In this page, you will find a detailed description of the four different engineering cycles we put in place as well as the multiple iterations we went through over the course of the project.
Introduction
BioMoon aimed to develop a biostimulant that would allow plant growth on the Moon using only resources available on a lunar space station. To this end, we engineered the plant growth-promoting rhizobacteria Pseudomonas fluorescens to be able to metabolize creatinine. Furthermore, we designed new paths that could further improve plant growth. We put into practice the D-B-T-L cycle (Design-Build-Test-Learn) and were able to learn from our tests, redesign and rebuild the different elements of our project.
Creatinine
DESIGN
The essence of our project lies in the engineering of different pathways in P. fluorescens to improve its metabolism to support plant needs in the adverse environment of lunar regolith. To effectively complement the project, the modeling strategy reflects this modular structure by testing each modification individually and then integrating them together alternatively to obtain meaningful insights (Figure 1).
In Pseudomonas putida, the creatinine metabolic pathway consists of three enzymes: creatinine amidohydrolase to convert creatinine intro creatine, creatinase to convert creatine into sarcosine (while forming urea), and sarcosine oxidase to convert sarcosine into glycine (with formaldehyde and hydrogen peroxide as byproducts)1. This amino acid can then be used as carbon and nitrogen sources. However, this metabolic pathway has not been found in Pseudomonas fluorescens. Thus, we aim to introduce into a plasmid three genes coding for the aforementioned enzymes: crnA, creA and soxA. Note that we are using the soxA gene, originating from Bacillus subtilis, because the enzyme is made of only one subunit contrary to the one of Pseudomonas putida consisting of four subunits (SoxABCD).
BUILD
As the genes crnA and creA are present in an operon in some Pseudomonas species, we mimicked this by joining the two genes under the same promoter Pm which is inducible by m-toluic acid. The third gene, soxA, was under transcriptional control of another promoter, Ptet, inducible by anhydrotetracycline.
In this construct, the two inserts tap-to-tail are under the control of the same terminator, LUZ7 T50 (BBa_K4757058)2, which is bidirectional. The three genes were codon-optimized for P. putida using the Codon Optimization Tool from Integrated DNA Technologies (IDT) to lower complexity and minimize secondary structures. The three genes were introduced in a single plasmid by In-Fusion Assembly, and we designed primers with floating tails overlapping adjacent regions for each insert.
TEST
We first tested the growth of our bacteria on M9 minimal medium without nitrogen source supplemented with each metabolite of the pathway individually: glycine, sarcosine, creatine, or creatinine. The goal was to test if these molecules can be used as a carbon and nitrogen source. Additionally, to further test the feasibility of our strategy, we performed in silico simulations of bacterial growth, which aided in designing better informed in vivo experiments. See modeling here.
After construction, our plasmids were sequenced by Sanger sequencing. Once the constructs were validated, we repeated the growth experiments with the modified P. fluorescens to quantify the effect of our plasmid. Finally, SDS-PAGE was performed to assess the expression of the three enzymes introduced.
LEARN
After testing the growth of P. fluorescens on a minimal medium with creatinine, we concluded that P. fluorescens WT is unable to utilize creatinine as its sole carbon and nitrogen source. See results here. In light of the results obtained in silico, which predicted growth of the WT strain on sarcosine, we decided to also test its growth on creatine and sarcosine, the downstream products of the metabolic pathway. While no growth was observed on creatine, the bacteria were able to grow on sarcosine, thus confirming the predictions of the in silico model. See modeling here
DESIGN 2.0
Based on these findings, we decided to simplify our construct. Instead of introducing all three genes as initially planned, we opted to first clone the operon containing creA and crnA into the plasmid pSEVA438-Ptet. This allowed us to focus on the initial steps of the pathway, from creatinine degradation up to sarcosine. If necessary, the gene soxA could be introduced at a later stage to complete the pathway and enable the conversion of sarcosine into glycine.
BUILD 2.0
The new plasmid, without the soxA gene was built following this representation:
TEST 2.0
After cloning and transformation of P. fluorescens with the pSEVA438-Ptet-creA-crnA plasmid, we conducted the same growth experiments to verify whether the engineered strain was able to grow on each of the four metabolites of the pathway: creatinine, creatine, sarcosine and glycine. In silico, the genes were added to the GSMM in order to simulate the growth of the new strain.
LEARN 2.0
The new round of growth tests, this time on the engineered strain, showed significant growth of bacteria on creatinine, creatine, or sarcosine as sole carbon and nitrogen sources. We have successfully engineered Pseudomonas fluorescens to metabolize a urinary compound! The fact that the strain was able to grow on creatinine without cloning of the soxA gene further supports that there was no real need for it in our initial design.
DESIGN 3.0
With this confirmation, we envisioned a simplified pathway that only required two genes for sustainable growth on creatinine: crnA of the creatinine amidohydrolase, and creA of the creatinase.
BUILD 3.0
Because the two genes are part of the same operon, we decided to clone them into plasmid pSEVA438, under the control of the Pm promoter which is inducible by m-toluic acid. The new construct looked like this:
In this construct, the two inserts tap-to-tail are under the control of the same terminator, LUZ7 T50 (BBa_K4757058), which is bidirectional. The three genes were codon-optimized for P. putida using the Codon Optimization Tool from Integrated DNA Technologies (IDT) to lower complexity and minimize secondary structures. The three genes were introduced in a single plasmid by In-Fusion Assembly, and we designed primers with floating tails overlapping adjacent regions for each insert.
TEST 3.0
The two genes were introduced in the plasmid by In-Fusion Assembly. Further experiments will be required to assay bacterial growth with this new strain.
Nitrates
Like bacteria, plants need nutrients to grow. Their carbon source is gaseous CO2 and their main source of nitrogen is located in the soil3. In a lunar space station, CO2 will be provided directly by astronauts’ respiration. However, lunar regolith is not composed of any nitrogenous compound. This is why nitrates have to be provided externally. The chosen solution is to have our Pseudomonas fluorescens-based biostimulant producing these nitrates by nitrifying the urea resulting from creatinine metabolization. In short, this design aims to implement nitrate-synthesis metabolic pathways into P. fluorescens.
DESIGN
Nitrification in soil is typically viewed as a two-step process, where ammonia, in our case provided by creatinine degradation, is initially oxidized to nitrite by ammonia-oxidizing bacteria (nitritation), followed by the oxidation of nitrite to nitrate by nitrite-oxidizing bacteria (nitratation).
To choose the best ammonia- and nitrite-oxidizing genes, we looked into the nitrifying subsystem of the MELiSSA4 loop, consisting of a co-culture of Nitrosomonas europaea ATCC® 19718 and Nitrobacter winogradskyi ATCC® 25391.“MELiSSA (Micro-Ecological Life Support System Alternative) is the European project of circular life support system5. (...) The MELiSSA Project is to develop a closed system, inspired by Earth ecosystem, aiming to reproduce its main functions with highly reduced mass and volume, higher kinetics and of course under extreme safety conditions.”
- We decided to select three nitrifying genes from these microorganisms and to engineer them in our Pseudomonas fluorescens-based biostimulant:
- First, the ammonia monooxygenase (Amo, iGEM13_DTU-Denmark BBa_K1067003)6 from N. europaea, consisting of three subunits: CAB. This enzyme catalyzes the conversion of ammonia originating from the degradation of creatinine into hydroxylamine.
- Then, the hydroxylamine dehydrogenase (Hao, iGEM13_DTU-Denmark BBa_K1067004)7 from N.europaea. This enzyme catalyzes the conversion of hydroxylamine into nitric oxide, this compound is then spontaneously oxidized into nitrite.
- Finally, the nitrite oxidoreductase (Nxr) from N.winogradskyi, consisting of three subunits AXB. This enzyme catalyzes the final oxidation of nitrites into nitrates.
BUILD
The amoCAB and hao existing parts (iGEM13_DTU-Denmark BBa_K1067003 and BBa_K1067004) and the three nxr subunits AXB were optimized for expression in P. putida (the closest host organism from P. fluorescens available on the sequence optimization tool)thanks to the Codon Optimization Tool from IDT. The optimized amoCAB and hao sequences were cloned in antisense orientation into the same pSEVA438-Ptet vector and separated by the LUZ7 T50 bidirectional terminator (iGEM23_Heidelberg BBa_K4757058)8. They are under the control of Pm and Ptet promoters (BBa_R0040)9, respectively. The three optimized nxrAXB ORFs were cloned into the pSEVA244 vector under the control of the Ptrc promoter (iGEM20_XMU-China BBa_K3332038)10. Within the same plasmid, each ORF was cloned downstream of a different Pseudomonas sp. RBS (BBa-K510800611, BBa-K510800712, BBa-K510800813), selected to be effective in our strain. amoCAB and nxrAXB subunit ORFs were cloned in operons, as found in N. europaea and N.winogradskyi, respectively.
TEST AND LEARN
In order to test the nitrate pathway, in silico simulations were performed using the adapted metabolic model of P. fluorescens. In doing so, we identified the presence of the gene for the enzyme nitric oxide dioxygenase. This enzyme is able to catalyze the conversion of NO to NO3-, therefore simplifying the design for the engineering of this pathway. See modeling here.
DESIGN 2.0
Since P. fluorescens is capable of catalyzing the oxidation of nitric oxide into nitrate and, we decided to not implement the nxrAXB genes. In addition, hao ORF (iGEM13_DTU-Denmark BBa_K1067004) and its RBS (BBa-K5108006) were eventually cloned into the pSEVA244 vector under the control of the Ptrc promoter (iGEM20_XMU-China BBa_K3332038) to simplify the construction. Optimized amoCAB ORFs (iGEM13_DTU-Denmark BBa_K1067003) and their RBS (BBa-K5108006, BBa-K5108007, BBa-K5108008) were cloned into the pSEVA438 vector without the LUZ7 T50 bidirectional terminator (iGEM23_Heidelberg BBa_K4757058).
STRESS
DESIGN
Another key aspect of our project is the survival of our bacteria to the hostile environment of the Moon. Our initial hypothesis was that our bacteria will be exposed to two major sources of stress.
On the one hand, lunar regolith will impact the bacteria's growth in different ways, the most notable one being osmotic pressure due to solubilization of ions when water is added. Other causes are not well characterized, which is why we decided to focus our efforts on the global stress response to increase our chances of success.
On the other hand, the creatinine metabolization pathway that we implemented leads to the production of H2O2, which can be toxic for the cell at high concentration. This can be added to the oxidative stress caused by regolith itself. Therefore, we aimed at engineering an efficient H2O2 detoxification pathway.
BUILD
Firstly, we designed a construct to overproduce the KatB enzyme, a native catalase to P. fluorescens, whose function is to decompose hydrogen peroxide into water and oxygen. Overexpression of KatB would promote degradation of H2O2 produced by the creatinine metabolism. We amplified by PCR the katB gene from the bacterial genome and cloned it into the pSEVA 244 vector, under control of the promoter Ptrc which is inducible with IPTG.
The second construct consisted of overexpressing two global stress response regulators : Hfq, a small chaperone protein involved in the regulations of mRNA, and RpoS, a specific polymerase subunit. The two genes were cloned on the same transcription cassette, using the bidirectional terminator TerLuz7 (insert registry reference), and cloned into the vector that we previously designed : pSEVA438 -Ptet. Hfq was under control of the promoter Pm, inducible with m-toluic acid and rpoS under control of the promoter Ptet, inducible with tetracycline.
TEST
The plasmids were cloned via In-fusion Assembly and separately transformed into P. fluorescens SBW25, creating two new strains, each one tackling a different source of stress.
After transformation, we noticed that none of the two strains was able to effectively produce the proteins of interest.
LEARN
Sequencing of the pSEVA438-Ptet-hfq-rpoS plasmid revealed multiple mutations, suggesting that either there had been an error during the cloning process or during proliferation.
Sequencing of the pSEVA244-katB plasmid validated the correct sequence. However, no protein expression in the soluble fraction was detected. Doing additional research we found that KatB is organized in a homotetramer, which may complicate its production as a soluble form.
A sequence alignment also revealed the presence of a peptide signal, possibly changing the location of the enzyme into the periplasm, a paper released on KatB from Pseudomonas aeruginosa14 (an homologous protein of our P. fluorescens catalase) suggests that the enzyme is in the periplasm.
In the meantime we assessed that P. fluorescens growth was not fully inhibited on regolith and could slowly grow on creatinine, despite the toxic byproducts. Therefore, our stress module may improve survival but is not absolutely necessary.
DESIGN 2.0
We decided to redo all the steps of the pSEVA438-Ptet-hfq-rpos cloning, in case the mutations occurred somewhere along the process. As for the overexpression of KatB, testing the different hypotheses was not possible within the timeframe of the project.
Future prospects - Biofilm
Moon regolith’s low water retention limits plant growth. Biofilm provides a potential solution to this problem by enhancing water-holding capacity and improving regolith’s suitability for cultivation. Therefore, we developed a strategy to optimize biofilm formation in Pseudomonas fluorescens.
DESIGN
Introduction
Biofilm formation involves several steps: initial reversible and irreversible attachment to a surface, biofilm maturation, and finally, dispersion15. This transition allows bacteria to shift from a free-swimming planktonic form to a sessile, biofilm-forming state. The extracellular polymeric substances (EPSs), which compose the biofilm, enhance resistance to antibiotics and environmental stress16. Biofilm from soil organisms can alter soil properties, such as water retention, pore size distribution, surface tension, and viscosity17, 18, — key factors in environments like regolith, which poorly retains water. Notably, Pseudomonas fluorescens is capable of biofilm formation, making it relevant to our project19.
Pathway description
At the molecular level, c-di-GMP is a key bacterial second messenger regulating biofilm formation. In Pseudomonas, the Wsp (wrinkly spreader phenotype) signaling pathway is a c-di-GMP-dependent mechanism for biofilm production20. The membrane protein WspA is activated by surface growth or pressure. This causes WspE to phosphorylate itself. Phosphorylated WspE then adds phosphate groups to WspF and WspR. Phosphorylated WspR increases c-di-GMP production, which promotes biofilm formation. Meanwhile, phosphorylated WspF inactivates WspA, reducing WspE activity and lowering c-di-GMP levels to decrease biofilm production. Therefore, inhibiting the transcription of wspF could increase intracellular c-di-GMP levels, thereby promoting biofilm formation.
BUILD
Elevated levels of c-di-GMP in biofilm drive bacteria to rapidly consume large amounts of energy for EPS production, resulting in a reduced metabolic state21. To address this, we chose to employ the CRISPR interference method, which allows for gene knockdown rather than complete knockout22, enabling targeted gene suppression after reaching the desired biomass. This strategy has also been successfully validated in Pseudomonas aeruginosa23. We will use it to prevent the expression of wspF.
Two components are needed: a dead Cas9 (dCas9) and a guide RNA (sgRNA). The dCas9 binds to the sgRNA, which is complementary to a target DNA sequence near the gene's promoter, using a PAM (Protospacer Adjacent Motif) sequence. This binding sterically prevents the promoter from being activated or directly blocking the transcription elongation. It is important to note that the gene is not deleted from the genome since Cas9 is dead; this inactivation is then reversible. To apply this method, two plasmids are required: one for the dCas9 and another for the sgRNA.
The protein dCas9 we used is from Streptococcus pasteurianus, and it proved to be efficient in P. fluorescens24. We will test two different sgRNA, one targeting the supposed promoter sequence and one at the beginning of the coding sequence of wspF. As a control to validate the approach, we will use an mScarlettI plasmid as the gene target and another plasmid containing a designed sgRNA. This control will allow us to quantify the efficiency of the CRISPRi method by fluorescence.
FUTURE TEST
Two parameters need to be assessed: water retention and biofilm formation. First, water retention tests will be conducted on fertile soil and regolith without bacteria to show regolith's poor water retention.
After establishing conditions for biofilm overexpression, we will compare water retention between raw regolith and regolith inoculated with P. fluorescens (wild-type and engineered strains). We expect that engineered bacteria will enhance biofilm formation and, thus, water retention. Biofilm formation assays will use crystal violet dye, which binds to negatively charged molecules, such as acidic polysaccharides and nucleic acids of the extracellular polymeric substance of the biofilm25. This test will be performed on wild-type P. fluorescens and engineered strains with sgRNAs targeting wspF.
Plant experiments
The aim of the tests on plants was to provide proof-of-concept experiments showing that plants can survive on regolith, while exhibiting severe signs of stress and are slowed down in their development. Inoculating P. fluorescens should help plants grow stronger and live longer.
There are many problems to keep in mind when designing plant biology experiments. Here, we summarized the main issues we encountered, how we built a set-up to do our tests, and what we learned from our results.
DESIGN
Arabidopsis thaliana, a model plant
The first problem we encountered when designing our project was time. We knew that it would take weeks to have any meaningful results when it came to the different conditions we wanted to test. We needed to choose a plant that could grow fast, survive well, and still provide valuable results. We found that in most peer-reviewed articles dealing with plant assays, the model plant used was Arabidopsis thaliana. This plant, generally considered as a weed, has a short life cycle (6 weeks, from germination to maturation of the first seeds) and its genotype and phenotype are well characterized26. Discussions with various plant biologists also encouraged us to choose Arabidopsis thaliana for our experiments.
How to obtain significant results
Another issue we anticipated thanks to our extensive bibliographic research was the need for multiple replicates. Assays on plants are extremely variable from one subject to another and it is of the utmost importance to multiplex assays in order to obtain exploitable results. In early May, we had the chance to visit the LIPME (Laboratoire des Interactions Plantes-Microbes-Environnement). During this visit, we gained valuable insight about how to grow our replicates in the limited space of our lab that is neither designed nor equipped for this. Their precious advice coupled with bibliographic research27made us opt for a microplate set-up.
When deciding the number of seeds to plant for each condition to be tested, we had to choose a number high enough to limit the impact of variability in our data analysis, while staying reasonable because of time and space issues. It was essential to ensure the significance and the relevance of the statistical models we wanted to use. Thanks to an online sample size calculator28, we estimated that 6 plants per condition was the minimum to get relevant results. To fill the 24-well microplates we chose, we planted 12 plants per condition.
Sterilizing seeds
The next challenge we faced was the need to sterilize the Arabidopsis thaliana seeds that were donated to us by Sandra Belshiem, a researcher at the LIPME. Because we would add the bacterium Pseudomonas fluorescens in the plant medium to test its impact on growth, it was necessary to start from sterile material, to be sure that the impact we would observe would come from our bacterium only, and not from differences in the microbiomes of the seeds. The protocol we ended up deciding on included one ethanol wash, one bleach wash and several physiological-water washes to remove any trace of bleach.
A controlled environment
Finally, we designed a setup to control the environmental parameters. The purpose was to limit bias and standardize conditions as much as possible.
Firstly, we wanted the plants to grow in a closed place where it would be easy to control temperature, humidity and light. Because the experiments were conducted over the Summer, it was important to ensure that the temperature will not rise too much, which can be deadly for plants.
To be sure to start with only viable seeds we planned to make them germinate before planting only the germinated seedlings, after one week.
BUILD
We decided to place the plants in a room, where the temperature always stays between 21 and 23°C and humidity varies only between 55 and 70 % (controlled with a thermometer and hygrometer). We used a Greenception GC 4 Led Module in order to provide optimal UV lighting for the plants, along with an automatic time–switch in 10-14 hour cycles. The lighting system was hung one meter above the plants.
The seeds were sterilized, germinated, and then planted in 6-well microplates. They were watered everyday, once the lighting had gone off. During the building of our set-up we were lucky enough to receive support from Antoine Berger, a researcher at TBI, our host-lab, who has done extensive work in plant biology before. In 2017, he published an article29 with a plate cultivation method on which we adapted our own. Here is a schematic of the experimental set-up of the growth system:
TEST
Three conditions were tested during our first round of experiments on plants: A. thaliana growing in commercial grade soil, A.thaliana growing in lunar regolith simulant and A.thaliana growing in lunar regolith simulant inoculated with Pseudomonas fluorescens SBW25.
After 14 days, there was a significant difference between the plants on fertile ground and those on regolith. In the latter, plants were clearly smaller and less developed. Most importantly, there was a clear difference between the inoculated plants on regolith and the non-inoculated. This means that P. fluorescens helps the plant grow stronger and live longer.
LEARN
We were able to familiarize ourselves with ImageJ30 (version 1.0) which we used to extract data on the color and surface area of the different plants, with R (version 4.4.1 and the package grDevices, version 4.4.1) which we used to process all our data, and with the watering protocol.
DESIGN 2.0
Having constructed our first engineered strain, we wanted to test the effect it would have on plant growth. This meant that we had the opportunity to change our experimental set-up. For starters, we decided to move the plants outside of the room with the perfectly controlled humidity and temperature. In addition, we decided to change the photoperiod of plants to 16 hours of light and 8 hours of darkness.
BUILD 2.0
The new set-up consisted of a glass aquarium filled with 5 centimeters of water. The plants were inside with a non-sealed lid that allowed air to flow in and out of the aquarium. This allowed humidity to stay between 70 and 80 %. In the new room, temperatures fluctuated between 25 and 30 °C.
Because we were constrained by the space available in the aquarium and we had more conditions to consider, we decided to switch to 24-well microplates. This time, the wells did not have holes at the bottom to aid retain water in the wells.
TEST 2.0
In order to test our engineered bacteria, several changes needed to be made to the medium. First, we had to add antibiotics, to avoid our strain to lose the plasmid. Secondly, we needed the presence of m-toluic acid, the inducer, to ensure transcription of our genes. Finally, we added creatinine to observe whether the new metabolic pathway had an effect on plant growth. The three components were mixed together and the plants were watered with this mix every three days.
Four conditions were tested. We tested the growth of plants on regolith inoculated with the engineered strain with pSEVA438-Ptet-creA-crnA. Secondly, we used a P. fluorescens strain transformed with the pSEVA438-Ptet empty plasmid, also resistant to the antibiotic. To attest the influence of the antibiotic, inducer and creatinine, the third condition consisted of plants on non-inoculated regolith treated with the mix. Finally, we did the same control of plants in non-inoculated regolith without the mix-treatment. See results here
LEARN 2.0
After 9 days, almost all the plants without bacteria being treated with the mix were dead. In fact, Streptomycin (the antibiotic chosen as a selection marker) was detrimental to plants. In contrast, the plants treated with the mix and inoculated with the antibiotic-resistant bacteria showed significantly higher survival rates than the non-inoculated ones. We concluded that the effect of the bacteria was beneficial, probably because the bacteria was inactivating the antibiotic. In conclusion, any further testing of our engineered bacteria on plants would require chromosomal integration of our genes of interest, thus eliminating the need for the antibiotic.
Conclusion
During this process, we learnt that P. fluorescens did promote plant growth on regolith. Furthermore, we were able to engineer the bacteria so that it could grow using only a urine-compound, creatinine. Finally, we determined that, for our engineered strain to support plant-growth on the Moon, there would need to be chromosomal integration of our target genes.
- Tsuru, D., Oka, I., & Yoshimoto, T. (1976b). Creatinine Decomposing Enzymes inPseudomonas putida. Agricultural And Biological Chemistry, 40(5), 1011‑1018. Click
- Part:BBa_K4757999 Click
- Fredes, I., Moreno, S., Díaz, F. P., & Gutiérrez, R. A. (2018). Nitrate signaling and the control of Arabidopsis growth and development. Current Opinion In Plant Biology, 47, 112‑118. Click
- Nitrification. (s. d.). Melissa Foundation Click
- Melissa Foundation. (s. d.). Click
- Part:BBa_K1067003 Click
- Part:BBa_K1067004 Click
- Part:BBa_K4757999 Click
- Part:BBa_R0040 Click
- Part:BBa_K3332038 Click
- Part:BBa_K5108006 Click
- Part:BBa_K5108007 Click
- Part:BBa_K5108008 Click
- Panmanee, W., & Hassett, D. J. (2009). Differential roles of OxyR-controlled antioxidant enzymes alkyl hydroperoxide reductase (AhpCF) and catalase (KatB) in the protection ofPseudomonas aeruginosaagainst hydrogen peroxide in biofilm vs. planktonic culture. FEMS Microbiology Letters, 295(2), 238‑244. Click
- Rather, M. A., Gupta, K., & Mandal, M. (2021). Microbial biofilm : formation, architecture, antibiotic resistance, and control strategies. Brazilian Journal Of Microbiology, 52(4), 1701‑1718. Click
- Baum, M. M., Kainović, A., O’Keeffe, T., Pandita, R., McDonald, K., Wu, S., & Webster, P. (2009). Characterization of structures in biofilms formed by a Pseudomonas fluorescens isolated from soil. BMC Microbiology, 9(1), 103. Click
- Zheng, W., Zeng, S., Bais, H., LaManna, J. M., Hussey, D. S., Jacobson, D. L., & Jin, Y. (2018). Plant Growth‐Promoting Rhizobacteria (PGPR) Reduce Evaporation and Increase Soil Water Retention. Water Resources Research, 54(5), 3673‑3687. Click
- Rosenzweig, R., Shavit, U., & Furman, A. (2011b). Water Retention Curves of Biofilm‐Affected Soils using Xanthan as an Analogue. Soil Science Society Of America Journal, 76(1), 61‑69 Click
- O’Toole, G. A., & Kolter, R. (1998). Initiation of biofilm formation in Pseudomonas fluorescens WCS365 proceeds via multiple, convergent signalling pathways : a genetic analysis. Molecular Microbiology, 28(3), 449‑461. Click
- Francis, V. I., Stevenson, E. C., & Porter, S. L. (2017b). Two-component systems required for virulence in Pseudomonas aeruginosa. FEMS Microbiology Letters, 364(11). Click
- Lichtenberg, M., Kragh, K. N., Fritz, B., Kirkegaard, J. B., Tolker-Nielsen, T., & Bjarnsholt, T. (2022). Cyclic-di-GMP signaling controls metabolic activity in Pseudomonas aeruginosa. Cell Reports, 41(3), 111515 Click
- Zhang, R., Xu, W., Shao, S., & Wang, Q. (2021). Gene Silencing Through CRISPR Interference in Bacteria : Current Advances and Future Prospects. Frontiers In Microbiology, 12. Click
- O’Neal, L., Baraquet, C., Suo, Z., Dreifus, J. E., Peng, Y., Raivio, T. L., Wozniak, D. J., Harwood, C. S., & Parsek, M. R. (2022). The Wsp system ofPseudomonas aeruginosalinks surface sensing and cell envelope stress. Proceedings Of The National Academy Of Sciences, 119(18). Click
- Tan, S. Z., Reisch, C. R., & Prather, K. L. J. (2018). A Robust CRISPR Interference Gene Repression System in Pseudomonas. Journal Of Bacteriology, 200(7). Click
- Preparation for endoscopy. (2011). Dans Elsevier eBooks (p. 52‑83). Click
- Meinke, D. W. (1998). Arabidopsis thaliana: A Model Plant for Genome Analysis. Science, 282(5389), 662–682. Click
- Anna Lisa-Paul (2022). Plants grown in Apollo lunar regolith present stress-associated transcriptomes that inform prospects for lunar exploration. Nature, 382(2022). Click
- Schoenfeld DA. Sample-size formula for the proportional-hazards regression model. Biometrics 1983;39:499-503. Click
- Pierre Piétracq (2017). Metabolite profiling of non‐sterile rhizosphere soil. The Plant Journal, 92(1): 147–162. Click
- Rasband, W.S., ImageJ, U. S. National Institutes of Health, Bethesda, Maryland, USA Click