Engineered variants of InlB with an additional leucine-rich repeat discriminate between physiologically relevant and packing contacts in crystal structures of the InlB:MET complex.

ABSTRACT

The physiological relevance of contacts in crystal lattices often remains elusive. This was also the case for the complex between the invasion protein internalin B (InlB) from Listeria monocytogenes and its host cell receptor, the human receptor tyrosine kinase (RTK) MET. InlB is a MET agonist and induces bacterial host cell invasion. Activation of RTKs generally involves ligand-induced dimerization of the receptor ectodomain. The two currently available crystal structures of the InlBMET complex show the same arrangement of InlB and MET in a 11 complex, but different dimeric 22 assemblies. Only one of these 22 assemblies is predicted to be stable by a computational procedure. This assembly is mainly stabilized by a contact between the Cap domain of InlB from one and the Sema domain of MET from another 11 complex. Here, we probe the physiological relevance of this interaction. We generated variants of the leucine-rich repeat (LRR) protein InlB by inserting an additional repeat between the first and the second LRR. This should allow formation of the 11 complex but disrupt the potential 22 complex involving the Cap-Sema contact due to steric distortions. A crystal structure of one of the engineered proteins showed that it folded properly. Binding affinity to MET was comparable to that of wild-type InlB. The InlB variant induced MET phosphorylation and cell scatter like wild-type InlB. These results suggest that the Cap-Sema interaction is not physiologically relevant and support the previously proposed assembly, in which a 22 InlBMET complex is built around a ligand dimer.