Researching about microbial bioremediation, we quickly came across the Pseudomonas genus. Multiple Pseudomonas species show very promising abilities as chassis for biodegradation of pollutants [1]. They are known to be stress resistant, providing us with a resilient chassis to modify. Additionally, they show great metabolic diversity and promising abilities to degrade PAHs [2]. Naphthalene degradation, for instance, is studied best in Pseudomonas [3]. On top of that, they populate marine and freshwater environments ubiquitously [2, 4], meaning our chassis will be comfortable living in riverine conditions.
A number of Pseudomonas species show impressive PAH degradation abilities, like Pseudomonas fluorescens, putida or stuzeri [5]. We decided to work with Pseudomonas vancouverensis DSM 8368 (aka. NCIB 9816, formally listed as Pseudomonas putida, from now on referred to as P. vancouverensis) [6, 7]. P. vancouverensis degrades naphthalene so efficiently, that it can even feed on it as sole carbon source. On top of that, it completely metabolizes fluorene and phenanthrene [8]. This makes it attractive as a degrader for a large variety of PAHs, effectively cleaning rivers from multiple pollutants.

As P. vancouverensis is a little studied, non-model organism, all experiments were carried out with Pseudomonas putida KT2440 [9] as well. As P. putida is intensively studied as chassis for bioremediation [10], it served to us as a reference organism. We would like to thank Prof. Victor de Lorenzo for kindly providing us this strain.

As PAHs are highly stable and poorly soluble in water, biodegradation is difficult. The degradation pathway’s initial step incorporates an oxygen molecule into the compound, breaking aromatic resonance and increasing solubility in water. This reaction is catalyzed by a dioxygenase enzyme and subsequently cis-dihydro-diol intermediates are created (see Figure 2) [5]. This first step is crucial for biodegradation, as aromatic resonance gets broken, and a more reactive compound is formed. In addition, the product’s higher water-solubility facilitates further degradation [11].
Regarding naphthalene degradation, in this way naphthalene is converted to cis-1,2-dihydro-naphthalene-1,2-diol. Next a dehydrogenase creates 1,2-dihydroxynaphthalene. Reaction cascade leads further via salicylate, which is subsequently fully degraded to tricarboxylic acid cycle (TCA) intermediates [5] (see Figure 2).

Analogous phenanthrene gets oxidized to cis-3,4-dihydro-phenanthrene-3,4-diol first, which is then converted into 3,4-dihydroxy-phenanthrene. Most likely in Pseudomonas further oxidation and ring cleavage create 1,2-dihydroxy-naphthalene thus leading into naphthalene degradation pathway. [12, 13] (see Figure 3)

In this project we introduced a pyrene degradation pathway in P. vancouverensis and linked it to its native phenanthrene pathway [14]. Similar to naphthalene- and phenanthrene-pathway, in pyrene-degradation the important initial step creates cis-4,5-dihydro-pyrene-4,5-diol [15]. This step is catalyzed by nidA and nidB, a rieske non-hemen iron ring hydroxylating oxygenase [16]. Though for its activity phtAc and phtAd from phthalate degradation are needed as extern electron carrier [15]. The pathway proposed in [15] shows further degradation involving one ring-cleavage leading to 3,4-dihydroxy-phenanthrene (see Figure 4).

3,4-dihydroxyphenanthrene is an intermediate of P. vancouverensis’ native phenanthrene pathway (see Figure 3). In this way pyrene degradation gets channeled into our chassis’ native metabolic pathways. Thereby we only need to genetically introduce nine protein coding sequences forming five enzymes/enzyme complexes. As complete pyrene degradation includes around twenty enzymes [15], this is crucial to make our project feasible and prevent genetic overload.

To express genes necessary for pyrene degradation, we decided to arrange all nine protein coding sequences in one operon. In this way sequence lengths can be shortened, as transcriptional regulators are only needed once. All proteins will be translated from one mRNA-molecule resulting in spatial proximity to each other. This might increase the degradation pathway’s efficiency, as intermediate’s diffusion paths from one enzyme to the other are shrunk [17]. A scheme of the operon is displayed in figure 5.

Protein coding sequences were mainly taken from Mycobacterium vanbaalenii Pyr-1, the first bacterium reported to completely break down pyrene [18]. As shown in [15], for some proteins already characterized homologous proteins in other bacteria exist. Therefore some proteins in our pathway were derived from different organisms. We decided to add a hexahistidine-tag to every protein, to enable easy protein production verification via western-blot using just one primary antibody [19].
Initially we planned to use a strong promoter, hoping to thereby get high protein concentrations and hence good degradation rates. Yet Professor Victor de Lorenzo from CSIC advised us to use a low or medium constitutive promoter as PAH concentrations will be low and protein production costs lots of resources for our chassis. Therefore we decided to use Anderson promoter J23110 which shows medium expression in Escherichia Coli and Pseudomonas [20]. In this way the promoter shall be the rate limiting part for protein production. Hence, we decided to use an RBS showing good translation efficiency, namely BBa_B0034 [21]. As a terminator we decided to use phage derived Luz7-T50, which shows good efficiency in E. Coli and Pseudomonas as well [22]. We used pSEVA231 from SEVA-collection as plasmid backbone [23], providing broad-host-range origin pBBR1 and kanamycin resistance marker, as advised by Prof. de Lorenzo. The backbone was generously provided to us by Prof. de Lorenzo. (See our parts page)

Genes were ordered from IDT. RBS and his-tags were added via overhang-PCR. Iterative assembly was carried out using GoldenBraid method [24]. We used Colabfold as protein structure prediction tool, to evaluate if the his-tag would be best placed at N- or C-terminus, in order to not interfere with protein’s native structure [25].
To make the pSEVA-backbone GoldenBraid-compatible, a DNA-fragment was ordered from IDT carrying promoter, lacZ-cassette, terminator as well as GoldenBraid-BsaI-restriction sides. By cloning of this fragment into the backbone, pSEVA231 served as a pseudo-GB-alpha-vector. A scheme of our cloning procedure is displayed in figure 6.

We would like to thank M. Sc. Jacob Mejlsted for helping us planning the cloning procedure.

Utilizing electroporation, we successfully introduced the ligated SEVA plasmid into Pseudomonas vancouverensis. After carefully evaluating various protocols, we selected the most effective. The transformation was confirmed when the modified bacteria grew on LB medium containing 50μg/ml Kanamycin. As negative control wildtype cells without plasmid were plated out (see Figure 7). In addition, plasmid miniprep and control digest using BamHI was carried out to verify correct plasmid uptake.

At first, we tested our expression system consisting of promoter J23110, RBS B0034 and terminator Luz7-T50. For this purpose, amilGFP as reporter protein was used (see BBa_K5043011). Protein production could be verified by observing green fluorescence of transformed cell pellets (see Figure 8).

The transformed Pseudomonads strains were then characterized regarding their growth, pyrene minimal bactericidal concentration (MBC) and pyrene degradation using HPLC analysis.
However the HPLC-assay of liquid cultures does not indicate successful pyrene degradation (see Figure 9). For details visit our results page.

In addition, growth curves and MBCs for P. vancouverensis do not indicate pyrene degradation. In contrast, P. putida possessing pyrene degradation plasmid shows significantly higher pyrene and phenanthrene tolerance and grows faster than unmodified bacterium, when pyrene is present (see Figure 10). This could be an indication for pyrene degradation into less toxic intermediates.

Pyrene degradation could not be verified in our engineered strains. However, P. putida shows increased tolerance to pyrene and phenanthrene. This could indicate, our plasmid’s enzymes degrade phenanthrene and pyrene into less toxic intermediates in P. putida. Some of our chosen proteins can use a variety of PAHs as substrate [16], which could explain the increased phenanthrene tolerance.

We clearly need to characterize our engineered strains in more detail. At first a reliable verification of the production of all nine proteins is necessary, for example using western blot. Second, more reliable pyrene degradation assays are needed, as our method showed great fluctuations in data. Cultures are also required to be incubated longer, to be sure whether pyrene is degraded or not. In addition, characterization of all our used enzymes in vitro is necessary to verify the integrity of the pyrene pathway.

At the heart of our project lies the genetic engineering of P. vancouverensis to successfully degrade PAHs. However, for our advanced bacteria to work effectively against PAH pollution in rivers, it needs to be brought in close contact with as many PAHs as possible. To achieve this, the drylab team worked alongside the wetlab team throughout the project to develop an effective platform for a successful P. vancouverensis deployment in polluted areas.
One of the main concerns when working on the deployment of GMOs is to comply with all regulations at hand and avoid an uncontrolled release of the GMOs. To achieve a strong degree of containment of the GMOs combined with a high level of exposure to the contaminated water, we came up with the idea to immobilize P. vancouverensis on the inner surface of a specialized device, which could be deployed at suitable sites in rivers, such as bridges, sluices or buoys.

From the beginning, it was clear to us that the interior design of the device, especially the interaction area of our chassis, would play a major role in the effectiveness of our solution.

Design and conceptualize: The membrane concept - a platform combining filtering and degradation of PAHs

The first idea that we got to work featured a membrane within the device. On this membrane, P. vancouverensis would be immobilized. It would also be able to filter the oil-micelles out, letting them accumulate on the same side of the membrane where P. vancouverensis would be placed, while letting other particles in the water pass through the entire system. This way, we could both separate the PAHs from the water and degrade them at the same place and time.

Present and learn: Receiving input from experts at BFH Meetup and scrapping the membrane concept

We presented this design at the BFH Meetup in Bielefeld, where the jury and other experts were not convinced that this system could work. They saw the main problems in finding or developing a suitable membrane as well as in not clogging our system with too many PAHs. They instead encouraged us to look for a solution without a membrane, focusing more on water flow and the maximization of surface area inside the device. After further research into this and the uncertainty of how the immobilization could work on a membrane in the end, we decided to scrap the membrane and work on new concepts.

Design: Labyrinths, Pillars and Balls - maximizing contact efficiency

We came up with three base configurations for an updated interior design of the device. One featured a labyrinth (see Figure 11). The focus of this design was really to get the maximum surface area out of the small inside of our device. If all sides of such a labyrinth were covered in silica, P. vancouverensis would be immobilized by the non-polar interactions of its membranes with the surface. Using the labyrinth structure we could reduce the probability of PAHs flowing through the entire system without ever encountering a bacterium. This would make for a very efficient design.

A second design featured large pillars filling the inside of the device (see Figure 12). P. vancouverensis would in this case be immobilized on the surface of these pillars. We planned on finding the best configuration of pillars through simulation, to make sure that the maximum amount of water would come in contact with one of the pillars at some point. An advantage of this system over the labyrinth would be the much smaller amount of forces the water would exert onto the interior of the device, since the water from the river would not be slowed as much through the pillars as through the harsh turns of a labyrinth.
The third design was based on small beads/balls containing P. vancouverensis (see Figure 13). When researching methods for immobilization, we discovered that many common methods result in permeable beads filled with bacteria. These beads would be free to move around inside the device. The creation of such beads is relatively simple and cost efficient. The materials used to create the alginate beads that turned out to be sufficient for this use case, are stable and non-toxic (learn more in results). A big advantage of this design would be the modularity of it. One could simply open the box and replace broken balls. However, it was unclear how we could stop the balls from simply getting pushed to the end of the box and being stuck there.

Build: Creating 3D-models and printing a prototype

Before we could start our simulations to test the viability of our ideas, we designed them using common computer aided design (CAD) software. The resulting 3D-models were not only the basis for the following simulations, but also gave us the opportunity to 3D-print our first prototype models to learn more about the sizing of our device.

After designing all three structures virtually, we proceeded with a flow analysis simulation. In the simulations we focused mainly on the velocity and pressure of the fluid traversing the device. A low velocity at the contact points with the surface would be ideal for P. vancouverensis to work most efficiently. A constant decrease in pressure would then secure an optimal continuous flow throughout the device.
While the simulations were running, the different methods for immobilization were tested in the laboratory. As the method of immobilizing bacteria on a silica surface was verified, as well as the method of enclosing the bacteria in alginate beads, all three models would work in terms of biological feasibility (see our results page). Therefore, we could rely fully on flow analysis to show us which design would turn out optimal for our device.
Through the simulations we found that the ball design works best for our application. It also features to easy accessibility and replaceability of the beads containing the GMOs. This means that applications for other bacteria, targeting other riverine contaminations, could easily be installed as well. Through these advantages we concluded that the balls make for the best interior design of our device.

Our approach to a real-life flow simulation with a simple 3D-printed model of the pillar design. To make the path the water takes through the box visible, we used fluorescein.

[1] E. Martínez-García and V. de Lorenzo, "Pseudomonas putida as a synthetic biology chassis and a metabolic engineering platform," Current opinion in biotechnology, vol. 85, p. 103025, 2024, doi: 10.1016/j.copbio.2023.103025.
[2] D. C. Volke, P. Calero, and P. I. Nikel, "Pseudomonas putida," Trends in microbiology, vol. 28, no. 6, pp. 512–513, 2020, doi: 10.1016/j.tim.2020.02.015.
[3] K. M. Yen and C. M. Serdar, "Genetics of naphthalene catabolism in pseudomonads," Critical reviews in microbiology, vol. 15, no. 3, pp. 247–268, 1988, doi: 10.3109/10408418809104459.
[4] C. Valenzuela, V. L. Campos, J. Yañez, C. A. Zaror, and M. A. Mondaca, "Isolation of arsenite-oxidizing bacteria from arsenic-enriched sediments from Camarones river, Northern Chile," Bulletin of environmental contamination and toxicology, vol. 82, no. 5, pp. 593–596, 2009, doi: 10.1007/s00128-009-9659-y.
[5] B. Mohapatra and P. S. Phale, "Microbial Degradation of Naphthalene and Substituted Naphthalenes: Metabolic Diversity and Genomic Insight for Bioremediation," Frontiers in bioengineering and biotechnology, vol. 9, p. 602445, 2021, doi: 10.3389/fbioe.2021.602445.
[6] DSMZ, Pseudomonas vancouverensis. [Online]. Available: https://www.dsmz.de/collection/catalogue/details/culture/DSM-8368 (accessed: Aug. 20 2024).
[7] W. W. Mohn, A. E. Wilson, P. Bicho, and E. R. Moore, "Physiological and phylogenetic diversity of bacteria growing on resin acids," Systematic and applied microbiology, vol. 22, no. 1, pp. 68–78, 1999, doi: 10.1016/S0723-2020(99)80029-0.
[8] Y. Yang, R. F. Chen, and M. P. Shiaris, "Metabolism of naphthalene, fluorene, and phenanthrene: preliminary characterization of a cloned gene cluster from Pseudomonas putida NCIB 9816," Journal of bacteriology, vol. 176, no. 8, pp. 2158–2164, 1994, doi: 10.1128/jb.176.8.2158-2164.1994.
[9] M. Bagdasarian et al., "Specific-purpose plasmid cloning vectors. II. Broad host range, high copy number, RSF1010-derived vectors, and a host-vector system for gene cloning in Pseudomonas," Gene, vol. 16, 1-3, pp. 237–247, 1981, doi: 10.1016/0378-1119(81)90080-9.
[10] Z. Zuo et al., "Engineering Pseudomonas putida KT2440 for simultaneous degradation of organophosphates and pyrethroids and its application in bioremediation of soil," Biodegradation, vol. 26, no. 3, pp. 223–233, 2015, doi: 10.1007/s10532-015-9729-2.
[11] T. Pantsyrnaya et al., "Biodegradation of Phenanthrene by Pseudomonas putida and a Bacterial Consortium in the Presence and in the Absence of a Surfactant," Indian journal of microbiology, vol. 52, no. 3, pp. 420–426, 2012, doi: 10.1007/s12088-012-0265-z.
[12] W. C. EVANS, H. N. FERNLEY, and E. GRIFFITHS, "OXIDATIVE METABOLISM OF PHENANTHRENE AND ANTHRACENE BY SOIL PSEUDOMONADS. THE RING-FISSION MECHANISM," The Biochemical journal, vol. 95, no. 3, pp. 819–831, 1965, doi: 10.1042/bj0950819.
[13] S. Gao, J.-S. Seo, J. Wang, Y.-S. Keum, J. Li, and Q. X. Li, "Multiple degradation pathways of phenanthrene by Stenotrophomonas maltophilia C6," International biodeterioration & biodegradation, vol. 79, pp. 98–104, 2013, doi: 10.1016/j.ibiod.2013.01.012.
[14] EPA, Priority Pollutant List. [Online]. Available: https://www.epa.gov/sites/default/files/2015-09/documents/priority-pollutant-list-epa.pdf (accessed: Aug. 21 2024).
[15] S.-J. Kim, O. Kweon, R. C. Jones, J. P. Freeman, R. D. Edmondson, and C. E. Cerniglia, "Complete and integrated pyrene degradation pathway in Mycobacterium vanbaalenii PYR-1 based on systems biology," Journal of bacteriology, vol. 189, no. 2, pp. 464–472, 2007, doi: 10.1128/JB.01310-06.
[16] O. Kweon, S.-J. Kim, J. P. Freeman, J. Song, S. Baek, and C. E. Cerniglia, "Substrate specificity and structural characteristics of the novel Rieske nonheme iron aromatic ring-hydroxylating oxygenases NidAB and NidA3B3 from Mycobacterium vanbaalenii PYR-1," mBio, vol. 1, no. 2, 2010, doi: 10.1128/mBio.00135-10.
[17] R. E. Svetic, C. R. MacCluer, C. O. Buckley, K. L. Smythe, and J. H. Jackson, "A metabolic force for gene clustering," Bulletin of mathematical biology, vol. 66, no. 3, pp. 559–581, 2004, doi: 10.1016/j.bulm.2003.09.008.
[18] A. A. Khan, S.-J. Kim, D. D. Paine, and C. E. Cerniglia, "Classification of a polycyclic aromatic hydrocarbon-metabolizing bacterium, Mycobacterium sp. strain PYR-1, as Mycobacterium vanbaalenii sp. nov," International journal of systematic and evolutionary microbiology, vol. 52, Pt 6, pp. 1997–2002, 2002, doi: 10.1099/00207713-52-6-1997.
[19] W. N. Burnette, ""Western blotting": electrophoretic transfer of proteins from sodium dodecyl sulfate--polyacrylamide gels to unmodified nitrocellulose and radiographic detection with antibody and radioiodinated protein A," Analytical biochemistry, vol. 112, no. 2, pp. 195–203, 1981, doi: 10.1016/0003-2697(81)90281-5.
[20] A. N. Pearson et al., "The pGinger Family of Expression Plasmids," Microbiology spectrum, vol. 11, no. 3, e0037323, 2023, doi: 10.1128/spectrum.00373-23.
[21] S. G. Damalas, C. Batianis, M. Martin-Pascual, V. de Lorenzo, and V. A. P. Martins Dos Santos, "SEVA 3.1: enabling interoperability of DNA assembly among the SEVA, BioBricks and Type IIS restriction enzyme standards," Microbial biotechnology, vol. 13, no. 6, pp. 1793–1806, 2020, doi: 10.1111/1751-7915.13609.
[22] E.-M. Lammens, L. Putzeys, M. Boon, and R. Lavigne, "Sourcing Phage-Encoded Terminators Using ONT-cappable-seq for SynBio Applications in Pseudomonas," ACS synthetic biology, vol. 12, no. 5, pp. 1415–1423, 2023, doi: 10.1021/acssynbio.3c00101.
[23] E. Martínez-García et al., "SEVA 4.0: an update of the Standard European Vector Architecture database for advanced analysis and programming of bacterial phenotypes," Nucleic acids research, vol. 51, D1, D1558-D1567, 2023, doi: 10.1093/nar/gkac1059.
[24] M. Vazquez-Vilar et al., "GB3.0: a platform for plant bio-design that connects functional DNA elements with associated biological data," Nucleic acids research, vol. 45, no. 4, pp. 2196–2209, 2017, doi: 10.1093/nar/gkw1326.
[25] M. Mirdita, K. Schütze, Y. Moriwaki, L. Heo, S. Ovchinnikov, and M. Steinegger, "ColabFold: making protein folding accessible to all," Nature methods, vol. 19, no. 6, pp. 679–682, 2022, doi: 10.1038/s41592-022-01488-1.