Design
When we first decided to pursue the Anatoxin-a biosensor project, we envisioned to combine a part of a receptor that can bind Anatoxin-a with an intracellular signalling system in E. coli to result in an expression of a reporter protein. Since Anatoxin-a strongly and irreversibly binds nicotinergic acetlycholine receptors (nAchRs), we first anticipated to employ the soluble homologue, namely acetylcholine binding protein (AChBP). Even so, AchBPs are usually expressed by eukaryotic systems, there is evidence that it can also be efficiently expressed in E. coli (2).
However, after literature review and discussion with our Principal Investigator Libera lo Presti we decided AChBP would not be suitable for the use in our biosensor because of its pentameric structure and an unusual way of signal transduction compared to bacterial sensing systems.
Instead, we chose to employ the recently characterized PctD chemoreceptor that shows acetylcholine detection for ligand binding(3). We contacted Victor Sourjik (MPI Marburg), an expert in the field of creation of hybrid receptors, to advise us on the design of our receptor hybrid. He and Wenhao Xu, a postdoc in his group, confirmed that a combination between PctD and EnvZ, a histidine kinase part of bacterial two-component signalling systems, could be feasible. They offered us templates for PctD, as well as a reporter plasmid that expresses GFP in response to EnvZ activation.
We then approached Joachim Schultz (University of Tuebingen) to help us with some details of the fusion between PctD and EnvZ. We designed two versions of the chimeric receptor differing in the fusion site between the two receptors. Cells carrying the reporter plasmid, as well as these chimeric receptors in another plasmid, were envisioned to induce the expression of GFP upon ligand binding to PctD.
Build
The first step to be performed at the building stage in the wetlab was to assemble our selected hybrid receptors in DH5a cells. For this cloning step, we decided to use Gibson Assembly for its ability to produce scar-free fusion proteins, critical in our case, since the fusion point is crucial for the hybrid receptor functioning. Then, we amplified the different linear fragments via PCR that later were used for the Gibson Assembly and transformation in DH5a cells. Later on, to finish our device we transformed the correct hybrid receptor into the reporter strain VS1007 harbouring a reporter plasmid with GFP expression under the promoter OmpC. The chemoreceptors and EnvZ genes were deleted previously from this strain.
Test - Prediction of Anatoxin-a binding to PctD
The main goal we set in the drylab was to test the binding of Anatoxin-a to PctD. Besides, we wanted to predict potential mutations in the ligand binding domain that enhance the binding. We obtained the experimentally determined structure of PctD, the chemical structure of Anatoxin-a, the natural ligands acetylcholine, choline and betaine to compare predicted binding.
First we tested the potential binding site of Anatoxin-a to PctD via SSnet (4). To verify the results obtained from SSnet, we employed SeeSar (5) which ouputs predictions of affinities for potential binding sites.
Finally,we collaborated with iGEM Team iGEM Team Patras to utilize Gromacs (6) for Molecular Dynamics Simulation to predict binding while accounting for dynamic changes in the receptor structure.
Test - Anatoxin-a binding to PctD-LBD
Despite the unpromising results from drylab, we still wanted to test Anatoxin-a binding to PctD in vitro as well. The interaction characterization between PctD and Anatoxin A is a quantitative result that would determine whether our idea would fulfil the original purpose of measuring this harmful cyanotoxin. To do this, we had planned a Isothermal Titration Calorimetry Experiment with the natural ligands Acetylcholine and Choline as controls, previously performed for the characterization of PctD by Dr. Tino Krell, and Anatoxin-a. The first step for this experiment was the expression of the PctD-Ligand Binding Domain (LBD), in E. coli , that was purified via an Immobilized Metal Affinity Column (IMAC). The purification was successful and tested through with an SDS-PAGE gel. Then, we measured the binding of the PctD-LBD with the natural ligands previously mentioned in the microcalorimeter device.
Test - Biosensor construction
During the construction of our biosensor plasmids, the first test was getting the correct PCR fragments for the Gibson Assembly, which took some troubles due to the overlapping regions required for this cloning technique. After the transformation, the screening for the correct hybrid receptor construct was performed with colony PCR to see the length of the 2 fragments assembled with the backbone vector. Within the positive colonies, a representative number of isolated plasmid samples were sent to Sanger sequencing to confirm the right sequence.
For our engineered strain containing the reporter plasmid and the hybrid receptor expression plasmid, we tested different three conditions compared to the strain with only the reporter plasmid. First, we tested the induction with acetylsalicylic acid to produce the hybrid receptor at 2 µM and 4 µM, which was not noticeable in a SDS-PAGE where we analysed the cell extracts after 0, 2 and 6 hours, this being consistent with previous experiments by other iGEM teams with this promoter in a mid-low copy number plasmid. Second, we checked at the fluorescence microscope the presence of GFP expresion. To our surprise, the strain harbouring the reporter plasmid VS1007, derived from pUA66, showed fluorescence despite being under the regulation of the inducible promoter OmpC, as well as the strain containing the hybrid receptor with and without the ligand present in the media. Showing us a possible leakiness of the OmpC promotor. Third and foremost, with the intention to test the final application of our engineered strain, we performed a plate reader experiment where we tested different concentrations of the natural ligands during 4 hours in minimal media. The testing with Anatoxin-a was not possible since the provider had issues for the production, acquisition and delivery of the product, which they informed us on a short notice that they would no longer provide the compound.
Learn
For the learning stage, several setbacks pushed us back and taught us what to improve. First, the planification of experiments, including member or material availability and documentation plays a crucial role in the success of an experiment. Another underestimation we did was setting up the lab, including the material organization or antibiotics stocks, which are essential for the appropriate selection of the plasmids and strains. These together with a small team working part time, we learnt that everything takes longer than originally planned.
One of the wetlab experiments we planned and executed was the purification of PctD and the ITC measurement. The results we got from ITC suggest that, even though we successfully purified PctD, most of the protein is not functional. This means that we need to verify that our protein is denatured or misfolded and then evaluate and redesign our purification for future experiments and design measures to refold the protein, if needed. It also made us think about the importance of checking the protein functionality before conducting further experiments like ITC.
Another issue we had with the ITC measurement and also our biosensor testing was the fact that we were not able to get Anatoxin-A on time, because our deliverer postponed the delivery and then cancelled in on short-notice. From this we learned that it is important to have multiple sources for crucial compounds and that we act earlier when faced with delivery issues in the future.
One of the biggest issues we had to deal with in our initial experiments was the amplification of the DNA fragments by PCR. At the beginning we used the pfu DNA polymerase in all our PCR reactions, but it turned out that it only worked for some of them. We thought that the issue could come from the annealing temperature of the primers, so we did a gradient PCR to cover a wide range of annealing temperature as well as added Q5 GC enhancer buffer, but it was not successful either. Surprisingly, by using the Taq polymerase all the PCRs were successful, but due to its high rate of error we decided to not use it for our experiment. After searching in the literature and consulting with members of the lab, we finally decided to use the Q5 DNA polymerase for our PCRs, and it was a success as all the PCR reactions worked from the first time. From this we have learned that PCR troubleshooting is an important part to know and probably we will have to deal with it in the future.
At advanced stages of our experiments, we found out that our chimeric receptor has an unexpected point mutation close to the binding pocket that might affect to the affinity and specificity. This was really unfortunate as the mutation was already existing in the pctd sequence that we used as template, but we did not realise in the beginning. Therefore, we have learned from this experience that checking the starting material in the beginning of a project is a crucial step to do because otherwise you might find some unexpected events at later stages that could break all your plans.
Upon testing of our biosensor qualitatively and quantitatively we found that our reporter plasmid introduces some background fluorescence, and that our biosensor candidate VS1007 PEH shows high fluorescence without ligands present. But we could not observe a clear change in fluorescence when acetylcholine or choline where added, therefore not verifying any sensing of these substrates. This indicates that our chimeric receptor might not work. The testing of the biosensor therefore needs to be redone to verify whether the substrates have no effect on the biosensor. Then we need to start another engineering cycle.
Drylab results predicting the binding of Anatoxin-a to PctD were also not promising. In all three used tools, Anatoxin-a binding was considerably weaker at the acetlycholine binding site or the optimal binding site was predicted in the intracellular part of the receptor, which in our hybrid receptor design is replaced by EnvZ.
When starting another engineering cycle and re-designing our chimeric receptor, one way to do it is by prediction of structural changes to the ligand binding site affecting the structure in the wetlab. Initially, the plan was to use the drylab results to predict how binding of Anatoxin-a to PctD could be improved. However, we learned talking to Prof. Thiel, that prediction of effects of mutations on binding is difficult to conduct reliably. Therefore, we would need to shift this part entirely to wetlab, by having a lybrary of different variants of the chimeric receptor synthesized which could then be screened.
Other steps we want to take in our re-designed project plan is to generate primers to use PCR to remove of the mutation in our current “PEH” chimeric receptor and also reassess the expression and purification of our PctD-LBD to redo the binding study of acetylcholine and choline to the protein. When this has been successful, testing with Anatoxin-a is the next step.
Overall, when looking back at our project, we can say that we gained a lot of insights into the different stages of engineering, that we could use to improve our project and future projects. We learned that establishing methods for building and testing a biosensor takes a lot of time and that controls are crucial on every step of the process.