Multivalent site-specific phage modification enhances the binding affinity of receptor ligands

Abstract

High-throughput screening of combinatorial chemical libraries is a powerful approach for identifying targeted molecules. The display of combinatorial peptide libraries on the surface of bacteriophages offers a rapid, economical way to screen billions of peptides for specific binding properties and has impacted fields ranging from cancer to vaccine development. As a modification to this approach, we have previously created a system that enables site-specific insertion of selenocysteine (Sec) residues into peptides displayed pentavalently on M13 phage as pIII coat protein fusions. In this study, we show the utility of selectively derivatizing these Sec residues through the primary amine of small molecules that target a G protein-coupled receptor, the adenosine A1 receptor, leaving the other coat proteins, including the major coat protein pVIII, unmodified. We further demonstrate that modified Sec-phage with multivalent bound agonist binds to cells and elicits downstream signaling with orders of magnitude greater potency than that of unconjugated agonist. Our results provide proof of concept of a system that can create hybrid small molecule-containing peptide libraries and open up new possibilities for phage-drug therapies.

Figure 1.

Schematic of the M13KE bacteriophage with incorporation of selenocysteine for site-specific modification. A Sec insertion sequence and UGA codon were inserted at the 5′ end of M13 gIII, which encodes for the pIII coat protein, providing five Sec residues for small molecule tethering under appropriate conditions.

Figure 2.

Site-specific modification of seleno-pIII coat protein of s12d phage. (A) Phage modified with the indicated reagents were analyzed via western blotting for the presence of biotin (top) and the pIII coat protein (bottom). Bands are seen at molecular masses near 60 kDa (the apparent molecular weight of the pIII protein) for biotin-tethered and NOAMI-Bt-tethered s12d phage but not for negative control, Sec-free M13KE phage. (B) Phage–small molecule complexes were reacted and quantified using qWestern blotting (left) to determine the relative efficiency of small molecule tethering (right). (C) Conjugation of additional small molecules to s12d or M13KE phage. Phage were incubated with the indicated molecules and subsequently incubated with Bt. As the small molecules did not contain a biotin group, the decrease in a band corresponding to the presence of biotin indicates successful tethering of small molecules.

Figure 3.

NOAMI-Bt-conjugated phage retain specificity for A₁. (A) Cell binding assays using human adenosine A₁ receptor-expressing CHO cells and modified s12d phage. Cells were incubated with NOAMI-Bt-modified phage and the presence of bound phage detected via ELISA. NOAMI-Btmodified phage bind with high affinity (ED₅₀ = 0.17 nM), whereas control phage (unmodified or biotin-conjugated only) demonstrated negligible binding. (Inset) Western blot of hA₁-CHO cells for the human adenosine A₁ receptor reveals a strong band at approximately 36 kDa, the reported molecular weight of the A₁ receptor. (B) Competition assays. hA₁-CHO cells were incubated with increasing concentrations of free NOAM followed by 0.5 nM NOAMI-Bt- or Bt-modified phage. As in panel A, phage binding was detected via ELISA. (C) Radioligand competition assays. ¹²⁵I-ABA incubated with hA₁-CHO membrane was reduced by 18, 54, and 77% with increasing concentrations of s12d-NOAMI-Bt (bottom) but not s12d-Bt (top).

Figure 4.

Binding kinetics of modified s12d phage. (A) Aliquots containing 0.5 nM of phage were incubated with hA₁-CHO cells at 4 °C for the indicated times. Due to deviation from standard binding kinetics (red box), association kinetic curve fits of s12d-NOAMI-Bt binding (squares) excluded the first 10 min were fit with an R² value of 0.868. Unmodified (triangles) and Bt-modified (circles) phage are fit without omission. (B) Phage were incubated for 1 h with hA₁-CHO cells and then removed. The cells were washed at various time, and remaining phage was quantified via ELISA. Data beyond 45 min were fit to a single-exponential decay curve (R² = 0.879) as illustrated for s12d-NOAMI-Bt.

Figure 5.

NOAMI-Bt is functional when tethered to s12d phage. (A) AKT pathway activation by small molecules and modified phage. hA₁-CHO cells were incubated with the indicated concentrations of CHA, NECA, and NOAMI-Bt molecules or modified phage for 15 min. Activation of AKT was analyzed by qWestern blotting for phospho-AKT and normalized to β-actin. (B) Activation plot fits based on the qWestern blots and normalized to baseline for AKT in response to increasing concentrations of s12d-NOAMI-Bt (gray squares, EC₅₀ = 3.20 pM), free NOAMI-Bt (black squares, EC₅₀ = 510 nM), CHA (triangles, EC₅₀ = 2.44 nM), or NECA (circles, EC₅₀ = 1.88 nM).