About 10% kinase domains lack catalytic residues required for catalytic activity. Many have been shown experimentally to lack activity, while others have been associated with weak kinase activity, with varying degrees of confidence.
Defining a pseudokinase
The first catalog of pseudokinases Manning looked for the presence of three residues in ePK domains: K72 (salt-bridge and ATP binding), D166 (catalytic aspartate) and D184 (Mg-binding). These were thought to be essential for catalysis. 50 of 478 human ePKs had pseudokinases by this definition, including 5 that had a second, active, kinase domain.
Other highly conserved residues that may be important for defining pseudokinases include E91 (salt bridge), N171 (catalytic loop and metal ion binding) and the GxGxxG ATP-binding motif (G-rich loop), though this sequence is quite variable, a more inclusive consensus is [G/A/S]xGxx[G/A/S].
Finding activity in predicted pseudokinases
Several pseudokinases have recently been described to have catalytic activity.
CASK lacks D184 (and N171), indicating that it cannot bind Mg ions. However, it was shown to have a low rate of Mg-independent biochemical activity - in fact it is in hibited by Mg Mukherjee. This correlates with its activity in neurons, where resting Mg concentration is low, but massively increased during synaptic transmission, when CASK is inhibited. A phylogenetic study Mukherjee2 showed that the earliest branching form, in Trichoplax adherens, is predicted to be fully active as a Mg-dependent kinase, and a Nematostella (cnidarian) homolog is possibly active, while metazoans - with established nerve tissue - all lack the N171 and D184 residues. Mutations of human CASK to the Trichoplax form in these key residues restored Mg-dependent catalysis to the enzyme.
Jak kinases have a catalytically active kinase domain (JH1) followed by a highly degenerate pseudokinase domain (JH2), which has changed D166 to an N. However, JH2 from human Jak2 was recently shown to autophosphorylate on both Ser and Tyr, at about 10% of the catalytic rate of JH1 Ungureanu. The autophosphorylation increased the activity of JH2 to inhibit JH1 activity, and three disease point mutations in JH2 that were associated with hyperactive Jak2 were shown to block the Ser autophosphorylation.
V617F, results in constitutively active JAK2 and is responsible for >95% of cases of polycythemia vera and ~50% of cases of essential thrombocythemia and primary myelofibrosis17, 18, 19
"ErbB3" is one of four members of the EGFR family of receptor tyrosine kinases. It signals by heterodimerization with other EGFRs (including HER2, which has an active kinase domain but no known ligand-binding activity). Weak activity (~0.1% that of EGFRs) was seen when purified protein was concentrated by lipid association, but the physiological relevance of this activity is not known.
SCYL2 is a member of the SCYL family which are predicted pseudokinases found in all eukaryotes. This family typically lacks all three catalytic residues and has a very divergent sequence. However, one report Conner demonstrated a kinase activity associated with bacculovirus-purified SCYL2 (CVAK104) in the presence of poly-L-lysine. Given the degree of degeneration of the catalytic motifs of SCYL, and the false positives sometimes seen in biochemical experiments (see e.g. H11), this activity is probably in need of further validation.
What does it take to kill a kinase?
Several studies have shown that single mutations designed to kill kinase activity (typically K72R or K72M) can result in some residual activity. It may be that many dead kinases are no longer *designed* to be active, but may still retain some substrate or ATP binding ability, and so act as weak catalysts. A major question for such partially-active kinases is whether their activity is biologically relevant. It may be that some weak activities can be amplified by proximity: for instance, CASK phosphorylates substrates recruited by its PDZ domain, Jak autophosphorylates, and ErbB3 phosphorylation may be meaningful in the context of EGFR family heterodimers.
- Conner pmid=15809293
- Manning pmid=12471243
- Ungureanu pmid=21841788
- Mukherjee pmid=18423203
- Mukherjee2 pmid=20424264