Implementation of a synthetic system to realize placement, detection and reset of orthogonal epigenetic marks in mammalian cells in MATLAB (& SimBiology). A proof-of-principle finite-state machine is also provided.
Note: This is a team-project during the course "Introduction to Biological Computers" in ETH Zurich.
Specifically, we describe the molecular architecture and provide dynamical models of a dCas9-based epigenetic Write module capable of writing orthogonal epigenetic tags on specified loci and in a targetable manner in mammalian cells, a Read module, as well as a novel Reset module.
To construct an orthogonal system, the DNA epigenetic marker of choice that we selected is N6-methyladenine (m6A), which is extremely rare in eukaryotes and an E.coli DNA adenine methyltransferase (Dam) was employed, which catalyzes the methylation of adenines in GATC sites.
- Write module similar to Park et al., however the Dam methylase was fused to dCas9 instead of a ZF protein
- Read module based on a restriction enzyme (DpnI) specific to the sequence GATC, coupled to an activator (VP64) to induce the production of a reporter fluorescent protein (GFP)
- Reset module including an ALKBH1 demethylase and with analogous design to the Write module. The targeting function is not performed by a dCas9-gRNA complex but with a ZF protein, in order to achieve independent targeting by the two modules.
The example finite-state machine is shown in the following figure:
The automaton consists of:
- 2 states (S0, S1), defined as the unmethylated (S0) and methylated (S1) version of the target site
- 2 inputs (I1, I2), encoded as ”pulses” of two gRNAs (gRNA 1, gRNA 2)
The model is constructed by incorporating the Write-Read-Reset modules together with a plasmid (genome integrated) that inducibly produces the second input gRNA (gRNA 2). The production of the Reset module is gRNA 2-dependent, as gRNA 2 drives a dCas9 protein fused to a VP64 transcriptional activation domain to activate expression of the Reset module. To couple the expression of this dCas9-VP64-gRNA activation complex to the current state of the finite automaton, namely unmethylated (S0) or methylated (S1), the synRead protein (of the Read module) leads to the VP64-dependent activation of the fusion protein dCas9-VP64.
The choice was made to represent all the concentrations in molarity, transforming units of molecules/cells by assuming a median mammalian cell volume of 2 * 10-12 L. The plasmids were considered to be integrated inside the genome (two copies/cell) and thus not diluted or consumed by reaction. We assumed that recognition sites for the binding of the fusion proteins ZF-ALKBH1 and dCas9-Dam are the same. The engineered loci contained 60 GATC sites for potential methylation, which is acquired from Park et al. The models have versions with and without dilution, with not genome-integrated species being diluted during every cell division, with a dilution rate taken from Park et al., which corresponds to a doubling time of 20 h.