Team:NUS Singapore/Measurement
Overview
This year, we at NUS have developed a novel growth control system consisting of a growth switch and a tunable growth knob to control cell growth and protein production, suitable for use in fundamental studies and real-world applications. To assist users in the implementation of our systems, we have developed a comprehensive workflow and rigorous measurement protocols to ensure the robustness and reproducibility of our system.
Our growth switch mainly works by inhibiting cellular growth to promote the accumulation of cellular reserves, resulting in the preservation of the functional viability of cells for long-term usage. As such, to fully characterize the system, there is a need to come up with a long-term workflow to detail the entire process from inducing growth arrest to quantifying protein production.
Here, we are proud to present our experimental and data analysis workflow tailored for the implementation of our growth switch. In this workflow, we mainly employ microplate readers for characterization, although alternative approaches such as using flow cytometry or microscopy could be adopted as well.
Experimental workflow
We characterized all our cells in the MG1655 strain of E. coli. To characterize the growth switch, consisting of toxin-antitoxin systems, we grew up cells co-transformed with both toxin and antitoxin plasmids in LB containing appropriate antibiotics. The next day, we refreshed the overnight culture to an OD600 of 0.1 in 10mL LB media, before inducing the expression of toxin gene by adding 1mM IPTG one hour later. Afterwards, the cells were grown in a 37°C shaking incubator - commencing day 0 of our experiment. For every five days, we would aliquot the cells into a 96-well microplate and add 0.0133M arabinose into each well to induce the expression of LuxCDABE and antitoxins. After the addition of arabinose, we measured the OD600 and luminescence of each well in a microplate reader for over 12hrs, with continuous double orbital shaking throughout (Fig. 1).
For more information about our protocol, do check out our experimental page!
Fig. 1: Pictorial representation of the long-term characterization workflow performed to measure luminescence and OD600 in normal-growing cells and growth-arrested cells.
Samples were plated as triplicates. We included three blank replicates to normalize the background luminescence of LB media and wells without any arabinose added as negative controls. To compare the performance of our system, we used the same set up for cells that similarly contained the two plasmids, but without the genes activated. These thus allowed us to understand whether our system could improve the performance of normal wild-type cells. By including these controls, we account for different variables that might affect our interpretations and analysis. An example of a typical plate layout used in our workflow is shown in Fig. 2 below.
Fig. 2: A 96-well microplate layout during luminescence measurement with a microplate reader. For both normal-growing control MG1655 and growth-arrested MG1655, we included a negative control treatment (without arabinose added) to study the effect of toxin on growth and protein production in growth-arrested MG1655 while using normal-growing control MG1655 to compare against. For the treatment group, we added 0.0133M arabinose into each well containing the cells, on the day of protein induction. A LB blank was included to normalize the background luminescence present emitted by the media itself.
To ensure the precision of our liquid handling, we utilized the Opentrons OT-2 robot, generously awarded to us by Opentrons this year. We wrote a generalized protocol in Python that can be easily customizable to plate out different volumes, whilst taking into account nuances such as avoiding the wells of the border of the plate and using a larger volume of cells. These were all measures that we discovered significantly improved reproducibility in our results, thus spurring us to create protocols with these considerations in mind. With this partially automated workflow, we were able to generate consistent data within and across different sets of experiments to further optimize the precision of our measurement.Data visualisation
We established a systematic workflow for handling the data obtained from experiments in order to ensure efficiency and reproducible data. The modelling team wrote standardised MATLAB scripts that allowed them to examine all the raw data sets generated over the course of the project. The raw data was then examined at the level of individual replicates to identify sources of variability within a single experiment. The wet lab team was then prompted to make changes to their experimental workflow that allowed for more robust data in future experiments. For example, the modelling team made the aforementioned observation that wells along the border tend to display high level of deviation and spurred us to avoid them, allowing for much more reliable triplicates in future experiments. In addition to processing data for an experiment, the team also recognised trends across different experiments and used techniques such as plotting them on the same graph to study the aspect of reproduciblity. Furthermore, the modelling team also provided visual plots and aids that assisted the wet lab team in understanding their results, leading to fast and iterative experimental designs that greatly improved our workflow. Processing and presenting the experimental data in different formats allowed the entire team to identify hidden trends and gain better insights into the systems.