STAT3 is activated by various cytokines and growth factors through phosphorylation of Tyr-705 (pY-STAT3); pY-STAT3 plays a key role in numerous cellular processes such as cell growth and apoptosis resistance. Increased levels of pY-STAT3 have been observed in many cancers and have been associated with poor prognosis. The STAT3 SH2 domain is required for two essential steps in STAT3 activation: recruitment of latent STAT3 to the activated, kinase-containing receptor complex, which leads to its phosphorylation on Tyr-705, and STAT3 tail-to-tail dimerization. Structure-based drug design has been used to identify small-molecule inhibitors that target the STAT3 SH2 domain with the goal of developing potent anti-cancer drugs. However, the few small-molecule STAT3 inhibitors reported to date have, at best, moderate potency, blocking STAT3 activation with IC50s in the 5–10μM range. One reason for this modest success may be the highly flexible nature of the STAT3 SH2 domain, which results in unreliable modeled structures of the STAT3 SH2 domain bound by inhibitors that are based on X-ray crystallography. In order to improve structure-based drug design, it will be helpful to examine the protein dynamics of the STAT3 SH2 and its structural details in the context of the STAT3 dimer, unbound monomer, and ligand-bound monomers. Here, we performed homology modeling, flexible docking, and molecular dynamics (MD) simulations on the STAT3 SH2 domain alone and in a complex with a STAT3 partner (STAT3 dimer), a phosphotyrosyl peptide derived from gp130, and the peptidomimetic CJ-887. The STAT3 dimer and the STAT3-gp130 peptide complexes share many features, including the H-bond network that binds the phosphorylated tyrosine, the double H-bonds between the binding peptide chain and residue 636–638 backbones, and hydrophobic contacts between the binding residue side chains and the SH2 domain. The gp130 peptide-binding system contains an extra H-bond network by virtue of its Gln-908 side chain being engaged by the STAT3 Gln-644 side chain and the Glu-638 backbone carbonyl. Furthermore, MD simulation (1-μs) of the ligand-free STAT3 monomer indicates that the phosphotyrosine binding pocket and nearby regions are stable in the absence of ligand binding although the extended parts of the SH2 domain are considerably flexible. The structural details in the CJ-887 binding are largely similar to those of the gp130-binding system but extra H-bonds and hydrophobic contacts are identified that may contribute to its increased binding affinity. These findings provide new support at the atomic level to the design and optimization of novel non-peptide STAT3 inhibitors, especially organic molecules obeying Lipinski’s rule of five.
Read full abstract