"

From 2012.igem.org

Motivation

Synthetic biology has grown as a field over the past few decades to feature increasingly complex circuit topology, modular parts, and a myriad of application spaces. However, the increase in complexity is limited by the number of specialized parts available to the community. The proteins used for cellular processing must be orthogonal to one another and cannot place too high a metabolic load on the host cell.

At the same time, a field of similar origins, nucleic-acid based molecular computation, has seen an exponential growth in the ability to make more and more complex structures and logic circuits. In particular, recent publications feature square-root calculators and neural network implementations using strands of DNA and their natural thermodynamic reactions to process information and make calculations.

We are motivated by the advances made in both synthetic biology and molecular computation and aim to bring the best of both together into one project.

Nucleic-acid based circuitry has several advantages over traditional transcription-based topologies:

• a much smaller nucleotide footprint
• no need for specialized parts
• a huge combinatorial space which is easily parametrized
• minimal metabolic load; DNA parts don’t need to be translated into proteins
• direct interface with mRNA, miRNA, etc.

By using mammalian cells, we have access various levels of regulation, for example being able to use the RNA-interference (RNAi) pathway.

Let's look at a potential synthetic biology application: detecting cancer state in mammalian cells. What would this require?

You can imagine the first step, a cancer cell sensor, can be achieved by creating an mRNA sensor. The state of the art circuit with this function, (1 Multi-Input RNAi-Based Logic Circuit for Identification of Specific Cancer Cells. Xie et al. Science 2011) requires at least five composite parts for sensing high and low mRNA concentrations. The next part would need to process information from these five separate inputs, invert some of it, and send a signal to up regulate fluorescent protein production. This would require another set of promoters for each sensed mRNA and a repression system to produce the correct logic. The last step would be the induced expression of the signal protein; another unique promoter and protein pairing.

We'll compare traditional, promoter-based synbio logic with a novel strategy from the field of DNA computing: strand displacement.

Parts required (see below for explanation):

• 5 input sensor modules
• 5 processor modules (one for each input)
• 1 actuation module

Promoter-based logic Maximum number of promoters found in published synbio circuits, over time.	Strand displacement-based logic Maximum number of gates found in published strand displacement-like systems, over time.
Promoter-protein pairs required: 11 (one for each module) Max achieved in literature: 6	Strands of nucleic acid required: 80
Size of each promoter-protein pair: ~1,600 bp	Length of each strand: ~40 bp
Total size of circuit: ~17,600 bp	Total size of circuit: ~3,200 bp

This hypothetical circuit requires at least eleven unique promoters and proteins to function. At 1600 nucelotides per composite part, that means inserting 17,600 new bases into the cellular DNA. As shown in the graph above, the current maximum number of promoters in a cellular circuit is six.

Clearly, a novel method of information processing is needed if we want to create complex circuits in vivo. For inspiration, we turned to "Scaling up digital circuit computation with DNA strand displacement cascades." Qian, L., Winfree, E. Science 2011. That research shows that it is possible to use the method of DNA strand displacement for complex circuits in vitro. The team built a circuit which calculated square roots with inputs of different short DNA strands representing binary numbers. The same hypothetical cancer sensing and highlighting circuit, designed using the strand displacement motif, requires only 3200 nucleotides of coding, creating over eighty strands which are roughly 40 bases in length.

Integrating this into a ceullar system is non-trivial, but possible.