ENGINEERING

In order to adapt the ilux operon for use in the iGEM backbones, standard iGEM prefixes and suffixes had to be added; the EcoRI (E) and XbaI (X) restriction sites must be positioned upstream, and SpeI (S) and PstI (P) downstream of the operon. We therefore had to remove the EcoRI, XbaI, SpeI, PstI restriction sites from the ilux coding sequence, so the operon would not be destroyed when using these enzymes. The ilux operon contains 3 PstI restriction sites, 1 EcoRI site and 1 XbaI site that needed to be removed.

First, the 5 sites were removed by directed mutagenesis PCR.

The DNA construction was digested and analyzed by agarose gel electrophoresis. It was unsuccessful: the restriction sites were indeed mutated and the correct fragments were obtained (Figure 2), but due to technical limits, we were not able to reconstruct the whole operon at once.

As we were not able to reconstruct the operon, another method to remove the restriction sites needed to be tested.

As the problematic restriction sites were not present in the genes iluxA, iluxB and iluxE, we performed a classical PCR to reconstruct fiatluxABE. We decided to have mutated fiatluxCD synthesized in a pACYDuet-1 vector containing fiatluxCD and a constitutive and weak BBa_J23117 promoter (without restriction sites).

The amplified sequence of fiatluxABE was inserted in pSB1C3 (already in iGEM biobrick format) and pBAD18 (high-copy vector with an arabinose inducible promoter). The synthesis of fiatluxCD was ordered.

The DNA constructions were digested and analyzed by agarose gel electrophoresis. Restrictions profiles were done: the results were in agreement with the expected sequence.

The restriction map shows that fiatluxABE was well reconstructed. The synthetized plasmid containing fiatluxCD is correct. In the end, we obtain fiatluxABE in pBAD18 and pSB1C3, and fiatluxCD in pACYDuet-1. The corresponding parts are BBa_K4239007 and BBa_K4239006.

Silent mutations effect on luminescence production had to be tested. If light is still emitted, it would mean that the operon is still functional despite our mutations.

In order to test the effect of the previously induced silent mutations, E.coli DH5 alpha were transformed with pACYDuet-1-fiatluxCD and pBAD18-fiatluxABE (that are compatible).

Light emission was observed with ChemiDoc XRS+, Bio-Rad (a luminescence detection apparatus).

After observation with ChemiDoc XRS+, Bio-Rad, it appears that we can still observe bioluminescence after having introduced silent mutations (Figure 3).

Having both fiatluxABE and fiatluxCD in the same conjugative vector is essential if we want to be able to transfer it to other bacterial strains in future experiments. We had to assemble both parts in the same conjugative plasmid: pSEVA521.

fiatluxCD and fiatluxABE were then assembled together in the conjugative vector pSEVA521 by HiFi assembly.

pSEVA521-fiatluxCDABE was amplified, extracted and verified with a restriction profile.

The restriction profile was the one expected: the fiatlux operon is now complete and well inserted in a conjugative plasmid: pSEVA521.

fiatlux was inserted in two conjugative vectors, pSEVA521 and pSEVA531. They both have different origins of replications: pSEVA531 is thus present in a higher number of copies. To study the stability of the plasmids in the long term, it is necessary to determine the optimal storage temperature and tetracycline concentration of E.coli transformed with fiatlux.

E.coli bacteria were transformed by pSEVA521-fiatlux and pSEVA531-fiatlux separately and grown in a medium supplemented with 10 µg/mL of antibiotic. Several temperatures were tested for the storage conditions.

E.coli transformed strains were stored at different temperatures (-80°C, 4°C, room temperature) and then revivified on a medium containing an antibiotic concentrations of 10 µg/m, to see how they recover from storage.

E.coli bacteria did not grow again after freezing at -80°C.

We chose to test different antibiotic concentrations to check if it had an impact on this issue.

Different antibiotics concentrations were tested, after storage at different temperatures.

E.coli transformed strains were stored at different temperatures (-80°C, 4°C, room temperature) and then revivified on media containing several antibiotic concentrations (2, 5 and 10 µg/mL).

E. coli strains transformed by pSEVA531-fiatlux showed slower growth than those transformed by pSEVA521-fiatlux, which means a too high strength of transcription may exhaust the bacteria. Plus, E.coli transformed with fiatlux showed better growth in a 5µg/mL tetracycline medium, and were able to grow again after any storage temperature at 5µg/mL, but not at 10µg/mL.

After successfully transforming E.coli DH5α bacteria with our plasmids and characterizing the tool, the objective was to insert it in D.solani. The four plasmids used in the experiment are pSEVA521, pSEVA521-fiatlux, pSEVA531 and pSEVA531-fiatlux. The two empty vectors (pSEVA521 and pSEVA531) are used as a control to compare the effects of the fiatlux operon in D.solani.

We first tried to insert the plasmids by electroporation.

Dickeya solani colonies did not grow on a selective medium.

D.solani does not respond to electroporation.

We used the conjugation method to transfer the plasmids into D.solani.

An intermediate donor bacteria E.coli MFDpir was transformed, and then the conjugation was performed.

After 48h of incubation, some colonies grew on the selective media, and luminescence was observed with a high sensitivity camera.

The four plasmids were successfully inserted in D.solani, and the colonies were luminescent when fiatlux was present (Figure 4). The colonies were re-isolated on Petri dishes and the most luminescent colonies were selected to be used for the rest of the experiments.

To compare the growth between bacteria with or without plasmid, and to compare the effect of the different plasmids, we needed to model each situation. To do so, we used the software R and several mathematical models to determine the one that best fitted the experimental growth curves, get the parameters for each situation, and compare them thanks to a statistical analysis. The purpose of the analysis was to detect if there were significant differences between the different growth conditions and the different backbones tested.

We modeled the curves to find the appropriate mathematical model thanks to the nlstools package on R. Thanks to this package, we could then visualize them and the model adjusted to them, but also get the parameters of this model. We modeled both the O.D. 600 nm as a function of time and the Luminescence/OD ratio as a function of time. Once we did it for all of our conditions and replicates, we compared them using the emmeans library.

After extracting data from our experiments, we compared the AIC criteria of our different models for each condition (antibiotic concentration and the nature of the plasmid). We checked that the best model really fitted the curve properly and returned consistent parameters values. We statistically compared the parameters of each model. It helped us determine which condition (or backbone) was significantly different from the other.

The model that best fitted our curves turned out to be the Baranyi model. We observed that pSEVA531-fiatlux produced more luminescence than pSEVA521-fiatlux in Dickeya solani. Moreover, the fiatlux operon didn’t affect bacterial growth in the pSEVA531 backbone.

The first step was to extract the proteins contained in our bioluminescent bacteria without denaturing them.

We chose to extract our luciferase using E.coli MFDpir x pSEVA531-fiatlux. We chose E.coli MFDpir transformed with pSEVA531-fiatlux and not with pSEVA521-fiatlux (lower copy number plasmid than pSEVA531-fiatlux) because we wanted the maximum possible concentration of our luciferase. We chose to use the “French Press” method for the lysis of our bacteria because it disrupts cell walls without altering cytoplasmic proteins. The solution obtained is centrifuged and the proteins remain in the supernatant.

We verified the presence of proteins in the supernatant by the measurement of the absorbance at 280 nm. The absorbance at 280 nm is typical of the aromatic rings of the amino acids composing proteins.

The French Press method works.

We wanted to extract our luciferase, in order to purify it, from the solution obtained after the French Press. To do so, we had to separate our luciferase from the other proteins. As our protein is not tagged, the easiest way was to perform a gel exclusion chromatography.

Our enzymatic complex weighs 80kDa. We used a Sephadex G-100 gel because it will exclude all proteins above 100kDa and separate proteins under that size. The size of our column is 20cm tall to have a good separation without using too much Sephadex gel. We added 1mL of our solution (whose O.D. at 280 nm was 3.301 originally) to be purified. We measured the O.D. at 280 nm of each collected tube to draw a chromatogram.

The chromatogram indicated that all our proteins were eluted in the first collected tube, which means we overloaded our column.

We need to add a smaller volume of our solution to be purified or we need to dilute our solution.

We decided to add a smaller volume of our solution in a 20cm tall Sephadex G-100 column.

We added 350µL of our solution to be purified, instead.

We obtained a satisfactory chromatogram with 6 different tubes of 2.5mL containing proteins.

350µL of a solution with an O.D. at 280 nm of 3.301 is sufficient to not overload a 20cm tall Sephadex G-100 column.

To know in which one of the 6 tubes the luciferase is present, we chose to do a polyacrylamide gel electrophoresis (PAGE).

We did a PAGE non-denaturing with a sample load of 10µL.

No bands other than the protein ladder (molecular-weight size marker) appeared on the PAGE.

Our collected tubes didn’t contain enough proteins. Indeed, the limit of detection of our gel is 50µg. Hence, we assumed that our collected tubes had a concentration of protein smaller than 5µg/µL. We needed to quantify the protein concentration of each tube to know if our hypothesis was true.