Results

This page represents our highlighted and core experiments that allowed iGEM Guelph to produce our AmilCP and Violacein Biosensors. This is not an exhaustive list or presentation of the work done. For any further clarification, please refer to our lab notebook as all of our experiments are catalogued there. In order to create our biosensor, we began by setting out to determine whether the common lab Escherichia coli strains - DH5α and BL21(DE3) - would be able to survive high enough concentrations of tetracycline to be useful as a biosensor. This was done by performing minimum inhibitory concentration (MIC) assays in a 96-well plate. We also mapped a growth curve of both strains in the presence of tetracycline in order to see what quantities of the drug inhibited growth.

Initial Work with E. coli in Tetracycline

To test the tolerance of E. coli to tetracycline, DH5α and BL21 (DE3) strains were grown in a 96-well plate, each in 200 μL of LB, in serially diluted tetracycline from 40 μg/mL to 0.04 ng/mL. The plates were shaken at 900 rpm at 37 °C, and the optical density at 600nm was recorded after 24hours using the Tecan Infinite Series Pro 200 96-well plate reader.

When the ODs were normalized to the no drug control, we were able to calculate our MIC₅₀, which is the value of where the growth of the cells are inhibited by 50%. We saw that our MIC₅₀ was at 312.5ng/mL as described where the normalized OD values fell below 0.5. This means that half of the cells from both strains were able to survive in up to 312.5 ng/mL tetracycline compared to cells growing in wells without tetracycline. This was great news as 312.5ng/mL is slightly more than three times the regulatory limit for tetracycline in dairy products in Canada (100 ng/mL). With this information, we felt confident that both DH5α and BL21 (DE3) strains would survive above the regulatory limit and will be used for the development and initial testing of our tetracycline biosensor.

Now that we know our strains were fine, we ordered the plasmid pJKR-H-TetR (p-Tet) from addgene that is the backbone on which we based our tet-O DNA synthesis¹. We then needed to test the rate of growth of our strains in ampicillin and tetracycline as ampicillin will be needed to maintain the plasmid. We transformed our pTet plasmid that contained ampicillin resistance into our DH5α strain. We then created growth curves by growing E. coli in LB broth at 37°C shaking at 250 rpm overnight. The cultures were then diluted 1:1000 in 50 mL LB media. Based on the paper we obtained p-Tet from, the highest dose of tetracycline the researchers used was 267ng/mL. We wanted this experiment to quantify the impact that tetracycline would have on growth and capture the start of log phase for these cells. In our future work, we will be using lower volumes of tetracycline; So we wanted to capture growth in the highest concentration of tetracycline (and therefore the longest lag phase). So for this experiment we added tetracycline to a final concentration of 267 ng/mL and ampicillin to a final concentration of 100 μg/mL. For our negative control, we diluted the cultures into flasks that contained 100 μg/mL of Amp, without any tetracycline. The flasks were incubated shaking at 250 rpm in 37°C for 14 hours, taking absorbance readings at 600 nm every 2 hours.

From the analysis of our growth curves, we were able to determine that E. coli in tetracycline begins log phase after 4h of growth. Our control strain had almost no lag phase and started logarithmic growth almost immediately. This means that after 4 hours, with the highest concentration of tetracycline that we would use to induced expression of our pigments, we should clearly see colour being produced. We anticipate based on these results that the longest time is should take the cells to reach log phase would be 4 hours and with lower concentrations of tetracycline, it should take even less time. This information was used when determining when to harvest cells for expression testing.

Validating the Responsiveness of BBa_K3189001 to Tetracycline

The next step was to determine the sensitivity of our new biobrick, BBa_K3189001, a tetracycline-activated phage lambda promoter. To do this, DH5α and BL21(DE3) cells containing the pTet plasmid were grown in 50 mL LB media containing tetracycline to a final concentration of 50 ng/mL, as well as 100 μg/mL ampicillin to maintain the plasmid. We chose 50 ng/mL as it was under the regulatory limit and we wanted to push the limits of our test so we also measured 1ng/mL which previously showed no additional fluorescence beyond background levels (data not shown). The flasks were incubated and shaken at 37°C at 250 rpm for 6 hours. The 6 hour time point was chosen based on the point where we were confident that the cells were past log phase from our previous growth curve experiments. After six hours, the cells grown in flasks were washed and resuspended in PBS to reduce the autofluorescence by LB media. When assessed under UV light, it was clear that the 50 ng/mL of tetracycline showed a visible increase in fluorescence compared to the same cells grown in only 1 ng/mL tetracycline (Figure 3).

Since GFP was visibly produced with tetracycline present at only 50 ng/mL, half the maximum concentration of tetracycline allowed in dairy products in Canada, we concluded that BBa_K3189001 is sensitive enough to tetracycline to be useful in our biosensor.

E. coli BL21 and DH5α cells containing pTet were also grown in a 96-well plate with a decreasing gradient of tetracycline in the same manner as a MIC plate. We did this attempting to establish a linear relationship between concentration of tetracycline and fluorescence by growing pTet containing BL21(DE3) and DH5α cells in sequentially diluted concentrations of tetracycline in a 96-well plate. Unfortunately, we had difficulties with our Tecan Infinite Pro200 plate reader measuring fluorescence, so we were not able to establish a quantitative relationship between the two variables. However, we were able to visualize the 96-well plate under UV light to qualitatively assess this relationship (Figure 4).

What we observed was that as expected, fluorescence generally increased with higher tetracycline concentrations. The brightest fluorescence was seen at 312.5 ng/mL highlighted in red. This was interesting as this is the same value that is the determined MIC₅₀ of E. coli in tetracycline) This may be due to tetracycline affecting the cell’s metabolism and above their MIC50 that they grow so poorly or no longer have a functioning metabolism that allows them to adequately respond to the tetracycline by producing GFP.

Improvement of BBa_K1343022: Addition of BBa_K3189001

To create a biosensor we first wanted to create a basic sense and respond gene circuit that would produce a visible colour change in response to tetracycline. We needed a basic reporter gene and decided to use the chromoprotein AmilCP. We initially wanted to do work with BBa_K592009, the BioBrick created from the iGEM Uppsala 2011 team that has been well characterized. However this part in our distribution kit came without a RBS that we needed for our biosensor to transcribe amilCP. So instead we PCR amplified the AmilCP from the Biobrick BBa_K1343022 from Distribution Kit Plate 5 Well 22I. We realise that this means that the functional unit we are improving is the AmilCP BBa_K592009 in BBa_K1343022, but to not deviate from the iGEM rules, we have listed our improvement under the Biobrick BBa_K1343022. We placed, the Amil chromoprotein gene (amilCP) from BBa_K1343022, downstream of the BBa_K3189001, our tetracycline-responsive promoter. Initially, we attempted to transform the part in pSB1C3 as found in the distribution kit into E. coli DH5α, but no colonies grew on the LB+chloramphenicol plates (not shown). The amilCP of the BBa_K1343022 composite part, including the ribosome binding site immediately upstream (position 84-820), was then amplified by PCR, to put compatible restriction enzyme ends on it. The amilCP was then digested, along with pTet using EcoRI and PacI restriction enzymes according to our protocol listed to remove the GFP and give both DNA pieces sticky ends.The AmilCP amplicon was then ligated into pTet in the place of GFP.

The identity of the newly created BioBrick, BBa_K3189015, was confirmed by transforming it into E. coli DH5α, selecting colonies that grew on LB+Ampicillin. A colony PCR that amplified our BBa_K3189001 and AmilCP from BBa_K1343022 to give an amplicon of 759bp was done using Frogga Taq. We then ran the samples in a 1% agarose gel electrophoresis. This confirmed that the AmilCP was present in the right orientation, downstream of BBa_K3189001, our pLTetO sequence, creating the new plasmid pTA or our composite part BBa_K3189015.

To see whether BBa_K3189015 was functional, we grew up the cells in 50 mL LB flasks with 50 ng/mL tetracycline (as described above with GFP and in Experiments). After 6h of growth, the cells were washed with PBS and the characteristic brightblue colour was not observed.

We tried multiple tests with many PCR confirmed colonies however, we never saw the blue. To see why our part was not working, we sequenced an archive we made the day after we PCR confirmed the mutants from Figure 5 of strains with BBa_K3189015, using Sanger sequencing. What we found was a premature stop codon introduced at Gly 129, that mutated the first G in Glycine’s GGA to a TGA, resulting in a premature stop codon. This stop codon would result in protein structure of the AmilCP chromoprotein to be truncated, potentially explaining our results from Figure 6.
The forward and reverse sequences can be respectively found here and here

These findings lend support to what has already characterized about BBa_K592009 by iGEM Uppsala in 2018. They have observed that the amilCP protein is mildly toxic to bacteria, and thus the functional gene imposes a selective pressure that encourages the survival of nonfunctional mutants. However, this theory can be challenged by the fact that this mutation was observed when the gene was not suppose to be induced and expressing the toxic amilCP. The sequencing was done on strains that were archived after confirming their initial transformation (Figure 5). So another reasoning may be that the amilCP that was amplified out of BBa_K1343022 innately possesses a stop codon.

After several failed expression tests with our first version of BBa_K3189015 (based on BBa_K1343022) from iGEM plate we tried to circumvent the problem by synthesizing the sequence found in the registry from IDT. We repeated the same work in order to recreate a new BBa_K3189015, and this was sent for Sanger sequencing which returned an error-freesequence.

After PCR confirming and sequencing our BBa_K3189015, used a positive E. coli strain and commenced our initial expression testing.

We took an overnight of BBa_K3189015 and added it to LB-Amp with 100 ng/mL of Tetracycline and used no tetracycline in our negative control. They were grown at 37°C, shaking at 250 rpm. After 24 hours, a dark blue pigment was observed that increased in intensity after longer incubation. Figures 8a show the tubes that were incubated and Figure 8b shows a dark blue pellet that is observed when 2mL of each culture is spun down in the presence of tetracycline.

A more sensitive expression assay was then conducted on the functioning BBa_K3189015 by, growing multiple E. coli BBa_K3189015 transformants in a 96-well plate in varying concentrations of tetracycline labeled in Figure 9. Based on our characterization of BBa_K1349002, that showed amilCP expression is better at lower temperatures, the plate was left at 16°C for three days, and then moved to 4°C. After two days, pigments were visible, but they intensified after one week at 4°C. The cell density of the E. coli was quantified by its recorded OD 600nm after three days and the AmilCP produced was measured at a wavelength of 588 nm. The pigment production at 588nm, was quantified by normalizing the 588nm reads to the 24 hour cell density.

A dark pigment was seen at 100 ng/mL tetracycline, with less colour visible at 50 ng/mL for some of the samples, and no colour visible in the absence of tetracycline.

Based on the results in Figure 9, it would appear that our biosensor is sensitive enough to distinguish between 50 ng/mL and 100 ng/mL tetracycline, which is useful for seeing if a dairy sample is above the Canadian regulatory limit of 100 ng/mL. We demonstrated the improvement of BBa_K1343022 and characterised its capacity of having amilCP introduce stop codons that truncate the translated protein. However, further work will have to be done in order to create a standard curve relating the intensity of the pigment produced to the concentration of substrate, including normalizing to the number of cells present at the time of sampling.

Contribution by Characterizing the Expression of BBa_K1349002 at Different Temperatures

Given the instability of BBa_K1343022, the source of amilCP in BBa_K3189015, we also used BBa_K1349002 - a version of amilCP codon optimized for expression in E. coli by iGEM Uppsala in 2018 - to combine with the BBa_K3189001 promoter and create a tetracycline biosensor with E. coli optimized AmilCP, submitted to the registry as BBa_K3189014, using the same molecular cloning procedures as for BBa_K3189015 as outlined in the Experiments section of our Wiki.The identity of the part was confirmed by Sanger sequencing.

We used this construction in the characterization of the expression of BBa_K1349002 at different temperatures.

We decided to choose temperature characterization as through our literature review, we found that the source of AmilCP is the coral Acropora millepora. We wondered based on the Violacein’s need for colder conditions for pigment accumulation, if growing these strains at colder conditions allowed for the stronger expression of this pigment. This characterization is new for this version of amilCP (codon optimized for expression in E. coli) as previously temperature different did not impact the original AmilCP from biobrick BBa_K592009. We hoped that with the codon-optimization of this amilCP iGEM Guelph would obtain different results. Transformants were grown on LB plates containing 50 ng/mL tetracycline, and the plates were placed at 25 °C and 37 °C overnight. The next day, the colonies grown at 25°C were a dark colour, while the colonies grown at 37°C were visibly less pigmented (Figure 10).

From these results we have characterized the expression of BBa_K1349002 and can say that at 24 hours at 25°C it produces a stronger visual colour change when compared to colonies grown at 37°C for the same amount of time. This is most likely a result of reduced expression or stability of AmilCP at higher temperatures. This seemingly thermal instability makes sense given the fact that this gene was originally form coral which grows at lower temperatures in the ocean.

Validation of Violacein Biosensor: BBa_K3189005

In order to make our final biosensor design involving the violacein pathway, we first transformed the vioABE genes under control of the lac operator (taken from the plasmid pETM6-G6-vioABE (henceforward: BBa_K3189013) into E. coli BL21 (DE3) and tested their expression². BL21 was chosen rather than DH5α because the promoter for the genes in this fragment required a T7 RNA polymerase for transcription, which DH5α does not express.

BL21(DE3) BBa_K3189013 transformants were spread on LB agar plates containing 1 mM IPTG to induce VioABE expression. The plates were grown at 37°C for one day then transferred to 4 °C. After one day, a very dark green pigment was visible, and this pigment became darker after one week. It was noticed that the colour was darkest in isolated colonies and at the edges of densely populated areas, and decreased in a gradient towards the middle of lawn-like growth.

In order to investigate the uneven production of prodeoxyviolacein, we took colonies from the plate and performed both streak and spread plates on agar containing concentrations of IPTG ranging from 0.1 mM to 1 mM, to see if a lower concentration of IPTG would give a less intense colour. Instead, what we saw was that on all of the plates, the cells, which grew in a lawn, did not produce any clearly visible green pigment, even after a day of growth at 37 °C and several days of incubation at 4 °C. We hypothesised that the cells could have lost the plasmid containing vioABE so to test this we streaked to isolation from these plates onto a fresh set of LB agar plates with 0.75 mM IPTG, incubated them at 37 °C for a day, and then placed them at 4 °C. After 1 week, a dark green colour became visible, but only in isolated colonies and at the edges of lawn-like growth indicating the colonies growing in the lawn had not lost the plasmid and some other mechanism was inhibiting pigment production.

The consistent inverse relationship between cell density and pigment intensity indicated to us that some other mechanism was at play affecting the production of prodeoxyviolacein: either a quorum sensing mechanism was inhibiting the expression of vioABE in areas of high cell density, or competition for resources especially tryptophan (the initial reagent in the violacein pathway) in these areas was causing the cells to only use the nutrient for essential metabolic processes and preventing it from being used to synthesize the pigment. Further investigation into the literature showed us that in Chromobacterium violaceum, the organism in which the violacein pathway was first isolated, can have the expression of vioA inhibited by C10-homoserine lactone produced in the extracellular environment as a quorum sensing molecule.

We then cloned our new BioBrick, vioC (BBa_K3189000) together with the PLO-1 tetO promoter (BBa_K3189001) and many of our other basic parts (please see our parts overview). This part was then submitted to the registry as the composite part BBa_K3189005 which was then synthesized by IDT. The part was amplified using PCR, digested with EcoRI and NheI, and ligated into the pTet plasmid. The newly created pTV plasmid was then transformed into E. coli DH5α and colonies growing on LB+Ampicillin were verified using PCR to contain vioC. The expected band size of vioC is 1.509 kb and in the gel image in Figure 12, multiple colonies having a bright band at the 1.5 kb mark on the ladder which is also seen in the positive control thus confirming the presence of vioC in these colonies.

pTV was purified from the successful colonies using the MiraPrep protocol outlined on our Experiments page, and transformed into E. coli BL21(DE3) already containing BBa_K3189013 (plasmid containing vioABE). The cells were then grown on LB+Ampicillin at 37°C for 24 hours at which point they were moved to 4°C for 48 hours. The resulting raw transformation plates are pictures in Figure 14. Figure 14a is a raw picture of the plate with faint colours able to be seen clearly by the naked eyes. To more clearly distinguish the pigments being produced, Figure 14b has had the saturation increased, brightness decreased, and contrast increased. This image more clearly shows the pigmentation of the colonies ranging from green to pink. It can also be seen that the colour of the colonies is dependent on cell density with the more isolated colonies being green and the more closely associated ones being pink.

To ensure that these transformed colonies did not lose either of the plasmids, one pink colony and one green colony were re-streaked on a LB agar plate containing 100 μg/mL ampicillin, 100 μg/mL tryptophan, 1mM IPTG, and 50 ng/mL tetracycline. For both the pink and green colonies in Figure 15, the outer edge of the re-streak was a dark green colour whereas the interior was a light pink colour. For both pigments to be present means that a majority of the cells in these lawns have kept both plasmids. The difference in pigmentation is similar to what was seen in Figure 15, where more isolation leads to a green colour and the denser the colonies are the more pink is present until no colour is able to be seen.

From cells that contained vioABE and vioC, small colonies were dotted onto LB agar containing 100 μg/mL ampicillin, 100 μg/mL tryptophan, 1mM IPTG, and 50 ng/mL tetracycline. After 1 week at room temperature, colonies produced a dark, almost black pigment seen in Figure 16. What is interesting to note is that the colonies produced both green and pink halos around them. The colour enhanced colonies in Figure 16c show green halos pointed towards the outside of the plate when the part of the colony facing the inside produced a pink halo. The colony in the center of the plate however only produced a pink halo in every direction. Since the interior of the plate is more populated than the outside, this lends more evidence to the density dependent pigmentation seen in previous tests.

Point to note is that we tried another construct, on the advice and feedback from our oGEM presentation that suggested adding linkers between between the tetO->RBS and T7Term->S-Tag. This new construct (BBa_K3189006), after being synthesized by IDT, and cloned into our BL21 cells did not have any promising results (data not shown).

Finally, five different samples of E. coli BL21(DE3) containing vioABE and vioC were grown at 37 C for one day and then incubated at 4 C for one day on media containing both 23.5 ng/mL tetracycline, and 191 ng/mL tetracycline, with no IPTG or tryptophan added (Figure 17). The pigmentation of the colonies appeared to be more pink on the plate with the higher tetracycline concentration, and more green on the plate with the lower tetracycline concentration. This demonstrates the desired phenotype for the biosensor, with the colony colour being responsive to tetracycline concentration, although the intensity of the colour differs between the different samples.

Conclusions

Over the course of the summer, we have accumulated a wide range of data concerning the feasibility of developing a tetracycline biosensor using AmilCP or violacein.

We have validated the function of the tetracycline-inducible phage lambda promoter, BBa_K3189001, clearly demonstrating that induction by tetracycline causes it to increase the transcription of gfp, amilCP (BBa_K1343022 and BBa_K2669002), and vioC (BBa_K3189000) when these genes are placed downstream of it.

We have improved BBa_K1343022 by combining it with BBa_K3189001, creating a biosensor which produces AmilCP in response to tetracycline, and we have shown that in broth it expresses differently in response to 100 ng/mL tetracycline compared to 50 ng/mL tetracycline - a very useful trait in the context of Canada, where maximum tetracycline concentration permitted in milk are 100 ng/mL.

We have added characterization data to the relatively new E. coli-optimized AmilCP biobrick, BBa_K2669002, demonstrating that it is strikingly more visible when produced at room temperature, compared to when produced at 37 °C.

Future Directions - Extended Biosensor

Based on the results to date, it would be important to continue with robust categorization of the biosensor for tetracycline. Once this categorization has occurred for the biosensor to detect levels of tetracycline, the natural progression would be to extend the biosensor to detect other antibiotics. By utilizing other Violacein proteins and inserting them into the E. coli bacteria, the biosensor could detect multiple antibiotics at once.

This could be further extended by creating a suite of modified E. coli bacteria, each designed to detect one or two antibiotics. By combining these into a product package, the biosensors could act as a single test to detect all the necessary antibiotics found in agricultural run-off.

Future Directions - Other Projects

iGEM Guelph was fortunate to have the opportunity to explore multiple projects this year, before deciding to pursue Tetracycline Biosensor as our primary project. However, this left us with two major projects in limbo.

One of these projects involve utilizing CRISPRa in plants, and was halted due to time constraints, as the growth cycle of our plants were too long compared to the 4-month iGEM season. However, this gave us the experience we would need to begin the project next year and time our results to coincide with the 4-month iGEM season. This project will likely be pursued by the 2020 iGEM Guelph team.

The Beerstone project was halted for a variety of reasons, though none involving the infeasibility of the project. The team members who worked on the project retain their passion for the project and believe that it could develop into a potentially marketable project, or at the very least a strong research paper. iGEM Guelph intends to approach University of Guelph yeast labs with our data and see if there is a professor who will take this project on in their lab. This would ideally involve at least one or two iGEM Guelph members acting as student researchers in this new lab to continue the work on this project, creating research opportunities for iGEMers that may not have existed before.

References

1. Rogers, J. K. et al. Synthetic biosensors for precise gene control and real-time monitoring of metabolites. Nucleic Acids Res. 43, 7648–7660 (2015).
2. Alieva, N. O. et al. Diversity and evolution of coral fluorescent proteins. PLoS One 3, e2680 (2008).
3. Cress, B. F. et al. Rapid generation of CRISPR/dCas9-regulated, orthogonally repressible hybrid T7-lac promoters for modular, tuneable control of metabolic pathway fluxes in Escherichia coli. Nucleic Acids Res. 44, 4472–4485 (2016).
4. Stauff, D. L. & Bassler, B. L. Quorum Sensing in Chromobacterium violaceum: DNA Recognition and Gene Regulation by the CviR Receptor. Journal of Bacteriology 193, 3871–3878 (2011).

Team:Guelph/Results