This section presents our wet lab achievements. Our main goal was to engineer the bacterium Pseudomonas fluorescens - our BioMoon biostimulant - so that it could be grown from a yet non-recycled substrate in a space base. After successfully completing this task, we considered improving the bacterial capacity to facilitate plant growth on lunar regolith. We eventually demonstrated the ability of our strain to act as a plant biostimulant on regolith.
Engineering a Pseudomonas fluorescens strain that can grow on creatinine
The wild-type strain Pseudomonas fluorescens SBW25 does not grow on creatinine
Objectives
The pathway for creatinine degradation is well characterized in some Pseudomonas bacteria, e.g., P. putida 1. However, creatinine metabolization by our strain of interest P. fluorescens has not been investigated so far. Our objective was to assess whether the wild-type strain P. fluorescens was able to metabolize creatinine and the downstream compounds of the pathway.
Method
We studied growth of P. fluorescens on each compound of the metabolic pathway separately, namely creatinine, creatine, sarcosine, and glycine on a M9 minimal medium deprived of NH4Cl (i.e., no additional nitrogen source). The growth was monitored by measuring OD at 600 nm in a 96-well microplate with a microplate reader.
Hypothesis
In silico simulations obtained from the GSMM (Genome-Scale Metabolic Model) of P. fluorescens helped us make predictions about bacterial growth under different compounds. The modeling predicted that glycine supports growth of the wild-type (WT) P. fluorescens strain. Sarcosine should also be usable as a substrate. Moreover, the presence of a sarcosine oxidase gene, catalizing the sarcosine degradation into glycine, was uncovered in the genome of P. fluorescens. However, the model predicts that growth on creatinine and creatine is not possible in the WT strain.
Results
We experimentally tested these hypotheses and found that P. fluorescens SBW25 was only able to grow on sarcosine as a sole source of carbon and nitrogen (56 mM) with a maximal growth rate of 0.16 ± 0.01 h–1 (mean ± standard deviation, from 9 biological replicates), but not on creatinine (44 mM), creatine (38 mM), nor glycine (67 mM) (Figure 1).
To quantify the degradation of sarcosine by the WT P. fluorescens strain, we monitored sarcosine uptake using Nuclear Magnetic Resonance (NMR) spectroscopy (500 MHz NMR spectrometer Bruker). NMR analysis was conducted on the supernatants collected during the exponential growth phase of the bacteria. A sarcosine uptake rate of 5.2 mmol.h⁻¹.gDW⁻¹ was obtained (Fig. 2), while the measured growth rate under this condition was 0.19 h⁻¹ (n=1). The cultures used for this analysis were grown in Erlenmeyer flasks, providing more effective aeration and agitation compared to microplate experiments, which may explain the higher growth rate.
To validate the predictability of our metabolic model, we used the experimentally measured sarcosine uptake rate as an input parameter and the model predicted a growth rate of 0.20 h⁻¹, which is in perfect agreement with the experimentally observed rate of 0.19 h⁻¹. These results strongly suggest that the sarcosine degradation pathway is functioning optimally in the WT strain, and that our whole-cell model is able to give accurate predictions about metabolic fluxes.
The results from Figure 1 also demonstrate that P. fluorescens is unable to utilize glycine as a sole carbon and nitrogen source in vivo. However, P. putida is able to degrade creatinine following the creatinine metabolic pathway which ends with glycine conversion and provides a source of carbon and nitrogen2, 3 (Figure 3). The obtained results from P. fluorescens are different, except for the sarcosine oxidase (SoxA) which catalyzes the conversion of sarcosine into glycine with co-production of formaldehyde and H2O2:
It is important to note that, although P. fluorescens cannot grow on creatinine, creatine, or glycine as sole carbon and nitrogen sources, it may still be capable of utilizing creatinine or glycine as nitrogen sources. To explore this hypothesis, we supplemented the M9 minimal medium with sarcosine, creatine, or creatinine along with citrate, which served as primary carbon source. Citrate, a key intermediate in the TCA cycle, allows bacteria to produce energy and biosynthetic precursors. In addition, modeling simulations demonstrated that citrate can be co-consumed with creatinine.
This was experimentally demonstrated. The WT strain significantly grew on creatinine supplemented with citrate (Figure 4, left), exhibiting a clear exponential phase. In contrast, no growth was observed when citrate was provided along with creatine (Figure 4, center). Bacterial growth was also observed on glycine and citrate (0.28 ± 0.01 h⁻¹ (n=9), Figure 4, right), with a growth curve slightly delayed compared to that of creatinine and citrate (0.29 ± 0.01 h⁻¹ (n=9), Figure 4, left).
These findings suggest that P. fluorescens is unable to use creatinine, creatine, or glycine as sole carbon and nitrogen sources. The WT strain can utilize creatinine and glycine only as nitrogen sources when citrate is provided as the primary carbon source, but not creatine.
New questions arise
Further experiments will be needed to explain why the WT strain does not grow in the presence of citrate and creatine. Bacteria could either be unable to effectively metabolize creatine as primary nitrogen source, or creatine diffusion/transport across the bacterial membrane may be limiting.
According to the creatinine metabolic pathway in P. putida4 and in silico results, sarcosine is degraded into glycine and formaldehyde (Figure 3). We have demonstrated that glycine cannot be used as a carbon and nitrogen source but only as a nitrogen source (Figure 1 and 4). As P. fluorescens can grow on sarcosine, we suggest that formaldehyde, a by-product of sarcosine degradation was further converted into formic acid, which was then used as carbon source, while glycine was used as nitrogen source. We tested this hypothesis in silico.
Formaldehyde was indeed observed to have a large impact on P. fluorescens growth rate. To verify the in silico predictions, we cultured bacteria on M9 minimal medium supplemented with 10 mM formate and 67 mM glycine. P. fluorescens was not able to grow under this condition. Further experiments are required to understand how sarcosine is metabolized by P. fluorescens and what the influence of the reaction products is on growth.
To enable P. fluorescens to grow on creatine or creatinine as sole carbon and nitrogen sources, we proposed to introduce the creatinine metabolization pathway from P. putida into P. fluorescens. The genes creA and crnA were cloned in P. fluorescens to convert creatinine into creatine and then into sarcosine. Simulation results from our in silico model showed that P. fluorescens expressing CreA and CrnA enzymes should be able to grow on creatinine alone. This prompted us to engineer a strain with the creA and crnA genes.
Construction of the plasmid pSEVA438-Ptet cloned with the operon creA-crnA and soxA genes
Objectives
The initial design aimed at cloning the genes creA, crnA, and soxA into pSEVA438-Ptet vector to obtain the pSEVA438-Ptet-creA-crnA-soxA plasmid for transformation of P. fluorescens.
Method
The strategy to construct the pSEVA438-Ptet-creA-crnA-soxA plasmid was to perform a three-step cloning through In-Fusion Assembly, transformation into E. coli Stellar Competent Cells, vector isolation and linearisation by PCR (Fig. 5). The first cloning step introduced the Ptet promoter from the pHD_SS9_PLtetO1_TLt0_eutC-GFP plasmid into the pSEVA438 vector. Then, the creA-crnA operon (BBa_K5108009) was cloned. Finally, the bidirectional terminator LUZ7 T50 (BBa_K4757058) and the soxA gene were cloned. The final construct was transformed into P. fluorescens.
We experimentally tested these hypotheses and found that P. fluorescens SBW25 was only able to grow on sarcosine as a sole source of carbon and nitrogen (56 mM) with a maximal growth rate of 0.16 ± 0.01 h–1 (mean ± standard deviation, from 9 biological replicates), but not on creatinine (44 mM), creatine (38 mM), nor glycine (67 mM) (Figure 1).
Results
We succeeded in cloning the Ptet promoter into pSEVA438 and the creA-crnA operon (BBa_K5108009) into the pSEVA438-Ptet backbone, as shown in Figure 6. The construct was confirmed by Sanger sequencing (Genewyz, Germany). However, we failed to introduce the bidirectional terminator LUZ7 T50 (BBa_K4757058) and the soxA gene into the pSEVA438-Ptet-creA-crnA plasmid. As we demonstrated that P. fluorescens could grow on sarcosine, the soxA gene should be unnecessary for bacteria to grow on creatinine. This is why we decided to directly transform P. fluorescens with the pSEVA438-Ptet-creA-crnA plasmid.
The engineered strain Pseudomonas fluorescens SBW25 can express creatinine amidohydrolase (CrnA) and creatinase (CreA)
Objectives
The objective was to verify the overproduction of the creatinine amidohydrolase (CrnA) and creatinase (CreA) in P. fluorescens by activation of the Pm promoter with the inducer m-toluic acid.
Method
The pSEVA438-MBPeGFP plasmid, originally used in P. putida KT2440, was employed as positive control of Pm promoter's inducibility in P. fluorescens. This construct encodes the maltose-binding protein (MBP) fused to eGFP (enhanced Green Fluorescent Protein) under the control of the Pm promoter and is issued from Vogeleer et al. (2024)5. We cultured the pSEVA438-MBPeGFP- and pSEVA438-Ptet-creA-crnA-transformed P. fluorescens in M9 minimal medium supplemented with glucose (28 mM), with or without 0.5 mM of m-toluic acid inducer. After incubation, a whole-protein extraction was performed for each strain to assess the level of expression, as well as the solubility of our proteins.
Hypothesis
We expected that MBPeGFP, creatinine amidohydrolase (CrnA) and creatinase (CreA) expression should be activated in the presence of m-toluic acid. These proteins should be soluble with an expected molecular weight of 69 kDa, 28 kDa and 46 kDa, respectively.
Results
The obtained SDS-PAGE is presented in Figure 7.
Both soluble and insoluble fractions contained MBPeGFP, with the majority of protein being in the soluble fraction independently of the presence of the inducer. Although transcriptional leakage was clearly observed without the inducer, MBPeGFP was overproduced when the Pm promoter was activated with 0.5 mM of m-toluic acid, confirming the functionality of the Pm promoter in P. fluorescens. The presence of insoluble MBPeGFP can be caused by its overexpression leading to protein aggregation. SDS-PAGE analysis of the cell lysate derived from P. fluorescens transformed with pSEVA438-Ptet-creA-crnA revealed a clearly visible band at the expected size of CreA in both soluble and insoluble fractions when its expression is induced. As expected based on the leaky expression of MBPeGFP, CreA is also produced in lower quantities without the inducer in the soluble protein fraction. In contrast, there is no visible band at the expected size of CrnA, suggesting that the crnA gene is not or poorly expressed.
New questions arise
Both genes are expressed from an operon system, i.e., from a single transcription cassette. In such genetic systems, the expression level of the downstream gene, here crnA, can be lower than that of the upstream gene, creA.
The engineered strain Pseudomonas fluorescens SBW25 expressing creatinine amidohydrolase (CrnA) and creatinase (CreA) grows on creatinine
Objectives
Our goal was to implement the metabolic pathway of creatinine degradation in P. fluorescens and to assay the growth of the engineered strain on creatinine, creatine, or sarcosine.
Method
We studied the growth dynamics of P. fluorescens on each compound of the metabolic pathway of creatinine degradation into sarcosine separately, namely creatinine, creatine and sarcosine. The M9 minimal medium did not contain NH4 Cl and growth was monitored by measuring OD at 600 nm in a 96-well microplate with a microplate reader. Production of CreA and CrnA was induced during the pre-culture to avoid a too long lag phase during growth curve measurements.
Hypothesis
We expected growth of our engineered strain on creatine after induction with m-toluic acid. We did not expect growth on creatinine even after induction with m-toluic acid because of the absence of CrnA on the SDS-PAGE (Figure 7). Bacteria transformed with the empty plasmid should not grow either (negative control).
Results
Growth curves of three different strains of Pseudomonas fluorescens were monitored on M9 minimal medium supplemented with creatinine, creatine, or sarcosine as sole carbon and nitrogen sources:
- the wild-type strain (negative control),
- a strain transformed with the empty plasmid pSEVA438-pTET (negative control),
- a strain transformed with the plasmid pSEVA438-pTET creA-crnA (engineered strain).
No growth was observed in either the WT or the negative control strains when cultured with creatinine, confirming that our strain P. fluorescens SBW25 does not naturally metabolize this compound. In contrast, the engineered strain harboring the creA and crnA genes exhibited clear growth on creatinine, demonstrating that the introduced metabolic pathway is functional with a growth rate of 0.06 h⁻¹ (Figure 9).
In analogy with the results obtained with creatinine, neither the WT nor the negative control strains grew on creatine. However, the strain engineered with creA-crnA operon showed moderate growth, confirming that the transformed bacteria could also metabolize creatine. The same growth has been observed for growth on creatine as a sole source of carbon and nitrogen.
However, the growth rate on sarcosine (0.10 h⁻¹) was higher than on creatinine, indicating that the pathway for creatinine degradation may not be as efficient as for sarcosine. This suggests potential possibilities to optimize the metabolic pathway, as differences in growth rates point to bottlenecks or suboptimal expression of the enzymes involved in creatinine degradation.
Using sarcosine as a substrate, all three strains could grow, as expected. Growth rates (Figure 9) of the strains transformed with the empty plasmid (0.11 h⁻¹) or with the pSEVA438-Ptet-creA-crnA plasmid (0.10 h⁻¹) were lower than that of the WT strain (0.15 h⁻¹), indicating that plasmid introduction had an adverse effect on the growth rate. This could be explained by a metabolic burden on the bacteria caused by the production of CreA and CrnA , along with the cost of maintaining antibiotic resistance.
Overall, these results clearly demonstrate that transformation of P. fluorescens with the plasmid pSEVA438-pTET-creA-crnA enabled the bacteria to grow on creatinine and creatine, whereas the WT and negative control strains could not utilize these nutrients. This validates our engineering approach and the constructed DNA creatinine parts.
New questions arise
The slower growth rates on creatinine compared to sarcosine suggests room for further optimization to improve the pathway efficiency. For instance, we suggest to increase the activity of creatinase (CreA).
CrnA, the first enzyme in the creatinine pathway (Figure 3), catalyzes the hydrolysis of creatinine into creatine with a catalytic efficiency (kcat/Km) of 1.48 s⁻¹·mM⁻¹. This high efficiency makes CrnA6 highly effective in processing creatinine. In contrast, CreA7, which catalyzes the conversion of creatine to urea and sarcosine, has a much lower catalytic efficiency of 0.02 s⁻¹·mM⁻¹, mainly due to its low catalytic constant kcat (0.246 s⁻¹). This significant difference creates a bottleneck in the pathway, where creatine accumulates more quickly than it can be metabolized by CreA.
The genetic organization as an operon has a notable effect on the expression levels of the enzymes. CreA, located upstream of CrnA in the operon, is expressed at higher levels, as demonstrated by SDS-PAGE analysis (Figure 7). This difference in expression compensates for the lower catalytic efficiency of CreA compared to CrnA, helping to balance the overall metabolic process. This likely explains why in spite of the poor expression of CreA, the construction successfully allowed creatinine consumption.
To further optimize the pathway, we propose redesigning the genetic organization of our construct (Figure 5). Currently, both creA and crnA are expressed from the same operon, which limits our control over the individual expression levels of these enzymes. By placing creA and crnA under the control of their own distinct promoters and terminators, we could independently regulate the expression levels of each enzyme.
To better understand the pathway and optimize the genetic construct, it would be crucial to gain more insights into the uptake fluxes of each substrate (sarcosine, creatine and creatinine) in the transformed bacteria. By quantifying how efficiently sarcosine, creatine, and creatinine are imported into the cells, we could identify potential bottlenecks that affect growth rates, which could further inform us to redesign the metabolic pathway.
NMR Analysis of Creatinine, Creatine, and Sarcosine Uptake and Degradation in engineered strain Pseudomonas fluorescens SBW25
Objectives
The objective of these experiments was to monitor and quantify the degradation of sarcosine, creatine, and creatinine by Pseudomonas fluorescens transformed with the creA-crnA plasmid, and to compare the experimental growth and degradation rates to those predicted by our in silico metabolic model.
Method
Supernatants from bacterial cultures were collected during the exponential growth phase and analyzed using 500 MHz NMR spectroscopy to monitor the concentrations of sarcosine, creatine, and creatinine over time. Growth rates and uptake rates were calculated using Physiofit8.
Hypothesis
We expected that Pseudomonas fluorescens carrying the creA-crnA plasmid efficiently degrades sarcosine, creatine, and creatinine. Moreover, we expected the uptake flux to be similar in every condition.
Results
Sarcosine degradation:
As shown in Figure 10, the experimental data reveals a clear decrease in sarcosine concentration over time, which can be attributed to the uptake and subsequent degradation by the cells. A sarcosine uptake rate of 3.6 mmol.h⁻¹.gDW⁻¹ was obtained and the measured growth rate under this condition was 0.10 h⁻¹ (n=1). The lower growth rate observed in the transformed Pseudomonas fluorescens strain on sarcosine (0.10 h⁻¹) compared to the wild-type (0.19 h⁻¹) in the same condition (growth in Erlenmeyer flasks) can likely be attributed to the metabolic burden imposed by the plasmid carrying the creA-crnA genes.
To validate the predictability of our metabolic model, we used the experimentally measured sarcosine uptake rate as an input parameter and the model predicted a growth rate of 0.11 h⁻¹, which is in perfect agreement with the experimentally observed rate.
Creatine degradation:
The degradation of creatine as the only substrate was also monitored by NMR (Figure 11). The experimental data fit well with the simulated kinetics, further validating our metabolic model for creatine degradation. A creatine uptake rate of 2.05 mmol.h⁻¹.gDW⁻¹ was obtained and the measured growth rate under this condition was 0.05 h⁻¹ (n=1). To validate the predictability of our metabolic model, we also used the experimentally measured creatine uptake rate as an input parameter and the model predicted a growth rate of 0.08 h⁻¹, which remains in agreement with the experimentally observed rate.
Creatinine Degradation:
Lastly, the degradation of creatinine was measured (Figure 12). Although the uptake rate of creatinine was higher than that of sarcosine and creatine, the observed growth rate was, in fact, lower. This indicates a lack of optimisation of the creatinine degradation, which may be due to a low expression of the gene crnA, as discussed above. A creatinine uptake rate of 3.58 mmol.h⁻¹.gDW⁻¹ was obtained and the measured growth rate under this condition was 0.03 h⁻¹ (n=1). To validate the predictability of our metabolic model, we also used the experimentally measured creatinine uptake rate as an input parameter and the model predicted a growth rate of 0.14 h⁻¹, which is higher in comparison to the experimentally observed rate.
Conclusion:
These results demonstrate that Pseudomonas fluorescens carrying the creA-crnA plasmid can degrade sarcosine, creatine, and creatinine, and utilize each of them as a sole source of carbon and nitrogen. As summarized in Table 1, the experimental data from NMR were in good agreement with the simulations for sarcosine and creatine uptake, validating our model and overall approach.
New questions arise
Take a look at the Figure 3 for a better understanding.
- The 3 co-products of the creatinine degradation pathway are:
- Urea, produced during the conversion of creatine to sarcosine, is naturally degraded by P. fluorescens9. This presents an opportunity to exploit the ammonia released during urea degradation for the production of nitrate — a critical nutrient for plants. This strategy has been integrated in the design of a nitrate production module as a perspective to enhance plant growth.
- Hydrogen peroxide, another by-product, is known to induce oxidative stress in bacteria, potentially inhibiting growth. To mitigate this, P. fluorescens could be engineered to become more tolerant to oxidative stress by incorporating stress-resistance genes. We designed a stress resistance module to help bacteria cope with elevated H₂O₂ levels.
- Formaldehyde, a toxic by-product, may pose a growth challenge. However, according to Dr. Charles Cockell, with whom we met on the 30th of May, formaldehyde may be absorbed by regolith, reducing its impact on bacterial survival. Therefore, formaldehyde toxicity might not be a significant issue.
Engineering a Pseudomonas fluorescens strain that can cope with oxidative stresses
To go further in optimizing the project, we wanted to reduce oxidative stress-related damages by overexpressing the catalase (KatB) that degrades H2O2 which is produced during glycine synthesis from sarcosine. H2O2 generates superoxide radicals, which can damage cellular components, including DNA, proteins, and lipids. In addition, we knew that regolith is a hostile environment for both plants and bacteria. To help P. fluorescens overcome the stress it could be exposed to, we intended to stimulate the global stress response by overexpressing the native stress factors hfq and rpoS.
Oxidative stress has an impact on bacterial growth
Objectives
The first objective was to observe how P. fluorescens reacted to oxidative stress.
Method
We monitored growth of Pseudomonas fluorescens during the exponential phase on M9 minimal medium supplemented with sarcosine (56 mM) under varying concentrations of paraquat, a molecule known to induce oxidative stress.
Hypothesis
We expected Pseudomonas fluorescens to be sensitive and to stop growing at high paraquat concentrations.
Results
As shown in Figure 13.A, growth of Pseudomonas fluorescens was affected by paraquat in a concentration dependent manner, with a strong inhibition from 1 mM.
Calculated growth rate dropped from 0.16 h⁻¹ at 0 mM of paraquat (control) down to 0.03 h⁻¹ at 10 mM (Figure 13.B), following the reduction of biomass measured at OD600. This result indicates that the WT strain is significantly affected by oxidative stress. Interestingly, bacteria still exhibited measurable growth at 0.1 mM paraquat, suggesting some natural ability to cope with moderate levels of oxidative stress, likely through the activation of detoxification mechanisms mediated by catalases.
New questions arise
To mitigate the oxidative stress caused by H₂O₂ during creatinine degradation, an effective strategy would be to enhance the bacterial strain’s resistance to radical oxidative stress. P. fluorescens could be engineered to overexpress the katB gene which encodes a catalase, an enzyme responsible for the degradation of hydrogen peroxide (H₂O₂).
Construction of the plasmid pSEVA244 cloned with the gene katB and transformation of Pseudomonas fluorescens SBW25
Objectives
The aim was to clone the gene katB into the plasmid pSEVA244 for transformation of P. fluorescens.
Method
katB was amplified from P. fluorescens genomic DNA via PCR. pSEVA244 served as a backbone for cloning of katB by In-fusion prior transformation of E. coli Stellar Competent Cells.
Hypothesis
Transformation of P. fluorescens with pSEVA244-katB was expected to yield overexpression of the KatB protein.
Results
After transformation in E. coli Stellar Competent Cells and isolation, the plasmid was digested using the enzymes EcoRI and HindIII, which validated the cloning (Figure 15). The correct sequence was validated by Sanger sequencing.
Attempts to overexpress a catalase in Pseudomonas fluorescens SBW25
Objectives
The objective was to verify the overproduction of the catalase KatB in P. fluorescens by activation of the Ptrc promoter with the inducer IPTG.
Method
The pSEVA438-MBPeGFP plasmid was employed again as positive control for heterologous protein expression in P. fluorescens SBW25. The pSEVA438-MBPeGFP- and pSEVA244-katB-transformed P. fluorescens SBW25 strains were cultured in M9 minimal medium supplemented with glucose (28 mM), with or without 0.5 mM of m-toluic acid or IPTG (1 mM). After incubation, a whole-protein extraction was performed for each strain to assess the level of expression, as well as the solubility of our proteins..
Hypothesis
We expected that MBPeGFP and KatB were activated in the presence of m-toluic acid and IPTG, respectively. These proteins should be soluble with an expected molecular weight of 69 kDa and 57 kDa, respectively.
Results
The obtained SDS-PAGE is presented in Figure 16. Both soluble and insoluble fractions contained inducible MBPeGFP, with the majority of protein being in the soluble fraction independently of the presence of the inducer. SDS-PAGE analysis of the cell lysate derived from P. fluorescens transformed with pSEVA244-katB revealed a faint but visible band at the expected size of KatB in both soluble and insoluble fractions when its expression was induced (Figure 16).
.
New questions arise
The presence of proteins in the insoluble fraction suggests that the conditions of production (e.g., inducer concentration and temperature) are not optimal for overexpressing KatB in the soluble form. After aligning the sequence of KatB with database PFAM, we noticed the presence of a peptide signal, suggesting the targeting of KatB in the periplasm. In our lysis protocol, we could not differentiate periplasmic proteins from cytoplasmic ones. To solve this problem, we could change the KatB production conditions and our lysis protocol to purify exclusively periplasmic proteins.
If the protein is indeed located in the membrane, it could help our P. fluorescens strain metabolizing creatinine to survive the increasing concentrations of H2O2. In order to prove that the engineered bacterium could better survive certain conditions, the oxidative impact test should be repeated, as well as a viability assay on regolith. Plant growth should also be assayed in the presence of P. fluorescens pSEVA244-KatB. We expect plants will live longer after overexpression of the catalase.
To mitigate stresses, an effective strategy could be to enhance the bacterial strain’s resistance to general stress. P. fluorescens could be engineered to overexpress the hfq and rpoS native genes which encodes respectively for Hfq, a chaperon protein, and RpoS, a transcriptional factor involved in the stress response. We performed preliminary experiments in this direction but we failed to construct the expression plasmid. Likewise, we intended in the beginning of the project to improve the biofilm production of P. fluorescens to counter the poor water retention of regolith, but we encountered cloning problems.
Testing the biostimulant properties of Pseudomonas fluorescens for plants on regolith
In this section, we describe our experiments with plants.
- We addressed three questions:
- How do plants grow on regolith?
- Does P. fluorescens have a positive effect on plant growth on regolith?
- Do our modifications enabling creatinine consumption by P. fluorescens impact the growth of plants?
Plants grow poorly on regolith
Objectives
This first experiment aimed to observe the negative effect of lunar-simulant regolith on plant growth. It was an important step to ensure that our setup to grow plants was effective.
Method
We monitored the growth of plants on fertile ground or regolith (LSP-2 Lunar South Pole Simulant). Arabidopsis thaliana Col-0 was grown for 22 days after germination, watered every day with 3 mL of osmotic water. Pictures of plants were taken during the process.
Hypothesis
Based on literature10, we expected that the regolith condition would result in a strong development delay compared to the soil condition.
Results
On regular potting soil, we obtained green and healthy plants which grew vigorously (Figure 17). The plant on our lunar regolith simulant showed a slower development and a browner color. The smaller size and area of the rosette is due to the lack of nutrients present in the regolith. Under starving conditions, the leaves tend to be smaller11 and the color browner. The change in color is due to the production of anthocyanins, a pigment indicative of abiotic stresses. Because the lightning, humidity, and temperature were designed to be as optimal as possible for the plant, the production of anthocyanins could be explained by the lack of nutrients on regolith12.
This experiment confirmed that regolith is an unfavorable environment for plant growth. The use of a biostimulant to improve the development and limit stress is relevant.
Pseudomonas fluorescens WT increases the fraction of healthy leaves and plant hue on regolith
Objectives
The objective of this second experiment was to observe if P. fluorescens WT could relieve the negative effect of lunar-simulant regolith on plant growth.
Method
We monitored the growth of plants under three different conditions by varying the soil (fertile ground or regolith), and the presence or not of the WT strain on the regolith.
-
This experiment also allowed us to define which metrics to use to quantify the growth of plants:
- Proportion of healthy leaves
The proportion of healthy leaves were counted every day and used to analyze the plant’s survival. - Plant area
We determined the plant area for each plant. Pictures were taken every day and analyzed using ImageJ 2.7.0 version software. To calculate the plant area, the contour of each plant was cut out precisely on the image and the area was measured with the Measure tool in the Analyze tab. - Color hue
We determined the color for each plant. To measure the color, we used the Color_Histogram-2.7.0.jar plugin from ImageJ to extract the mean value of Red, Green and Blue pixels present in each pixel of the cropped plant. Using the rgb2hsv command from the grDevices package on R Studio, we converted RGB space (Red/Green/Blue) to HSV space (Hue/Saturation/Value). Only the Hue value was used in our analysis, which correlates with chlorophyll content and is suitable to estimate the stress13.
Hypothesis
Based on literature14, we hypothesized that the presence of P. fluorescens should improve the growth and/or health of the plants on regolith.
Results
Images of plants at different stages are shown for the tested conditions in Figure 19. Addition of P. fluorescens on regolith moderately increased the number of leaves per plant after 14 days compared to the control without bacteria.
To collect more quantitative information about the impact of stress on the phenotype of plants, we also analyzed the proportion of healthy leaves per plant, the rosette area, and the plant hue. While the percentage of healthy leaves per plant decreased to 40 % after 8 days in the absence of bacteria on regolith, it remained higher than 80 % for plants growing on soil or on a regolith containing bacteria (Figure 20).
Therefore, the addition of Pseudomonas fluorescens WT on regolith results in plants that wilts significantly less than on regolith only. Furthermore, after 14 days, plants grown on regolith without bacteria exhibited the browner color and the smallest size (Figure 21), two characteristics of stress.
From these results, we conclude that P. fluorescens improves the growth and health of the plants on regolith.
Transformed P. fluorescens may have a protecting role on plant when grown on regolith
Objectives
This last experiment aimed to examine whether (i) our modification of P. fluorescens affects plant growth and (ii) the use of chemical agents to maintain the plasmid in the strain and to trigger the creatinine pathway expression could have some impact too.
Method
We monitored the growth of plants under four different conditions by varying the bacteria (no bacteria, the strains transformed with the empty plasmid or creA-crnA plasmid), and the watering solution (water or a mix composed of creatinine 44 mM and streptomycin (1 µg/mL)).
Table reporting the tested conditions:
Condition |
Comment |
|---|---|
| Regolith + No bacteria + Water | Negative control |
| Regolith + No bacteria + Mix | Negative control for the “Mix” condition. This experiment showed us the impact of the mix alone on plant growth. |
| Regolith + Empty plasmid + Mix | Negative control for the “creA-crnA” condition. This experiment showed us the impact of bacteria in the presence of creatinine on plant growth. |
| Regolith + creA-crnA plasmid + Mix | This experiment showed us the impact of bacteria in the presence of creatinine that was metabolized to sustain bacterial growth on plant growth. |
Hypothesis
It is possible that the implementation of the creatinine pathway in P. fluorescens results in oxidative stress because of the production of H2O2 associated with glycine biosynthesis. Likewise, streptomycin and creatinine addition could have some impact too. Another scenario could be that the wild-type strain inoculated with water only and the bacteria transformed with the empty plasmid could have a lower positive effect since both strains are not able to metabolize creatinine, hence to grow.
Results
After 8 days on regolith, it appeared that the four conditions led to different phenotypes (Figure 22). This required investigating each measured parameter separately (healthy leaves, plant surface, and hue).
.
The proportion of healthy leaves per plant was clearly lower without P. fluorescens and in the presence of creatinine and streptomycin mix (Figure 23), pointing to a negative effect of one or both compounds. Notably, a clear positive effect of bacteria can be seen, leading to >60 % of healthy leaves after 9 days. The strain with the functional creatinine pathway did not protect more than the strain with the empty plasmid. This suggests a possible detoxification of the soil through the degradation of streptomycin by the cells.
To strengthen these observations, we carried out a survival analysis by using the Kaplan-Meier estimator (Figure 24). This allowed us to estimate the probability to reach an event in function of time. Here, the event of interest we considered is the time an individual reached less than 45 % of healthy leaves. The survival curves indicated that after 8 days, the probability to have less than 45 % of healthy leaves was between 0.8 and 0.9 for all conditions, except the No bacteria+Mix condition that led to a probability of 0.25 (Figure 24A). This indicates that the probability to have a healthy plant is a lot lower without bacteria and with the mix.
We coupled this analysis with a Cox proportional Hazard model (Figure 24C), which allowed us to evaluate simultaneously the effect of several factors on survival. The results are given by the exp(coef), which corresponds to the Hazard Ratio (HR). HR gives an estimation of how often a condition reaches more than a half (55 %) of wilted leaves over time. Here, the conditions No bacteria+Mix, the Empty plasmid+Mix and the creA-crnA plasmid+Mix were compared to the control No bacteria+Water.
- For interpretation:
- HR = 1: No difference in the wilting compared to the control group over time.
- HR < 1: Reduction of the risk it gets 55 % of wilted leaves compared to the control group over time.
- HR > 1: Increased risk of getting 55 % of wilted leaves compared to the group control over time.
For the No bacteria+Mix, we obtained an HR=8, which means that this condition gives a high risk of getting 55 % of wilted leaves compared to the No bacteria+Water condition. For the bacteria conditions, the HR < 1 indicates that transformation with either the empty plasmid or creA-crnA plasmid reduces the risk of getting 55 % of wilted leaves compared to the No bacteria condition. This result suggests a negative effect of streptomycin.
Because the Cox proportional hazards model makes several assumptions, we needed to validate them for each of the covariable sets (Figure 24B). We proved the validity of our model by reaching p-values < 0.05 for each covariable.
The second metric used was the plant hue (Figure 25). The hue works like a binary metric: plants are either stressed (hue < 0.15) or not (hue > 0.15). For this study, dead plants have been arbitrarily set to a value of -0.025.
The negative control No bacteria+Water may have led to a high production of anthocyanins, hence the purple color (Figure 25, day 8). The three other conditions presented a more pronounced yellowing. With this color analysis, it is impossible to say whether yellowing or purpuling is a stronger stress mark. After 8 days, every condition showed signs of stress (mean hued < 0.15) on regolith. Notably, the conditions with bacteria transformed with either the empty-plasmid or creA-crnA displayed less stress than the condition No bacteria+Mix, with an averaged hue value at 0.14, corroborating the results described above.
To get a complete response of the plan phenotype to stress exposure, we combined the hue metric with plant surface (i.e., the third metric used in this analysis) (Figure 26). This provided a better discrimination of each plant state. At day 1, the hue indicates an absence of stress with some variability in the plant area. At day 8, the profiles are a lot more varied, indicating specific responses. The condition No bacteria+Mix clearly led to death. Condition No bacteria+Water resulted in a strong stress and lesser plant development. Conditions creA-crnA-plasmid+Mix and especially empy-plasmid+Mix resulted in greener hue and larger plant area.
These results demonstrate that our engineered strain of P. fluorescens is endowed with protective properties for the plant development on regolith. Further improvements will be necessary to take full advantage of the implemented creatinine pathway as it seems to create some more stress to plants than the strain transformed with the empty plasmid.
New questions arise
In order to keep the plasmid into the bacteria, Streptomycin was present in the Mix. However, studies demonstrated the toxicity of Streptomycin for plants. The plasmid pSEVA438 contains the aadA gene coding an adenylyltransferase enzyme that O-adenylates the position 3″ of streptomycin and deactivates its antimicrobial activity. Therefore, bacteria transformed with this plasmid may have a protecting role on plants by inactivating streptomycin.
During the metabolization of creatinine, H2O2 is produced by bacteria. If H₂O₂ is released into the medium, it would trigger the plant stress response and would explain the observed sign of stress compared to the strain transformed with the empty plasmid.
New potential research
To mitigate the oxidative stress caused by the release of H₂O₂ during creatinine degradation, an effective strategy would be to enhance the bacterial strain’s resistance to Reactive Oxygen Species. Overexpressing the catalase gene katB in P. fluorescens would be a way to degrade H₂O₂ and avoid its adverse effect on plant growth.
In order to avoid having to add Streptomycin to preserve the plasmid in the bacteria, chromosomal integration of creA-crnA should be considered.
Conclusion
In this BioMoon project, we choose to optimize the properties of Pseudomonas fluorescens as a biostimulant for growth in poor soil conditions like the lunar regolith.
Our main goal was to engineer the strain so that it could utilize non-recycled resources in a space base. We identified creatinine as a high-value compound unconsidered so far in spatial programs and that could serve as both carbon and nitrogen source. We demonstrated that Pseudomonas fluorescens is not able to naturally grow from creatinine. A synthetic creatinine pathway was successfully implemented in the bacterium and the properties of the resulting Pseudomonas fluorescens were thoroughly characterized both in silico and in wet lab experiments.
We set up a home-made system to ensure efficient and reproducible plant assays. We qualitatively and quantitatively analyzed the deleterious effect of regolith on plant growth, as well as the positive impact of Pseudomonas fluorescens. We managed to observe that our strain engineered with the creatinine pathway has kept most of its protective properties. We discussed how it could be optimized further to provide stronger stress resistance. Several strategies to reduce Pseudomonas fluorescens oxidative stress have been proposed and are currently in progress.
This work paves the way to future developments in using synthetic biology for the production of efficient biostimulants, both on Earth and in space. Please visit our entrepreneurship page to learn more about it.
- Tsuru, D., Oka, I., & Yoshimoto, T. (1976). Creatinine Decomposing Enzymes in Pseudomonas putida. Agricultural And Biological Chemistry, 40(5), 1011‑1018. Click
- Tsuru, D., Oka, I., & Yoshimoto, T. (1976). Creatinine Decomposing Enzymes in Pseudomonas putida. Agricultural And Biological Chemistry, 40(5), 1011‑1018. Click
- Cloning, Expression and Purification of Pseudomonas putida ATCC12633 Creatinase. (2017, 1 décembre). PubMed. Click
- Tsuru, D., Oka, I., & Yoshimoto, T. (1976). Creatinine Decomposing Enzymes in Pseudomonas putida. Agricultural And Biological Chemistry, 40(5), 1011‑1018. Click
- Vogeleer, P., Millard, P., Arbulú, A. O., Pflüger-Grau, K., Kremling, A., & Létisse, F. (2023). Metabolic impact of heterologous protein production in Pseudomonas putida : Insights into carbon and energy flux control. Metabolic Engineering, 81, 26‑37. Click
- Information on EC 3.5.2.10 - creatininase - BRENDA Enzyme Database. (s. d.). Click
- Information on EC 3.5.3.3 - creatinase - BRENDA Enzyme Database. (s. d.). Click
- PhysioFit | Anaconda.org. (s. d.). Click
- Information on EC 3.5.1.5 - urease - BRENDA Enzyme Database. (s. d.). Click
- Paul, A., Elardo, S. M., & Ferl, R. (2022). Plants grown in Apollo lunar regolith present stress-associated transcriptomes that inform prospects for lunar exploration. Communications Biology, 5(1) Click
- Kolář, J., & Seňková, J. (2008). Reduction of mineral nutrient availability accelerates flowering of Arabidopsis thaliana. Journal Of Plant Physiology, 165(15), 1601‑1609. Click
- Diaz, C., Saliba-Colombani, V., Loudet, O., Belluomo, P., Moreau, L., Daniel-Vedele, F., Morot-Gaudry, J., & Masclaux-Daubresse, C. (2005). Leaf Yellowing and Anthocyanin Accumulation are Two Genetically Independent Strategies in Response to Nitrogen Limitation in Arabidopsis thaliana. Plant And Cell Physiology, 47(1), 74‑83. Click
- Information on EC 3.5.3.3 - creatinase - BRENDA Enzyme Database. (s. d.-b). Click
- Paul, A., Elardo, S. M., & Ferl, R. (2022). Plants grown in Apollo lunar regolith present stress-associated transcriptomes that inform prospects for lunar exploration. Communications Biology, 5(1). Click