We plan to enable our chassis Pseudomonas vancouverensis DSM8368 to degrade pyrene. For this purpose, we put the nine proteins necessary to channel pyrene into its native phenanthrene pathway, under medium level constitutive expression. To achieve this, at first, we needed to verify our expression system. We did so by introducing amilGFP as reporter protein, regulated by our chosen promoter J23110, RBS B0034 and terminator Luz7-, forming BBa_K5043011, into our chassis. For test amilGFP coding plasmid was also introduced into P. putida KT2440. Cell pellets of transformant were analyzed under blue light (see Figure 1).

As clearly visible in the pictures, green fluorescence can be observed for cells carrying amilGFP-coding-plasmid. This indicates that our chosen expression system works both in P. putida as well as P. vancouverensisThese results also prove our transformation method succesfull.
We then cloned an operon coding for all nine necessary enzymes into backbone pSEVA231 [1] using GoldenBraid [2] method and verified it by sequencing. It was transformed into P. putidaand P. vancouverensis by electroporation. Correct plasmid uptake was in addition verified by plasmid miniprep and control digest using BamHI (see Figure 2).

Description

To better understand the growth dynamics of P. putida and P. vancouverensis, OD600 measurements were plotted against the time. During the growth curve three key points were plated out to obtain viable colony counts, as seen in the standard curve in Figure 4 and 5. This data helps to establish a more accurate correlation between OD600 and colony forming unit (CFU), allowing for better prediction of cell density during future experiments. For P. vancouverensis, a non-model organism with limited available data, these results are particularly interesting and valuable as they provide crucial insights into the growth characteristics of the strain.

Results

The growth dynamics in Figure 4 and 5 of P. putida and P. vancouverensis were analyzed using regression formulas to estimate colony-forming units (CFU) per milliliter based on optical density (OD600). For P. putida, the regression formula of 4×10¹¹ x indicates that at an OD of 0.1, the CFU count is approximately 4×10¹⁰ CFU/ml. Similarly, for P. vancouverensis, the regression formula of 2×10⁷ x results in a CFU count of 2×10⁶ CFU/mL at an OD of 0.1.
In terms of growth rates, P. putida exhibited a growth rate (µ) of 0.63 1/h, with a generation time of 1.10h hours (66 minutes) in exponential phase. Comparatively, P. vancouverensis shows a slightly lower growth rate of 0.59 1/h and a longer generation time of 1.17 hours (70 minutes). Additionally, P. vancouverensis showed an extended lag phase.

Discussion

While P. putida exhibits a slightly faster growth rate, the differences to P. vancouverensis do not affect experimental outcomes in most cases. Both strains could be used effectively for various biotechnological and environmental applications, with strain choice depending more on specific needs rather than a significant difference in growth performance.
P. vancouverensis performs comparably well after a longer lag phase. However, P. putida has an advantage in processes requiring quicker biomass accumulation with a CFU/ml count that is 1000-fold higher than for P. vancouverensis. While cell size differences could contribute to variations in CFU counts, since smaller cells would result in more cells per unit volume at the same OD600 [3], this is less likely because both species belong to the same genus (Pseudomonas), suggesting comparable cell sizes which would not result in such a high difference. It seems more likely that P. vancouverensis experiences a higher rate of cell death compared to P. putida, resulting in lower CFU/ml count.
Nonetheless, the measurement of CFU/mL is cumbersome, and various factors contribute to variability in CFU counts. This is reflected in the correlation coefficients, which is 0.916 for P. vancouverensis and 0.726 for P. putida. The greater variability is likely due to pipetting errors during the serial dilution, particularly for P. putida a plating between the dilutions 10^-8 and 10^-12 was required, meaning more dilution steps. Each additional step increases the chance for errors. As seen in the colony counts for both strain, these were not consistent with the expected values based on the dilution factor. Furthermore, at higher CFU the software OpenCFU [4] was used which introduces additional variability, because the threshold for identifying a colony introduced another error factor.
Overall, despite having similar growth rates, P. putida exhibited a significantly higher CFU count than P. vancouverensis. This difference is due to greater cell viability in P. putida, whereas P. vancouverensis experienced higher cell death, contributing to optical density without forming viable colonies. Despite variability introduced by factors like pipetting errors withing the dilution step an overall picture for both bacteria, especially for P. vancouverensis could be taken.

The OD600 measurements with different pyrene concentration were plotted over incubation time for P. vancouverensis and P. putida including the unmodified and the engineered strain bearing the pyrene degradation plasmid. The primary objective was to compare the growth rates in the presence of pyrene, determining if the plasmid enhances the strain's tolerance to pyrene. Furthermore, the minimal bactericidal concentration (MBC) was determined.

Description

To test pyrene degradation abilities of transformed P. vancouverensis and P. putida, HPLC-analysis of bacterial cultures determining pyrene concentration was carried out.

Method

Liquid cultures of P. vancouverensis and P. putida in LB and M9 media with different pyrene concentrations were incubated at 28°C, 180 rpm. Probes were taken regularly and centrifuged. As pyrene is poorly water-soluble [8], most of it should be found in pellets after centrifuging. Cell pellets were resuspended in dimethyl sulfoxide (DMSO) and cells were lysed using an ultra sonic device. Lysed cells were centrifuged again, and supernatant was applied to Zorbax SB-C18 HPLC column.

Results

Pyrene showed 7.8 to 8.1 minutes retention time on HPLC column. Pyrene peak was best visible at 270nm absorbance. In Figure 10 and 11 results of HPLC-analysis are shown. Pyrene A270 peak area correlating to pyrene concentration in probe is plotted against incubation time. A time error of 1h is assumed. For pyrene peak area an error of 10% as result of pipetting errors is assumed.

HPLC-analysis for P. putida in LB-medium indicates constant pyrene concentration in media. P. vancouverensis analysis shows also constant pyrene concentration for transformed bacteria but great fluctuations for unmodified control. Both exhibit unusual high peak area (>3500mAU*s) after 21h of incubation (not shown in diagram).
All probes in M9-medium show initial pyrene decrease which is followed by constant pyrene amounts. Data for unmodified P. putida strain shows great fluctuations.

Discussion

Data cannot provide proof of pyrene degradation for any of the examined bacterial cultures. Pyrene decrease in all M9-probes including sterile-control could indicate spontaneous decomposition of pyrene. Though this seems unlikely as pyrene is thermodynamically stable [9] and decline could not be observed in LB-cultures.
It stands out, that especially graphs for P. vancouverensis in LB medium and unmodified P. putida in M9 medium show great fluctuations, beyond assumed error tolerances. This most likely indicates that the chosen extraction method does not work quantitatively reliable. This could also explain, why pyrene peak area does not correlate well with pyrene concentration in medium.
In summary, data suggests that no bacterial strain, neither unmodified nor bearing pyrene degradation plasmid, is capable of degrading pyrene. Therefore graphs in all cultures, except P. vancouverensis in LB-medium, show approximately constant pyrene concentrations. This could be due to several reasons.
At first, due to time constraints, we were not able to prove protein expression by transformed strains. As amilGFP was successfully produced using the same expression system, most likely at least some of pyrene pathway’s proteins get produced. Yet there is no proof, the whole operon gets expressed.
In addition, as mentioned earlier, pyrene is poorly water soluble (135μg/l at 25°C) [8]. This means only few supstrate will be available for degradation resulting in overall slow pyrene degradation, so that no pyrene decrease could be observed during our chosen incubation times. In this case, production of biosurfactants enhancing PAH-solubility like shown by 2015 iGEM-team from Uppsala, could be helpful [10].
Furthermore, cellular uptake of pyrene could also be a problem. As all expressed degradation enzymes are cytosolic, pyrene is required to pass bacterial cell membrane. As pyrene shows high octanol-water partition coefficient [8], it can be suspected to diffuse passively through membranes. However for naphthalene active uptake by Pseudomonas fluorescence has been reported [11], indicating an active PAH import system could be necessary for efficient degradation.

In conclusion, most data shows that the transformed bacteria are not capable of degrading pyrene. Only growth curves and MBCs of transformed P. putida indicate a higher pyrene tolerance. This could be due to pyrene degradation or at least detoxification. In summary further analysis, improvements and iterations of the DBTL-Cycle are necessary to implement successful pyrene degradation.

To get more insight into pyrene degradation, we decided to characterize some of our pathway’s enzymes in vitro. It is established that phtAc and phtAd, as electron transfer components, along with a ring-hydroxylating dioxygenase system formed by pdoA2 and pdoB2, aggregate to create a complex exhibiting dioxygenase activity. [13–15] This complex is capable of converting phenanthrene-4-carboxylate, an intermediate product of pyrene degradation, into cis-3,4-dihydroxy-phenanthrene-4-carboxylate (see Figure 12) [14, 15]. Given this reaction’s importance in pyrene degradation, we opted to conduct a more detailed characterization. The components of the electron transport chain, ferredoxin phtAc and ferredoxin reductase phtAd, were produced and purified to elucidate their kinetic parameters. Subsequently, we aimed to analyze the complex formation with pdoA2B2 through HPLC analysis.

Enzyme coding sequences were cloned into pQE bacterial expression vector with a N-terminal, hexahistidine tag. Proteins were expressed in E. coli BL21 (DE3). Main cultures were incubated at 37°C until an OD600 of 0.5 was reached. Subsequently, cultures were induced with a final concentration of 0.5 mM IPTG and incubated overnight at 30°C. Purification was performed using Immobilized Metal Affinity Chromatography (IMAC). Both production and purification samples were analyzed via SDS-PAGE and Coomassie staining (data not shown). Enzyme concentration was determined using the Bradford Assay [17].

The characterization of phtAd activity was performed by evaluating its ability to reduce 2,6-dichlorophenolindophenol (DCPIP), which serves as an electron acceptor [16]. In its oxidized form, DCPIP displays a blue color, which transitions to colorless upon reduction, facilitating the evaluation of phtAd activity (see Figure 13). The decrease in absorbance was monitored at 600nm [16].

The activity of phtAc was assessed using a coupled assay. The increase in absorbance resulting from an electron transfer to cytochrome c signifies an interaction between ferredoxin phtAc and ferredoxin reductase phtAd. The increase in absorbance was monitored at 550nm. Both reactions needed NADH and FADH as cofactors [16].

Method

The optimal reaction time and enzyme concentration for the assay were determined by measuring the decrease in absorbance at 2-minute intervals over a 30-minute period. This approach was necessary due to the uncertainty regarding the concentration at which the reaction would proceed and the duration for which the enzyme could effectively catalyze the reaction before reaching substrate saturation. Following the determination of these parameters, the reaction was assessed using varying concentrations of NADH while maintaining a constant enzyme concentration to determine the kinetic parameters. The reaction was initiated by the addition of a specified amount of phtAd and monitored at 600nm using a UV/Vis spectrometer.

Results

Determination of the optimal enzyme concentration
Initial optical analysis indicates that the assay conducted with the highest concentration of phtAd, 1.25 µM, resulted in the complete reduction of DCPIP, as evidenced by the solution's transition to a fully colorless state (see Figure 14). On the other hand, Figure 15 shows that there is no significant difference in the absorbance measurements between the assays conducted with 1 µM and 1.25 µM phtAd. However, the reaction conducted with 0.5 µM phtAd exhibited a lesser decrease in absorbance over the 30-minute period, indicating that the reaction proceeded at a relatively slow rate.

Regarding the optimal reaction time, Figure 15 illustrates that within the initial 10 minutes, absorbance decreases rapidly, indicating that the reduction of DCPIP occurs predominantly during this period. Beyond the 10-minute mark, absorbance measurements stabilize, suggesting that the reaction has reached a saturation point. This phenomenon may be attributed to the complete reduction of DCPIP or the inability of the reaction to proceed further due to the lack of NADH regeneration.

Based on the results obtained, we selected a concentration of 1.25 µM phtAd for the subsequent assay, as this concentration demonstrated a complete reduction of DCPIP.

Kinetic parameters of phtAd

For the determination of kinetic parameters the assay was run with different NADH concentrations ranging from 40 µM to 250 µM. Data obtained from these assays was then analyzed using Michaelis-Menten kinetics (see Figure 16) and a Lineweaver-Burk plot (see Figure 17) was constructed to provide a linear representation of enzyme kinetics to determine key parameters such as Vmax and Km.

The Vmax, Km, kcat, kcat/Km values of the ferredoxin reductase phtAd were 0.0011 mM/min, 0.028 mM, 1.12 min^-1 and 39.46 mM^-1 x min^-1, respectively. Notable discrepancies from the values reported in the literature [16] were identified, which can be attributed to variations in enzyme concentration. The reaction was performed using a higher enzyme concentration due to the observed diminished activity of the produced enzymes.
Furthermore, the R² value of 0.4668 indicates that the Lineweaver-Burk plot does not perfectly describe the relationship between 1/[NADH] and 1/V. While the fit is moderate, it suggests that factors like experimental variability, biological complexity, or non-linearity in enzyme behavior might be contributing to deviations from the ideal linear relationship.

Method

Initial measurements were conducted in accordance with the methodology outlined by Wu et al. (2020) [16]. However, the assay produced inconclusive results, as no variation in absorption was observed over time. This lack of change was attributed to the high concentration of cytochrome c (600 µM) utilized in the initial trials. Consequently, we opted to conduct further assays employing varying concentrations of cytochrome c and NADH to identify the optimal concentrations that would facilitate a measurable increase in absorption. In alignment with the approach taken by Wu et al. (2020) [16], subsequent assays were performed using a phtAd to phtAc ratio of 1:3. The increase in absorbance was measured over a 30-minute period at 2-minute intervals at 550 nm. After the optimal cytochrome c concentration was established, further assays were performed varying the NADH concentration to determine the kinetic parameters.

Results

Determination of the optimal substrate concentration
In the initial assay conducted with 250 µM cytochrome c, no change in absorbance was observed (see Figure 18A). Considering this outcome, the subsequent assay was performed using 100 µM cytochrome c in conjunction with a higher concentration of NADH (200 µM). This combination resulted in a measurable increase in absorbance, demonstrating a linear progression over a duration of 26 minutes (see Figure 18C) indicating the oxidation of cytochrome c. To validate this outcome, additional assays were conducted utilizing 100 µM cytochrome c alongside varying concentrations of NADH, both higher and lower. The assay employing 250 µM NADH also demonstrated a linear increase in absorption (Figure 18B) in the first 26 minutes. Conversely, the assay with 100 µM NADH (See Figure 18D) also exhibited a linear increase in absorption, albeit at a reduced rate. Notably, no saturation point was observed at this concentration, in contrast to the higher concentrations, which indicated saturation occurring approximately 30 minutes into the assay.

Based on the results obtained, we chose a concentration of 100 µM cytochrome c for the subsequent tests, as this concentration showed a measurable increase in absorbance. For the reaction time, we decided to run further reactions for 34 minutes to confirm the saturation point.

Kinetic parameters of phtAc

For the determination of kinetic parameters the assay was run with different NADH concentrations ranging from 40 µM to 400 µM. Data obtained from these assays was then analyzed using Michaelis-Menten kinetics (see Figure 19) and a Lineweaver-Burk plot (see Figure 20) was constructed to provide a linear representation of the enzyme kinetics to determine key parameters such as Vmax and Km.

The Vmax, Km, kcat, kcat/Km values of the ferredoxin phtAc were 0.00043 mM/min, 0.31 mM, 0.12 1/min and 3.9 1/mM x min, respectively. Significant deviations from the literature values [16] were observed; however, the reaction was conducted with a substantially lower concentration of cytochrome c and elevated concentrations of enzymes, as the produced enzymes exhibited reduced activity.
Moreover, the R² value of 0.5378 derived from the kinetic analysis of phtAc is marginally higher than that of phtAd. This marginally elevated value may indicate a greater reliability of the data or a more consistent enzymatic response across varying NADH concentrations. Nevertheless, the Lineweaver-Burk plot does not accurately characterize the relationship between 1/[NADH] and 1/V. Although the fit is moderate, it implies that factors such as experimental variability, biological complexity, or non-linear enzyme behavior could be responsible for the observed deviations from the expected linear relationship.

We successfully characterized and demonstrated the activity of phtAc and phtAd both individually and as a complex. This complex functions as an electron carrier for pdoA2B2 [13–15]. The subsequent step involved conducting an activity assay in conjunction with HPLC analysis with the aim to evaluate the function of pdoA2B2 both independently and in complex with phtAcAd, forming a tetramer.

Initially, the activity assay was conducted, wherein the enzymes were incubated at 30°C for approximately 15 to 20 minutes, both in the presence and absence of NADH to examine the effect of NADH on enzymatic aggregation. Subsequently, phenanthrene-4-carboxylate (P4C) was added to initiate the reaction. The reaction mixture was then incubated for 30 minutes at 30°C, after which the reaction was terminated by the addition of 100% methanol. Following this, HPLC analysis was performed using a Zorbax SB-C18 column, employing an acetonitrile/water gradient at a flow rate of 0.4 ml/min, with the column oven maintained at room temperature. The gradient elution program commenced with an initial mobile phase of 60:40 (v/v) acetonitrile to water, transitioning linearly to 100% acetonitrile over 14 minutes, followed by a return to the initial phase (60:40) after 5 minutes. The total duration for each analysis was 35 minutes. For the negative control, an assay was conducted using denatured enzymes, achieved by heating the enzymes at 95°C for 10 minutes. Peaks were analyzed at 270nm absorbance.

Initial activity assays were performed solely with pdoA2 and pdoB2 to form a dimer, thereby validating their enzymatic activity and evaluating their aggregation characteristics. Monomers were incubated together for aggregation in the presence and absence of NADH to determine whether NADH, as a cofactor, influences the aggregation process.
The efficacy of the reaction was assessed by measuring the absorbance of phenanthrene-4-carboxylate (P4C) and NADH using HPLC. The area under the resulting peaks corresponding to the levels of P4C and NADH was determined, enabling a comparative analysis of the consumption of P4C and NADH against established standards and between the samples. In this context, phenanthrene-4-carboxylate consumption refers to the extent of P4C degradation, as indicated by a reduction in peak area compared to the standard. In the absence of additional samples with different concentrations, it is not possible to quantify the remaining P4C in the samples. The only conclusion that can be drawn is whether P4C has been degraded, as evidenced by the decrease in the area of the corresponding peak; this applies also to NADH. Furthermore, to eliminate potential statistical errors in the assessment of our results, we employed triplicate measurements for each sample and calculated their averages. The results are presented in Figure 21. Results indicated that the assays in which NADH was added prior to aggregation exhibited enhanced P4C degradation and, correspondingly, increased NADH consumption compared to the assays where NADH was added after aggregation. However, the observed NADH consumption was relatively low, which may be attributed to the absence of the electron carrier dimer (phtAcAd) in the sample, thereby limiting NADH consumption.

Considering the initial results, the analysis of the tetramer aggregation process and the degradation of phenanthrene-4-carboxylate was conducted by introducing NADH prior to the aggregation process.

Degradation of phenanthrene-4-carboxylate by enzyme complex with dioxygenase activity using NADH as a cofactor
The presence of the functional enzyme complex with dioxygenase activity is anticipated to result in a reduction in the quantity of phenanthrene-4-carboxylate. This reduction can be quantified by determining the decrease in the area of the corresponding peak in comparison to the standard. Given the unknown stoichiometry of this enzyme complex, we conducted assays utilizing two distinct stoichiometric ratios. The first assay employed a 1:1 stoichiometry of all monomers, while the second utilized a ratio of pdoA2¹:pdoB2¹:phtAc³:phtAd¹. This latter stoichiometry was selected based on previous kinetic analyses of phtAc, which were performed with a ratio of three equivalents phtAc to one phtAd for determining kinetic parameters [16]. This ratio was employed in the present assays to assess its influence on the degradation of phenanthrene-4-carboxylate and to determine whether the presence of three equivalents phtAc enhanced the efficiency of the electron transfer process, thereby accelerating the breakdown of phenanthrene-4-carboxylate. However, results showed (see Figure 22) that the breakdown of phenanthrene-4-carboxylate was slightly more efficient with the 1:1 ratio of the monomers; this could suggest that a higher ratio of phtAc does not enhance the electron transfer process. And that the complex with 1:1 ratio of all enzymes aggregates better.
To confirm the aggregation of the enzyme complex and the successful degradation of P4C by this complex, negative controls were prepared. For this, assays were prepared with denatured enzymes with the stoichiometries described above. Results showed (see Figure 22) that the peak area of P4C within these assays was higher in comparison with the assays with the active enzyme complex. The comparison of these results indicates that the enzyme complex was successfully formed and demonstrated the capacity to degrade P4C.

Another negative control was conducted by excluding any enzymes from the assay. In comparison to samples containing denatured enzymes, the results showed a reduced peak area in the peak corresponding to phenanthrene-4-carboxylate, suggesting either a diminished quantity of P4C in the sample or its degradation. However, the possibility of cross-reactivity with other components in the assay can be discounted, as such interactions would also manifest in the other negative controls. This observation indicates the presence of an independent source of error. Potential issues may involve the unintentional introduction of enzymes or inadequate substrate concentration. The presence of enzymes in the assay can be corroborated by taking into consideration the amount of NADH; notably, the peak area of NADH in this sample was diminished suggesting the possibility that a reaction has taken place, and the results are more consistent with those obtained from the assay conducted with a 1:3 enzyme ratio (see Figure 22). While numerous errors could have arisen, a systematic issue such as cross-reactivity, which would render the enzyme ineffective, can be dismissed by comparing the sample to those with non-functional enzymes.
Moreover, as anticipated, the samples containing the active enzyme complex demonstrated a decrease in the peak area associated with NADH relative to the standard (see Figure 22). This finding corroborates the functionality of the electron carrier dimer, which collaborates with pdoA2B2 in the degradation of P4C. In contrast, the assay involving denatured enzymes at a 1:1 ratio revealed no consumption of NADH, indicating the absence of any reaction.
However, the results of NADH consumption from the assay involving denatured enzymes, utilizing three equivalents of phtAc in conjunction with a 1:1 ratio of the other enzymes, demonstrate a reduction in NADH consumption, as evidenced by the diminished area of the corresponding peak relative to the standard. This observation indicates the presence of an independent source of error, as the other assay performed with denatured enzymes shows no NADH consumption. Potential issues may include insufficient NADH concentration. If the NADH concentration was low, degradation could have occurred over time, as the activity assays were conducted days prior to the HPLC analysis. The thawing process and exposure of the sample to temperature fluctuations may have contributed to NADH degradation due to its inherent instability. Additionally, the possibility of incomplete denaturation should be considered. If the enzymes were not fully denatured during the assay, residual active enzyme could remain, resulting in NADH consumption. The denaturation conditions may not have been adequate to completely inactivate the enzymes. This is further supported by the observation that P4C consumption was slightly lower in comparison to the sample with denatured enzymes with the 1:1 ratio.

In conclusion, the enzyme complex with dioxygenase activity demonstrated its ability to degrade phenanthrene-4-carboxylate (P4C), with slightly greater efficiency observed in the assay utilizing a 1:1 ratio of monomers compared to the 1:3 ratio of phtAc to pdoA2:pdoB2:phtAd. This suggests that a higher ratio of phtAc does not enhance electron transfer efficiency. Negative controls, including assays with denatured enzymes and no enzymes, confirmed that the active enzyme complex was responsible for P4C degradation, while the reduction in NADH in some denatured enzyme assays points to possible errors, such as NADH degradation or incomplete enzyme denaturation. Further assays with stricter controls are needed to confirm these findings and minimize potential sources of error.

Immobilization of our chassis is important, to make a self-maintaining device of riverine deployment possible. For this we chose five evidence-based methods by the criteria feasibility and sustainability. We included silica beads and a larger silica surface as P. vancouverensis was immobilized on silica in preceding research [28]. The other three materials (modified alginate, carboxy methyl cellulose and rice bran) were chosen due to them having proven to work in the related model organism P. putida. The physical properties were tested, as well as viability and immobilization efficiency. All tests were done with the P. vancouverensis transformant containing the pSEVA231-plasmid.

Physical properties

1. Alginate beads do not melt at temperatures up to 37°C so they would work even in very hot rivers. They are difficult to destroy with physical force because they withstand being squished to some degree and then dodge the force exerted by the object by slipping away.
2. Silicabeads are almost indestructible by the physical force a human can induce. They do not melt or disintegrate with high temperatures (in our experiments they had no problem with 120°C and we could have gone much higher).
3. Silica pieces with larger surface than the beads, are slightly easier to crush but still a very stable material.
4. Carboxy methyl cellulose (CMC) beads are easy to crush, even a water flow can crush them, and they melt at 28 °C once damaged. They are fully solved in the watery solution after 48 hours. Therefore, they are not suitable for our application, but we only figured that out later while doing biological tests.
5. Rice bran turned out to be impractical for our use case because the formed beads disintegrate in water, especially fast when there is water flow. Therefore, the idea was discarded after testing the physical properties.

Viability tests

1. The creation of alginate beads does not kill the bacteria. The method is established and often used with a variety of bacteria [25]. For our viability tests, the beads were destroyed by physical force and put in liquid culture as well as being plated out over 48 h. The tests proved that P. vancouverensis is no different and survives the procedure.
2. Immobilizing bacteria on silica beads is also known to work [26-28]. The nonpolar surface of silicon dioxide leads to P. vancouverensis immobilizing itself. The bacteria are just cultivated in liquid culture with the silica beads, making it the gentlest treatment. Therefore, the viability of the cells is expected to remain fully intact. Our results prove these expectations, showing that immobilized bacteria are still viable.
3. The immobilization on the larger silica surface is done in the same way as on the beads. Therefore, our expectations were the same and the results turned out the same as well.
4. The bacteria in CMC beads showed viability as well, even though the beads were not physically stable.

Immobilization efficiency

To determine immobilization efficiency, meaning how difficult it is to wash bacteria off, the materials with immobilized cells were first washed multiple times in water and then inverted in LB medium, simulating water flow. The LB medium was subsequently plated out. The less viable cells could be observed, the more effective immobilization worked.

1. As the bacteria are enclosed in the alginate beads, its not possible to wash off all the bacteria. After washing there were still some bacteria coming of. In the first test there were around 900 colonies, in the second test 16 colonies could be counted.
2. Silica beads showed a lot of bacteria remaining on the surface after washing five times, but there were still bacteria that came off when simulating water flow. When plated out the washing medium, around 10.000 colonies could be counted after several days. There were no second tests of this method due to a lack of material.
3. The Silica surface also showed a lot of bacteria remaining on the surface after washing, but there were still bacteria that came off when simulating water flow as well. When plated out the washing medium, around 2.500 colonies could be counted after several days. There were no second tests of this method due to a lack of material.
4. In the CMC beads, the bacteria are enclosed as well. When simulating water flow and plating the medium in the first test, only 15 colonies could be observed. In the second test there were no colonies visible.

Conclusion

Our tests show that rice bran and carboxy methyl cellulose are insufficient for our application, due to their physical properties. On the other hand, using alginate or silica is suitable for our purposes. Both methods have good physical properties and immobilize the bacteria without killing them. When testing the efficiency of the immobilization, it turned out that after several washes, most of the bacteria remained immobilized, but many were still washed off.

It is not our desired outcome, that a lot of the bacteria were still washed off after several washes. But when talking to Julia Polat, whose research is centered around immobilization with alginate, she reassured us there is no way to prevent bacteria from leaving their immobilized state, especially as time progresses. For us, this means we need a very reliable kill switch to prevent biocontamination. We can be sure that the only bacteria that grew is P. vancouverensis because a transformed variant was used that contained a plasmid with kanamycin resistance, and all tests were done in LB medium with 50μg/ml kanamycin. The creation of the beads and further tests were also carried out in a sterile environment so the chances of another kanamycin resistant bacterium entering the experiment are very low. Nevertheless, we additionally analyzed the visual properties of the colonies, and they all looked like the typical P. vancouverensis colonies. To conclude, our tests of the two methods using alginate and silica showed promising results. They should undergo more testing in the future, especially simulations of a river with a typical current, to see which one is the best method for our use case.

As in our application in rivers no antibiotic resistance pressure can be applied and our chassis would lose pyrene degradation plasmid, it is necessary to integrate the pyrene pathway into the host’s genome.
Furthermore, our advisor M. Sc. Christian Hake pointed out, that PAHs are very stable molecules and therefore represent unattractive nutrition sources. Hence our chassis might not produce naphthalene and phenanthrene degradation enzymes, as more nutritious carbon sources are around in rivers. Therefore we plan to put naphthalene and phenanthrene degradation genes under low constitutive expression as well. This could be done for example using CRISPR systems [18].

In order to avoid contamination or an unwanted upsurge of GMOs in the environment implementing a kill switch is crucial. An light activated kill switch could be promising, like shown by iGEM Team Ashesi 2022 [19], which kills the engineered bacteria upon leaving our device. Though the light in the depth of a riverbed could not be enough to activate the kill switch. Another possibility could be implementing an auxotrophic marker to create a dependency on a substance only supplied by the surface inside the container on which the bacteria grow, but that requires attendance and care for the system which is supposed to function autonomously.
Therefore, the best cut out would be one designed especially for this purpose. A kill switch activated by not being immobilized on a surface like a RNAse or toxin bound to an extracellular receptor. Even here might be a downside, as not every receptor in the membrane of a cell could be attached to the surface and would cause an RNAse or toxin to be moving freely inside the cell. A promising possibility could be a toxin-antitoxin system, where the bacteria express a toxin constitutively and selectively express an antitoxin or inhibitor based on the containment in our device. In the project of TU Darmstadt 2020 [20] they thought of a kill switch connecting their bacteria to a biofilm. They did so by giving the gene for the toxin to one colony and the gene for the antitoxin to another. When they coexisted in a biofilm they survived and otherwise were killed by an excess amount of the toxin. The idea would need to be tested long term on P. vancouverensis colonies to see if it works for us. This complex and difficult aspect of our project might inspire future generations of iGEMers to find a solution to our kill switch problem.

PAHs are poorly water soluble [8], which limits substrate availability for degradation and probably thereby shrinks bioremediation’s efficiency [21]. This is why Prof. Victor de Lorenzo from CSIC Madrid advised us to use biosurfactants. Biosurfactants are amphipathic, surface-active molecules which can increase solubility of hydrophobic substances [21]. Though recent research shows doubts, it is suspected that in this way PAH degradation could be enhanced [22]. Therefore we plan to produce biosurfactants in situ by introducing its biosynthesis pathway genetically into our chassis.
iGEM teams Sorbonne 2023 and Columbia-University 2016 already used Pseudomonas putida to produce sophorolipids [23] and rhamnolipids [24]. As we would be happy to build on previous iGEM team’s success, we are eager to try out biosurfactant production in P. vancouverensis and study its effects on PAH degradation.

Since the engineering cycle is an iterative process, continuous improvements can be made to the design of the device. Additional simulations could help us to further optimize the internal fluid dynamics of the device, maximizing the contact between PAHs in the water and Pseudomonas bacteria in the immobilization material. The results of these simulations would then be validated using 3D-printed prototypes.
The next phase would involve constructing full-scale prototypes with the material that is already used for buoys, which would need to be tested under real-world conditions, along with the engineered bacteria.
Several potential issues need to be identified and addressed through testing, including the risk of clogging due to floating debris, the stability of the device under adverse weather conditions and animal interactions, the ability of Pseudomonads to adhere to the interior surfaces as intended, and whether they can effectively degrade PAHs in this scenario. These tests would show us potential for improvement, which can be used to develop a market-ready product.
When the fundamental functionality of the device has been demonstrated, the focus would shift towards its longevity and ease of maintenance. This would require long-term testing, which must still be planned and conducted.

Once the device has been fully tested and its functionality confirmed, the next step would involve transitioning it into a complete, market-ready product. This process would begin with sourcing a supplier for the necessary materials and identifying a manufacturer capable of producing the device on a large scale. With these partnerships established, we could accurately calculate production costs and use this information to determine an appropriate price point for the product. Mostly practical would be the use of the material of already used buoys, which are proved already as a durational, cheap and water friendly option.
Subsequently, efforts would be directed towards customer acquisition. Potential customers include companies seeking to remediate polluted waterways they are responsible for, as well as municipalities or local governments aiming to clean their rivers. Given the versatility of our design, the device can be deployed globally, provided the water body does not have excessively strong currents and there is a suitable method for securing the device in place.
To anchor the device, we propose using existing buoys, which are already widely available in many river systems. Buoys offer a practical solution for maintaining the device’s position and are typically well-distributed, making it feasible to deploy multiple units if required for larger-scale water treatment efforts. To use an already existing infrastructure would be wasteless and a cheap opportunity.

Our hope is for future iGEMers to build upon our project, expand and improve it, as it represents an innovative and low-effort method for cleaning rivers from dangerous pollutants. We wish to inspire many new projects no matter if it’s in bioremediation, upscaling, foundational advance or anything else. Either way, one thing’s for sure. There’s much to do for generations of iGEMers to come.

[1] 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.
[2] 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.
[3] J. A. Myers, B. S. Curtis, and W. R. Curtis, "Improving accuracy of cell and chromophore concentration measurements using optical density," BMC Biophys, vol. 6, no. 1, p. 4, 2013, doi: 10.1186/2046-1682-6-4.
[4] Q. Geissmann, "OpenCFU, a new free and open-source software to count cell colonies and other circular objects," PLoS ONE, vol. 8, no. 2, e54072, 2013, doi: 10.1371/journal.pone.0054072.
[5] R.-H. Peng et al., "Microbial biodegradation of polyaromatic hydrocarbons," FEMS Microbiol Rev, vol. 32, no. 6, pp. 927–955, 2008, doi: 10.1111/j.1574-6976.2008.00127.x.
[6] L. Wang et al., "Improving Degradation of Polycyclic Aromatic Hydrocarbons by Bacillus atrophaeus Laccase Fused with Vitreoscilla Hemoglobin and a Novel Strong Promoter Replacement," Biology, vol. 11, no. 8, 2022, doi: 10.3390/biology11081129.
[7] T. J. Kim, "Degradation of polyaromatic hydrocarbons by Burkholderia cepacia 2A-12," (in En;en), World Journal of Microbiology and Biotechnology, vol. 19, no. 4, pp. 411–417, 2003, doi: 10.1023/A:1023998719787.
[8] M. M. Miller, S. P. Wasik, G. L. Huang, W. Y. Shiu, and D. Mackay, "Relationships between octanol-water partition coefficient and aqueous solubility," Environmental science & technology, vol. 19, no. 6, pp. 522–529, 1985, doi: 10.1021/es00136a007.
[9] 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.
[10] iGEM Uppsala 2015, Decyclifier. [Online]. Available: https://2015.igem.org/Team:Uppsala/Biosurfactants (accessed: Sep. 23 2024).
[11] T. Bugg, J. M. Foght, M. A. Pickard, and M. R. Gray, "Uptake and active efflux of polycyclic aromatic hydrocarbons by Pseudomonas fluorescens LP6a," Applied and environmental microbiology, vol. 66, no. 12, pp. 5387–5392, 2000, doi: 10.1128/aem.66.12.5387-5392.2000.
[12] 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.
[13] C. Pagnout, G. Frache, P. Poupin, B. Maunit, J.-F. Muller, and J.-F. Férard, "Isolation and characterization of a gene cluster involved in PAH degradation in Mycobacterium sp. strain SNP11: expression in Mycobacterium smegmatis mc(2)155," Research in microbiology, vol. 158, no. 2, pp. 175–186, 2007, doi: 10.1016/j.resmic.2006.11.002.
[14] K. Yuan et al., "Transcriptional response of Mycobacterium sp. strain A1-PYR to multiple polycyclic aromatic hydrocarbon contaminations," Environmental pollution (Barking, Essex : 1987), vol. 243, Pt B, pp. 824–832, 2018, doi: 10.1016/j.envpol.2018.09.001.
[15] S. Krivobok, S. Kuony, C. Meyer, M. Louwagie, J. C. Willison, and Y. Jouanneau, "Identification of pyrene-induced proteins in Mycobacterium sp. strain 6PY1: evidence for two ring-hydroxylating dioxygenases," Journal of bacteriology, vol. 185, no. 13, pp. 3828–3841, 2003, doi: 10.1128/jb.185.13.3828-3841.2003.
[16] Y. Wu, Y. Xu, and N. Zhou, "A newly defined dioxygenase system from Mycobacterium vanbaalenii PYR-1 endowed with an enhanced activity of dihydroxylation of high-molecular-weight polyaromatic hydrocarbons," Front. Environ. Sci. Eng., vol. 14, no. 1, 2020, doi: 10.1007/s11783-019-1193-5.
[17] M. M. Bradford, "A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding," Analytical biochemistry, vol. 72, pp. 248–254, 1976, doi: 10.1006/abio.1976.9999.
[18] P. L. H. Vo et al., "CRISPR RNA-guided integrases for high-efficiency, multiplexed bacterial genome engineering," Nature biotechnology, vol. 39, no. 4, pp. 480–489, 2021, doi: 10.1038/s41587-020-00745-y.
[19] iGEM Team Ashesi, Gold Alliance. [Online]. Available: https://2022.igem.wiki/ashesighana/design (accessed: Sep. 25 2024).
[20] iGEM Team Darmstadt, B-Tox. [Online]. Available: https://2020.igem.org/Team:TU_Darmstadt/Project/Kill_Switch (accessed: Sep. 25 2024).
[21] Kaustuvmani, M. C. KALITA, and S. DEKA, Degradation of polycyclic aromatic hydrocarbons (PAHs) employing biosurfactant producing Pseudomonas aeruginosa KS3, 2015.
[22] Ł. Chrzanowski, Ł. Ławniczak, and K. Czaczyk, "Why do microorganisms produce rhamnolipids?," World Journal of Microbiology and Biotechnology, vol. 28, no. 2, pp. 401–419, 2012, doi: 10.1007/s11274-011-0854-8.
[23] iGEM Team Sorbonne, PseudoDetox. [Online]. Available: https://2023.igem.wiki/sorbonne-u-paris/description (accessed: Sep. 25 2024).
[24] iGEM Team ColumbiaU NYC, A Mosquito repellent for everyone. [Online]. Available: https://2016.igem.org/Team:ColumbiaU_NYC/Project (accessed: Sep. 25 2024).
[25] S. Khashei, Z. Etemadifar, and H. R. Rahmani, “Immobilization of Pseudomonas putida PT in resistant matrices to environmental stresses: a strategy for continuous removal of heavy metals under extreme conditions,” Ann Microbiol, vol. 68, no. 12, 2018, doi: 10.1007/s13213-018-1402-7.
[26] B. R. Mutlu, J. K. Sakkos, S. Yeom, L. P. Wackett, and A. Aksan, “Silica ecosystem for synergistic biotransformation,” Sci Rep, vol. 6, 2016, doi: 10.1038/srep27404.
[27] J. K. Sakkos, D. P. Kieffer, B. R. Mutlu, L. P. Wackett, and A. Aksan, “Engineering of a silica encapsulation platform for hydrocarbon degradation using Pseudomonas sp. NCIB 9816-4,” Biotechnol Bioeng, vol. 113, no. 3, 2016, doi: 10.1002/bit.25821.
[28] G. Hwang, C. H. Lee, I. S. Ahn, and B. J. Mhin, “Analysis of the adhesion of Pseudomonas putida NCIB 9816-4 to a silica gel as a model soil using extended DLVO theory,” J Hazard Mater, vol. 179, no. 1–3, 2010, doi: 10.1016/j.jhazmat.2010.03.101.