Engineering Bacterial Finite State Machines

John T. Sexton1, Jeffrey J. Tabor1,2

1. Department of Bioengineering, Rice University, Houston, TX, USA
2. Department of BioSciences, Rice University, Houston, TX, USA
Cells use genetic circuits to make complex decisions such as differentiation, sporulation, or the prioritized utilization of sugars to optimize energy payoffs. The ability to engineer cellular decision making would have widespread applications in basic research and engineering. Finite state machines (FSMs) may serve as a useful tool for engineering complex cellular decisions1. FSMs are a framework whereby decisions
can be explicitly specified as a collection of states and a mapping of the transitions among states. In computer engineering, FSMs are implemented using layered networks of digital logic gates (AND, OR, NOT). Digital-like logic gates can be implemented inside of cells using repressors (or activators) to modulate transcription from promoters. Thus, decision-making circuits can be implemented inside of cells as layered networks of promoters and repressors.
The implementation of complex decision-making circuits inside of cells requires many orthogonal promoter-repressor modules. We use the CRISPRi transcriptional repression system2 as a scalable platform for generating a large number of mutually orthogonal repressors in E. coli. CRISPRi allows for programmed repression of target promoters via the expression of a protein (dCas9) and a small guide RNA (sgRNA) with ~20 base pairs of homology to the target promoter DNA sequence. To ensure that CRISPRi repressors uniquely target only their intended cognate promoters, we developed a standardized procedure whereby an sgRNA target site (operator) is randomly generated in the spacer region of an otherwise invariant promoter. We demonstrate that this approach allows for the construction of families of promoters that can be strongly repressed by partner sgRNAs with complementar y
sequences, while retaining very similar overall transcription rates. We have used this approach to construct 9 highly orthogonal promoters and cognate sgRNAs. We achieve fold repression values ranging from ~25x to several hundred or more, and present a set of 7 promoter-repressor units with minimum orthogonal range of 50x (as defined by Nielsen et al.3). We also extended the above design to incorporate a second input sgRNA, resulting in NOR logic.
Working towards more complex logic circuits, we next incorporated one NOT gate and three
NOR gates to construct a 2-to-1 multiplexer circuit (MUX) in E. coli. The MUX is a signal selection device commonly used in electrical engineering which selects one of several possible input signals and forwards it as the output signal. In a biological context, the MUX allows us to assess the activities of two different promoters using a single reporter gene (e.g. GFP) based on the presence of an input selector signal such as a chemical inducer.
In addition to the combinatorial circuits (i.e. circuits whose outputs depend only on the current inputs) we have constructed to date, we also plan to use the CRISPRi system to construct sequential digital logic circuits with feedback and memory (e.g. toggle switches and flip-flops). Together, these technologies will be leveraged to implement finite state machines inside of bacteria.

Funding Source: Supported in part by the National Science Foundation Graduate Research Fellowship Program under Grant No. DGE-0940902.

References: [1] K. Oishi, E. Klavins, ACS Synth. Biol. 3, 652-665 (2014). [2] L. S. Qi et al., Cell 152, 1173-1183 (2013).

[3] A. A. Nielsen, T. H. Segall-Shapiro, C. A. Voigt, Curr. Opin. Chem. Biol. 17, 878-892 (2013). [4] A. A. Nielsen, C. A. Voigt, Mol. Syst. Biol. 10, 763 (2014).

[5] B. C. Stanton et al., Nature Chem. Biol. 10, 99-105 (2014).