Team:Toulouse/Result/experimental-results

We performed several assays to demonstrate the chemotaxis ability of our engineered Bacillus subtilis strain to move towards N-acetylglucosamine (NAG), the base unit of chitin. The ability of the wild-type strain to move towards glucose was the positive control.

1. Petri dishes Test

The first chemotaxis assay was done in Petri dishes filled with a growth medium containing 0,3% agar. This semi-solid medium was supposed to favor bacterial motility. A paper disk soaked with the attractive compound was placed in the middle of the dish, then cells were loaded into the medium (see Figure 1). This protocol was adapted from the Imperial College 2011 iGEM team.

Figure 1: Petri dishes chemotaxis assay. (A) A pipet was used to inject cells into the semi-solid medium. (B) Bacteria would move toward the attractive compound diffusing from a paper disk.

With this assay, we however failed to see any chemotaxis of the wild-type control toward glucose (Figure 2-A). As B. subtilis is attracted by many other glucides and amino-acids that can be found in rich medium such as in LB medium. Therefore, with the hope of improving the experimental conditions, we have challenged the cells with paper discs soaked in LB medium containing glucose (Figure 2-B).

Figure 2: Chemotaxis test with paper discs soaked in either a Glucose solution (A) or in Glucose containing LB medium (B) as attractive compound. Paper discs soaked with water were used as negative controls.

No difference was observed between the Petri dishes with or without glucose. With glucose containing LB medium, a large halo around the paper disk was observed. This halo might be due to cells attracted by the solution, as it was not observed when cells were not inoculated in the Petri dish (data not shown). However this result was barely reproducible. Moreover with the addition of LB medium, it was hard to make the distinction between attractive effects and cell growth resulting from random diffusion. We have therefore given up this protocol and tested alternative protocols.

2. Plug in Pond system

This protocol was adapted from a PhD work (see references [1]). B. subtilis was grown overnight to a density of 8.10⁸ cells/mL. 10 mL of the culture was mixed with 15 mL of LB medium containing 1.5% agar kept at 45°C. The final agar concentration was 0.9%. Tetracycline (25 µg/ml) was added to inhibit growth in order to only observe the chemotaxis phenomenon. Plates were cooled down and dried, before digging wells with either a punch or 1mL tips. The wells were then filled with the attractive compound (Figure 3). After one hour at room temperature, pictures of the plates were taken and the results analyzed.

Figure 3: Plug-in-pond test design.

On our first try with B. subtilis, we made three wells per plate (Figure 4). The wells were filled with glucose at different concentrations and tetracycline was not added in one of the plates.

Figure 4: Plates after 12h at room temperature.

After one hour, no tangible results were obtained. After 12h we observed halos around the 1 M glucose containing wells in the plates without tetracycline but not in the plates with tetracycline. Again, because cells could use glucose for growth, we could not distinguish between growth and chemotaxis. Making the hypothesis that the concentration of tetracycline could have been too high and inhibited any bacterial activity, we thereafter lowered the tetracycline concentration to 15 µg/mL. We repeated this protocol with this new experimental condition. We made two wells per plate (Figure 5), one with either Glucose or n-acetyl-glucosamine and one with LB medium. As previously, no results were achieved after 1h, but after 12h we could notice halos.

Figure 5: Chemotaxis test with WT B. subtilis. The upper well contains attractive compound and the lower well contains medium without any attractive compound.

Results were not as clear as in the previous assay (Figure 4), but halos around the wells with glucose at 250 mM with and without tetracycline were observed. With N-acetyl-glucosamine (NAG) as the attractive compound, halos were observed for a concentration of 25 mM with tetracycline and for a concentration of 250 mM without tetracycline, thus suggesting that our B. subtilis 168 strain is attracted toward NAG and uses it to grow.

References:
[1]: Claudine Baraquet. Etude de la réponse adaptative à l'oxyde de triméthylamine et de son mécanisme de détection chez Escherichia coli et Shewanella oneidensis. 2008, Université de la Méditerranée Aix-Marseille II.

4. Capillary test between two tubes also called the tubes test

After the experiment of the plug-in-pond, we decided to construct a system by welding two Eppendorf tubes with a glass capillary and using an electric burner.

Figure 6: Photography of the first tubes system

We tested this system with a fuchsin dye and water and we were able to observe the diffusion of fuchsin dye towards water. However this construction had a leakage next to the welding seam that we could not stop. Thus, the Toulouse 2014 iGEM team asked the help from the INSA glass blower, Patrick Chekroun. He designed two systems composed of two tubes linked by a capillary.

Figure 7: Scheme of the tubes system

This new system was tested with the fuchsin dye and the assay was made with WT B. subtilis and N-Acetylglucosamine as the attractive compound.

NB: We could not see any diffusion of the fushin dye from one tube to the other. We made the hypothesis that it was not visible by sight because of the small diameter of the capillary.

The following strategy was used to avoid disturbance due to pressure and liquid movement through the capillary:
- The first step was the addition of Wash Buffer until the capillary was full to prevent the presence of any air bubbles which could have impeded diffusion.
- Then, the tube 2 was plugged and maintained with the thumb while another iGEM mate was adding the bacterial suspention of WT B. subtilis into the tube 1.
- The tube 1 was also plugged and only then the thumb could be removed from the tube 2.
- In the same way, the N-Acetylglucosamine was added in the tube 2.
- The same process was made with xylose as a positive control.

NB: According to the paper "Chemotaxis towards sugars by B. subtilis", (Ordal et al., 1979), glucose and xylose have the same attractant power. We have privileged a positive control instead of a negative one as we were not sure that this system was efficient.

- The system was kept straight for 2 hours. Every 40 minutes, samples from each were removed and streaked on solid medium (dilution 1/1,000) in order to estimate the bacterial concentration.

Figure 8: Photography of the tubes system

Unfortunately, the dilution was too high to detect any chemotaxis movement. As we did not find any information in the litterature and did not have enough time to optimize this protocol, we decided to test again and possibly ameliorate the first protocol from the Imperial college 2011 iGEM team : the tips capillary test.

5. Tips capillary system

First tips capillary system

This protocol was adapted from The Imperial College 2011 iGEM team (See chemotaxis protocol).
In the first tips capillary system, we used parafilm to avoid any kind of air disturbance in the tips. The different steps are described below:
- 15 µL of each chemo-attractant was pipetted.
- The bottom of tip with the pipette was then put on a piece of parafilm and the pipette was removed from the top of the tips.
- The tip was then sealed with a piece of parafilm in order to keep the liquid sterile and inside the tip.
- To finish, the level of the solution in the tip was marked.

Figure 9: Sealing of a tip with parafilm

- When all the chemo-attractants were added, the tips were fixed on a green support. The whole process can be seen on Figure 10.
- Each tip was then immersed into 300 µL of a bacterial suspension in the wells of an Elisa plate.

Figure 10: First tips capillary system

NB: the yellow carton was used to stabilize the system and kept it straight.

- After one hour, the tips were removed from the bacteria suspensions and the bacteria content within the tips was monitored with a Thoma cell under the microscope.

We experienced several problems with this system:
- The liquid level decreasing so badly during the course of the experiment that we did not have enough liquid to fill the Thoma cell for counting.
- The bacteria were moving, therefore preventing us from accurately counting them (but that's the point of swimming, no??? ;-)). Taking into account this problem, we decided to estimate the bacterial concentration by streaking the tips content on agar plate instead of using Thoma cell and microscopy.

Second tips capillary system

And then the revolution came! We found a multichannel pipette :D The same protocol was performed except that the parafilm was used to avoid the air entrance between the tips and the pipette and therefore the loss of liquid.

Figure 11: Second tips capillary system

Improvement of the second tips capillary system

However this system was not optimal it is why we decided to use blu tack instead of parafilm:

Figure 12: Improvement of the second tips capillary system

At that point, the protocol was approved and the final test could finally start! :-)

There was just one tiny problem… we did not have our optimized bacterium transformed with the chemotaxis module!!! That is why we concentrated our efforts on WT B. subtilis strain.

The main goal was to find an optimized control and to analyze the eventual chemotaxis of the WT strain. To avoid osmolality bias, we wanted to find a molecule which was non-attractant and with a similar molecular weight than that of the N-Acetylglucosamine (221.21 g/mol). Our first idea was to use fuchsin (Molecular weight: 337.85 g/mol).

At the beginning, the experiment was conducted with only one negative control, the fuchsin and different NAG concentrations: 25 mM, 250 mM and 500 mM. The tested strain was B. subtilis 168:

Figure 13: Fuchsin - negative control (dilution 1/50)	Figure 14: NAG (25mM) (dilution 1/50)

The average number of colonies with the negative control is 121. In contrast, a cell layer was observed for the NAG plates with every concentration, indicating that a large number of cells have been inoculated in the plates.

Thus, we assumed that WT B. subtilis was more attracted by NAG than by fuchsin. In addition we could neglect the bacterial growth because the test only lasted one hour. We also neglected diffusion and osmolality phenomena for the reasons explained above.

Unfortunately for us we forgot one major effect… Can you believe that fuchsin solution contains about 15% of ethanol?!!! This concentration can lead to the death of some cells and thus make the fuschin dye an appropriate negative control.

This incredible and dramatic discovery destroyed all of our hopes about the God of chemotaxis! :-(

However, our team did not give up on synthetic biology! :-) Indeed, after days of disappointment and no time left for lab work, we raised from ashes and tried to find another negative control.

Hopefully, we managed to find a negative control: galactose (25 mM). The article "Chemotaxis towards sugars by B. subtilis" (Ordal et al., 1979) proved that it was a poor attractant.

We made our tests again with this new molecule and glucose (25 mM) as positive control.

Figure 15: Final results (dilution : 1/10,000)

The miracle arrived! We managed to prove that our WT B. subtilis was indeed naturally attracted to NAG.

NB: It was our last experiment. Unfortunately we were running out of time and we could not do much more test. We would like to do the experiment with a lower dilution and repeat it several times.

Our results are not statistically significant however this result has been described in the litterature.

Collapse

1. Preliminary experiments

For the first experiment we wanted to check if the buffer of the binding assay was compatible with B. subtilis survival. To do that, we tested four bacterial concentrations (from OD 0.1 to OD 0.01). These B. subtilis suspensions were incubated 1 hour at 4°C with 500 µL of either Chitin Binding Buffer (CBB) or water. 100 µL of cell suspension were plated on LB medium in order to count surviving cells.
Cells do not seem affected by the presence of CBB or water with estimated ODs of 0.05 or 0.025.
We observed similar survival rates between cells treated with CBB or water (these data are not shown on figure 16). Thus, the experimental conditions of the chitin binding assay are compatible with the bacterial life.

Figure 16: We don't want to bother you with useless data

2. Binding test using engineered B. subtilis

Purpose

B. subtilis transformed with the binding module should produce a chimeric protein allowing the bacterium to bind chitin. It is composed of the Cell Wall Binding Domain of a B. subtilis protein LycT, and of the GbpA domain 4 of Vibrio cholerae able to bind chitin. The capacity to bind chitin was assessed by bringing together either the WT strain or the engineered bacterium with chitin beads. After several washes, bacteria still bound to the beads were counted.

Results

We measured the quantity of bacteria before the binding assay (Direct), in the eluted fraction of the first (Wash A) and the second (Wash B) washing steps, and associated with the beads.
Both bacterial solutions of WT and engineered bacterium used for the binding assay had the initial same concentration before the assay (Direct on Figure 17). There is also no significant difference between both strains after the first wash (Wash A on Figure 17). However, the concentration of cells quantified in the eluted fraction after the second wash was significantly higher for the wild-type strain. This suggests that the engineered strain is more retained on chitin beads. This was confirmed by a 40 times higher number of cells associated with the chitin beads for the engineered over the wild-type strain (Beads on Figure 17).
Thus, we successfully engineered B. subtilis to promote its binding onto chitin.

Figure 17: Attachment of WT B. subtilis and engineered bacterium to chitin. The WT bacteria or the bacteria with the binding system concentrations have been determined during the different steps of the binding test. The stars represent a significant difference observed with a Student test with p<0.05.

3. Microscopic observations

Purpose

We wanted to observe SubtiTree bound on the chitin coated beads. In order to perform a 3D reconstruction showing this interaction, we used confocal laser scanning microscope. Through the use of a green fluorochrome (Syto9), we highlighted the presence of bacteria on the surface of the beads (individualized by phase-contrast). A first calibration step determined the minimum threshold to remove the background noise and the natural fluorescence.

Results

We can notice the engineered bacterium is well attached to the surface of beads coated with chitin.

Figure 18: Microscopic view of engineered strain associated with beads surfaces coated with chitin

Using ImageJ software, we created 3D pictures and movies of those images.

Figure 19: A short movie of 3D bead surfaces coated with chitin and the engineered strain (emotional sequence for Subtitree: first movie apparition, before Cannes festival…)

We then performed a wash step on the chitin beads. We measured the release of bacteria on the washing solution. When our bacterium has the binding module (Right on Figure 20), there is less release than without the module (Right on Figure 20). Therefore, the engineered bacterium is retained by the beads.

Figure 20: Microscopic view of elution fraction of WT B. subtilis (Left) and engineered bacterium (Right).

To conclude, all the results are consistent with the successfull integration of a functional chitin binding system in B. subtilis. We thus validated the second module.

Collapse

1. Preliminary experiments

Tests with commercial peptides and controls

The first tests were accomplished with commercial GAFP-1 and D4E1 peptides at different concentrations (12.5 µM, 25 µM, 100 µM). As Ceratocystis Platani is pathogenic, we could not perform tests directly with this fungus. These tests were therefore performed with different non-pathogenic fungal strains from the same phylum as Ceratocystis Platani.
After 1 to 6 days at 30°C depending on the fungal strain, the PDA (Potato Dextrose Agar) plates covered with fungus and containing a paper disk soaked with a commercial peptides solution were analyzed.

An inhibition halo was noticeable with commercial D4E1 peptide at 100 µM on Aspergillus brasiliensis (Figure 21). Less bright halos were also present with lower concentrations. Concerning commercial GAFP-1, we did not notice any effect in the tested conditions. Copper sulfate, a well-known chemical fungicide was used as a positive control. Inhibition of the fungal growth was complete with 20 mg/ml copper sulfate, and at 10 mg/ml a darker halo appeared around the pad as can be seen on Figure 21. This corresponds to a sporulating halo in response to the stress generated by the fungicide.

Figure 21: Results of the preliminary tests of antifungal compounds

Regarding these results, we concluded that very high fungicide concentrations are required to inhibit the fungal growth in the tested conditions. Following these tests, new conditions were adopted in order to avoid too much fungal growth over bacterial growth. The culture medium was adjusted to fit our objective and to approximate the conditions found in the trees, and a 'sap-like' medium was elaborated (See protocols for more informations). Incubations were performed at room temperature. These new conditions were used as standard for the the next experiment.
Camille also concluded that turning the water of the Canal du Midi to deep blue using high copper sulfate concentrations is not such a good idea... Thereby strengthening our faith in SubtiTree :-) !

2. Test with antifungal bacteria

In order to test our B. subtilis engineered strains, it was essential to find the right balance between the fungal growth and the bacterial one which was achieved with the previous modifications. Furthermore, in our genetic constructions, the antifungal peptides were designed to be exported in the extracellular medium.

The transformed B. subtilis strains were grown at 37°C during 72h before testings. Inhibitory effect of supernatant and cell pellets were tested separately. After centrifugation, the supernatant and the resuspended pellet were placed on pads and disposed on plates previously seeded with a defined number of conidia (see protocols to have more details). After several days at room temperature, an inhibition halo of Trichoderma reesei growth was clearly observable for the strain expressing the D4E1 peptide. The inhibition was even more noticeable with the strain carrying the GAFP-1 + D4E1 operon (Figure 22).
However, no effect was detected for the strain expressing the GAFP-1 gene, we thus suppose a putative synergistic effect between these two peptides.
Regarding EcAMP-1, no effect has been detected on the fungal growth. Several factors can explain these results: a number of post-transcriptional modifications are required to have a functional EcAMP-1 and in addition, the sequencing results for these constructs showed several discrepancies with the original designed sequence.

Inhibition halos are not visible with supernatants, probably because of either low concentrations or of instability in the extracellular medium. Another effect was noted with the same strains expressing D4E1 and GAFP-1 + D4E1 on another fungus, Aspergillus brasiliensis (figure 22). This effect is comparable to the one previously noted with a low concentration of copper sulfate (figure 21).

Figure 22: Results with transformed bacteria.

After this set of experiments, the strains expressing D4E1 or GAFP-1 + D4E1 have been shown to be the best candidates to play a major role in the fight against fungal diseases such as Canker stain. Keeping in mind our objective, we decided to tests these strains in model plants: Nicotiana benthamiana and Arabidopsis thaliana.
These tests were performed in association with Sylvain Raffaële and Marielle Barascud in the National Institute for the Agronomic Research.

3. In planta tests with SubtiTree

Figure 23: Injection of antifungal B. subtilis in a model plant

The goal of the project is to introduce the transformed bacteria in a diseased tree. So it is necessary to perform preliminary in planta tests to judge the fungus-killing abilities of the two strains selected after the previous sets of experiments.
The strain expressing fungicides was first inoculated in two model plants (Arabidopsis thaliana and Nicotiana benthamiana). After this step, a phytopathogenic fungus (Sclerotinia sclerotiorum) was placed on the leaves.

Twenty-four hours after SubtiTree inoculation, no phenotypic modification of the leaves could be detected. We can conclude that our bacterium, its introduction and the fungicides production in plants do not have deleterious effects.
Without proper treatment, the drop of the phytopathogenic fungus on Nicotiana benthamiana leaves caused a necrosis halo which could be measured after 40 h. The number of necrotic sites and the lesion sizes appeared reduced by B. subtilis expressing DE41 or GAFP1-D4E1, unlike the WT bacterium. Two independant replicate of this experiments were performed successfully


We did not observe any significant results for Arabidopsis thaliana because of the use of two plants batches with different ages.
We can therefore conclude that when SubtiTree is in plant's physiological conditions, it is harmless to the plant, and that the production of fungicides is effective, reducing the leaves necrosis.

Figure 24: Results of in planta test

We now expect that bringing altogether the three modules (chemotaxis, binding and antifungal) should even improve the performance of SubtiTree. Thus, these results pave the way towards the testing of SubtiTree in infected plane trees
More than ever, let's save our trees with SubtiTree!

Collapse