gruppen-photo

Engineering

abstract-icon Abstract

With our project, we aim to establish a cell-free endolysin expression platform that enables the reliable and rapid identification of lead variants, particularly against Gram-negative bacteria. In addition, we seek to develop a toolkit based on this system to facilitate endolysin modification and the generation of target-optimized variants.

Considering the potential application of our endolysins in human infections, we identified biofilms as a major obstacle and therefore aimed to disrupt its integrity via a synergistic approach employing phage-based endolysin transport in combination with antimicrobial and signal-modulating peptides.

AMPs (antimicrobial peptides) can also enable endolysins to diffuse through the OM (outer membrane). This process is still not well understood, thus, we developed AMP-MD, a molecular dynamics (MD) platform to study how AMPs interact with the OM.

To successfully implement our research idea, we used the engineering cycle based on the design-build-test-learn (DBTL) model.

Endolysin Expression Platform

Cytosolic endolysin constructs

As every project ours too started with a design phase. Here we created our experimental plasmid design and chose our endolysin candidates. With a clear outline of that, a project blueprint was established comprising all characteristics of the biological system and parts. This was then implemented in the build phase where we assembled and verified our plasmids and started the protein expression. Subsequently, the expression samples of our endolysins were tested with regard to antimicrobial activity against diverse organisms. With the testing results in at hand we could learn from the generated data and use the gained insights for further improvement and shaping of our project in the next iterations of this engineering cycle.

eng-cycle

Design: Experimental & Plasmid design

During the preliminary design phase, we selected the endolysin PlyEc2 (bba-25h8ekrx). This phage lysin demonstrates promising cell lysis for a broad spectrum of Gram-negative organisms, including Pseudomonas aeruginosa, Acinetobacter baumannii, and Escherichia coli (Xu et al., 2021). Consequently, it served as an ideal starting point for developing our production system using a cell-free expression kit.

In the selection of a cell-free expression kit, the ALiCE® kit demonstrated notable potential. This kit is based on the crude lysate of Nicotiana tabacum Bright Yellow 2 (BY-2) cells and exhibits a high amount of producible protein (Buntru et al., 2014; LenioBio®, 2024). Since protein expression in bacteria is often difficult for endolysins due to their toxic properties (Cremelie et al., 2024) cell-free expression with high yield, like ALiCE®, enables a promising alternative. Furthermore, the kit was kindly provided to us by the AG Prüfer, rendering a promising and accessible expression system, also for establishing our planned toolbox, the details of which will be explained in the AMP-Endolysin fusion constructs section.

During the first cycle, our primary objective was to assess the potential of producing functional protein using ALiCE®. To accomplish this, we designed the necessary plasmids and decided on growth inhibition assays as our primary screening method.

Additionally, the necessary preparations were made to work with the biosafety level 2 (BSL-2) strain of Pseudomonas aeruginosa PAO1 since growth inhibition by PlyEc2 had already been described for this model strain (Xu et al., 2021). P. aeruginosa PAO1 was deliberately selected because strain-specific differences of the endolysin would limit the significance of results obtained with other S1 Pseudomonas strains. Furthermore, our project contributes to the fight against antibiotic resistance, with ESKAPE pathogens being particularly relevant. In front of this background, the inclusion of a clinically relevant strain such as P. aeruginosa was of utmost importance.

Build: Cloning of expression plasmids and subsequent protein expression

The pALiCE01 standard vector was utilized as a backbone in cloning PlyEc2 (Figure 1), as this endolysin is neither membrane-bound nor subject to post-translational modifications, theoretically enabling the cytosolic production of our protein of interest (POI), using the ALiCE® system (LenioBio®, 2024).

For the assembly, we used restriction-ligation cloning, as the supplied plasmid is optimized for this. We utilized E. coli Turbo for transformation and plasmid amplification. The nucleotide sequence of the endolysin was ordered as a G-block from IDT. View the results page to discover more about our exact cloning procedures.

fig8

Figure 1: Plasmid map of the pALiCE01_PlyEc2 plasmid.

After successful plasmid sequencing, we purified the plasmids using phenol-chloroform purification to ensure a high degree of purity, essential for the ALiCE® system. We then started with expressing our desired proteins.

Test: Growth inhibiton assay with Pseudomonas aeruginosa PAO1

A growth inhibition assay was conducted with Pseudomonas aeruginosa PAO1 to assess the functionality of the produced endolysin (Figure 2). No growth inhibition was observed for the PlyEc2 reaction supernatant or for the non-template reaction control (negative control). All the assay variants exhibited comparable growth to that of the positive control without any added component.

fig8

Figure 2: Initial growth inhibition assay of PlyEc2 on Pseudomonas aeruginosa PAO1. PlyEc2 supernatant (SN) with an estimated concentration of 20 µg/ mL and 60 µg/ mL were tested on 106 P. aeruginosa PAO1 cells. Additionally, the non-template controls (NTC) and the LB control were used as negative controls. LB + P. aeruginosa PAO1 itself and in combination with gentamecin were utilized as positive control. Cells were grown in LB medium, at 37 °C and 250 rpm.

Learn: Findings for the next cycle iteration

Contrary to our expectations, our initial tests revealed no signs of growth inhibition and the lack of a reliable and feasible detection method for our non-tagged PlyEc2 prevented the determination of the exact protein concentration. Initially, we were thus unable to rule out an insufficient protein concentration, its total absence, or the lack of endolysin functionality despite it being present.

Based on these findings, we decided to integrate a StrepII-tag into our vector constructs in order to be able to confirm the absence or presence of our protein and to clarify the concentration of our POIs.

In addition to the possibility that our lysin was not produced at all, or at least not produced functionally, we also investigated whether the strain specificity of the tested endolysin could be a possible cause for the lack of growth inhibition.

StrepII-tag endolysin constructs
eng-cycle

Design: Experimental & Plasmid design 

The modified pALiCE01 vector now contained a StrepII-tag combined with a (GGGGS)x2 linker. This vector was additionally adapted for both restriction-ligation and Golden Gate cloning, provided that specific primers are utilized.

Furthermore, we decided to employ an additional endolysin, designated as PlyYouna2 (bba-25uoafah). This endolysin originates from the Youna phage and is specialized for the lysis of Gram-positive bacteria, including Bacillus megaterium and B. subtilis. However, studies demonstrate its additional efficacy in the lysis of specific Gram-negative organisms, including e.g., E. coli or Pseudomonas putida (Son et al., 2023).

While the lysis of Gram-positive bacteria by externally applied endolysins is specific, it is facilitated by the lack of an outer membrane, which simplifies access and disruption of the peptidoglycan layer from the exterior (Zheng and Zhang, 2024). The bacterial strains utilized in our experiments are BSL-1 strains, which can be easily integrated into our standard laboratory protocols due to their RG1 classification and their well-established utilization in neighbouring working groups.

Additionally, we decided to switch to plate spot assays instead of growth inhibition assays as the initial screening method, to enable a faster but still reliable workflow.

Build: Cloning of the plasmids and protein expression

Using the modified pALiCE01 vector including a StrepII-tag and the (GGGGS)2 linker we cloned the following plasmids for each endolysin via Golden Gate Cloning (Figure 3):

  • pALiCE01_StrepII_PlyEc2
  • pALiCE01_StrepII_PlyYouna2
fig8

Figure 3: Plasmid map of A) pALiCE01_StrepII_PlyEc2 and B) pALiCE01_StrepII_PlyYouna2.

The correct assembly was confirmed by plasmid purification, followed by sequencing. Afterwards protein expression via the ALiCE® system was performed.

Test: Plate spot assays and validation of purification and expression methods

Initially, our SDS-PAGE results did not allow a clear determination of the presence or absence of our endolysins due to contamination of the non-template control (NTC), which was of the same expected size as that of the endolysins.

Additionally, the plate spot assay for PlyYouna2 on B. megaterium revealed lysis in both the supernatant and the NTC. However, lysis was no longer detectable after native protein purification with the MagStrep Strep-Tactin™ XT Beads. The same was observed for PlyEc2 on P. aeruginosa (Figure 4).

fig8

Figure 4: Plate spot assay of PlyYouna2 on Bacillus megaterium (A) and of PlyEc2 on Pseudomonas aeruginosa PAO1(B). For A) the supernatant (SN), pellet and natively purified PlyYouna2 as well as 50 µg/µL ampicillin, the non-template control (NTC) supernatant and the elution buffer of the purification (1x BXT) were tested for lysis. For B) the PlyEc2 supernatant, 50 µg/µL ampicillin and the non-template control (NTC) supernatant were tested regarding lysis.

Repeating the SDS-PAGE followed by Western blot revealed that our endolysin had not been produced (Figure 5).

To ensure the functionality of the ALiCE® kit, another Western blot was performed to verify the successful production of the positive control protein (eYFP) using our StrepII-tag (Figure 5). This experiment herein also confirmed the successful purification of proteins, such as the labelled eYFP, using the MagStrep Strep-Tactin™ XT Beads. Furthermore, phenol-chloroform plasmid purification was excluded as a potential causative agent for our failed endolysin expression. Plasmids purified using this method could be used effectively to produce and purify eYFP successfully.

fig8

Figure 5: Ponceau staining on the left and Western blot on the right. A) shows the results of the StrepII-tagged eYFP derived from the standard pALiCE01 plasmid already delivered purified. Further it exhibits the PlyEc2 samples of the whole lysate (WL), supernatant (SN), resuspended pellet (P), and the natively purified protein (A). B) shows the results of the StrepII-tagged eYFP derived from the standard pALiCE01 that was previously purified with the phenol-chloroform purification method. It contains the WL, SN, P and A as well as the input (IP) and the wash of the native protein purification. Additionally, IP and wash of PlyEc2 and the non-template control (NTC) are visible. The NTC is an ALiCE® reaction sample where no plasmid was used as template. It exhibits the negative control of the ALiCE® reaction. Further, the SDS-PAGE gel of B) contained a leaky pocket. As a result, a sample flowed across the gel, resulting in the slight continuous staining visible in the WB.

Learn: Findings from the troubleshooting process

During the troubleshooting process, the functionality of the ALiCE® kit was confirmed, as well as the efficacy of the phenol-chloroform plasmid purification and the MagStrep Strep-Tactin™ XT Beads purification. Furthermore, the accuracy of the plasmid assembly was again validated through sequencing.

Through advanced research and consultation, we discovered that the lysis, which was observed in the NTC, could be attributed to antibacterial substances present in the lysate of the tobacco root cells utilized in the kit (Al-Lahham et al., 2020).

Finally, this indicated that the kit cannot be used to establish an adequate negative control for the expression of antimicrobial proteins. Consequently, we deemed the kit to be unsuitable for our intended application. However, the implementation of an alternative cell-free kit based on a prokaryotic lysate, which was actually planned since May, was hindered by the protracted regulatory process involved, rendering further efforts in this direction infeasible due to time constraints.

AMP-Endolysin fusion constructs

Based on our Integrated Human Practices work, we identified antimicrobial peptides (AMPs) as a possibility for improving endolysins to enhance their effectiveness against Gram-negative bacteria. Because of the broad range of AMPs that could potentially be fused to endolysins, we developed a molecular-dynamics-based (MD) model to understand their insertion mechanism into the outer membrane of Gram-negative bacteria and evaluate which properties may enhance their effectiveness against these bacteria when fused to an endolysin. Based on these studies, we selected the AMPs K14 (BBa-25xu6jmw) and Sushi S1 (BBa-k5057004) to design our fusion constructs. Both the cationic CM15 derivate K14 as well as the Sushi S1 AMP exhibit low or no haemolytic activity (Yau et al., 2001; Ding, Li and Ho, 2008; Sato and Feix, 2008).

For an efficient and rapid assembly, we initially aimed to create a toolbox approach based on the cytosolic ALiCE® vector using Golden Gate Assembly. The individual building blocks, consisting of AMPs, linkers, and endolysins with respective position tags, could then be easily exchanged to effectively create enhanced endolysin variants. In silico designs for both the ALiCE® and pEt28b(+) vectors with these building blocks had already been designed at that point (see Contribution for more details).

However, due to non-reliable protein expression results with the ALiCE® system, the absence of the second cell-free expression kit, and time constraints, we had to stop at this point.

For more information about the model, please visit our model page and view the model cycles.

Biofilm destruction

Identifying the needs for a holistic alternative to antibiotics

Our initial approach aimed to research and develop a comprehensive alternative to conventional antibiotics in light of rising antimicrobial resistance (AMR). During literature review, phage therapy appeared especially promising, as well as phage-encoded endolysins, identified as potent active agents due to their ability to degrade bacterial cell walls and disrupt biofilms. However, besides our work on utilizing endolysins potential, our human practices endeavors also led us to focus on potential for clinical applications in humans. In this regard, biofilms were repeatedly pointed out as key complicating factors during infections caused by resistant bacteria. It became evident that biofilms pose a significant barrier to bacterial control across multiple settings. Thus, in line with our objective of finding a holistic alternative potentially surpassing conventional antibiotics, we expanded our focus to include the targeted eradication and disruption of biofilms in respect to future application.

Theory of biofilm infiltration and destruction by phages and auxiliary substances
eng-cycle

Design:

During our literature research on the targeted destruction of bacterial biofilms, we developed the idea of introducing modified endolysin variants to biofilm-forming bacteria using infiltrating phages. To increase this infiltration capacity of the phages into the biofilm, we planned to fuse biofilm-damaging peptides to their capsid proteins. At the same time, we discovered that pathogenic and non-pathogenic biofilms differ structurally and functionally. To test our concepts on clinically relevant strains like enteroaggregative Escherichia coli, we initiated a collaboration with the Dobrindt research group. We chose M13 phages as the vector, because they are well-established for phage display and enable clear analysis of endolysin effects due to their lysogenic cycle.

Build:

For the implementation, we selected peptide sequences on the condition that they would be suitable for possible later use in humans. Accordingly, we did not allow the selected peptides to be hemolytic or otherwise toxic. We decided to test different modes of action, including antimicrobial peptides (AMPs), anti-biofilm-active peptides (ABPs), and peptides that disrupt bacterial growth signals within the biofilm specifically. We derived some sequences from human proteins in the hope of minimizing immunological reactions in clinical applications. We then designed suitable primers for these peptides and produced the first bacteriophages.

Test:

The production of the phages was successful, as was the DNA extraction using phenol-chloroform purification. However, difficulties arose during DNA amplification of the phage vector using Q5 polymerase for overlap extension PCR, a step that was also intended to enable the cloning of the peptides.

Learn:

After several attempts, we evaluated alternative polymerases for PCR. We finally achieved the desired amplification with the Phusion polymerase. At the same time, we conducted several human practices interviews on the regulatory situation of phages in Germany and decided to continue working with wild-type phages rather than genetically modified (GMO) phages.

Planning of BSL-2 work on pathogenic bacteria and initial testing 
eng-cycle

Design:

Therefore, we decided to focus on the combined effect of the peptides on biofilm destruction, both individually and in combination with wild-type M13 phages. Although these phages are exclusively lysogenic, they can also inhibit horizontal gene transfer, preventing the exchange of resistance genes within the biofilm. In preparation for the upcoming experiments, we examined the legal and safety requirements for working in BSL-2 laboratories. We prepared the necessary documentation, consulted the company doctor, and clarified all necessary aspects of the safety briefings to ensure our personal safety.

Build:

For the first series of tests, we selected peptide concentrations based on the minimum inhibitory concentrations (MIC values) specified in the literature and designed combinations of phages and peptides for use in a 96-well plate.

Test:

The first experiment was performed on a PVC 96-well plate. The biofilm was first cultured for 24 hours, followed by further 24 hours of incubation with the peptide/phage combinations. After staining with crystal violet, the amount of biofilm was quantitatively determined by absorption measurements. All peptides were used at double the MIC concentration.

Learn:

In the first run, only slight biofilm formation was observed. The enteroaggregative E. coli strain 55989 exhibited limited growth, but no statistically reliable conclusions could be drawn from the results.

Testing biofilm destruction and reflecting on the results
eng-cycle

Design:

As the results of the previous test cycle were only of limited significance, we analyzed potential sources of error and possible revisions to the test design.

Build:

During this analysis, we determined that, under certain conditions, biofilms require more than 24 hours to achieve sufficient density and stability for reproducible measurements. We also decided to use an alternative plate material since biofilm adhesion can be influenced by the plastic surface. We also planned to use higher peptide concentrations to quantify a more pronounced effect.

Test:

The optimized experiment was performed on polystyrene 96-well plates. The biofilms were allowed to form for 48 hours before being treated with the peptide/phage combinations. In addition to twice the MIC concentration, we also tested the 10-fold concentrations of the respective peptides from the first testing.

Learn:

The modified test conditions led to significantly more usable results, which we then analyzed in detail. This enabled us to identify plausible explanations for the observed effects. However, the measured standard deviations were still relatively high in some cases. For subsequent cycles, it would be essential to repeat the experiments in order to be able to make more statistically reliable statements. In addition, possible synergistic effects and the adaptability of the biofilm to different environmental conditions should be investigated.

Modeling of AMP-membrane interactions

Our initial goal was to quantify the effectiveness of different antimicrobial AMPs based on their insertion free energy into the membrane through Umbrella Sampling. This approach failed because of the sensitivity of Umbrella Sampling in large systems. Therefore, we decided to focus more on understanding the mechanism of the insertion of peptides into the membrane. This new approach of our modeling led to exciting results, and our model could even be used to explain phenomena in our wetlab.

Umbrella Sampling
eng-cycle

Design:

Our first idea was to quantify the effectiveness of different antimicrobial AMPs based on their insertion free energy into the membrane through Umbrella Sampling.

Build:

We built the membrane system (E. coli R3 O111/CM15) with CHARMM-GUI (Jo et al., 2008) and inserted the CM15 peptide on top of the membrane in the extracellular space.

Test:

Then we performed a pull simulation and extracted the Umbrella Sampling starting windows. For each starting window, a biased simulation was performed with a harmonic potential restricting the peptide in the z direction. From the biased distributions of each umbrella window, the insertion free energy was calculated. We also tested different starting configurations, yielding different free energies.

Learn:

A quantitative description of the insertion free energies from different peptides in the same membrane is not possible with umbrella sampling due to the high sensitivity of umbrella sampling to small changes in the starting configurations.

Unbiased simulations
eng-cycle

Design:

After our failed goal to quantify the effectiveness of different AMPs with umbrella sampling, we decided to focus on the mechanism of AMP insertion into the membrane and how it increases the membrane permeation.

Build:

We built a new membrane/peptide (E. coli R3 O111/6 x CM15) system and equilibrated the system.

Test:

We performed unbiased simulations of the system and observed a free insertion into the outer membrane. Further, we studied the interactions between the peptide and outer membrane and quantified the increase of membrane permeation with the “PackMem” package (Gautier et al., 2018).

Learn:

The unbiased simulations helped us understand the mechanism of membrane insertion of peptides. We observed an increase in lipid packing defects after the peptide was inserted into the Core-saccharide region of the membrane.

References

Al-Lahham, S. et al. (2020) “Antioxidant, antimicrobial and cytotoxic properties of four different extracts derived from the roots of Nicotiana tabacum L.,” European Journal of Integrative Medicine, 33, p. 101039. Available at:
https://doi.org/10.1016/j.eujim.2019.101039 .

Buntru, M. et al. (2014) “Tobacco BY-2 cell-free lysate: an alternative and highly-productive plant-based in vitro translation system,” BMC Biotechnology, 14(1), p. 37. Available at:
https://doi.org/10.1186/1472-6750-14-37 .

Cremelie, E., Vázquez, R. and Briers, Y. (2024) “A comparative guide to expression systems for phage lysin production,” Essays in Biochemistry, 68(5), pp. 645–659. Available at:
https://doi.org/10.1042/EBC20240019 .

Ding, J.L., Li, P. and Ho, B. (2008) “The Sushi peptides: structural characterization and mode of action against Gram-negative bacteria,” Cellular and Molecular Life Sciences, 65(7), pp. 1202–1219. Available at:
https://doi.org/10.1007/s00018-008-7456-0 .

Gautier, R. et al. (2018) “PackMem: A Versatile Tool to Compute and Visualize Interfacial Packing Defects in Lipid Bilayers,” Biophysical Journal, 115(3), pp. 436–444. Available at:
https://doi.org/10.1016/j.bpj.2018.06.025 .

Jo, S. et al. (2008) “CHARMM-GUI: A web-based graphical user interface for CHARMM,” Journal of Computational Chemistry, 29(11), pp. 1859–1865. Available at:
https://doi.org/10.1002/jcc.20945 .

LenioBio® (2024) “Instruction Manual for Cell-Free Protein Expression with ALiCE® for Research.”

Sato, H. and Feix, J.B. (2008) “Lysine-Enriched Cecropin-Mellitin Antimicrobial Peptides with Enhanced Selectivity,” Antimicrobial Agents and Chemotherapy, 52(12), pp. 4463–4465. Available at:
https://doi.org/10.1128/AAC.00810-08 .

Son, B. et al. (2023) “Isolation and Characterization of a Weizmannia coagulans Bacteriophage Youna2 and Its Endolysin PlyYouna2,” Journal of Microbiology and Biotechnology, 33(8), pp. 1050–1056. Available at:
https://doi.org/10.4014/jmb.2303.03021 .

Xu, S. et al. (2021) “Decontamination of Escherichia coli O157:H7 on fresh Romaine lettuce using a novel bacteriophage lysin,” International Journal of Food Microbiology, 341, p. 109068. Available at:
https://doi.org/10.1016/j.ijfoodmicro.2021.109068 .

Yau, Y.H. et al. (2001) “High Therapeutic Index of Factor C Sushi Peptides: Potent Antimicrobials against Pseudomonas aeruginosa,” Antimicrobial Agents and Chemotherapy, 45(10), pp. 2820–2825. Available at:
https://doi.org/10.1128/AAC.45.10.2820-2825.2001 .

Zheng, T. and Zhang, C. (2024) “Engineering strategies and challenges of endolysin as an antibacterial agent against Gram-negative bacteria,” Microbial Biotechnology, 17(4), p. e14465. Available at:
https://doi.org/10.1111/1751-7915.14465 .

Results
Description
Experiments
^