Fabian Buller

Selection of Carbonic Anhydrase IX Inhibitors from One Million DNA-Encoded Compounds

DNA-encoded chemical libraries, i.e., collections of compounds individually coupled to distinctive DNA fragments serving as amplifiable identification barcodes, represent a new tool for the de novo discovery of small molecule ligands to target proteins of pharmaceutical interest. Here, we describe the design and synthesis of a novel DNA-encoded chemical library containing one million small molecules. The library was synthesized by combinatorial assembly of three sets of chemical building blocks using Diels-Alder cycloadditions and by the stepwise build-up of the DNA barcodes. Model selections were performed to test library performance and to develop a statistical method for the analysis of high-throughput sequencing data. A library selection against carbonic anhydrase IX revealed a new class of submicromolar bis(sulfonamide) inhibitors. One of these inhibitors was synthesized in the absence of the DNA-tag and showed accumulation in hypoxic tumor tissue sections in vitro and tumor targeting in vivo.

A Portable Albumin Binder from a DNA‐Encoded Chemical Library

Albumin represents the most abundant protein in human plasma, at a concentration of 45 mgmL . To keep physiological production rates to a minimum, albumin displays a long circulatory half-life in mammals thanks to its size above the renal filtration threshold and its unique ability to interact with the neonatal FcRn receptor. Fusions of biopharmaceuticals to albumin or to albumin-binding peptides have been devised to expose the body to adequate concentrations of the therapeutic agent for a sufficiently long period of time, thus improving efficacy and reducing the number of injections. In principle, small organic albumin-binding molecules could be used as functional analogues of albumin-binding peptides. However, although many small molecules are known to bind to albumin, the success in isolating small molecules as portable albumin-binding moieties has been limited, mainly because most albumin binders (for example, ibuprofen) lose binding affinity upon chemical modification. Myristoylation of insulin has been shown to significantly prolong the circulatory half-life, but this modification is not applicable to a broader set of molecules because of its negative effect on solubility. In another example, a 4,4diphenylcyclohexyl moiety has been connected through a phosphodiester bond to the metal chelator diethylenetriaminepentaacetic acid (DTPA) for magnetic resonance imaging (MRI) applications and to short peptides. These compounds display dissociation constants (Kd) from human serum albumin in the 100 mm range and are susceptible to hydrolysis in vivo. Thus, there is a considerable scientific and biotechnological interest in the identification of small portable binders that display a stable noncovalent interaction with serum albumin. Herein, we report the discovery and characterization of a class of 4-(p-iodophenyl)butyric acid derivatives from a DNA-encoded chemical library, which display a stable noncovalent binding interaction with both mouse serum albumin (MSA) and human serum albumin (HSA). One of these portable albumin-binding moieties was used to improve the performance of the contrast agents fluorescein and GdDTPA. HSA-binding molecules were selected from a DNAencoded chemical library consisting of 619 oligonucleotidecompound conjugates carrying a six-base-pair code for identification. After selection, the DNA sequences of the enriched compounds were amplified by PCR and decoded on oligonucleotide microarrays displaying the complementary sequences (Figure 1a), normalizing the signal intensities after selection against the intensities of compounds selected on empty resin (Figure 1b). Some of the identified binding molecules were excluded from further evaluation based on being promiscuous binders or because of the high standard deviations of the signal intensities on the microarrays (64, 313, 453, 454, 619). Several of the selected molecules (428, 533, 535, 539) displayed striking structural similarities. The basic structure featured a 4-phenylbutanoic acid moiety, with different hydrophobic substituents on the phenyl ring. To obtain further insights into structure–activity relationships, DNA-modified analogues containing propanoyl or pentanoyl skeletons, and/or carrying various substituents on the phenyl ring (Figure 1c; 536, 622–632), were characterized in a radioactivity-based chromatographic albumin-binding assay, which allowed a first classification of the potential binders (Retention: 428> 539> 624> 535> 533> 536> 326> others; see the Supporting Information). The absence of retention of compounds with propanoyl (625) and penta[*] S. Tr ssel, F. Buller, Dr. F. Bootz, Dr. Y. Zhang, L. Mannocci, Dr. J. Scheuermann, Prof. Dr. D. Neri Institut f r Pharmazeutische Wissenschaften Departement f r Chemie und Angewandte Biowissenschaften ETH Z rich Wolfgang-Pauli-Strasse 10, 8093 Z rich (Switzerland) Fax: (+41)44-633-1358 E-mail: neri@pharma.ethz.ch

New Strategy for the Extension of the Serum Half-Life of Antibody Fragments

Antibody fragments can recognize their cognate antigen with high affinity and can be produced at high yields, but generally display rapid blood clearance profiles. For pharmaceutical applications, the serum half-life of antibody fragments is often extended by chemical modification with polymers or by genetic fusion to albumin or albumin-binding polypeptides. Here, we report that the site-specific chemical modification of a C-terminal cysteine residue in scFv antibody fragments with a small organic molecule capable of high-affinity binding to serum albumin substantially extends serum half-life in rodents. The strategy was implemented using the antibody fragment F8, specific to the alternatively spliced EDA domain of fibronectin, a tumor-associated antigen. The unmodified and chemically modified scFv-F8 antibody fragments were studied by biodistribution analysis in tumor-bearing mice, exhibiting a dramatic increase in tumor uptake for the albumin-binding antibody derivative. The data presented in this paper indicate that the chemical modification of the antibody fragment with the 2-(3-maleimidopropanamido)-6-(4-(4-iodophenyl)butanamido)hexanoate albumin-binding moiety may represent a general strategy for the extension of the serum half-life of antibody fragments and for the improvement of their in vivo targeting performance.

Design and synthesis of a novel DNA-encoded chemical library using Diels-Alder cycloadditions.

DNA-encoded chemical libraries are increasingly being employed for the identification of binding molecules to protein targets of pharmaceutical relevance. Here, we describe the synthesis and characterization of a DNA-encoded chemical library, consisting of 4000 compounds generated by Diels-Alder cycloaddition reactions. The compounds were encoded with unique DNA fragments which were generated through a stepwise assembly process and serve as amplifiable bar codes for the identification and relative quantification of library members.

High-throughput sequencing for the identification of binding molecules from DNA-encoded chemical libraries.

DNA-encoded chemical libraries are large collections of small organic molecules, individually coupled to DNA fragments that serve as amplifiable identification bar codes. The isolation of specific binders requires a quantitative analysis of the distribution of DNA fragments in the library before and after capture on an immobilized target protein of interest. Here, we show how Illumina sequencing can be applied to the analysis of DNA-encoded chemical libraries, yielding over 10 million DNA sequence tags per flow-lane. The technology can be used in a multiplex format, allowing the encoding and subsequent sequencing of multiple selections in the same experiment. The sequence distributions in DNA-encoded chemical library selections were found to be similar to the ones obtained using 454 technology, thus reinforcing the concept that DNA sequencing is an appropriate avenue for the decoding of library selections. The large number of sequences obtained with the Illumina method now enables the study of very large DNA-encoded chemical libraries (>500,000 compounds) and reduces decoding costs.

Drug discovery with DNA-encoded chemical libraries.

DNA-encoded chemical libraries represent a novel avenue for the facile discovery of small molecule ligands against target proteins of biological or pharmaceutical importance. Library members consist of small molecules covalently attached to unique DNA fragments that serve as amplifiable identification barcodes. This encoding allows the in vitro selection of ligands at subpicomolar concentrations from large library populations by affinity capture on a target protein of interest, in analogy to established technologies for the selection of binding polypeptides (e.g., antibodies). Different library formats have been explored by various groups, allowing the construction of chemical libraries comprising up to millions of DNA-encoded compounds. Libraries before and after selection have been characterized by PCR amplification of the DNA codes and subsequent relative quantification of library members using high-throughput sequencing. The most enriched compounds have then been further analyzed in biological assays, in the presence or in the absence of linked DNA. This article reviews experimental strategies used for the construction of DNA-encoded chemical libraries, revealing how selection, decoding, and hit validation technologies have been used for drug discovery programs.

Automated screening for small organic ligands using DNA-encoded chemical libraries

DNA-encoded chemical libraries (DECLs) are collections of organic compounds that are individually linked to different oligonucleotides, serving as amplifiable identification barcodes. As all compounds in the library can be identified by their DNA tags, they can be mixed and used in affinity-capture experiments on target proteins of interest. In this protocol, we describe the screening process that allows the identification of the few binding molecules within the multiplicity of library members. First, the automated affinity selection process physically isolates binding library members. Second, the DNA codes of the isolated binders are PCR-amplified and subjected to high-throughput DNA sequencing. Third, the obtained sequencing data are evaluated using a C++ program and the results are displayed using MATLAB software. The resulting selection fingerprints facilitate the discrimination of binding from nonbinding library members. The described procedures allow the identification of small organic ligands to biological targets from a DECL within 10 d.