Research Article: Peptide barcoding for establishment of new types of genotype–phenotype linkages

Date Published: April 23, 2019

Publisher: Public Library of Science

Author(s): Kana Miyamoto, Wataru Aoki, Yuta Ohtani, Natsuko Miura, Shunsuke Aburaya, Yusei Matsuzaki, Kaho Kajiwara, Yoshinori Kitagawa, Mitsuyoshi Ueda, Maxim Antopolsky.


Measuring binding properties of binders (e.g., antibodies) is essential for developing useful experimental reagents, diagnostics, and pharmaceuticals. Display technologies can evaluate a large number of binders in a high-throughput manner, but the immobilization effect and the avidity effect prohibit the precise evaluation of binding properties. In this paper, we propose a novel methodology, peptide barcoding, to quantitatively measure the binding properties of multiple binders without immobilization. In the experimental scheme, unique peptide barcodes are fused with each binder, and they represent genotype information. These peptide barcodes are designed to have high detectability for mass spectrometry, leading to low identification bias and a high identification rate. A mixture of different peptide-barcoded nanobodies is reacted with antigen-coated magnetic beads in one pot. Peptide barcodes of functional nanobodies are cleaved on beads by a specific protease, and identified by selected reaction monitoring using triple quadrupole mass spectrometry. To demonstrate proof-of-principle for peptide barcoding, we generated peptide-barcoded anti-CD4 nanobody and anti-GFP nanobody, and determined whether we could simultaneously quantify their binding activities. We showed that peptide barcoding did not affect the properties of the nanobodies, and succeeded in measuring the binding activities of these nanobodies in one shot. The results demonstrate the advantages of peptide barcoding, new types of genotype–phenotype linkages.

Partial Text

Binders such as antibodies are biologics that are useful for application as experimental reagents, diagnostics, and pharmaceuticals. To measure the binding properties of binders, one approach is to purify each binder and measure its kinetic parameters using surface plasmon resonance or ELISA. This approach enables precise measurement of kinetic parameters, although its throughput is low. An alternative approach is to use various display technologies: phage display [1,2], bacterial display [3], yeast display [4,5], mammalian display [6], or ribosome display [7]. In these display technologies, a gene encoding a chimeric protein of a binder and a cell surface protein is designed and introduced into phages or cells. The produced chimeric proteins are immobilized at the cell surface, establishing genotype–phenotype linkages at the single-cell level. Using randomized display libraries, we can isolate various binders after several cycles of panning assays and genetic sequencing of positive clones. Display technologies can evaluate a large number of binders in a high-throughput manner, and have been used to isolate functional antibodies including various antibody medicines [8,9].

High-throughput evaluation of the binding properties of non-immobilized free binders can be a valuable approach to develop good binders. In this paper, we propose a novel strategy for this, peptide barcoding. In this strategy, we generate a binder fused with a unique peptide barcode, which includes genotype information, and the simultaneous detection of peptide barcodes by LC–MS enables the high-throughput evaluation of the binding properties of multiple free binders. This approach has several advantages. First, peptide barcoding is devoid of adverse effects derived from the immobilization effect. Conventional display technologies require the immobilization of binders on the host cell surface. This immobilization on large carriers could affect the binding properties of binders [10,11] and the accessibility to antigens [12]. Peptide barcodes are very small and expected to have less adverse effects. In fact, our results showed that peptide barcoding did not affect the binding properties of the nanobodies (Fig 3 and Table 1). However, it will be necessary to obtain definitive evidence on whether peptide barcoding affects the properties of nanobodies by using peptide barcodes with a wide range of chemical properties. The second advantage of peptide barcoding is the improvement of detection sensitivity. In our approach, we only detect pre-designed peptide barcodes confirmed to possess high detection sensitivity for LC–MS, leading to the possibility of identifying rare binders. The third advantage is the low identification bias, which is also derived from the use of pre-designed peptide barcodes. In previous studies using shotgun proteomics to identify functional binders [17–19], tryptic digests of binder ensembles are introduced to LC–MS. Such an experimental scheme is very simple, but suffers from identification bias due to the existence of peptides with low ionization and fragmentation efficiency. Furthermore, high sequence homology of binder ensembles leads to ambiguous identification of functional binders. In contrast, our approach detects only peptide barcodes that are pre-designed to have high ionization and fragmentation efficiency and 1:1 genotype–phenotype matching. These characteristics result in the unambiguous identification of functional binders. The fourth advantage is the ability to simultaneously evaluate the productivity of binders. In our experimental scheme, we used nanobodies secreted by P. pastoris. Hence, we can simultaneously evaluate the productivity and binding properties by LC–MS, enabling the simplification of procedures of binder generation.