Development of Methodologies for Target Protein Identification and Their Application

Abstract

Identifying target proteins of a bioactive small molecule is a starting point of mechanism-of-action (MoA) study in chemical biology. However, target identification is often considered as a bottleneck in drug discovery based on phenotypic approach because there is a gap between highly complex biological nature and limited analytical methodology. This study aims to expand methodologies for target identification and to apply them for compounds with novel bioactivity. Photoaffinity linkers (PLs) have been used to make a covalent bond between small molecule and target protein, but nonspecific bindings to off-targets increase due to the incorporation of PLs. In Chapter 1, a relation between molecular structure of PLs and nonspecific binding was systematically explored. Small and branched form showed minimal nonspecific binding for both benzophenone- and diazirine-embedded PLs. With branched form PLs, efficiency of the target identification increased due to the reduced artifacts coming from nonspecific off-target bindings. Chemical modification of bioactive compounds with functional groups such as PLs or electrophile is sometimes impossible for compounds having very complex or simple structure, which are often found in natural products. Therefore, chemical labeling-free method for target identification has been needed. In Chapter 2, a label-free target identification method (named as TS-FITGE) was developed. Based on the phenomenon that melting temperature of target protein is shifted by drug binding, fluorescence difference in two-dimensional electrophoresis revealed the target protein among cellular proteome. After the proof-of-concept study with methotrexate, the generality of this method was shown by identifying target proteins of bryostatin 1 (complex structure) and hordenine (simple structure). In Chapter 3, a comprehensive MoA study was conducted with a HeLa-specific cytotoxic compound, A08. As this compound lost its activity by minor modification, label-free target identification methods including TS-FITGE and thermal proteome profiling (TPP) were applied. Seventeen proteins that showed either difference in fluorescence spot from TS-FITGE or shift in melting temperature from TPP were selected as target candidates. Based on siRNA-mediated gene knockdown and in vitro enzymatic assay, MutT homolog 1 (MTH1) was validated to be functionally related to the phenotype of HeLa-specific cytotoxicity. MTH1 protein sanitizes oxidized nucleotides to inhibit their insertion into DNA and prevent DNA damage. However, the binding between MTH1 and A08 was not specific in HeLa cell because thermal stabilization of MTH1 was observed not only in HeLa cell but also in CaSki cell where A08 did not show cytotoxicity. Proteome expression profiling claimed that MTH1 expression level was not distinct, but expression levels of proteins involved in DNA mismatch repair were suppressed in HeLa cell compared to CaSki cell. Thus, it was inferred that HeLa cell tends to be susceptible to oxidative stress that would results in DNA damage. Fluorescence imaging showed increase of intracellular 8-oxo-dG, and subsequent DNA damage and apoptosis markers were activated by A08 in HeLa selectively. In summary, in an effort to contribute to the field of target identification and chemical biology, this study analyzed impact of the molecular shape of PLs on nonspecific binding, developed a label-free method for target identification, and investigated MoA of a HeLa-specific cytotoxic compound with the label-free methods. As the target identification has been a hurdle for phenotype-based drug discovery, increase of current knowledge and technique for target identification would promote phenotypic screening. As a result, this virtuous cycle may accelerate translational research via discovery of novel druggable targets and first-in-class drugs.