Hepatocyte Growth Factor Receptor c-Met

Introduction

The Hepatocyte Growth Factor Receptor plays a large role in embryonic development, as its activation leads to events such as cell growth, motility and invasion. This receptor is a Tyrosine Kinase, and is one of the most well studied RTKs, as mutations in the c-met proto-oncogene can lead to the formation of tumors. The ligand for this receptor is Hepatocyte growth factor/scatter factor (HGF/SF), and upon binding of this ligand, the receptor becomes auto phosphorylated, causing downstream signaling events such as cell growth. ^[1] C-Met is an αβ heterodimer with extracellular and intracellular domains. These two domains are disulfide linked together. ^[2] This particular structure is one of a mutated hepatocyte growth factor tyrosine kinase domain, which is part of the intracellular β subunit. ^[3]

Structure

The A loop of the wild type receptor contains two tyrosines at position 1234 and 1235. When these two residues become phosphorylated, the kinase can become active. A unique part of the c-met structure is the pair of . Studies have shown that these tyrosines are necessary for normal c-met signaling. When these two tyrosines were substituted with with phenylalanine, in mice, the mice had an embryonically lethal phenotype and defects were found in placenta, liver, muscles and nerves. ^[4] In a wild type c-met, these sites will become phosphorylated and act as docking sites for many different transducers and adapters. ^[5] Upon phosphorylation, these tyrosines can bind with Src homology 2 (SH2) domains and phophotyrosine-binding (PTB), and therefore bind many effectors that will cause downstream effects such as cell proliferation, scattering and inhibition of apoptosis. ^[6]

This receptor follows the typical structure of a protein kinase, with a bilobal structure. The N-terminal contains and is linked through a hinge to the C lobe, which is full of α helices. This particular kinase domain is very similar to the domains of the insulin receptor kinase and fibroblast growth factor receptor kinase.^[7]

Helices

This structure is made up of many α helical structures that move in the transformation from inactive to active kinase. Some of these helices are conserved in many different tyrosine kinases. C-met does show a divergence from other tyrosine kinases (such as IRK and FGFRK) in the helix formed at the N-terminus, before the core kinase domain, in residues . ^[8] The αA is in contact with αC and so causes αC to be in a slightly different orientation than in FGFRK and IRK. Residues Leu-1062, Val-1066, and Val-1069 of αA with with residues Leu-1125 and Ile-1129 of αC. There is another between the residues Ile-1053, Leu-1055 and Leu-1058 of αA and Ile-1118 and Val-1121 of αC. Because of the movement of αC during activation of the kinase, it is an assumption that αA is also part of the kinase activation upon ligand binding. ^[9]

Mutation

This particular structure of the hepatocyte growth factor tyrosine kinase domain is one harboring a human cancer mutation. The two are replaced by a phenylalanine and aspartate, respectively. This mutation normally causes the receptor to be constitutively active, and is found in HNSC (Head, Neck squamous cell) carcinoma. Although there is no longer phosphorylation at these sites, it is believed that the negative charge of the aspartate resembles the negative phosphate that would normally cause activation, and therefore keeps the protein in its active form. ^[10] There is a third mutation at Tyr-1194 which is substituted for a . This is shown to point into the formed by Lys-1198 and Leu-1195 from αE. ^[11] This structure is conserved in the wild type protein, suggesting that the mutation at residue 1149 is not changing the structure at this position.

K-252a

is a staurosporine analog. Staurosporine is an inhibitor of many Ser/Thr Kinases, and has been shown to also inhibit c-Met activation by inhibiting its autophosphorylation. The structures of K-252a and staurosporine are very similar, with the main difference being that K-252a has a furanose instead of a pyranose structure. The binding of K-252a causes the c-Met to adopt an inhibitory conformation of the A-loop, specifically with residues . This segment blocks the place where the substrate tyrosine side chain would bind, if the protein were in an active conformation. Residues also enhance this inhibitory conformation, as they constrain αC into a conformation that does not allow the catalytic placement of keeping αC in an inactive conformation. In an active kinase, Glu-1127 would form a salt bridge with Lys-1110. Residues 1229-1230 pass through the triphosphate subsite of bound ATP blocking ATP binding. The K-252a itself binds in the adenosine pocket, therefore also inhibiting the binding of ATP. The binding of K-252a is very favorable (enthalpy change of -17.9 kcal/mol). This is probably due to polar interactions as well as a change in conformation upon binding. ^[12]

There is a concerted conformational change in the complex upon K-252a binding. One of these changes involves the A-loop, specifically residues . In the Apo-Met structure, the side chain of Met-1229 would pass through the ring of the inhibitor, and so, in order to make room for K-252a, the segment must move, with residues 1229 and 1230 moving by 3-4 Å. In order to make room for the side chain of Tyr-1230, moves by 8 Å toward . Arg-1208, which in the uninhibited complex would stack with tyr-1230, now stacks with

^[13]

K-252a binds in the adenosine pocket. It has four hydrogen bonds to the enzyme, with of these mimicking hydrogen bonds of an adenine base. There is a hydrogen bond between the K-252a nitrogen and the carbonyl oxygen of Pro-1158, and another between the K252-a carbonyl oxygen and the hydrogen of the amide of Met-1160. There are two more hydrogen bonds between the 3' hydroxyl and carbonyl oxygen and the of the A loop. ^[14]

There are also many hydrophobic interactions between the interface of the enzyme and K-252a. The residues involved in this are Ile-1084, Gly-1085, Phe-1089, Val-1092, Ala-1108, Lys-1110, and Leu-1140 (); Leu-1157, Pro-1158, Tyr-1159, and Met-1160 (); and Met-1211, Ala-1226, Asp-1228, Met-1229, and Tyr-1230 (). ^[15]

Met-1229, Met-1211 and Met-1160 all make up the for the indolocarbazole plane as they are all within van der waals distance of it. ^[16]

C-Terminal Docking Site

In c-Met, there are two tyrosines located in the C-terminal tail sequence, which, upon phosphorylation, act as the docking sites for many signal transducers. These tyrosines correspond to residues . Both of these sites interact with SH2, MBD and PTD domains of signal transducers. The residues form an extended conformation, which is seen in other phosphopeptides that bind to SH2 domains. Residues form a type I β turn, which is similar to sequences that bind to Shc-PTB domians. Whether binding to SH2 domains or PTB domains, upon binding, these motifs would move to avoid clashes with the C lobe. The third binding motif is found in residues , which form a type II β turn, and is similar to pohsphopeptides that bind Grb2. When comparing the unphosphorylated conformation of the motif to one that is phosphorylated, and bound to the Grb2 complex, there is a peptide flip between the bind of . This suggests that when Grb2 docks onto c-Met, there is a change in orientation of this motif. These three binding motifs of the mutated structure are very similar to binding motifs that would be recognized by their binding partners, implying that the C-terminal supersite of this structure is very similar to that of an active c-met. ^[17]

Biological Significance

Many human cancers such as HNSC carcinoma, papillary renal carcinoma and hepatocellular carcinoma, can be traced back to mutations in the c-met kinase domain. The mutations found in c-Met often lead to an over activation of this kinase. Some of the following mutations are known to cause an over active c-met: V1092I, H1094L/Y/R, H1106D, M1131T, V1188L, L1195V, V1220I, D1228H/N, Y1230H/C, Y1235D, K1244R, and M1250T/I. Many of these mutations most likely affect the A loop conformation of the wild type receptor, causing it to become constitutively active. This is done by either stabilizing the active form of the enzyme or destabilizing the inactive form.

The significance of this particular structure is that it shows the A loop of c-Met is flexible and will adapt in order to bind to an indolocarbazole (K-252a). This gives insight on designs for specific c-Met inhibitors, as this specific site can be targeted by drugs to block c-Met activity.

This structure also shows the binding motifs of c-Met in an unphosphorylated form, giving insight on how the motifs may move when interacting with their respective binding domains (Grb2, SH2, PTB domains).