Interlab Study: The Results
Rationale and Aims
The interlab study investigates the replicability of data by providing all participating iGEM teams with protocols for measuring the fluorescence output of a set of genetic devices that have been designed to provide varying levels of Green Fluorescent Protein (GFP) expression. Team Newcastle 2017 followed the Interlab plate reader protocol.
Aim: Carry out the InterLab Study as specified by iGEM.
The iGEM InterLab Study is the currently the largest interlaboratory study in synthetic biology (Beal et al., 2016). The overall goal of the InterLab is to create a reproducible fluorescence assay protocol, of which anyone can follow and produce similar results to another. The published data of the 14/15 InterLab studies identified areas for improvement, with this year's protocol leaving little room for misinterpration/error.
InterLab Measurement Kit
This year, teams were provided with the InterLab Measurement Kit containing fluorescein and LUDOX stocks, along with the devices stored in the Distribution kits. The devices are:
- A positive control: BBa_I20270
- A negative control: BBa_R0040
- Test Device 1: BBa_J364000 (J23101.BCD2.E0040.B0015)
- Test Device 2: BBa_J364001 (J23106.BCD2.E0040.B0015)
- Test Device 3: BBa_J364002 (J23117.BCD2.E0040.B0015)
- Test Device 4: BBa_J364003 (J23101+I13504)
- Test Device 5: BBa_J364004 (J23106+I13504)
- Test Device 6: BBa_J364005 (J23117+I13504)
OD600 Reference Point
LUDOX-HS40 was used as a single point reference to obtain a radiometric conversion factor to convert absorbance data into a standard OD600 measurement. The Reference OD600 divided by the Abs600 from four replicates of LUDOX was used to obtain a correction factor for use against the cell based assays.
Fluorescein Standard Curve
A dilution series of fluorescein in four replicates was prepared and measured in the plate reader to obtain a standard curve of fluorescence for fluorescein concentration. This was used to correct cell-based readings to an equivalent fluorescein concentration, and to then convert this to a GFP concentration.
Competent E. coli DH5α cells were transformed with each of the devices and plated onto LB+Chl agar, and two colonies from each transformation plate were grown overnight in 10 ml LB + Chl in 50 ml Falcon tubes. Protocols for making competent cells and cell transformation can be found here (link to protocol). Fluorescence was measured as specified in the InterLab protocol on a Synergy H1 plate reader.
Raw results can be found here
Figure 1 InterLab results – (A) Growth, (B) fluorescence and (C) calculated µM fluorescein/OD600 of InterLab devices grown at 37°C in LB for 6 h. Measurements from 4 individual plate readings at 0, 2, 4 and 6 h.
Bacterial growth - We found that the negative control, Tests 3, 5 and 6 grew the most in the 6 h experiment, with the positive control and Test 2 growing less so. Test 1 had the least growth.
Fluorescence - Test 2 showed the most fluorescence after 6 h, with Test 4 showing slightly less, followed closely by the positive control, then Tests 1 and 5. Tests 3 and 6 showed very little fluorescence.
FL/OD - Test 1 gave a significantly higher overall uM fluorescein/OD600 ratio compared to all other devices with values of 0.91 and 0.94 from the separate isolates, compared to a range of <0.01 and 0.18 from all other devices.
Variability - As part of the Interlab study, we analysed two separate bacterial transformants in quadruplicate. This allows an examination of the variability between replicates and between colonies. Between replicates, variation was minimal. However, between colonies containing the same device we saw a degree of variability in growth and fluorescence, thus affecting FL:OD. Figure 2 shows high levels of variability between max OD and Abs measurements of Tests 2 and 4. Fluorescence and subsequently FL:OD are directly affected by growth, but factors including inaccuracies in growth set up and the user taking the samples could have also affected these. As plasmid copy number will not be equal, this will also manifest in readings over time.
Figure 2 Box and whisker plots of (A) max OD and (B) Absorbance values for fluorescence. Maximum values were obtained from all colonies and replicates for each device.
Conclusions and Future Work
After carrying out the InterLab Study we decided to take a high, medium and low expressing device and analyse their sensitivity to changes in environmental conditions. We also looked into how automation could reduce variation between results.
Beal, J. et al. (2016) ‘Reproducibility of Fluorescent Expression from Engineered Biological Constructs in E. coli’, PLOS ONE. Edited by D. D. Jones. ACM, 11(3), p. e0150182. doi: 10.1371/journal.pone.0150182.
Interlab Study Improvements: The Results
Interlab Devices in Different Contexts
We looked at the impact of temperature, pH and media on how high, medium and low expressing devices performed. Tests 2, 5 and 6 were selected from the Interlab Study after initial results were analysed. Single colonies of each transformant were selected and grown in LB+Chl, then washed and diluted to an OD600 of 0.05 in 100 ul media. Transformants were analysed in quadruplicate, with bacteria and media pipetted onto the plate using a pipetting robot for maximum accuracy. Once set up, the plate was incubated at specified temperatures in a plate reader with double orbital shaking, taking readings every 10 min for 24 h.
Growth in Test 2 was affected differently in LB and SOC (Fig 3). In LB, there appears to be a slight increase in max OD reached as temperature increases, however at an alkaline pH the device appears unable to grow well in higher temperatures. In SOC, there is no distinct pattern in how temperature affects growth, but max growth is improved as pH is increased, suggesting pH and temperature have a combined effect.
Figure 3 Max OD reached by Test 2 in (A) LB and (B) SOC media over 24 h at 31C, 37C and 43C.
Fluorescence levels were affected differently at different pH levels as temperature was increased (Figure 4). 37C appears to be optimum temperature for both media, as there are decreases between 37C and 43C at all pH levels apart from pH 7.20 in LB media. However, we can see a dramatic increase in overall max fluorescence levels in both LB and SOC when pH is increased to an alkaline level.
Figure 4 Max fluorescence reached by Test 2 in (A) LB and (B) SOC media over 24 h at 31C, 37C and 43C.
For the overall FL:OD, in LB media 37C was the optimum temperature, but in SOC a higher temperature appeared to be more favourable. In both media, the exception was seen at an alkaline pH, where FL:OD increases with temperature in LB, and 37C was optimum in SOC (Figure 5).
Figure 5 Max FL:OD reached by Test 2 in (A) LB and (B) SOC media over 24 h at 31C, 37C and 43C.
Growth was affected significantly by changes in pH, which can be seen in figure 6. In SOC media, a distinct increase is seen in max OD as pH increases. In LB media the opposite is seen, where as pH increases, max growth decreases.
Figure 6 Max OD reached by Test 2 over 24 h in (A) LB and (B) SOC media adjusted to varying pH levels.
Fluorescence was affected by pH particularly in SOC where an increase was seen with an increase in pH; In LB there was less of a pattern seen; increases can be seen at 31C and 37C with increases in pH, however at 43C the max FL is reduced dramatically. The combination of high temperature/pH may have been toxic to the cells.
Figure 7 Max FL reached by Test 2 over 24 h in (A) LB and (B) SOC media adjusted to varying pH levels
Despite clear correlations in both growth and fluorescence, there appears to be no distinct pattern in the increase of pH and the overall FL:OD (Figure 8).
Figure 8 Max FL:OD reached by Test 2 over 24 h in (A) LB and (B) SOC media adjusted to varying pH levels
Overall, it is clear pH and temperature have an impact on the growth and fluorescence of the devices, and this raises the importance of maintaining conditions specified in InterLab instructions. It is difficult to draw definite conclusions without more thorough experimentation. It should also be noted that whilst each replicate of isolates was added to the plate using a robot for maximum accuracy, there was still a degree of variability. These variances will have increased over time, and the InterLab method of taking samples at specific time points may be more laborious, but will yield a more accurate result of overall culture behaviour.
Standard Assembly Methods
We investigated whether different assembly methods influenced performance and the expression of downstream products. This was done by designing and assembling the same GFP reporter device (promoter (BBa_B0032) and RBS (BBa_B0034) driving expression of a GFP (BBa_I746916) using the iGEM standards RFC10 and MoClo-based Phytobricks and the non-standard, scarless Gibson assembly method. The Biobrick method was unsuccessful in assembly the desired devices, and we therefore compared the Gibson (Figure 9) and Phytobrick (Figure 10) assembled constructs.
During growth in LB media it was found that the presence of sequence ‘scars’ introduced by Phytobrick assembly impacted strongly on the device operation; the constructs containing sequence ‘scars’ had a significantly lower expression of the reporter GFP gene compared to the constructs containing no scars.
Figure 9 Max FL:OD reached by Test 2 over 24 h in (A) LB and (B) SOC media adjusted to varying pH levelDiagrammatic representation of two DNA parts being annealed and inserted into a linearised vector backbone using the Gibson DNA assembly method. Each DNA part and the vector backbone were synthesised to overlap by about 30 bp. An exonuclease, ligase and polymerase are then used to generate complementary single strands at the ends of each DNA fragment, and seamlessly join them together in a single isothermal reaction.
Figure 10 Example of a type IIS restriction endonuclease used in Phytobrick DNA assembly. The BsaI restriction endonuclease cuts the DNA outside of the restriction site producing DNA fragments with 4 bp overhangs. The 4 bp overhangs are used to join to overhangs on another DNA fragment through complementary base pairing.
Different assembly methods have a significant impact on performance of the genetic devices.These findings clearly demonstrate the importance of teams (and the wider scientific community) accurately reporting how devices are made, and the final sequence of the device. The presence or absence of assembly scars can impact on function.