Min Yao, Katsumi Maenaka, Fuyuhiko Inagaki, Yoshikazu Tanaka, Toyoyuki Ose, Koji Kato, Kimiko Kuroki,
Shintaro Mikuni, Shunsuke Kita
[Shionogi & Co., Ltd.]
Takeshi Shiota, Yoshinori Yamano, Masashi Deguchi, Yoshito Numata, Kenichi Higashino, Yutaka Yoshida,
Akira Naito, Hiroshi Takemoto, Akira Ino, Kenji Matsuo
To establish an infrastructure for structure-based drug design, we construct a streamlined workflow from ligand screening based on X-ray crystallography and NMR analyses to functional characterization, and discover drugs using structure-function correlation. Major subjects are 1) Development of rapid screening technology platform based on X-ray crystallography including overexpression system, 2) Development of an assay system for protein-protein interaction based on fluorescence cross-correlation spectroscopy (FCCS), 3) Development of drug screening based on NMR analyses, 4) Development of antibody drugs for human Killer cell immunoglobulin-like receptors.
1) Development of rapid screening technology platform based on X-ray crystallography
Structure-Based Drug Design (SBDD) has been considered as an effective method for genomic drug discovery, while the difficulty in handling induced-fit of protein is pointed out. Fragment-Based Drug Design (FBDD), which potentially overcomes this difficulty, is considered as a promising method. FBDD involves the identification of low molecular weight inhibitors and subsequent optimization of chemical structure to reach to the best lead compound. Smaller molecules are readily accommodated to the active site(s) of target protein, hence advantageous in finding the best-fitted compounds. Expeditious structural analysis in complex with fragments is necessary for the successful application of FBDD. We have developed an automated X-ray crystal structure analysis system LAFIRE, and extended for FBDD (LAFIRE_FBDD) to make its active use for the screening of lead compounds in "Drug Discovery Platform Hub". We also developed a program called POCASA to predict ligand binding site based on the structure, and it has been applied to target proteins of antibiotics. Furthermore, we are extending POCASA to be a virtual ligand screening software package POCASA_LS for searching lead compounds.
We also focus on the development of overexpression system using Escherichia coli for membrane proteins or human proteins because soluble overexpression is essential and the first step of SBDD/FBDD experiment.
<Click to enlarge>
2) Development of an assay system for protein-protein interaction based on fluorescence cross-correlation spectroscopy (FCCS)
A novel assay system using fluorescence cross-correlation spectroscopy(FCCS), an optical imaging method, was developed to screen chemical compounds that inhibit protein activity. In the field of drug discovery, powerful strategies such as X-ray crystallography and/or NMR have been used to investigate the affinity of low molecular weight compounds for target proteins. However, analysis of protein function in solution is required to reveal whether the compound really inhibits the function of the target protein in vivo. FCCS has the unique property of being able to measure protein-protein interactions and protein-ligand interactions quantitatively with high sensitivity in aqueous solution. In addition, FCCS has advantages as an assay system in that the measurement can be carried out at a low concentration and in a small amount of sample protein solution with high throughput that needs only several tens of seconds per sample. Such FCCS is carried out to detect the association and dissociation of target proteins and inhibitors of the interactions in a chemical library. Finally, we quantify the inhibitory potency of synthesized compounds by FBDD methods from a seed compound using the system, and provide feedback for the synthesis strategy.
A diagram of the FCCS setup and the example of protein-protein interaction analysis.
3) Development of drug screening based on NMR analyses
Although an in silico screening and HTS have been widely used for drug discovery, the discovery rate of HIT compounds is very low. This is mainly due to difficulties to distinguish a specific binding from a non-specific binding especially in the primary stage of drug screening. The difficulty also arises from the evidence that the ligand-binding induces conformational changes in both ligands and target proteins. This makes it difficult to discover drugs with high affinity binding. NMR is quite sensitive to the weak interaction and therefore is useful for the primary screening. Moreover, NMR detects the conformational changes in both ligands and proteins upon their interaction. In the present project, we will make use of advantage of NMR and develop a novel ligand-screeing method using paramagnetic lanthanoid ions. We will apply this method to explore anti-cancer drugs and antibiotics. Paramagnetic lanthanoid ions fixed in the target protein induces shifts and line broadening of the observed nucleus according to the relative position between the lanthanoid ion and the observed nucleus and are useful to detect the binding of the ligand to the specific site of the target protein and detect the structural change upon complex formation. First, we develop a fast and efficient ligand screening method to screen a bound compound from mixtures utilizing broadening effect of Gd3+. Next, we apply shift inducing lanthanoid ions to detect the structural change of both ligands and proteins.
Here, we searched for binding ligands to Grb2 SH2 by using paramagnetic lanthanoid ions. First, binding ligands were screened using the relaxation effect induced by Gd3+. We screened one ligand from the mixture of 10 compounds. Next, the induced shifts of both Grb2 SH2 and the ligand were analyzed that enabled to determine the ligand binding site and binding mode on the protein. The merit of the present strategy is to obtain a structural information of the bound ligand at the primary stage of drug screening.
4) Development of antibody drugs for human Killer cell immunoglobulin-like receptors
To maintain immune system homeostasis, important cell surface receptors selectively recognize specific ligands. However, in the development of therapeutic agents against infectious diseases and immune diseases, the discovery of low-molecular-weight compounds targeted to such cell surface receptors is often difficult, because the interfacial surfaces involved in protein-protein interactions are flat. In this study, we aim to develop antibody drugs specific for immune cell surface receptors, based on three-dimensional structural analyses. We intend to properly control the function of the Killer cell Ig-like receptors (KIRs), which are expressed on natural killer cells and some subsets of T cells and are important for the elimination of virus-infected and cancer cells.
Cell surface receptors, unlike common soluble proteins, are very difficult to manipulate because of glycosylations and disulfide bonds. We have developed multiple preparation techniques for cell surface receptors: 1) E. coli expression as inclusion bodies and refolding, 2) transient expression in human HEK cells, 3) silkworm expression by direct inoculation of baculovirus plasmid DNA. Using these expression technologies, we can prepare large amounts of antibody drugs and their target proteins. In addition, since we have determined the three-dimensional structures of these cell surface receptors and their complexes, the elucidation of the structural and interaction parameters for the KIR recognition mechanism by monoclonal antibodies will confer enhanced specificity in the development of high-performance antibody drugs.