BME494s2013 Project Team3
Overview & Purpose
By modifying the input and output for the Lac switch, it may be possible to produce materials such as plastics. The switch could be triggered by another environmental factor other than [IPTG], and instead of producing GFP, more useful materials like plastic could be an output. Currently, production of plastics is a very energy intensive process, and by using bacteria for the production, we can save energy and limit waste into the environment.
The natural Lac-operon has 2 controls that tightly regulate the production of the proteins necessary for the breakdown of lactose. In the presence of glucose, the Lac Operon inhibits the production of those proteins. When glucose is present, the lac repressor is bound to the operator, prohibiting the transcription of the proteins. In the presence of lactose, lactose is able to bind to the lac repressor initiating a conformation change in the repressor protein, causing it to release and allow for transcription. This, however, is not the only control in place. It is incredibly energy intensive to produce the proteins necessary for the breakdown of lactose, so when there is glucose present, the glucose will be metabolized first, even with some lactose present. This is accomplished by the presence of a second regulatory device, cyclic AMP (cAMP). cAMP is present only when there are very low levels of glucose found in the environment. cAMP serves as an activator and binds to RNA polymerase allowing for transcription of the output proteins.
Design: Our genetic circuit
OUR GENE SWITCH:
The functionality of our genetic switch resembles that of an "AND" logic gate: the device requires two conditions to be true in order for an output to be produced. One conditional requirement is that IPTG must be present in the device's environment. When IPTG is present, it binds to the LacI repressor, thus allowing for transcription to continue. The other conditional requirement is that glucose levels in the device's environment must be low. Glucose levels inversely affect production of cyclic AMP (cAMP): when glucose levels are low, cAMP production increases and when glucose levels are high, cAMP production decreases. cAMP binds to catabolite activator protein (CAP) to form the CAP-cAMP complex. In order for this complex to form, cAMP must be present and, thus, glucose levels must be low. The CAP-cAMP complex is an input that our device requires in order to produce an output.
In the natural lac operon, the CAP-cAMP complex leads to enhanced activation of gene expression from the lac operon. However, if glucose is present, cAMP levels will in turn be low and the host will preferentially metabolize glucose even if lactose is present.
Building: Assembly Scheme
To form the build the lac switch, the group used Type IIs Assembly, which allows the the parts to be assembled in one step. For this type of assembly, forward and reverse primers needed to be created and placed in the system so as to create sticky ends that can bind various parts together in a specified order. Additionally, site-directed mutagenesis need be performed so as to remove the BsmBI cut-site in the promoter region. To put the pieces together, PCR is implemented, which allows all the parts to be replicated thousands of times to yield a pure product. Subsequently, digestion and ligation is implemented, during which BsmBI cuts the DNA fragments and creates complementary overhangs that anneal via base pairing.
Testing: Modeling and GFP Imaging
A LAC SWITCH MODEL
We used a previously published synthetic switch, developed by Ceroni et al., to understand how our system could potentially be modeled and simulated. Mathematical modeling is a method of representing via the language of math how a system is expected to behave. The various components of a system, referred to as system parameters, are assigned variable names. These variables hold values that can be adjusted during testing without incurring what would otherwise be additional production costs. Modeling is very beneficial in that multiple tests and simulations can be performed on the system so that a predictable pattern of behavior can be developed, thus reducing production costs.
We modeled our device after the Ceroni switch, using similar variables and parameters to represent the various components and processes of our device. The graphic to the right depicts these components and the variables used to represent them in a mathematical model. The following table contains the variables and parameters used both in our model and the Ceroni et al., model.
AN INTERACTIVE MODEL
When interacting with the MATLAB model, we were able to analyze three outputs: concentration of IPTG over time, concentration of β-galactosidase (Bgal) over time, and mRNA concentration over time. Our focus was on how the concentration of Bgal changed over time as we adjusted the input concentration of IPTG. The default concentration of IPTG was set to 0.32 mM. The Bgal concentration over time when the input IPTG was set to this value can be seen in Image 1 to the right.
Next, we adjusted the concentration of IPTG to attempt to find a value that would allow for sustained output of Bgal over time. We determined the best estimate of IPTG concentration to sustain a relatively constant output of Bgal over time was 0.0639 mM of IPTG. At this concentration of IPTG, the sustained output value of Bgal was approximately 2.54 x 10-4 mM. The resulting data can be seen in Image 2 on the right.
Finally, we adjusted the concentration of IPTG to attempt to find a value that would result in no Bgal being produced. We were able to determine that the lowest value of IPTG concentration that did produce a Bgal output was 0.021 mM of IPTG. Any concentration of IPTG below this value resulted in the lowest observable output of Bgal: approximately 0.4 x 10-5. However, even at this concentration of IPTG, trace amounts of mRNA were still present in the system. These amounts decreased immediately after our model started to run but were still high enough to continue to initially drive transcription. Thus, an initial spike of Bgal output was observed at 0.021 mM of IPTG. Shortly after almost all mRNA had degraded, the total amount of Bgal had degraded to the minimum threshold value of 0.4 x 10-5. Because the mRNA in the system never completely degraded, we never observed a complete lack of Bgal output. The resulting data can be seen in Image 3 to the right.
COLLECTING EMPIRICAL VALUES TO IMPROVE THE MODEL
We explored how one technique, imaging via microscopy could be used to determine the production rate of an output protein, in this case GFP in yeast, could be used to determine a "real" value for maximum GFP production rate under our own laboratory conditions. To accomplish this, we used ImageJ to identify specific cells in a series of images that ranged from maximum GFP output to virtually no GFP output. We tracked these cells throughout the image series and used ImageJ selection tools to give us an average brightness, or simulation of GFP output, per cell per image. We used these brightness values to generate a plot in MATLAB that showed a rough value of how the GFP output in our cells increased over time. Applying a best fit curve to our data, we found that a cubic equation generated the closest fit to our data. These data can be seen in Image 4 to the left.
The equation that best represented our data was:
y = -0.011x3 + 0.4x2 - 2.1x + 1.5
To find the maximum rate of GFP production, we took the derivative of our best fit equation. This gave us the equation:
y' = -0.033x2 + 0.8x - 2.1
Plotting this equation in MATLAB gave us Image 5 on the left. By taking the derivative of our quadratic, we were able to find the maximum GFP production rate of 2.75 mM/hour.
Ideally, the GFP production rate measured by this method could be entered as a value for GFP rate of synthesis in the Ceroni et al. model.
-If we were to make the switch dependent on a tumor antigen input, the system could potentially detect a cancer cell and switch on, at which point it could produce synthetic antibodies as outputs, which could potentially speed up the host's immunologic response. Additionally, researches could experiment using different output proteins to see which ones would be most effective in neutralizing cancer cells while reducing damage to neighboring cells.
1. Brown, W., Ralston, A. & Shaw, K. (2008) Positive transcription control: The glucose effect. Nature Education 1(1). Web.