Team:Vilnius-Lithuania/Results
RESULTS
Prevention
Promoter characterization
To assure the most efficient possible naringenin production pathway, we had to select the most suitable promoters for the expression of naringenin synthesis genes. This was done by evaluating super folder green fluorescent protein (sfGPF) expression rates under the promoters of interest and dividing the intensiveness of the signal by the OD600 of the medium during the course of 6 hours. The data showed that (Fig. X).
Evaluation of transcription efficiency dependency on genomic site
We seek to create naringenin producing probiotics. For this reason, we decided to insert naringenin metabolic pathway encoding genes into E. coli Nissle 1917 genome. This experimental decision helps to overcome the problem of additional antibiotic usage and reduce the fluctuations gained because of unstable plasmid copy numbers in cells. Firstly, to measure the transcription activity from two genomic regions, we have inserted sfGFP into colicin and nupG genes (fig. X, X) and compared the amount of fluorescence (fig. X).
mRNA cyclization system evaluation
To evaluate the quantity of synthesis by our constructs/enzymes, we employed HPLC-MS to find naringenin and intermediate compounds. All enzymes were subjected to analysis first by themselves and further in different combinations. Both control for native cellular metabolism and with additional substrates were taken into account.
Metabolic pathway construction
To construct the pTRKH2 vector containing all four genes of the naringenin metabolic pathway we amplified pTRKH2 vector and all four naringenin synthesis genes using primers that contain specific restriction endonuclease recognition sites. This way we should have been able to digest each sequence with appropriate restriction enzymes and create a library of inserts with sticky ends, that can be ligated into the target vector as the ending part of composite insert or as a part of the whole naringenin synthesis cassette. However, we were only able to construct plasmids containing only TAL and TAL+4CL, cassettes that later were found to have been mutated by Sanger sequencing.
To get the construct containing all four genes we chose the strategy of Gibson assembly. By amplifying the pTRKH2 vector and all naringenin synthesis genes with primers containing flanking regions that form homologous pairs with each other in the manner that a complete naringenin synthesis cassette should be constructed in a single tube reaction. Nevertheless, we were not able to obtain the desired construct, only the plasmids.
Furthermore, to enhance naringenin synthesis in E. coli Nissle 1917 we created ackA-pta double knockout. Firstly, we knockouted ackA (fig. X), and pta (fig. X) genes separately. Later on, we used ackA knockout to generate ackA-pta double knockout (fig. X). As we see in (fig. X) sgRNA designed for ackA knockout creation shows 100 % efficiency as all randomly selected colonies had desired changes in ackA gene. For further experiments chosen ackA knockout have been verified by ackA gene sequencing. pta knockout also have been generated with 60 % efficiency (fig. X). In addition, ackA-pta knockout have been created by generating pta knockout from ackA knockout with 80 % efficiency of pta knockout generation.
Our next move was to create tyrP knockout. Firstly, we have successfully obtained tyrP knockout (fig. X).
However, we have not succeeded in creating double or triple knockouts (ackA-tyrP or ackA-pta-tyrP). As you can see in the fig. X, restriction of cPCR product from randomly selected transformants do not show genomic modification in the tyrP gene. Interestingly, in the positive control line we can see very large DNA fragment (> 10 kbp). We could not explain this result without additional genomic analysis.
We have repeated PCR from previously obtained tyrP knockouts and all knockouts had this one sharp fragment above 10 kbp ladder line (fig. X). These surprising results might be obtained because of some unknown genomic reorganization of edited genomic locus. The exact reorganization output can be determined by genome sequence which was not in our focus. As we seek to avoid usage of undetermined changes containing E. coli Nissle 1917 strain, we decided to not use tyrP knockout in further experiments. In addition, we do not succeed in obtaining adhE gene knockout even after testing two different sgRNAs.
Kill-switch
To quantitatively evaluate VapXD kill-switch performance, OD600 measurements were performed. First of all, VapD toxin (BBa_K3904000) activity was characterized while regulating its production with cold-induced promoter (BBa_K3904003). Graphs at the top of figure X illustrate bacteria growth without toxin and graphs at the bottom with toxin in different temperatures. While comparing obtained data in 37 and 24°C, temperature change can be seen as a factor inducing greater toxin production and cell death. On the other hand, VapX activity is not fully accurate due to the leakage of the promoter.
Furthermore, VapXD with the bile-induced promoter before antitoxin and with the cold-induced promoter before toxin was characterized. Graphs in the top of X figure demonstrate bacteria growth with and without bile salts supplementation in media at 24°C, as graphs in the bottom at 37°C. It can be seen that OD600 in the presence of bile salts and 37°C bacteria grow more exponentially than without bile salts and in 24°C. In the ideal case, no antitoxin should be produced in the absence of bile salts and 24°C, and toxin synthesis should be induced. However, the results indicate that in such conditions, bacteria growth is only slightly repressed.
When results in 37°C obtained with different OD600 values were averaged (Fig. X. Mean comparison with/no bile 37°C), the difference between measurements with and without bile salts appeared to be mathematically insignificant.
What is more, the activity of different promoters before VapX toxin was compared without bile salts supplementation in media at 37°C. From promoters' strength evaluation measurements (fig. X), it was seen that the p-slpA promoter ( BBa_K3904712) is the strongest in our inventor. VapXD assessment also showed that under this promoter, toxin production is more significant than under other promoters. The sloping graph rise illustrates this because more toxin is produced, and bacteria growth is inhibited. However, after the results were averaged, no significant difference between different promoters is seen.
While comparing results with different promoters in 24°C no significant difference can be seen (fig. X).
Furthermore, we successfully inserted the VapXD system construct into E. coli Nissle 1917 genome (fig. X, fig. X).
Naringenin evaluation
To evaluate the quantity of synthesis by our constructs/enzymes, we employed HPLC-MS to find naringenin and intermediate compounds. All enzymes were subjected to analysis first by themselves and further in different combinations. Both control for native cellular metabolism and with additional substrates were taken into account.
First, we created control chromatograms for naringenin and first enzymatic intermediate - p-coumaric acid (product of Tyrosine ammonia lyase (TAL) from naringenin synthesis pathway). By dissolving technical grade compounds in pure water we found retention times:
| Naringenin | p-coumaric acid |
|---|---|
| 6.88 min | 6.17 min |
This information enabled us to search for compounds in more complex mixtures, in particular LB medium from overnight cultures. Furthermore, we were able to distinguish our products based on retention time, m/z and UV-Vis absorption spectrum.
Using the HPLC-MS method we analyzed the media samples of cultures containing pTRKH2 vectors with TAL gene and J23101 Anderson or surface layer protein A (slpA) promoters. The data of the experiments showed that plasmid with J23101 Anderson promoter and TAL encoding sequence determines an efficient synthesis of p-coumaric acid in our transformants, nevertheless analogous process has not been identified in the samples containing slpA promoter. We hypothesized that the reason for this data non-reproducibility may be a possible mutation in pTRKH2 vector containing slpA promoter, the hypothesis later was approved by Sanger sequencing.
Similar intermediate compound detection strategy was applied for constructs containing 4CL and CHS encoding sequences. We supplied cultures with p-coumaric acid as a substrate for their specific reactions and used HPLC-MS method to detect the consumption of p-coumaric acid. The experiments were conducted with cultures containing pTRKH2 vectors with 4CL gene and J23101 Anderson or slpA promoters, as well as linked (linkers: GSG, GGGGS, (GGGGS)2, (GGGGS)3, EAAAK, (EAAAK)2, (EAAAK)3) 4CL and CHS genes under the same promoters. However, none of the aforementioned constructs have been found to demonstrate distinct enzymatic activity by consummation of the given substrate.
In the hopes of finding further intermediates we searched for a few additional m/z as a result of accumulation. p-coumaroyl-CoA and naringenin chalcone were chosen as the ones who could give us more information. However, fusion protein samples did not show any signs of naringenin chalcone. Moreover, we found the same m/z of p-coumaroyl-CoA in both control and sample from the desired construct medium when supplied with additional p-coumaric acid. This suggested to us that we could not precisely determine the quantity of synthesized p-coumaroyl-CoA because we do not know detailed information about internal processes. We hypothesize that control E. coli DH5α has 4CL homology enzymes for forming carbon-sulfur bonds as acid-thiol ligases and thus synthesis of p-coumaroyl-CoA by both recombinant and native enzymes overshadow one another. We even cannot be sure about synthesis of a particular compound because it needs further analysis by NMR as chromatograms and UV-Vis spectrum lack structural information.
Detailed reports from HPLC-MS are referred to in table X.
| Name | Description | File |
|---|---|---|
| Technical grade p-coumaric acid | Technical grade p-coumaric acid was dissolved in water for reference chromatogram and other specifications. | Download |
| TAL1 lysate | First enzyme Tyrosine ammonia lyase (TAL) in pTRKH2 plasmid with supplied p-coumaric acid to LB medium. Culture was lysed to check for compounds inside cells. | Download |
| TAL1 LB medium | First enzyme Tyrosine ammonia lyase (TAL) in pTRKH2 plasmid with supplied p-coumaric acid to LB medium. Only LB medium was subjected to HPLC-MS analysis. | Download |
| DH5alpha with p-coumaric acid | Only DH5alpha without any plasmid cells were grown in LB medium with supplied p-coumaric acid. | Download |
| LB with p-coumaric acid | LB medium with p-coumaric acid. | Download |
| LB without p-coumaric acid | LB medium without p-coumaric acid. | Download |
| GS1 LB medium without p-coumaric acid | Fusion protein (4CL and CHS) construct with GGGGS linker in pTRKH2 plasmid without supplied p-coumaric acid to LB medium. | Download |
| GS1 lysate with p-coumaric acid | Fusion protein (4CL and CHS) construct with GGGGS linker in pTRKH2 plasmid without supplied p-coumaric acid to LB medium. | Download |
| GS1 lysate with p-coumaric acid | Fusion protein (4CL and CHS) construct with GGGGS linker in pTRKH2 plasmid without supplied p-coumaric acid to LB medium. | Download |
| GS1 LB medium with p-coumaric acid | Fusion protein (4CL and CHS) construct with GGGGS linker in pTRKH2 plasmid with supplied p-coumaric acid to LB medium. | Download |
| GSG LB medium with p-coumaric acid | Fusion protein (4CL and CHS) construct with GSG linker in pTRKH2 plasmid with supplied p-coumaric acid to LB medium. Culture was lysed to check for compounds inside cells. | Download |
| Naringenin | Technical grade naringenin dissolved in distilled water. | Download |
Detection
Entamoeba histolytica recombinant protein synthesis
In order obtain our aptamers through the SELEX process we first needed to produce recombinant Entamoeba histolytica proteins - pyruvate, phosphate dikinase (PPDK)[6] and cysteine proteinase 5 (CP5) [7]. The synthesised CP5 gene was cloned into a pET-28a(+) vector plasmid which is used to tag proteins with a histidine tag on both C and N terminus. The PPDK gene was first cloned into a pET-28a(+) we modified to not contain extra amino acids before the histidine tags but the constructs kept mutating so we decided to clone it into the default pET-28a(+) plasmid to keep the chance of mutation to a minimum. The conditions of biomass growth were similar for both proteins, the induction was carried out in the BL21 strain of E. coli for 3 hours in 37°C with a concentration of 0.6 mM IPTG for CP5 and 1 mM for PPDK. The bacteria containing the PPDK plasmid also needed to be grown in TB medium while the CP5 containing bacteria could be grown in simple LB medium[8, 9].
Since an active protease would lyse our cells the CP5 gene contains an inactive and insoluble pro-enzyme that has an N-terminal pro-sequence that can be cut off by the enzyme itself. The purification of CP5 was supposed to be done in three distinct steps: denaturation, renaturation and activation. This strategy failed because the whole process itself was too costly as there were too many steps and the yield of the protein was too low. We could only see the same pro-enzyme band in the SDS gel after the final step, but we think that the most important step in the purification is the ratio between the protein and the refolding buffer because we found different methods of refolding where the ratio ranged from 1:100 to a 1:1000 [7, 8] and that is the only thing we could not check as we could not manage to concentrate the protein again from a volume this large.
Although we failed to produce the CP5 protein, we still had chosen another biomarker - pyruvate, phosphate dikinase (PPDK). The purification of this protein was supposed to be much simpler based on the literature[9], but we faced the same problems here too. Although we found some of our target protein in the soluble fraction, the amount was miniscule compared to what we needed to evolve the aptamers. It seemed like it also formed insoluble inclusion bodies just like the CP5, we hypothesise that it happens because of the extra amino acids on the pET-28(a+) backbone. We increased the solubility somewhat by losing the cells in the presence of a couple of detergents - Triton X-100 and NP-40 but the protein was still barely visible on a western blot membrane.
We tried out the same stabilising strategy with PPDk too [10] and this time it worked perfectly, we found out that denatured PPDK can be transferred and stabilised to a buffer containing 0.375 M arginine through dialysis[11]. PPDK was then further purified through nickel affinity ch chromatography and used in the SELEX process.
SELEX
There are plenty of different ways to complete a SELEX in vitro evolution of aptamers using nitrocellulose, microfluidics, whole cells and other materials as the scaffolding for the biomarkers[12, 13], but we decided to use the basic selex protocols using magnetic beads[14]. A protein with a His-tag can be attached to the beads and after an incubation with the aptamer pool we can wash off the ones with weak interactions and do a PCR reaction by loading the beads directly to the PCR mixture to amplify the target aptamers. The selex process itself seemed to be successful. We did have to optimise the selection parameters a bit, the PCR reaction only worked when we substituted the TE buffer for water as the aptamer storage solvent because even miniscule amounts of EDTA did not let the Taq polymerase amplify the strands. The magnetic beads themselves also seemed to inhibit the multiplication process when it was done by Taq polymerase so we decided to compare different polymerases and settled on Phusion because it worked the best.
Emulsion PCR
We found creation of emulsion an easy task, but nonetheless few problems occured. First, version 1 emulsion showed irresistance for thermal cycling and micelles broke even after 15 cycles. For this we tried creating emulsions in colder conditions and by mixing for longer. None of these showed better results. Furthermore, comparison of products between ePCR and oPCR was done. In figure X different cycle count was used and on this basic data we can see that open PCR started generating non-specific fragments after 25 cycles and emulsion PCR lagged at overall production of fragments but in the end did not create same longer fragments as seen in oPCR.
ePCR v1
ePCR v1
We observed produced micelles using fluorescent microscopy (400x Magnification) with purified GFP shown in figures X and X.
Micelles were stable at room temperature while observing them.
ePCR v2
We tested updated composition emulsion and it did not show any signals of breakage even after 50 PCR cycles. The main problem with this is when we want to check nonspecific PCR products in electrophoresis. It is not an easy task to break emulsion with neither 1-butanol, nor isopropanol. However when used in PCR purification kit the emulsion has gone from cloudy to clear from binding buffer and centrifugation at 20.000 rcf.
To recreate both versions of emulsions
Observations
Distribution among PCR tubes by 50 µl leaves quite visible mineral oil smear on pipette tips. It is better to produce more of the overall mixture for higher yield. As per visual examination emulsion breaks even after 10 cycles of PCR which suggests that different emulsifiers should be used.