Microfluidics and Microscopy1. Microfluidic chamber construction
We constructed microfluidic chambers of different channel sizes and dimensions, by laser etching on acrylic sheets using the Epilog LASER FUSION M2™ laser etching machine. The chambers were designed on Corel Draw X6 and fed to the machine, which then etched the channel out. Through holes and the outer rectangular cut was also made using the same machine. High resolution microscopy shows the etch marks when focused on the channels (figure 2). These marks were, however, found to not be inhibitory to the imaging of the cells in the channel in any way, and we were able to take images of the cells in a satisfactory manner.
2. Fluorescence imaging of negative and positive controls shows the working of the mRFP1 protein
We did the fluorescence imaging of the positive control (self-amplifying mRFP1/sfGFP under the control of the lux promoter) and the negative control (aiiA gene without reporter) for both the Danino oscillator and the iDanino oscillator using the Nikon Eclipse T-1 S fluorescence microscope. The positive control showed a high level of fluorescence in both the cases, with similar high fluorescence levels for both mRFP1 and sfGFP, confirming that the mRFP1 protein, as well as the LuxI and LuxR quorum sensing system that we cloned. The negative control showed negligible fluorescence (figure 3). All images were taken on the Nikon Eclipse Ti-S fluorescence microscope under 10X and 20X magnification.
3. Microfluidic testing shows synchronized oscillations in the iDanino circuit, cultured at 37oC
In the microfluidic device, E. coli cells were loaded from the cell port while keeping the media port at sufficiently higher pressure than the waste port below to prevent contamination. Cells were loaded into the cell traps by manually applying pressure pulses to the lines to induce a momentary flow change. The flow was then reversed and allowed for cells to receive fresh media with 0.075% Tween20 which prevented cells from adhering to the main channels and waste ports. The images of the cell trap in the center were taken over different time points (typically images were taken every 7-10 minutes), and the mean fluorescence of each of the images was analysed and plotted, which shows the oscillatory trend as predicted by Danino et al. (figure 4).
The images were analysed using ImageJ. Within each image, to test synchronization, we picked up 6-10 sample squares of similar dimensions, and measured the mean fluorescence within each square, and then took our the average and standard deviation and variance from all of these squares combined. The values of each individual square was very close to the moving average of all the squares combined.
4. Image Analysis results confirm synchronization
Image analysis using ImageJ showed a very small value of variance as compared to the mean oscillation value, which confirmed that the oscillations among the cells were synchronized, ie all the cells were in the same/similar phases at every time point. The results are shown in table 2.
For this, we took 18 images of the oscillating culture in the microfluidic chamber (18 different time points). Within each image, we picked up 10 square segments of equal area and calculated the mean fluorescence value as well as the standard deviation of the same.
Next, we fit these 18 points in increasing order of mean fluorescence rather than time, and used curve fitting using support vector regression and interpolated the data to generate the points in between, making a total of 100 points. The mean value of these 100 points was then plotted, as were the 10 squares over these 100 points individually (Figure 5). The graph showed that the individual squares did not differ too much from the mean value, confirming synchronization of these cells over the period of time taken.
Switching of temperature of the iDanino circuit causes oscillations to stop and toggles to a 0 or 1 state, depending on the reporter type used.
Cultures of the iDanino system were grown at 37oC until it reached a cell concentration (O.D.600) value of ~1, following which the temperature was switched to 30oC. Samples from this culture were taken at different times, and fluorescence microscopy for these samples was done at 10X and 20X magnification (figure 6). We saw the following trends in the images:
- If the reporter (mRFP1) was placed downstream of a Plux promoter, when the temperature was brought down to 30oC, the oscillations stopped and an increase in the level of mRFP1 was seen, until a final saturation level was reached. This corresponds to a constant “ON” state in the toggle switch.
- If the reporter (mRFP1) was placed downstream of a Plux-ʎ hybrid promoter, when the temperature was brought down to 30oC, the oscillations stopped similar to case (a) and an decrease in the level of mRFP1 was seen, until a near zero level of protein expression was reached. This corresponds to a constant “OFF” state in the toggle switch.
Hence, these results show that we have successfully designed and tested the iDanino in such a manner that it has a temperature dependent switch from a natural oscillator to a bistable toggle switch.