Introduction

While monoclonal antibodies continue to dominate the biologics market, there is significant activity in the discovery of non-mAb scaffolds as biological products for several reasons such as lower half-life, smaller size and ease of engineering. Out of the different protein scaffold families, one is the 10th Fibronectin Type III (10Fn3) domain, popularly known as ‘monobodies’ or ‘Adnectins™’ (Koide et al., 2012). These molecules also possess flexible loops similar to the CDRs in antibodies but are nearly 10-fold smaller than monoclonal antibodies (~10-12 kDa in size compared to monoclonal antibodies which are ~140-150 kDa in size) (Lipovsek, 2011). It is well known that protein A targets the conserved Fc region of IgGs and is a well-established platform affinity ligand for the purification of monoclonal antibodies (Huse et al., 2002). However, unlike IgGs, small scaffold proteins such as Adnectins contain only a single domain. A 70-80% homology was observed from the sequence alignment of existing Adnectin molecules in the literature. These homologous regions can be potentially targeted for a generalizable purification platform for Adnectins.
Peptides have shown considerable potential as affinity ligands for the purification of biological targets due to several favorable attributes such as structural flexibility, stability, sequence diversity and ease of production. While rational approaches have been used for the design of peptide ligands against protein targets (Chandra et al., 2013), combinatorial display techniques have also resulted in the identification of affinity peptides against protein targets. Peptide-based phage display has been widely used for the identification of peptide affinity ligands for drug discovery (Nixon et al., 2014), biosensing (Wu et al., 2011), as well as in the affinity purification of biologics (Kelley et al., 2004). Phage display-aided peptide selection employs a procedure known as biopanning which involves challenging a library of bacteriophage against an immobilized target. Strategic washes, negative selections and multiple biopanning rounds result in the discovery of affinity peptide ligands specific to the target. Peptides have been previously employed for the chromatographic purification of mAbs (Yang et al., 2008) as well as various non-mAb targets such as Factor VIII (Kelley et al., 2004), human growth hormone (Chandra et al., 2019), erythropoietin (Kish et al., 2018), Fab fragments (Nascimento et al., 2019) and viruses (Heldt et al., 2008). The moderate affinities of peptide affinity ligands to their corresponding target result in both effective purification as well as relatively milder conditions for elution and subsequent target recovery.
The class of phase transitioning smart biopolymers – elastin-like polypeptides (ELPs) – has been investigated over the last two decades for the downstream processing of proteins (Yeboah et al., 2016). ELPs are typically composed of multiple repeats of the pentapeptide unit V-P-G-X-G (where guest residue X can be any amino acid except proline) and can undergo precipitation upon the change in stimuli such as temperature and salt concentration (Meyer & Chilkoti, 1999). Different approaches involving ELP-protein fusion systems have been effectively employed for protein purification (MacEwan et al., 2014). However, these often require proteolytic cleavage or the introduction of self-splicing inteins (Fong et al., 2009) to dissociate the ELP-tag from the product, resulting in the requirement of additional purification steps. Therefore, an alternate strategy – affinity precipitation – involving fusions of ELPs and affinity ligands have been employed for the tag-free purification of proteins (Madan et al., 2013). Previous work by Cramer and Chen has successfully demonstrated the purification of monoclonal antibodies from complex mixtures using ELP-Z-based affinity precipitation (Sheth et al., 2013). This process was found to be comparable to protein A purification and also readily scalable when performed with membrane filtration steps (Sheth, Bhut, et al., 2014; Sheth, Jin, et al., 2014). The Cramer and Karande labs have also recently demonstrated the purification of model proteins such as Anti-Flag Antibody M2 and streptavidin using peptide-ELP fusions (Mullerpatan et al., 2020).
In this work, we specifically focused on an industrially relevant small non-mAb scaffold protein, P-Adnectin (AdP). At present the purification of AdP has been carried out using affinity tag fusions (Mitchell et al., 2014) or with a series of chromatographic steps. In the current work, we first employed phage panning to identify peptide affinity ligands against AdP. Lead candidates were then synthesized and their binding behavior to the product was evaluated using fluorescence polarization. A peptide-ELP fusion of the best performing candidate was then produced inE. coli and various conditions of stoichiometry and pH were examined to establish appropriate conditions for binding and precipitation of the AdP. Elution conditions were then established using a combination of pH and fluid phase modifiers. Finally, the optimized capture and elution conditions were employed in a proof of concept experiment demonstrating the utility of a single-step affinity precipitation process that was capable of capturing ~80% of AdP from a crude E. coli lysate at a purity of ~90%.