Arizona_State_University_iGEM_2017_Results
Engineering Variable Regulators for a Quorum Sensing Toolbox - EVR-QST
*************************************************************************************************************
Sender Quest, Battle of the AHLs
Some of the specific questions this research was aiming to answer were: How do combinations of senders affect gene output? Can I find any combinations of senders that increase the overall GFP expression? Can I find any combinations that do not affect the GFP expression?
The specific senders I have chosen for the induction tests were selected because previous research showed that they have either a very low or very high rate of GFP induction when used in a single sender/ receiver circuit. In other words, the chosen senders tend to either work very well or not very well at all and we need more data on how well these senders express the gene when used in combination with another. By combining two senders at a time, sometimes with senders that have shown to induce a high GFP expression and sometimes with senders that have shown a weak induction, I wanted to see if I could increase or decrease the GFP expression on demand as needed. The controls I used for the experiment were testing single sender inductions on the same plate as the combinations, the use of blank wells (LB AMP 100%), a positive GFP control, and a negative control with negative receiver cells and negative sender supernatant.
The first sets of tests that utilize the LuxR were duplicates from last years results, there were some discrepancies in their final data so, all tests were duplicated to ensure the results were accurate before moving forward with the testing of further receivers.
Additional information regarding error calculations and additional graphs for the data presented below is available in the results summary section.
Test #1 with LuxR: Sender A: LasI, B: EsaI, C: RpaI
The receiver being used for the below results is the Lux receiver. The second set of senders that was tested is shown below, these are all the combinations and percentages of the AHLs for the test including the controls. Each data point was tested in triplicate. The colors will coordinate with the graphs for each set of tests. The graphs for each set of data will include the overall average GFP signal, the average OD 600 and the normalization of the GFP over the OD 600. The number of data points used made adding individual error bars ineffective as the data was not able to be read. Error was calculated on the controls and added as separate bar graphs below the full data set. There was also Hill curve (trans equations) made that include error/ standard deviation if more information is needed for any notable results.
Results from test #1 with LuxR:
In this set of tests with the Lux receiver the results showed that the LasI expressed the highest, EsaI 2nd highest and RpaI 3rd highest. The graphs concluded that the higher the Las and EsaI combination, the higher the overall GFP expression. No combinations pushed the GFP expression higher than any 50% sender alone.
Test #2 with LuxR: Sender A: RhlI, B: CerI, C: AubI
The second set of senders that was tested is shown below, these are all the combinations and percentages of the AHLs for the test including the controls. Each data point was tested in triplicate. The colors will coordinate with the graphs for each set of tests. The graphs for each set of data will include the overall average GFP signal, the average OD 600 and the normalization of the GFP over the OD 600. The number of data points used made adding individual error bars ineffective as the data was not able to be read. Error was calculated on the controls and added as separate bar graphs below the full data set. There was also Hill curve (trans equations) made that include error/ standard deviation if more information is needed for any notable results.
Results from test #2 with LuxR:
This test showed some notable results. As seen clearly in the last graph, the AubI showed a higher expression when mixed with 10% of a second sender (even when that sender was a negative control sender). The 40% AubI mixed with 10% negative sender and the 40% AubI mixed with 10% RhlI both expressed higher than the 50% AubI by itself. This result was confirmed in another test where 40% Aub mixed with 10% EsaI and 40% AubI mixed with 10% CerI both expressed higher than the 50% AubI alone.
Test #3 with LuxR: Sender A: LasI, B: EsaI, C: RpaI
The third set of senders that was tested is shown below, these are all the combinations and percentages of the AHLs for the test including the controls. Each data point was tested in triplicate. The colors will coordinate with the graphs for each set of tests. The graphs for each set of data will include the overall average GFP signal, the average OD 600 and the normalization of the GFP over the OD 600. The number of data points used made adding individual error bars ineffective as the data was not able to be read. Error was calculated on the controls and added as separate bar graphs below the full data set. There was also Hill curve (trans equations) made that include error/ standard deviation if more information is needed for any notable results.
Results from test #3 with LuxR:
In this set of tests with the Lux receiver the results showed that the RpaI expressed the highest, LasI 2nd highest and EsaI 3rd highest. The graphs concluded that the higher the RpaI and LasI combination, the higher the overall GFP expression. No combinations pushed the GFP expression higher than any 50% sender alone.
Test #4 with LuxR: Sender A: AubI, B: EsaI, C: CerI
The fourth set of senders that was tested is shown below, these are all the combinations and percentages of the AHLs for the test including the controls. Each data point was tested in triplicate. The colors will coordinate with the graphs for each set of tests. The graphs for each set of data will include the overall average GFP signal, the average OD 600 and the normalization of the GFP over the OD 600. The number of data points used made adding individual error bars ineffective as the data was not able to be read. Error was calculated on the controls and added as separate bar graphs below the full data set. There was also Hill curve (trans equations) made that include error/ standard deviation if more information is needed for any notable results.
Results from test #4 with LuxR:
This test showed some notable results. As seen clearly in the last graph, the AubI showed a higher expression when mixed with 10% of a second sender (even when that sender was a negative control sender). The 40% AubI mixed with 10% negative sender and the 40% AubI mixed with 10% EsaI both expressed higher than the 50% AubI by itself. This result was confirmed in the previous test #2 with the LuxR (40% Aub mixed with 10% EsaI and 40% AubI mixed with 10% CerI both expressed higher than the 50% AubI alone).
Test #5 with LuxR: Sender A: LuxI, B: BraI, C: RpaI
The fifth set of senders that was tested is shown below, these are all the combinations and percentages of the AHLs for the test including the controls. Each data point was tested in triplicate. The colors will coordinate with the graphs for each set of tests. The graphs for each set of data will include the overall average GFP signal, the average OD 600 and the normalization of the GFP over the OD 600. The number of data points used made adding individual error bars ineffective as the data was not able to be read. Error was calculated on the controls and added as separate bar graphs below the full data set. There was also Hill curve (trans equations) made that include error/ standard deviation if more information is needed for any notable results.
Results from test #5 with LuxR:
In this test there were some results that were not able to be replicated, meaning that there may have been some contamination or other unaccounted for error. The 40% LuxI + 10% negative sender expressed higher than the 50% LuxI alone, this result was not able to be replicated. Again the 40% LuxI + 10% RpaI expressed higher than the 50% LuxI alone but the result was unable to be replicated and lastly the 40% LuxI + 10% BraI expressed higher than the 50% LuxI and again, the result was unable to be replicated.
Test #6 with LuxR: Sender A: LuxI, B: RhlI, C: LasI
The sixth set of senders that was tested is shown below, these are all the combinations and percentages of the AHLs for the test including the controls. Each data point was tested in triplicate. The colors will coordinate with the graphs for each set of tests. The graphs for each set of data will include the overall average GFP signal, the average OD 600 and the normalization of the GFP over the OD 600. The number of data points used made adding individual error bars ineffective as the data was not able to be read. Error was calculated on the controls and added as separate bar graphs below the full data set. There was also Hill curve (trans equations) made that include error/ standard deviation if more information is needed for any notable results.
Results from test #6 with LuxR:
In this set of tests with the Lux receiver the results showed that the LuxI expressed the highest, RhlI 2nd highest and LasI 3rd highest. The graphs concluded that the higher the LuxI combination, the higher the overall GFP expression. No combinations pushed the GFP expression higher than any 50% sender alone.
This next section of results is for the tests done with the next receiver, LasR.
Test #1 with LasR: Sender A: AubI, B: EsaI, C: CerI
The first set of senders that was tested is shown below, these are all the combinations and percentages of the AHLs for the test including the controls. Each data point was tested in triplicate. The colors will coordinate with the graphs for each set of tests. The graphs for each set of data will include the overall average GFP signal, the average OD 600 and the normalization of the GFP over the OD 600. The number of data points used made adding individual error bars ineffective as the data was not able to be read. Error was calculated on the controls and added as separate bar graphs below the full data set. There was also Hill curve (trans equations) made that include error/ standard deviation if more information is needed for any notable results.
Results from test #1 with LasR:
This test showed some notable results. As seen clearly in the last graph, the AubI showed a higher expression when mixed with 10% of a second sender (even when that sender was a negative control sender). The 40% AubI mixed with 10% negative sender and the 40% AubI mixed with 10% EsaI or 10% CerI both expressed higher than the 50% AubI by itself.
Test #2 with LasR: Sender A: BraI, B: LasI, C: AubI
The second set of senders that was tested is shown below, these are all the combinations and percentages of the AHLs for the test including the controls. Each data point was tested in triplicate. The colors will coordinate with the graphs for each set of tests. The graphs for each set of data will include the overall average GFP signal, the average OD 600 and the normalization of the GFP over the OD 600. The number of data points used made adding individual error bars ineffective as the data was not able to be read. Error was calculated on the controls and added as separate bar graphs below the full data set. There was also Hill curve (trans equations) made that include error/ standard deviation if more information is needed for any notable results.
Results from test #2 with LasR:
This test showed some notable results. As seen clearly in the last graph, the BraI showed a higher expression when mixed with 10% of a second sender (even when that sender was a negative control sender). The 10% BraI + 40% AubI mixed expressed higher than the 50% AubI by itself. Close behind the 50% AubI was the 50% LasI. Interesting result because the AubI expresses higher than the matching sender LasI to its own LasR receiver.
Test #3 with LasR: Sender A: AubI, B: RpaI, C: BjaI
The third set of senders that was tested is shown below, these are all the combinations and percentages of the AHLs for the test including the controls. Each data point was tested in triplicate. The colors will coordinate with the graphs for each set of tests. The graphs for each set of data will include the overall average GFP signal, the average OD 600 and the normalization of the GFP over the OD 600. The number of data points used made adding individual error bars ineffective as the data was not able to be read. Error was calculated on the controls and added as separate bar graphs below the full data set. There was also Hill curve (trans equations) made that include error/ standard deviation if more information is needed for any notable results.
Results from test #3 with LasR:
This test showed some notable results. As seen clearly in the last graph, the 40% AubI showed a higher expression when mixed with 10% of a second sender (even when that sender was a negative control sender). The 10% BjaI + 40% AubI mixed expressed higher than the 50% AubI by itself. Interesting result because the AubI expresses higher when mixed with another sender, seemingly with both LuxR and LasR.
All the below tests are with the TraR Receiver.
Test #1 with TraR: Sender A: AubI, B: EsaI, C: LuxI
The first set of senders that was tested is shown below, these are all the combinations and percentages of the AHLs for the test including the controls. Each data point was tested in triplicate. The colors will coordinate with the graphs for each set of tests. The graphs for each set of data will include the overall average GFP signal, the average OD 600 and the normalization of the GFP over the OD 600. The number of data points used made adding individual error bars ineffective as the data was not able to be read. Error was calculated on the controls and added as separate bar graphs below the full data set. There was also Hill curve (trans equations) made that include error/ standard deviation if more information is needed for any notable results.
Results from test #1 with TraR:
This test showed some notable results. As seen in the last graph, the AubI and the EsaI expressed much higher than the LuxI. This is a result that does not match other tests performed by the team. This was also not a result that was able to be replicated and it is thought that there may have been an issue with the LuxI that caused it to not induce the TraR.
Test #2 with TraR: Sender A: LasI, B: RhlI, C: BjaI and SinI
The second set of senders that was tested is shown below, these are all the combinations and percentages of the AHLs for the test including the controls. Each data point was tested in triplicate. The colors will coordinate with the graphs for each set of tests. The graphs for each set of data will include the overall average GFP signal, the average OD 600 and the normalization of the GFP over the OD 600. The number of data points used made adding individual error bars ineffective as the data was not able to be read. Error was calculated on the controls and added as separate bar graphs below the full data set. There was also Hill curve (trans equations) made that include error/ standard deviation if more information is needed for any notable results.
Results from test #2 with TraR:
There are not any notable results in this test, the sender's did not induce the Tra receiver. The overall GFP expression (in A.U.) showed no induction and it may be orthogonal pathways when using filtered supernatants from cultured cells.
Test #3 with TraR: Sender A: CerI, B: BraI, C: RpaI
The third set of senders that was tested is shown below, these are all the combinations and percentages of the AHLs for the test including the controls. Each data point was tested in triplicate. The colors will coordinate with the graphs for each set of tests. The graphs for each set of data will include the overall average GFP signal, the average OD 600 and the normalization of the GFP over the OD 600. The number of data points used made adding individual error bars ineffective as the data was not able to be read. Error was calculated on the controls and added as separate bar graphs below the full data set. There was also Hill curve (trans equations) made that include error/ standard deviation if more information is needed for any notable results.
Results from test #3 with TraR:
There are not any notable results in this test, the sender's did not induce the Tra receiver. The overall GFP expression (in A.U.) showed no induction and it may be orthogonal pathways when using filtered supernatants from cultured cells.
************************************************************************************
Noteworthy (successful) results summary.
Here you will find all supporting data for the conclusions described in the results section including the error calculations and all additional graphs. Each section is broken down by the receiver and labeled at the start of the section.
Summary for Lux Receiver:
When testing the LuxR the AubI showed a higher expression when mixed with 10% of a second sender (even when that sender was a negative control sender). The 40% AubI mixed with 10% negative sender and the 40% AubI mixed with 10% RhlI both expressed higher than the 50% AubI by itself. This result was confirmed in another test where 40% Aub mixed with 10% EsaI and 40% AubI mixed with 10% CerI both expressed higher than the 50% AubI alone. (See graphs below).
Summary for Las Receiver:
Below we see an experiment where the AubI expresses higher when mixed with 10% of any other sender, these results are with the Las receiver. This evidence further confirms that AubI simply works best when mixed versus being used alone. The 40% AubI mixed with 10% negative sender, 10% EsaI and 10% CerI all expressed higher than the 50% AubI alone.
Below, another test showed some notable results. As seen clearly in the last graph, the BraI showed a higher expression when mixed with 10% of a second sender (even when that sender was a negative control sender). The 10% BraI + 40% AubI mixed expressed higher than the 50% AubI by itself. Close behind the 50% AubI was the 50% LasI. Interesting result because the AubI expresses higher than the matching sender LasI to its own LasR receiver.
Below, another test showed some notable results. As seen clearly in the last graph, the 40% AubI showed a higher expression when mixed with 10% of a second sender (even when that sender was a negative control sender). The 10% BjaI + 40% AubI mixed expressed higher than the 50% AubI by itself. Interesting result because the AubI expresses higher when mixed with another sender, seemingly with both LuxR and LasR.
barely induces the Las receiver. May be orthogonal (below).
AubI expresses higher than the LasI with the LasR (below)
Summary of the Tra Receiver:
The TraR receiver is orthogonal with most senders, even the TraI synthetic AHL barely induces its own receiver. The only two senders that induced the Tra receiver (aside from the TraR synthetic AHL) within a noticeable range, was the AubI and the EsaI.
Unsuccessful results:
In the below test with the LuxR, there were some results that were not able to be replicated, meaning that there may have been some contamination or other unaccounted for error. The 40% LuxI + 10% negative sender expressed higher than the 50% LuxI alone, this result was not able to be replicated. Again the 40% LuxI + 10% RpaI expressed higher than the 50% LuxI alone but the result was unable to be replicated and lastly the 40% LuxI + 10% BraI expressed higher than the 50% LuxI and again, the result was unable to be replicated.
Future plans for the project:
In consideration for the foundational advance of this project in the future, it would be advisable to add additional concentrations of the sender combinations with each receiver. These experiments tested combinations of 50%, 40%, 25%, 12.5%, and 10% and the notable results from these experiments could be used as a base for creating new experiments with additional concentrations and combinations. As new receivers are developed replicating these tests with them would be wise as it would grow the library of currently characterized receivers with their senders.
Considerations for replicating the experiments:
A few important details to take into consideration when replicating these experiments are, noting that you get the best results from freshly transformed bacteria versus using bacteria from existing agar plates and that it is best to filter the supernatant from the sender cells immediately prior to running your experiment. Highest GFP induction occurred when freshly transformed samples were gathered and grown in liquid culture, using the shaking incubator, on the day of the experiment.
Synthetic AHL Concentration Inductions of LasR, LuxR, and TraR
Questions these induction tests answer was how GFP expression is affected by different synthetic AHLs over a range of concentration. Conducting these experiments will also provide insight on what the optimal AHL concentration a sender is able to communicate a signal to the receiver to better express GFP. It will also be seen which systems of AHL and receiver are orthogonal.
The synthetic AHLs and range of concentrations chosen for these inductions came from the paper "Quorum Sensing Communication Modules for Microbial Consortia" written by Scott Spencer and Jeff Hasty at the University of California, and from last year's iGEM team. This year, it was decided to expand the characterization of the new and improved receivers of this year to include a broader concentration range, and more synthetic AHLs. To see these relationships, the synthetic chemicals induced receiver cells in liquid culture and analyzed the growth curves in a 96-well plate done in triplicate. The controls for these inductions included blank LB wells, a positive GFP control, a negative receiver control, a negative sender supernatant control, and a positive sender control.
The experiments performed were to characterize the Las, Lux, and Tra receivers included the synthetic AHLs of 3-oxo-C12 (Las), 3-oxo-C6 (Lux), 3-oxo-C14 (Sin), Coumaroyl (Rpa), C4 (Rhl), and 3-oxo-C8 (Tra). The following graphs show the maximum (DGFP/OD)/DT across a concentration range of 1E-14M to 1E-4M. Each data point was done in triplicate then averaged. These points determine where the GFP is changing the fastest, and represents the maximum rate of change at a specific synthetic AHL concentration.
The graphs below depict the GFP production from the F2620 receiver to be higher at a higher concentration of the synthetic AHL of Las. The second graph, depicts the same nature however at a higher concentration of the synthetic AHL of Lux. It is interesting to note that the Las sender, at a lower concentration, promotes a higher GFP expression than the Lux. Our group hypothesized that it would be opposite considering the LuxI sender comes from the same system as F2620. These results show differently.
As Las receiver and Tra receiver are also induced by the synthetic AHLs, a similar trend emerges. There is little to no GFP expression for the lower concentration, between 1E-14 through 1E-9, 1E-8, or 1E-7 in some cases. Then a dramatic increase in GFP expression as the AHL concentration increases. This indicates that the system shows significant GFP expression at higher concentrations.
Another interesting result was the with the synthetic AHL Rpa. As depicted in the graph below, it shows an almost steady state of GFP production, independent of the concentration of the AHL signal. This trend can also be seen in multiple systems such as the induction of LasR with a Rhl AHL, LuxR with Rhl AHL, and LuxR with a Rpa AHL. This might be a useful finding if researchers wanted to use a smaller amount of signal due to limiting resources of RpaI or RhlI sender. However, it did not show an orthogonal pathway since it did in fact promote GFP expression.
The trend outlined above is also expressed in the below graphs. From concentrations 1E-14M through 1E-5M there is a steady production of GFP that is independent of the AHL signal concentration. However, as the receiver is induced by the 1E-4M signal, an outlier of GFP expression is produced. This is an unusual result as the 1E-4M concentration in the LasR and Lux AHL system is expected to be around 2-6 arbitrary units. This unexpected point is also present in the TraR and Rhl AHL system. The max GFP expression of 1E-4M is an outlier as the trend indicates orthogonality between the receiver TraR and RhlI sender. Within our experiments, we implied that orthogonality is defined as not expressing GFP or having no communication within the systems. This might be an important find for synthetic biologist and genetic engineering to use this orthogonal quorum sensing circuit within their genetic circuit to be able to, with confidence, produce results of a protein or phenotypic expression. This orthogonality does not apply to the system of LasR and Lux synthetic AHL, however, as there was still induction and GFP was shown to express. As stated above, this system could be beneficial to researchers who have limited resources of the Lux sender and wish to express the same amount of GFP that would occur in higher concentrations.
The graph of LasR and Las AHL displayed below is a combination of different sender concentrations of LasI in respect to the GFP expression of LasR. It is interesting to note that the corresponding sender and receiver system, Las, does not express as highly in GFP expression as other experiments with different senders. Our group hypothesized that this sender/receiver system would still promote a higher GFP expression, even at a more diluted concentration of synthetic AHL. This trend can also be seen in the graph of LasR and Sin AHL. The 1E-7M expresses a higher GFP of around 20 arbitrary units than the 1E-4M, which is an unexpected result. This is beneficial to researchers looking for maximum GFP expression using the Las receiver and a small supply of sender.
A notable conclusion can be drawn from the below graph of the Tra receiver and Lux AHL system. It is seen from 1E-14 to 1E-7 that there is little to no GFP expression, while the 1E-4M concentration reached 6 arbitrary units with high error. This type of fluctuation also occurs in the LuxR with Sin AHL, although it is more varied throughout the concentration range. Both graphs appear to have an upward trend despite the fluctuations as concentration increases.
Future plans for the experiment:
For the foundational advance of this experiment in the future, it would be advisable to expand the concentrations used for inductions. Currently, the concentrations used include 1E-14M, 1E-13M, 1E-12M, 1E-11M, 1E-10M, 1E-9M, 1E-8M, 1E-7M, 1E-6M, 1E-5M, and 1E-4M. From the experiments, it is seen that the most GFP expression occurs in higher concentrations from 1E-8M to 1E-4M. If the range is narrowed between the values to include 1E-4.5M, 1E-5.5M, 1E-6.5M, and 1E-7.5M, a better idea of how GFP is expressed can be analyzed. Also if more synthetic AHLs are able to be tested with these receivers, the characterization can be expanded and these receivers can be further analyzed on how they work in multiple systems.
Considerations for experimental replication
When replicating these inductions, it is important to note that these AHLs are dissolved in ethyl acetate and serial dilutions were performed to get the desired concentrations. These dilutions were done the day of induction, and kept at -20 degrees Celsius for optimum results and lowest chance of degradation before testing. Also when plating, the final concentration of each AHL in the wells were the concentrations listed, with 1% ethyl acetate. Freshly transformed bacteria cells and grown cultures also gave optimum results for induction.
