Selection is a key mechanism of evolution like Darwin defined it. Biodiversity and life as we know it in all its complexity is scientificly not explicable without it. We wanted to design a system able to generate an Evobody that can gain affinity towards any target protein it is confronted with. So what better way is there to design our system than copying one of evolutions core principles and letting the bacteria with the best Evobody emerge by itself?
Having a library of Evobodies as well as a mutation system is a great way to guarantee a high variety of binding proteins. However, there is still the need of a screening method to separate Evobodies that show affinity to the target protein from those that do not. As we wanted our whole system to work in vivo utilizing directed evolution we settled for three different approaches to screen for functional Evobodies:

The bacterial two-hybrid system, our main approach, consists of two fusion proteins and a reporter gene that grants selective advantage. One fusion protein is the target protein fused to a DNA-binding domain while the other comprises the Evobody consisting of a binding protein with a RNA polymerase subunit. The DNA-binding domain that is connected to the target protein binds upstream of the reporter gene, in our case coding for beta-lactamase. If there is a certain affinity between the Evobody and the target protein, the polymerase subunit of the Evobody recruits the full enzyme and subsequently the beta-lactamase is expressed (Badran et al. 2016). Cultivated in an increasing concentration of ampicillin, only those bacteria expressing enough beta-lactamase will prevail. Since the expression of beta-lactamase correlates to the affinity between Evobody and binding protein we can use the best grown bacteria to determine the sequence of the Evobody which then could be used for production.

The split-protein approach is also based on beta-lactamase as a resistance to ampicillin. The general idea is, to split a reporter protein into two non-functional subunits. Then, a translational fusion of one half of the split-protein respectively and one of the two allegedly binding proteins is created. In case of successful binding between the proteins, the fused split-protein halves should therefore restore the functional reporter protein. As afore mentioned we chose beta-lactamase as a reporter, as it allows the application of selective pressure. Both, the Evobody and the target protein are fused to a different half of the split-beta-lactamase. In case of a successful binding Evobody the bacterium in concern should be granted an ampicillin resistance due to a restored and functional beta-lactamase (Galarneau et al. 2002).

The TAT-hitchhiker selection works on a similar premise as the split-protein. Yet, in this case the beta-lactamase is not divided into two subunits but instead its natural export signals are removed. this modified beta-lactamase forms a translational unit with the target protein whilst a signal peptide for the Tat (Twin-Arginine Translocation) pathway is attached to the Evobody. Only in case of a funtionally binding Evobody to the target protein is the beta-lactamase exported into the periplasm where it can counteract the ampicillin. Since the Tat pathway only exports correctly folded proteins, this approach also provides an option to check if the proteins are expressed as we predicted (Waraho, Dujduan & DeLisa, Matthew P. 2009).

In summary, we wanted a selection system to screen for Evobodies with high affinity to the target protein while being functional in vivo. We chose three approaches, of which the bacterial two-hybrid system proved to be the most promising system.

• Badran, Ahmed H.; Guzov, Victor M.; Huai, Qing; Kemp, Melissa M.; Vishwanath, Prashanth; Kain, Wendy et al. (2016): Continuous evolution of Bacillus thuringiensis toxins overcomes insect resistance. In: Nature 533 (7601), S. 58–63. DOI: 10.1038/nature17938.
• Galarneau, Andre; Primeau, Martin; Trudeau, Louis-Eric; Michnick, Stephen W. (2002): Beta-lactamase protein fragment complementation assays as in vivo and in vitro sensors of protein protein interactions. In: Nature biotechnology 20 (6), S. 619–622. DOI: 10.1038/nbt0602-619.
• Waraho, Dujduan; DeLisa, Matthew P. (2009): Versatile selection technology for intracellular protein-protein interactions mediated by a unique bacterial hitchhiker transport mechanism. In: Proceedings of the National Academy of Sciences of the United States of America 106 (10), S. 3692–3697. DOI: 10.1073/pnas.0704048106.