Team:CSU Fort Collins/Results
COLORADO STATE
UNIVERSITY iGEM
Project Results
Key Results
Key Achievements- Construction and submission of multiple team parts
- Proof of the composite lac promoter:fadD:fadL part function
- Characterization of KillerRed in E. coli for use as an induced lethality switch
Key Challenges
- Reworking breakdown construct ideas after fadD experiment
- Inconclusive results about trans-zeatin production
- Proper handling and experimental design for testing KillerRed
Breakdown Results
Figure 1: First growth curve for the breakdown process. Growth comparison between lac promoter control and lac promoter:fadD. We concluded fadD addition was insufficient for increased fatty acid breakdown.
Figure 2: Our second growth curve for the breakdown process. Growth comparison between lac promoter control, lac promoter:fadD, lac promoter:fadL, and lac promoter:fadD:fadL.
Once we discovered that having just the fadD wasn’t enough we went back and did more research to see what else we could add to improve its performance. We found the the transport of fatty acids into the cell was another rate limiting step. This step is facilitated by fadL. So we then constructed plasmids with just fadL and a promoter as well as one with both fadD, fadL, and a promoter. We then ran a very similar experiment as before on all of those constructs with the lac promoter plasmid as a control. Instead of using steric acid, we grew the cells on 0.04% used frying oil. The results of this experiment can be seen in Figure 2. We found that combining the fadD and fadL resulted in the best growth rate by far.
With the concept proven, we moved on to testing all of these constructs in real used frying oil. The growth on 100% frying oil was quickly too much to measure using ODs so we switched to measuring cell weight. The 50% oil: 50% Brij58 solution remained at low enough densities that we continued to measure growth using ODs. The results from this experiment can be seen in Figures 3 and 4.
From this data we can determine that when grown on 50% Frying, 50% Brij 58 solution both of our constructs reach greater cell density than our control. It is not completely clear which construct is preferable as they overlap and even cross. The fadD + fadL construct had higher initial growth but lower overall growth than just the fadL by itself. More experiments would need to be run to determine if this is typical behavior or just something in this particular run. If it is typical then this is useful data for eventual production design. The optimum point in the cells’ growth cycle for trans-zeatin growth would dictate which construct we would want to use. The fadD +fadL cells likely as a whole reached stationary phase much sooner than the fadL cells. If trans-zeatin is best produced in stationary phase then fadD +fadL would be the better choice. However if trans-zeatin is better produced while the cells are in exponential growth the fadL by itself would be more ideal.
The data of the growth on 100% frying oil is less conclusive. It seems that the fadD + fadL does the best with fadL not far behind. Both the constructs do better than the control. This method of measurement is very noisy. The viscosity of the oil makes it very hard to work with. There is lots of room for error, including: getting a consistent volume of oil in the draw, spinning down the sample all the way to collect all the cells, and pulling off all the oil so only the cell weight is left. Our results (shown in Figure 4) indicate at least an initial observation that the addition of these constructs is beneficial to cell growth.
Figure 3: Growth comparison of strains grown on 50% used frying oil solution. Compared to our control, the lac promoter:fadL strain was not consistently at higher ODs, but the lac promoter:fadD:fadL strain grew much better.
Figure 4: Growth comparison of strains grown on 100% used frying oil solution. Both experimental strains were able to obtain increased cell growth compared to the control strain.
The data that actually says more about the cells’ usage of the oil is the qualitative data. As the experiment proceeded the color of the samples changed. From this we conclude that the cells were doing something to alter the composition of the media. Another observation that points to the cells using the fatty acids came from the 50% solution. When no cells were added the water separated from the Brij 58:oil solution into very clear layers. A little mixing quickly re-homogenizes the sample. After about two hours of growth the media separated significantly less and by the 12 hour data point stayed homogenous even without shaking. This told us that the media was becoming more soluble in water, implying that the cells took up and used the insoluble fatty acids.
Trans-zeatin Production Results
Our team was able to successfully create two constructs, a basic part and a composite part which included an inducible promoter, which encode for the trans-zeatin biosynthesis pathway. We ran multiple growth experiments to assess how much metabolic stress, if any, our construct introduced into our strain. We tested our strain, with the tzS:LOG operon behind a lac promoter, against a plasmid just containing the lac promoter. We compared their growth over 24 hours and 72 hours in shake flasks, as well as in bioreactors over 72 hours. From the results, we can conclude that the presence of the genes does not increase metabolic stress on the cell.
Figure 5: Growth comparison between a control strain and our trans-zeatin strain in shake flasks over 24 hours.
Figure 6: Growth comparison between a control strain and our trans-zeatin strain in shake flasks over 72 hours.
Figure 7: Growth comparison between a control strain and our trans-zeatin strain in bioreactors over 72 hours.
There are two possible reasons for these results. It is possible that trans-zeatin is being created, but is below our limit of detection using HPLC-UV. To resolve this, we will use a more sensitive process, liquid chromatography with dual mass spectrophotometry, which can detect trans-zeatin presence in the pico- to nanogram per mL range. Alternatively, it is possible that the E. coli is not producing trans-zeatin at all. To determine if this is the issue, we will run qPCR to detect mRNA expression.
Kill Switch Results
We ran a lot of experiments to try and characterize the KillerRed part. Working with this part was incredibly difficult. Our first issue was the strain of cells we were using did not have a strong lac repressor, therefore KillerRed was always being expressed. This meant that any light exposure resulted in cell death. This made constructing a kill curve almost impossible, as we lost our control every time at the same rate we lost the cells that were supposed to be expressing KillerRed. So we transformed into E. coli F’I2 cells, which are modified to have a strong lac repression system. We ran multiple experiments on these cells with only small details changed, however we could not find any significant difference between our control and our KillerRed cells. We suspected that it might have to do with our jerry-rigged dark room (cells got light exposure before we even had a chance to plate), or maybe we did not have high enough intensity light (so we tried exposing the plates to direct sunlight, the heat killed all our cells) or maybe our entire premise on how to measure the kill rate was off.
To prevent unintended cell death, we kept our KillerRed strains in a jerry-rigged darkroom. It didn't work, so we found a real darkroom to work in.
Experimental Data
Click to see our data tables
References
- Ma, Zhen et al. "Simultaneous analysis of different classes of phytohormones in coconut (Cocos nucifera L.) water using high-performance liquid chromatography and liquid chromatography–tandem mass spectrometry after solid-phase extraction." Analytica Chimica Acta 610:274-281. January 2008.