Team:NTNU Trondheim/InterLab Study
All iGEM teams were invited and encouraged to participate in InterLab Measurement Study. iGEM teams following Measurement Track were asked to participate in the study as well.
All teams were asked to measure the fluorescence of three different standardized devices. These devices were constructed by three different promoters and GFP genes. A given promoter was supposed to have a certain strength of GFP expression. The main purpose of these measurements is to compare results of different teams around the world, and to research equipment and methods that can be used. The suggested method was a standard plate reader. In addition to this, our team utilized a flow cytometer and quantitative image analysis by confocal microscopy. For the latter, the team developed a script that automatically analyzes the images - a tutorial is provided here!
- Device 1: J23101 + I13504 (B0034-E0040-B0015)
- Device 2: J23106 + I13504 (B0034-E0040-B0015)
- Device 3: J23117 + I13504 (B0034-E0040-B0015)
All devices had to be built in the pSB1C3 backbone with three biological replicas.
III.1) Plate Reader
The measurements were executed on 96 well plates with the plate reader. For the plate reader, we had two experimental settings:
- Measurement of Fluorescence by OD600 = 0.5 +/-10 %
- Measurement of Fluorescence and OD600 every 30 minutes over 6 hours
III.2) Flow Cytometry
The experiments conducted by means of flow cytometry were executed on a BD Accuri C6 sampler.
III.3) Quantitative Imaging By Confocal Microscopy
From the iGEM webpage, we got the information of the relative strength of the promoters:
- Device 1 with promotor J23101: 0.70
- Device 2 with promotor J23106: 0.47
- Device 3 with promotor J23117: 0.06
Hence, we expected our experiments to show a GFP expression of the devices that is similar to the relative strength of the promoters.
IV.1) Plate Reader
The plate reader results are different than expected (Figure 1). The mean fluorescence of all Device 1 replicas was lower than the one of Device 2 replicas. However, the deviation between the three biological replicas of Device 1 was remarkably high. Device 3, with a relative strength of only 0.06 is significantly lower than the other two devices and the positive control. The negative control shows low levels of fluorescence, as expected. The standard deviation of the controls is less than the standard deviation of the devices.
Figure 2 shows an extensive experiment: Measurement of the optical density at 600 nm as well as the fluorescence of GFP. The dashed lines display the OD, the solid lines the fluorescence. The values shown are the means of the technical replicates. The biological replicates of one device/control are each shown in shades of the same color. Even though the optical density increases similarly in all samples (except Device 2A), the increase in fluorescence over time differs between the devices and the controls. Device 3 and the negative control display low or no expression of GFP, as expected. The fluorescence of the positive control is clearly visible and shows again less variation between biological replicates. This is also the case for device 1 in this experiment. However, the biological replicates of device 2 show a higher variation in fluorescence level. The reason for the sample fluorescing the most could be, that its OD is also higher.
IV.2) Flow cytometry
Analysis of the GFP expression pattern by means of flow cytometry reveals more exactly the expected results (Figure 3). Even though the fluorescence of the biological replicates of the devices, especially Device 1, vary to a greater extent than the controls, there are significant differences in GFP expression patterns between the devices. As expected, device 1 shows the highest fluorescence, Device 2 a medium fluorescence, and Device 3 only very low fluorescence. The negative control does not show any fluorescence, as expected.
IV.3) Quantitative Imaging by Confocal Microscopy
Images for all the devices and controls were acquired by confocal microscopy (Leica SP5) for subsequent analyses using ImageJ.
A macro was programmed to select only the bacteria of interest automatically, without the need for drawing and selection tools. The macro labels each bacterium individually with a number and calculates the average fluorescence intensity for each bacterium, as show in Figure 4. We have made the tool available for download from Quantitative Analysis Tool page accompanied with a detailed tutorial and sample images.
The bright field image was analysed to find the bacteria of interest, and the fluorescence image was analysed to find the total fluorescence per bacterium. To find fluorescence per cell, the sum of all the individual fluorescence intensities was divided by the number of identified cells.
For bright field image segmentation, we used background subtraction with rolling factor equal to 50, automatic threshold detection, particle analysis with sizes larger than 2E-8 cm^2 and circularity of 0.1 to 0.7. The images were 0.2 cm by 0.2 cm (2048 px by 2048 px), 12-bit bit-depth, and containing both fluorescence and bright field images. A script has been written to analyze all the images automatically without any user input parameters.
Figure 5 shows the result from the image analyses, which compares satisfactorily with the results obtained by flow cytometry and the plate reader.