protein kinase RAF (rapidly accelerated fibrosarcoma) comprises three isoforms,
A-RAF, B-RAF and C-RAF, which despite their homology differ in the mechanism
of their activation. The highly studied family member B-RAF dimerizes in a
typical receiver/acceptor asymmetric dimer. Interestingly, kinase inhibitors
that bind to the active state of B-RAF stabilize the activator kinase resulting in
paradoxical activation of MAPK signalling [59, 60]. The activation of the pathway
is also thought to be mediated by C-RAF though B-RAF/C-RAF heterodimerization
[61–63]. Kinase inactive mutants of B-RAF, but not C-RAF, can therefore still
activate the MAPK cascade by acting as allosteric B-RAF activators [64]. The
structural reasons explaining the inability of C-RAF kinase dead mutants activating
MAPK signalling have recently been elucidated [65]. Full activity of RAF requires
phosphorylation at the activation segment [66, 67] at two sites as well as at the
N-terminal acidic (NtA) motif [68, 69]. In B-RAF the NtA motif is acidic (sequence
SSDD) and constitutively phosphorylated [69]. In contrast C-RAF and A-RAF
lack the two acidic aspartate residues requiring phosphorylation by upstream
kinases on their SSYY and SGYY NtA motifs [70]. Hu et al. revealed that NtA
motif phosphorylation is only required for the “activator” but not the “receiver”
kinase offering a rational why B-RAF can activate C-RAF but not vice versa
[65]. Importantly, the oncogenic mutant B-RAF(V600E) does not require dimerization for activity explaining the efficacy of B-RAF inhibitors supressing MAPK
signalling in tumours harbouring this mutant but not in wild-type tissue where
activation is observed. The paradoxical activation of MAPK signalling in wildtype tissue has been associated with the development of both benign and malignant
cutaneous manifestations, ranging from seborrheic dermatitis-like rashes to eruptive
keratoacanthomas and squamous cell carcinomas [71]. These examples demonstrate
how the complex activation mechanisms of protein kinases may lead to unexpected
adverse clinical manifestations. A schematic of the B-RAF/C-RAF activation model
is shown in Fig. 4c.
3.1 Kinase Activation by Interacting Domains and Proteins
The dimerization models of kinase activation highlight the importance of protein
interactions stabilizing the kinase active state and suggest that also other proteins
and domains may act as kinase activators or inhibitors. Indeed, a large number
of interactions regulating kinase activity have been described which include
flanking domains, for instance, the SH2 which plays a role stabilizing the inactive
[72, 73] as well as active state [74, 75], and these interactions might be exploited
therapeutically [76].
One of the best studied examples of kinase regulation by an interacting protein
is the cyclin-dependent kinases (CDKs) which are stringently regulated by their
interaction partners the cyclins. CDKs are master regulators of the cell cycle, and
dysfunction of CDK regulation is a major driver of tumourigenesis and attractive
drug targets [77–79]. CDKs typically require activation segment as well as binding
of a cyclin for full activity [80–82]. In addition, ATP and peptide binding contribute
Function, Structure and Topology of Protein Kinases
9
A-RAF, B-RAF and C-RAF, which despite their homology differ in the mechanism
of their activation. The highly studied family member B-RAF dimerizes in a
typical receiver/acceptor asymmetric dimer. Interestingly, kinase inhibitors
that bind to the active state of B-RAF stabilize the activator kinase resulting in
paradoxical activation of MAPK signalling [59, 60]. The activation of the pathway
is also thought to be mediated by C-RAF though B-RAF/C-RAF heterodimerization
[61–63]. Kinase inactive mutants of B-RAF, but not C-RAF, can therefore still
activate the MAPK cascade by acting as allosteric B-RAF activators [64]. The
structural reasons explaining the inability of C-RAF kinase dead mutants activating
MAPK signalling have recently been elucidated [65]. Full activity of RAF requires
phosphorylation at the activation segment [66, 67] at two sites as well as at the
N-terminal acidic (NtA) motif [68, 69]. In B-RAF the NtA motif is acidic (sequence
SSDD) and constitutively phosphorylated [69]. In contrast C-RAF and A-RAF
lack the two acidic aspartate residues requiring phosphorylation by upstream
kinases on their SSYY and SGYY NtA motifs [70]. Hu et al. revealed that NtA
motif phosphorylation is only required for the “activator” but not the “receiver”
kinase offering a rational why B-RAF can activate C-RAF but not vice versa
[65]. Importantly, the oncogenic mutant B-RAF(V600E) does not require dimerization for activity explaining the efficacy of B-RAF inhibitors supressing MAPK
signalling in tumours harbouring this mutant but not in wild-type tissue where
activation is observed. The paradoxical activation of MAPK signalling in wildtype tissue has been associated with the development of both benign and malignant
cutaneous manifestations, ranging from seborrheic dermatitis-like rashes to eruptive
keratoacanthomas and squamous cell carcinomas [71]. These examples demonstrate
how the complex activation mechanisms of protein kinases may lead to unexpected
adverse clinical manifestations. A schematic of the B-RAF/C-RAF activation model
is shown in Fig. 4c.
3.1 Kinase Activation by Interacting Domains and Proteins
The dimerization models of kinase activation highlight the importance of protein
interactions stabilizing the kinase active state and suggest that also other proteins
and domains may act as kinase activators or inhibitors. Indeed, a large number
of interactions regulating kinase activity have been described which include
flanking domains, for instance, the SH2 which plays a role stabilizing the inactive
[72, 73] as well as active state [74, 75], and these interactions might be exploited
therapeutically [76].
One of the best studied examples of kinase regulation by an interacting protein
is the cyclin-dependent kinases (CDKs) which are stringently regulated by their
interaction partners the cyclins. CDKs are master regulators of the cell cycle, and
dysfunction of CDK regulation is a major driver of tumourigenesis and attractive
drug targets [77–79]. CDKs typically require activation segment as well as binding
of a cyclin for full activity [80–82]. In addition, ATP and peptide binding contribute
Function, Structure and Topology of Protein Kinases
9