Crystal structure of the kinase domain of a receptor tyrosine kinase from a choanoflagellate, Monosiga brevicollis

Genomic analysis of the unicellular choanoflagellate, Monosiga brevicollis (MB), revealed the remarkable presence of cell signaling and adhesion protein domains that are characteristically associated with metazoans. Strikingly, receptor tyrosine kinases, one of the most critical elements of signal transduction and communication in metazoans, are present in choanoflagellates. We determined the crystal structure at 1.95 Å resolution of the kinase domain of the M. brevicollis receptor tyrosine kinase C8 (RTKC8, a member of the choanoflagellate receptor tyrosine kinase C family) bound to the kinase inhibitor staurospaurine. The chonanoflagellate kinase domain is closely related in sequence to mammalian tyrosine kinases (~ 40% sequence identity to the human Ephrin kinase domain EphA3) and, as expected, has the canonical protein kinase fold. The kinase is structurally most similar to human Ephrin (EphA5), even though the extracellular sensor domain is completely different from that of Ephrin. The RTKC8 kinase domain is in an active conformation, with two staurosporine molecules bound to the kinase, one at the active site and another at the peptide-substrate binding site. To our knowledge this is the first example of staurospaurine binding in the Aurora A activation segment (AAS). We also show that the RTKC8 kinase domain can phosphorylate tyrosine residues in peptides from its C-terminal tail segment which is presumably the mechanism by which it transmits the extracellular stimuli to alter cellular function.


Introduction
Receptor tyrosine kinases are crucial signaling molecules in animal cells, regulating inter-cellular responses to extracellar signals via numerous different pathways dependent on the organism involved. These receptors consist of an extracellular ligand-binding domain, a single transmembrane helix, and a cytoplasmic tyrosine kinase domain [1,2]. Binding of ligands to the extracellular domain of the receptor leads to dimerization and activation of the kinase domain. The activated kinase domain then autophosphorylates tyrosine residues present in the activation loop in the kinase domain, thus sustaining or amplifying the activation. Other sites of phosphorylation in the C-terminal tail of the receptor usually are involved in recruitment of further components of the signaling pathway. In some instances, loops in the kinase domain, or in the juxtamembrane segment bridging the transmembrane helix and the kinase domain are phosphorylated to alter the kinase activity. Thus, in summary, phosphorylated tyrosine residues switch on kinase activity, or act as docking sites for proteins that transmit the signal further downstream [3][4][5].
The discovery of numerous genes encoding receptor tyrosine kinases in choanoflagellates was remarkable [6][7][8][9][10]. Phylogenetic studies suggest that choanoflagellates are the closest living relatives to metazoans, indicating that receptor tyrosine kinases were present in the common ancestor of choanoflagellates and metazoans. The choanoflagellate M. brevicollis contains 128 protein kinases, including 88 receptor tyrosine kinases [11]. These choanoflagellate receptor tyrosine kinases are classified into 15 families (denoted A through M, FGTK and LRTK) [11]. They are not direct orthologs of metazoan receptor tyrosine kinases as they have extracellular domains that appear unrelated to those of metazoan receptor tyrosine kinases [11]. The structural and functional aspects of choanoflagellate receptor tyrosine kinases are still relatively unexplored.
The kinase described in this study is from the M. brevicollis receptor tyrosine kinase C (RTKC) family. This family has 10 members (denoted RTKC1 to RTKC10). The RTKC family is a distinct clade of kinases and shares a common ancestor with the Ephrin kinases but does not share all of the Ephrin-specific conserved motifs [12]. Most of the members of this family contain Cys-rich and Hyalin-rich (HYR) extracellular domains, a transmembrane domain, a kinase domain and a C-terminal tail with a CAP-Gly domain. The HYR domain is predicted to be related to immunoglobulin (Ig) and fibronectin type3 (FN3) domains in metazoans, and might be involved in cell adhesion [11,13]. The metazoan CAP-Gly domains bind to microtubules [14]. The presence of a CAP-Gly domain in M. brevicollis C family receptor tyrosine kinases therefore suggests that these kinases may interact with the cytoskeleton [11]. The M. brevicollis RTKC8 gene codes for a 2031 amino acid protein, It consists of an extracellular domain of~1487 residues containing two Cys-rich domains and four HYR domains, a transmembrane domain (TM) domain of~19 residues and an intracellular domain consisting of a juxtamembrane domain (JM,~37 residues) a kinase domain (~254 residues) and a C-terminal tail containing a 40 residue CAP-Gly domain and 10 tyrosine residues (Fig 1).
We determined the crystal structure of the RTKC8 kinase domain to 1.95 Å resolution. The RTKC8 kinase domain is co-crystallized with the kinase inhibitor staurospaurine in two positions, one the canonical active site position and one in the peptide binding site or Aurora A activation segment (AAS). We also demonstrated the kinase is active on tyrosine residue containing peptides derived from its C-terminal tail segment.

Cloning of the RTKC8 kinase domain
The sequence of the RTKC8 kinase matches that reported in the Kinase.com database (http:// kinase.com/kinbase) [11]. The DNA (NCBI accession id: XM_001750163.1), encoding for the RTKC8 kinase domain (1543-1796) flanked by a 13 residues of the preceeding sequence and 15 residues of the following sequence, residues 1530-1811, was amplified using standard PCR protocols from a Monosiga cDNA library using the forward primer 5'AGGGCCATATGCA GCTTTCCAAGGAGCCGCGCG3' and the reverse primer 5'CCTTTGAATTCTCAGCTTT GGAAGAGGCTGGACATGTCCGAATCAGT3'. The PCR product size was confirmed using agarose gel electrophoresis. The amplified DNA was ligated into the pFastbac1 insect-cell expression vector using EcoRI (5') and NdeI (3') restriction enzyme sites. The ligated product was transformed into E. coli Top10 chemically competent cells and was selected with Ampicillin on LB Agar. Next day, overnight cultures were set up from the Top10 bacterial colonies to isolate the plasmid and the plasmid was sequenced to confirm the cloned construct. The expressed construct consisted of an N-terminal hexahistidine tag and a TEV protease site followed by the M. brevicollis TRKC8 core kinase domain with short flanking sections from the RTKC8 protein sequence either side (S1A Fig).

Purification of the RTKC8 kinase domain
To purify the protein, sf9 cells were thawed and lysed by passing three times through an Avestin EmulsiFlex C50 cell homogenizer with 1500 psi pressure at 4˚C. The suspension was clarified by ultracentrifugation at 40,000 rpm for 1 hour at 4˚C. The supernatant was filtered using a 0.45 μM syringe filter, loaded onto a Ni-NTA affinity column (GE Healthcare Life Sciences) pre-equilibrated with Buffer A (50 mM Tris-HCl pH 8.0, 500 mM NaCl, 10% (v/v) glycerol, 25 mM imidazole, 0.5 mM TCEP) and the flow-through was collected. The resin was washed with 20 column volumes of Buffer A to remove the non-specific proteins. Then, the RTKC8 protein was eluted in Buffer B (50 mM Tris-HCl pH 8.0, 200 mM NaCl, 10% (v/v) glycerol, 250 mM imidazole, 0.5 mM TCEP). To cleave the His-tag, TEV protease was added to the pooled fractions containing the protein and dialyzed into Buffer C (50 mM Tris-HCl pH 8.0, 200 mM NaCl, 10% (v/v) glycerol, 0.5 mM TCEP) overnight at 4˚C. A subtractive Ni-NTA purification was run to remove TEV protease and other impurities. The flow through was collected, concentrated and purified on a Superdex 200 16/600 column (GE Healthcare Life Sciences) in Buffer C. Fractions containing RTKC8 kinase were either pooled and concentrated to 105 μM for biochemical experiments and stored at -80˚C or concentrated to 200 μM and used immediately for crystallization experiments. The purification yield was~1 mg of purified RTKC8 kinase from 1 liter of Sf9 insect cells. The elution volume from size exclusion was consistent with the molecular weight of a monomeric protein (28.5 KDa) (S1B Fig

Phosphorylation and dephosphorylation analysis of RTKC8 kinase domain
To determine the phosphorylation state of the purified RTKC8 kinase domain, the protein was probed using an anti-phosphotyrosine antibody (4G10) by western blot. The protein was run on 12% (w/v) acrylamide SDS-PAGE gel at a final concentration of 1 μM and transferred to polyvinylidene difluoride (PVDF) membrane (Millipore) using a semi-dry transfer apparatus. Protein samples were transferred using Towbin transfer buffer (25 mM tris, 192 mM glycine, 20% (v/v) methanol). The membrane was blocked using blocking solution (5% (w/v) BSA dissolved in Tris-buffered saline (20 mM Tris pH 7.5, 150 mM NaCl), 0.1% (v/v) Triton X-100). The membrane was probed using anti-phosphotyrosine 4G10 primary antibody (Millipore, 05-321) (1:3000), diluted in blocking solution overnight at 4˚C. Next day, the blot was washed with 1X Tris-buffered saline containing 0.1% (v/v) Triton X-100 four times for 5 minutes each with gentle shaking at room temperature. The blot was probed using anti-Mouse IgG secondary antibody linked with horse radish peroxidase (Cell Signaling Technology, #7076) (1:3000) for 1 hour at room temperature on shaker. The blot was washed again with 1X Tris-buffered saline solution containing 0.1% (v/v) Triton X-100 four times for 5 minutes each with gentle shaking at room temperature. The blot was developed using the Western Bright kit (Advansta) and imaged using a Bio-Rad Gel Doc.
For dephosphorylation assays, the RTKC8 kinase domain and YopH phosphatase, each at a final concentration of 1 μM, were incubated together at room temperature and the dephosphorylation reaction was stopped at different time intervals by adding SDS-buffer. Phosphotyrosine levels were detected by western blot as described above.

In vitro kinase activity assays
Kinase activity measurements were carried out using a coupled kinase assay (Pyruvate Kinase/ Lactate Dehydrogenase (PK/LDH) ATPase assay) in presence of different substrate peptides [15,16]. This assay measures ADP production, which is coupled to NADH oxidation. In this experiment, the reaction solution contained 50 mM Tris-HCl pH 8, 100 mM NaCl, 10 mM MgCl2, 1 mM phosphoenolpyruvate (PEP), 300 mg/mL NADH, 2 mM sodium orthovanadate (Na 3 VO 4 ), 100 μM ATP, and an excess of pyruvate kinase and lactate dehydrogenase (approximately 120 units/mL and 80 units/mL, respectively, added from a commercially available mixture of the two enzymes). Human c-Src kinase domain phosphorylation of a preferred peptide substrate, PKCδ-Y313 (SSEPVGIYQGFEKKT) and a non-preferred peptide, PDGFRα-Y762 (SDIQRSLYDRPASAK), in which the Tyr 768 in the wildtype PDGFRα sequence was mutated to alanine in the peptide so as to present a single substrate tyrosine residue, were used as controls in these kinase assays [17]. The tyrosine phosphorylation kinetics for RTKC8 with these two previously mentioned peptides and three peptides from the RTKC8 C-terminal tail (purchased from Elim Biopharmaceuticals, Inc.), corresponding to the residues surrounding Tyr 1828 (EPETDEVYGNEGNVS), Tyr 1897 (GTVGEHEYFDCADQH), and Tyr 1990 (ASD-NELLYDMGRAEA) were measured. The reactions were initiated by the addition of either c-Src at 1 μM or RTKC8 kinase domain at 5 μM, to 250 mM peptide and were conducted in duplicate. The reaction progress was monitored by measuring absorbance at 340 nm every 30 s at 25˚C on a SpectraMax plate reader.

Crystallization and data collection
The freshly purified RTKC8 kinase domain (10-15 mg/mL, in Buffer C) was mixed with a final concentration of 1 mM staurosporine (10 mM stock, dissolved in 100% (v/v) DMSO) and incubated on rotator for 2 hours at room temperature. The crystallization trays (96 well) were set up using sitting-drop vapor diffusion with a Mosquito crystallization robot (TTP Labtech) using commercial sparse matrix screens. 0.1 μL of the protein was mixed with 0.1 μL of the reservoir solution. The drops were equilibrated against 50 μL of reservoir volume and the trays were incubated at 20˚C. A condition from the PACT screen (Qiagen) yielded crystals suitable for diffraction (0.1 M succinic acid, sodium dihydrogen phosphate and glycine (SPG) buffer pH 6.0, 25% (w/v) polyethylene glycol (PEG) 1500 [18]. The crystals directly from the 96 well screen were cryoprotected with a solution of the well condition with 20% (v/v) glycerol and Xray diffraction data were collected at Lawrence Berkeley National Laboratory Advanced Light Source, Beamline 8.2.1 with 0.9998 Å wavelength X-rays.

Structure determination
The reflections were indexed and integrated by XDS [19,20]. The data were scaled and merged with Pointless [21] and Aimless [22], in the CCP4 suite [23]. The structure was solved by molecular replacement with Phaser-MR [24] using the structure of tyrosine kinase AS-a hypothetical protein with a sequence corresponding to a common ancestor of Src and Abl with~43% identity to the RTKC8 kinase domain (PDB 4UEU [25]) as a search model. The structure was rebuilt using Phenix-AutoBuild and refined with multiple cycles of refinement with phenix.refine [26] and model building with Coot [27]. We used Molprobity [28] statistics provided in the Phenix suite to validate the structure. The calculated difference Fourier map (Fo-Fc) revealed clear electron density for two molecules of bound staurosporine per kinase domain and one phosphate, resumed to be from the crystallization buffer. The phosphate (PO4) and staurospaurine (STU) ligands were refined with full occupancy using the standard cif dictionaries included in the phenix suite. The B-factors for the ligands were comparable to the B-factors of the protein amino acid residues co-ordinated with them. The coordinates and structure factors of the final structure have been deposited in the Protein Data Bank (PDB) with the PDB ID 8E4T. Software used in this project was curated by SBGrid [29].

RTKC8 kinase domain is phosphorylated
Western blot analysis was performed to check the phosphorylation status of the purified RTKC8 kinase. The immunoblot was probed with an anti-phosphotyrosine antibody and showed that the purified RTKC8 kinase had at least one phosphorylated tyrosine residue which could be dephosphorylated by the tyrosine-protein phosphatase YopH (Fig 2A). The band intensity was decreased to~30% of the initial level after five minutes and to less thañ 5% after an hour.

Substrate specificity of the RTKC8 kinase
The phosphorylation rates of the peptides human c-Src preferred peptide (PKCδ-Y313) and non-preferred peptide (PDGFRα-Y762) [17] by c-Src (1 μM) and RTKC8 kinase domain (5 μM) were assayed (Fig 2B and 2C). RTKC8 kinase appears to have a different substrate specificity than c-Src, showing more rapid phosphorylation of PDGFRα-Y762 than PKCδ-Y313. Kinase assays were also performed with three RTKC8 tail peptides comprised of the sequences surrounding Tyr 1828, Tyr Y1897, or Tyr Y1990. These experiments showed that the different tail peptides were phosphorylated at different rates. The RTKC8 peptide containing Tyr 1990 had the highest rate of phosphorylation by RTKC8 kinase (Fig 2B and 2D). Both the c-Src non-preferred peptide (PDGFRα-Y762) and the tail peptide (RTKC8-Y1990) have Leu before the Tyr and Asp after the phosphosite tyrosine residue. This result suggests that the preferred substrates of the RTKC8 kinase contain a hydrophobic residue at the position immediately before the phosphosite and an acidic residue immediately after it.

Structure determination
Crystallization trials of the RTKC8 kinase domain with staurospaurine yielded triangular shaped crystals, 70-100 μm in diameter. The crystals were of the orthorhombic space group P2 1 2 1 2 with cell dimensions a = 81.1, b = 55.5, c = 60.6 (Å) with a solvent content of 46.91% and a Matthews coefficient of 2.32 and one molecule in the asymmetric unit. The crystal structure of the kinase domain was determined to 1.95 Å resolution and the model was built with a final Rwork/Rfree (%) of 20.4/25.5 ( Table 1). The crystallographic model includes residues 1532 to 1804, two staurosporine molecules, 100 water molecules and one phosphate group ( Fig  3A). The first three residues (1530-1531), the last seven residues (1805-1811) and a region of the activation loop (1694-1698) in the expressed protein construct were not modeled as there was no interpretable electron density for these residues. The tyrosine residues which would be expected to be phosphorylated in the activation loop (Tyr 1697 and Tyr 1698) were in the disordered activation loop region. The staurosporine inhibitor bound to the kinase domain at two sites, one in the ATP binding site (Fig 4A) and other on the C-lobe in the peptide-substrate binding site (Figs 3C and 4B).

The choanoflagellate RTKC8 kinase domain crystallized in an active conformation
The choanoflagellate kinase domain exhibits the conserved structural features of a typical kinase domain [30,31] with two lobes, the N-lobe and C-lobe. The activation loop is a stretch of 29 amino acids in the C-lobe, beginning with the conserved Asp-Phe-Gly (DFG) motif and ending at Ala-Pro-Glu (APE). The activation loop of the RTKC8 kinase domain contains two consecutive tyrosine residues for which there is no observable electron density.
The RTKC8 kinase crystallized in an active conformation with a RMSD of 1.0 Å over 201 Cα atoms to the structure of Lck in an active conformation [32]. The conserved Glu 1592 of the αC-helix is rotated towards the active site to form a salt bridge with Lys 1575 of the β3-strand, oriented as "αC-helix-in". The distance between αC-helix-Glu-Cβ and β3-Lys-Cβ atoms is 8.1 Å (C-helix-in and C-helix-out structures are defined with the distance of �10 Å or >10 Å, respectively) [33]. The kinase is crystallized in "DFGin" conformation where the DFG-Asp is pointed towards the active site and the DFG-Phe interacts with the αC-helix. The Asp backbone and its sidechain is oriented in such a manner that the carbonyl group of the Ala prior to the DFG motif forms a hydrogen bond with the nitrogen of the histidine sidechain in HRD motif [33]. The distance between the sidechain nitrogen of His 1664 and the carbonyl group of the Ala 1683 is 3.1 Å, consistent with formation of a hydrogen bond between them. The activation loop is extended away from the C-lobe of kinase domain. In this extended conformation, the cleft is accessible to ATP and peptide substrates [34,35]. The orientation of the activation loop leads to the formation of a hydrogen bond between the backbone N atom of the sixth residue in the loop (DFGxxX) and the backbone carbonyl of the residue prior to the HRD motif in the catalytic loop. The amide nitrogen of Arg 1689 (DFGxxR) and the carbonyl of Ile 1663 form a hydrogen bond of 3 Å. Another major structural feature of this conformation is the presence of type I β-turn at the Phe in the DFG motif, which forms a hydrogen bond between the carboxyl group of the Phe and the NH of the Ala two residues after the DFG motif [33]. The hydrogen bond length between Phe 1685 and Ala 1688 is 3.2 Å. This interaction plays an important role in orientating the activation loop to an extended conformation. All these interactions and structural features present the signature of the active conformation of the kinase domain (Fig 3C).

Staurosporine interactions with the RTKC8 kinase domain
Two molecules of staurosporine are bound to the RTKC8 kinase. One is in the ATP-binding site as seen in several other structures of kinase:staurospaurine structures (eg. CDK2 (1AQ1 [36]) and ZAP-70 (1U59 [37]) (Fig 3A), and the other is bound in the peptide substrate binding site in the C-lobe (Fig 3B and 3C). The active site inhibitor is positioned in the large groove   of staurosporine and the carbonyl oxygen on the backbone of Arg1670, in the ribose-binding pocket of the kinase domain ( Fig 4B). This is a typical binding mode for staurospaurine in the active site of a kinase. Another molecule of staurosporine was found bound to the C-lobe of the kinase domain in a hydrophobic groove between the activation loop, the αEF helix and the αG helix (Fig 4C). It also forms pi-pi stacking crystal contacts with its symmetry molecule, which implies that soaking of an apo or ATP/ADP bound crystal would not have revealed this secondary site, since the protein would likely have been different in crystals form without the inhibitor. This staurosporine is bound at the peptide-substrate binding site as revealed in the structure of the insulin receptor tyrosine kinase domain in complex with a peptide substrate (1IR3) [39] (Fig 3C). This site has been identified as a druggable site, [40] and is the site of a trans interaction in Aurora kinases, where the activation loop from one kinase binds to and presents as a substrate for phosphorylation by the other aurora kinase, so is designated the AAS site. This potential binding site has been identified in Src, Zap-70, and other tyrosine kinases using solvent mapping algorithms [40], and it is intriguing to see the inhibitor bound to a tyrosine kinas AAS site. The ligand interacts with hydrophobic residues (Gly 1701, Gly 1702, Val 1704, Val 1706) in the activation loop and Trp 1752, Asn 1748 and Gln 1749 in the αG-helix (Fig 4D).

Structural comparison of the RTKC8 kinase domain with the Ephrin kinase domain
A Dali search [41] of the RTKC8 kinase model (268 residues) from Choanoflagellate against the PDB database shows the highest structural similarity with the kinase domain of the Ephrin type A receptor 5 (2R2P, 295 residues) with the z-score of 36.2 and 38% sequence identity. The two structures were aligned over 214 atoms with a root mean square deviation of 0.84 Å ( Fig  5). The Ephrin receptor juxtamembrane region modulates the kinase activity [42]. Two tyrosine residues in the juxtamembrane segment, when phosphorylated, release inhibition, and provide docking sites for SH2 domain-containing proteins to transmit the downstream signal [43]. However, the juxtamembrane domain of RTKC8 kinase does not contain tyrosine residues. In addition, while the kinase domains are structurally similar, the extracellular receptor on the Ephrin receptor bears no resemblance to that of the RTKC8 receptor. The structure of RTKC8 kinase demonstrates that the kinase domain remained highly conserved throughout the evolution of receptor tyrosine kinases from choanoflagellates to human but with divergent fused extracellular domains.

Concluding remarks
Choanoflagellates provide an excellent opportunity to study the evolution of receptor tyrosine kinases as they are the only group shown to have receptor tyrosine kinases outside of metazoans [9]. It is an enigma why such an apparently primitive organism has such an extensive kinome. We report the expression, purification, and initial characterization of the RTKC8 kinase domain from a choanoflagellate receptor tyrosine kinase. The isolated kinase domain is monomeric, which is usual for an isolated kinase domain, since it is usually the dimerization of the extracelluar sensor domains which dimerizes the kinase domains to activate them. We also present the crystal structure of the kinase domain in complex with staurospaurine. The purified RTKC8 kinase is catalytically active and can phosphorylate tyrosine residue containing peptides generated from its C-terminal tail sequence. This suggests that the tail tyrosine residues act as substrates for the kinase. These phosphorylated tyrosine residues most likely act as recruitment sites for a yet unrealized signaling pathway in choanoflagellates. What the role of these pathways in such a primitive organism is an intriguing question, open for further investigation. That Choanoflagellates have an Ephrin like kinase with a completely different sensor domain implies a divergence, in which either metazoans or choanoflagellates repurposed the kinase domain for their different signaling requirements. The unexpected secondary binding site for staurospaurine in the substrate binding site or Aurora activation segment (AAS), which is likely a non-optimal ligand for the site, shows the potential for this site in developing tyrosine kinase inhibitors.