Affinity Maturation of a Cyclic Peptide Handle for Therapeutic Antibodies Using Deep Mutational Scanning

Abstract

Meditopes are cyclic peptides that bind in a specific pocket in the antigen-binding fragment of a therapeutic antibody such as cetuximab. Provided their moderate affinity can be enhanced, meditope peptides could be used as specific non-covalent and paratope-independent handles in targeted drug delivery, molecular imaging, and therapeutic drug monitoring. Here we show that the affinity of a recently reported meditope for cetuximab can be substantially enhanced using a combination of yeast display and deep mutational scanning. Deep sequencing was used to construct a fitness landscape of this protein-peptide interaction, and four mutations were identified that together improved the affinity for cetuximab 10-fold to 15 nm Importantly, the increased affinity translated into enhanced cetuximab-mediated recruitment to EGF receptor-overexpressing cancer cells. Although in silico Rosetta simulations correctly identified positions that were tolerant to mutation, modeling did not accurately predict the affinity-enhancing mutations. The experimental approach reported here should be generally applicable and could be used to develop meditope peptides with low nanomolar affinity for other therapeutic antibodies.

Keywords: antibody; cyclic peptide; directed evolution; epidermal growth factor receptor (EGFR); protein engineering.

FIGURE 1.

Improving cyclic peptide binding to cetuximab using yeast display. A, structure of the Fab fragment of cetuximab indicating the position of the meditope in the central cavity. B, schematic representation of the yeast display system. Simultaneous labeling of the HA tag and cetuximab with different fluorophores allows selection for binding to be normalized for variations in expression.

FIGURE 2.

Display of cyclic meditope peptide on yeast cells. A, cetuximab titration to 10 nm FITC-labeled peptides (supplemental Table S2, numbers 1–4) in PBS solution, pH 7.4, with 1 mg ml⁻¹ BSA. Fluorescence polarization was used as a readout. Data represent mean ± S.D. from duplicate experiments. B, titration of cetuximab to yeast displaying the indicated cyclic peptides. Cells were labeled with the indicated concentrations of cetuximab conjugated to Alexa Fluor 647 as well as mouse anti-HA and an Alexa Fluor 488-conjugated goat anti-mouse secondary antibody. The flow cytometric Alexa Fluor 647 mean fluorescence of the Alexa Fluor 488-positive cells was used as a readout.

FIGURE 3.

Selection of improved mutants from single amino acid substitution libraries. A, FACS sorting of pooled libraries. Libraries were pooled prior to sorting and incubated with either 30 or 100 nm Alexa Fluor 647-conjugated cetuximab, mouse anti-HA, and an Alexa Fluor 488-conjugated goat anti mouse secondary antibody. Plots show the library and sorting gate used in the first (left), second (middle), and third (right) sorting rounds. B, FACS histograms of cells expressing either the original meditope or mutants Q1V, S5G, or K10R. Cells were labeled with 30 nm Alexa Fluor 647-conjugated cetuximab. Populations on the left are non-expressing cells, whereas populations on the right are expressing cells. C, fluorescence polarization assay. Cetuximab was titrated to 10 nm of the indicated synthetic peptides in 50 mm sodium phosphate, pH 7.0, 100 mm sodium chloride, 1 mg ml⁻¹ BSA. Indicated mutations are with respect to Md1. For exact sequences with linkers and flanking residues see supplemental Table S2 peptides 5–12. Error bars represent mean ± S.D. from duplicate measurements.

FIGURE 4.

Deep sequencing analysis of separately sorted single NNK libraries. A, schematic overview of the deep mutational scanning strategy. B, heat map of robust enrichment ratios (see under “Materials and Methods”) of all single amino acid substitutions of the original cetuximab meditope. All RERs above 1 are indicated in the respective squares. At each position, the original residue is encircled.

FIGURE 5.

Deep sequencing analysis of separately sorted double NNK libraries. A, heat map of robust enrichment ratios (see under “Materials and Methods”) of double amino acid substitutions of the original cetuximab meditope. All RERs above 1 are indicated in the respective squares. At each position, the original residue is encircled. B, cetuximab titration to 10 nm FITC-conjugated synthetic peptides in PBS, pH 7.4 1 mg ml⁻¹. Fluorescence polarization was used as a readout. Data points represent mean ± S.D. of triplicate measurements. The indicated mutations are with respect to Md1. For exact peptide sequences, see supplemental Table S2 peptides 13–16.

FIGURE 6.

In silico modeling of meditope mutations. A, heat map of interaction free energies of cetuximab with single amino acid substitutions of the meditope generated by Rosetta modeling. Negative values (improved binding) is represented with dark shades of gray, and positive values (weaker binding than the original meditope) are represented with light shades of gray. Positions are shown on the horizontal axis. B, heat maps of interaction free energies of cetuximab with amino acid substitutions at position 6 of the meditope where position 5 is mutated to tryptophan (left), phenylalanine (middle), or tyrosine (right), generated by Rosetta modeling. Values are normalized to the respective single amino acid substitution at position 5. C, model of double mutant S5Y/T6M showing the position of the side chains that fill the empty space in the binding pocket. D, model of double mutant S5Y/T6M showing the possible hydrogen bond between tyrosine 5 of the meditope and valine 152 of the antibody. E and F, model of mutant D3R showing how the side chain of arginine 3 of the mediotope may fill empty space within the binding pocket (E) and possibly form hydrogen bonds with glutamine 150 and alanine 176 of the antibody (F). G, fluorescence polarization assay. Cetuximab was titrated to 10 nm of the indicated mutant peptides in PBS, pH 7.4, 1 mg ml⁻¹ BSA. Error bars represent mean ± S.D. of triplicate measurements. The indicated mutations are with respect to Md1. For exact sequences see supplemental Table S2 peptides 13, 14, and 18–21.

FIGURE 7.

Cetuximab-mediated targeting of meditopes to EGFR-overexpressing cancer cells. A, A431 cells were incubated with 5 nm unlabeled cetuximab and 50 nm FITC conjugated peptide (supplemental Table S2, numbers 13 and 15) and analyzed by flow cytometry. Mutations are with respect to Md1. B, concentration-dependent meditope binding. A431 cells were incubated with cetuximab, and the indicated concentrations of cyclic peptides as in A and analyzed by flow cytometry. Data represent mean ± S.D. from duplicate experiments. The control consisted of 500 nm of the respective meditope peptides without cetuximab and represents background binding to the cells due to the hydrophobic nature of the peptide.