Team:US AFRL CarrollHS/Model
Overview
The goal of the modeling for our project is to determine what ribosomal binding site (RBS) provides the optimal amount of CheZ expression, as too much CheZ can negatively impact the microbes chemotactic ability, but a high enough concentration is required to initiate chemotaxis. Although the modeling below uses arbitrary units and therefore only useful for determining relative amounts of CheZ produced, it provides a good idea of which RBS’s result in much too little protein expression. Additionally, in the future, we can measure the protein expression obtained with one of the RBS’s and then use the modeling to predict the absolute production that each of the other RBS’s would yield.
Method
We modeled the effect of RBS strength on expression of CheZ, the protein that causes our engineered microbe to move in a straight line. We used COPASI software for our modelling, and the differential equations we used are shown below. Data for relative RBS strength came from the iGEM registry. The RBS strength was changed in the model by changing k1R3.
Reactions in Model
R1) (Transcription) DNA -> mRNA + DNA
R2) (Degradation of mRNA) mRNA -> mRNA0
R3) (Ribosome binding to mRNA) mRNA + ribo = mRNA_ribo
R4) (Translation) mRNA_ribo -> peptide + mRNA_ribo
R5) (Degradation of peptides) peptide -> peptide0
R6) (Maturation) peptide -> protein
R7) (Degradation of proteins) protein -> protein0
Kinetic Constants
Km - Michaelis-Menten constant
VR1 - Max rate of reaction for reaction 1
k1RN - Rate constant for the forward reaction N
k2RN - Rate constant for the reverse of reaction N
Vc1 - Volume of the compartment
Equation (1) models transcription and uses Michaelis-Menten kinetics. The Law of Mass Action was used for Equations (2), (3), (5), and (6), which describe the rate of change of the concentrations of the peptide, protein, mRNA-ribosome complex, and ribosome respectively. Equations (4), (7), and (8) are also based on the Law of Mass Action and account for the degradation of mRNA, peptides, and the proteins respectively. The initial concentrations of DNA and ribosomes were set at 1, and all other initial concentrations were 0.
Results and Discussion
Fig 1. Levels of CheZ expression for RBS 30,31, 32, 33, 34, and 64 using Data Set 1
Fig 2. Levels of CheZ expression for RBS 29, 32, 33, 34, and 35 using Data Set 2
The results from the first data set indicate that RBS 33 provides very little expression of CheZ, and is most likely not suitable for our purposes. RBS 30, 31, 32, and 64 all provide relatively high amount of expression, with RBS 34 resulting in the greatest expression of CheZ. The amount of CheZ expressed when using RBS 31 falls between the two extremes.
The second data set, like the first, indicates that RBS 32 and 34 yield relatively large expression of CheZ. RBS 34 and 35 (34 is covered by 35 in the graph) result in the greatest production of CheZ of the five RBS’s in the set, and RBS 32 and 29 yield slightly less CheZ than RBS 34 and 35. Relative to RBS 34, RBS 33 provides an even lower level of expression in Data Set 2 than in Data Set 1.
Since the data from the modeling can only be used to determine relative amounts of CheZ expression, we cannot determine with certainty which RBS would provide the optimum level of CheZ expression for our engineered microbe. However, as RBS 33 provided very low levels of expression in both data sets, it is unlikely RBS 33 would be suitable. Furthermore, RBS 31, which was the one used in our project, does not provide an extremely high or low level of CheZ, making it a likely candidate as a viable RBS for our project.
While the modelling by itself does not give any information concerning the absolute levels of CheZ expression obtained, experimentally determining the amount of CheZ produced when one of the RBS’s is used would allow us to use the model to make predictions concerning the absolute levels of CheZ expression for the other RBS’s as well. This would be particularly convenient with RBS 31, as we have already created a construct with RBS 31 and would only have to measure the amount of CheZ produced.
Modeling Collaboration and Attributions
NUS-iGEM
Last year, NUS-iGEM had modeled our project for us, so we began by asking them about their methods. We spoke with Russell and Marcus of the NUS-iGEM about their process for modeling and they gave some tips for modeling our own. They gave us some of their techniques for creating a model using experimental data and spoke about the applications about our project. When modeling our project, we referenced their parameters and equations from their project last year.
George Smith
We next spoke to George Smith, a member of the 2017 CLSB-UK iGEM team. He provided feedback on our proposed modeling ideas, outlined how he used known parameters to model, and explained some of his methods.
Dr. Yaroslav Chushak
One of our mentors pointed us towards Dr. Chushak, the modeler in one of our labs. We were able to meet with him several times so he could explain the process of how to model our project. He guided us to help us find the parameters we needed, along with walking us through his process in detail and providing feedback in detail. He also reviewed and provided feedback on our modeling write up.