Team:UCSF/Modeling2

From 2013.igem.org

Revision as of 01:48, 29 October 2013 by Khelmke (Talk | contribs)

Modeling: Conjugation

The primary goal of the modeling portion for the CRISPRi conjugation project is to create a model that will help us identify the behavior of our transferable targeting system and identify critical parameters, given our assumptions. We used this model to check whether our CRISPRi conjugation system would have the desired behavior under biologically relevant parameters.

This system is designed to be transferred from cell to cell, targeting a specific gene of interest within the community. We model the activity of three strains of E. coli, an engineered donor strain containing a conjugative plasmid with genes coding for dCas9 and a gRNA, a recipient strain containing the RFP gene (target gene) in its genome, and upon successful conjugation, a transconjugate strain that no longer expresses RFP as well as the community's total RFP concentration.

Our objective is to simulate this community over time to see if our engineered system causes RFP levels to decrease.


ASSUMPTIONS: While creating the model for our system, we made four assumptions in order to simplify some of the aspects of the model. Many similar assumptions have been made in the literature.
1) RFP, dCas9, and gRNA are produced and degraded at a constant rate;
2) conjugation rate is linearly dependent on the concentration of donor and recipient cells (mass action);
3) CRISPR expression efficiency in a cell is 100%;
4) The growth rates of the three strains are equal.

EQUATIONS

Given these assumptions we have the following equations for the system:
Fluorescent proteins:

In these equations, the amount of fluorescent protein depends on a production term and a degradation term. The parameters for the degradation terms are:

dx – rate of degradation of GFP based on the “low” function
dy – rate of degradation of RFP based on the “high” function

The production term consists of a function that relates inducer and repressor complex to mRNA production rate and conversion factor that relates mRNA production to protein production rate.

px – conversion factor (mRNA/protein) for x (GFP)
py – conversion factor (mRNA/protein) for y (RFP)
The mRNA production functions depend on the amount of inducer and the amount of repressor complex as follows:

There is one function for the low sensitivity promoter (fL) and another for the high sensitivity promoter (fH) . The other parameters for the production functions are:
A – basal expression level of promoter
B – maximal expression level of promoter
k – Activation coefficient for low/high promoter
kR – half maximal effective concentration of R:C
n – hill coefficient for induction.
nR – hill coefficient for repression
I – Inducer concentration
R:C – repressor/dCas9 complex.

The subscripts indicate whether the parameter is a property of the low or high sensitivity promoter (L or H) or related to protein x or y. Now we have to figure out what the equations for the gRNA/Repressor complexes are (Rx:C and Ry:C). The binding and unbinding of the gRNA and dCas9 complex is represented by the following chemical reactions:

Given these chemical reactions, we can write the following equations for the gRNA/dCas9 Complex:

Where KRxf represents the forward reaction rate for reaction (5) and KRxb represents the reverse reaction rate. The equations for the gRNA/dCas9 complex depend on the rate at which the gRNA complex with dCas9 and also the rate at which the complex breaks apart. As mentioned earlier, we made an assumption that the binding and unbinding of the gRNA/dCas9 complex happens much quicker than the production and degradation of gRNAs and fluorescent proteins. Since the binding and unbinding happen very quickly, we evaluate the equations above at steady state (equal to zero) for the given values of the other parameters.

Under that assumption (setting equations (7) and (8) to zero – known as the quasi steady state assumption), we can solve for the complex in terms of the unbound repressor concentrations:

Where the amount of dCas9 available in the system is given by:

The available amount of dCas9 depends on the amount of the two different complexes and also the amount of free dCas9. It has the following parameters:

The equations for the gRNAs depend on the amount of the gRNAs that is produced, the degradation rate, and also the rate at which the gRNA complexes with dCas9. With the quasi-steady state assumption, the terms for complexing with dCAS9 drop out and the final equations for the gRNAs are similar to equations (1) and (2) for the fluorescent proteins:


PARAMETERS

This model has many parameters, so in order for it to be more useful, we need to reduce the number of parameters that are undetermined. To accomplish this, we gathered some values from literature and also did experiments to find other parameters (Table1).

How did we fit parameters? A few parameters are properties of the promoters (A, B, k, n). By determining the dosage response of a promoter to inducer we are able to fit those parameters using a curve fitting tool to get the following values.
How is the hill function affected by the amount of gRNA/dCas9 complex? The previous plot helped us determine a few parameters, but there are still a few parameters we have no values for from experimental data or from literature. The following plots show how the parameter for amounts of the gRNA/dCas9 complex affects the behavior of the model.
From given values for inducer concentrations and amounts of complex, we can calculate the amount of fluorescent protein that should be present. Our model can help us design an experiment that helps us calculate parameters that are still unknown.

How does the model look with our actual “low” and “high” promoters?
If the only change in the low and high functions (FH and FL) is the K values (which determine the sensitivity of the promoters), then we don’t get our desired behavior. However, there are other parameters that might give us the desired behavior for the low and high promoters.
If we set BL to 443.7 and BH to 443.7*1.25, and if we set the half max values to kL = 11.45 and kH=17, the promoters have the following profile:
And they generate the following behavior in the full model:
Thus, based on our model, if certain conditions are met, our synthetic circuit will work as expected.


How does the system change when the hill coefficient is manipulated? In this first plot, the hill coefficients for both the low and the high function are the same number: 2.551. This number is the one we determined from our experimental data.

nL = 2.551
nH = 2.551

We can see that the switch from GFP to RFP is relatively sharp, and that RFP seems to be expressed in higher concentrations of inducer, while GFP is being expressed in lower concentrations. In this next plot, the only thing that changed was the hill coefficient for the high function. In the first plot, it was 2.551, in the second plot, it is 1.551.

nL = 2.551
nH = 1.551

It is similar to the first graph in that the switch between GFP and RFP is sharp. However, RFP is being expressed in both low and high concentrations of inducer, while GFP is being expressed in medium concentrations.