Introduction
In order to establish a basis for biosensor development, a series of experiments were conducted using the library of promoters and the ten molecules we selected. First, each promoter and molecule combination were tested for fluorescence using strains started up in 96 well plates. We then processed the data to identify the promoters that produced the most fluorescence for each of the molecules. After the screen, we tested the levels of fluorescence the promoter produced with different concentrations of the molecule using titrations, trouble-shooting to determine the most promising combinations. A labeling system was used for each of the molecules: PBA represents 3-Phenoxybenzoic Acid, LOV represent Lovastatin, PRO represents Propoxur, DEP represents Diethyl Phthalate, TAR represents Tartaric Acid, CAR represents Carbaryl, BHL represents Butanoyl-Homoserine Lactone, PGA represents Phenylglyoxylic Acid, PFS represents Perfluorooctane Sulfonate, and CND represents Cis-Naphthalene Dihydrodiol.
Initial Screening
An initial screening round was done where each promoter was tested with each molecule at 1mM concentration to identify which ones produced the highest fold increase in fluorescence.
Table 1. Promoters and fold that produced the highest fold in fluorescence (sfGFP/OD600 value greater that 1.35) for each of the ten molecules
From the screening of 2000 promoters, around 8 to 12 were selected for each of the molecules that produced the highest levels of fluorescence following normalization by the density of cells in the well and by the background fluorescence of each promoter.
Titrations
Once we identified the top fold increases for each molecule, those promoter strains were titrated across 8 concentrations of their corresponding molecules to examine how the fluorescence varies with concentration. The titration graphs are shown below, with the error bars representing ±1 SD.
Figures 1-10. Titration graphs for all the molecules with trends for each strains
We predicted that the fluorescence would increase with increasing concentrations of the respective molecules since the molecules serve as an inducer for GFP transcription. However, many of the strains actually showed a linear or decreasing trend and had overlapping error bars of standard deviation, indicating the results aren’t statistically significant. The level of fluorescence in the titration at 1 mM was also much lower than that of the original screening, so we took several steps to troubleshoot for these issues.
Redone Strains
Certain titrations were redone using the same procedure if they had an OD600 value less than 0.1, as this value is outside the threshold of the plate reader, in order to correct the molecule concentrations or obtain better cell growth for the plates. The titration graphs are shown below, with the error bars representing ±1 SD.
Figures 11-13. Titration graphs for each molecule containing updated data from the redone strains.
These results indicated a slightly better increasing trend that aligned with our predictions and also made the results more reliable, as fluorescence from a higher cell density can be measured better and trusted more. However, there still weren’t any strains that showed a strong increase or could potentially be used to make an effective biosensor.
Fresh and Old Strains
One possible source of deviation between the screen and titration was the freshness of the strain, as the time in the incubator and left on the bench differed between the two procedures, so the titration for the top 4 strains for 3 randomly selected molecules were redone using strains that were incubated for only one day (fresh), just as the initial screening had been. The titration graphs are shown below, with the error bars representing ±1 SD.
Figures 14-16. Comparison of fluorescence levels in fresh and old strains for certain strains grown in both conditions with the molecules
The fresh stains notably increased the level of fluorescence produced by each of the molecule-promoter pairs, as shown in the figures above. It also largely eliminated the sudden peaks at random concentrations that were present with the old strains. This may be due to better efficiency of biological processes when the strains are recently grown. However, the deviation between the initial screen and titration still wasn’t completely resolved.
Media Comparison (old media compared to new media)
After discovering that the M9 minimal salt media was contaminated, the titrations for the top 4 strains of each molecule (except PFS and LOV due to lack of molecule left) were redone. The titration graphs are shown below, with the error bars representing ±1 SD.
Figures 17-24. Comparison of titration fluorescence levels in new and old media for certain strains grown in both conditions with the molecules
Using the new media, there was a lack of contamination to affect the results of the research or interfere with the production of a fluorescent signal. Largely, the new media further increased the level of fluorescence for each of the strains; however there were some exceptions to this. These exceptions may be due to interference of other environmental factors or experimental error. The new media also changed the trend of the data from linear to slightly increasing with increasing concentration of the molecule, which is what was originally expected. The increasing amount of fluorescence provides promising results for the development of a biosensor since the concentration of a molecule present can be estimated from the amount of fluorescence produced.
Corrected Strains
Our team also realized that some of the promoters that were matched to wells with a high fluorescent signal in the initial screen were mismatched. As a result, we retested these molecule-promoter pairs with the correct promoter and observed the results of the titration. The titration graphs are shown below, with the error bars representing ±1 SD.
Figures 25-26. Titration graphs for each strain with corrected matching
While the newly titrated strains had a largely decreasing or linear trend rather than the expected increasing trend, it did produce higher levels of fluorescence with respect to a single molecule. This means that only one of the molecules induced the fluorescent response in a 1 mM solution, which is expected based on our background research.
Final Strains
After several iterations through designing, conducting, and reevaluating our experiment, we finally selected some strains with promising titration results, as shown below. We primarily selected these strains due to a high fold increase across different concentrations, significantly different results, and an increasing trend with increasing concentrations of the respective molecule. The error bars represent ±1 SD.
Figures 27-30. Titration graphs for selected strains that have the best potential for future biosensor development based on data analysis
BHL with the ydel promoter produced a 1.7 times fold increase. CND with the ybcK promoter produced a 2.2 times fold increase. CND with the aegA promoter produced a 2.8 times fold increase. Finally, DEP with the yfiF promoter produced a 2.0 times fold increase. While the level of fluorescence has some drops and inconsistencies at certain concentrations, the overall significant, increasing trend indicates that these strains can be further researched to be eventually developed into effective biosensors to measure the concentrations of their corresponding molecules.
