Generating FN3‐Based Affinity Reagents Through Phage Display

Antibodies produced by immunizing animals with foreign antigens have been invaluable tools for various detection methods. While the importance of their utility is widely accepted, there are limitations regarding the production, and reproducibility of monoclonal and polyclonal antibodies (Bradbury, 2015; Jones, 2016). To overcome these limitations, recombinant antibody development has received increasing attention. Antibody fragment or scaffold protein libraries can be displayed on bacteriophage and binders identified through affinity selection (Kehoe, 2005; McCafferty, 2015; Shim, 2016). Because the DNA sequences of isolated, recombinant antibodies are easily obtained, reagents can be expressed in Escherichia coli and their affinity and specificity can be improved through directed mutagenesis to yield high-quality, specific affinity reagents. By this approach, affinity reagents have successfully been generated to recognize a wide range of targets, including cell signaling proteins, membrane proteins, transcription factors, peptides, and post-translational modifications (Kim 2011; Kummer, 2012; Pershad, 2012; Koide, 2013; Horsby, 2015; Jones, 2016; Gustafsson, 2017). Successful generation of affinity reagents is highly dependent on the choice of library (Hosse, 2006). Fibronectin type III (FN3) domain scaffold libraries serve as valuable options. Protein engineering experiments have shown that it is possible to randomize residues within three loops (BC, DE, FG) on one side of the FN3 94-amino acid domain (Fig. 1) without loss of stability or folding (Koide, 1998; Batori, 2002). FN3 sequence variants, also known as ‘monobodies,’ have been selected from phage display libraries that bind tightly and selectively to a wide variety of proteins through these randomized regions, such as Abl (Wojcik, 2010), β-catenin (Yeh, 2013), EphA2 (Park, 2015), estrogen receptor (Koide, 2002; Huang, 2006), Fyn (Huang, 2012), integrin (Richards, 2003), Pak1 (Huang, 2012), Ras (Spencer-Smith, 2017), VEGF-R (Fellouse, 2007), and several other human cell-signaling proteins (Huang, 2015). In addition to its target recognition versatility, the FN3 has many practical advantages. It lacks cysteines, it can be overexpressed (≥ 50 mg/L culture) in E. coli, it is thermally stable (Tm = 88°C), and it retains binding when absorbed onto microtiter plate wells (unlike 95% of monoclonal antibodies, which lose functionality when adsorbed onto plastic (Butler, 1993)). These make the FN3 an ideal scaffold for robust affinity reagent generation and production. This article describes the procedure for isolating FN3 monobodies via phage display, expressing and purifying soluble SUMO-tagged variants, and removing the SUMO-tag to produce affinity reagents that are ready for use in future immunoassay applications. Open in a separate window Figure 1 PyMol representations of the FN3 monobody Left: cartoon, with β-sheets and loops (color). Middle: surface representation. Right: surface representation, viewed from the top. Red, purple, and yellow correspond to FG, BC, and DE loops that have been randomized.