"
Team:Peking R/Project/Application/BS
From 2011.igem.org
Template:Https://2011.igem.org/Team:Peking R/bannerhidden Template:Https://2011.igem.org/Team:Peking R/back2 Template:Https://2011.igem.org/Team:Peking R/Projectbackground
To further demonstrate the power of genetic softcoding, we applied our previously described methodology to a more complex, nonlinear and milestone genetic device – the bistable switch. Usually, modulating such a device requires overwhelming amounts of workload using conventional screening approaches. When modulating with genetic softcoding methodology as depicted below, however, the behavior of bistable will be flexibly and quantitatively tuned and predicted.
Figure 1 Detailed scheme of the bistable switch.
The straight arrows and the circles represent genes and their corresponding RBSs respectively. Curved arrows represent the transcription strength and the direction of the promoters. The PRM promoter is activated by its own downstream gene product, CI, while repressed by CI434 whose expression is controlled by the PR promoter. The PR promoter is repressed by CI, thus forming a double-negative feedback necessary for the bistable switch property.
The bistable switch, which was inherited from the iGEM 2007 Peking Team, mainly consistsofa positive feedback loop and a double-negative feedback loop. The expression of two mutually repressingrepressorsgenes cI434 and cIare controlled by the promoters PR and PRM respectively.(Figure1)Promoter RM can be activated by CI and repressed by CI434. GFP and mRFP are placed downstream cI434 and cI as reporters of two states, respectively. In the state when CI is dominant, it can activate its own gene’s transcription and repress that of cI434, thus developing and stabilizinga stable high CI/low CI434 state, in which a red fluorescent protein (mRFP)gene co-transcriptedwith CI is expressed. Alternatively, when CI434 is dominant, a stable high CI434/low CI state will be established and GFP co-transcripted with CI434 is expressed to represent thethisstate. Each cell that bears the bistable switch is expected to express GFP or mRFP exclusively. The two states are believed to be stabilized over a long period while under certain circumstances one state may be turned over to the other.
When the bistable switch on pSB1C3 plasmid is transformed into DH5α strain, green colonies and red colonies could beobserved. Interestingly,several mixed colonies could also be observed, which implied the random steady-state characteristic of the bistable switch (Figure 2).A ratiometric of the green colonies to the red colonies(G/R ratio) was calculated on the LB agar plate. We proposed that the G/R ratio is relevant to the translation strength of CI & CI434 genes, meaning that modulating the translation strength of one or more could result in different ratios of G/R under current architecture of bistable switch.
Figure 2 Images of colonies and individual cells bearing the genetic bistable switch.
(A) A red colony and a mixed colony captured by fluorescence stereomicroscope. (B) Individual cell images in the mixed colony captured by laser confocal microscope. Each rod represents a single cell, expressing GFP or mRFP exclusively
Model of the Bistable Switch
We first demonstrate through mathematical modelling that the genetic device we’ve constructed is indeed able to display properties characteristic of a bistable switch within a particular parameter range. Since we are focusing on the translation level, the only variable will be the translation strength of the gene-of-interest(In our case, CI434, with reasons stated below). The original contruct did not display a bistable property because the promoter and the strength of the CI434 gene is too strong compared to that of CI, rendering the system monostable in the high CI434/low CI state. To endow the system with bistability, the strength of CI434 production can be down-regulated by reducing translation strength. To quantitatively describe the system, we employed a set of ordinary differential equations(ODEs) to represent the transcription and translation control of the system(See Appendix)1. Taking into consideration the stochastic nature of the system2, we used a stochastic algorithm to produce the probability distribution of the system in every possible state(Figure 3). If the system is a monostable one(e.g., high CI434/low CI), its states will be tightly distributed around a single peak value. As the translation strength of CI434 gradually decreases, the peak falls and another peak indicating the high CI/low CI434 state starts to grow, until the first peak disappears and high CI/low CI434 becomes the predominant state, bringing the system back to a monostable one. This can also be represented as the proportion of the “green”(high CI434/low CI) state versus translation strength of the cI434 gene(Figure 4).
Figure 3
Changes in distribution of the number of CI434 molecules in response to changes in translation strength(reflected by βSdcro, a parameter in the ODE model that describes translation rate of cI434 gene. For more details, see Appendix).
The surface plot was generated using a stochastic algorithm that simulates the behaviour of 1000 cells. z-axis(height of the surface) correspond to the proportion of cells that express a certain number of CI434 molecules. It can be seen that when translation rate is low, the system’s state is tightly distributed around a low-CI434 state, i.e., the system is monostabe. As translation strength grows higher, another distribution peak starts to appear, indicating a high-CI434 state. The system thus enters the bistable region. When translation rate is too high(βSdcro above 0.6), the system returns to a monostable state again, with only the high-CI434 distribution peak.
Figure 4 Proportion of cells in the high CI434/low CI(green) state derived from data in Figure 3 and experimental results.
The black data points were derived by counting the cells in the arbitrarily defined “green” state(cells with more than 300 CI434 molecules) under each translation rate. It can be clearly seen that as translation strength(Arbitrary △G values determined from ODE parameters and normalized to computationally determined values from the mutation library) increases, the system travels from a low CI434, monostable state to a transitional, bistable state and then to a high CI434, monostable state. The simulation data is normalized and fitted to experimental results of hardcoding(site-directed mutagenesis), shown in red data points(with error bar). The two sets of data showed fair fitness with each other, validating our previous assertion that translation strength regulates the device’s behavior.
Experimentally modulating the device using conventional procedures(hardcoding) involves laborious work as derivation of parameters(e.g., translation strength) in the differential equations require complex experimental measurement and calculation, and large amounts of mutagenesis to produce sequences with desired translation strength may be needed. Our softcoding methodology saves huge amounts of time and effort by obviating the need for numerous rounds of measurement and selection.
Before manipulating the bistable module, a library mutating the RBS of cI434 gene is constructed using site-directed mutagenesis method to verify the conformity of our model with experimental results. Each plasmid of the library is transformed into E. coli DH5α strain separately harboring the bistable switch logic device.(Figure 5)
Figure 5 Construction of bistable switch library via site-directed mutagenesis. Binding sites for forward and reverse primers are schematized.
After growing on agar plates, G/R ratio is calculated for each of the sequences in the library, several images of which are captured by the fluorescence stereomicroscope. (Figure 6) The ability of translation strength to regulate the switch’s behaviour is thus verified.
Figure 6 Experimental results of varying translation rate by altering the RBS sequence upstream of cI434 gene.
Images were acquired by applying excitation light with wavelengths of 470nm(for eGFP) and 580nm(for mRFP) to the agar plate respectively and merging two emission images together. (A)Synthetic RBS with △G of about -3.029kJ/mol. Calculated proportion of “green” colonies is 0.95. (B)Synthetic RBS with△G of about -2.08kJ/mol. Calculated proportion of “green” state is 0.35. (C)Synthetic RBS with △G of about 0.79kJ/mol. Calculated proportion of “green” state is 0.21. Apparently, translation strength indeed has a significant role in regulating the device’s behavior.
Avoiding laborious screening and assays, a down-regulating TPP ligand responsive hammerhead ribozyme TPP 2.5 is applied to the bistable switch modifying the translation rate of cI434 gene to regulate the G/R ratio and a corresponding genetic device is constructed. (Figure 7)
Figure 7 Construction of the bistable switch device carrying the RNA controller(TPP ribozyme, shown as the tuning switch named TPP2.5 in the figure). The rest of the device is the same as that shown in Figure 5.
Since the bistable module’s behaviour can only be quantified using agar plate culture, it is difficult to apply a precise concentration gradient of TPP to the host E.coli. Therefore we set two experiment groups: one without addition of TPP and another with TPP sufficient for full induction of the RNA controller’s funtions(self-cleavage of ribozyme). The experimental results are shown in Figure 8. It can be seen that the group with excess TPP(down-regulated translation strength of cI434 gene) displayed bistability. Therefore, the RNA controller(TPP ribozyme) is indeed capable of regulating the device’s behavior.
Figure 8 Fluorescence stereomicroscopic images of E.coli colonies with and without TPP treatment.
(A) E.coli colonies without TPP treatment(no decrease in translation rate) are all green(high CI434/low CI state), displaying monostability of the genetic device. (B)E.coli colonies with TPP treatment(no decrease in translation rate) are a mixture of green(high CI434/low CI state) and red(low CI434/high CI state) colonies, displaying bistability of the genetic device. (C)Experimental results are mapped to the simulated “green” proportion–△G curve in Figure 4.
The final step then, is fixation of the RBS sequence upstream of cI434 gene with the desired translation rate. Using the forward engineering curve(see the RBS Calculator section for more details), we are able to pick the right RBS sequence from the library and integrate it to the device to endow it with satisfactory bistability.
1. Lou, C., Liu, X., Ni, M., et al. (2009). Synthesizing a novel genetic sequential logic circuit: a push-on push-off switch. Nature Molecular Systems Biology 6, 350.
2. Tian, T., and Burrage, K. (2006). Stochastic models for regulatory networks of the genetic toggle switch. PNAS 103, 8372-8377.
