Results

The experimental results described here provide insights of the functionality of our project and clues about how our synthetic consortium will work.

1. Mimicking Vibrio sp. presence with an engineered E. coli

Background

For safety reasons, we cannot manipulate V. cholerae. Therefore to mimic its presence, we engineered E. coli to produce a quorum sensing molecule from a non-pathogenic Vibrio species i.e. C8-CAI-1 from V. harveyi ( cloning ). C8-CAI-1 is an analogue of V. cholerae CAI-1 quorum sensing molecule and is produced by the enzyme CqsA. Besides, it allows to investigate our capacity to create synthetic communication between both engineered E. coli and V. harveyi.

Materials and methods

Strain construction can be found here ( Protein production and sampling, NMR analysis , Cultivation condition and Medium composition )

Results and discussion

The figure 1 presents the results obtained with the supernatants of wild type E. coli (as a negative control with no C8-CAI-1 produced), wild type V. harveyi (as a positive control producing C8-CAI-1), and our engineered E. coli strain expressing cqsA from V. harveyi (i.e. VhCqsA, Part:BBa_K2278001 ). These supernatants were analysed by ¹H Nuclear Magnetic Resonance (NMR). Characteristic and observable proton signals of C8-CAI-1 in ¹H NMR should appears at approximatively (δ, in ppm) 4.14 as a triplet dedoubled (CH in α of both carbonyl and hydroxyl groups), 2.50-2.36 (CH2 in α of carbonyl group) and 1.92-1.84 (1H of CH2 in β of hydroxyl group) as multiplets. Unfortunately, the NMR profiles of the three supernatants were similar and no characteristic ¹H NMR signal of C8-CAI-1 had been detected neither in wild type V. harveyi nor in E. coli-VhCqsA strains. So, NMR approaches failed to confirm the production of C8-CAI-1 molecule in both E. coli-VhCqsA and in V. harveyi WT strains. This does not mean the tested construction is not functional but more probably that C8-CAI-1 production was below the sensitivity limit of NMR. As a more sensitive alternative, mass spectrometry analyses were also applied on the same samples. However, no conclusive results were obtained (data not shown) and to the optimization of the method required extra time. We thus cannot exclude that the expression of CqsA is not functional but a more sensitive approach such as in vivo strategy should be tested to check the functionality of this part.

Perspectives

Culture parameters, induction by IPTG or extraction protocol could be optimized to increase C8-CAI-1 production. Alternatively, MS method could be improved to detect the C8-CAI-1. Alternatively, we can focus on the expression of the cqsA gene from V. cholerae to create a safe V. cholerae effector.

2. E. coli producing C8-CAI-1 molecules can be sensed by V. harveyi

Background

In response to its quorum sensing molecule, i.e. C8-CAI-1, V. harveyi becomes bioluminescent. Here we used it as a sensor to detect the production of C8-CAI-1 by our engineered E. coli-VhCqsA strain (i.e., Part:BBa_K2278001 ) .

Materials and methods

Protocols can be found here: Medium composition , Solid Bioluminescence assay

Results and discussion

We used the JMH626 mutant strain of V. harveyi, a strain with deletion of all its quorum sensing production genes, including cqsA¹. Thus, it is unable to produce its own C8-CAI-1. However, the whole apparatus for response to quorum sensing is still present in this strain. So, the supernatant of the E. coli-VhCqsA strain was applied on this strain to check if it was able to trigger luminescence, as a natural property of V. harveyi exposed to C8-CAI-1. Results are presented in Figure 2. In the negative control with V. harveyi ∆cqsA supernatant applied to the same strain, a very low basal bioluminescent is observed (likely from promoter leaking). The same very low basal luminescence level was observed with supernatant of wild type E. coli strain. As expected, when the supernatant of V. harveyi wild type strain was applied to V. harveyi ∆cqsA, bioluminescence could be restored. When supernatant of E. coli-VhCqsA strain was applied to the V. harveyi ∆cqsA strain, bioluminescence was observed with a comparable level to the one observed with wild type V. harveyi. Together, these data demonstrate that C8-CAI-1 is efficiently produced by E. coli-VhCqsA and thus validate our BBa_K2278001 part. More importantly, this result also demonstrates that we successfully created synthetic communication between engineered E. coli and V. harveyi strains.

Perspectives

Next step will be to proof the communication between E coli producing the CAI-1 from V. cholerae ( Part:BBa_K2278002 ) and V. harveyi expressing the gene encoding the modified sensor CqsS* that could sense both C8-CAI-1 and CAI-1.

3. Modification of V. harveyi to detect both C8-CAI-1 and CAI-1

Background

V. harveyi has all the assets to be a good sensor for V. cholerae: it possesses its own pathway of detection for C8-CAI-1, very similar to the V. cholerae pathway to detect CAI-1. It has been shown that only a single point mutation on its receptor CqsS allows V. harveyi detecting V. cholerae CAI-1 molecule¹. Because genetic in V. harveyi is limited (i.e. classical transformation does not work), we first had to set up a triparental conjugation protocol.

Materials and methods

Strain construction can be found ( here ) and protocols used can be found here: Triparental conjugation , Fluorescence microscopy , Cultivation condition , Medium composition ).

Results and discussion

A protocol of triparental conjugation to transfer plasmids to V. harveyi was set up from an original conjugation method provided by M. Arlat (LIPM, Toulouse) usually used with Xanthomonas campestris. It was adapted to our V harveyi strain. It requires an E. coli donor with the conjugative plasmid pBBR1MCS-5 + insert (BBa_J04450), an E. coli helper with the helper plasmid pRK2073, and a receiver V. harveyi. Transformants were analysed by fluorescence microscopy and results are presented in Figure 3. The pictures show that the strain of V. harveyi conjugated with the plasmid pBBR1MCS-5 containing the BBa_J04450 construction is able to express RFP using a promoter and terminator from E. coli. This raised several important conclusions. (i) It clearly demonstrated that V. harveyi can be engineered. This is the first use of V. harveyi as a genetic chassis in the IGEM competition. (ii) It proved that iGEM registry elements could be used in V. harveyi (promoter and terminator for instance). (iii) To the best of our knowledge, this is the first reported use of RFP in V. harveyi.

Perspectives

Next step will be to integrate the plasmid containing the engineered CqsS* receptor into V. harveyi to create a sensor able to detect both V. cholerae CAI-1 and V. harveyi C8-CAI-1. Likewise, the Als construction tested in E. coli could be now tested in V. harveyi (see next module).

4. Production of diacetyl to establish communication between prokaryotic and eukaryotic cells

Background

The quorum sensing signal detected by V. harveyi has to be transmitted to the yeast P. pastoris. For that, we chose to engineer V. harveyi so that it will conditionally produce diacetyl as the communication molecule between the bacterium and the yeast. Here we tested the functionality of our construction containing the als gene, encoding the acetolactate synthase that enable diacetyl production ( Part:BBa_K2278011 ).

Materials and methods

Strain construction can be found here ( cloning ) and protocols used can be found here (Link to the following protocol: Cultivation condition , NMR analysis and Protein production and sampling.

Results and discussion

To check the functionality of the plasmid expressing als, we first used E. coli. Figure 4 presents the results obtained for the ¹H NMR analysis of the supernatant of E. coli strains with empty pSB1C3 or pSB1C3-als. Diacetyl specific shift was confirmed by spiking experiment. A small peak of diacetyl could be detected in the strain expressing als but not in the control. However, the amount seems very low. This could be related to two punctual mutations observed in the als sequence of this clone. However, only a single sequencing run was performed for this part, which is insufficient to be affirmative about the reality of these mutations.

Perspectives

Next step will be to verify the sequence of als and to express the plasmid encoding for the right sequence of Als in V. harveyi. Diacetyl production will be then checked again by NMR.

5. Diacetyl detection by Pichia pastoris:

Background

To allow a communication between V. harveyi and P. pastoris mediated by diacetyl, the Odr-10 receptor (design) had to be expressed in P. pastoris. To check if Odr-10 was functional in vivo, we used a pFUS1-RFP reporter system. Indeed when diacetyl binds to Odr-10 a cascade of activation of Ste proteins (endogenous to P. pastoris) will lead to the binding of Ste12 on pFUS1 promoter, and the expression of RFP should be activated.

Materials and methods

Strain construction can be found here ( cloning ) and protocols used can be found here (Link to the following protocol: Cultivation condition , Medium composition and Plate reader).

Results and discussion

As a control, we firstly checked the activity of the pGAP promoter (i.e. used for both overexpressing Odr-10 and testing the production of AMP (see next section) present in the yeast vector pPICZα we used. We tested the functionality of the complete system Odr-10/PFUS by growing the cells on medium specifically design to induce the activation of Ste proteins (Johnston et al., 1977) and that contain diacetyl. Absorbance and fluorescence production by P. pastoris strain having integrated the empty plasmid or the plasmid containing Odr-10/pFUS1-RFP system was followed over the time in a microplate reader. Results are presented in Figure 5. In the conditions w/o diacetyl or with 500µM of diacetyl no difference was observed between the control and the strain expressing Odr-10/pFUS1-RFP. However, when diacetyl is present at higher concentration (i.e. 1000µM), differences are observed between both strain. The RPF fluorescence is higher in the P. pastoris strain having integrated the plasmid containing Odr-10/pFUS1-RFP system than the one having integrated the empty plasmid. These data demonstrate the functionality of the complete detection pathway and proof the concept that our engineered P. pastoris is ready to communicate with Vibrio species if this former one produce diacetyl.

Perspectives

Coding sequence of RFP should be optimized to improve the fluorescence ratio. For pGAP, same experiment should be done in rich medium, i.e. the medium used in our device. Next step will be to growth together our V. harveyi expressing Als and our P. pastoris expressing Odr-10/pFUS-RFP to demonstrate the synthetic communication between both.

6. P. pastoris is able to produce functional antimicrobial peptides

Background

The final goal of our synthetic consortia was to kill V. cholerae with newly described antimicrobial peptides from crocodile. Here we tested the production by P. pastoris of three selected antimicrobial peptides (AMP), i.e. D-NY15, cOT2 and Leucrocin I (respectively Part:BBa_K2278021, Part:BBa_K2278023 and Part:BBa_K2278022 )

Materials and methods

Strain construction can be found here ( cloning ) and protocols used can be found here (Link to the following protocol: Cultivation condition , Medium composition , Plate reader, Semi quantitative RT-PCR and On plate toxicity assay).

Results and discussion

• Expression assay

Because our AMPs have never been produced by P. pastoris and their toxicity on this yeast are not assessed, we first validated the expression of the AMPs by RT-qPCR. Figure 6 showed the results obtained for D-NY15 expression in P. pastoris. The amount of fluorescence provided by the RT-qPCR with the D-NY15 amorces (curves A1, A2 and A3) is raising after few cycles (8.32 +/- 0.03) whereas the negative control (pPICZα only, curves C1, C2 and C3) starts to be amplified at over 29 cycles (i.e. non-specific amplification). This means that the D-NY15 encoding gene is expressed in P. pastoris and non-lethal.

• Toxicity assay

To test the toxicity of the three AMPs, disks soaked with the supernatant of P. Pastoris with genomic integration of pPICZα, pPICZα-D-NY15, pPICZα-cOT2 and pPICZα-Leucrocin I were put on freshly inoculated solid culture of V. harveyi. Results are presented on Figure 7. Inhibition halos were observed around both the positive control and the patch containing the supernatant from a D-NY15 expressing strain. However no growth inhibition were observed around the disks containing the supernatant from a COT2 and Leucrocin I expressing strains. These data nicely demonstrated the capacity of the engineered crocodile peptide D-NY15 to inhibit V. harveyi growth but also the capacity of P. pastoris to produce antimicrobial peptides. This is the first demonstration of AMP production by P. pastoris by an IGEM Team and the first use of sequence coming from crocodile in iGEM!

Perspectives

Next step will be to prove the inhibition of AMPs in a liquid medium containing both P. pastoris expressing AMP together with V. harveyi and V. cholerae. For cOT2 and Leucrocin I for which cytoxicity cannot be yet demonstrated, a higher concentrations should be tested. However, for this, AMPs should be first extracted from the supernatant.

7. Co-cultivate P. pastoris and V. harveyi is possible:

Background

Ultimately our synthetic consortia will be freeze-dried and placed into a plastic containing dried cultivation medium. As each microorganism of the consortia has its own cultivation medium, the first difficulty is to find a common media on which V. harveyi and P. pastoris can grow together. The second one is to test the capacity of those strain to recover growth on the selected medium after freeze drying.

Materials and methods

Protocols used can be found here (: Cultivation condition , Medium composition ).

Results and discussion

• Selection of the cultivation medium

404 text not found P. pastoris and V. harveyi were grown on seven different medium (i.e. LB, LB acetate 2%, LB NaCl, LM, LM acetate 2%, YPD, YPA 1%, and YPA 2%) corresponding to the standard media used to grow them (i.e YPD for P. pastoris or LM for V. harveyi) or media arranged to fit both (i.e LB + salt). Results are presented in table X.

ADD TABLE 1 As expected both species are able to grow on their favourite medium, however on X and Y none of them can growth while on Z and FF, both can growth. As V. harveyi is a marine bacterium, it cannot grow without salt therefore it is not surprising that the best suited medium is LB+NaCl On this medium both microorganism can grow at fast rate (1 hour for V. harveyi and less than 5 hours for P. pastoris) . This is a really interesting result for this project since LB medium is cheap and easy to produce. All these data have been used to fit our model see “Model - Simulation” page https://2017.igem.org/Team:INSA-UPS_France/Model/Simulation) in order to help in the design of the device (Link to the device).

• Freeze-dried microorganism

Next, we checked the growth recovery on LB+NaCl of P. pastoris wild type or expressing D-NY15 and V. harveyi wild type. Figure X shows the growth of the microorganisms on LB+NaCl after they were freeze dried during X hours.

Whatever the strain and microorganism, all can recover growth on LB+NaCl after having being freeze-dried. For V. harveyi, no clear result can be draw from this experiment because no antibiotics were used during the rehydration of the bacteria next to freeze-dried, and so, contaminant can’t be excluded.

Perspectives

Next step will be to test our freeze dried synthetic consortia in our device (Link to device section) and place it into water to check to growth recovery in real condition.

8. Permeability assay of the membranes used for the devices:

Background

We selected two types of membrane materials, i. e. TPX® X44B 50 PVDF 0.1 µm for making the sachet containing our synthetic biology system composed of V. harveyi and P. pastoris with nutrient medium. Here we tested the permeability of both to check whether this membrane could to contain our synthetic organisms while allowing the antimicrobial peptides and quorum sensing molecules go through from the device compartment to the treated water.

Results and discussion

We first used the TPX® X44B 50µm material used by iGEM Groeningen 2013 and iGEM Toulouse 2014. To assess the porosity of the material, we performed a liquid/liquid transfer assay with bengale pink dye. Bengale pink dye was chosen because of its high molecular mass for a dye (977 g/mol) to get closer to the size of D-NY15 (1767 g/mol). The test was done with a Hele-Shaw measurement cell (two parallel planes spaced by 2 mm) with a camera (EDMUND OPTIC) to monitor the diffusion through the membrane (figure X). The cell was filled with water and bengale pink was introduced from the top in a cone of membrane (or absorbent paper as a control), in contact with the cell water.

We did not observe diffusion of the dye through the TPX membrane in this experiment (Data not shown). Since we were concerned with the hydrophobicity of the material, we also tried to check if water can go through by putting the material with a drop of water on top of absorbent paper. No moisture on the absorbent paper was observed during the course of the experiment. We conclude that the chemical characteristics of the TPX material does not suit to our project. Next we tested the material proposed by our partner Sunwaterlife, i.e. a membrane used in their water cleaning system (PVDF 0.1 µm from Orelis Environnement SAS, France). We tested this material with the same Hele-Shaw measurement cell. As shown on figure X, the dye is able to go through the material. Diffusion was observed in less than 1 minute. This means this material is entirely suitable for our project since the 0.1 µm porosity prevents bacterial diffusion but allows peptides diffusion.

Perspectives:

Next step will be to test this membrane with our synthetic consortia to check whether bacterial diffusion is prevented while peptides diffusion allowed.