Components of the Genetic Circuit
Once we had come up with the idea for what we wanted to achieve, we had to figure out how to practically make it happen. First of all, we needed to find a mechanism for introducing mutations to the binding site of the receptor, since without this component there would be no evolution occurring to the binding site. Secondly, we needed to find a mechanism that could be triggered upon successful target binding to give a detectable signal, indicating success. Lastly, we needed to choose a sequence for the basic receptor we would introduce into the circuit, as well as a way of extracting the mutated sequence once it had successfully bound to the target.
Component 1: The Mutation Mechanism
What we wanted to achieve with this component was to create a way of inducing random mutations to the sequence encoding the binding domain of our base receptor. We needed the mutations to be random, in order to simulate an evolutionary process, but needed to control the region to which they occurred so that the rest of the organism’s genome did not get disrupted. Initially, we took inspiration from Guilinger, Thompson & Liu (2014) and intended on using dCas9-Fok1, a deactivated Cas9 unit coupled via a flexible linker to a Fok1 endonuclease.
2 Fok1’s need to dimerise in order to make a double stranded break in the DNA. Since they are coupled to Cas9, we could control the regions that would be susceptible to the Fok1 dimers by making the sgRNA in the Cas9 units specific to binding sites in the receptor binding sequence. Once a Fok1 dimer has made a double stranded break, non-homologous end joining (NHEJ) would re-join the broken DNA. This process is incredibly error prone and introduces indel mutations to the join site, which in this case is exactly what we wanted. Of course, this also meant we needed to choose a host organism that is capable of NHEJ, but preferably with very error-prone machinery.
We also needed a way of ensuring that the Cas9 bindings were located only in the sequence encoding the receptor binding site. Additionally, we faced a problem with making sure that the binding sites did not become mutated, as Cas9 would no longer be able to bind, without this limiting the mutation potential of the entire sequence. After much thought, we came up with an elegant solution to both of these problems: introns! By placing the Cas9 binding sites inside artificial introns that could be inserted at regular intervals into the receptor sequence, we could make sure that the binding sites didn’t get mutated, and that they were spliced out and removed from the mRNA transcript so as to not interfere with the structure of the receptor.
This brought us onto another reason to re-consider the host organism. Since we now needed something capable of NHEJ and splicing out introns, we finally decided to use S. cerevisiae, or baker’s yeast. This organism is easy to culture and transform with plasmids and has endogenous cellular machinery for error-prone NHEJ and intron removal, everything we needed in a host cell.
It wasn’t until a few weeks into the project that we came up with a second possible mutation mechanism. Upon being introduced to the enzyme Active-Induced Cytidine Deaminase, or AID, we decided it would be worth investigating this alongside the Fok1 to determine which would give us the best results.
AID induces mutations through a very different mechanism to Fok1. Instead of making a double stranded break, it removes an amine group from cytosine and turns it into uracil, resulting in a point base-substitution mutation. We decided than instead of changing the whole system based around intron-binding, we would link the AID with flexible linkers to its own dCas9 complex. AID does not need to dimerise in order to work, so the idea was that a single dCas9-AID complex would bind and mutate the surrounding cytosine residues, but not those that are in the site bound by the Cas9.
Component 2: Target Binding Signal
The second component in the genetic circuit we needed to create was a mechanism that could give a signal to indicate successful binding of the target. This would signal to us that the system had worked, and that the receptor had evolved an affinity for the target molecule, so that we could proceed with the sequence extraction. After many hours of research and thinking, we finally found a paper by Li, Xu, Chupalov & Marchisio (2018) about something called an anticrispr protein. This is exactly what it sounds like: it’s a protein that can be activated by an active receptor and bind to the dCas9 PAM sequence to prevent it binding to the DNA. Our idea was to have the anticrispr become active only when the base receptor is activated by target binding, whereupon it would deactivate the dCas9’s in the dCas9-Fok1 complexes to stop the mutations.
But what about the signal?
This was where it got really complicated, but also really interesting. We planned to introduce another plasmid containing the gene for green fluorescent protein, or GFP. The promotor of this GFP gene would containing a binding site for a single Cas9-Fok1. No cleavage would occur because for that, 2 complexes would need to bind. Instead, the gene would be supressed by the bound dCas9, essentially acting as an inhibitor. When the receptor activated the anticrispr protein, the bound dCas9-Fok1 would be displaced and the GFP would be transcribed, giving a green fluorescent signal that we could detect.
Component 3: Base Receptor
The original idea was to use the VEGFR-1 receptor as the base receptor sequence. This receptor is capable of recognising and binding to VEGF-A and -B, but not -C. As a proof of concept, we planned to start by trying to mutate the VEGFR-1 receptor just enough to be able to recognise the VEGF-C ligand. VEGF receptors are homodimers: 2 receptors need to come together and form a dimer when the ligand binds in order to trigger a signalling cascade. When the VEGFR-1 receptor was inactive and unbound, it would be attached to the inactive anticrispr protein. When 2 receptors were able to bind to a target, the anticrispr would be cleaved off at a TEV cleavage domain on the intracellular side and would therefore become activated and released to act upon dCas9.
After talking extensively with science representatives from our sponsor Promega, we stared to consider using NanoLuc® Binary Technology (NanoBiT) as our signal component instead of anticrispr. The NanoBiT system consists of 2 subunits, Large BiT (18kDa) and Small BiT (1.3kDa). These units can be independently fused to 2 target proteins to study their intereaction. If the 2 target proteins interact, the Small and Large BiT combine to give off a fluorescent signal
NanoLuc® Luciferase: One Enzyme, Endless Capabilities, 2019
Our first idea was to fuse one NanoBiT unit to each VEGFR-1 receptor monomer, so that when they dimerised upon binding to the target molecule, a fluorescent signal would be given off. However, we quickly realised this was a flawed approach, since the VEGFR-1 monomer have the same sequence it would be impossible to make sure they were both fused to different NanoBiT units. This led us to consider changed the base receptor altogether to one that does not require dimerization. We could then fuse the Large BiT to the receptor sequence and use the Small BiT as the actual target molecule for a proof of concept experiment. When the receptor bound to the target, the Small and Large BiTs would combine to give off a signal to indicate successful binding. We could still use the activation of anticrispr to stop the mutations but would not need to complicate the genetic circuit by introducing GFP in a separate plasmid.
References
Guilinger, J., Thompson, D., & Liu, D. (2014). Fusion of catalytically inactive Cas9 to FokI nuclease improves the specificity of genome modification. Nature Biotechnology, 32(6), 577-582. doi: 10.1038/nbt.2909
Li, J., Xu, Z., Chupalov, A., & Marchisio, M. (2018). Anti-CRISPR-based biosensors in the yeast S. cerevisiae. Journal Of Biological Engineering, 12(1). doi: 10.1186/s13036-018-0101-z
NanoLuc® Luciferase: One Enzyme, Endless Capabilities. (2019). Retrieved 2 October 2019, from https://nld.promega.com/resources/technologies/nanoluc-luciferase-one-enzyme-endless-capabilities/
