Design

THE PROBLEMS we had to solve were...

Problem 1

“How to get DNA extracted from a sample pure enough and in sufficient concentration without the need of laboratory equipment”

For solving this problem, we were researching on different novel inventions. One of our favorites was the hand-centrifuge. It was invented by Devin Powell in 2017 as an adaptation of a children’s toy. [1] Another idea was to design and program a heating/cooling-block that is solar powered and uses a peltier-element to change the temperature. This way one could perform experiments that need cycles of changing temperature without the need of a laboratory.

Problem 2

“How to get a readout visible to the naked eye that will only appear if dCas9 is bound to its target DNA”

We designed several solutions:

1. Two different fusion proteins, each has one dCas9 and one domain of a reporter protein dimer. This dimer could for example be Alkaline phosphatase, which converts different colorless substrates to a colorful products through oxidation. The dCas9s would have guideRNAs with target sites next to each other on the genome. If they bind close to each other the monomers of the reporter protein could find each other, form a dimer and become active. Problems with this approach: The dimers could form in solution, giving a strong background signal and the guideRNA target sites need a certain basepair sequence, the PAM region, which might be unlikely to find in close proximity and thereby limit the applicability of our tool.

2. Upon binding of the protein to the target DNA the recognition lobe (REC) and the nuclease lobe (NUC) of dCas9 undergo a strong conformational change. We could design a fusion protein, that depends on that conformational change to become an active protein. Problems with this approach: Very difficult to realize.

TWO PROBLEMS – ONE SOLUTION!

While we tried to solve these two problems independently, to our big surprise, the solution came in a combination of both. It was brought to us by an old friend and college of our teammate Sebastian, Dr. Paul Szekely . He is working in the amazon rainforest in describing and identifying new species of frogs and other amphibians. As a potential stakeholder that could benefit from our project, we got in contact with him and he told us about a new DNA extraction method that he read about. He sent us the link to the paper of Zou et al. [2] and we realized that we had the solution to both of our problems at hand. The method that Zou et al. developed uses a paperstrip on which the DNA binds and is then transferred into a PCR reaction. If we could modify this novel DNA extraction method to be applicable for microbes and human cells (Zou et al. used it for plants) and if we could increase the yield to not only be enough for a PCR reaction but for our purposes we would have both of our problems solved.

The DNA bound to the paperstrip became our solution. This way you would only have to dip a paperstrip into a tube with the DNA, a tube with the dCas9 bound the the reporter and a tube with the reporter substrate. Simple, perfect and elegant. DipGene was born.

New Solution - New Problems

Several smaller subproblems came up with the solution of our first big problem. We were determined to solve them all, because we loved the idea of the DipGene method. This determination sent us back to the drawing board, we had to redesign the project and structure our upcoming work. For this we defined three new subproblems. To solve each of these we had to go through many cycles of planning, testing and redesigning. Please check the lab notebook to follow this process in detail.

Subproblem 1
Understand dCas9 - paper - interaction

Of course our new method would only work if the dCas9 fusion protein would not bind to the paper itself and would still be able to detect its target in the DNA when it is bound to the paper. To investigate this, we studied the interaction of a dCas9-GFP fusion protein with cellulose and nitrocellulose paper in combination with different blocking solutions (please read more about this in our lab notebook, results and experiments section). We tried to analyze this interactions with different methods, like loading the cellulose and nitrocellulose disks treated with different protein blockers into SDS-PAGE gels to check if we could block nonspecific binding of dCas9 to the paper.

Subproblem 2
Understand DNA - paper - interaction

To improve the method developed by Zou et al. (2017) we had to go deeper into understanding the interaction between different kinds of paper and nucleic acids. Therefore, we tried multiple buffers for cell lysis and further binding of DNA to cellulose paper or nitrocellulose membrane. The cell lysis buffers were tailored to epithelial cells from a buccal swab sample and E. coli from the GB05 strain, due to the different composition of the plasma membrane and cell wall respectively. DNA binds to cellulose via electrostatic interactions between its negatively charged sugar-phosphate backbone and the -OH groups of the cellulose, this interactions are mediated by cations (salts) which we add to the binding buffer.

Subproblem 3
Designing a fusion protein

This part proved to be quite problematic and time consuming, since dCas9 is a difficult protein to work with. We were incredibly lucky to get in contact with Aliona Bodganova from the protein facility of the Max Planck Institute (MPI-CBG), she has a lot of experience in expressing dCas9 fusion proteins and could warn us from all the pitfalls that can happen in dealing with this protein. She also provided us with a plasmid from which we could amplify the dCas9 protein for cloning. The first problem that came up in this process was a forbidden restriction enzyme site in the middle of the coding region of dCas9, which we had to remove by site directed mutagenesis PCR. Designing the final construct needed a lot of time and in silico cloning as well, since we needed N- and C-terminal purification tags and a linker between the dCas9 and the reporter. As a reporter we had decided to use HRP, since it is a well-established and robust reporter protein.

In total our fusion protein construct was made of six different parts, which all had to be cloned in plasmids as separate Biobricks and then assembled together one by one. To read more about this process click here.

Amplifying the Signal

Another problem we had to solve was that the DNA concentration will always be relatively low. Since only one dCas9 can bind to one target site, the amount of bound enzyme will be low. Since only one reporter protein is bound to the dCas9, also the signal will be very low. On the drawing board we were designing a thermocycler that can be solar powered. But we were never completely convinced of this idea since it defied our idea of making DipGene as simple and technology-free as possible. We played with many ideas and in the end came up with a simple and elegant solution: We would not use a reporter protein like eGFP, where the amount of signal depends on the amount of protein, rather we chose a reporter that is able to amplify the signal over time. This way we avoided the need for amplifying our sample DNA and could instead amplify the signal itself in the end. The reporter protein we decided to use was HRP.

Choosing the right target

Even if we could make all these separate parts work, we would still need to demonstrate somehow that our protein is active and working and we do not get false positive or false negative results. For the proof-of-principle in microbial cells we decided to target eGFP with a well-established guide-RNA line that Aliona provided to us. By using this we can prove in an EMSA if our fusion protein is able to bind microbial plasmid DNA. Our newly developed method for DNA extraction from microbial cells can be applied in this example as well. eGFP targeting will be examplatory for all potential applications in microbial cells, which could be testing for pathogenic strains, for antibiotic resistances or successful insertion of an insert while cloning.

For testing in human cells we were thinking for an extensive amount of time about which locus to target. There are many disease loci that would be of interest, but for a student project the scope of iGEM it would cause many problems of safety concerns to test for an actual disease gene, furthermore, it would be nearly impossible to get positive controls. And the information about predisposition for a genetic disease should be absolutely private and no team member should feel pressured to give samples for the project. To have a target that would give us easy positive and negative controls, we decided to target the SRY-locus. This is a gene on the Y-chromosome. To be sure that this does not get in conflict with ethical, societal or our own values of equality, we contacted the local transgender community and met with representatives to talk about how we could design our project in a way that would make it very unlinkely or ideally impossible to abuse our project to discriminate against people.

References

[1] https://www.nature.com/news/spinning-toy-reinvented-as-low-tech-centrifuge-1.21273
[2] Zou, Y., Mason, M. G., Wang, Y., Wee, E., Turni, C., Blackall, P. J., Traut, M., Botella, J. R. (2017) Nucleic acid purification from plants, animals and microbes in under 30 seconds. PLoS biology, 15(11): e2003916.