Modeling

Achievements

We built a new model for protein synthesis including DNA, RNA and protein species
We predicted performance of our Basic Part BBa_K2588004 in a generic circuit setup
We based our choice of a constitutive promoter for our NOT-gate of MlcRE, BBa_K2588002, on insights gained from this model

Modelling in synthetic biology serves the purpose to predict functions and behaviours that are difficult or impossible to characterize experimentally, and to drive design decisions that otherwise would take much time to gather information on. In our project, we implemented modelling for both aspects.

Shigella flexneri icsA is coding for an outer membrane protein and virulence factor, and is highly regulated by a unique mechanism involving a short noncoding RNA (ncRNA) which is complementary to icsA 5’ untranslated region (5’ UTR), called RnaG120 hereafter. When binding to nascent icsA mRNA, RnaG120 induces stem loop formation, leading to termination of transcription before icsA coding sequence is transcribed1.

While RnaG120 was studied extensively in situ, to our knowledge it has never been used in an engineering context. We identified the vast potential of RnaG120 as a compact, extremely energy efficient NOT-gate. Assuming that binding of RnaG120 to icsA 5’ UTR is sufficient to induce stem loop formation and termination of transcription, expression of any gene cloned downstream of icsA 5’ UTR could be inhibited by expression of RnaG120. Based on this assumption, we created BBa_K2588002 as a proof of concept of a general RnaG120-based NOT gate.

BBa_K2588002 – An RnaG120-based NOT gate of glucose-inhibited promoter MlcRE

To build iGEM registry’s first ever transcription regulator that is induced by glucose, we employed RnaG120 to create a NOT-gate of glucose-inhibited promoter MlcRE (BBa_K2588000). It is comprised of multiple subparts, in 5’ to 3’ order: BBa_B0015 terminates transcription of RnaG120. A constitutive promoter provides constant gene of interest (GOI) mRNA transcription. icsA 5’ UTR, which is complementary to RnaG120, acts as 5’ UTR for any GOI downstream of BBa_K2588002. MlcRE is cloned in opposite reading direction, downstream of icsA 5’ UTR, activating transcription of RnaG120. Downstream of BBa_K2588002, any gene of interest could be placed (Fig. 1).

**Fig. 1 Map of BBa_K2588002.** Parts in order are: terminator BBa_B0015, a constitutive promoter, icsA 5’ UTR BBa_K2588004, MlcRE BBa_K2588000 (coded backwards). Note that only one DNA molecule is depicted, but annotated in two rows to accommodate for RnaG120/icsA operator, which has one function in each reading direction.

The Objective

We assume that the performance of BBa_K2588002 heavily relies on the ratio of transcribed mRNA and present RnaG120 depending on current induction. The easiest way to optimize this ratio is to select a constitutive promoter based on its strength relative to MlcRE. The Anderson collection offers a wide selection of constitutive promoters which have been characterized extensively. Testing all of them to select the one best suited for our application would be extremely laborious. Therefore, we decided on building a model that is able to predict the GOI expression based on constitutive promoter strength and glucose input.

Our aim was to select the Anderson promoter which provides the best GOI expression signal to noise ratio.

Building a model of RNA/RNA interactions using iBioSim

iBioSim is a java-based desktop application designed to create and analyse genetic circuit models2. Models built with iBioSim simulate cellular agents commonly used to predict kinetics of genetic circuits. Classic genetic circuits are built with regulated promoters inducing production of proteins. While iBioSim computes both transcription and translation using internal mRNA, it does not support manual editing regulatory functions of mRNA. BBa_K2588002 is a genetic circuit that performs logic purely on RNA level. Therefore, iBioSim does not support models of our application of the shelf.

As a workaround, we built our model to simulate both mRNA and RnaG120 as protein-level species behaving like RNA by setting stoichiometry of production, which usually is used to model the number of proteins produced from one mRNA molecule before its degradation, to 1. To model translation, icsA mRNA was set up to activate an implicit promoter acting as ribosome binding site to induce translation of a GOI. The transcription-terminating function of RnaG120 was modelled as complex formation of RnaG120 with icsA mRNA. The resulting complex would not activate translation of GOI, accounting for transcription termination before transcription of GOI coding sequence (Fig. 2).

**Fig. 2 Setup for kinetic model of BBa_K2588002.** MlcRE was set up as activateable promoter with an activated RNAP binding equilibrium smaller than its basal RNAP binding equilibrium to account for the repressing function of glucose. Constitutive promoter and MlcRE induce transcription of GOI mRNA and RnaG120, respectively. GOI mRNA induces translation of GOI. RnaG120 and GOI mRNA undergo complex formation. The resulting complex does not induce translation of GOI.

Despite being a workaround, this model should kinetically perform exactly like the real construct, given that correct values are used for the relevant parameters of both constitutive promoter and MlcRE.

Genetic circuit models rely on many measured parameters to be accurate. We previously characterized MlcRE in BBa_K2588039, comparing MlcRE-induced GFP expression to InterLab positive control BBa_I20270, in which strong Anderson promoter BBa_J23100 drives GFP expression. Using this data, relative strengths of all Anderson promoters compared to MlcRE were computed and used as input values for the model. RnaG120 was extensively characterized by Tran et al., including binding behaviour of RnaG120 to icsA 5’ UTR. All model input values are summarized in Table 1.

Table 1: User-set values of BBa_K2588002 kinetic model

Species	Parameter	Value	Reference
Glucose	Degradation rate	0.0075	Natarajan et al3
RnaG120	Degradation rate	0.068	Selinger et al4
GOI mRNA	Degradation rate	0.068	Selinger et al4
RnaG120/GOI mRNA complex	Degradation rate	0.068	Selinger et al4
RnaG120/GOI mRNA complex	Complex formation equilibrium	0.7	Tran et al1
GOI	Degradation rate	0.012	Set arbitrarily

Promoters
MlcRE	Basal production rate	0.81	MlcRE characterization
	Activated production rate	0.19	MlcRE characterization
	Stoichiometry of production	1	model-intrinsic
Constitutive Promoter	Basal production rate	1	BBa_I20270 compared to MlcRE in characterization5
	Activated production rate	1	BBa_I20270 compared to MlcRE in characterization5
	Open complex productio rate	varying (0.2-2)	Anderson promoter family6
	Stoichiometry of production	1	model-intrinsic
RBS	Activated production rate	1	model-intrinsic
RBS	Stoichiometry of production	20	Set arbitrarily

Results

BBa_K2588002 was modelled using a hierarchical Runge-Kutta-Fehlberg simulation7. The simulation was run for 1000 time units, with addition of glucose after 300 time units. Glucose was set up to be degraded over time to simulate consumption. Levels of RnaG120, GOI, and RnaG120-mRNA complex was monitored over time (Fig. 3A). The model predicts basal expression of RnaG120 and GOI, as well as complex formation of RnaG120 with GOI mRNA. Due to the RnaG120/GOI mRNA binding constant, RnaG120 is predicted to not fully bind to GOI mRNA. Upon glucose addition, RnaG120 and RnaG120-GOI mRNA levels are predicted to decline before returning to normal upon glucose consumption. Consequently, GOI translation from free GOI mRNA is predicted to raise GOI levels. Since GOI protein stability is set arbitrarily, it declines over time.

Consecutively, constitutive promoter strength was varied between 0.2 times measured MlcRE strength to 2 times measured MlcRE strength. Predicted signal to noise ratio was calculated as the ratio of basal GOI level before addition of glucose and peak GOI level after induction (Fig. 3B). A constitutive promoter with roughly half the strength of MlcRE without glucose was predicted to be optimal for BBa_K2588002.

**Fig. 3 Results of BBa_K2588002 model. A:** Predicted RnaG120, GOI, and RnaG120-GOI mRNA complex levels over time, simulated using a hierarchical Runge-Kutta-Fehlberg simulation. After addition of glucose, RnaG120 and RnaG120-GOI mRNA complex are depleted, and GOI level is raised before returning to basal level after depletion of glucose. B: Predicted signal to noise ratio of GOI expression depending on constitutive promoter strength relative to basal MlcRE.

Conclusions

A constitutive promoter inducing transcription half as strongly as MlcRE without glucose repression was predicted to be optimal for use in BBa_K2588002. Consequently, we chose BBa_J23106 as constitutive promoter.

Our sponsor Integrated DNA Technologies kindly synthesized BBa_K2588002. We assembled it in pSB1C3 as basic part, however we were unable to assemble it with a GFP reporter. Therefore, we could not experimentally verify this model. We are positive that this part would have a tremendous value for future synthetic biology applications, and encourage every team to look further into it.

Team:Hamburg/Model