This page is intended to document the experiments we have directly performed, describe how difficult we found that they were to perform and explain the reasoning behind the choices we made. Full protocols and buffer recipes can be found on our protocols page. A full list of results can be found on our wet lab overview page. For experiments relating specifically to the Interlab study see our Interlab page.
We wanted to test our copper-sensitive promoter parts at a range of different copper concentrations over time. We decided that our ideal promoter system would show a large change in expression over the region 0.01-0.05mM because this is the range under which copper concentration would rise after a meal. Our ideal system would also induce quickly so the chelator could be produced and the free copper chelated before it could be absorbed.
However we decided to test the promoters to copper concentration outside this range (0mM to 2mM) and for a longer period of time. This was to better understand how our promoter systems operated, provide a more complete data set for our dry lab team to parameter fit and to provide data for others wishing to use our parts for our specifications. In order to increase the chance of our data being applicable to a probiotic bacterial in the small intestine we set the plate reader to human body temperature of 37°C and gently shaking at 225 rpm to stimulate gastric movement. We also verified that our growth medium was close to pH7 which is around the pH small intestine.
Two different fluorescent proteins were used in our project. The fluorescence our each our samples is presumed to be proportional to the expression level our our promoters. Our CueR-linked systems use the GFP variant sfGFP allowing these to directly compared. (Excitation filter 485-12nm, emission filter 520nm). Our pCusC RFP sequence however was provided with the red fluorescent protein mKate and when we attempted to improve the part by adding a positive feedback loop with CusR we used the same protein so we could compare them on the same plate and see if there was any improvement. (Excitation 580-10nm, emission 620-10nm)
To account for the number of cells present at different copper concentrations and different times we measured the optical density (OD) as a proportional measure of the number of cells present. For our parts using GFP, the optical density at 600nm (orange light) was used as at this wavelength has less interference with the yellow broth. For our mKate parts however, measuring at cell density at 600nm wouldn’t be accurate because of the emission from the fluorescent protein at approximately this wavelength. Consequently we measured the optical density at 700nm for these parts.
When comparing our copper promoter systems at different copper concentrations we chose to compare the data after they had been in the plate reader for four hours. We found that this was just before the optical density reached maximum so the cells were still about in the exponential growth phase.
Plate reader experiments were prepared by picking individual colonies off stored plates into 5ml of LB with 1 in 1000 chloramphenicol and grown overnight (at least 8 hours).
A range of copper concentrations were prepared from stock solutions. A large volume plate was then prepared with 10μl of copper solution, 10μl of overnight culture and 980μl of broth with antibiotic. This resulted in a 1 in 100 dilution of the copper solutions prepared:
This large-volume plate was then centrifuged to mix the solutions and then 200μl transferred to a small-volume plate with a clear lid and then placed in the plate reader. The delay between mixing the cells, broth and copper solution and the starting of the plate reader was found to be less than 30 minutes.
In each of our plates we had 8 rows of 12 columns (96 wells in total). In the first column we included the negative control part from the Interlab study inside our testing MG1655 strain at 0mM copper. This gave us the background growth curve of the cells and acted as baseline to compare the increase in fluoresce to and show how leaky our constructs were at 0mM copper. Two repeats from each of four separate picked colonies was used.
Similarly in the second column we used the positive control part from the Interlab study. Whilst using a different form of GFP it confirmed to us that the plate reader was performing correctly. Using these parts rather than untransformed, plasmid-free MG1655 strain allowed us to use antibiotic-containing broth in both lanes to account for any effect the antibiotic had on measurement. As we lacked a positive control expressing red fluorescent protein for our CusSR-linked systems we still included the GFP positive control to see see if growth was different when expressing a large amount of protein with no benefit to the cell.
In the remaining wells of the plate we included four biological repeats of two parts we wished to compare (e.g. pCopA sfGFP and pCopA CueR sfGFP) across the range of copper concentrations. Four biological repeats was considered sufficient for our purposes especially when combined with our flow cytometry data. By putting the parts in the same plate we ensured that the conditions were identical for both.
The plate reader measured the fluorescence and OD every ten minutes for at least 12 hours, shaking between measurements.
Overall we found this experiment to be reasonably simple albeit time consuming to perform. With practice the plate can be set up in about two hours. Although the experiment was designed to use large volumes and the same volume in each well, accurately pipetting exact volumes into each well was found to be quite monotonous and it was easy to loose track of which well you were working on. If a single volume was added incorrectly the entire plate would need to done from the start. Consequently we found it easiest to do prepare as much as possible (e.g. the copper solutions) the night before and perform the experiment early in the morning or just after lunch when relatively attentive.
The easiest method to keep track of position in the plate was to pipette against the sides of the wells to be more visible and then shaking the plate after each component was added. Audibly counting rows and columns also helped. The easiest method of filling the plate was to add the all the copper solutions first down the columns then add the positive and negative control cells into their respective columns followed by each of the parts from left to right across each row. Going from low to high copper concentrations meant that the same pipette tip could be used for the part overnight cultures with relatively tiny change in copper from contamination between wells. Using the same pipette tip reduces the number of motions necessary for each row. Using a multi-pipette to go from the large volume plate to a small volume plate greatly increases the speed of the procedure.
Flow cytometry is a technique whereby cells are passed individually into the path of a beam of light. The frequency range of incident light can be adjusted to allow excitation of specific fluorophores in the sample of cells. Downstream detectors can measure fluorescent emission, and this data can be used to quantify the amount of fluorophore in each cell. Using the Attune™ NxT Flow Cytometer from Invitrogen™, we measured the GFP (sfGFP) and RFP (mKate) fluorescence in bacteria containing all of our plasmids for which the protein contained a fluorescent protein tag. By performing these experiments in the presence of several concentrations of inducer, we were able to determine the population dynamics with respect to the expression of the protein product under different inducing conditions.
To ensure comparability of experiments, all cells were grown for 3-4hrs (until entrance into the exponential growth phase) at 37°C and 225rpm shaking in the presence of the inducer before measurement. This allowed adequate time for activation of expression by the promotor systems. The negative control used in all cases were MG1655 bacteria containing an empty shipping plasmid. As these did not contain any fluorescent molecules, this population could be used to set the negative “gate” (i.e. the background fluorescence of the bacterial cells). Although the experiments were tedious as every sample had to be measured manually, the results were of remarkably high quality, were clearly interpretable, and fit very well with the other experimental data.
Microscopy was done in order to visually confirm the plate reader and flow cytometer experiments and to study the protein’s cellular distribution.
The experiment started with 5ml overnight cultures containing the appropriate antibiotic. Then in the morning 100μl of each colony was pipetted into 5ml of fresh LB with antibiotic and inducer (copper or arabinose) and grown till the OD reached 0.4-0.6.
A flask of 1% agarose made with MilliQ was melted. 200μl of this was placed on a slide between two coverslips, flattened to get a nice smooth surface where the bacteria are immobile. 20μl of the culture are then added. The slide can then be viewed under a fluorescence microscope.
After finding the correct focal plane the slide can be moved to find as many cells as possible to image. After focusing again an image of the DIC and fluorescence channels can be obtained.
Whilst finding the correct focal plane was often quite difficult and the preparation of slides quite fiddly, this experiment was very relaxing to perform. A single person can easily do it in a day.
Alginate Bead Preparation
We needed a delivery system for our copper-chelation system so turned to the public for their guidance. Survey 4 was centred around delivery methods, how one would like to take a probiotic treatment and at which frequency. 27% of the responses were from Wilson’s Disease patients so the design of our delivery method truly reflects patients’ wishes. ‘Gel-like bead’ and ‘tablet/pill’ came out as the two most popular delivery choices, hence we started here. You can read more about this survey, and the others we carried out, on our surveys page.
Having read about the Oxford iGEM 2015 team’s project and spoken to many of the team’s members we started to look into the possible use of alginate-beads as a delivery method. Through their research on alginate beads they showed that GFP-producing bacteria could be encapsulated inside the beads and that the bacteria stayed alive in storage for at least twenty days. You can read more about their alginate bead research on their wiki.
For our project we need the delivery method to encase the bacteria and successfully deliver them to the intestine, while maintaining their integrity in the stomach. To investigate whether alginate beads would achieve all of these aims we prepared a 2% (w/v) sodium alginate solution and then added 2ml of dye. The dye was used as a way of seeing if anything would leave the beads, i.e. if the beads disintegrate, the dye would leave the beads and enter the solution. We chose crystal violet because last year’s team used this dye and we used a similar experimental design.
To test whether these beads would leak dye (i.e. maintain their structural integrity) in the stomach and intestine, we produced solutions which would stimulate these conditions. The simulated stomach solution was made up using 0.2% (w/v) NaCl solution and then made up to pH 2.0 with 1M HCl. The simulated intestine solution was made up using 0.68% (w/v) monobasic potassium phosphate solution, and then made up to pH 7.2 with 1M NaOH.
The alginate beads were placed in either solution and then into an incubator shaking at 37°C for 90 minutes, as this was said to most closely simulate the stomach and intestine (see references on our overview page. Every 10 minutes 100μl of both solutions were removed and placed into wells of a 96 well plate, to then measure absorption of each well once the 90 minutes was completed. The dye absorbs at 595nm, hence you can see an increase of dye in the solution by measuring the absorption (so can see if the beads have disintegrated throughout their 90 minutes in the simulated stomach/intestine).
We don’t want our beads to disintegrate in the stomach because then the stomach conditions would damage our copper chelator. We do however want the beads to disintegrate in the intestine, to then release our system which can start copper chelation. The data from the alginate-only beads showed the fluorescence of the stomach solution increased significantly over the 90 minutes, hence the alginate-only design is not appropriate for our project.
We then turned back to primary literature to see how we could improve our beads design. Many papers expressed the benefits of a layered chitosan-alginate bead. If you would like to read these papers, please see the references on our overview page. These beads are created by preparing the beads as before, then after 30 minutes of hardening (the “stabilisation phase”), you filter them and alternatively dip them into 0.4% (w/v) chitosan solution for 10 mins and 0.04% (w/v) alginate solution for 10 mins. To test the stability of these beads in the stomach and intestine we carried out the same experiment with the simulated solutions as described above. Success! These beads didn’t show a significant increase in fluorescence in the simulated stomach solution, meaning they maintained their integrity in this environment. In the simulated intestine solution however the fluorescence increased steadily over the 90-minute period, meaning they slowly start to disintegrate in this environment. This is exactly what we need for our delivery method!
Cu Absorbance Assays
The copper assay used was the Bathocuproine disulphonic Acid (BCS) absorbance assay. BCS is colourless in the absence of Cu+ but upon exposure to Cu+, BCS forms complexes with Cu+ and absorbs strongly in the 480nm range. In our assays, at concentrations of 50µg/mL, the 480nm absorbance varied linearly with the Cu+ concentration from the detection limit of around 1µM Cu+, to approximately 20µM Cu+. MymT and Csp1 bind copper in the Cu+ form and the assay also requires singly charged copper. Therefore the assays were optimized to include a suitable about of mild reducing agent to ensure reduction of the added CuSO4 (releases Cu2+) to Cu+. After trying L(+)-Ascorbate and DL-Dithiothreitol, and L-Glutathione as candidates, L-Glutathione was selected as it was both mild enough to not damage biological material and efficient at reducing Cu2+. At >2-3 times the concentration of Cu2+ in solution, L-Glutathione had maximum reductive activity against Cu2+. This assay was used to measure the ability of purified MymT-sfGFP, purified Csp1-sfGFP, and whole bacteria expressing the chelators to take up free copper from the medium and sequester it. Unfortunately, none of these assays could find appreciable copper chelation activity in any circumstance.