By reading the last final report (Arsenic-Prototype-3-0), we saw that the prototype detected much more fluorescence in the most concentrated bioreporter solution than the lab equipment. We did the assumption that this problem could be due to light scattering of the blue LED because bacteria's shape might deviate this blue light. So, our part of the project was to quantify, observe and test the light scattering in the prototype.
To determine if the sensor also detects blue light, we wanted to compare eGFP bacteria and lysed eGFP bacteria. To test this assumption we realized a cell lysis protocol. We chose the Three-Phase Partitioning method inspired by "William W Ward of Brighter Ideas, Three-Phase Partitioning for Protein Prufication" " pdf
Once the cell lysis protocol was performed, we analysed the light spectrum of different samples using spectruino to compare the spectrum of green fluorescense and blue light in eGFP bacteria and free GFP from lysed cell. The different samples were: Free GFP from cell lysis, eGFP bacteria, fluorescent dextran as negative control because we know that dextran does not scatter and fluorescent beads as positive control.
This figure shows what we would observe if the light scattering were present:
The green fluorescense disappears in the blue light.
This figure shows what we would observe if the light scattering were small:
We would be able to observe two well separated peaks.
If the light scattering were significative, the eGFP bacteria spectrum would look like the first figure. And if cell lysis would decrease light scattering significatively, the free GFP from cell lysis spectrum would look like the second figure.
We tried to test directly with the prototype, after having exchanged the sensor by the spectruino, but unfortunately that didn’t work because the length between the spectruino and the vial was too high. To solve this problem, we built a box serving only to test the light scattering.
Light scattering is a physical process where light is forced to deviate from a straight trajectory. The bacteria's shape deviate the blue light and induces light scattering. We would like to know if the blue light, which excites GFP in bacteria at 475nm, is also detected by the sensor because it can’t distinguish between green and blue light. If it's the case the light detection would be higher and arsenic concentration would be distorted.
To determine if the sensor also detects blue light we would like to compare eGFP bacteria and lysed eGFP bacteria, since bacteria's shape is the first source of light scattering. To test this assumption we realized a cell lysis protocol.
We drew our inspiration for the protocol from the following paper: « William W Ward of Brighter Ideas, Three-Phase Partitioning for Protein Purification ». We chose the Three-Phase Partitioning method because it seemed to be an easy and interesting method to separate GFP from bacteria. We were following the protocol from the paper but at the step where we should have seen if the concentration of the starting buffer was good we didn't obtain the good result and each concentration seemed to be higher. We decided to go our way and to try with three different concentrations (1.6 M, 3M and 4M) of starting buffer. Finally the result gave that the GFP was not visible to the naked eye but with fluorescence analysis the GFP was existing in the 3 samples. The 1.6M of starting buffer has the higher green fluorescence value so we decided to perform our protocol with this concentration. We performed again the experiment with more bacteria to obtain more green fluorescence with this following protocol:
1.6 M starting buffer composed of: 10.57g of ammonium sulphate 2.5 ml of 50nM of Tris Buffer pH 8 add water until 50ml eGFP bacteria Ars reporter control bacteria tert-Butanol (ref: 19460)
• Compare absorbance of eGFP bacteria and bacteria control (to see if we can put the same volume) • Take 50ml of eGFP bateria and 50ml of control bacteria • Centrifuge 2 tubes at 10 000 rpm(or the max on the bug centrifugeuse) during 10 min (to convert rpm in rcf RCF = 1.1118 x 10-5 x r x rpm^2) • Remove supernatant, we only keep the pellet bacteria • Add 5ml of 1.6M of starting buffer and 5ml of t Butanol in the 2 tubes • Vigourous shaking during 1min (vortex) • Centrifuge during 10min at 1000rpm • Observe 3 phases: butanol, membrane and constituent cells, GFP phase for eGFP bacteria and without GFP for control. • Keep only the lowest phase, which is GFP phase for eGFP bacteria and without GFP for control and remove the rest • Add 500μL of t Butanol • Centrifuge 2 tubes at 2000rpm during 10min • We obtained again 3 same phases 20150421_110409 • Keep only the lowest phase, which is GFP phase for eGFP bacteria and without GFP for control and remove the rest • Measure green fluorescence , excitation at 488nm
absorbance of eGFP bacteria: 1.96 OD absorbance of control bacteria: 2 OD
so we can put the same volume (50ml) because the OD values are close enough
control bacteria: 873
1.6M starting buffer: 1
We measured the fluorescence of 1.6M starting buffer to compare the green fluorescence between bacteria with GFP, bacteria and sample without bacteria. The fluorescence of control bacteria of 873 can be explained because lots of things fluoresce in green in biology. The difference between eGFP and control bacteria fluorescence is significant. We obtained about 4ml of free GFP (from lysis of eGFP bacteria).
 William W Ward of Brighter Ideas, Three-Phase Partitioning for Protein Prufification pdf
We would like to determine if blue light is also present in green fluorescence detection. With this different test we want to quantify the blue light scattering. We analysed the light spectrum of the following samples to compare the amount of blue light and green fluorescence:
- Free GFP from cell lysis protocol
- eGFP bacteria
- Fluorescent dextran as negative control because we know that dextran does not scatter
- bacteria lysed without GFP (control)
We first tried to test this light scattering with the Arsenic-Prototype-3-0 and finally we found and a better solution and built a new box.
With the prototype
To determine if blue light scattering affects green fluorescence detection which would distort arsenic concentration, we analysed the light spectrum of different samples using spectruino.
We wanted to test our sample with the first prototype.
For that, we removed the detector, drilled a small hole and placed the spectruino to test the blue light and the green fluorescence.
We took care of the spectruino angle and spectruino height with regard the prototype.
But Spectruino didn’t detect any light spectrum!!! To solve it we tried:
- To close the box.
- To change spectruino position.
- We test only with blue light and after only with red light by changing spectruino code.
Nothing happened with spectruino!!!
Finally we tried the following tests:
We can see two peaks, one blue peak with an intensity about 80 and a green peak with an intensity about 80.
We can see one blue peak with an intensity about 90
With Free GFP: We inserted free GFP sample between the blue led and the spectruino at 10cm.
We observed only one little blue peak with an intensity about 30
The green peak disappeared
We placed the blue light in front of spectruino at 4cm and increased the distance until 15cm (to be sure that the problem doesn’t come from the angle).
CONCLUSION with Prototype
The prototype has a length about 23 cm and the light is not in front of the spectruino but it deviates. So the spectruino is not sensitive enough to detect the blue light and the green fluorescence. We thought to use the second prototype, but its length is 12 cm, still too far away, and the light is deviated so same conclusion : spectruino cannot detect the spectrum. So we decided to create a box to test our sample.
With our Test Box
We are using the spectruino to quantify how much light scattering influences the detection of green fluorescence.
We would like to find if the blue light, which excites GFP in bacteria, is also detected by the sensor because it can’t distinguish between green and blue light.
In order to compare the spectra and observe scattering, we tested several samples:
- Free GFP from cell Lysis
- eGFP within bacteria
- fluorescent dextran and cell lysed without GFP (control)
As described in the previous section, when we tried to test with the prototype, we noticed the length between the sample and the spectruino detector was too big and the spectruino did not detect any spectrum.
To address this, we decided to build a custom box.
The blue LED goes through the vial and spectruino detects the resulting spectrum. The spectruino is connected to a computer which displays the spectrum.
We chose a closed box to eliminate stray light.
The second step was to fix the vial, the blue LED and the spectruino with the right angle, to have the best detection of the spectrum.
We used polystyrene to stabilize the system because it can be found anywhere for free.
We also added a hole to facilitate the changing of the vial.
Now that we have the samples AND the test box, we can finally quantify the light scattering. We obtained different spectra with our 4 samples.
eGFP bacteria sample:
We observe one big blue peak which is a little bit extended on the green spectrum.
Free GFP sample:
We observe two peaks: one on blue spectrum, one on green spectrum. The two peaks have the same amplitude.
We observe two peaks: one on bigger blue spectrum, one on smaller green spectrum
lysed cell without GFP (control):
We observe one blue peak.
We would like to observe two peaks in eGFP bacteria sample, because it would mean that the scattering is not important. But we observe only one peak and the green peak is hidden by the big blue peak.
We estimate that the blue peak in Control sample has a smaller intensity than eGFP bacteria sample. That means the bacteria scattered the blue light and therefore the spectruino detected more blue light from the source that there is in reality, the signal from eGFP.
We obtained two peaks with the dextran sample, as well as with the free GFP sample. We know that the dextran does not scatter so we can conclude that free FITC does not scatter as well. The bigger blue peak in dextran sample can be explained by the structural difference between dextran, which is a sugar polymer, and eGFP. So the light does not go through the same way.
To conclude, the free eGFP sample gives two distinct peaks, which demonstrate that there is no scattering anymore. There is more green fluorescent and less blue light without scattering and so the detector would detect the right amount of green fluorescent without stray blue light. That is very satisfying! Cell lysis is a good way to prevent light scattering. Unfortunately cell lysis has to take place after GFP expression and so after the mixing with the water sample.
We need to correct for the scattering in our prototype device, otherwise the assay will require an extra step of cell lysis.
Thanks to the several tests we performed we were able to see the blue light scattering and to quantify it. We observed two distincts peaks, a blue peak and a green one, with the free GFP sample and only one blue peak with eGFP bacteria sample. These results shows that the blue light hides a part of the green fluorescence, our signal we would like to detect from eGFP.
Since we obtained one blue peak, which masked our signal, this is what we want to avoid.
The cell lysis protocol allowed us to quantify this light scattering, and obtain a clear eGFP signal, but is not a solution that we can apply in the field.
To conclude, our spectra shows that the green fluorescence was hidden by the blue light. This will be a problem in the final value of arsenic concentration because the sensor detects more light that there is in reality.