Software

In this iGEM project we faced the need to characterize a promoter, the NahR/Psal system (Schell et al, 1986) which is induced by salicylate. We decided to compare it with different amounts of salicylate concentrations and also with two constitutive promoters (the BBa_J23101 and BBa_J23110). Among the existing approaches for measuring the strength of a promoter, we chose to use fluorometry. We did not have access to a fluorometer in our lab, so we decided to build our own.

Fluorometry

Fluorometry is an analytical technique for identification and measurement of fluorescent compounds. It works by emitting light in a specific wavelength that excites electrons in the molecules of a respective compound. That causes the compound to emit light, typically within the visible spectrum. Instruments take advantage of this emitted light in order to measure the fluorescence level of the compound.

Custom Fluorometer Design

Since there are a lot of instruments that can measure fluorescence, we had to choose one that suits our case. Dependent on our low budget, we had to choose the best combination of a cheaper and easier approach, which as we concluded is the filter fluorometer. This type of fluorometer uses optical filters in order to let only the desirable wavelengths to pass, while blocking the rest of the spectrum (Curry et al., 1973)

Originally, filter fluorometers have one excitation filter between the light source and the sample, in order to let pass only the wavelength that excites the compound. Also, they have an emission filter in order to let only the wavelength emitted by the compound to pass out to a light sensor. The actual measurement is performed by the light sensor that captures the light’s intensity, that comes out of the emission filter. Based on that and taking advantage of existing materials, we had as a base of our fluorometer project the Arduino microcontroller (Banzi, M., 2011). The rest of the main fluorometer components are two green LEDs used as light source, one longpass optical filter at 590nm (FGL590) used as an emission filter, and a light sensor (TSL250R) that outputs an analog signal, with the voltage being directly proportional to the incoming light intensity. All of those components are attached in a 3D printed cuvette holder chamber (designed by Dr. Erik Gullberg), made specifically for this purpose.

dTomato

The dTomato protein is a fluorescent protein dimer, created by direct evolution of the wild-type DsRed, from Discosoma sp (Shaner et al, 2004). It emits orange-red light when it is excited by green-yellow light. It is preferable to use – especially in self-made fluorometry tests – because the excitation wavelengths and the emission wavelengths do not overlap as much as in other fluorescent proteins. The dTomato excitation peak is at 554 nm and 50% of it is at 510 nm. The emission peak is at 581 nm and the 50% emission peak is at 629 nm (Figure 1). Characterization has been made with the Anderson promoter BBa_J23110 (in construct BBa_K1688011).

To test the reliability of our constructed fluorometer, we wanted to compare the measurement with an existing method for fluorometry. MACS, Magnetic-Activated Cell Sorting is a method where the fluorescence of individual cells in a culture can be measured. When taking the average fluorescence of a large number of cells, the mean fluorescence of the culture can be deduced. Four individual colonies of DH5α containing the constitutive Anderson promoter BBa_J23110 and dTomato ( BBa_K1688011) on a pSB1K3 plasmid were transferred to overnight cultures as biological replicates. These were then diluted 1:300 in PBS buffer and measured on a MACSQuant VYB, Miltenyi Biotec. Excitation wavelength was set at 561 nm, with a 615/20 nm band pass filter for the emission. Parallel measurements were performed on the same biological replicates (not diluted) using our fluorometer.

The comparison between the fluorometer and the MACS showed that we can distinguish between the fluorescence of different biological replicates. The variance between the biological replicates were larger than the variance of the measuring method. The large variance between the replicates might have to do with the high-copy number of the PSB1K3 plasmid. This, in combination with the strength of the BBa_J23110 promoter results in a high load put on the cells. The stronger BBa_J23101 promoter was also tested, but showed unreliable results. The fluorescence of the cultures with dTomato under the BBa_J23101 promoter was lower than the ones containing BBa_J23110 , with higher variation between the biological replicates, which made the results unreliable. This may be solved through expression in a lower copy plasmid, but in this case a weaker promoter provided us with more reliable results. This experiment allowed us to scale our output data from the fluorometer and see that these results follow the expected values when compared to an existing measuring method.

The design of the custom fluorometer included two buttons and two extra indication LEDs. The function of one button is to start the measurement that is about to take place, while the other button is used in order for the user to change mode. One of the available modes is the Comparison mode and the other one is the Timelapse mode. The indication LEDs are used to indicate which mode is currently running. The Comparison mode is used in order to compare several samples interchangeably, by inserting a new cuvette with a new sample each time and pressing the start button. The Timelapse mode is almost the same piece of code, with the difference that its running time and the intervals between each read are way longer than the ones in the Comparison mode. The Timelapse mode is useful when we want to measure the changes in the fluorescence intensity of one sample over time.

Below in the Figure 3, is data gathered from the Comparison mode representing the relative fluorescence of each sample.

Finally, the fact that we use an analog light sensor means that the output values are very sensitive to external or internal noise. For this reason the circuit should be as insulated as possible and the output data should sampled and filtered. Since our fluorometer is in a prototyping phase all this period, we were not able to have a very good insulated circuit. Thus, we really had the need to have a good sampling and filtering algorithm. So, we sacrificed some speed in order to capture more samples per measurement, and we also implemented a combination of a median and a mean filter in our sampling data. The median part of the Median-­Mean filter, is modified to capture not a single median value, but several values around the median. Then, the mean part of the Median-­Mean filter calculates the average of those median values, for the final result.

A picture of the actual construct is presented in Figure 4.

Java

To read more about our java model, please click here.

References

Schell, Mark A. "Homology between nucleotide sequences of promoter regions of nah and sal operons of NAH7 plasmid of Pseudomonas putida." Proceedings of the National Academy of Sciences 83, no. 2 (1986): 369­373.

Curry, Robert E., Harry L. Pardue, Glen E. Mieling, and Robert E. Santini. "Design and Evaluation of a Filter Fluorometer That Incorporates a Photon­-Counting Detector." Clinical chemistry 19, no. 11 (1973): 1259-­1264.

Banzi, Massimo. Getting started with Arduino. " O'Reilly Media, Inc.", 2011.

Shaner, Nathan C., Campbell Robert E., Steinbach Paul A., Giepmans Ben N. G., Palmer Amy E. & Tsien Roger Y,. “Improved monomeric red, orange and yellow fluorescent proteins derived from Discosoma sp. red fluorescent protein”, 2004