The Newcastle iGEM team aimed to develop a suite of biosensors using fluorescence proteins as a reporter. Our biosensors may integrate a combination of fluorescence proteins and chromoproteins in future development. As a potential diagnostic tool, it was highly important for our project to measure fluorescence levels accurately. We conducted a series of experiments to increase measurement reliability of fluorescent proteins and make measurements more comparable between researchers in Parkinson's Disease and in NHS lab settings should this be developed for diagnostics in the future. Our goal was to create a standard curve of protein fluorescence levels to correlate to cellular fluorescence levels. Additionally, we found that plate types i.e a 384 well plate vs a 96 well plate affected the standard curve for fluorescence intensity. We also investigated if fluorescence intensity can be measured accurately when fluorescence proteins and other reporters such as chromoproteins are combined.
Promoter strength is commonly reported as a fluorescence value relative to the optical density at OD600 nm. Arbitrary values are not reproducible and are not suitable for an engineering field like synthetic biology. Whilst relative values of fluorescence are more suitable than arbitrary fluorescence, discrepancies in relative values still limit comparison of results and lead to incorrect conclusions on promoter strength. Our goal was to calculate absolute values, the gold standard for measuring fluorescence to reduce variability in reproducibility. Standardised absolute values can greatly improve data quality by allowing a direct measurement of results against a set of expected standard values .
Characterising the strength of the Bba_J23100 Promoter
Newcastle 2019 have characterised the strength of the constitutive promoter BBa_J23100. The activity of BBa_J23100 was determined using eGFP as a reporter. A standard curve of the fluorescence level of eGFP was determined using eGFP with a His6-tag under the BBa_J23100 promoter. This allowed us to purify the eGFP protein to determine fluorescence at protein level. Our goal was to calculate absolute values of eGFP protein concentration in cells.
eGFP and mCherry Protein Purification
Transformed Escherichia coli Top10 cells with His6-Tagged eGFP and His6-Tagged mCherry were grown overnight at 37 °C at 200 rpm. Cells were harvested and the cell pellets were lysed for protein purification (Figure 1). eGFP and mCherry proteins were manually purified using a HisTrap HP 5 mL column (GE Healthcare, UK) (Figure 2). Proteins were washed in HisA buffer (50 mM Tris-HCl, 500 mM NaCl, 50 mM imidazole) and the purified protein was eluted in HisB buffer (50 mM Tris-HCl, 500 mM NaCl, 500 mM imidazole) at different concentrations of imidazole (100, 200, 300, 400 mM).
Figure 1. Cell Lysate of eGFP and mCherry. A) Cell Lysate in normal lighting B) UV imaging of eGFP and mCherry cell lysate. Cells were pelleted at 4200 rpm for 10 mins and the cell pellet was resuspended in 1 x PBS buffer. Cell pellet was sonicated in HisA buffer and cells were clarified by Ultrafuge for 30 mins at 45,000 rpm.
Figure 2. HisTrap HP 5 mL column (GE Healthcare, UK) with His6-tagged mCherry (left) and His6-tagged eGFP (right)
SDS-PAGE of purified protein analysis
Sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) analysis of purified eGFP of these fractions showed that there was high contamination in the wash step and flow through (Figure 3). However, there was a minimal amount eGFP eluted during binding due to the presence of a band at approximately 28 kDa in both lanes. With increasing imidazole concentrations, a higher yield of protein was eluted in 400 mM imidazole. Although, with the increase of imidazole, there was evidence of contamination due to the presence of bands at approximately 55 kDa and 15 kDa, suggesting the HisTrap column may not be sufficiently cleaned. Due to the high yield of proteins purified, a secondary purification was deemed not necessary.
Analysis of the purified mCherry showed that there was also high contamination in the wash step and flow through. With increasing imidazole concentrations, a higher yield of protein was eluted in 300 mM imidazole. However, two bands were present at approximately 29 kDa and 19 kDa. The predicted molecular weight of mCherry was 27.98 kDa but no band was present at the predicted size (Figure 3). The band at approximately 29 kDA is likely to be mCherry but this may be due to issues with purifying red fluorescent proteins. It was clear mCherry was purified as the cell lysate and His-trap was visually a bright purple (Figure 1 and 2). Due to these findings and the limitation in time, experimental work on mCherry was not continued and a standard curve was not performed for mCherry. Purified proteins were then stored in Gel Filtration Buffer (50 mM Tris-HCl, 150 mM NaCl).
Figure 3. 12% SDS-PAGE showing resolve of Marker (ThermoFisher PageRulerTM Unstained Protein Ladder), wash through, flow through and elution fractions at 100, 200, 300 and 400 mM imidazole of His6-Tag eGFP followed by wash through, flow through and elution fractions at 100, 200, 300 and 400 mM imidazole of His6-Tag mCherry. Stained using Coomassie Brilliant Blue. His6-Tag eGFP predicted to be 28.179 kDa
The protein concentration of the purified eGFP protein was determined using a ThermoFisher BCA assay kit. The concentration of the purified eGFP was 5.97 mg/ml as determined using the standard curve of bovine serum albumin (BSA). The purified protein was diluted at a concentration range from 3.125 to 100 µM. However, due to an error in the calculation of the protein concentration initially, the correct concentration range was 2.0625 to 66 µM. These serial dilutions were plated in a 384 well plate and 96 well plates to measure the fluorescence level to produce a standard curve to quantify the amount of protein present in cells.
Standard Curve of eGFP
The fluorescence level was measured in a ThermoFisher Varioskan LUX Multimode Microplate reader using black welled 384 well plate and black welled 96 well plates to observe variation in plate type. Black welled plates were chosen for this experiment to reduce errors in measuring fluorescence from surrounding wells. An additional 96 well plate was prepared and measured in a BioTek Synergy H1 Microplate Reader to observe variation in measurement equipment. Firstly, as a preliminary experiment, the standard curve of the diluted eGFP was determined in a 384 well plate. The results showed there was no linear regression but instead a hyperbola (Figure 4). This maybe due to the emission of eGFP reaching threshold at higher concentrations. It was clear that above 24.75 µM, the fluorescence level deviated from the linear trend. This suggests higher concentrations above 24.75 µM are not quantifiable as the fluorescence level plateaus and fluorescence above this threshold will not be distinguishable.
Figure 4. Standard Curve of the mean fluorescence level of purified eGFP protein diluted to concentrations from 0 to 66 µM. Excitation wavelength = 488 nm. Emission wavelength = 507 nm. Measured in Greiner black welled 384 well plate in a ThermoFisher Varioskan Lux Microplate Reader
Based from the results of the original dilution range, the purified eGFP was diluted further in lower concentrations up to 24.75 µM to widen the range of dilutions in the lower dilutions range. These serial dilutions were measured in 96 well plates in both microplate readers and a 384 well plate in the ThermoFisher Varioskan. Similarly, we did not observe a linear regression as the fluorescence level plateaued at the higher concentrations in a 96 well plate measured in the ThermoFisher Varioskan plate reader (Figure 5A). However, we observed a linear regression in the 384 well plate. The results showed that there was a slight deviation between the readings from a 384 well plate compared to a 96 well plate (Figure 5B) suggesting variations in readings dependent on the plate type.
Figure 5. Standard Curve of the mean fluorescence level of purified eGFP protein diluted to concentrations from 0 to 24.75 µM. A) Measured in Greiner black welled 96 well plate B) Measured in Greiner black welled 384 well plate. Excitation wavelength = 488 nm. Emission wavelength = 507 nm. Fluorescence readings measured in a ThermoFisher Varioskan Lux Microplate Reader. R2 values are displayed.
The same parameters were set in the BioTek Synergy H1 Microplate Reader but the gain was set at 100. The standard curve produced shows the values at 100 gain (Figure 6). These results showed a linear regression of the fluorescence level in the increasing serial dilutions.
Figure 6. Standard Curve of the mean fluorescence level of purified eGFP protein diluted to concentrations from 0 to 24.75 µM. Measured in Greiner black welled 96 well plate. Excitation wavelength = 488 nm. Emission wavelength = 507 nm. Fluorescence readings measured in a BioTek Synergy H1 MicroPlate Reader. R2 value is displayed.
It was concluded that the standard curve should be plotted from the results in a 96 well plate because the fluorescence intensity and OD600 of the cell culture will be measured in 96 well plates to allow for a higher surface area for growth. This would allow our results to be consistent and minimise technical error from varying plates as observed previously (Figure 5). We analysed the data from the 96 well plate measured in the ThermoFisher VarioSkan Lux Plate Reader and plotted a standard curve of the dilutions up to 8.25 µM (Figure 7). This produced a linear regression equation of Y=302.4x + 72.79 and an R2 value of 0.9951.
Figure 7. Standard Curve of the mean fluorescence level of purified eGFP protein diluted to concentrations from 0 to 8.25 µM. Linear regression equation of Y = 302.4x + 72.79 and an R2 value of 0.9951. Excitation wavelength = 488 nm. Emission wavelength = 507 nm. Measured in Greiner black welled 96 well plate in a ThermoFisher Varioskan Lux Microplate Reader
Measuring optical density and fluorescence intensity
Assembled eGFP under the J23100 promoter in E. coli Top10 cells were cultured overnight (figure of plate). The overnight culture was re-inoculated and grown to OD600 of 0.3. Once an OD600 of 0.3 was reached, the cell culture was diluted to OD600 0.05 and the cell culture was loaded into a 96 well plate with 100 µL of culture per well. The OD600 and fluorescence intensity was determined in the ThermoFisher Varioskan plate reader for 24 hours. The growth curve showed an exponential phase and death phase however, the lag and stationary phase was not observed (Figure 8A). This was unexpected as we could expect the curve to be similar to that of a typical E. coli growth curve. This may be due to an issue with the plate reader and cell clumping at the lower optical densities. However, the fluorescence level reflected a similar pattern to a typical E. coli growth curve (Figure 8B).
Figure 8. A) The optical density (OD600) of E. coli Top10 cells assembled with eGFP over 24 hours B) The fluorescence intensity of E. coli Top10 cells assembled with eGFP over 24 hours. Excitation wavelength = 488 nm. Emission wavelength = 507 nm. Measured in Greiner black welled 96 well plate in a ThermoFisher Varioskan Lux Microplate Reader
The OD600 and fluorescence intensity was also determined in the BioTek Plate Reader, at a gain of 100. Whilst the OD600 showed the growth of E. coli cells, the fluorescence intensity reading was unsuccessful due to the limitation of the gain at 100 and fluorescence intensity beyond one hour was not read.
Using the linear regression equation generated from the standard curve, we converted the fluorescence values into protein concentration in mg/ml (Figure 9). The graph shows that the maximum protein concentration is approximately 3mg/ml after 10 hours. The protein concentration (mg/ml) showed a similar pattern as observed in the fluorescence level (Figure 8B).
Figure 9. Protein concentration in mg/ml of eGFP assembled in E. coli Top10 cells over 24 hours. Protein concentration was determined by converting fluorescence figures in Figure 8 using the equation Y = 302.4x + 72.79.
In addition to growing our transformed cells, E. coli Top10 cells were grown to measure the OD600 and CFU of untransformed cells. This allows us to calculate the Colony Forming Unit (CFU) per mL of our cells expressing eGFP (Figure 10). The number of viable cells in our culture can be measured over the 24 hour period. Using these values combined with the protein concentration in mg/ml, we are able to calculate the mg of protein per CFU (mg/CFU) by dividing the protein concentration (mg/ml) over the CFU/ml value at the equivalent time.
Figure 10. Colony forming units (CFU) per mL (CFU/mL) of eGFP assembled in E. coli Top10 cells over 24 hours.
Whilst mg/CFU is much better than a relative fluorescence value, our goal was to calculate absolute values of eGFP protein concentration in cells - the Number of Proteins per Cell. We divided the mg/CFU values by the molecular weight (MW) of the protein in mg. The MW of eGFP is 28.179 kDa, which is 28179•1.66 x 10-21 in mg. This resulted in figure 11 showing the Number of eGFP proteins per cell over 24 hours. The results suggest that cells express up to approximately 8 x 10^6 of eGFP over the 24 hour period. However, there was a spurious increase in the number of protein per cell, this may be due to the issue in OD600 readings as seen in figure 8. The OD600 reading should be repeated and measured in different plate readers to observe the variability in measurement equipment, should this be the issue we observed with our growth curves.
Figure 11. The Number of eGFP proteins per cell over 24 hours. E. coli Top10 cells were assembled with eGFP.
In conclusion, we achieved the goal to calculate absolute values to measure promoter strength of BBa_J23100 using eGFP fluorescent expression as a reporter. A standard curve of the number of proteins per cell over time was achieved. The results showed that protein expression was high per cell reflecting the strength of the promoter. We were able to accurately measure proteins for which there are no good external standards. The method we applied to achieve these results put into practice the expectations of absolute values as a standard measure .
The experiments conducted also took into account technical variances in measuring equipment to minimise variability in standardised values. However, this experiment should be repeated with increased automation to further reduce variability in technical errors to allow a comparison of our absolute values to automated absolute results. Whilst we took into account reproducibility of fluorescent expression to measure the strength of a promoter by measuring absolute fluorescence, further efforts are necessary to produce the highest reproducible results. We encountered issues with measuring the optical density of cells whereby we saw unexpected growth curves. A quantified measurement of cell population has not been heavily investigated but such efforts would allow for further reduction in variability .
Mixing reporter proteins experiment
sfGFP and mCherry
We obtained sfGFP (BBa_K515105) and mCherry (BBa_J04450) from iGEM distribution kits. The resuspended DNA was transformed into E. coli DH5alpha cells as shown in figure 12. The successfully transformed cells were inoculated in LB with the appropriate antibiotic to prepare overnight cultures at identical conditions of 37 °C shaking at 200 rpm for 16 hours. The cell cultures were then inoculated in 10 ml of LB to grow cells up to an OD600 value of 0.6. Half of the culture was diluted to OD600 0.3 and the individual fluorescence levels were measured at: sfGFP - excitation 480nm and emission 507 and mCherry – excitation 580, emission 610 in a BioTek Synergy H1 Microplate Reader with a gain of 100.
Figure 12. UV imaging of successfully transformed cells expressing A) sfGFP in pSB1AT3 (BBa_K515105) B) mCherry (BBa_J04450) in pSB1AT3 on LB + Ampicillin agar plates C) fwYellow (BBa_K1033907) in pSB1C3 and cjBlue (BBa_K864404) in pSB1C3 on LB + Chloramphenicol agar plates
The sfGFP and mCherry cultures at OD600 0.6 were mixed together in equal amounts and the fluorescence level of the combined mixture was measured. Theoretically, the OD600 of each fluorescence protein in the mixture will be 0.3. By comparing the individual fluorescence level at OD600 of 0.3 to the fluorescence level of the equivalent fluorescence protein minus the opposing fluorescent protein, the effect of mixing fluorescent proteins can be observed. It is hypothesised that the fluorescence intensity will be quenched when mixed with another fluorescent protein or chromoprotein.
The emission spectra for sfGFP ranges from 469 nm to 628 nm and the emission spectra for mCherry ranges from 551 nm to 800 nm. There is an overlap of 77 nm in the emission spectra of the two fluorescent proteins. This small overlap suggests that there will not be much effect on the fluorescence intensity of the individual fluorescent proteins as the overlap will be at an emission with a low percentage emission.
The results of mixing sfGFP and mCherry at sfGFP excitation and emission (Figure 13) showed that there was no significant difference in fluorescence intensity of sfGFP when mixed with mCherry (Figure 13A). The mean fluorescence of sfGFP individually was 168008 ± 2503.4 compared to the mean fluorescence of sfGFP mixed with mCherry at 172678.2 ± 6826.9. Similarly, at mCherry excitation and emission, similar results were observed and there was no significant difference in fluorescence intensity (Figure 13B). The mean fluorescence of mCherry individually was 4524.8 ± 5.1 compared to the mean fluorescence of mCherry mixed with sfGFP at 4693.6 ± 162.5.
Figure 13. A) Bar chart showing the mean fluorescence intensity of E. coli DH5alpha cells expressing sfGFP individually at an optical density of 0.3 with standard deviations compared to the mean fluorescence intensity when mixed with cells expressing the fluorescent protein mCherry Excitation 480 nm, emission 507. B) Bar chart showing the mean fluorescence intensity of E. coli DH5alpha cells expressing mCherry individually at an optical density of 0.3 with standard deviations compared to the mean fluorescence intensity when mixed with cells expressing the fluorescent protein sfGFP. Excitation 580 nm, emission 610. Fluorescence intensity measured in BioTek Synergy H1 MicroPlate Reader
sfGFP with chromoproteins fwYellow and cjBlue
We also wanted to investigate the effect of chromoproteins on the fluorescence levels of the fluorescent proteins and conducted the same experiment. fwYellow (BBa_K1033907) and cjBlue (BBa_K864404) in pSB1C3 plasmids, were also transformed into E. coli DH5alpha cells (Figure 12).
The absorption wavelength maximum of cjBlue is approximately 610 nm . The maximum absorption wavelength of fwYellow is 523 nm . Given that the emission spectra for sfGFP ranges from 469 nm to 628 nm and the emission spectra for mCherry ranges from 551 nm to 800 nm, we would expect the mixing of fluorescent proteins with chromoproteins to decrease the fluorescence intensity of the fluorescent protein.
The culture of sfGFP expressing cells was mixed cultures of cells expressing fwYellow and cjBlue separately and the signal intensity of each mixture was measured (Figure 14). When sfGFP was mixed with fwYellow, there was no significant difference in fluorescence intensity. Although, when sfGFP was mixed with cjBlue, there was a slight increase in fluorescence intensity but this was not significantly increased.
Figure 14. Bar chart showing the mean fluorescence intensity of E. coli DH5alpha cells expressing sfGFP individually at an optical density of 0.4 with standard deviations compared to the mean fluorescence intensity when mixed with cells expressing the chromoproteins fwYellow and cjBlue. Excitation 480 nm, emission 507. Fluorescence intensity measured in BioTek Synergy H1 MicroPlate Reader
mCherry with chromoproteins fwYellow and cjBlue
The mCherry cell culture was also mixed with fwYellow and cjBlue cultures and the signal intensity was measured (Figure 15). When mCherry was mixed with fwYellow, there was a slight decrease in the fluorescence intensity of mCherry. Similarly, when mCherry was mixed with cjBlue, a slight decrease in the fluorescence intensity of mCherry was observed. Whilst there was a slight decrease in the fluorescence intensity of mCherry when mixed with the chromoproteins fwYellow and cjBlue, the difference was insignificant.
Figure 15. Bar chart showing the mean fluorescence intensity of E. coli DH5alpha cells expression mCherry individually at an optical density of 0.4 with standard deviations compared to the mean fluorescence intensity when mixed with cells expressing the chromoproteins fwYellow and cjBlue. Excitation 580 nm, emission 610. Fluorescence intensity measured in BioTek Synergy H1 MicroPlate Reader
Mixing of sfGFP and mCherry did not significantly alter the fluorescence intensity of each individual fluorescent protein. When fluorescent proteins are mixed with chromoproteins, this also did not affect the fluorescence intensity of the fluorescent proteins. It can be concluded that despite the mixing of fluorescent proteins, when measuring fluorescence intensity, it can be confidently assumed that the results for each fluorescent protein is accurate.
- Content: Karen Fung
- Figures: Karen Fung
- Supervisory advice: Dr Jon Marles-Wright, Dr Thomas Howard, Jasmine Bird, Bradley Brown, Dr Cecilia Piergentili, Andrea Giachino
- Proofreading: Connor Trotter, Bradley Brown, Jasmine Bird, Karen Fung
- Newcastle iGEM 2018. [Cited 20 Oct 2019] Available at: https://2018.igem.org/Team:Newcastle/Measurement
- Beal J, Haddock-Angelli T, Gershater M, de Mora K, Liarazo M, Hollenhorst J, Rettberg R, I GEMISC. Reproducbility of Fluorescent Expression from Engineered Biological Constructs in E. coli. PLOS one. 2016;11(3): e0150182 DOI:10.1371/journal.pone.0150182.
- Beal J, Haddock-Angelli T, Baldwin G, Gershater M, Dwijayanti A, Storch M, de Mora K, Lizarazo M, Rettberg R, I GEMISC. Quantification of bacterial fluorescence using independent calibrants. PLOS one. 2018;13(6): e0199432. https://doi.org/10.1371/journal.pone.0199432.
- Chan MC, Karasawa S, Mizuno H, Bosanac I,Ho D, Privé GG, Miyawaki A and Ikura M. Structural characterization of a blue chromoprotein and its yellow mutant from the Sea Anemone Cnidopus Japonicus 2006;281:37813-37819
- NWU-China iGEM 2019. [Cited 20 Oct 2019]. Available at: https://2019.igem.org/Team:NWU-China/Characterization