Disulfide-constrained peptide scaffolds enable a robust peptide-therapeutic discovery platform

Abstract

Peptides present an alternative modality to immunoglobulin domains or small molecules for developing therapeutics to either agonize or antagonize cellular pathways associated with diseases. However, peptides often suffer from poor chemical and physical stability, limiting their therapeutic potential. Disulfide-constrained peptides (DCP) are naturally occurring and possess numerous desirable properties, such as high stability, that qualify them as drug-like scaffolds for peptide therapeutics. DCPs contain loop regions protruding from the core of the molecule that are amenable to peptide engineering via direct evolution by use of phage display technology. In this study, we have established a robust platform for the discovery of peptide therapeutics using various DCPs as scaffolds. We created diverse libraries comprising seven different DCP scaffolds, resulting in an overall diversity of 2 x 1011. The effectiveness of this platform for functional hit discovery has been extensively evaluated, demonstrating a hit rate comparable to that of synthetic antibody libraries. By utilizing chemically synthesized and in vitro folded peptides derived from selections of phage displayed DCP libraries, we have successfully generated functional inhibitors targeting the HtrA1 protease. Through affinity maturation strategies, we have transformed initially weak binders against Notch2 with micromolar Kd values to high-affinity ligands in the nanomolar range. This process highlights a viable hit-to-lead progression. Overall, our platform holds significant potential to greatly enhance the discovery of peptide therapeutics.

A phage-displayed disulfide constrained peptide discovery platform yields novel human plasma protein binders

“…The phage libraries were quality-controlled by next-generation sequencing (NGS) analysis before proceeding to the selection step. Combined with the previously engineered libraries [ 35 , 36 ], we have developed in total fifty-nine individual DCP phage libraries based on thirteen DCP scaffolds (S1 Table in S2 File ). To streamline the screening process, these DCP phage libraries were pooled into three groups (Lib E, containing EETI-II libraries; Lib 1, containing AVR9, Circulin-A, conotoxin-MVIIA, Huwentoxin, Charybdotoxin, CBD, CBD amylose libraries; and Lib 2, containing Mch1, gurmarin, Asteropsin-A, AMP-1, CPI libraries) for phage panning (S1 Table in S2 File ).…”

“…The DCP phage libraries were generated as previously described [ 35 , 36 ]. Five naturally occurring DCPs (Mch1, gurmarin, Asteropsin-A, AMP-1 and CPI) were selected as scaffolds for display on the surface of M13 bacteriophage.…”

“…Each DCP was fused via the C-terminus to the N-terminus of M13 major coat protein (pVIII), with a gD (HSV-2 envelope glycoprotein D)-tag fused to the N-terminus of the DCP sequence. To prevent introduction of extra Cys within the loop region, libraries were generated with trinucleotide oligos where residues denoted with “X” were replaced with equimolar of 19 amino acids (Cys excluded) [ 36 ]. The extended DCP phage libraries contain 13 DCP frameworks as shown in S1 Table of S2 File .…”

“…The DCP family offers an attractive opportunity for phage library construction with high diversity. Taking advantage of the loop regions of these peptides, we have previously engineered phage libraries using eight DCP scaffolds as templates, including EETI-II (loop 1, 5, or 1/5 randomized), AVR9 (loop 4 randomized), Circulin-A (loop 4 randomized), conotoxin-MVIIA (loop 2 randomized), Huwentoxin (loop 2 randomized), Charybdotoxin (loop 1/2/3, loop 4, or C-term + loop 3/4/5 randomized), cellulose binding protein (CBD) (loop 1/3 or N-term + loop 3 randomized), and CBD amylose (loop 1/3 or N-term + loop 3 randomized) [ 35 , 36 ]. Six out of the eight DCPs are CKPs, while CBD and CBD amylose, containing only four cysteines, do not form knotted conformation.…”

“…Six out of the eight DCPs are CKPs, while CBD and CBD amylose, containing only four cysteines, do not form knotted conformation. These DCP phage libraries were used to identify binders against different protein targets successfully [ 35 , 36 ]. Here, we expanded the molecular diversity of the DCP phage libraries through engineering additional 26 new DCP phage libraries based on five different DCP scaffolds, namely Momordica charantia 1 (Mch1) [ 37 ], gurmarin [ 38 ], Asteropsin-A [ 39 ], antimicrobial peptide-1 (AMP-1) [ 40 ], and potato carboxypeptidase inhibitor (CPI) [ 41 ].…”