl alcohol (PVA) and agarose, and different lipid compositions. Initially, we started with 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) due to its availability and its common use as a model lipid in biophysical studies. We also used the cationic lipid 2-dioleoyl-3-trimethylammonium propane (DOTAP), to enhance the encapsulation of negatively charged plasmid DNA and optimize fusion with bacterial membranes. As a neutral helper lipid, we included 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE). The vesicles were produced either entirely from POPC or from various DOTAP/DOPE molar ratios, and then characterized regarding their size and quantity using microscopy and flow cytometry. In addition, the vesicles were stained with the lipophilic dye DiO (excitation: 483 nm, emission: 501 nm) and imaged using fluorescence microscopy. In a separate experiment, we incorporated small amounts of the fluorescent lipid Atto 647N DOPE (excitation: 647 nm, emission: 667 nm) into the lipid composition and visualized the liposomes using confocal microscopy.

Results:

To potentially reduce material costs, we initially tested swelling with POPC on agarose in comparison to PVA. As shown in Figure 33, the liposomes produced on agarose were generally smaller and less abundant than those produced on PVA.

We characterized the liposomes formed with the following lipid compositions: 100% POPC, 100% DOPE, 100% DOTAP, and different molar ratios of DOTAP/DOPE: 52:48, 26:74 and 76:24, using microscopy (Figure 34) and flow cytometry (Figure 35). As can be seen in Figure 34, PVA swelling of 100% POPC results in a high abundance of GUVs with diameters up to 100 µm. Similar results can be observed for the DOTAP/DOPE molar ratios of 52:48 and 26:74. In contrast, when using 100% DOTAP or a DOTAP/DOPE molar ratio of 76:24, the resulting GUV diameters reach a maximum of approximately 30 µm. The use of 100% DOPE results in the formation of negligible amounts of GUVs.

Using flow cytometry, we demonstrated that liposomes were measurable as individual particles. This was confirmed by comparing our samples before and after the addition of Triton X-100, a detergent known for its ability to disrupt lipid membranes. The addition of Triton X-100 caused the liposome population to disappear, as shown in Figure 35, indicating that the measurement points corresponded to liposomes that were effectively destroyed by the detergent.

Given our aim to observe processes such as the fusion of carriers with their target later in the experimental phase, we first assessed the visualization of liposomes using both fluorescence and confocal microscopy. In one experiment, we stained the liposomes with the fluorescent dye DiO, as shown in the fluorescence microscopy image (Figure 36). In a separate experiment, we incorporated small amounts of the fluorescent lipid Atto 647N DOPE into the lipid composition and visualized the liposomes via confocal microscopy (Figure 37).

Conclusion:

Through our initial tests on liposome formation, we learned that GUV formation on PVA works more efficiently than on agarose and decided to continue with that.

After the evaluation of GUV formation using POPC and various DOTAP/DOPE molar ratios, we decided to proceed with liposomes prepared from POPC and DOTAP/DOPE at a 52:48 molar ratio. This choice was based on effective formation and the need to balance DOTAP, the cationic lipid essential for enhancing encapsulation and fusion, with DOPE, which serves as a helper lipid. We opted against using 100% DOPE to avoid instability, as excessive DOPE could compromise GUV integrity. Additionally, DOTAP/DOPE at a 76:24 molar ratio was excluded because the resulting GUVs were smaller and appeared less stable under the microscope, suggesting they were prone to rapid destabilization.

Flow cytometry analysis provided valuable quantitative data on the liposomes’ integrity. The ability to detect individual GUVs and observe their disappearance upon the addition of Triton X-100 validated the presence of intact liposomes in the samples.

Additionally, the results demonstrate that both fluorescent staining and the incorporation of fluorescent lipids into the lipid composition serve as effective techniques for the visualization in fluorescence and confocal microscopy.

Extrusion for Liposomal Size Control

Key findings:

• Extrusion can be used to create uniform large unilamellar vesicles (LUVs) from GUVs, resulting in vesicles approximately 200 nm in size.
• The optimal number of extrusion cycles is nine.
• LUV size distributions can be measured using Dynamic Light Scattering (DLS).

Aim:

Since the liposomes should be significantly smaller than the target bacterium (1.5–3 µm), we needed to reduce the size of the vesicles to approx. 200 nm. For downsizing, we selected a method called extrusion, followed by measuring the liposomal size using DLS. After the reduction in size, our liposomes would be classified as large unilamellar vesicles (LUVs, 100 - 1000 nm).

Experimental Setup:

GUVs prepared from various lipid compositions (POPC and DOTAP/DOPE in different molar ratios) were extruded using membrane pore sizes of 1000, 200, and 100 nm. The samples underwent a range of extrusion passages, specifically 0, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, and 21. The size distribution of the resulting GUVs was analyzed after each passage using DLS to identify the optimal extrusion conditions for our intended application. Read more about why we choose DLS on our Measurement page.

Results:

DLS measurements indicate that while there is a general trend of decreasing liposome size with smaller polycarbonate membrane pore sizes, the final average diameter of the liposomes is not directly reflecting the filter pore size used during extrusion (Figure 39 A). More specifically, using a 100 nm filter results in liposomes with a mean diameter of approximately 130 nm, a 200 nm filter yields liposomes approx. 180 nm, and a 1000 nm filter produces liposomes with a mean diameter of about 270 nm.

The measurement also provides the polydispersity index (PdI), which aids in determining whether the sample is more monodisperse, indicating a narrow size distribution, or polydisperse, suggesting a broader distribution. For our instrument, PdI values ranging from 0.1 to 0.4 are regarded as indicative of moderate dispersion[8]. As the data show, smaller pore sizes not only reduce the mean diameter but also lead to a more homogeneous liposome population, as evidenced by the lower PdI values (Figure 39 B).

Figure 40 illustrates the empirical modeling approach we used to analyze the relationship between the number of passages through the polycarbonate membrane and the resulting liposomal size distribution. This model was developed by fitting the limited exponential growth function

to the experimental data, correlating the number of extrusion passages with the observed changes in liposome size. Find the tool on our Contribution page and read more about DLS modeling on our Measurement page. The curve generated from the data highlights the optimal trade-off between the sample loss in the filter during extrusion and the effectiveness of size reduction. Specifically, the model shows that after seven passages, the improvements in size distribution begin to plateau, with additional passages offering only minimal gains. This empirical insight allowed us to conclude that nine extrusion passages are sufficient for creating a uniform sample, making further passages unnecessary despite the initial testing of 21 cycles.

We additionally explored the influence of various lipid compositions on the average diameter and polydispersity index (PdI) (Figure 41). GUVs were prepared from different lipid compositions: POPC, DOPE, DOTAP, and varying DOTAP/DOPE molar ratios (52:48, 26:74, and 76:24). These GUVs were extruded through a 1 µm pore size membrane, and the DLS measurements for average diameter (A) and PdI (B) were taken after different numbers of passages. Extrusion resulted in liposomal sizes quickly converging to a final diameter between 200 and 300 nm across all samples. However, the DLS measurements for liposomes prepared from 100% DOPE had to be terminated prematurely due to insufficient liposome concentration in the sample, preventing valid data collection. This may be attributed to DOPE’s limited ability to form stable lipid bilayers independently, as it is commonly used as a helper lipid to induce liposome destabilization in membrane fusion processes.

Conclusion:

We successfully used the liposome extrusion method to transform GUVs with a broad and heterogeneous size distribution ranging from 5 to 100 µm, identified under the microscope, into LUVs with a much more uniform size distribution of 200 to 300 nm. Using empirical modeling, we determined a number of nine extrusion passages through the membrane for an optimized trade-off between effort and effectiveness. These findings will be integrated into the refinement of the experimental protocol.

The optimal size of liposomes and their uniform size distribution are crucial for the biosafety and efficacy of the delivery system. The collected data can serve as a foundation for further enhancements and precise adjustments of the experimental protocol, tailored to specific applications.

Plasmid Encapsulation in Giant Unilamellar Vesicles

Key findings:

• Plasmid encapsulation using PVA swelling is achieved by incorporation of plasmid into or onto PVA gel.
• Encapsulation efficiency is higher in liposomes composed of POPC compared to DOTAP/DOPE.

Aim:

Ultimately, our liposomes should not only be produced in the right lipid composition, number and size, but above all, be able to serve as a carrier system for our plasmid. Hence, we had to establish how to encapsulate a plasmid into liposomes. To do this, we first experimented with unextruded GUVs, as these are large enough to be visualized using microscopy.

Experimental Setup:

We purified the high-copy-number plasmid pUC19 through a Maxiprep and then stained the plasmid with DAPI, a dye that exhibits fluorescence upon binding to A-T rich regions of dsDNA, allowing visualization through fluorescence microscopy (excitation: 350 nm, emission: 465 nm). PVA swelling was carried out as described. The plasmid was incorporated at different stages of the GUV production process (Figure 42) either mixed with the rehydration buffer before PVA swelling, mixed into the PVA gel before it was dried, or applied onto the dried PVA gel before the addition of the lipid solution. All samples were then analyzed using fluorescence microscopy.

Results:

Figure 43 A shows POPC-GUVs prepared with a plasmid-containing rehydration buffer. The bright field image shows that GUVs were formed, but no fluorescence signal from inside the liposomes was visible. Only the surrounding solution emitted a signal, indicating that the plasmid was not encapsulated during liposome formation. However, the results of the other two approaches, in which the stained plasmid is added into or onto the PVA gel before swelling, were promising. Some of the GUVs exhibited fluorescence, indicating that the stained plasmid may have been encapsulated. This effect was more pronounced when the plasmid was incorporated within the PVA (Figure 43 B) compared to when it was applied on top of the PVA (Figure 43 C).

In the control, a DAPI signal can also be seen, but this only comes from the membrane of the liposomes and not from their interior. In addition, this signal quickly decays within 30 to 60 seconds of exposure to UV light, in contrast to the plasmid-containing sample, which has a stable fluorescence signal. DAPI has been found to interact with phospholipid vesicles and micelles in the absence of DNA. The fluorescence properties of DAPI have been demonstrated to undergo a change upon binding to phospholipids as opposed to DNA. Specifically, the emission wavelength has been observed to shift slightly. A photobleaching effect comparable to that observed in our liposomes has yet to be documented. However, we suspect that a non-specific interaction between DAPI and the phospholipids of the liposomes may have occurred. The photostability of DAPI may be reduced in lipid environments due to the reduced stabilization of the dye, rendering it more susceptible to photobleaching.

As the results of the two approaches (where the stained plasmid was added into or onto the PVA gel before swelling) were promising, we also tested these methods with liposomes formed from DOTAP/DOPE in a 52:48 molar ratio. Similar to the POPC liposomes, some of the DOTAP/DOPE liposomes formed with plasmid-containing PVA (Figure 44 A) also exhibited a fluorescent signal. However, when the plasmid was added onto the PVA (see Figure 44 B), less GUVs were formed, and instead, clump formation was observed. This is consistent with the results obtained with POPC, where fewer liposomes were formed when the plasmid was added onto the PVA rather than into it. Overall, microscopy images suggest that the fluorescence intensity and quantity of fluorescent liposomes are generally higher with POPC compared to DOTAP/DOPE in the 52:48 molar ratio.

Conclusion:

Adding the plasmid into or onto the PVA gel before adding the lipids and the rehydration buffer, facilitated the encapsulation of the plasmid. It seems that adding the plasmid into the gel is the best approach to obtain the highest possible amount of intact GUVs with encapsulated plasmid. Unlike POPC, DOTAP is cationic, which could be the reason the encapsulation did not work out in the same way as with POPC.

Our observations suggest that the PVA swelling method may not be highly efficient in encapsulating plasmid DNA within liposomes. However, a quantitative comparison with other methods would be necessary to confirm this impression. We therefore wanted to explore a different type of lipid-based DNA carrier system with a potentially higher encapsulation rate: lipid nanoparticles.

Production of Lipid Nanoparticles

Key findings:

• Lipid nanoparticles (LNPs) can be produced using simple pipetting-mixing method or microfluidic mixing (MFM).
• DLS reveals the formation of LNPs with an average diameter of approximately 200 nm.
• LNPs can be identified using flow cytometry.
• LNPs could not be analyzed by Mass photometry due to their size.
• Gel electrophoresis can be used to verify plasmid encapsulation.

Aim:

Our observations in previous experiments suggest that the PVA swelling method may not be highly efficient in encapsulating plasmid DNA within liposomes. Therefore, we investigated a new type of lipid-based DNA carrier system with a potentially higher encapsulation rate: lipid nanoparticles. Liposomes have one or more phospholipid bilayers, and form a spherical structure to encapsulate the cargo in an aqueous solution. LNPs (100 - 400 nm) have a lipid monolayer and their payload is enveloped by reverse micelles. Their size ranges from about 100 - 400 nm.

LNPs were produced using two methods: a simple pipetting-mixing protocol and microfluidic mixing. Since LNPs are too small to be effectively characterized using standard microscopy, we instead used DLS for size measurement, flow cytometry analysis, and mass photometry for identification of LNPs as well as gel electrophoresis to verify plasmid encapsulation.

Experimental Setup:

To produce LNPs, we used two methods: a simple pipetting-mixing method and a more advanced method called microfluidic mixing (MFM) (Figure 45). The lipids dissolved in ethanol were mixed with water for the production of empty LNPs (eLNPs) or plasmids dissolved in water for the production of plasmid-containing LNPs (pLNPs). After LNP formation, remaining ethanol is removed by centrifugation using a spin concentrator. We hypothesized that MFM might offer better size distribution, plasmid encapsulation and fusion performance. However, this method is also significantly more expensive and time-consuming than simple pipetting. This is why we characterized and compared LNPs produced by both methods.

Find the general protocol we established for microfluidic mixing here. As with the pipetting, the aim is to mix the plasmid and lipids in solution to form LNPs. To find optimal conditions we adjusted two parameters, namely the total flow rate (TFR) and the flow rate ratio (FRR) while keeping the initial nitrogen to phosphorus (N/P) ratio constant with the N/P ratio from our pipetting experiments. To learn more about the effect of the N/P ratio on the plasmid encapsulation efficiency see our results from the PicoGreen assay below. We used different TFRs ranging from ca. 10 µL/s to ca. 80 µL/s for the aqueous phase with the pUC19 plasmid. The FRRs of the aqueous phase to the phase with DOTAP/DOPE/POPC 50:30:20 in ethanol were set as 4:1, 6:1 and 8:1, making the TFR of the ethanol phase subject to change for each experiment. From our results we concluded a FRR of 4:1 with a TFR of 80 µL/s worked best with our experimental setup.

Results:

Both production methods, simple pipetting-mixing and MFM, led to the formation of LNPs of similar size distribution (Figure 46). The average diameter is 150 nm for pipetting-mixing and 170 nm for MFM. The addition of Triton X-100 led to the destruction of both LNP types, which can be recognised by the subsequent shift in the size distribution towards very small particles. The detergent appears to cause the LNP structures to be broken apart, which can be interpreted as an indication that they are indeed LNPs and not other contaminants or aggregates in the sample. However, this disruption is not observed following sonication or heat treatment, where the LNPs appear to remain intact. Dr. Wohlwend suggested that during sonication, the LNPs may temporarily disassemble and then reassemble after the procedure ends.

Using flow cytometry, we demonstrated that pLNPs were measurable as individual particles. This was confirmed by comparing our samples before and after the addition of Triton X-100, a detergent known for disrupting lipid membranes. The addition of Triton X-100 caused a certain share of the pLNP population to disappear, as shown in Figure 47, indicating that the measurement points corresponded to pLNPs that were effectively destroyed by the detergent. However, results also indicate some particles in the structural range of pLNPs remain intact even with Triton X-100 treatment. These might still be complexes of DNA with lipids, which would require further investigation to actually quantify the effectiveness of Triton X-100. Especially to determine errors for downstream encapsulation assays, which implement Triton X-100 as a detergent, with the assumption that all pLNPs get destroyed.

At the Institute of Biochemistry at the University of Freiburg, in collaboration with the lab of Prof. Dr. Thorsten Friedrich, and with the assistance of Dr. Daniel Wohlwend, we conducted mass photometry analysis of eLNPs (read more on our Integrated Human Practices page). The analysis encountered several challenges, such as background noise from other particles, along with the large size and movement of the eLNPs, which affected measurement accuracy and made the technique unsuitable for determining their individual mass. However, we benefited from the integrated microscope of the mass photometer, typically employed to detect light scattering as nanoparticles or molecules pass through a laser beam. Due to the eLNPs being so large, over the measuring limit most likely which means 60 MDa or even bigger, the scattering of the whole particles could be seen. We confirmed their appearance as small vesicles, consistent with expectations (Figure 48 A). The observed particle morphology matched the moderately polydisperse distribution obtained from DLS measurements, further supporting the characterization of the eLNPs. Figure 48 B displays the ratio view calculated from two frames, the current and the frame before. White pixels indicate a particle dissociated from the glass surface, black pixels indicate a particle associated. Due to the size of the eLNPs and their property to move along the glass surface the software can not distinguish between dissociation and association of a single eLNP. Also, scattering from one particle interferes with the detection of surrounding particles.

A

B

To verify plasmid encapsulation, we performed agarose gel shift assays with the pLNPs (Figure 49). In lane 2, the pUC19 plasmid control produced a clear band, confirming that the electrophoresis process worked correctly and showing the expected location of the plasmid band. However, a smear was observed, likely due to the high amount of plasmid, that was used to make the controls applicable to the LNP sample conditions. This smear did not interfere with the clarity of the main band.
In lane 3, no band was observed, indicating successful degradation of non-encapsulated plasmid by DNAse I treatment, as expected.
In lane 4, the pUC19 plasmid treated with heat also produced a distinct band, indicating that heat treatment did not affect the integrity of the plasmid.
In lane 5, DNAse was inactivated by heat and a new plasmid was added. Because a clear plasmid band was observed, it confirms that the heat-stop method successfully inactivated the DNAse and preserved the plasmid. A smear, again attributable to plasmid supercoiling, was also noted.
In lane 6, the stop solution control showed a band similar to lane 5, indicating that the DNAse stop solution effectively halted DNAse activity and preserved the plasmid.
In lane 7, which contained LNPs with both encapsulated and non-encapsulated plasmid, a band was visible, confirming the presence of free plasmid in the surrounding solution. Additionally, a signal was observed in the pocket, suggesting that the encapsulated plasmid remained in the pocket at the top of the gel, unable to migrate due to being trapped within the pLNPs.
In lane 8, where the pLNPs were treated with DNAse, no band was detected, as expected, indicating that all non-encapsulated plasmid was successfully degraded by the DNAse, and only encapsulated plasmid remained within the pLNPs, which stayed trapped in the well.
In lanes 9 and 10, where pLNPs were treated with DNAse, followed by heat stop or stop solution and Triton X-100 to disrupt the LNPs, clear bands appeared as expected. This confirmed that previously encapsulated plasmid was protected from DNAse degradation and released from the pLNPs after their disruption and could migrate through the gel. In both lanes, faint smears were observed above the bands, which were likely due to the presence of Triton X-100 having caused signal interference.

Conclusion:

Using pipetting-mixing and MFM, we successfully produced eLNPs and pLNPs, as confirmed through DLS, flow cytometry and mass photometry. Plasmid encapsulation was effectively verified by gel electrophoresis, demonstrating the suitability of this method for assessing encapsulation success. In terms of production techniques, pipetting-mixing offers a simpler, faster, and more cost-effective approach, whereas MFM provides enhanced scalability, improved control over production parameters, and potentially higher plasmid encapsulation efficiency, which will be further explored in the following chapter.

Fluorescence-Based Nucleic Acid Quantification in Lipid Nanoparticles

Key findings:

• The PicoGreen assay reveals an encapsulation efficiency of 98% for pLNPs composed of DOTAP/DOPE/POPC and 82% for DOTAP/DOPE pLNPs.
• The molar ratio of DOTAP in the lipid composition and resulting N/P ratio impact encapsulation efficiency in a dose dependent manner.
• In a direct comparison experiment, the encapsulation efficiency determined by PicoGreen assay was 70%, while the Midori Green assay showed a consistent value of 63%.
• The close alignment between the Midori Green and PicoGreen assays demonstrates the potential of the Midori Green assay as a robust and reliable alternative.

Aim:

After verifying plasmid encapsulation in our pLNPs, we aimed to quantify encapsulation efficiency using the PicoGreen assay, a highly sensitive and widely used method for DNA quantification in LNPs. However, due to its high cost and limited availability in many laboratories, we sought a more accessible and cost-effective alternative. We developed a rapid and inexpensive DNA quantification method utilizing the widely available DNA stain, Midori Green, offering a practical solution for future iGEM teams and researchers.

Experimental Setup:

Plasmid encapsulation in our pLNPs was quantified using both the PicoGreen assay (Figure 50) and the Midori Green assay (Figure 51), allowing for a direct comparison of the two methods. For more details, visit our Measurement page.

Results:

PicoGreen Assay

The efficiency of plasmid encapsulation in pLNPs (produced with pipetting-mixing) was determined by the PicoGreen assay (Figure 52). In the final step of pLNP production, the removal of ethanol, in which the lipids were originally dissolved, is essential for preparing the pLNPs for potential therapeutic applications. To evaluate the impact of ethanol removal via centrifugation in a spin concentrator on the detection accuracy of the PicoGreen assay, we analyzed the encapsulation efficiencies post-treatment. The results revealed notable differences based on lipid composition. For the DOTAP/DOPE formulation, the encapsulation efficiency with ethanol removed was approximately 82%. In contrast, the DOTAP/DOPE/POPC formulation achieved a higher encapsulation efficiency of about 98%. This suggests that the addition of POPC enhances the ability of pLNPs to encapsulate plasmid DNA, indicating a potential benefit of this lipid composition in improving delivery efficiency.

The results indicated that the presence of ethanol did not significantly influence the assay’s accuracy, as no substantial difference was observed in the encapsulation efficiency of the DOTAP/DOPE formulation (83%). However, for the DOTAP/DOPE/POPC formulation, a lower encapsulation efficiency of approximately 84% was measured when ethanol was not removed. Although this decrease is worth noting, it could also be attributed to statistical variation, as the measurement was performed in duplicate rather than in triplicate like the DOTAP/DOPE formulation. Overall, the findings suggest that the impact of ethanol removal on the assay’s reliability and on encapsulation efficiency are minimal.

After evaluating encapsulation efficiency, we proceeded to quantify the concentration of encapsulated plasmid DNA using a calibration curve. To ensure accuracy in our quantification, we examined the suitability of various DNA standards for use in the PicoGreen assay with our pLNPs. The assay kit provides a Lambda DNA standard, which was initially employed as a reference for DNA quantification. However, inconsistencies were observed when calculating the concentrations of pLNPs encapsulating the plasmid pUC19 using the Lambda DNA standard, indicating that it may not be the most suitable reference for our specific application.

To address this, we tested the pUC19 plasmid, used for the production of our pLNPs, as a DNA standard. This experiment revealed significant differences in the behavior of the two standards (Figure 53). Specifically, the plasmid DNA standard produced a much flatter calibration curve compared to the Lambda DNA standard, highlighting that the two DNA types do not behave equivalently in the assay. This discrepancy demonstrates that the Lambda DNA standard does not accurately reflect the behavior of the plasmid DNA encapsulated in pLNPs, thus validating the importance of using a plasmid-specific standard for our applications.

Additionally, we measured the plasmid DNA standard in the presence of eLNPs to determine whether the lipids themselves might interfere with DNA quantification. The results showed that the presence of lipids had no significant influence on the measurements, suggesting that any differences observed are due to the DNA standards and not the lipid components.

In Table 1, the amount of DNA was calculated following the general protocol, comparing results for the different standards. Both standards, implementing the provided lambda DNA standard and implementing the pUC19 plasmid standard were measured in parallel. The concentrations of the stock solutions were determined in the Nanodrop prior to dilution. Both standards were then created using dilutions ranging from 40 ng/mL to ca. 400 ng/mL. In parallel controls for each DNA stock were measured. Specifically 1 µL of the 70 ng/µL lambda DNA standard and 5 µL of a plasmid control solution for pLNP preparation, containing 25 µg of pUC19 in total. Notably, the accuracy of the DNA quantification depended on using the corresponding standard for each type of DNA (Lambda or pUC19), with minimal deviations observed when the correct standard was applied. This highlights the importance of selecting the appropriate standard when adapting the assay to different DNA templates.

The results also indicate a clear enhancement in encapsulation efficiency with increasing amounts of DOTAP in the lipid composition, as demonstrated in Figure 54. Specifically, the encapsulation efficiencies range from 5% at a molar ratio of 0:30:70 (DOTAP/DOPE/POPC) to 95% at 70:30:0, illustrating a strong positive correlation between the amount of DOTAP and encapsulation efficiency.

Two types of ratios are relevant in this context:

The molar ratio of the lipid composition refers to the molar ratios of the lipids used, here DOTAP/DOPE/POPC. Increasing the proportion of DOTAP, a cationic lipid, enhances the ability to encapsulate negatively charged plasmid DNA, as cationic lipids interact more effectively with DNA, thereby improving encapsulation efficiency.

The nitrogen-to-phosphorus (N/P) ratio is critical for determining DNA encapsulation efficiency. It is calculated based on the amount of nitrogen in the cationic lipid (DOTAP) relative to the phosphorus in the DNA. When the lipid composition is altered, the N/P ratio also changes, assuming a consistent amount of added DNA. Literature suggests that an optimal N/P ratio of approx. 3 is ideal for effective nucleic acid encapsulation [10], corresponding to the DOTAP/DOPE/POPC molar ratio of 50:30:20 in this study. An excessively high N/P ratio can lead to aggregation, which negatively impacts the stability of the formulation.

Balancing the lipid composition and N/P ratio is essential for achieving optimal encapsulation efficiency. For effective fusion with target bacteria residing in the lung, it is beneficial to maintain a higher proportion of DOPE, as this enhances membrane fusion. However, an increased amount of DOTAP can be cytotoxic to lung cells (as demonstrated here), necessitating careful adjustment of lipid percentages. Thus, while increasing DOTAP can improve encapsulation efficiency, it is crucial to fine-tune both the lipid composition and N/P ratio to achieve a formulation that maximizes encapsulation while minimizing potential cytotoxic effects.

After assessing plasmid encapsulation efficiency in pLNPs produced through pipetting-mixing, we aimed to determine whether microfluidic mixing could further enhance encapsulation efficiency. As shown in Figure 55, this method appears to result in higher encapsulation efficiency. The data also suggest that the flow rate ratio (FRR) had minimal effect on encapsulation efficiency under our experimental conditions.

Additionally, the results highlight the potential to further optimize the N/P ratio for improved encapsulation. By incorporating advanced formation techniques, such as microfluidic mixing, we might also explore reducing the ratio of DOTAP to potentially enhance clinical outcomes.

Midori Green Assay

The fluorescence measurements shown in Figure 56 revealed a significant decrease in fluorescence intensity after plasmid encapsulation in pLNPs, compared to the free plasmid DNA, even though the same amount of plasmid was used in both samples. Since the Midori Green staining occurred prior to encapsulation, this suggests that the reduced fluorescence is due to the encapsulation in the nanoparticle. This observation is critical because it forced us to revise our original plan for the Midori Green assay. Initially, we planned to stain the plasmid, encapsulate it, and then use DNAse to destroy any residual non-encapsulated DNA outside the nanoparticles. The fluorescence of the encapsulated DNA would be measured and compared to a standard curve generated from free plasmid DNA. However, the experiment demonstrated that the fluorescence intensity of the encapsulated plasmid is not directly comparable to that of free plasmid, making this approach inefficient.

To address this, we revised the assay. We now determine the fluorescence of the residual non-encapsulated DNA outside the nanoparticles and compare it to the standard curve of free DNA, as explained in detail on our Measurement page. This allows us to calculate the percentage of non-encapsulated DNA, and by subtracting this from 100%, we can determine the encapsulation efficiency of the lipid nanoparticles.

To evaluate the performance of the Midori Green and PicoGreen assays, we compared both methods using the same batch of pLNPs encapsulating plasmid DNA that was pre-stained with Midori Green. This batch was divided into two samples: one was tested using the PicoGreen assay, while the other was tested with the Midori Green assay. This approach allows for a direct comparison of the assays using the exact same sample.

The encapsulation efficiency results (Figure 57) show a mean value of 63% for the Midori Green assay and 70% for the PicoGreen assay (Figure 57 A). The means are relatively close, and a comparison of the individual replicates (Figure 57 B) shows only minor variation between the two assays for the same samples. For example, Replicate 1 exhibited values of 56% (Midori Green) and 70% (PicoGreen), while Replicate 2 was almost identical, with 71% (Midori Green) and 71% (PicoGreen). Replicate 3 showed a slight difference as well, with 61% (Midori Green) compared to 68% (PicoGreen). Despite these small variations, the overall trends remain consistent, with both assays producing highly comparable results. The small differences observed likely reflect minor day-to-day or assay-specific factors, rather than any inherent limitation of the Midori Green method.

In the preliminary PicoGreen assay, an encapsulation efficiency of 84% (Figure 52) was observed for pLNPs composed of DOTAP/DOPE/POPC at a molar ratio of 50:30:20 (ethanol removed), which served as a benchmark for subsequent experiments. However, a direct comparison experiment, in which pLNP samples were stained with Midori Green and split for testing with both the Midori Green and PicoGreen assays, revealed encapsulation efficiencies of 63% and 70%, respectively. The close alignment of these values demonstrates the potential of the Midori Green assay to produce results consistent with those obtained from PicoGreen, despite the reduced efficiency observed in this particular experiment. The discrepancy between the initial PicoGreen result (84%) and the later measurement (70%) could be attributed to various factors, including sample handling, experimental timing, or operator-related differences. It is also possible that the presence of Midori Green during the encapsulation process influenced the efficiency, or that the fluorescence properties of Midori Green interfered with the PicoGreen measurements. While the possibility of such interference cannot be fully excluded, the comparable results obtained from both assays in the direct comparison support the reliability of the Midori Green assay for assessing encapsulation efficiency.

Conclusion:

The PicoGreen assay revealed high encapsulation efficiencies for pLNPs formulated with DOTAP/DOPE, with particularly notable results for those including POPC, suggesting that the addition of POPC may enhance encapsulation. Additionally, ethanol removal had minimal impact on both assay reliability and encapsulation efficiency, an important finding given that ethanol was not removed from the samples tested with the Midori Green assay. To further optimize the production in regards to potential sample loss, dialysis could be implemented for ethanol removal instead of centrifugation with a spin concentrator.

Comparative experiments using a Lambda DNA versus a plasmid-specific standard underscore the need for plasmid-specific standards in pLNP DNA quantification, as differences in DNA behavior affect the accuracy of encapsulation measurements and the optimization of pLNP formulations for therapeutic applications.

The comparison of DOTAP molar ratios and resulting N/P ratios showed a strong correlation between DOTAP proportion and encapsulation efficiency, likely due to DOTAP’s cationic nature, which enhances its ability to encapsulate negatively charged plasmid DNA. However, it is crucial to balance these ratios, as high concentrations of DOTAP can be cytotoxic [11], while sufficient levels of DOPE are needed to promote fusion and improve encapsulation efficiency [12].

The Midori Green assay produced results closely aligned with those of the PicoGreen, suggesting it is a reliable and cost-effective alternative for measuring nucleic acid encapsulation efficiency in pLNPs. While further studies are needed to confirm its robustness under a broader range of conditions, the results demonstrate significant potential for future applications in nucleic acid quantification and pLNP development.

Fusion with P. aeruginosa

Key findings:

• LNPs and extruded liposomes attach to the membrane of P. aeruginosa.

Aim:

The aim of this experiment was to achieve fusion between GUVs, LUVs, and LNPs with the target bacterium P. aeruginosa. This was explored by integrating the glycosphingolipid Gb3 into the lipid formulation to enable interaction with LecA, which is present on the surface of P. aeruginosa [13,14,15] This approach aimed to enhance selectivity for bacterial cells.

Experimental Setup:

Considering the biosafety level of Pseudomonas aeruginosa, the experiment was carried out under S2 conditions by Dr. Makshakova, a member of the Römer lab at the University of Freiburg. Read more about our collaboration on our Integrated Human Practices page. We provided the samples and later participated in the imaging process to observe potential fusion events.

The lipid-based nanocarriers were prepared both with and without Gb3 and encapsulated or non-encapsulated with pUC19 plasmid DNA. The lipid mixtures included DOTAP, DOPE, POPC, and the fluorescent lipid Atto 647N DOPE (excitation: 647 nm, emission: 667 nm) for fluorescence tracking. Further details on the molar ratios are available in the Notebook. The P. aeruginosa lab strain PA-01, expressing GFP (excitation: 489 nm, emission: 510 nm) in the membrane was placed in a Greiner CellView chamber. The culture was diluted 1:100 in PBS to a volume of 100 µl and 5 µl of the GUVs, LUVs, and LNPs were added directly to the bacterial culture before loading the chambers. Images were acquired using a Nikon A1 plus confocal microscope with a 60x objective. To prevent GUVs attaching to the surface, glass of the CellView chamber was coated with casein. Images were taken after 30 min incubation to allow potential fusion interactions to take place.

Results:

The image analysis showed that, colocalization of red fluorescence (from the Atto 647N-labeled LUVs) and green fluorescence (from the GFP-expressing bacterial membrane) was observed when the bacteria were incubated with LUVs containing Gb3 (Figure 58 A). This data suggests that fusion between the bacteria and the LUVs has taken place. In contrast, LUVs without Gb3, were present in close proximity to the bacteria but no colocalization was observed (Figure 58 B).

For the LNPs, we were only able to successfully image the LNP sample containing Gb3. However, the sample contained a high number of small particles, which made it challenging to observe specific attachment or fusion events (Figure 59). After 30 minutes of incubation, the bacteria and LNPs were found to accumulate in the same areas on the slide. This could suggest potential interaction, though the exact nature of the interaction remains unclear due to the difficulty in resolving individual particles. Also in lower focal planes more bacteria were visible, indicating the LNPs might have aggregated and settled on top of the bacteria. However, in the focal plane of Figure 59 the GFP signal of the bacteria and LNPs overlapped.

The Gb3 and non-Gb3 GUVs adhered to the glass slide and ruptured, preventing any conclusive imaging of interactions or fusion with the bacteria. However, small particles remained and we captured their movement attached to a P. aeruginosa (Figure 60).

Conclusion:

While the GUV samples encountered technical difficulties that led to rupture, the LUV samples demonstrated promising results, particularly with the incorporation of Gb3 in the lipid formulation. Here, potential fusion was observed. For the LNPs, optimization of both the sample and bacterial concentrations is necessary for effective imaging. Furthermore, incorporating DNA staining within the vesicles could provide valuable insights into whether the DNA is delivered through the fusion process.

Effect of Lipid-based Nanocarriers on Mammalian Cells

Key findings:

• As evidenced by the MTT assay, LUVs exhibited significant cytotoxic effects on human A549 lung cells. In contrast, LNPs showed no significant cytotoxicity at low concentrations, but a dose-dependent cytotoxic effect was observed at higher concentrations.
• According to the SYTOX Green assay, neither LUVs nor LNPs demonstrated significant cytotoxicity on A549 lung cells, suggesting minimal impact on cell membrane integrity.
• Preliminary fusion experiments indicated that pLNPs might transfect A549 lung cells.

Aim:

Our nanocarrier system is intended for aerosol delivery to the lungs, where the nanocarriers selectively fuse with target pathogens while avoiding interaction with the patient’s lung cells. Therefore, a key part of our study involves testing the nanocarriers on human lung cells. Specifically, we aim to assess the biocompatibility of our nanocarriers (LUVs, eLNPs, and pLNPs) with human lung epithelial cells. Furthermore, we want to determine the potential for unintended fusion and genetic material transfer between the nanocarriers and human lung cells.

Experimental Setup:

A549 lung epithelial cells were cultured in DMEM under standard conditions. The cells were exposed to varying concentrations of the nanocarriers: LUVs (suspended in 10% sucrose solution), eLNPs, and pLNPs (both suspended in water). This range of concentrations was designed to identify the threshold at which nanocarriers might affect cell viability.

Several controls were implemented to ensure result validity. As positive control H₂O treatment (expected to induce cell death), solvent controls included 10% sucrose solution and water (to account for potential solvent effects) and viability control (DMEM) representing 100% cell viability. Two complementary assays were performed in 96-well plates to evaluate the effects of nanocarriers on A549 lung cells:

MTT Assay: This colorimetric assay assessed metabolic activity as an indicator of cell viability.

SYTOX Green Assay: This fluorescence-based assay evaluated membrane integrity, providing a measure of cytotoxicity

MTT Assay Procedure

Following overnight incubation with nanocarriers, MTT solution was added to each well (Figure 61). In this assay, metabolically active cells reduce the yellow MTT ((4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) to purple formazan crystals, providing a colorimetric indication of cell viability. The intensity of the purple color was quantified spectrophotometrically, with higher absorbance indicating better cell viability. Results were normalized to viability control to calculate relative cell viability percentages.

SYTOX Green Assay Procedure

Treated cells were stained with SYTOX Green dye (Figure 62). Fluorescence intensity was measured, which correlates with the degree of membrane damage. Percentage cytotoxicity was calculated using a viability and a 100% cell lysis control as reference points [15].

Fusion Assay Procedure

An eGFP-encoding plasmid with a human-specific promoter was encapsulated into the pLNPs using a pipetting-mixing technique. Post-incubation, the culture medium was replaced with phenol red-free DMEM to optimize fluorescence detection. Two complementary methods were used to assess eGFP expression. Fluorescence microscopy for qualitative visual confirmation of eGFP-expressing cells, and fluorescence intensity measurements for quantitative analysis of eGFP expression levels.

As a positive control, eGFP-encoding plasmid was transfected in A549 cell using lipofectamine according to manufacturer protocol.

The presence and intensity of eGFP fluorescence were used as indicators of successful pLNP-cell fusion events and subsequent plasmid uptake and expression.

Results:

MTT Assay

The MTT assay results, as shown in Figure 63, revealed a dose-dependent effect of LUVs on A549 cell viability. Positive control, treatment with 100 µL H₂O, resulted in complete cell death, as expected. To determine whether the sucrose solution affected viability, a control with 100 µL DMEM containing 1% sucrose was included, showing a reduction in viability, suggesting a cytotoxic effect.

In comparison, cells treated with various volumes of LUVs (composed of DOTAP/DOPE in a 52:48 molar ratio) showed a dose-dependent cytotoxicity. At the lowest volume (0.8 µL LUVs) viability remained relatively high, suggesting minimal cytotoxicity. Similarly, 2.0 µL LUVs resulted in a comparable viability. A notable decrease in viability was observed at 3.9 µL and 7.7 µL of LUVs. This trend mirrors the sucrose control results, suggesting that the observed reduction in cell viability might be partially attributed to the sucrose solution used to suspend LUVs, in addition to any direct effect of LUVs themselves.

The MTT assay results shown in Figure 64 illustrate the effects of eLNPs on A549 cell viability, comparing two lipid compositions: DOTAP/DOPE in a 52:48 molar ratio (Figure 64 A) and DOTAP/DOPE/POPC in a 50:30:20 molar ratio (Figure 64 B). Controls included water and DMEM dilutions to assess any impact of solvent on cell viability, as well as a positive control with 100 µL of H₂O, which predictably led to complete cell death. Water dilutions also showed a dose-dependent cytotoxicity, with cells treated with 30 µL H₂O + 70 µL DMEM maintaining high viability, while 60 µL H₂O + 40 µL DMEM decreased viabilities, indicating mild cytotoxic effects at higher water volume.

Cells treated with eLNPs formulated from either DOTAP/DOPE (52:48 molar ratio) or DOTAP/DOPE/POPC (50:30:20 molar ratio) exhibited a dose-dependent decrease in viability. At lower eLNP volumes (6 µL and 15 µL), both formulations showed minimal cytotoxicity. As the volume increased to 30 µL, a moderate reduction in viability was observed for both formulations. At the highest volume (60 µL), both formulations caused a more pronounced reduction in viability, reflecting stronger cytotoxic effects. The cytotoxicity observed at this level may be partially attributed to water-related stress, as the corresponding water controls also exhibited some cytotoxicity. Overall, both formulations demonstrated a similar pattern of decreasing cell viability with increasing eLNP volumes.

The MTT assay results, as shown in Figure 65, revealed a dose-dependent effect of pLNPs on A549 cell viability. As expected, treatment with 100 µL of H₂O (positive control) resulted in complete cell death. Controls with varying volumes of water mixed with DMEM were included to evaluate any effects of the dilution solvent. Again, consistent with the previous experiment, the water controls behaved as expected; cells treated with 30 µL H₂O + 70 µL DMEM maintained high viability, while the 60 µL H₂O and 40 µL DMEM group showed reduced viability, indicating mild cytotoxicity from the higher water content.

In comparison, cells treated with pLNPs encapsulating an eGFP plasmid showed varying effects on viability depending on the concentration. Lower volumes of pLNPs (3 µL, 6 µL, and 15 µL) had only slight effects on cell viabilities.

At higher concentrations, a decrease in cell viability was observed. Cells treated with 30 µL of pLNPs exhibited a reduced viability, possibly due to the higher pLNP dose or the combination of pLNPs and water. Finally, the highest concentration of 60 µL pLNPs noticeably reduced cell viability, indicating a cytotoxic response at this level. Notably, the corresponding water control (60 µL H₂O) already showed some cytotoxicity, which was exacerbated by the addition of pLNPs.

SYTOX Green Assay

The SYTOX Green assay results shown in Figure 66 demonstrated no dose-dependent effect of LUVs on A549 cell viability. Treatment H₂O resulted in the highest fluorescence intensity, indicating maximum membrane disruption. DMEM containing 1% sucrose showed no increase in fluorescence compared to untreated cells, suggesting no effect on membrane integrity.

LUV treatments at various volumes resulted in the following observations: a similar low fluorescence intensity was measured for all volumes of GUVs, with no difference to negative controls.

The SYTOX Green assay results, as shown in Figure 67, revealed no cytotoxic effects of eLNPs on A549 cell, with a lipid composition, DOTAP/DOPE/POPC in a 50:30:20 ratio. Controls included water and DMEM dilutions to assess any impact of solvent on cell viability, as well as a positive control with 100 µL of H₂O, which predictably led to almost complete cell death, and was taken as 100% cytotoxicity value. Water dilutions showed no dose-dependent cytotoxicity, suggesting minimal impact on cell death from the samples with 30 and 60 µL of water. Overall, the results indicate no correlation between LUVs or eLNPs and cytotoxicity.

Fusion Assay

eGFP expression was assessed using two complementary methods: fluorescence microscopy for visual confirmation of eGFP-expressing cells (Figure 68) and fluorescence intensity measurements for quantitative analysis (Figure 69).

The lipofectamine-mediated transfection of the eGFP-encoding plasmid (positive control) successfully resulted in eGFP expression in A549 cells, validating the functionality of the assay. Treatment of cells with the eGFP-encoding plasmid alone showed no fluorescent signal, confirming that the plasmid cannot enter cells without a delivery system.

For DOTAP/DOPE (52:48 molar ratio) pLNPs some fluorescent signal was observed, indicating limited fusion or interaction with A549 cells and subsequent plasmid uptake (Figure 69 A). Fluorescent measurement of DOTAP/DOPE/POPC (50:30:20 molar ratio) pLNPs showed no detectable fluorescent signal, suggesting minimal to no fusion with A549 cells or plasmid uptake (Figure 69 B).

No significant difference in fluorescence was observed between different volumes of pLNPs used in the assay, suggesting that the observed effects were not dose-dependent within the tested range. It is important to note that these observations are based on preliminary data. Further experiments, improvements in methodology, and additional control experiments are necessary to draw conclusive results.

Conclusion:

In conclusion, our experiments revealed that Large Unilamellar Vesicles (LUVs) exhibited no detectable cytotoxic effects on A549 cells when used at low concentrations, as measured by the MTT assay. Lipid Nanoparticles (LNPs), both with and without encapsulated plasmid, demonstrated only mild dose-dependent cytotoxicity. This effect was primarily correlated with the volume of nanocarriers used and likely reflects the influence of the solvent rather than inherent toxicity of the nanocarriers themselves.

Importantly, these findings were further confirmed by the SYTOX Green assay, which showed minimal or negligible cytotoxicity for both eLNPs and LUVs at the tested concentrations. This additional data supports our conclusion about the low cytotoxicity profile of these nanocarriers.

Preliminary data of the eGFP fusion assay suggest minimal or no fusion of pLNPs with human lung tissue cells, addressing a key objective of preventing unwanted expression of payload in human cells.

These findings suggest that both LUVs and LNPs have potential as safe delivery vesicles, particularly at lower concentrations, though careful consideration of dosage and solvent effects is necessary when designing applications for these nanocarriers.

Outer Membrane Vesicles

OMV Production and Characterization

Introduction

We have utilized the natural ability of outer membrane vesicles (OMVs), produced by Gram-negative bacteria, to deliver an antimicrobial peptide (AMP) encoding plasmid. To achieve both maximum levels of plasmid encapsulation and hypervesiculation, we have used the E. coli BL21(DE3) Omp8 strain provided to us by Prof. Dr. Daniel Müller and Dr. Nico Strohmeyer (See our Human Practices). This strain has been modified through the deletion of the following membrane proteins: OmpA, OmpC, OmpF, LamB; in order to achieve greater display of a recombinant outer membrane proteins [16].

Key findings:

• E. coli BL21(DE3) Omp8 exhibit reduced growth rate compared to wild type E. coli BL21(DE3).
• Successful OMV isolation confirmed by DLS.
• Complete OMV lysis was achieved using a combination of detergent treatment and sonication.
• Enhanced OMV production in Omp8 strain quantified by total protein concentration.

Aim:

To establish and validate a robust protocol for the production, isolation, and characterization of outer membrane vesicles (OMVs) from Omp8 cells, including the development of methods for OMV lysis to facilitate accurate protein quantification.

Experimental Setup:

1. Growth Curve: 50 mL of LB media, supplemented with or without antibiotics, was inoculated with overnight cultures of both strains E. coli BL21(DE3) and the E. coli BL21(DE3) Omp8 to reach OD₆₀₀ of 0.1. The cultures were incubated at 37°C with shaking at 200 rpm. Optical density OD₆₀₀ measurements were performed every hour, for a total of 13 hours, with LB as a blank. Two further measurements were performed after 24 hours and 29 hours. Resolution limitations associated with the nanodrop device necessitated the following dilutions:
   1. Dilutions of 1:5 once cultures reached an optical density of 0.5
   2. Dilutions of 1:20 once cultures reached an optical density of 1.0

2. OMV Extraction : LB medium was inoculated with an overnight culture and incubated for 20 hours. Cells were removed by centrifugation at 10,000 x g, 4°C for 45 minutes. The supernatant was then filtered through 0.45 µm filters followed by ultracentrifugation at 60,000 x g for 1.5 hours. The pellets were resuspended in cold PBS and centrifuged once more at 60,000 x g for 1.5 hours. Resulting final pellets were resuspended once more and filtered through 0.22 µm filters. Extracted OMVs were stored at -80°C.

3. DLS: Samples were prepared by diluting OMVs to a volume of 60 µL. Size distribution of both untreated samples and samples treated with 2.5% Triton-X followed by 10 minutes sonication, were analyzed via DLS (for an explanation see our Measurement page). A control with 2.5% Triton-X alone was included. DLS was performed for OMVs extracted from both WT and Omp8 strains.

4. Flow Cytometry Analysis: Preparation of both treated and untreated OMVs were done as described previously for DLS measurements. OMV samples were then analyzed using a FACS machine.

5. BCA Protein Assay Pierce™ BCA Protein Assay-Kit from ThermoFisher was used according to the manufacturer’s instructions: OMVs were lysed as mentioned in previous experiments, working reagents are added and the solution is incubated at 37°C. The resulting color change allows for the quantification of protein concentrations. The BSA sample provided was used as a standard. Protein concentration of OMVs extracted from both WT and Omp8 strains were analyzed.

These experiments are crucial for establishing a reliable OMV production pipeline. Verifying the purity of isolated OMVs and developing a validated lysis method are essential steps for accurate protein quantification, which is fundamental for downstream applications and quality control in OMV-based research.

Results:

Growth Analysis of E. coli BL21(DE3) Omp8

The maximum growth rate was calculated using the following equation :

Growth rate µ = (y2 - y1) / (x2 - x1)

For WT, the maximum rate was measured after 5 hours at 2.49 divisions per hour, whereas the Omp8 strain showed a rate of 1.32 divisions per hour.

The doubling time was calculated by inserting the above value into the following equation: td = ln(2) / µ

The doubling time for the WT strains was measured to be 16.7 minutes, whereas the Omp8 strain required 31.5 minutes.

Both strains, WT and Omp8, entered the logarithmic phase after 2 hours, the stationary phase after 10 hours and death phase after 24 h.

OMV Extraction and Characterization

OMVs could successfully be extracted from both WT and Omp8 cultures, and appear as a quickly-dissolving black pellet.

DLS analysis confirmed the successful isolation of OMVs from both E. coli strains. As shown in Figure 76, the isolated OMVs exhibited average diameters of 120 nm for E. coli BL21(DE3) Omp8 and 150 nm for E. coli BL21(DE3). These size distributions fall within the expected range for bacterial OMVs, typically 20-250 nm, reported in the literature [17].

In order to facilitate accurate protein concentration measurements, OMVs had to be lysed as encapsulated proteins would be inaccessible to BCA reagents. After having tested various methods to achieve lysis (See our Notebook), we decided on treatment involving 2.5% Triton X-100 followed by 15 min sonication. A significant reduction in average diameter was observed for both E. coli BL21(DE3) (Figure 76 (A)) and E. coli BL21(DE3) Omp8 (Figure 76 (B)). However, Triton X-100 control samples showed a peak at approx. 9 nm, which might be associated with formation of Triton X-100 micelles (FIG). This could not be conclusively correlated with lysis due to the ~10 nm peak associated with Triton X-100 micelles as observed in Figure 77..

Due to the aforementioned Triton X-100 peaks observed via DLS, the obtained results could not be used to definitively prove OMV lysis with the tested conditions. Proper verification of lysis was performed via flow cytometry, wherein the region of interest (ROI) associated with OMVs disappeared upon treatment with both Triton X-100 and Sonication as observed in Figure 78. These results could decisively prove OMV lysis using the tested conditions.

BCA assays performed using BSA as standards, indicated that protein concentrations of 0.625 mg/mL and 0.745 mg/mL were obtained for OMVs isolated from E. coli BL21(DE3) Omp8 and E. coli BL21(DE3) respectively.

Conclusion:

Possibly as a result of the four deletions, the Omp8 strain showed a twofold increase in doubling time, however this could possibly be attributed to its growth in a medium containing kanamycin. We were nonetheless able to utilize standardized protocols to extract OMVs from both the Omp8 and WT strains. The isolated OMVs could then be analyzed via DLS in order to determine their average sizes. In order to characterize total protein concentration, lysis of OMVs was necessary, which was achieved by combining Triton X-100 and sonication. DLS results of the same were inconclusive due to the peak associated with Triton X-100, however further verification through flow cytometry could definitively associate the observed decrease in diameter with OMV lysis. Accordingly, BCA assays were performed in order to quantify protein concentrations. With these experiments we were able to establish a standardized set of protocols that could be used for OMV isolation, characterization and quantification.

Development of a Modular Targeting System for OMVs

Our approach for precise delivery is built around modularity. We achieve this by displaying a SpyTag on the OMV surface through the membrane protein ‘eCPX’. This SpyTag is designed to be “caught” by fusion proteins consisting of a SpyCatcher and the P2 Phage-Tail protein (derived from the PRD1 bacteriophage [18]).

Initially, the plasmid for eCPX-SpyTag expression was driven by an IPTG inducible Trc promoter. However, this induction method resulted in a heterogeneous OMV population in terms of surface proteins. To address this issue, we explored the use of constitutive promoters for eCPX-SpyTag expression: LacIq (BBa_K3257003) and Amp (BBa_K2569033). We have evaluated the two promoters based on their impact on cell growth/burden and the level of eCPX-SpyTag display.

Key findings:

• Recombinant outer membrane protein eCPX-SpyTag was successfully displayed on OMV surfaces.
• LacIq promoter (BBa_K3257003) showed highest levels of eCPX-SpyTag expression while maintaining cell viability.

Aim:

To establish a modular targeting system for OMVs through surface functionalization with SpyTag, enabling versatile and efficient coupling of targeting moieties or cargo proteins.

Experimental Setup:

To develop and evaluate our modular targeting system for OMVs, we designed a series of experiments to assess the efficiency of SpyTag display and the impact of different promoters on OMV production and cell viability. The following experiments were conducted:

1. Cloning and Transformation

• We cloned two vectors for the expression of eCPX-SpyTag under the control of constitutive promoters using cloning techniques.
• Final constructs were transformed into E. coli BL21(DE3) Omp8 cells (See our Notebook).

2. Growth Rate Analysis

• Growth rates were compared between E. coli strains harboring:
  • pTrc99a-eCPX-ST (original construct with IPTG-inducible Trc promoter)
  • pTrc99a-LacIq-eCPX-ST (LacIq promoter construct)
  • pTrc99a-Amp-eCPX-ST (Amp promoter construct)
• Optical density (OD₆₀₀) was measured hourly for 24 hours (See Protocols).

3. OMV Isolation and Characterization

• OMVs were isolated from each strain using our established protocol.
• Purified OMVs were characterized by DLS to determine size distribution.
• Total protein content was quantified using BCA assay to estimate OMV production efficiency (See Protocols).

4. Protein Expression Verification

• Expression of eCPX-SpyTag was verified by SDS-PAGE analysis of OMV samples from all strains (See Protocols).

Results:

1. Cloning of Constitutive Promoters:

The constitutive LacIq (BBa_K3257003) and Amp (BBa_K2569033) promoters were cloned into the pTrc99-eCPX-SpyTag plasmid in place of the IPTG inducible trc promoter. The resulting plasmids were then sent for sequencing for verification.

2. Bacterial Growth Analysis:

Comparing the growth of all strains, all E. coli BL21(DE3) Omp8 cultures are growing slower compared to E. coli BL21(DE3) for the first 10 hours, likely due to the deletion of outer membrane proteins. Interestingly, the LacIq promoter construct demonstrated a higher maximum growth rate.

For WT, the maximum growth rate was observed with an OD₆₀₀ increase of 2.49/h after 5h, for Omp8 1.33/h after 5 h, for Omp8 with the amp promoter 1.28/h after 6 h and for Omp8 with the LacIq promoter 2.19/h after 9h.

The resulting doubling times are 16 min for WT, 31 min for Omp8, 32 min for Omp8 with amp promoter and 19 min for Omp8 with the LacIq promoter.

All 4 bacteria cultures entered the logarithmic phase after 2 h, the stationary phase after 10 h and death phase after 24 h.

3. OMV Characterization:

DLS analysis confirmed the successful isolation of OMVs from all four E. coli BL21(DE3) Omp8 strains. The isolated OMVs exhibited average diameters of 105.1 nm.

The following average diameters and protein concentrations were obtained :

To verify the expression and incorporation of eCPX-SpyTag into OMVs, we performed SDS-PAGE analysis. This allows us to visualize and compare the protein profiles of OMVs, with a particular focus on the presence of eCPX-SpyTag. A distinct band corresponding to eCPX-SpyTag was detected at approximately 20 kDa in OMVs from strains with the inducible construct demonstrating effectiveness of the induction system (Figure 82). This band was absent in the control strains, confirming its specificity to eCPX-SpyTag expression. The constitutive LacIq construct also successfully expressed eCPX-SpyTag, indicating that this promoter can drive continuous production of the protein. Unexpectedly, no specific band for eCPX-SpyTag was detected in OMVs from the constitutive Amp construct. This suggests potential issues with expression or incorporation of eCPX-SpyTag in this strain.

Conclusion:

The constitutive Amp and LacIq bacterial cultures did not exhibit significant growth inhibition compared to E. coli BL21(DE3) or E. coli BL21(DE3) Omp8 cells. This suggests that the continuous expression of eCPX-SpyTag does not impose a substantial metabolic burden on the host cells.

We successfully purified OMVs from all strains, including those with constitutive promoters. This demonstrates that the modification of the expression system does not interfere with the process of OMV formation and release. While both constitutive promoters allowed for OMV production, we observed differences in protein expression levels. The LacIq construct showed clear eCPX-SpyTag expression, whereas the Amp construct results were less conclusive and may require further optimization.

Growth curves performed with sequenced plasmids containing promoters for the constitutive expression of eCPX-SpyTag seemingly indicated that neither the presence of a plasmid nor the constitutive expression of eCPX-SpyTag had a significant effect on cell growth. OMVs from E. coli BL21(DE3) Omp8 strains containing eCPX-SpyTag plasmids controlled by the (uninduced) Trc promoter, (induced) Trc promoter, LacIq promoter and Amp promoter were characterized in terms of size, protein content, and SpyTag display. Results indicated that iTrc-OMVs showed the highest expression of eCPX-SpyTag, followed by LacIq-OMVs. Amp OMVs showed the lowest expression of eCPX-SpyTag.

Functionalization of eCPX-SpyTag OMVs with SpyCatcher-P2 Target Protein

Key findings:

• OMVs exhibiting eCPX-SpyTag on their surface could be functionalized with SpyCatcher bound fusion proteins.

Aim:

To demonstrate the functionality of the SpyTag-decorated OMVs by coupling them with a SpyCatcher-fused target protein (P2), thus validating the modular targeting system.

Experimental Setup:

We cloned the SpyCatcher-P2 fusion protein (BBa_K5057011) under control of a Doxycycline inducible promoter. E. coli BL21(DE3) were transformed with pBb2Ac-SpyCatcher-P2. For expression, a larger culture was grown until OD₆₀₀: 0.6-0.8 and induced with 200 ng/ml Doxycycline. Cells were incubating for 24 h at 25°C. SpyCatcher-P2 Protein was purified using Immobilized Metal Affinity Chromatography (IMAC) with the help of our Primary Investigator, Dr. Pavel Salavei (See Protocols).

Results:

We successfully express and purify SpyCatcher-P2 (Figure 83). Following IMAC purification, we revealed a total protein concentration of about 4.0 mg/mL

For functionalization, purified OMV samples were diluted to a concentration of 10 µg, and incubated with 1.5 µg of SpyCatcher-P2 fusion protein for 1 hour at 37°C. Samples were then separated using a SDS Page. (See Protocols) On SDS-page (Figure 83) we could identify the express membrane protein eCPX-SpyTag (~20kDa), the purified SpyCatcher-P2 protein (~74kDa) and the functionalized complex of eCPX-SpyTag with SpyCatcher-P2 (~94kDa).

To confirm the specific expression of SpyCatcher-P2 and its fusion to eCPX-SpyTag on OMVs, we conducted Western blot analyses using two different antibodies (See Protocols):

The anti-His antibody specifically detected the His-tagged SpyCatcher-P2 fusion protein. A clear band was observed at ~74 kDa, corresponding to the unbound SpyCatcher-P2. For OMV samples derived from inducible (Ind. eCPX) and from constitutive lacIq expression an additional band at ~94 kDa was detected, representing the eCPX-SpyTag:SpyCatcher-P2 complex. This result confirms the presence and successful coupling of the His-tagged SpyCatcher-P2 to OMVs.

The anti-SpyCatcher antibody (Figure 84) also detected the SpyCatcher-P2 fusion protein and its complex with eCPX-SpyTag. However, we observed a significant degree of non-specific binding with this antibody.

Conclusion:

Taken together, our experiments successfully demonstrated the functionalization of purified OMVs exhibiting eCPX-SpyTag with the SpyCatcher-P2 fusion protein. This achievement represents a crucial first step in validating our modular approach towards targeting-ligand functionalization of OMVs. Our next crucial step is to evaluate the efficiency of OMV-mediated transformation. This process begins with the critical task of verifying plasmid encapsulation within the OMVs.

Plasmid Encapsulation in OMVs and Fusion

Key findings:

• Presence of DNA within isolated OMV solutions could be detected, however both DNA encapsulation and OMV-mediated transformation could not be proved.

Aim:

To demonstrate the capability of OMVs to encapsulate plasmid DNA and to investigate their potential for fusion with target cells, laying the groundwork for their use as a delivery system.

Experimental Setup:

E. coli BL21(DE3) Omp8 cells were transformed with pUC19 plasmid. OMVs were purified from these transformed cells (pUC19-OMVs) using our established protocol (See Protocols).

Presence of pUC19 plasmid in OMVs was analyzed via PCR using pUC19-specific primers. Various treatments including DNAse (used to digest DNA outside the OMVs) and Triton X-100 (used to lyse OMVs and expose the encapsulated DNA) were used, alongside DNAse digest and stop controls.

After verifying the presence of plasmids within the isolated OMV solutions, we tried to transform E. coli BL21(DE3) strains using the extracted OMVs without the use of targeting-ligands. (See Protocols).

Results:

1. Plasmid Encapsulation by pUC19-OMVs

The following results (Figure 87 and 88) were obtained when samples were amplified via PCR and visualized by Gel-Electrophoresis:

pUC19-OMVs showed a weaker band compared to Empty-OMVs + 10 ng pUC19. This suggests that pUC19-OMVs contain less than 10 ng of DNA.

DNase treatment of Empty-OMVs + 10 ng pUC19 and pUC19-OMVs resulted in no bands, indicating that the amount of DNase used (2 units) was sufficient to digest external DNA.

In the case of pUC19-OMVs treated with both DNase and Triton X-100, no bands were observed either, suggesting that pUC19 encapsulation was not successful. This stands in stark contrast to the results observed by Tran. et al.

Reasons for this could be due to imperfections stemming from DNase treatment or from the isolation of OMVs encapsulating plasmids.

2. Non-Specific Fusion

Results of non-specific fusion experiments performed with non-competent E. coli BL21(DE3) cells and OMVs without targeting-ligands were inconclusive. Growth was observed in overnight cultures containing both OMVs and bacteria incubated in LB + ampicillin. However, further inoculations in fresh media yielded no growth. The controls that we performed eliminated the possibility of contamination. We were therefore unable to determine if OMVs could deliver plasmids.

Conclusion:

Plasmid localization inside OMVs could not be clearly verified, but due to complicated procedures and many unknown factors not excluded either. With regards to the confusing results obtained within non-specific fusion experiments, we hypothesize that these observations were due to encapsulation of the ampicillin degrading enzyme, β-lactamase, as observed in literature [19].

Verification of β-Lactamase Encapsulation

Experimental Setup:

Our hypothesis was tested through the following experiments:

1. OMVs encapsulating pUC19 or pET (both Amp plasmids) were incubated overnight in LB media containing ampicillin at 37°C. Subsequently, E. coli BL21(DE3) cultures were inoculated into the same media containing OMVs.

2. A further fusion experiment was performed according to the protocol established previously. However, bacteria were pelleted after co-incubation (stationary) with OMVs, in order to remove the OMVs present in the supernatant. The pelleted bacteria were then resuspended in LB and inoculated in test tubes and cultured overnight as previously described (See Notebook).

Results:

1. Growth (increase in optical density) was observed indicating that OMVs had successfully degraded antibiotics, allowing bacterial growth. No growth (no increase in optical density) was observed when E. coli BL21(DE3) was inoculated in media containing just ampicillin.

2. No growth was observed in any of the test tubes containing LB containing ampicillin, indicating that bacteria were not transformed when co-incubated with OMVs. This experiment indicates that OMV mediated plasmid transfer did not take place.

Conclusion:

This experiment suggested that OMVs had encapsulated β-lactamase alongside the plasmid. The resulting OMVs could then degrade ampicillin present in LB media by themselves, which has not yet been demonstrated. In conclusion, we were not able to prove that OMVs could fuse with and transform plasmid in E. coli BL21(DE3) cells.

Effect of OMVs on Mammalian Cells

Key findings:

• Minimal cytotoxic effects of OMVs on A549 lung cancer cells

Aim:

To quantify and evaluate the cytotoxic effect of OMVs on A549 lung epithelial cells, we are using two assays: SYTOX Green and MTT.

The SYTOX Green assay measures the amount of dead cells when treated with OMVs. In contrast, the MTT assay aims to observe effects on general metabolism and measures living cells. By combining the two methods, we are able to support the clinical perspective.

Cytotoxicity measurement with the SYTOX Green Assay

Experimental Setup:

Cytotoxicity of OMVs was analyzed using A549 lung epithelial cells cultured in DMEM (dulbecco’s modified eagle’s medium) with 10% FBS (fetal bovine serum), at 37°C and 5% CO₂10⁴ cells were transferred per well, into a 96-well tissue-culture plate, the day before the cytotoxicity experiment.

The following day, media was replaced by OMV samples in DMEM without the phenol red pH indicator. Positive and negative controls were done with just water and DMEM respectively, further controls were made using PBS. The samples are according to the following scheme in Figure 90.

Cells were incubated with OMV samples for 24 hours, after which 0.5 µM SYTOX Green was added to the cells. Measurements were performed using an excitation wavelength of 483 nm and an emission wavelength of 523 nm. Technical duplicates of the experiment were performed.

Results:

SYTOX Green stains cells with reduced membrane integrity, quantification of the resulting fluorescence allows for the evaluation of damage caused by OMVs (Figure 91). Cells treated with 100% water exhibited the highest fluorescence intensity, confirming the sensitivity of the assay to complete membrane disruption. Cells incubated with 100% PBS showed the second-highest fluorescence intensity. This suggests some baseline level of membrane permeability or cell stress. Cells treated with varying volumes of OMVs (5 µL, 20 µL, and 30 µL) demonstrated similar fluorescence intensities compared to cells cultivated in DMEM. The differences in fluorescence intensity were small and within the range of experimental variation.

These results seemingly indicate the OMVs do not significantly affect membrane integrity of A549 cells.

Based on the Sytox Green assay results, we can conclude that our OMVs do not induce substantial membrane damage in A549 cells at the tested concentrations. However, the observed differences are subtle, and further replication and statistical analysis would be necessary to confirm any potential effect of OMVs on mammalian cell membranes.

A factor that needs to be taken into account is, the assay with SYTOX Green measures the cytotoxicity which refers to the ability of a substance to damage or kill cells. What is not able to quantify is the viability of how many cells are alive and functioning. Meaning, OMVs may exert minimal influence on apoptosis of the mammalian cells, but on the viability of cells.

Measure cell viability with MTT assay

Experimental Setup:

For the assay, an amount of 10⁴ A549 cells 100 µL DMEM were seeded per well in a 96 well plate and incubated overnight at 37°C and 5% CO₂. Different amounts of OMV and PBS (solvent control) were added to the cells and incubated overnight at 37°C and 5% CO₂. MTT assay was carried out according to the instruction manual. Measurement is performed in the plate reader with absorbance measured at 570 nm and a baseline at 650 nm. To eliminate standard errors, duplicate measurements were performed.

Results:

The viability of A549 cells treated with OMVs ranged from approximately 75% to 100% compared to untreated control cells. We observed a slight dose-dependent decrease in A549 cell viability as OMV concentration increased (Figure 92). The maximum effect on cell viability was about 25% reduction at the highest OMV concentration tested. The influence of PBS can be excluded, as there is almost no effect with 30 µL PBS. Only when adding 100 µL of PBS, this leads to a high mortality of about 60%.

These results suggest that OMVs have a moderate impact on A549 cell viability at the concentrations tested.

Conclusion:

The results of the MTT assay indicate that OMVs affect A549 cell viability in a dose-dependent manner, with a maximum reduction of 25% at the highest concentration tested. Since reduced cell viability is caused by stress to the epithelial cells, there must be components in our OMV solution that cause low-level stress to the cells. One reason for this could be the peptidoglycan layer still present on the surface of the OMVs. Our BCA assay confirmed high total protein in the purified OMV solution.

For our future approach, we need to find a concentration where the OMVs are able to fuse with the target cells in sufficient amounts, but do not alter the viability of the lung cells. This will require further experiments.

General Discussion

Our goal was the delivery of AMP encoding plasmids by utilizing the natural propensity of gram negative bacteria to exchange genetic material via OMVs. To this end, we have established standardized methods for the isolation and characterization of OMVs. We have also been able to constitutively express eCPX-SpyTag on the OMV surface, which we have also successfully functionalized with our targeting ligand; SpyCatcher-P2 fusion protein.

We were able to show the presence of plasmids within isolated OMV solutions, but could not show plasmid encapsulation.

Likewise, we were also not able to illustrate the plasmid-delivery capabilities of OMVs in the performed experiments but were able to prove the ability of OMVs to both encapsulate antibiotic degrading enzymes and utilize said enzymes to degrade antibiotics.

With regards to biosafety, we were able to successfully show the non-toxicity of our OMVs when used against human epithelial cells in low concentrations. Given more time, precise delivery of encapsulated plasmids via OMVs could be demonstrated.

References

---