Missense mutations in the tyrosine kinase domain of the MET proto‐oncogene occur in selected cases of papillary renal carcinoma. In biochemical and biological assays, these mutations produced constitutive activation of the MET kinase and led to tumor formation in nude mice. Some mutations caused transformation of NIH 3T3 cells. To elucidate the mechanism of ligand‐independent MET kinase activation by point mutations, we constructed several 3D models of the wild‐type and mutated MET catalytic core domains. Analysis of these structures showed that some mutations (e.g., V1110I, Y1248H/D/C, M1268T) directly alter contacts between residues from the activation loop in its inhibitory conformation and those from the main body of the catalytic domain; others (e.g., M1149T, L1213V) increase flexibility at the critical points of the tertiary structure and facilitate subdomain movements. Mutation D1246N plays a role in stabilizing the active form of the enzyme. Mutation M1268T affects the S+1 and S+3 substrate‐binding pockets. Models implicate that although these changes do not compromise the affinity toward the C‐terminal autophosphorylation site of the MET protein, they allow for binding of the substrate for the c‐Abl tyrosine kinase. We provide biochemical data supporting this observation. Mutation L1213V affects the conformation of Tyr1212 in the active form of MET. Several somatic mutations are clustered at the surface of the catalytic domain in close vicinity of the probable location of the MET C‐terminal docking site for cytoplasmic effectors. Proteins 2001;44:32–43. © 2001 Wiley‐Liss, Inc.