In this thesis, a polypeptide binder concept is illustrated. By conjugation to a set of sixteen polypeptides, a small binding molecule can evolve into a polypeptide binder with increased affinity and selectivity. The concept was applied to 2-oxo-1,2-dihydroquinoline-8-carboxylic acid (DQ) and acetazolamide (AZM) for development of high affinity binders targeting human C-reactive protein (CRP) and human carbonic anhydrase (HCA) II and IX respectively. In addition, peptididic macrocycles were developed as inhibitors of lysine specific demethylase 1 (LSD1). CRP is a well-known biomarker of inflammation in humans and binders recognizing it are therefore of large interest as medical diagnostics. Until now, phosphocholine (PCh) and derivatives are the only known small molecule binders for CRP, but they have low μM affinity and bind CRP in a Ca2+ dependent manner. The small molecule DQ was designed as a CRP binder that is structurally unrelated to PCh. Its polypeptide conjugate, 4-C25l22-DQ, was demonstrated as a strong, Ca2+ independent binder for CRP, and had an affinity approximately three orders of magnitude higher than DQ itself. HCA IX is a protein that is interesting for diagnosis of cancer. AZM is a small molecule inhibitor of HCAs with a dissociation constant of 38 nM for HCA II and 3 nM for HCA IX. Interestingly, polypeptide conjugate 4-C10L17-AZM displayed stronger binding to both HCA II (KD 4 nM) and HCA IX (KD 90 pM). This result provided evidence that the binder concept can be applied also for small molecules which already have high affinity for their protein receptors. LSD1 is an enzyme that regulates the methylation of Lys 4 of histone 3 via a PPI-like interaction and which is of therapeutic interest in certain cancers. Based on the structures of two peptidic ligands bound to LSD1, we sequentially prepared truncated, mono-substituted and macroyclic peptides in order to develop reversible inhibitors of LSD1. Some stapled cyclic peptides bound to LSD1 with 10-fold higher affinity than the corresponding linear parent peptide. Changing the staple into a lactam further improved the binding potency and the best lactams inhibited the enzymatic activity of LSD1 at low μM Ki values.

A novel concept for protein recognition has been developed. The recognition unit is a hybrid molecule obtained by conjugation of a small organic molecule to a synthetic polypeptide selected from a 16-membered set of 42 amino acid residue sequences. The sequences are unordered and have no prior relation to the target proteins. The concept is based on the hypothesis that a small set of sequences capable of hydrophobic interactions, hydrogen bonding and electrostatic interactions can yield a binder for any selected protein, provided that the small molecule shows medium affinity or better and is reasonably selective. The concept has been illustrated by the design, synthesis and evaluation of binders for three different proteins, the C-reactive protein, CRP, human Carbonic anhydrase II, HCAII, and Acetylcholine esterase, AChE. Highly efficient binders for CRP have been developed by conjugation of a derivative of the natural ligand, phosphocholine, to the side chain of one of the amino acids in each polypeptide. The binders in the set show a wide range of affinities for CRP and the tightest binder, 4-C10L17-PC6, binds almost irreversibly. Selected binders have been evaluated in human serum, where they capture CRP with high selectivity.High-affinity binders have been developed for HCAII, and the selectivity evaluated by extraction of the protein from blood. The binder 4-C37L34-B, a polypeptide conjugated to a spacered benzenesulphonamide residue, was able to extract Carbonic anhydrases specifically and to discriminate between the two isoforms of human Carbonic anhydrase. The conjugation of an acridine derivative to a polypeptide via a 14 atom spacer has been shown to yield a binder with high affinity and selectivity for AChE. The selectivity was demonstrated by extraction of AChE from Cerebrospinal fluid. This thesis focuses on the development of a fast and reliable procedure for the construction, selection and evaluation of protein binders, with the ambition to develop a technology that is applicable to the development of binders for all proteins.

In this thesis molecular interactions between polypeptide conjugates and protein targets were investigated. Polypeptides were derivatized with small organic molecules, peptides and oligonucleotides. New strategies were developed with the aim to increase affinities for proteins of biological interest. A 42-residue polypeptide (4-C15L8) conjugated to a small organic molecule 3,5-bis[[bis(2-pyridylmethyl)amino]methyl]benzoic acid (PP1), was shown to bind glycogen phosphorylase a (GPa) in the presence of zinc ions. Under the assumption that hydrophobic interactions dominated the binding energy, the hydrophobic residues of 4-C15L8-PP1 were systematically replaced in order to study their contribution to the affinity enhancement. The replacement of the Nle, Ile and Leu residues by Ala amino acids reduced affinities. The introduction of non-natural L-2-aminooctanoic acid (Aoc) residues into the peptide sequence enhanced the binding affinity for GPa. A decreased KD of 27nM was obtained when Nle5, Ile9 and Leu12 were replaced by Aoc residues, in comparison to the KD value of 280nM obtained for the unmodified 4-C15L8-PP1. It is evident that there are non-obvious hydrophobic binding sites on the surfaces of proteins that could be identified by introducing the more hydrophobic and conformationally flexible Aoc residues. The downsizing of the 42-mer peptide to an 11-mer and the incorporation of three Aoc residues gave rise to a KD of 550 nM, comparable to that of  4-C15L8-PP1 suggesting that bioactive peptides can be downsized by the introduction of Aoc. Aiming to improve in vivo stability, the affinity for human serum albumin (HSA) of hydrophobic, positively and negatively charged polypeptide-PP1 conjugates was evaluated. Increased hydrophobicity due to the introduction of Aoc residues did not significantly increase the affinity for HSA. No binding was observed in the case of the most negatively charged polypeptides whereas the slightly negatively and positively charged polypeptides conjugated to PP1 bound HSA with affinities that increased with the positive charge. It was found that polypeptide-PP1 conjugates target the zinc binding site of the HSA. Affinity enhancement was obtained due to the incorporation of PP1 and increased by charge to charge interactions between the positively charged amino acids of the polypeptide and the negatively charged residues of HSA, in close proximity to the HSA zinc binding site. The survival times of the peptide-PP1 conjugates in human serum were extended as a result of binding to HSA. Zn2+ ion chelating agents can be incorporated in potential peptide therapeutics with a short plasma half-life, without increasing their molecular weights.