Developing a library of recognition proteins using FimH as a protein scaffold

Abstract

High specificity recognition proteins, like antibodies, are important biological tools with many applications, including diagnostics, therapeutics, and imaging. However, there are many challenges with antibody recognition proteins, including cost of manufacturing and limited tissue penetration due to antibody size. Alternative protein scaffolds solve many of the problems associated with antibodies, but none have activatable binding affinity. Activatable binding would increase specificity by allowing tunable binding and release for therapeutic, diagnostic, and imaging applications. Bacterial adhesion protein FimH switches from high- to low-affinity binding of its ligand mannose by mechanical activation, and the binding of FimH may be altered without affecting the allosteric regulation of binding, making FimH a potential activatable protein scaffold. In this work, the main hypothesis tested was that the binding domain of high-affinity FimH could be mutated without altering the conformation. To test this, I randomized 9 consecutive amino acids in one of the binding loops of FimH-Hi, which is locked into the high-affinity conformation by a single point mutation. I created a highly diverse bacterial display library of 4.4x10<super>3</super>-2.0x10<super>4</super> variants with incidence of stop codons and frame-shifts not significantly higher than expected. A significant percentage of the library is expressed and maintained in the high-affinity conformation despite extensive randomization, suggesting that FimH may be a good protein scaffold. The results presented here demonstrate that these methods can be used to randomize all three CDR loops on the FimH-Hi lectin domain.