Our lab in the Institute of Molecular Systems Biology at ETH Hönggerberg is specialized in investigating metabolic processes in bacteria, yeast and mammalian cells. Therefore, we had access to a wide range of instruments, machinery and tools, as well as chemicals. There are no biosafety cabinets in our lab, however. We carried out all sterile work under the laminar flow hood to ensure no contamination from the air.
It is not uncommon for bioengineering projects to involve the use of dry ice or liquid nitrogen for certain steps of the project. In our project, we managed to avoid the use of either, to the relief of our lab supervisors and neighbors, avoiding potential severe frost injuries when handling items in extremely low temperatures. However, working with samples (old glycerol stocks, competent cell stocks) and reagents (e.g. Gibson mix, certain enzymes), stored in freezers at -80 and -70°C, for which we used special cryo-gloves.
Changing the big minus to a big plus - we’re referring to the autoclave here - we received appropriate training from our lab technician Dr. Emmanuel Matabaro for the sterilization of liquids, glassware and waste (including antibiotics).
The chemicals that we used from the shelves and cupboards of our lab’s media kitchen are not sterile. To ensure no contamination in media and other solutions prepared by ourselves, we either autoclaved the solution before use, or used filter sterilization. Most of our media, including all LB and YPD, as well as PBS, was provided by the building’s biologics shop. That way we would reduce our use of the autoclave and ensure consistent quality of our media. M9 and MSG -ura media were the only ones that we made ourselves.
The main chemical our project revolved around was hydrogen peroxide, which we diluted weekly or sometimes even daily from a 30% stock. This concentration is very dangerous to living organisms, bleaching and damaging the skin barrier, and we made sure to only interact with the bottle and its contents while wearing gloves. We never took more than 100 ul of the concentrate and always diluted our stock to a 100 times weaker concentration before we started further dilutions to add to our Tecan assays.
In the gel electrophoresis of our PCR products, plasmid digests and other DNA fragments, we used the SYBR® Safe dye, which is much less hazardous than ethidium bromide and the SYBR® Green dye. Still, we made sure to wear gloves and keep containers for storing gels and their components away from other material. Our “gel and PCR room” is also the space for working with mammalian cells, and any live bacterial or yeast material is strictly prohibited from entering the room.
Our lab in ETH Hönggerberg is strictly a BSL1 lab, which means that certain organisms such as Pantoea agglomerans and Erwinia amylovora would be excluded from having any live role in our project as they are classified as BSL2 organisms in Switzerland. Despite that, we managed to find organisms that suit our project’s needs and do not pose a safety risk to us and our lab.
We chose to work with the E. coli strain DH5α for all cloning purposes - this means that all of our Gibson-assembled and ligated fragments and plasmid were first inserted into E. coli DH5α, isolated from overnight cultures via miniprepping and inserted into all other organisms that we were intending to transform. The DH5α strain was developed specifically for use in cloning and plasmid retention, and is generally considered safe to handle. Although we did try out making our own competent cells (see Protocols), we found much more success in using commercial-grade competent DH5a cells from NEB.
E. coli Nissle 1917 (often abbreviated as EcN) is not only an established probiotic, but was successfully transformed with our constructs. This strain lacks virulence factors, making it a useful tool in fighting pathogens without risking an unintentional infection by EcN itself[1]. Its resilience in the intestinal microenvironment is heavily reliant on the ability to scavenge iron as the mechanism of colonization and way of outcompeting other pathogens such as Salmonella, combined with the production of various microcins to ensure decreased colonization of intestinal pathogens. Its short O6 polysaccharide side chain ensures low immunotoxicity, yet doesn’t stop the strain from retaining immunomodulating properties[2].
The fact that we were able to successfully engineer and measure the construct’s activity in EcN gives promise for the potential use of our sensor system in real-life applications for treating various intestinal diseases that are characterized by high levels of ROS. Our construct could allow for versatile ROS-treating “add-ons” to an already powerful pathogen-fighting microbe.
P. protegens Pf-5 has been worked on extensively at ETH, where we received protocols for cultivation details. It is a bacterium known to colonize seeds and roots, so it could be an optimal chassis to use against pathogens that affect plant roots or that can generally be found in/on the ground. Although we did successfully transform the OxyR-based construct into this bacterium in the lab, we did not manage to obtain enough measurements to see clearly whether the construct works in this strain.
S. cerevisiae BY4741 is our chosen yeast strain for the eukaryotic proof of concept. This organism has deletions in genes responsible for the production of the amino acids histidine, leucine, methionine and the nucleic uracil, making it very reliant on outside resources. The GFP-X0 plasmid enables the production of uracil, which we use as the selection marker for cultivation in minimal media with amino acids. The likelihood of an engineered BY4741 strain yeast escaping the lab and surviving against competition in the wild is expected to be low due to the additional deficiencies in amino acid production.
As this is a designer strain is mostly used in functional genomics studies and has no real-life application[3] for our eukaryotic proof of concept, and, once more measurements confirm the effectiveness of the mScarlet-X0-based ROS-sensing constructs, would be engineered into different fungi that have been suggested to us by experts.
Final product considerations - which strains are known to (a) fare well under oxidative stress, (b) not be outcompeted easily by pathogens and other members of the microbiome (c) are expected to not stay alive for long in the chosen environment.
Research done in the Maurhofer lab at ETH suggests that P. protegens Pf-5, although certainly a non-pathogenic option for plant-based applications, might not be resilient against ROS bursts that we are aiming to capture with our sensor. Our measurements so far are insufficient to prove or disprove this information, at least not for the specific concentrations we are measuring.
Based on discussions with experts as well as literature research, the most promising bacterial strains for use in biocontrol applications, other than P. protegens Pf-5 are:
- Pantoea agglomerans strain E325 and P10c, already used in the commercial products Bloomtime Biological and Blossom Bless, respectively[4];
- Pantoea vagans strain C9-1W ("white")[5], a substrain of the C9-1 strain with a characteristic greyish-white color that is a result of the loss of a megaplasmid containing carotenoid production genes. Due to the inability to produce vitamin B1 (thiamine), its lifespan is limited without supplementation. If combined with a CRISPR-based kill-switch repressed by a TPP (thiamine pyrophosphate) riboswitch, this could ensure that the engineered DNA in this organism couldn't be taken up in a functional form by other members of the microbiome[6].
Additionally, we have considered fungal chassis organisms:
- Saccharomyces cerevisiae strain BCA61, mainly for antifungal applications.
- Aureobasidium pullulans strain C10, C40 or M13 for antifungal and antibacterial applications. These strains produce VOCs (volatile organic compounds) that have been shown to be active against a variety of fungal and bacterial plant pathogens. The strains C10 and C40 are used as a mixture in the commercial biocontrol agent BlossomProtect, which seems to have an immunomodulating effect on plants, leading to systemic acquired resistance to the fireblight pathogen E. amylovora[4].
Originally, we intended on attempting to engineer a A. pullulans strain, however, after difficulties with S. cerevisiae, this part of our project had to be scrapped.
As already elaborated above, the Nissle 1917 strain can already be potentially used in probiotic-based treatments of chronic inflammatory and other high-ROS-characteristic diseases due to its nonpathogenic nature. Further biocontainment measure could include adding a CRISPR-based kill-switch that gets activated by a combination of decreased temperature the ingestion of an antibiotic derivative, as described in the paper by Rottinghaus et al.[7]
Our developed sensor system detects high concentrations of reactive oxygen species, and we did not identify potential hazards in its function alone. Our tests of adding the catalase enzyme as an output in the OxyR construct in E. coli DH5a and Nissle show that our OxyR-based sensor is capable of capping its fluorescence output at lower values when combined with catalase compared to without catalase. However, we do not know what could arise combining this sensor with molecules that serve to heighten the release of reactive oxygen species or cause other biochemical phenomena which result in an increase of oxidative stress in the cell. We recognize that our developed sensor, if combined with toxic biomolecule expression, might be used to selectively cause more damage to a host organism experiencing oxidative stress.
Bibliography
- Schultz M, Burton J.P. (2017) The Microbiota in Gastrointestinal Pathophysiology: Implications for Human Health, Prebiotics, Probiotics, and Dysbiosis p.59-69, Elsevier.
- Sonnenborn U, Schulze J. (2009). The non-pathogenic Escherichia coli strain Nissle 1917 – features of a versatile probiotic. Microbial Ecology in Health and Disease, 21(3–4), 122–158. https://doi.org/10.3109/08910600903444267
- Baker Brachmann, C. et al. Designer deletion strains derived from Saccharomyces cerevisiae S288C: A useful set of strains and plasmids for PCR-mediated gene disruption and other applications. Yeast vol. 14 115–132 (1998).
- Zeng, Q. et al. Aureobasidium pullulans from the Fire Blight Biocontrol Product, Blossom Protect, Induces Host Resistance in Apple Flowers. Phytopathology® vol. 113 1192–1201 (2023).
- Smits, T. H. M. et al. Genomic and phenotypic characterization of a nonpigmented variant of Pantoea vagans biocontrol strain C9-1 lacking the 530-kb megaplasmid pPag3. FEMS Microbiology Letters vol. 308 48–54 (2010).
- Miranda-Ríos, J. The THI-box Riboswitch, or How RNA Binds Thiamin Pyrophosphate. Structure vol. 15 259–265 (2007)
- Rottinghaus, A.G. et al. Genetically stable CRISPR-based kill switches for engineered microbes. Nat Commun 13, 672 (2022).