Enabling structure-based drug design of Tyk2 through co-crystallization with a stabilizing aminoindazole inhibitor

Abstract

Background: Structure-based drug design (SBDD) can accelerate inhibitor lead design and optimization, and efficient methods including protein purification, characterization, crystallization, and high-resolution diffraction are all needed for rapid, iterative structure determination. Janus kinases are important targets that are amenable to structure-based drug design. Here we present the first mouse Tyk2 crystal structures, which are complexed to 3-aminoindazole compounds.

Results: A comprehensive construct design effort included N- and C-terminal variations, kinase-inactive mutations, and multiple species orthologs. High-throughput cloning and expression methods were coupled with an abbreviated purification protocol to optimize protein solubility and stability. In total, 50 Tyk2 constructs were generated. Many displayed poor expression, inadequate solubility, or incomplete affinity tag processing. One kinase-inactive murine Tyk2 construct, complexed with an ATP-competitive 3-aminoindazole inhibitor, provided crystals that diffracted to 2.5-2.6 Å resolution. This structure revealed initial "hot-spot" regions for SBDD, and provided a robust platform for ligand soaking experiments. Compared to previously reported human Tyk2 inhibitor crystal structures (Chrencik et al. (2010) J Mol Biol 400:413), our structures revealed a key difference in the glycine-rich loop conformation that is induced by the inhibitor. Ligand binding also conferred resistance to proteolytic degradation by thermolysin. As crystals could not be obtained with the unliganded enzyme, this enhanced stability is likely important for successful crystallization and inhibitor soaking methods.

Conclusions: Practical criteria for construct performance and prioritization, the optimization of purification protocols to enhance protein yields and stability, and use of high-throughput construct exploration enable structure determination methods early in the drug discovery process. Additionally, specific ligands stabilize Tyk2 protein and may thereby enable crystallization.

Figure 1

Structure of 3-aminoindazole inhibitors.

Figure 2

Tyk2 is protected from proteolysis by the addition of Compound 2. Shown are the combined, quantitated values of Tyk2 peaks of ~27 and ~29 kDa during digestion with thermolysin, in the absence or presence of 30 μM Compound 2. Intact Tyk2 protein runs as ~29 kDa. The ~27 kDa form produced by thermolysin digestion is unaffected by the presence of Compound 2.

Figure 3

Overlay of 4E20 (beige) with previously reported structure 3LXN (rose): The overall structural fold is preserved when comparing the two structures with an r.m.s.d of 0.5 Å. Compound 1 is shown in orange.

Figure 4

Tyk2 sequence variation enables novel crystal contacts.a: Sequence alignment of human and mouse Tyk2 using CLUSTALW [38]. The residues highlighted in the red box are the glycine rich loop. b: Location of mouse Tyk2 surface residue Gly928 (Asp935 in human Tyk2) permits a close, van der Waals crystal contact. (Picture generated with COOT) [36].

Figure 5

Mouse Tyk2/inhibitor crystal structures. a: Structure of Compound 1 (orange) complexed to mouse Tyk2 with experimental Fo–Fc electron density contoured at 2σ. The 3-aminoindazole moiety displays three hinge interactions, the m-chlorophenyl group is located underneath the glycine rich loop and the phenyl sulfonamide linker is stabilized by surrounding Arg1020, Asn1021 and Glu898, residues. b: Structure of the Tyk2/Compound 2 complex shows similar interactions with experimental Fo–Fc electron density contoured at 2σ.

Figure 6

3-Aminoindazoles stabilize an open conformation of the Tyk2 glycine-rich loop. The chlorophenyl group of Compound 1 occupies hydrophobic pocket under the glycine rich loop. Overlay of mouse Tyk2 /Compound 2 (tan) and human Tyk2/CMP-6 (PDB entry 3LXP, cyan) structures. Deviation in the positioning of the glycine-rich loop tip is demonstrated.

Figure 7

Step a. Synthetic route to 4-(3-Cyano-4-fluoro-phenyl) aniline.

Figure 8

Step b. Synthetic route to Compound 1.

Figure 9

Synthetic route to Compound 2.