Team:Newcastle/Proof/ElectricallyInducedLightBulb

Results for Heat induced 'lightbulb'

Initial testing our 'Lightbulb' constructs with exposure to different temperatures during growth

All of our genetic device constructs were placed in to the same pSB1C3 vector backbone and transformed in to BL21 (DE3) E.coli cells. We chose BL21 cells as these are optimized for protein expression, since they lack the ion protease involved in protein degradation (Rosano and Ceccarelli, 2014). The cells were grown up in liquid culture of LB broth (with chloramphenicol in strains containing a pSB1C3 housed construct) overnight at 37 °C. The following day, the bacterial cells were diluted down to an appropriate optical density in the region of 0.05 at 600 nm using LB broth with 0.034mg/ml chloramphenicol. We then pipetted 100 µl of the diluted cells into the corresponding wells, Diagram 1. The plate was laid out in this manner, with a border of wells containing only LB broth, to allow for any inaccuracies that may occur as the heating of the plate occurs from the outer to central wells. We tested both BBa_K1895000 and BBa_K1895006 lightbulb devices, with a positive control of device BBa_I20270 and a negative control of BL21 cells that does not produce GFP. In the figure below "sample" refers to the Interlab device 1, which we also included as a positive control.

Figure 1. The plate layout we used for these experiments.

The plate reader was then set at either 30°C, 37°C and 42 °C and measured for growth using OD₆₀₀ and fluorescence of GFP using 485 nm excitation wavelength and 520 nm emission wavelength with a gain of 250. The cells were left to grow for 18 hours with OD and fluorescence being measured every five minutes. In between measurements, the plate reader was programmed to shake, to ensure the cells didn’t clump together. The clumping of cells may have drastically effected the OD reading. The OD was also measured using a ‘well-scan’ function which takes 4 readings at different points within a well, resulting in a more accurate average reading.

Figure 2. Chart showing the normalised sfGFP expression over time for cultures grown at 30 °C in a BMG Fluostar plate reader. As can be seen, the signal detection is incredibly variable and unreliable. However, and initial hypothesis can be inferred that we are not seeing a significant sfGFP signal during the first 12 hours of growth, suggesting the system is in a low expression state at this temperature.

Figure 3. With the increase in temperature to 37 °C, we now begin to see both the P_htpG and P_dnaK regulated constructs beginning to produce a detectable sfGFP signal, distinct from the measured expression in our other tested strains. This was the first indication we had that sfGFP was being produced and that our lightbulb devices were inducible by heat. However caution must still be made with this data due to the high variability and noise in the readings.

Figure 4. At 42 °C we observed a different pattern of expression compared with at 37 °C. At 37 °C we saw an initial increase in sfGFP levels followed by a decline in expression after ~3 hrs followed by a more consistent signal for the remainder of growth. Here, at 42 °C, we saw an increasing sfGFP for the majority of the growth of the cultures. This data was more reliable and less variable than the previous two growth conditions, and adds further evidence that our lightbulbs constructs were both being expressed in response to increasing levels of heat. It is also clear that the response of P_htpG and P_dnaK promoters are broadly the same since - as observed at both 37 and 42 °C - the two line plots of these constructs overlay each other.

Further testing of lightbulb constructs with modified plate reader protocol

Although we could see a potential heating effect on sfGFP expression in our previous experiments, our data was highly variable and noisy. We examined what could be the causes of this. One the first observations made was the low normalised GFP RFU/OD signals. When looking at the data, we could see that there was not a high degree of difference in the recorded values between our tested devices and the positive and negative controls. In order to improve the signal measured, we increased the gain value of the fluorescence program on the plate reader to 1000, up from the initial value of 250.

We also looked at how often measurements were being taken by the plate reader. Our initial protocol defined that OD and fluorescence measurements should be recorded at 5 min intervals, with orbital shaking occurring in the time between measurements. As the measurement aspect of the protocol took up a high proportion of the 5 min cycle time, we felt that our cultures would not experience sufficient agitation. This would lead to clumping and impaired growth in the well, and may account for the noisy fluorescent signal. We therefore adjusted the cycle time to record only every 20 mins, over ~12 hours of growth.

Below we present the data for the normalised GFP RFU/OD600 data for our strains tested at our three growth conditions of 30, 37 and 42 ˚C respectively.

Figure 5. The normalised GFP RFU/OD600 when our strains were grown at 30 ˚C over 12 hrs. As can be seen, we do not see much of a variation in the fluorescence output of our two lightbulb constructs and the positive control.

Figure 6. The normalised GFP RFU/OD600 when our strains were grown at 37 ˚C over 12 hrs. At this temperature we begin to see an increase in the level of sfGFP production compared to the positive control, and which is higher than observed at 30 ˚C

Figure 7. The normalised GFP RFU/OD600 when our strains were grown at 42 ˚C over 12 hrs. At this temperature we really do begin to see a significant difference in sfGFP output when compared to both 30 and 37 ˚C. It is clear that a 42 ˚C temperature induces a high level of stress induction on the P_htpG and P_dnaK promoters, leading to a very high sfGFP expression and detected signal within our cells.

Figure 8. Summary of the data of all plate reader experiments performed with the modified protocol. It can be seen that temperature does indeed affect the activity of both the P_htpG and P_dnaK promoters, thus leading to increasing levels of sfGFP expression. We have annotated the three states ('OFF', 'LOW' and 'HIGH') at which our 'lightbulb' devices exist according to our three temperature conditions tested.

P_htpG lightbulb viewed using UV

As we have shown above that our lightbulbs are functioning correctly, we wanted to be able to prove that the fluorescent signal would be observable using equipment other than a plate reader or a fluorescent microscope. To do this, we decided to shine UV light on to one of our light bulb strains to show that this would allow the GFP signal to be observed 'by eye'. Overnight cultures of the P_htpG lightbulb strain were grown at 30 °C and then 'drawn' on to the surface of solid LB agar plates containing chloramphenicol. We then incubated the plates overnight in incubators set at each required temperature and observed the GFP signal the next day. The agar plates were placed on a UV transilluminator box within a dark room, and, using appropriate eye protection, we observed and photographed the agar plates with incident UV light. The result of this is shown below

Figure 9. The P_htpG lightbulb strain when grown on LB agar plates overnight at the indicated temperature. As can be seen, there is very little observable GFP visible on the plate grown at 30 °C, whereas the GFP signal is clearly brighter at 37 °C and increased further at 42 °C. As well as providing a very cool figure, this also allowed us to demonstrate our artistic flair! :-)

Exposure of E. coli containing 'lightbulb' constructs to an electrical current

Our take on the Stanford Paper Experiment:

Here we replicated the method carried out by a team of undergraduates at Stanford.

Again, all BioBricks were placed in the same pSB1C3 backbone and transformed BL21 E. coli cells. The cells were grown in liquid cultures 10 ml of LB broth with chloramphenicol overnight at 37 °C.

The next day, in a series of experiments, 5 ml of each overnight culture, along with 300 ml of LB broth (with chloramphenicol) was poured into a gel electrophoresis tank. The tanks were then run for 40 minutes at 28 V (400 mA). At each time point (0, 1, 3, 5, 7, 10, 13, 16, 19, 22, 25, 28, 31, 34, 37 and 40 mins), the electricity was switched off for roughly 30 seconds to allow us to obtain 9 samples of 100 μl. There were three samples taken from the positive end of the tank, three from the negative end, and three from the middle of the tank in order to measure the OD600 and fluorescence (in triplicate) on a plate reader. We measured the GFP fluorescence using 485 nm excitation wavelength and 538 nm emission wavelength.

Figure 10. Experimental set-up for carrying out the Stanford Experiment. A BioRad gel electrophoresis tank was filled with liquid LB broth and bacterial cell culture before being connected to a BioRad Power Pack and a current applied. The experiment was performed in a fume hood.

Figure 11. Graph showing the normalised GFP fluorescence reading divided by the OD600 value for our P_htpG construct over the 40 minutes where 28 V was applied

Figure 12. Graph showing the normalised GFP fluorescence reading divided by the OD600 value for our P_dnaK construct over the 40 minutes where 28 V was applied

Our Modified Stanford Experiment

After carrying out the initial experiment, we came up with some improvements which we believed would yield more efficient and reliable results.

We grew up fresh cultures for 4 hrs at 37 °C, to ensure that cells were in exponential growth. This means that the culture would not be under any form of stress from being in the stationary phase, and we would hopefully see more of a fluctuation in response when the cells were exposed to the electrical current.

We also felt that the 5 ml of culture in 300 ml of growth medium was too dilute a cell suspension, so we increased the concentration by adding 10 ml of culture to 250 ml of LB broth with chloramphenicol in the gel electrophoresis tank.

The new, more concentrated solution was added to the same gel tanks as used previosuly and this time we left the solution to run for an hour and a half, taking samples every 5 mins instead of every 2-3 mins as this gave us more time to accurately measure the samples.

We removed 1 ml of bacterial suspension from the positive end of the gel electrophoresis tank as we believed that the cells would migrate towards the positive end due to their negative charge and the overall flow of the tank. The OD600 was measured in a cuvette this time with an LB with chloramphenicol blank. When measuring fluorescence, 1 ml of culture was spun down at 14,000 rpm for 1 minute, 800 μl was removed and the pellet re-suspended in the remaining 200 μl liquid. Using this method increases the cell concentration, therefore amplifying the GFP signal and also any fluctuations. The fluorescence was again measured on the plate reader using 485 nm excitation wavelength and 538 nm emission wavelength.

Figure 13. Graph to show the normalised GFP fluorescence reading divided by the OD600 value for our P_htpG construct over the 90 minutes where voltage was applied with our modified protocol.

Figure 14. Graph to show the normalised GFP fluorescence reading divided by the OD600 value for our P_dnaK construct over the 90 minutes where voltage was applied with our modified protocol.

Figure 15. Growth Curve for our P_htpG lightbulb device over the 90 minutes where a voltage was applied to the growth medium

Figure 16 Growth Curve for our P_dnaK lightbulb device over the 90 minutes where a voltage was applied to the growth medium