NMR Solution Structure of Human Vaccinia-related Kinase 1 (VRK1) Reveals the C-terminal Tail Essential for Its Structural Stability and Autocatalytic Activity*

Vaccinia-related kinase 1 (VRK1) is one of the mitotic kinases that play important roles in cell cycle, nuclear condensation, and transcription regulation. Kinase domain structures of two other VRK family members (VRK2 and VRK3) have been determined previously. However, the structure of VRK1, the most extensively studied and constitutively active VRK member, is yet to be characterized. Here, we present the nuclear magnetic resonance (NMR) solution structure of a catalytically active form of human VRK1 with its extended C-terminal tail (residues 1–361). The NMR structure of human VRK1 reveals that the C-terminal tail orients toward the catalytic site and forms a number of interactions that are critical for structural stability and catalysis. The role of this unique C-terminal tail was further investigated by deletion mutant studies where deletion of the terminal tail resulted in a dramatic reduction in the autocatalytic activity of VRK1. NMR titration studies carried out with ATP or an ATP analog confirm that ATP/ATP analogs interact with all of the crucial residues present in important motifs of the protein kinase such as the hinge region, catalytic loop, DYG motif, and thereby suggest that the catalytic domain of VRK1 is not atypical. In addition to the conventional interactions, some of the residues present on the extended C-terminal tail also interact with the ligands. These observations also substantiate the role of the extended C-terminal tail in the biological activity of VRK1.

The coordinated action of protein kinases and phosphatases plays a major role in regulating signal transduction events in a multicellular organism (1)(2)(3). Kinases comprise one of the major members of the human genome (4), characterized by the presence of a conserved catalytic domain of approximately 300 amino acids which is involved in the phosphotransfer reactions (5)(6)(7). Vaccinia-related kinase 1 (VRK1) 4 belongs to a novel group of serine/threonine kinases that bear a high degree of sequence similarity with vaccinia virus B1 R kinase (8,9). Three members of VRK family are known in the human genome and show a significant sequence similarity among themselves with respect to their catalytic domains, but differ in their regulatory domains (8,10). VRK1 is the most well studied member of the family, known to participate in a number of biological activities especially in cellular proliferations as well as in management of cellular stress situations (11). A number of studies demonstrated that VRK1 phosphorylates several stress-related transcription factors e.g. p53, c-Jun, ATF2, which in turn play a major role in regulating cellular fate (10 -12).
Several studies have also reported the role of VRK1 in cellular proliferation both in normal cells as well as in cells with uncontrolled proliferation (13). Elevated levels of VRK1 protein have been observed in highly proliferative cell lines, indicating its role in cell division (14). Furthermore, VRK1 is a nuclear protein, known to play an important role in cell cycle progression through phosphorylation of barrier to autointegration factor (15)(16)(17), which helps in formation of the nuclear envelope during cell division (18,19) and in the chromatin condensation event during cell division by phosphorylating histone H3 both at threonine 3 and serine 10 residues (20,21).
The catalytic domain structures of VRK2 and VRK3 reveal the typical kinase fold but with a degraded catalytic site on VRK3 (22). It is noted that the catalytic domains of VRKs are flanked by the large noncatalytic regions either at N or C termini. VRK1 contains a C-terminal noncatalytic tail but is constitutively active. Currently, the underlying molecular mechanism of VRK1 activation is unknown. In this report, we present NMR solution structures of a catalytically active form of human VRK1 with the C-terminal noncatalytic fragment flanking the kinase domain, and we propose a molecular basis underlying the activation mechanism of VRK1.

EXPERIMENTAL PROCEDURES
Cloning, Expression, and Purification of the Wild Type and Deletion Mutants of Human VRK1 Protein-The complementary DNAs (cDNAs) encoding the full-length human VRK1 and the C-terminal truncation mutants (VRK1 1-391 , VRK1 1-381 , VRK1 1-371 , VRK1 1-361 , VRK1 1-351 , VRK1 1-341 , VRK1 1-331 ) were PCR-amplified using the corresponding forward and reverse primers containing NdeI and XhoI restriction enzyme cleavage sites. The resultant amplified products were gel-purified, digested using the restriction enzymes, and cloned into pET-29b plasmid (Novagen). For the protein purification, cells were grown in M9 medium containing 15 NH 4 Cl, 15 NH 4 Cl plus [U-13 C]glucose, and 75% or 100% 2 H 2 O at 37°C at 220 rpm until the culture optical density reached an absorbance of 0.7 at 600 nm. The culture was then induced with 1 mM isopropylthio-␤-D-galactopyranoside at 25°C for 4 h. After the cultured cells were harvested by centrifugation, the cell pellet was resuspended in a lysis buffer (20 mM phosphate buffer, pH 7.8, 500 mM NaCl, 1 mM phenylmethylsulfonate, lysed by sonication, and purified on a Ni 2ϩ -nitrilotriacetic acid-agarose column (Invitrogen). The eluted proteins from the Ni 2ϩ -nitrilotriacetic acid-agarose column were analyzed by 12.5% SDS-PAGE. The fractions containing VRK1 were collected, concentrated, and further purified on a Superdex S-200 column (GE Healthcare) preequilibrated with a buffer containing 20 mM Tris-HCl, pH 6.8, 2 mM dithiothreitol (DTT), and 0.01% NaN 3 . The purified proteins were verified by mass spectroscopy at Nanyang Technological University Proteomics Core Facility.
Nuclear Magnetic Resonance (NMR) Data Acquisition and Assignments-Samples for NMR experiments contained 0.3-0.6 mM protein in 20 mM Tris-HCl, pH 6.8, 2 mM DTT, 0.01% NaN 3 , and 10% or 100% D 2 O. For backbone assignments, transverse relaxation-optimized spectroscopy (TROSY)-based HNCACB, HN(CO)CACB, HNCA, HN(CO)CA, HNCO, and HN(CA)CO spectra were obtained using 100% fully deuterated and uniformly 13 C/ 15 N-labeled VRK1 (23). For side chain resonance assignments, TROSY-based 15 N-edited NOESY, HCCH TOCSY, and 13 C-edited NOESY experiments using fully or 75% perdeuterated protein samples were performed. Selectively labeled NMR samples on 13 C/ 15 N-labeled valine/isoleucine/ leucine residues were prepared, and HCCH TOCSY and 13 Cedited NOESY experiments were performed for assignments of side chain methyl resonances and deriving methyl-methyl or methyl side chain distance constraints. NMR experiments were performed on Bruker Avance 700, 800, and 900-MHz spectrometers equipped with a 5-mm triple resonance, z axis-gradient cryogenic probe at 298 K. A list of the NMR experiments is shown in supplemental Table 1. All spectra were processed with NMRPipe (24) and analyzed using SPARKY (25). The assigned chemical shift values of VRK1 have been deposited in the Biological Magnetic Resonance Bank (accession code 16738). Residual dipolar couplings (RDCs) were measured using polyethylene glycol/alcohol mixtures (26). 50 l of C12E5 was mixed in 530 l of NMR buffer containing 90% H 2 O and 10% D 2 O for stock solution preparation. 1-Hexanol was gradu-ally added by 2-l increments with vigorous shaking to a final molar C12E5:1-hexanol ratio of 0.96. After 1 h at room temperature, air bubbles were removed by centrifugation at 5000 ϫ g for several minutes. For the measurement of RDCs, 300 l of the C12E5:1-hexanol stock solution was added to 200 l of protein. The final concentration of C12E5:1-hexanol mixture in the NMR sample was about 5% (w/w). 15 N-1 H RDC constants were measured using combination of 1 H-15 N HSQC and 1 H-15 N TROSY-HSQC experiments at 25°C as described previously (27). The data analysis and calculation of alignment tensor were performed using REDCAT software (28) (see supplemental Fig. S6).
NMR Titration of VRK1 with ATP/ATP Analog-Molecular characteristics of ATP/ATP analog binding with VRK1 were studied using two-dimensional TROSY-HSQC spectra of 15 N-labeled VRK1 recorded on a Bruker Avance 700 spectrometer at 298 K, and the ligand binding sites were mapped by analyzing chemical shift perturbations. Initially, the NMR spectra of free sample was recorded using 0.1 mM 15 N-labeled VRK1 1-316 . The ligands were then added to the sample, and chemical shifts were compared before and after the addition of the ligand.
Structure Calculation and Refinement-Solution structures of VRK1 were calculated by simulated annealing in torsion angle space with combination of the program CYANA 2.1 (29) and CNS 1.2 (30,31). Distance constraints were derived by analyzing TROSY-based 1 H-15 N NOESY (80-ms mixing time) and 13 C-edited NOESY (100-ms mixing time) spectra of 50% random fractional deuterated or selective labeled (valine, isoleucine, and leucine residues) samples of VRK1. The secondary structure was predicted by program TALOSϩ (32) based on the analysis about sequential (Pi-jP ϭ 1), short range (Pi-jPϽ5) NH-NH and NH-aliphatic contacts on 15 N TROSY-NOESY spectrum. Dihedral angle (, ) restraints were also calculated from chemical shifts using TALOSϩ, and hydrogen bond restraints were obtained based on protein structure during structure calculations. Manually assigned NOE constraints were classified based on their intensities and applied with upper distance limit of 3.0 Å (strong), 3.5 Å (medium), 5.0 Å (weak), 6.0 Å (very weak), and total 200 conformers were generated as initial structures by CYANA 2.1 from 5193 NOE constraints. An additional 0.5 Å was added for NOEs involving methyl groups. After calculation, initial conformers were sorted by target function values, and the lowest 100 conformers were selected for further refinement using CNS 1.2. 292 backbone hydrogen bonds were identified on the basis of initial structures, and 457 torsion angle constraints derived from chemical shift values with the program TALOSϩ were applied in refinement stage. Also, 234 15 N-1 H residual dipolar coupling restraints were included in the final stage of calculation. The final structure was refined by using a simulated annealing protocol with combination of torsion angle space and Cartesian coordinate dynamics as published previously (33). Finally, 20 structures were selected by their total energy values for display and structural analysis. MOLMOL (34) and PyMOL (35) programs were used for structure visualization. The 20 NMR ensemble structures have been deposited in Protein Data Bank (PDB) with code 2LAV.
In Vitro Kinase Assay-The wild-type human VRK1 and its C-terminal truncation mutants were evaluated for their autophosphorylation activity according to the protocol described previously (36). Briefly, 1 g of the recombinant proteins were placed in a reaction mixture containing 20 mM Tris-HCl, pH 7.5, 150 mM KCl, 5 mM MgCl 2 , 0.5 mM DTT, 5 M ATP, and 5 Ci of [␥-32 P]ATP. The reaction mixture was then incubated at 37°C for 30 min. The protein mixtures were resolved by 12% SDS-PAGE. The gels were then exposed to x-ray film overnight and developed using X-OMAT2000 (Kodak).

RESULTS AND DISCUSSION
NMR Solution Structures of VRK1-The structural studies were performed on a catalytically active form of human VRK1 (residues 1-361) lacking 35 residues at C terminus. The structures were determined based on a total of 5193 nontrivial NMR-derived distance restraints and 234 15 N-1 H RDC restraints, which were included in the final step of structure calculation. The ensemble of 20 low energy structures calculated with CNS is shown in Fig. 1A. Excluding the N-terminal 20 residues and C-terminal 6 residues, the average root mean square deviation (r.m.s.d.) values relative to the mean coordinate of 20 conformers were 1.24 Å for the backbone atoms and 1.95 Å for all heavy atoms. Structural statistics for the NMR ensemble are given in Table 1. There are no distance violations greater than 0.5 Å or dihedral angle violations greater than 5°. Analysis of the average minimized structures with PROCHECK-NMR (37) showed that 76.2% and 20.0% of residues lie in the most favored and allowed regions, respectively.   The NMR structures of VRK1 reveal the well defined typical kinase fold. The overall fold of the kinase domain of VRK1 is similar to two other VRK kinase members (VRK2 and VRK3) adopting a bilobate structure (Fig. 1). The N-terminal tail (residues [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20] is unstructured in solution as evidenced by the lack of medium and long range nuclear Overhauser effects (NOEs) in this region. The characteristic additional helix (␣C4) is inserted between ␣C and ␤4 as predicted. The last helix in the C-terminal lobe (␣I) of VRK1 (316 -331) is more extended compared with VRK2 and stabilizes the entire protein conformation by forming hydrophobic spines with side chains of aliphatic and aromatic residues present in the protein core (Fig.  1B). Most notable and unique feature of the VRK1 solution structure is that the C-terminal tail (residues 341-361) orients around the catalytic site and forms a number of interactions with several important motifs such as G-loop, ␣C, the catalytic and activation loops ( Fig. 1C and supplemental Fig. S1).
The C-terminal Tail Interacts with the Catalytic Core and Shows That It Is Essential for VRK1 Structure and Catalysis-The C-terminal tail of VRK1 shows a high degree of sequence conservation across species (8), suggesting its potential regulatory role. The VRK1 NMR structures reveal that Ile 354 interacts with Pro 79 of ␣C, and the residues of ␣D helix (Leu 138 and Ile 141 ) also form hydrophobic interactions with Leu 341 and Val 343 of the extra C-terminal tail (Fig. 2A). The residue Leu 349 at the C-terminal tail forms a hydrophobic interaction with Phe 48 of the G-loop. We have investigated the role of the C-terminal tail on VRK1 stability and catalysis by performing chemical shift perturbation experiments on the uniformly 15 N-labeled VRK1 1-361 , which was used in our NMR structure determination and VRK1 1-341 construct where residues 342-361 were truncated. The two-dimensional 1 H-15 N HSQC spectra of VRK1 1-361 and VRK1 1-341 showed noticeable chemical shift perturbations on some of the residues in the activation (Ala 201 , Tyr 202 ) and catalytic loops (Glu 172 , Val 174 , Lys 179 , Ser 181 ) after the removal of the C-terminal 20 residue extension ( Fig. 2E and supplemental Fig. S2). In addition, the residues from ␣D (Leu 138 -Lys 140 , Glu 143 ) and the loop connecting ␣D and ␣E (Ala 146 , Lys 147 , Phe 149 , and Ser 150 ) and residues in ␤9 (Asp 191 -Val 196 ) showed chemical shift perturbations, suggesting a potential regulatory role by the C-terminal tail. Long range NOEs were observed between methyl protons and residues on the ␣D helix and C-terminal extension residues such as Leu 341 , Val 343 , and Val 344 (Fig. 2C). Our NMR structure reveals that Ala 201 and Tyr 202 are located near Thr 353 , Ile 354 , Thr 355 , and Lys 356 . This is also directly supported by the presence of long range NOEs between the ring protons of Tyr 202 and Thr 353 /Thr 355 (Fig. 2D). Fig. 2B shows the mapping of the interacting residues. These results suggest the molecular basis of VRK1 activation which requires the molecular interaction between the C-terminal tail and catalytic core of the enzyme. Thus, to delineate further the potential activation mechanism of VRK1, we constructed a series of truncation mutants of VRK1 and examined their characteristics (Fig. 3). Truncation of 75 residues from the C terminus (VRK1 1-321 ) causes the mutant VRK1 to become insoluble when expressed in bacterial cells (data not shown). Interestingly, VRK1 1-331 is soluble but inactive. VRK1 1-351 and VRK1 1-341 , which lack the C-terminal 45 and 55 residues, respectively, are soluble but show marginal kinase activities. We also evaluated the VRK1 catalytic activity by designing multiple point mutations in the C-terminal tail region. Our data showed that S342A and T353A mutants reduce the autocatalytic activities of VRK1, indicating that those residues play key roles in the kinase activity (supplemental Fig. S3). Taken together, these data suggest that the C-terminal region (residues 321-361) is important for VRK1 physical property as well as its catalysis (Fig. 3) and also support the hypothesis that the C-terminal tail of VRK1 interacts with the catalytic core of the enzyme and plays essential roles for VRK1 activation.  showing the key ion pairing between lysine (␤3) and glutamate (␣C) residues. VRK1 structure is shown in orange and VRK2 in blue. Key residues are highlighted with sticks, and hydrogen bonds are highlighted with lines. B, comparison of VRK1 NMR structure ion pairing pattern with VRK1 crystal structure. VRK1 NMR structure is represented in orange and VRK1 crystal structure is (PDB 3OP5) highlighted in cyan. C, structural alignment and comparison of the activation loops of VRK1 and VRK2 with conserved threonine residues highlighted as sticks. D, superimposition of the activation loops of VRK1 NMR, VRK1 crystal, and VRK2 structures. Residues that are mutated in the activation loop of VRK1 crystal structure (PDB 3OP5) are highlighted with sticks. The conformational differences in the activation loop of VRK1crystal structure with respect to VRK2 are shown with an arrow.

Structural Comparison of VRK1, VRK Family Members, and
Other Kinases-The r.m.s.d. value for 276 C ␣ carbons is 2.56 Å between VRK1 and VRK2 (Fig. 1C). Despite the similarity in their structures, they exhibit differences in their local structures. VRK1 is an active member of the VRK family, which encompasses all of the required features/motifs for kinase activity like VRK2. Previous crystallographic studies revealed that VRK3 is having a pseudo-kinase domain with no catalytic ability, unlike VRK1 and VRK2 (22). In the NMR structure of VRK1, ␣C is rotated inward, a key ion pair between Lys 71 and Glu 83 is formed, and the DYG motif (DFG in several other kinases) orients like VRK2 and other active kinases (Fig. 4A). The activation loop of VRK1 superimposed well on VRK2, and conserved threonine residues (Thr 224 and Thr 228 ) also oriented similarly to VRK2 (Fig. 4C). The ␣C4 helix, which is unique in all VRK family members, is also observed in VRK1 and contains the conserved residues Ile 96 , Trp 99 forming a hydrophobic interaction with Tyr 167 in ␣E, which is similar to what is seen in VRK2 and VRK3 (supplemental Fig. S4). Recently, the VRK1 crystal structure was deposited in the PDB with ID code 3OP5. The crystal structure consists of several mutations in key charged residues and lacking the electron densities for the FIGURE 5. Interaction between ATP analog AMP-PCP and VRK1. A, overlay of 1 H-15 N TROSY-HSQC spectra of uniformly 15 N-labeled VRK1 (red) and VRK1 in the presence of AMP-PCP (blue) at molar ratio of 1 (VRK1) to 5 (AMP-PCP). Expanded views of chemical shift perturbations upon ligand binding are shown on the right side of the panel. The perturbations of chemical shifts from free VRK1 (red) to AMP-PCP complex (blue) are displayed by arrows, and perturbed residues are labeled with a single letter. B, mapping of ATP binding site and interaction residues on VRK1 based on 1 H-15 N TROSY-HSQC titration data with AMP-PCP. Perturbed amino acids were displayed according to degree of chemical shift perturbations: red (⌬␦ Ͼ 0.5), blue (0.3 Ͻ⌬␦ Ͻ 0.5), and green (disappeared resonances). The orientation of VRK1 on the left side of the panel is the same as that in the Fig. 1, whereas that on the right side of the panel is rotated 180°along the y axis. C, plot of chemical shift perturbations of human VRK1 upon ATP analog AMP-PCP binding. The regions affected by binding of AMP-PCP are shown by a gray box, and secondary structure or amino acid residues are highlighted.