Functional Oligomerization of the EphA2 Receptor Tyrosine Kinase

Abstract

Eph receptor tyrosine kinases constitute the largest family of receptor tyrosine kinases (RTKs). Aberration of Eph/ephrin system leads to many pathologies, most notably in cancer. EphA2 is notorious for overexpression in a variety of human malignancies, in which the loss of ligand is often observed. Although the activation of EphA2 RTK is known to suppress the signaling of oncogenic kinases, such as Ras and Akt, the unliganded EphA2 receptor becomes a target for oncogenic kinases including Akt, ERK and PKA that phosphorylates EphA2 on serine 897 (S897). Remarkably the latter event turns EphA2 from a tumor suppressor into an oncogenic protein that drives tumor cell migration. We set out to understand the structural and cellular basis underlying the ligand-dependent tumor suppressive and ligand-independent pro-oncogenic functions of EphA2. We used a time-resolved fluorescence spectroscopy (PIE-FCCS) to study the assemblies of EphA2 in live cell membranes. The results show that EphA2 undergo multimerization prior to ligand binding, and polymerize upon ligand stimulation. We identified three sets of interfaces in the ecto-domain that drive the ligand-independent multimerization and ligand-induced polymerization of EphA2. These three sets of interfaces can be divided into two types of contact. The first one is symmetric contact driven by LBD-LBD and Sushi-Sushi interfaces. The second one is asymmetric contact via LBD interacting with FN2 domain on the neighboring molecule. Further studies reveal that the symmetric contact mediates the oligomerization and activation of kinases upon ligand stimulation, while the asymmetric contact keeps kinases separated facilitating the pro-migratory ligand-independent signaling via phosphorylation of S897. These results add a paradigm to the general framework of RTKs, and also suggest that the loss of ligands in many tumor tissues favors the asymmetric contact in the assemblies of overexpressed EphA2, promoting the ligand-independent oncogenic signaling.