Sports-related mild traumatic brain injury (mTBI), commonly referred to as concussion, affects anywhere from hundreds of thousands1 to millions of people every year2, leaving patients at increased risk for future brain trauma and a variety of cognitive issues. Concussions of any kind are dangerous, but mild concussions are arguably more dangerous since they routinely go undetected3. The main reason for this lack of detection is the subjective state of concussion testing: there is no reliable, evidence-based approach for concussion detection, even among professional sports teams4. The tests that can reliably detect concussions are expensive, inaccessible, and time consuming, so are only used in extreme cases5. For patients with severe enough impediments to get into the hospital, or whose coaches are diligent about administering concussion tests, traditional concussion tests can successfully detect concussions. Many borderline cases of concussion are pushed back onto the field, leaving them at greatly increased risk for future concussions and brain injuries6.
Our project seeks to fill this diagnostic blind spot with a rapid, biomarker-based test for concussion produced using synthetic biology techniques. The core of our project relies on split protein complementation systems which give a clear experimental readout, such as fluorescence or bioluminescence. Our proteins can be expressed in E. coli without using exotic techniques, which drastically reduces the cost of our system relative to existing products on the market. This straightforward approach also allows us to rapidly iterate on our designs as we get results from our experiments, since changes are just a matter of tweaking base pairs instead of buying new equipment or radically shifting mechanisms.
We are creating a six component system out of three major parts, a binder, a linker, and a split complementation system. These are grouped into two major complexes, each of which contain one binder, a linker, and a complementary half of a split complementation system. The binder will bind to our biomarker, giving us a molecular “hook” and forcing half of the split complentation system closer to the biomarker. The linker will be used to physically distance the binder from the substrate biomarker and the split complementation system in order to avoid steric hindrance and to allow our binders to bind as well as possible. The complementation system consists of a split protein (that produces a signal output) with each fragment connected to a different binder/linker part, resulting in two complexes, each with one binder and protein fragment with a linker connecting the two. Alone, these two systems produce no signal because the protein remains in two fragments, but when both complexes successfully bind to our biomarker of interest, the two halves of the complementation are forced together and a signal is produced.
We are focusing our efforts on a single promising biomarker, UCH-L1. This protein is introduced into biofluids (e.g., saliva, serum, plasma, blood) very quickly following an potentially concussive impact7. It has successfully been expressed in E. coli8, which allows us to express both our biomarker and detection system in the same cell. The advantage of this approach is that we can test hundreds of combinations of linkers and complementation systems in the same experiment and then select the most effective system based on the strength of the detection signal.
A UCH-L1/GFAP sensor has already been approved by the FDA for mTBI detection, the Abbott i-STAT Alinity analyzer. We believe that our project has the potential to be cheaper, more accessible, and easier to deploy, since no expensive equipment is required to deploy it. The Abbott i-STAT Alinity analyzer costs $10 0009 for the device itself and $16 per i-STAT cartridge. We haven’t costed out our system yet, but it is likely that a single dose will cost less than a single i-STAT cartridge.
In order to isolate variables when creating our system, we also researched binder controls. These binder controls are extremely well-characterized binders for common proteins, which will allow us to test if the rest of our system (linkers and complementation system) work, even if our binder doesn’t properly bind to the biomarker. The binder controls we decided on were for HCG10,11,12 and GFP13 dimers, since they are both extremely common in research and have a wide variety of high-affinity binders.
We chose a split complementation system to convert binding activity into a detectable signal. A split complementation system consists of a reporter protein split into two fragments, which produces a signal when recombined. The main ideas to keep in mind when using a split complementation system are the chosen split site and the binding affinities of the two halves.
A good split site allows the reporter to regain activity (i.e. create a detectable signal) upon recombination and remain inactivated when not combined. A poor split site can create a split protein which does not recombine properly and never produces a signal, or a protein which produces a signal without being docked to its complementary half.
The binding affinity of the two fragments must be kept low, in order to avoid recombination without the biomarker present. When binding affinities are sufficiently low, the only time the split protein will recombine is when the two halves are forced into close proximity by the two binding proteins.
Our goal was to create an easy to use, rapid test to be deployed immediately after a patient suffered a potentially concussive impact. We sought to find a reporter that would produce a signal as quickly as possible after the binders bound to our biomarkers, and whose readout could be detected without using specialized equipment. The final list of complementation systems tested includes luminescent proteins, fluorescent proteins, and enzymes that catalyze colorful reactions.
One of the distinguishing factors in our project is the use of combinatorial golden gate assembly. The advantage of this system is that we can test hundreds of combinations of binder-linker-complementation system complexes in a single experiment, without having to prepare each one individually. Assembly is a two step process; combining the four cassettes: backbone, binder/linker, linker/binder, and complementation systems with the riboJ spacer — using BbsI, then combining 4 cassettes to form plasmids using BsaI.
The full system will contain the following parts:
Combinatorial assembly allows us to mix and match cassettes A, B and C and transform the resulting plasmids into our E. coli. We ordered the sequences for our cassettes, amplified them, and thermocycled with Bbs1, then Bsa1 (taking care to keep UCH-L1 and GFAP binders in separate systems). We ordered the sequences for our cassettes, amplified them, and thermocycled with Bbs1, then Bsa1 (taking care to keep UCH-L1 binders and binder controls separate systems). This left us with a single tube of plasmids containing hundreds of different combinations, but ensured that the complementation systems were always with their synonymous half (N and C terminal matching), and also allowed us to control for which biomarker was accounted for based on binder pairing.
We then transformed these plasmids into 10-B E. coli from NEB, and plated on media containing chloramphenicol and kanamycin, as well as any ligand required for activity by the complementation systems. Any colonies present will contain our biomarker and the detection system, so we can actually detect the readout from inside the cell itself. We can then examine the strength of the readouts to determine the best combinations, culture the ones producing the strongest signals, extract the plasmids, and then have them sequenced in-house.
The only non-standard component of our system to be aware of is the beta-lactamase complementation system. It is a component of antibiotic resistance, notably to ampicillin and carbenicillin. In order to avoid interference with our system, we made sure to use backbones with resistance to different antibiotics.
Once we’ve identified the most effective combinations, we can create them in a cell-free system, which makes our system intrinsically safe and easy to reproduce. Once synthesized in a cell-free system, or even purified from E. coli, our system is ready to be combined with a sample from a potentially concussed patient. We suspect that blood serum will be the most reliable substrate, but based on previous research whole blood (i.e. from a finger prick) or saliva should contain detectable concentrations of our biomarkers14.