Multiple Bacteriophage Selection Strategies for Improved Affinity of a Peptide Targeting ERBB2

Abstract

Due to the heterogeneity of ERBB2-expression between tumors and over the course of treatment, a non-invasive molecular imaging agent is needed to accurately detect overall ERBB2 status. Peptides are a highly advantageous platform for molecular imaging, since they have excellent tumor penetration and rapid pharmacokinetics. One limitation of peptides however, is their traditionally low target affinity, and consequently, tumor uptake. The peptide KCCYSL was previously selected from a bacteriophage (phage) display library to bind ERBB2 and did so with moderate affinity of 295 nM. In order to enhance tumor uptake and clinical utility of the peptide, a novel phage microlibrary was created by flanking the parent sequence with random amino acids, followed by reselection using parallel strategies for high affinity and specific ERBB2 binding in an attempt to affinity maturate the peptide. One limitation of traditional phage display selections is difficulty in releasing the highest affinity phages from the target by incubation of acidic buffer. In an attempt to recover high affinity second-generation peptides from the ERBB2 microlibrary, two elution strategies, sonication and target elution, were undertaken. Sonication resulted in an approximately 50-fold enhancement in recovered phage per round of selection in comparison to target elution. Despite the differences in elution efficiency, the affinities of phage-displayed peptides selected from either strategy were relatively similar. Although both selections yielded peptides with significantly improved affinity in comparison to KCCYSL, the improvements were modest, most likely because the parental peptide binding cannot be improved by additional amino acids.

Keywords: Affinity Maturation; ERBB2; Peptides; Phage Display.

Fig. 1

A depiction of the parallel strategies used for affinity maturation of KCCYSL by phage display.

Fig. 2

Phage display selections utilizing excess ERBB2 or sonication as elution strategy were performed in parallel. A) The number of phage recovered after each round was quantified and the percent recovery was expressed as the total phage recovered/input phage. Black bars represent sonication eluted phage recoveries, while gray bars represent the recovery by ERBB2 elution. B) The amplification of each round, as calculated by dividing the total recovered phage of the current round by the total recovered phage of the previous round is shown.

Fig. 3

Individual phage supernatants were incubated with either ERBB2 or a negative control protein, and total bound phages were quantified by ELISA. From each selection strategy, 20 phages were chosen and their absorbance are plotted above, with black bars representing ERBB2 bound phage and white bars representing phage bound to control protein. Phages represented by an “S” were selected by sonication, while those with an “E” were selected by elution with excess ERBB2.

Fig. 4

The total phage recovered from round 3 of each selection strategy were analyzed for ERBB2 binding by phage ELISA and compared to insertless wildtype phage. Black bars represent ERBB2 binding, white bars negative control protein. Error bars represent a mean of three separate assays. Both ERBB2 elution and elution by sonication resulted in significantly higher binding to ERBB2 than negative control, in addition to higher ERBB2 binding than wildtype phage. * - P < 0.05; *** P <0.001

Fig. 5

Peptide binding curves of the four affinity maturated peptides: E6, E16, S13 and S16 in addition to the first genereation peptide KCCYSL were determined by incubating concentrations of peptides ranging from 100 nm to 50 μM with ERBB2 protein. Bound biotinylated peptides were detected by the addition of horseradish peroxidase-conjugated streptavidin. Bars represent the mean of triplicate wells, and the K_D of each peptide is averaged from three separate experiments

Fig. 6

Peptide-induced tryptophan fluorescence quenching of ERBB2. Changes in the fluorescence emission spectrum (320–450 nm) of 80 nM ERBB2 in DPBS (excitation at 290 nm) in the presence of increasing concentrations (0 – 6.65 μM) of S13. Spectra were corrected for Raman scattering. Inset: Binding isotherms for interactions of ERBB2 with peptides. ERBB2 (5–15 μg/ml) was titrated with KCCYSL (○), E6 (□) and S13 (Δ) in DPBS. ΔF were calculated as ·(F − F_min), F, and F_min are fluorescence emission at 334 nm (excitation at 290 nm) after the addition of peptide, and at saturation, respectively. F₀ corresponds to the fluorescence emission prior to addition of the peptide. Solid lines represent the best fit (r² >0.97) of the hyperbolic equation ΔF/(F₀−F_min) = [Peptide]/(Kd + [Peptide]), where [Peptide] is the concentration in nM and Kd is the apparent dissociation constant, to the data using SigmaPlot 12.0. The values of Kd, which correspond to the concentration of the peptide, which induced half of the maximal quenching of the fluorescence emission, were 180 ± 30, 150 ± 20 and 120 ± 15 nM for KCCYSL, E6 and S13, respectively

Fig. 7

Fluorescent microscopy was used to assay the binding of biotinylated peptides to either ERBB2-overexpressing BT-474 cells or ERBB2-negative MDA-MB-468 cells. Peptide bound to the 4% paraformaldehyde fixed cells was visualized by the addition of Cy3-conjugated anti-biotin antibody