have shown that binding of TANK to IKKε leads to its
phosphorylation and Lys63-linked polyubiquitination, both of which are required for IRF3 activation 23. Our data suggest that NAP1 can serve as substrate of IKKε as well. Whether phosphorylation and polyubiquitination of NAP1 and SINTBAD are also a prerequisite for IRF3 activation remains to be addressed. Heterodimerization with TBK1 appears to be mediated by a different region of IKKε since all IKKε isoforms could be coprecipitated with TBK1 (Fig. 9). Similarly, homodimerization of IKKε is not prevented buy BYL719 in the absence of the C-terminus (Fig. 8). An interesting candidate region possibly mediating these interactions is the ubiquitin-like domain (ULD). It has been shown that the ULD of IKKε and TBK1 bind to their respective kinase domains 33. Due to the high degree of homology between both kinase domains, it would be conceivable that homo or heterodimerization of these proteins might be mediated by an interaction between ULD and kinase domain as shown in Supporting Information HCS assay Fig. S4. The exact mechanism of NF-κB activation by IKKε
is still unclear 21. Initially, the proteins TANK and NAP1 have been described as IKKε-binding adapters mediating NF-κB activation 34, 35. Here, we could clearly rule out the involvement of TANK and NAP1 in IKKε-induced NF-κB activation since both proteins did not interact with the splice variant IKKε-sv1. Phosphorylation of p65/RelA has been described as another possible mechanism by which IKKε may activate NF-κB-mediated gene transcription. Here, we could confirm phosphorylation of Ser536 and Ser468 in cells overexpressing IKKε as reported previously 17, 18. However, our data suggest that phosphorylation of p65/RelA
even at both sites is insufficient to activate NF-κB-driven Selleckchem MG 132 gene expression (Fig. 4), indicating that most likely several mechanisms are involved in IKKε-mediated NF-κB activation. Recently, IKKε was shown to directly phosphorylate the deubiquitinating enzyme CYLD thereby inactivating its ability to suppress NF-κB activation 36. Whether phosphorylation of CYLD by some of the IKKε isoforms correlates with their capability to activate NF-κB-dependent transcription remains to be investigated. The protein domain(s) of IKKε that are required for NF-κB activation have not been identified. So far, we have demonstrated the requirement of a domain containing amino acids 647–684. Interestingly, a second coiled-coil region is located between residues 628 and 659 and could therefore be a motif either interacting with NF-κB proteins as direct substrates (such as p65/RelA), with the deubiquitinase CYLD, or with adapter proteins relaying the signal (Supporting Information Fig. S4). The IKKε mutant IKKε-Δ647 displayed reduced binding to TBK1 (Fig. 9A). Therefore, it is possible that TBK1 is partially involved in IKKε-induced NF-κB activation.