The best that most of us can hope to achieve in physics is simply to misunderstand at a deeper level.
The aim of the Interlab Measurement Study over the last several years has been to eliminate differences in results of fluorescence experiments between laboratories using Green Fluorescent Protein (GFP). This year we have carried out the Fifth International Interlab Measurement Study which had the same general goal but specifically focused on reducing variability by normalizing fluorescence measurement to absolute cell count and colony-forming units (CFUs).
Both absorbance at 600 nm had to be measured during this Interlab Study as well as fluorescence. Both types of measurements were done with a machine that can measure both: the Tecan Spark 10M. After our first measurements of Abs600 we saw that the graphs with results had ‘[OD]’ on the y-axis, but after scrolling through the manual and comparing results to OD values of a spectrophotometer we concluded the results of the Tecan Spark to be absorbance values. The manual also stated that the Tecan Spark had pathlength correction included. However, it could not be turned off. Furthermore, the temperature settings were variable, just like the excitation and emission settings for fluorescence measurements. The recommended excitation and emission values could be set. Lastly, our instrument used top optics to read samples in the plates.
For this study we transformed E. coli DH5-alpha bacteria with eight different plasmids. These plasmids have a pSB1C3 backbone (depicted in grey in figure 1). One plasmid functions as negative control only containing a TetR repressible promotor. Another plasmid is used as positive control containing a single basepair mutant of a weak Anderson promotor (BBa_J23151), ribosome binding site (BBa_B0032), a GFP construct (BBa_E0040) and two terminators (BBa_B0010 and BBa_B0012).
The other six plasmids, from here on called devices, all have a similar lay-out. Firstly an Anderson promotor, varying in strength from device to device, followed by a ribosome binding site (BBa_B0034), a GFP construct (BBa_E0040) and again two terminators (BBa_B0010 and BBa_B0012).
The negative control (left) is a plasmid with just a TetR respressible promotor, the blue arrow, on a pSB1C3 backbone depicted in gray. The positive control (middle) is a plasmid with a single basepair mutated Anderson promotor, a ribosome binding site, a GFP construct and two terminators (not shown) with a pSB1C3 backbone depicted in red, purple, green and gray respectively. The test devices (right) had varying Anderson promotors, again shown as a red arrow, a ribosome binding site as shown in yellow, a GFP construct shown in green and two terminators, that are not shown, with a pSB1C3 backbone depicted in gray.
More detailed information on the Interlab constructs along with the strength of the construct’s Anderson promotor can be found here:
Cell measurement experiment
We transformed our bacteria with every plasmid separately and plated them. We made overnight cultures of two colonies of every transformation plate. The next day the cultures were diluted ten times, their Abs600 was measured and they were diluted another time to a target Abs600 of 0.02. The two colonies separately were plated in two 96 well-plates in quadruplicates to measure both Abs600 and fluorescence after 0 and 6 hours of culturing. The data was recorded in the provided Excel sheet by the iGEM headquarters.
To produce a standard curve for our Abs600 and fluorescence measurements we used Silica beads (microspheres) and a small molecule called fluorescein respectively.
Prior to this experiment we obtained a conversion factor to calculate OD600 values from Abs600 values doing a small experiment with LUDOX-CLX.
We performed the Colony Forming Unit (CFU) experiment where we made small cultures of the positive control and negative control overnight in duplo. We diluted these cultures eight times and plated them. After that we diluted the cultures further, from the used 96 well plate, to match an OD600 of 0.1. This was done for both duplicates of the positive and negative control three times for a total of twelve culture dilutions. In the end, one sample of each duplicate was diluted three times twenty times to a final dilution factor of 8.000 times, one sample of each duplicate was diluted another ten times on top of that to a final dilution of 80.000 times and the last one of each duplicate was diluted again ten times for the final dilution of 800.000. These dilutions were plated in such fashion that the dilution factor increased another ten times while plating. After incubating the plates over night the amount of colonies on each plate were counted and recorded in the provided Excel sheet.
As said before, the measurement machine we used showed ‘[OD]’ on the y-axis on the graphs with results. This caused us to think that the output values were OD600 value instead of Abs600. We did the entire Interlab Study (both experiments) while assuming we measured OD. We used an elaborate variant of the conversion factor calibration experiment to try and convert OD values to absorbance. However, when talking to the responsible person for absorbance and fluorescence machines at our lab about the pathlength correction, we discovered this ‘[OD]’ was actually the absorbance value including pathlength correction. We found out about it one week before the deadline. This caused us to have to redo the entire Interlab Study, both the cell measurement and CFU experiment. We contacted iGEM HQ about it and got extension before the actual deadline. However, we managed to redo the Interlab Study in the last week before the deadline to our surprise. Everything went extremely well without any more major problems.
Next to the major problem we had described above, we also came across some minor trouble. Transforming the bacteria sometimes did not work the first timeand we had to retry for test device 2 and test device 5. We even had to retry test device 2 a third time. For the cell measurement experiment we had some hard time exactly understanding what we had to do. However, after reading over the protocol a few times we managed to figure it out. Lastly, for the CFU experiment we did not come across any problems besides that we diluted wrongly in one series. But in the end there was no trouble caused by the protocols or the setup of the Interlab Study. We were unlucky to come across the confusing results presented by our Tecan Spark 10M.