KIT1_in_complex_with_DP_2976

KIT1 in complex with DP-2976 (PDB ID: 6MOB) from Homo sapiens
Created By: Shomik Ghose

            KIT1 (PDB ID: 6MOB), also known as Mast/stem cell growth factor receptor KIT, is an important class III receptor tyrosine kinase and proto-oncogene in H. sapiens. KIT is a cell surface receptor for the cytokine stem cell factor (SCF); upon binding SCF, KIT forms an active phosphorylating homodimer which initiates signaling cascades essential for cell survival, proliferation, hematopoiesis, and stem cell maintenance (1 – 3). As a proto-oncogene, mutations in KIT are likely to lead to cancer. KIT mutations have been identified in over 90% of gastrointestinal stromal tumors and as such, kinase inhibitors are a commonly employed form of treatment; however, subsequent mutations specific to the kinase domain of KIT leave many conventional tyrosine kinase inhibitors ineffective, creating increasingly potent drug-resistant gastrointestinal tumors (1, 4, 5). A better understanding of the structure of the KIT kinase domain could reveal the origins of this drug resistance and improve cancer pharmaceuticals.

            In order to obtain the KIT protein for crystallography, the H. sapiens KIT gene was modified to create a synthetic gene containing only the KIT kinase domain (1). Bacterial transformation techniques were employed to amplify and purify the KIT. Prior to crystallography, the KIT was incubated for 8hr with a five-molar excess of DP-2976a switch control kinase inhibitor drug, to complex the protein with the drug. The crystals were grown through the vapor diffusion method set up with a 0.5 µL/0.5 µL protein to crystallant ratio, with the crystallant consisting of 2.5 M ammonium nitrate and 0.1 M sodium acetate with a pH of 4.6. As ammonium nitrate is used as a crystallant for the vapor diffusion process, non-functional nitrate ions are present as a ligand in the crystal structure. Structural data of KIT was obtained through X-ray diffraction using synchrotron radiation. The structure was derived using molecular replacement in PHASER, modified through COOT Crystallogr, and refined through REFMAC5 (1).

            Per the Expasy proteomics server, KIT possesses an isoelectric point of 7.16 and a molecular weight of 35289.90 Da (6). In its inactive form, a complete biomolecule of KIT consists of a single subunit with 976 total residues, which can be broken up into four functionally distinct domains (2, 3). Like all members of the receptor tyrosine kinase family, KIT consists of an extracellular domain, a transmembrane domain, an autoinhibitory juxtamembrane domain, and a kinase domain. KIT is specifically a class III receptor tyrosine kinase, thus the extracellular domain consists of five immunoglobulin-like loops. SCF binds to the second and third immunoglobulin domains, and the fourth is important in the dimerization of KIT into an active phosphorylating complex (3, 7). Class III receptor tyrosine-kinases are further differentiated by their distinct kinase domain structure. The kinase domain is interrupted by an 80 residue kinase insert in the ATP-binding region; this hydrophobic insert uniquely allows class III receptor tyrosine-kinases to bind important signal transduction molecules such as adaptor proteins, enabling them to interact with the kinase domain (1, 3, 7). In the distal portion of the kinase domain is an important region called the activation loop which helps to stabilize KIT in an active conformation (1, 7). In the crystallized KIT, the kinase domain has been selectively isolated for further study: the isolated section begins at amino acid Asn-565 in the kinase domain and extends to amino acid Glu-935, excluding the Gln-694 to Thr-753 kinase insert domain (1).

            The primary structure of the crystallized KIT is a 312 residue chain of amino acids linked with peptide bonds (1). The secondary structure of the crystallized KIT kinase domain consists 12 β-sheet strands from 55 residues (17%) and 16 helices with 8 α-helices and 8 3/10 helices from 120 residues (38%). The remaining 45% of the molecule is composed of random coil (2). All 12 β-sheets in the biomolecule are oriented in an anti-parallel fashion. Additionally, the β-sheets are composed of predominantly hydrophobic, nonpolar amino acids whereas the helices and random coil are largely amphiphilic. This finding is consistent with the fact that the kinase domain of the protein is present in the polar cytoplasmic environment (2). KIT’s tertiary structure closely matches the classic two-lobe structure of a protein kinase domain: residues Trp-582 to Glu-671 compose a small N-terminal lobe and residues Leu-678 to Ile-953 compose the larger C-terminal lobe. This lobe-based segregation is consistent with findings from the secondary structure’s emergent α+β architecture, as the small lobe is composed of primarily β-sheets whereas the large lobe is mostly composed of helices (7). While inactive, KIT is monomeric and exhibits no quaternary structure; however, when SCF binds to the extracellular domain, a homodimer forms to activate the phosphorylating activity of the tyrosine kinase (1, 3).   

            Prior to discussing the phosphorylating activity of active KIT, it is essential to understand the conformational changes which induce the active state and how the structure of KIT can bring about this functional change. As previously mentioned, the tertiary structure of KIT forms two lobes: the glycine-rich small lobe responsible for binding ATP, and the large lobe responsible for binding the protein substrate – these lobes can move relative to each other to expose or hide the active site and thus control the function of the kinase (7). KIT can exist in two conformations: a type II inactive form where the lobes are closed and KIT is a dormant monomer and a type I active form where the lobes are open, the active site is available and KIT is an actively phosphorylating homodimer (1). The distal kinase domain contains a region called the activation loop which is essential in determining KIT conformation via modulating the position of the large lobe (7). When the key Tyr-823 residue in this activation loop is unphosphorylated, it forms a hydrogen bond with the catalytic Asp-792 residue, preventing kinase activity in the type II conformation (1, 7). However, when Tyr-823 is phosphorylated, it changes the conformation of the activation loop, moving Phe-811 into position to allow for physical acceptance by a ring of amino acids (Val-643, Leu-647, Ile-653, and Val-654). This ring-Phe interaction helps to stabilize the activation loop and the active type I form of KIT (1,7). Through this mechanism, structural control over the activation loop exerts functional control over KIT activity.

            The key reaction carried out by KIT is the phosphorylation of protein tyrosine residues to enable downstream signal cascades (1, 3). This is achieved by a set of catalytic residues present in the active site: Lys-623 forms ionic bonds with β-, and γ- phosphates in ATP reactants, Asp-792 correctly orients the substrate tyrosine and abstracts away a proton to allow the side chain tyrosine to nucleophilically attack the γ-phosphate and complete the phosphorylation (7). It is important to note that this fairly simple mechanism is highly dependent on the correct positioning of Asp-792 – if the Tyr-823 is not phosphorylated, it will be hydrogen bonded to Asp-792 and thereby inhibit the catalysis of phosphorylation (1, 7). The kinase insert domain notably has the ability to bind different adaptor proteins when phosphorylated; for instance, phosphorylating Tyr-703 can attract the adaptor protein Grb2. These interactions allow for KIT to interact with a wide variety of signaling cascades (7). The dual lobe and activation loop structures thus clearly dictate the phosphorylating function of the kinase domain.

In the crystallized structure of KIT, the compound is complexed with the drug DP-2976, forming a structure-drug system (1). DP-2976 is a close chemical analog of a potential new tyrosine kinase inhibitor, DCC-2618 or Ripretinib. Ripretinib seeks to surpass existing tyrosine-kinase inhibitor treatments by not inducing drug-resistance in gastrointestinal stromal tumors through more robust kinase domain inhibition. To achieve this, Ripretinib occupies the ATP binding site and forms hydrogen bonds with the activation loop complex to lock KIT into the type II inactive form (1). Structurally, after occupying the ATP active site, Ripretinib’s carbonyl oxygen forms hydrogen bonds with the catalytic residue Asp-792 and surrounding residues (His-790 and Arg-796). These hydrogen bonds nucleated by the inhibitor then induce and stabilize the key Tyr-823 residue as a decoy substrate, locking the key tyrosine residue and the remainder of the activation loop into an inactive Type II conformation (1). Thus, the structural changes induced by the active site binding of Ripretinib result in functional inhibition of KIT kinase activity.

A position-specific iterated basic local alignment search tool (PSI-BLAST) query was performed using the amino acid sequence of the KIT kinase domain to find proteins with similar primary structure.  The PSI-BLAST query aligns protein amino acid sequences and assigns gaps where amino acids are present in the provided protein but not in the query proteins, then it quantifies the similarity between pairs of proteins as E-values (8). An E-value below 0.05 indicates a high degree of similarity in the primary structure of the protein pair. Fibroblast Growth Factor Receptor 4 (PDB ID: 5JKG_A) or FGFR4 from H. sapiens has an E-value of 2*10-116 indicating that it has a very similar primary structure to KIT (8). The Dali server was utilized to find proteins with similar tertiary structure to the KIT kinase domain. The Dali server uses the sum-of-pairs method to evaluate intermolecular distances and identify proteins with similar tertiary folding patterns to the given supplied protein (9). Note that this approach only works for proteins as it requires an amino acid backbone. The server calculates Z-scores to denote similarity of tertiary structure between proteins; a Z score of 2 or greater indicates a high degree of similarity in tertiary structures. FGFR4 from H. sapiens has a Z-score of 33.5 indicating that this biomolecule has a very similar folding pattern to KIT (9).  

Both KIT and FGFR4 are part of the broad receptor tyrosine kinase (RTK) family; KIT is a class III RTK whereas FGFR4 is a class V RTK. As such, FGFR and KIT share a common general structure, including an extracellular domain, a transmembrane domain, a juxtamembrane domain, and a kinase domain (7, 10). However, they differ slightly in the nuances of each – for instance, in FGFR4, the extracellular domain only has three immunoglobulin-like loops which uniquely present a set of acidic amino acids - this allows for increased regulation over the selectivity of its fibroblast growth factor ligand (10). Focusing on the kinase domains, mutations in this region of FGFR4 have been shown to cause several types of cancer (2, 10). Structurally, both proteins display a double lobe structure with an activation loop, but with slight differences in primary structure as FGFR4 notably lacks the defining kinase insert domain of KIT.There are also notable differences in secondary structure which manifest as differences in the function of the activation loop system. FGFR has 12 helices and thus lacks the large-lobe helix characteristic of a central activation loop residue like KIT’s Tyr-823  In FGFR4, inhibitory juxtamembrane residues are comparatively more important in the control of the two lobes and thus active site exposure. Finally, as a result of this different active site structure, FGFR tyrosine kinase inhibitors such as LY2874455, a pan-FGFR inhibitor drug, function slightly differently; rather than inducing a conformational change in the kinase domain, FGFR4 inhibitors create extensive hydrogen bonding in the ATP binding domain, sterically inhibiting the functioning of the kinase (10).

Receptor tyrosine kinases such as KIT and FGFR4 are increasingly important to the understanding of signaling pathways in H. sapiens and how these pathways can be disrupted. Early research has made clear the connection between cancers such as gastrointestinal stromal tumors and disruptions to kinase activity; however, the exact role of the nuances between differing kinase domains in the presentation and appropriate treatment of these diseases remains unclear. Moving forward, it is imperative that there be more investigation into how to best treat kinase-domain related cancers to prevent the proliferation of lethal drug-resistant tumors.