Protein phosphorylation, mediated by protein kinases, is a key event in the regulation of eukaryotic signal transduction. The majority of eukaryotic protein kinases perform phosphoryl transfer, assisted by two divalent metal ions. About 10% of all human protein kinases are, however, thought to be catalytically inactive. These kinases lack conserved residues of the kinase core and are classified as pseudokinases. Yet, it has been demonstrated that pseudokinases are critically involved in biological functions. Here, we show how pseudokinases have developed strategies by modifying amino acid residues in order to achieve stable, active-like conformations. This includes binding of the co-substrate ATP in a two metal-, one metal- or even no metal-binding mode. Examples of the respective pseudokinases are provided on a structural basis and compared with a canonical protein kinase, Protein Kinase A. Moreover, the functional roles of both independent metal-binding sites, Me1 and Me2, are discussed. Lack of phosphotransferase activity does not implicate a loss of function and can easily point to alternative roles of pseudokinases, i.e. acting as switches or scaffolds, and having evolved as components crucial for cellular cross-talk and signaling. Interestingly, pseudokinases are present in all kingdoms of life and their specific roles remain enigmatic. More studies are needed to unravel the crucial functions of those interesting proteins.

Background

Eukaryotic protein kinases (ePKs) are known to be essential mediators in signal transduction. Protein kinase action is correlated with their enzymatic function in catalyzing the transfer of a phosphoryl group of ATP to a substrate protein. Reversible protein phosphorylation is thereby a universal strategy known to be involved in almost every cellular function. Notably, ∼30% of all human proteins are covalently modified with phosphoryl moieties [1]. Deciphering the human genome and the subsequent categorization of human protein kinases (>500), however, uncovered various (∼50) human kinases, lacking residues critical for catalytic function. Those protein kinases are termed pseudokinases [2].

Although sharing only marginal overall sequence identity with ePKs, even eukaryotic-like kinases of the microbial kinome show high conservation of the catalytic core [3]. Specific residues involved in catalysis and the kinase fold itself seem to have evolved early in evolution.

All protein kinases share a highly conserved kinase core that spans ∼250 amino acids and folds up into two defined lobes [46]. Several motifs, known to be involved in the catalytic function, are distributed around the active site which is located in a cleft between both lobes [7,8]. In particular, the so-called (H/Y)RD and the DFG motifs are highly conserved where the respective aspartates in both motifs are thought to be invariant for catalysis [7]. Another important residue involved in ATP binding is located in β-strand 3 of the smaller N-lobe. This lysine (β3-Lys) not only stabilizes the α- and β-phosphate moieties of ATP, but also recruits another critical residue of the αC-Helix, the αC-Glu, thereby organizing an active conformation of the kinase (Figure 1). However, some protein kinases, known to be active, lack this Lys at β-strand 3. The kinase WNK (with no lysine) utilizes a Lys in β-strand 2, illustrating that other residues are able to compensate [9]. In this line, several pseudokinases (i.e. Tribbles 2 pseudokinase, Trb2), predicted to be catalytically inactive, were found to undergo at least autophosphorylation [10].

The protein kinase domain.

Figure 1.
The protein kinase domain.

All ePKs are defined by a bilobal structure. The C-terminal lobe (C-lobe) is built up by helices, whereas the smaller N-lobe mainly consists of β-strands. The activation loop phosphorylation (pThr197 in PKA) allows the interaction of the αC-Glu (E91 in PKA) with the β3-Lys (K72 in PKA) and thereby stabilizes nucleotide binding. The correct positioning of the polyphosphate of ATP for phosphoryl transfer is accomplished by two divalent metal ions (shown in black). Me1 is positioned by the DFG-Asp (D184 in PKA) of the Mg-binding loop (red) and Me2 is bound by an Asn (N171 in PKA) of the catalytic loop (blue). This loop also contains the catalytic base (D166 in PKA). Modified from ref. [8]. The structure (PKA, PDB 1ATP) was visualized using PyMOL v1.3 (Schrödinger LLC).

Figure 1.
The protein kinase domain.

All ePKs are defined by a bilobal structure. The C-terminal lobe (C-lobe) is built up by helices, whereas the smaller N-lobe mainly consists of β-strands. The activation loop phosphorylation (pThr197 in PKA) allows the interaction of the αC-Glu (E91 in PKA) with the β3-Lys (K72 in PKA) and thereby stabilizes nucleotide binding. The correct positioning of the polyphosphate of ATP for phosphoryl transfer is accomplished by two divalent metal ions (shown in black). Me1 is positioned by the DFG-Asp (D184 in PKA) of the Mg-binding loop (red) and Me2 is bound by an Asn (N171 in PKA) of the catalytic loop (blue). This loop also contains the catalytic base (D166 in PKA). Modified from ref. [8]. The structure (PKA, PDB 1ATP) was visualized using PyMOL v1.3 (Schrödinger LLC).

Besides these invariant residues in the kinase core, two nonlinear structural motifs have been identified. These two hydrophobic ‘spines’ span both lobes and are formed by noncontiguous residues from different parts of the kinase domain [11,12]. Assembled spines are considered a hallmark of an active protein kinase. Kinases toggle between an inactive and active state and, by assembly of both lobes, the equilibrium is drastically shifted toward the active state, with the kinase ready for catalysis [13]. With the integration of the adenine moiety of ATP, the catalytic spine (C-spine) is dependent on nucleotide binding, whereas the regulatory spine (R-spine) is mostly affected by the configuration of the activation loop (also known as T-loop). Many kinases need to be phosphorylated on this loop to gain activity by assembling the R-spine and thereby positioning the DFG motif for metal binding and catalysis [14].

Metal ion dependency of protein kinases

Canonical protein kinases are metalloenzymes utilizing divalent metal ions to effectively bind ATP and to assist phosphoryl transfer as the chemical step. The important role of divalent metal ions for protein kinase function was revealed at the same time that protein kinases were identified. In 1954, Burnett and Kennedy [15] showed that casein phosphorylation was enzymatically catalyzed only when Mg2+ was present in the reaction. In the following years, groundbreaking work by Fischer and Krebs revealed that the conversion of glycogen phosphorylase from the b-form to the a-form can be restored in a cell-free system without hormone response by adding Me2+ and ATP [16,17]. Remarkably, even before phosphorylase kinase was identified as the first genuine protein kinase, the dependency of divalent metal ions was established [17,18]. Later it was shown that, two metal-binding sites exist in the protein kinase domain, one coordinated by the already mentioned Asp of the DFG motif (metal-binding site 1, Me1) and a second by an Asn following the H/YRD motif (HRDXXXXN, catalytic loop Asn) of the assumed catalytic base (metal-binding site 2, Me2) (Figure 1) [5,19]. Both metal-binding sites have unequal affinities, and studies based on electron paramagnetic resonance investigating the catalytic subunit of cAMP-dependent protein kinase A (PKA) and kinetic studies on cyclin-dependent kinase 2 suggest that Me2 is the high-affinity binding site and Me1 is a weakly bound metal ion [1921]. Initial studies termed the weakly bound Me1 an ‘inhibitory’ metal; however, early studies revealed that, although it hampers product release, Me1 is necessary to accelerate the chemical step, phosphoryl transfer [19,21,22]. While it has not been proved yet, the conserved architecture of the kinase domain with two metal-binding sites hints in general for a two metal-binding mode.

This review focuses on the roles of metal ions for protein kinase function. Individual roles of the metal ions with respect to substrate and nucleotide binding and catalysis are considered.

Protein kinases are metalloenzymes

In principle, all metalloproteins are sophisticated metalorganic compounds consisting of a central (metal) atom which is surrounded by several other atoms or molecules. In proteins, polar amino acid side-chains (Asn, Gln, Cys, His, Asp and Glu) are found to act as ligands to coordinate a central metal ion [23]. While the central atom acts as a Lewis acid, the ligands offer binding electrons as Lewis bases. The number of atoms of a ligand that participate in metal binding is defined as denticity. Polydentate ligands are so-called chelators (Greek χελος chelos, claw), and carboxylate-containing amino acid side-chains like Asp or Glu are able to act as bidentate ligands [24]. Since most metalloenzymes are utilizing two or more residues to bind a metal, they are themselves chelators.

In the protein kinase core, both metals compensate the negative charges of the polyphosphate chain with the catalytic loop Asn binding Me2 as a monodentate ligand, whereas the DFG-Asp is a bidentate ligand and participates in Me1 and Me2 binding (Figure 1). This asymmetry also reflects the different functions of each metal. While Me1 seems to be critical for phosphoryl transfer, Me2 stabilizes nucleotide binding and is thought to bind first (in complex with ATP) [20]. Thermostability studies on the PKA catalytic subunit showed that ATP alone does not promote protein stabilization [25]. Low concentrations of metal ions, in order that only one metal site is occupied, improved thermostability (ΔTM = 2.5 K); yet only high metal concentrations, occupying both Me2 and Me1, resulted in a pronounced increase in ΔTM = 6 K [25]. Thus, both metal ions are required for stable nucleotide binding. AMP in the presence of high metal concentrations did not stabilize the kinase, probably due to being inefficient in metal chelation. Notably, adenine (lacking the ribose and the polyphosphate chain) and even more adenosine (only lacking the polyphosphate chain) were able to improve enzyme stability in the absence of metal ions (ΔTM = 2.5 and 5 K, respectively) [25]. As mentioned before, one feature of nucleotide binding is the completion of the catalytic spine by the adenine moiety.

Correct coordination of the metals not only promotes nucleotide binding, but also allows for precise positioning of the reactants for catalysis, congruent with the proposed in-line transfer of the phosphoryl group [26]. Different metal ions, although having the same charge (+2), prefer distinct coordination numbers, defined as the number of ligands bound to the metal [24]. This, at least in part, explains why only a few divalent metal ions are productive in steady-state catalysis [27]. Replacing the intracellularly most abundant metal Mg2+ with Ca2+, where Ca2+ prefers higher coordination numbers, produces an unproductive conformation of ATP [28]. In turn, this leads to an unstable enzyme substrate complex and is one reason for low substrate turnover in the presence of Ca2+ [29]. This clearly signifies the strong influence of the type of metal and substrate binding. Extensive studies on PKA with its inhibitor protein, the heat stable protein kinase inhibitor PKI, a pseudosubstrate, also demonstrate that substrate/pseudosubstrate binding is weak in the absence of either nucleotide or metal ion [25,29,30]. Thus both, metals and nucleotides prime the kinase for substrate binding.

For PKA, the overall influence of the metal ions on protein kinase dynamics was recently shown in silico by applying multiple microsecond-scale molecular-dynamics simulations [31]. The resulting community maps reflect regions of correlated motions and can be used to visualize alterations in protein dynamics upon binding of ligands, such as cofactors. Removing just one metal (Me1) was enough to induce specific communities while others vanished. Interestingly, most of the changes affected distant regions in the kinase domain [31]. This shows the allosteric effects of the metal ions on kinase dynamics and again demonstrates the strong interplay between metal and substrate binding.

Metal and nucleotide binding in pseudokinases

Any protein kinase needs at least two features that facilitate phosphoryl transfer. This is (i) the stable binding of the substrate as well as the co-substrate (usually ATP) and (ii) the ability to arrange both substrates allowing for an in-line transfer of the phosphoryl group.

While an enzymatic productive protein kinase needs to bind ATP as a phosphoryl donor and thereby needs metal ions, a catalytically inactive kinase might bypass metal binding and thus be expected to have altered nucleotide-binding properties. In this line, the nucleotide-binding ability of several pseudokinases was recently tested in a comprehensive screen based on thermostability [32]. Nucleotide binding is an indispensable part of kinase activity and along with the notion that protein kinases are stabilized by nucleotides, this assay allows to identify catalytically inactive pseudokinase domains (PKDs). Murphy and coworkers showed that more than half of the PKDs tested exhibited no detectable nucleotide or metal binding, although some of those contain the invariant β3-Lys and both residues involved in metal coordination. Besides stabilization by different nucleotides, the same authors tested the effect of divalent metal ions and could identify some PKDs (SgK269 and ROP2) that bind divalent cations in the absence of nucleotides. Congruently, the addition of nucleotides did not significantly alter thermal stability in these cases. Both SgK269 and ROP2 have degenerated DFG motifs; however, the catalytic loop Asn residues are still present.

The following examples reflect themes how specific protein kinase features (e.g. nucleotide binding) are preserved in pseudokinase function even in the absence of canonical metal-binding sites (Figures 24). Table 1 provides an overview covering resolution of the X-ray structure, ID of the metal binding site(s), metal binding mode and information about activity.

Protein kinases and pseudokinases utilizing a two metal-binding mode.

Figure 2.
Protein kinases and pseudokinases utilizing a two metal-binding mode.

(A) The PKA is shown as an example for a canonical protein kinase exhibiting a two metal-binding mode, where the catalytic loop Asn (N171) binds Me2 in a monodentate and the DFG-Asp (D184) Me1 in a bidentate manner. (B) The crystal structure of RNaseL bound to ATP reveals a metal-binding mode highly similar to PKA. However, a second Asp replacing the Gly in the DFG motif (DxD motif) also participates in Me1 binding, probably interfering with the phosphoryl transfer ability. (C) A similar Me2+-binding mode can be found in the (pseudo)active site of the glycan kinase POMK. (D) In the pseudokinase ROP5B of the parasite T. gondii, both metal ions are coordinated by Asp residues in a monodentate manner. Additionally, an Arg binds the γ-phosphoryl group of ATP, which probably interferes with a nucleophilic attack of a substrate. Structures (PKA, PDB 1ATP; RNaseL, PDB 4OAV; POMK, PDB 5GZA; ROP5B, PDB 3Q60) were aligned and visualized using PyMOL v1.3 (Schrödinger LLC). Mn2+ and Mg2+ ions are shown in violet and green, respectively.

Figure 2.
Protein kinases and pseudokinases utilizing a two metal-binding mode.

(A) The PKA is shown as an example for a canonical protein kinase exhibiting a two metal-binding mode, where the catalytic loop Asn (N171) binds Me2 in a monodentate and the DFG-Asp (D184) Me1 in a bidentate manner. (B) The crystal structure of RNaseL bound to ATP reveals a metal-binding mode highly similar to PKA. However, a second Asp replacing the Gly in the DFG motif (DxD motif) also participates in Me1 binding, probably interfering with the phosphoryl transfer ability. (C) A similar Me2+-binding mode can be found in the (pseudo)active site of the glycan kinase POMK. (D) In the pseudokinase ROP5B of the parasite T. gondii, both metal ions are coordinated by Asp residues in a monodentate manner. Additionally, an Arg binds the γ-phosphoryl group of ATP, which probably interferes with a nucleophilic attack of a substrate. Structures (PKA, PDB 1ATP; RNaseL, PDB 4OAV; POMK, PDB 5GZA; ROP5B, PDB 3Q60) were aligned and visualized using PyMOL v1.3 (Schrödinger LLC). Mn2+ and Mg2+ ions are shown in violet and green, respectively.

Pseudokinases with a two metal-binding mode

Ribonuclease L

Several recent crystal structures of PKDs illustrate potential noncanonical metal-binding modes. In the case of ribonuclease L (RNaseL), structures of porcine and human RNaseL have been solved in the presence of nucleotide and metal ions (Figure 2) [33,34]. The enzyme is involved in the degradation of viral RNA upon stimulation of the interferon pathway and consists of an Ankyrin domain, a catalytic active RNase domain and a PKD [33]. This PKD, although lacking phosphotransferase activity, displays a canonical two metal-binding mode, except that Me1 is not only coordinated by the DFG-Asp but also by another Asp which replaces the Gly of the DFG motif (Figure 2) [33]. The PKD itself and in particular both canonical metal-binding sites are important for RNase activity since mutants lacking either one of these sites were found to have no RNase activity [34]. While the DFG-Asp binds the Mg2+ ion in a bidentate manner, the additional Asp acts as a monodentate ligand. Interestingly, this DFD motif has been found previously in the mitogen-activated protein kinase-interacting kinases 1 and 2 (Mnk1 and Mnk2). These kinases are, however, functionally active, and reconstituting the canonical DFG motif by mutagenesis did not alter Mnk kinase activity or nucleotide binding [35].

Protein-O-mannose kinase (SgK196)

The kinase SgK196 (POMK, protein-O-mannose kinase) also harbors a DxD instead of the canonical DFG motif. Although SgK196 was initially classified as a pseudokinase and the structural analysis shows the characteristic bilobal structure of EPKs, this kinase was found to be a glycan kinase (POMK) and is catalytically active [36]. The recently solved crystal structure with Mg2+ADP and AlF3 resembles a transition state (Figure 2) [37]. The observed two metal-binding mode is similar to that of RNaseL, the canonical Asp of the DLD motif binds Me1 in a bidentate manner and the second Asp coordinates the same metal ion in a monodentate manner. Unlike RNaseL, the other metal-binding site, the HRDXXXXN Asn, is replaced by a Gln.

Rhoptry kinase ROP5B

Another pseudokinase which retains a two metal-binding mode is Rhoptry organelle protein 5B (ROP5B) of the intracellular parasite Toxoplasma gondii. ROP5B belongs to the rhoptry kinase (ROPK) family which includes several pseudokinases critical for the chronic infection of the host cell [38,39]. For ROP5B, although exhibiting both canonical metal-binding sites, no enzymatic activity has yet been reported [39]. The PKD lacks the catalytic base in the HRD motif (HGH); however, reintroducing an Asp at this position did not rescue phosphotransferase activity (Figure 2) [40]. Here, an unproductive conformation of ATP is responsible for the lack of activity. Crystal structures of ROP5B revealed a noncanonical two metal-binding mode, where the HRDXXXXN Asn no longer facilitates Me2 binding and, instead, an adjacent Asp fulfills this role (still in a monodentate manner). Owing to this, both metal ions are displaced compared with canonical active sites of protein kinases (i.e. PKA, Figure 2, upper panel left). Now, the corresponding DFG-Asp (DVS motif in ROP5B) solely binds Me1 and acts as monodentate instead of a bidentate ligand. While this still retains an active-like configuration, Me1 is not able to assist in phosphoryl transfer, due to the lost interaction with the γ-phosphoryl group. Notably, the bona fide protein kinase PKA and several other PKs [e.g. protein kinase C, protein kinase B (AKT1) or myosin light chain kinase] also have an acidic residue (Glu or Asp) adjacent to the catalytic loop Asn. Yet, this residue is positioned towards the substrate and is thereby important for substrate recognition (P-2 Arg/Lys of the substrate) [4,41]. Another difference of the (pseudo) active site is found in the Gly-rich motif of ROP5B. This contains an Arg at the tip of the loop, which interacts with the γ-phosphoryl group of ATP probably interfering with a nucleophilic attack of a substrate OH-group (Figure 2, lower panel right). In conclusion, the overall distorted active-site configuration of ROP5B seems to be composed for enzyme stability rather than catalysis.

Pseudokinases with a one metal-binding mode

Integrin-linked kinase

The pseudokinase integrin-linked kinase (ILK) lacks the catalytic base (H/YRD Asp) and no catalytic function has yet been reported; however, it is still able to bind ATP. The degenerated Me2 site (Asn → Ser) no longer coordinates a metal ion and thus, the arrangement of ATP differs from enzymatically active protein kinases. The unchanged binding site of Me1 still binds a metal ion and thereby helps to bind ATP in the pseudo-active site (Figure 3) [42]. Notably, nucleotide binding is a prerequisite for binding of ILK to its regulator α-parvin, clearly demonstrating that a pseudokinase maintains distinct features of a protein kinase also involving metal ions in order to fulfill their cellular function [42].

Pseudokinases utilizing a one metal-binding mode.
Figure 3.
Pseudokinases utilizing a one metal-binding mode.

(A) Crystal structures of the PKD of JAK2 exhibit a one metal-binding mode. The catalytic loop Asn still binds an Me2+ ion (most probably Me2), whereas the DFG-Asp binds the β3-Lys. (B) The DFG-Asp of the KSR2 interacts with an Arginine which replaces the β3-Lys and thereby the kinase binds just one metal ion, which is also coordinated by the catalytic loop Asn. (C) A highly unusual metal-binding mode is exhibited by the Poly(A) nuclease 3 (PAN3), which binds a metal ion utilizing amino acids of the extended Gly-rich loop. (D) ILK with its degenerated Me2-binding site solely binds Me1 with the DFG-Asp. In contrast with canonical protein kinases, the DFG-Asp functions as a monodentate instead of a bidentate ligand. Structures (JAK2JH2, PDB 4FVQ; KSR2, PDB 2Y4I; PAN3, PDB 4CYI; ILK, PDB 3KMW; Mg2+ ions are shown in green) were aligned and visualized using PyMOL v1.3 (Schrödinger LLC).

Figure 3.
Pseudokinases utilizing a one metal-binding mode.

(A) Crystal structures of the PKD of JAK2 exhibit a one metal-binding mode. The catalytic loop Asn still binds an Me2+ ion (most probably Me2), whereas the DFG-Asp binds the β3-Lys. (B) The DFG-Asp of the KSR2 interacts with an Arginine which replaces the β3-Lys and thereby the kinase binds just one metal ion, which is also coordinated by the catalytic loop Asn. (C) A highly unusual metal-binding mode is exhibited by the Poly(A) nuclease 3 (PAN3), which binds a metal ion utilizing amino acids of the extended Gly-rich loop. (D) ILK with its degenerated Me2-binding site solely binds Me1 with the DFG-Asp. In contrast with canonical protein kinases, the DFG-Asp functions as a monodentate instead of a bidentate ligand. Structures (JAK2JH2, PDB 4FVQ; KSR2, PDB 2Y4I; PAN3, PDB 4CYI; ILK, PDB 3KMW; Mg2+ ions are shown in green) were aligned and visualized using PyMOL v1.3 (Schrödinger LLC).

Pseudokinase-binding nucleotides in a Me2+-independent manner.
Figure 4.
Pseudokinase-binding nucleotides in a Me2+-independent manner.

(A) The pseudokinase CASK, although lacking both metal-binding sites, is able to bind ATP and was shown to be catalytically active. The DFG-Asp is replaced by a Gly and the catalytic loop Asn by a cysteine residue. This (pseudo)active site seems to be incapable of metal binding since kinase activity is suppressed by the presence of divalent metal ions. The introduction of four-point mutations resulted in a Me2+-dependent protein kinase, and the crystal structure revealed the presence of Me2 bound to the catalytic loop Asn and the DFG-Asp. (B) The STe-20-Related ADaptor (STRADα) utilizes several basic residues to effectively bind ATP. The β3-Lys is substituted by an Arg which binds the α-phosphoryl group, and Me1 and Me2 are functionally replaced by a His and Arg, respectively (CASK, PDB 3C0H; STRADα, PDB 3GNI; CASK(Mut), PDB 3MFU; the Mn2+ ion is shown in violet).

Figure 4.
Pseudokinase-binding nucleotides in a Me2+-independent manner.

(A) The pseudokinase CASK, although lacking both metal-binding sites, is able to bind ATP and was shown to be catalytically active. The DFG-Asp is replaced by a Gly and the catalytic loop Asn by a cysteine residue. This (pseudo)active site seems to be incapable of metal binding since kinase activity is suppressed by the presence of divalent metal ions. The introduction of four-point mutations resulted in a Me2+-dependent protein kinase, and the crystal structure revealed the presence of Me2 bound to the catalytic loop Asn and the DFG-Asp. (B) The STe-20-Related ADaptor (STRADα) utilizes several basic residues to effectively bind ATP. The β3-Lys is substituted by an Arg which binds the α-phosphoryl group, and Me1 and Me2 are functionally replaced by a His and Arg, respectively (CASK, PDB 3C0H; STRADα, PDB 3GNI; CASK(Mut), PDB 3MFU; the Mn2+ ion is shown in violet).

Janus kinase (JAK2)

ATP binding to the PKD (JAK homology 2, JH2) of the Janus kinase JAK2 is crucial for pathogenic activation. Janus kinases are nonreceptor tyrosine kinases carrying in tandem two kinase domains: a PKD and a canonical kinase domain [43]. JAKs are essential for cytokine-mediated signaling including the regulation of hematopoiesis and immune responses [43,44]. Interestingly, most disease-related mutations occur in the PKD, and it was shown recently that loss of ATP-binding in the PKD suppressed the effects of pathogenic JAK mutants [44]. While ATP binding to the PKD is necessary in a pathogenic variant (V617F), disruption of ATP binding in wild-type JAK2 showed no effect on JAK activation [44]. Although both metal-binding sites are present in the PKD, crystal structures resolved an unusual one metal-binding mode, where the DFG-Asp flips backwards to bind the β3-Lys (Figure 3) [45].

Kinase suppressor of Ras

The kinase suppressor of Ras (KSR) has been found in genetic screens utilized to identify components of the Ras-MAP kinase pathway. Although sharing profound similarities with RAF kinases, KSR function was thought to be kinase-independent, since kinase-dead mutants of KSR retained their regulatory function [46]. This kinase-independent function as a scaffolding protein is one hallmark of a pseudokinase; however, both KSR1 and kinase suppressor of Ras 2 (KSR2) show inefficient yet detectable enzymatic activity with MEK1 [47,48]. A crystal structure of the KSR2:MEK1 complex revealed a noncanonical metal-binding mode with a single Mg2+ (most probably Me2) bound to both metal sites (Figure 3) [47]. Owing to a second interaction of the DFG-Asp with an Arg replacing the β3-Lys, the second metal ion is missing and the polyphosphate chain of ATP is displaced. The accompanied incorrect positioning of the αC-Helix and the overall improper orientation of reactants seems to be responsible for the low catalytic efficiency. Despite that, ATP binding is stabilized by the metal ion and completing the C-spine is necessary for its association with MEK1 [49].

PolyANuclease 3

Binding of the pseudokinase PAN3 (PolyANuclease 3) to the exonuclease PAN2 is crucial for effective deadenylation and thereby degradation of mRNA [50]. PAN3 uses a unique one metal-binding mode to stabilize ATP binding at the pseudo-active site (Figure 3). Metal binding differs significantly from that observed in canonical protein kinases. Both metal-binding sites are degenerated and metal binding is accomplished by an Asp and a backbone carbonyl of a Ser residue in the extended Gly-rich loop [51]. This single magnesium ion is chelated by all phosphoryl groups of ATP and the canonical position of the HRDXXXXN Asn is occupied by a Lys, which directly interacts with the polyphosphate chain. The extended Gly-rich loop completely envelops the γ-phosphoryl group of ATP and thereby sterically hinders a nucleophilic attack of a substrate.

Metal-independent pseudokinases

Ca2+/calmodulin-dependent Ser–Thr kinase

The kinase domain of CASK (Ca2+/calmodulin-dependent Ser–Thr kinase) lacks both metal-binding sites and therefore was initially predicted to be a pseudokinase; however, phosphoryl transfer activity was reported with the synaptic substrate neurexin (Figure 4) [52]. The kinase does not require metal ions for catalysis; however, it is inhibited by the presence of Mg2+ already at micromolar concentrations [Ki (Mg2+) ∼60 µM]. This effect is most probably due to complexation of ATP by Mg2+, which, as a complex, is not able to bind the kinase [53]. With intracellular Mg2+ concentrations in the mM range, this kinase would be constitutively inactive. Therefore, this finding led to the conclusion that CASK activity may be regulated by synaptic activity-driven flux of metal ions [52]. The authors also created a mutant of CASK able to bind metal ions by introducing four-point mutations (Figure 4, lower panel). Reintroducing both metal sites was not sufficient to recover nucleotide (TNP-ATP) binding in the presence of Mg2+ [54]. Reconstituting the backbone interaction with the γ-phosphoryl group by replacing the Pro at the tip of the Gly-rich loop by Ala and changing the His immediately preceding the catalytic loop Asn with a Glu were also necessary to retrieve metal-dependent nucleotide binding. In the crystal structure of a CASK(4M) (CASK mutant with the following substitutions: G162D, C146N, P22A and H145E) mutant, the active site resembles a canonical kinase (e.g. PKA). Yet, the orientation of the polyphosphate chain differs slightly due to a one metal-binding mode where only Me2 (i.e. Mn2+) is bound. The loss of Me1 rearranges the position of the β3-Lys now binding the γ-phosphoryl group instead of the β-phosphoryl group.

STe-20-Related ADaptor

Another metal-independent pseudokinase is the STe-20-Related ADaptor (STRAD) which binds to and activates the tumor suppressor kinase LKB1 [55]. Several mutations in LKB1, leading to the Peutz–Jeghers cancer syndrome, interfere with the binding of STRADα [55,56]. The interaction of STRAD and LKB1 is strongly enhanced upon binding of the armadillo-repeat protein MO25 to STRAD [57]. MO25 shifts the pseudokinase into an active-like conformation due to positioning of the αC-Helix which is comparable to the effect of cyclins on CDKs [58]. This active-like conformation is also stabilized by ATP, thereby enhancing MO25 binding and vice versa [58]. ATP binding to STRAD is metal-independent and stabilization of the polyphosphate chain is accomplished exclusively by amino acid residues Lys, His and two Arg (Figure 3), which stabilize the binding of the nucleotide rather than promoting transfer of the terminal phosphoryl group [58,59]. This explains why no catalytic activity (i.e. substrate phosphorylation and autophosphorylation) of STRADα has been reported yet.

Concluding remarks

Canonical protein kinases employ divalent metal ions to bind ATP and to drive phosphoryl transfer. Likewise, several pseudokinases utilize metal ions to effectively bind ATP in order to enable specific functions (e.g. binding and activation of another kinase). However, due to displacement of the metal ions which lead to improper orientations of the polyphosphate chain, the phosphoryl transfer reaction is blocked while the active-like conformation is maintained. In this line, some pseudokinases seem to be able to bind metal ions in a nucleotide-independent manner and may utilize metal ions to stabilize specific conformations even without nucleotide binding.

Finally, several pseudokinases retain metal-binding properties to perform important physiological functions besides phosphoryl transfer by acting as switches and regulators, as scaffolds or signal transducers crucial in cellular cross-talk and signaling.BST-2016-0327CTB1 

Table 1
Comparison of metal ion-binding modes in protein kinase/pseudokinase structures
Protein kinase/pseudokinase PDB Resolution (Å) Me1 site Me2 site Binding mode Activity proved 
PKA 1ATP 2.2 Asp184 Asn171 Two metal 
RNaseL 4OAV 2.1 Asp503 + Asp505 Asn490 Two metal − 
POMK 5GZA 2.0 Asp225 + Asp227 Gln212 Two metal + (Glycan kinase) 
ROP5B 3Q60 1.72 Asp407 Asp393 Two metal − 
JAK2 JH2 4FVQ 1.75 1 Asn678 One metal (Me2) 
KSR2 2Y4I 3.46 2 Asn791 One metal (Me2) 
PAN3 4CYI 2.42 3 − Unique (one metal) − 
ILK 3KMW 2.0 Asp339 − One metal (Me1) − 
CASK 3C0H 2.3 − − No metal-binding + (Metal-independent) 
CASK(4M) mutant 3MFU 2.3 4 Asn146 One metal (Me2) 
STRADα 3GNI 2.35 − − No metal-binding − 
Protein kinase/pseudokinase PDB Resolution (Å) Me1 site Me2 site Binding mode Activity proved 
PKA 1ATP 2.2 Asp184 Asn171 Two metal 
RNaseL 4OAV 2.1 Asp503 + Asp505 Asn490 Two metal − 
POMK 5GZA 2.0 Asp225 + Asp227 Gln212 Two metal + (Glycan kinase) 
ROP5B 3Q60 1.72 Asp407 Asp393 Two metal − 
JAK2 JH2 4FVQ 1.75 1 Asn678 One metal (Me2) 
KSR2 2Y4I 3.46 2 Asn791 One metal (Me2) 
PAN3 4CYI 2.42 3 − Unique (one metal) − 
ILK 3KMW 2.0 Asp339 − One metal (Me1) − 
CASK 3C0H 2.3 − − No metal-binding + (Metal-independent) 
CASK(4M) mutant 3MFU 2.3 4 Asn146 One metal (Me2) 
STRADα 3GNI 2.35 − − No metal-binding − 
1

Canonical Asp (Asp699) is present; however, the structure of JAK2 JH2 shows no Me1.

2

In the available (low-resolution) structure, the canonical Asp (Asp803) binds only Me2.

3

Metal is bound by Asp257 and by the backbone carbonyl of Ser271 in the extended Gly-rich loop.

4

No Me1 present. Reintroduced DFG-Asp (Asp162) coordinates solely Me2.

Abbreviations

     
  • AlF3

    aluminium fluoride

  •  
  • CASK

    calcium/calmodulin (CaM)-activated serine–threonine kinase

  •  
  • CASK(4M)/CASKMut

    CASK mutant with the following substitutions: G162D, C146N, P22A and H145E

  •  
  • C-spine

    catalytic spine

  •  
  • ePKs

    eukaryotic protein kinases

  •  
  • ILK

    integrin-linked kinase

  •  
  • EPK

    eukaryotic protein kinase

  •  
  • JAK2(JH2)

    Janus kinase 2 (JAK homology 2)

  •  
  • KSR2

    kinase suppressor of RAS 2

  •  
  • LKB1

    liver kinase B1

  •  
  • Me1

    metal-binding site 1

  •  
  • Me2

    metal-binding site 2

  •  
  • Mnk1 and Mnk2

    mitogen-activated protein kinase-interacting kinases 1 and 2

  •  
  • MEK1

    mitogen-activated protein kinase kinase

  •  
  • PAN3

    poly(A) nuclease 3

  •  
  • PKI

    heat-stable protein kinase inhibitor of PKA

  •  
  • PKA

    protein kinase A

  •  
  • PKD

    pseudokinase domain

  •  
  • POMK

    protein-O-mannose kinase

  •  
  • RAF

    rapidly accelerated fibrosarcoma Ser/Thr protein kinase

  •  
  • RNaseL

    ribonuclease L

  •  
  • ROP5B

    Rhoptry organelle protein 5B

  •  
  • R-spine

    regulatory spine

  •  
  • STRADα

    STe-20-Related ADaptor alpha

  •  
  • TNP-ATP

    2′,3′-O-trinitrophenyl-adenosine-5′-triphosphate.

Funding

This work is supported by the Michael J. Fox Foundation [Grant ID: 11425] and the Deutsche Forschungsgemeinschaft [Grant ID: He1818/10] to FWH.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
Cohen
,
P.
(
2001
)
The role of protein phosphorylation in human health and disease
.
Eur. J. Biochem.
268
,
5001
5010
doi:
2
Manning
,
G.
,
Whyte
,
D.B.
,
Martinez
,
R.
,
Hunter
,
T.
and
Sudarsanam
,
S.
(
2002
)
The protein kinase complement of the human genome
.
Science
298
,
1912
1934
doi:
3
Kannan
,
N.
,
Taylor
,
S.S.
,
Zhai
,
Y.
,
Venter
,
J.C.
and
Manning
,
G.
(
2007
)
Structural and functional diversity of the microbial kinome
.
PLoS Biol.
5
,
e17
doi:
4
Taylor
,
S.S.
,
Buechler
,
J.A.
and
Yonemoto
,
W.
(
1990
)
cAMP-dependent protein kinase: framework for a diverse family of regulatory enzymes
.
Annu. Rev. Biochem.
59
,
971
1005
doi:
5
Knighton
,
D.R.
,
Zheng
,
J.H.
,
Eyck Ten
,
L.F.
,
Ashford
,
V.A.
,
Xuong
,
N.H.
,
Taylor
,
S.S.
et al. 
(
1991
)
Crystal structure of the catalytic subunit of cyclic adenosine monophosphate-dependent protein kinase
.
Science
253
,
407
414
doi:
6
Zheng
,
J.
,
Knighton
,
D.R.
,
Eyck Ten
,
L.F.
,
Karlsson
,
R.
,
Xuong
,
N.
,
Taylor
,
S.S.
et al. 
(
1993
)
Crystal structure of the catalytic subunit of cAMP-dependent protein kinase complexed with magnesium-ATP and peptide inhibitor
.
Biochemistry
32
,
2154
2161
doi:
7
Hanks
,
S.K.
,
Quinn
,
A.M.
and
Hunter
,
T
. (
1988
)
The protein kinase family: conserved features and deduced phylogeny of the catalytic domains
.
Science
241
,
42
52
doi:
8
Taylor
,
S.S.
,
Keshwani
,
M.M.
,
Steichen
,
J.M.
and
Kornev
,
A.P.
(
2012
)
Evolution of the eukaryotic protein kinases as dynamic molecular switches
.
Philos. Trans. R. Soc. B Biol. Sci.
367
,
2517
2528
doi:
9
Xu
,
B.-e.
,
English
,
J.M.
,
Wilsbacher
,
J.L.
,
Stippec
,
S.
,
Goldsmith
,
E.J.
and
Cobb
,
M.H.
(
2000
)
WNK1, a novel mammalian serine/threonine protein kinase lacking the catalytic lysine in subdomain II
.
J. Biol. Chem.
275
,
16795
16801
doi:
10
Bailey
,
F.P.
,
Byrne
,
D.P.
,
Oruganty
,
K.
,
Eyers
,
C.E.
,
Novotny
,
C.J.
,
Shokat
,
K.M.
et al. 
(
2015
)
The Tribbles 2 (TRB2) pseudokinase binds to ATP and autophosphorylates in a metal-independent manner
.
Biochem. J.
467
,
47
62
doi:
11
Meharena
,
H.S.
,
Chang
,
P.
,
Keshwani
,
M.M.
,
Oruganty
,
K.
,
Nene
,
A.K.
,
Kannan
,
N.
et al. 
(
2013
)
Deciphering the structural basis of eukaryotic protein kinase regulation
.
PLoS Biol.
11
,
e1001680
doi:
12
Kornev
,
A.P.
,
Haste
,
N.M.
,
Taylor
,
S.S.
and
Ten Eyck
,
L.F.
(
2006
)
Surface comparison of active and inactive protein kinases identifies a conserved activation mechanism
.
Proc. Natl Acad. Sci. U.S.A.
103
,
17783
17788
doi:
13
Shaw
,
A.S.
,
Kornev
,
A.P.
,
Hu
,
J.
,
Ahuja
,
L.G.
and
Taylor
,
S.S.
(
2014
)
Kinases and pseudokinases: lessons from RAF
.
Mol. Cell. Biol.
34
,
1538
1546
doi:
14
Steichen
,
J.M.
,
Iyer
,
G.H.
,
Li
,
S.
,
Saldanha
,
S.A.
,
Deal
,
M.S.
,
Woods
,
V.L.
et al. 
(
2010
)
Global consequences of activation loop phosphorylation on protein kinase A
.
J. Biol. Chem.
285
,
3825
3832
doi:
15
Burnett
,
G.
and
Kennedy
,
E.P.
(
1954
)
The enzymatic phosphorylation of proteins
.
J. Biol. Chem.
211
,
969
980
PMID:
[PubMed]
16
Krebs
,
E.G.
and
Fischer
,
E.H.
(
1955
)
Phosphorylase activity of skeletal muscle extracts
.
J. Biol. Chem.
216
,
113
120
PMID:
[PubMed]
17
Fischer
,
E.H.
and
Krebs
,
E.G.
(
1955
)
Conversion of phosphorylase b to phosphorylase a in muscle extracts
.
J. Biol. Chem.
216
,
121
132
PMID:
[PubMed]
18
Krebs
,
E.G.
and
Fischer
,
E.H.
(
1956
)
The phosphorylase b to a converting enzyme of rabbit skeletal muscle
.
Biochim. Biophys. Acta
20
,
150
157
doi:
19
Armstrong
,
R.N.
,
Kondo
,
H.
,
Granot
,
J.
,
Kaiser
,
E.T.
and
Mildvan
,
A.S.
(
1979
)
Magnetic resonance and kinetic studies of the manganese(II) ion and substrate complexes of the catalytic subunit of adenosine 3′,5′-monophosphate dependent protein kinase from bovine heart
.
Biochemistry
18
,
1230
1238
doi:
20
Bastidas
,
A.C.
,
Deal
,
M.S.
,
Steichen
,
J.M.
,
Guo
,
Y.
,
Wu
,
J.
and
Taylor
,
S.S.
(
2013
)
Phosphoryl transfer by protein kinase A is captured in a crystal lattice
.
J. Am. Chem. Soc.
135
,
4788
4798
doi:
21
Bao
,
Z.-Q.
,
Jacobsen
,
D.M.
and
Young
,
M.A.
(
2011
)
Briefly bound to activate: transient binding of a second catalytic magnesium activates the structure and dynamics of CDK2 kinase for catalysis
.
Structure
19
,
675
690
doi:
22
Jacobsen
,
D.M.
,
Bao
,
Z.-Q.
,
O'Brien
,
P.
,
Brooks
, III,
C.L.
and
Young
,
M.A.
(
2012
)
Price to be paid for two-metal catalysis: magnesium ions that accelerate chemistry unavoidably limit product release from a protein kinase
.
J. Am. Chem. Soc.
134
,
15357
15370
doi:
23
Dudev
,
T.
and
Lim
,
C.
(
2009
)
Metal-binding affinity and selectivity of nonstandard natural amino acid residues from DFT/CDM calculations
.
J. Phys. Chem. B
113
,
11754
11764
doi:
24
Dudev
,
T.
and
Lim
,
C.
(
2014
)
Competition among metal ions for protein binding sites: determinants of metal ion selectivity in proteins
.
Chem. Rev.
114
,
538
556
doi:
25
Herberg
,
F.W.
,
Doyle
,
M.L.
,
Cox
,
S.
and
Taylor
,
S.S.
(
1999
)
Dissection of the nucleotide and metal-phosphate binding sites in cAMP-dependent protein kinase
.
Biochemistry
38
,
6352
6360
doi:
26
Madhusudan
,
P.A.
,
Xuong
,
N.-H.
and
Taylor
,
S.S.
(
2002
)
Crystal structure of a transition state mimic of the catalytic subunit of cAMP-dependent protein kinase
.
Nat. Struct. Biol.
9
,
273
277
doi:
27
Sugden
,
P.H.
,
Holladay
,
L.A.
,
Reimann
,
E.M.
and
Corbin
,
J.D.
(
1976
)
Purification and characterization of the catalytic subunit of adenosine 3′:5′-cyclic monophosphate-dependent protein kinase from bovine liver
.
Biochem. J.
159
,
409
422
doi:
28
Gerlits
,
O.
,
Tian
,
J.
,
Das
,
A.
,
Langan
,
P.
,
Heller
,
W.T.
and
Kovalevsky
,
A.
(
2015
)
Phosphoryl transfer reaction snapshots in crystals: insights into the mechanism of protein kinase a catalytic subunit
.
J. Biol. Chem.
290
,
15538
15548
doi:
29
Knape
,
M.J.
,
Ahuja
,
L.G.
,
Bertinetti
,
D.
,
Burghardt
,
N.C.G.
,
Zimmermann
,
B.
,
Taylor
,
S.S.
et al. 
(
2015
)
Divalent metal ions Mg2+ and Ca2+ have distinct effects on protein kinase A activity and regulation
.
ACS Chem. Biol.
10
,
2303
2315
doi:
30
Zimmermann
,
B.
,
Schweinsberg
,
S.
,
Drewianka
,
S.
and
Herberg
,
F.W.
(
2008
)
Effect of metal ions on high-affinity binding of pseudosubstrate inhibitors to PKA
.
Biochem. J.
413
,
93
101
doi:
31
McClendon
,
C.L.
,
Kornev
,
A.P.
,
Gilson
,
M.K.
and
Taylor
,
S.S.
(
2014
)
Dynamic architecture of a protein kinase
.
Proc. Natl Acad. Sci. U.S.A.
111
,
E4623
E4631
doi:
32
Murphy
,
J.M.
,
Zhang
,
Q.
,
Young
,
S.N.
,
Reese
,
M.L.
,
Bailey
,
F.P.
,
Eyers
,
P.A.
et al. 
(
2014
)
A robust methodology to subclassify pseudokinases based on their nucleotide-binding properties
.
Biochem. J.
457
,
323
334
doi:
33
Han
,
Y.
,
Donovan
,
J.
,
Rath
,
S.
,
Whitney
,
G.
,
Chitrakar
,
A.
and
Korennykh
,
A.
(
2014
)
Structure of human RNase L reveals the basis for regulated RNA decay in the IFN response
.
Science
343
,
1244
1248
doi:
34
Huang
,
H.
,
Zeqiraj
,
E.
,
Dong
,
B.
,
Jha
,
B.K.
,
Duffy
,
N.M.
,
Orlicky
,
S.
et al. 
(
2014
)
Dimeric structure of pseudokinase RNase L bound to 2-5A reveals a basis for interferon-induced antiviral activity
.
Mol. Cell
53
,
221
234
doi:
35
Jauch
,
R.
,
Jäkel
,
S.
,
Netter
,
C.
,
Schreiter
,
K.
,
Aicher
,
B.
,
Jäckle
,
H.
et al. 
(
2005
)
Crystal structures of the Mnk2 kinase domain reveal an inhibitory conformation and a zinc binding site
.
Structure
13
,
1559
1568
doi:
36
Yoshida-Moriguchi
,
T.
,
Willer
,
T.
,
Anderson
,
M.E.
,
Venzke
,
D.
,
Whyte
,
T.
,
Muntoni
,
F.
et al. 
(
2013
)
SGK196 is a glycosylation-specific O-mannose kinase required for dystroglycan function
.
Science
341
,
896
899
doi:
37
Zhu
,
Q.
,
Venzke
,
D.
,
Walimbe
,
A.S.
,
Anderson
,
M.E.
,
Fu
,
Q.
,
Kinch
,
L.N.
et al. 
(
2016
)
Structure of protein O-mannose kinase reveals a unique active site architecture
.
eLife
5
,
e22238
doi:
38
Talevich
,
E.
and
Kannan
,
N.
(
2013
)
Structural and evolutionary adaptation of rhoptry kinases and pseudokinases, a family of coccidian virulence factors
.
BMC Evol. Biol.
13
,
117
doi:
39
Fox
,
B.A.
,
Rommereim
,
L.M.
,
Guevara
,
R.B.
,
Falla
,
A.
,
Hortua Triana
,
M.A.
,
Sun
,
Y.
et al. 
(
2016
)
The Toxoplasma gondii rhoptry kinome is essential for chronic infection
.
mBio
7
,
e00193-16
doi:
40
Reese
,
M.L.
and
Boothroyd
,
J.C.
(
2011
)
A conserved non-canonical motif in the pseudoactive site of the ROP5 pseudokinase domain mediates its effect on Toxoplasma virulence
.
J. Biol. Chem.
286
,
29366
29375
doi:
41
Johnson
,
D.A.
,
Akamine
,
P.
,
Radzio-Andzelm
,
E.
,
Madhusudan
,
M.
and
Taylor
,
S.S.
(
2001
)
Dynamics of cAMP-dependent protein kinase
.
Chem. Rev.
101
,
2243
2270
doi:
42
Fukuda
,
K.
,
Gupta
,
S.
,
Chen
,
K.
,
Wu
,
C.
and
Qin
,
J.
(
2009
)
The pseudoactive site of ILKIs essential for its binding to α-parvin and localization to focal adhesions
.
Mol. Cell
36
,
819
830
doi:
43
Babon
,
J.J.
,
Lucet
,
I.S.
,
Murphy
,
J.M.
,
Nicola
,
N.A.
and
Varghese
,
L.N.
(
2014
)
The molecular regulation of Janus kinase (JAK) activation
.
Biochem. J.
462
,
1
13
doi:
44
Hammarén
,
H.M.
,
Ungureanu
,
D.
,
Grisouard
,
J.
,
Skoda
,
R.C.
,
Hubbard
,
S.R.
and
Silvennoinen
,
O.
(
2015
)
ATP binding to the pseudokinase domain of JAK2 is critical for pathogenic activation
.
Proc. Natl Acad. Sci. U.S.A.
112
,
4642
4647
doi:
45
Bandaranayake
,
R.M.
,
Ungureanu
,
D.
,
Shan
,
Y.
,
Shaw
,
D.E.
,
Silvennoinen
,
O.
and
Hubbard
,
S.R.
(
2012
)
Crystal structures of the JAK2 pseudokinase domain and the pathogenic mutant V617F
.
Nat. Struct. Mol. Biol.
19
,
754
759
doi:
46
Stewart
,
S.
,
Sundaram
,
M.
,
Zhang
,
Y.
,
Lee
,
J.
,
Han
,
M.
and
Guan
,
K.-L.
(
1999
)
Kinase suppressor of Ras forms a multiprotein signaling complex and modulates MEK localization
.
Mol. Cell Biol.
19
,
5523
5534
doi:
47
Brennan
,
D.F.
,
Dar
,
A.C.
,
Hertz
,
N.T.
,
Chao
,
W.C.H.
,
Burlingame
,
A.L.
,
Shokat
,
K.M.
et al. 
(
2011
)
A Raf-induced allosteric transition of KSR stimulates phosphorylation of MEK
.
Nature
472
,
366
369
doi:
48
Zhang
,
H.
,
Koo
,
C.Y.
,
Stebbing
,
J.
and
Giamas
,
G.
(
2013
)
The dual function of KSR1: a pseudokinase and beyond
.
Biochem. Soc. Trans.
41
,
1078
1082
doi:
49
Hu
,
J.
,
Yu
,
H.
,
Kornev
,
A.P.
,
Zhao
,
J.
,
Filbert
,
E.L.
,
Taylor
,
S.S.
et al. 
(
2011
)
Mutation that blocks ATP binding creates a pseudokinase stabilizing the scaffolding function of kinase suppressor of Ras, CRAF and BRAF
.
Proc. Natl Acad. Sci. U.S.A.
108
,
6067
6072
doi:
50
Jonas
,
S.
,
Christie
,
M.
,
Peter
,
D.
,
Bhandari
,
D.
,
Loh
,
B.
,
Huntzinger
,
E.
et al. 
(
2014
)
An asymmetric PAN3 dimer recruits a single PAN2 exonuclease to mediate mRNA deadenylation and decay
.
Nat. Struct. Mol. Biol.
21
,
599
608
doi:
51
Christie
,
M.
,
Boland
,
A.
,
Huntzinger
,
E.
,
Weichenrieder
,
O.
and
Izaurralde
,
E.
(
2013
)
Structure of the PAN3 pseudokinase reveals the basis for interactions with the PAN2 deadenylase and the GW182 proteins
.
Mol. Cell
51
,
360
373
doi:
52
Mukherjee
,
K.
,
Sharma
,
M.
,
Urlaub
,
H.
,
Bourenkov
,
G.P.
,
Jahn
,
R.
,
Südhof
,
T.C.
et al. 
(
2008
)
CASK functions as a Mg2+-independent neurexin kinase
.
Cell
133
,
328
339
doi:
53
Kannan
,
N.
and
Taylor
,
S.S.
(
2008
)
Rethinking pseudokinases
.
Cell
133
,
204
205
doi:
54
Mukherjee
,
K.
,
Sharma
,
M.
,
Jahn
,
R.
,
Wahl
,
M.C.
and
Südhof
,
T.C.
(
2010
)
Evolution of CASK into a Mg2+-sensitive kinase
.
Sci. Signal.
3
,
ra33
doi:
55
Baas
,
A.F.
,
Boudeau
,
J.
,
Sapkota
,
G.P.
,
Smit
,
L.
,
Medema
,
R.
,
Morrice
,
N.A.
et al. 
(
2003
)
Activation of the tumour suppressor kinase LKB1 by the STE20-like pseudokinase STRAD
.
EMBO J.
22
,
3062
3072
doi:
56
Boudeau
,
J.
,
Scott
,
J.W.
,
Resta
,
N.
,
Deak
,
M.
,
Kieloch
,
A.
,
Komander
,
D.
et al. 
(
2004
)
Analysis of the LKB1-STRAD-MO25 complex
.
J. Cell. Sci.
117
,
6365
6375
doi:
57
Boudeau
,
J.
,
Baas
,
A.F.
,
Deak
,
M.
,
Morrice
,
N.A.
,
Kieloch
,
A.
,
Schutkowski
,
M.
et al. 
(
2003
)
MO25α/β interact with STRADα/β enhancing their ability to bind, activate and localize LKB1 in the cytoplasm
.
EMBO J.
22
,
5102
5114
doi:
58
Zeqiraj
,
E.
,
Filippi
,
B.M.
,
Goldie
,
S.
,
Navratilova
,
I.
,
Boudeau
,
J.
,
Deak
,
M.
et al. 
(
2009
)
ATP and MO25a regulate the conformational state of the STRADa pseudokinase and activation of the LKB1 tumour suppressor
.
PLoS Biol.
7
,
e1000126
doi:
59
Zeqiraj
,
E.
,
Filippi
,
B.M.
,
Deak
,
M.
,
Alessi
,
D.R.
and
van Aalten
,
D.M.F.
(
2009
)
Structure of the LKB1-STRAD-MO25 complex reveals an allosteric mechanism of kinase activation
.
Science
326
,
1707
1711
doi: