Functionally Distinct Effects of the C-Terminal Regions of IKKε and TBK1 on Type I IFN Production

Inhibitor of κB kinase ε (IKKε) and TANK binding kinase 1 (TBK1), so-called non-canonical IKKs or IKK-related kinases, are involved in the cellular innate immunity by inducing type I IFNs. Two kinases commonly phosphorylate transcription factors IRF3 and IRF7 in type I IFN production pathway. In contrast to TBK1, underlying mechanisms of IKKε activation and regions required for activation of downstream molecules are poorly understood. In this study, we investigated regions of IKKε required for the activation of type I IFN promoter specially, by focusing on the C-terminal region. To show the functional significance of the IKKε C-terminal region on type I IFN production, we employed various mutant forms of IKKε and compared to corresponding region of TBK1. We identified the specific regions and residues of IKKε involved in the activation of downstream signaling. Interestingly, corresponding region and residues are not required for activation of downstream signaling by TBK1. The results highlight the importance of the C-terminal region in the functional activity of IKKε in innate immune response and also the difference in activation mechanisms between IKKε and the closely related TBK1.


Introduction
Viral infection induces the cellular immune responses, which prevent viral propagation and pathogenic activities. A family of inhibitor of kB (IkB) kinase (IKK) is activated in response to pathogen infection, and these kinases control both innate and adaptive immunity [1,2]. Among them, TANK-binding kinase 1 (TBK1) and IKKe play a key role in the expression of cellular antiviral effects by inducing type I interferon (IFN) production [3,4,5]. Viral nucleic acids are sensed by cellular pattern-recognition receptors (PRRs), such as Toll-like and RIG-I-like receptors as well as cytosolic DNA receptors. These PRRs induce the autophosphorylation of Ser 172 located in the T loop of the TBK1 and IKKe kinase domains, which is essential for the enhancement of kinase activity [1,4,6,7,8,9]. Activated TBK1 and IKKe phosphorylate IRF3 and IRF7, leading to the nuclear translocation of these transcription factors and subsequent induction of type I IFN promoter activity [3,4,10,11].TBK1 is constitutively expressed in a broad range of cells, while the expression of IKKe is inducible and predominantly takes place in immune cells [12,13,14]. Despite these differences, TBK1 and IKKe are found together in a complex and are similarly targeted to the phosphorylation of the C-terminal Ser/Thr rich region of IRF3 and IRF7 [3,15]. The crystal structure of TBK1 revealed that it has a trimodular architecture with an N-terminal kinase domain (KD), followed by the internally located ubiquitin-like domain (ULD) and the C-terminal helical scaffold dimerization domain (SDD), similar to IKKb [16,17,18]. TBK1 takes on a dimer configuration in crystallized form and dimerization is also required for the activation through auto-phosphorylation of Ser172. The TBK1 dimer is stabilized by an extensive network of interactions among the KD, ULD and SDD, while IKKb which dimerization is mediated by a C-terminal region of SDD [16,17,18].In structural studies on TBK1, a C-terminally truncated fragment was used (residue 1-657), since this TBK1/1-657 fragment forms a dimer both in vitro and in vivo and is able to induce IRF3 phosphorylation, leading activation of the type I IFN promoter [17,18]. Therefore, the C-terminal region of TBK1 is dispensable for enhancing TBK1 kinase activity as well as activating downstream signaling, at least in this overexpression system [17,18,19,20]. These observations give rise to the question of whether the activation sequence of IKKe follows the same pattern as TBK1. To investigate this, we focused on the C-terminal region of IKKe and investigated its functional significance upon the dimerization of IKKe, phosphorylation of IRF3 and activation of the IFNb promoter, respectively.

Reagents and plasmids
Mouse monoclonal antibodies against FLAG M2 and a-tubulin and anti-FLAG agarose beads were purchased from Sigma (St Louis, MO). Mouse monoclonal antibody for V5-tag and Alexa Fluor 488 and 594-conjugated secondary antibodies were purchased from Invitrogen. Rabbit polyclonal antibody against IRF3 was from Becton Dickinson (Franklin Lakes, NJ). Rabbit antibodies against phospho-IRF3 (pSer396) and phospho-IRF3 (pSer386) were from Cell Signaling Technology (Danvers, MA) and EPITOMICS (Burlingame, CA), respectively. FLAG-human IKKe, FLAG-human TBK1 plasmids were a gift from Dr. Rongtuan Lin (McGill University, Montreal, Canada). To construct mutants for IKKe and TBK1, appropriate substitutions were introduced into the FLAG-human IKKe and FLAG-human TBK1 plasmids using the Quick Change site directed mutagenesis kit (Stratagene, La Jolla, CA). IFNb promoter-luciferase reporter was a gift from Dr. Takashi Fujita (University of Kyoto, Kyoto, Japan). ISRE-luciferase reporter was from Stratagene.

Immunoprecipitation
293T cells (3610 6 ) were transfected with a total of 3.3 mg of plasmids DNA using Lipofectamine 2000 (Invitrogen). Twelve h later, cells were lysed using lysis buffer. Lysates were centrifuged and supernatants were incubated with anti-FLAG agarose for overnight with gentle rotation at 4uC. Immune complexes were washed 4 times with lysis buffer. Samples were separated on SDS-PAGE and subjected to immunoblot analysis.

Luciferase reporter assay
293T cells were plated in 24 well plates at 3610 4 /0.5 ml and were transiently transfected with IFNb-luciferase reporter or ISRE-luciferase reporter, and control Renilla luciferase along with indicated plasmids using Lipofectamine 2000 (Invitrogen) according to the manufacturer's recommendations. Eighteen h post transfection, cells were lysed and luciferase activity was measured by using the Dual reporter luciferase assay Kit (Promega) according to the manufacturer's procedure. Renilla luciferase activity was used for normalization.

Quantitative (q) RT-PCR
Murine L cells (2610 5 ) were transfected with 2.5 mg of indicated plasmids using Lipofectamine 2000. Twenty four h later, total RNA were prepared by using Trizol reagent (Invitrogen) and were reversetranscribed with Transcriptor First Strand cDNA Synthesis Kit (Roche Diagnostic). The amount of IFNb and hypoxanthine guanine phosphoribosyltransferase (HPRT) cDNA were measured by using Universal ProbeLibrary and LightCycler 480 (Roche Diagnostic) according to the manufacture's instructions. Primers for q-RT-PCR were designed by the ProbeFinder software (Roche Diagnostic).

Confocal microscopy
293ET cells seeded on coverslips in 12-well plates were transfected with 1 mg of plasmids for FLAG-tagged wt or mutant forms of IKKe (L686S, L697S or K38A) using Lipofectamine 2000 reagent. At 20 h post transfection, cells were fixed and permeabilized with PBS containing 4% paraformaldehyde and 0.5% Triton X-100. Fixed cells were washed with PBS and incubated with antibodies against IRF3 and FLAG for 1 h at room temperature, followed by an Alexa Fluor 488 and 594-conjugated secondary antibodies for 1 h at room temperature. Nuclear DNA was stained with 49,69-diamidino-2-phenylindole (DAPI; Naca-laiTesque). Immunostained coverslips were mounted onto slides using ProlongGoldAntifade reagent (Invitrogen) and observed using an FV1000D Spectral Type confocal laser-scanning microscope (inverted microscope IX81) (Olympus, Tokyo, Japan).

C-terminal region of IKKe is required for IFNb promoter activity
We initially investigated whether the C-terminal region of IKKe is required for the induction of IFNb promoter activity. To this end, we generated a series of C-terminal truncated forms of IKKe and compared the ability to activate IFNb promoter with the corresponding fragments of TBK1 (Fig. 1A). As shown in Fig. 1B, the respective deletion of 37 and 80 residues at the TBK1 Cterminal (i.e. 1-692 and 1-649) induced IFNb promoter activity in the same manner as wt TBK1 (1-729). The further deletion of 99 residues (1-550) resulted in the loss of promoter activity, in the case of the kinase activity defective K38A mutant. These results are in good accord with the previous observation that TBK1 truncation 1-643 is still dimeric, which configuration is needed to activate TBK1 [17]. In contrast to TBK1, the C-terminal deletion mutants of IKKe failed to activate the IFNb promoter, even with a deletion of only 31 amino acids (aa) (1-685), indicating that the C-terminal region of TBK1 and IKKe plays a different role in the activation of downstream signaling, and this region in IKKe is indispensable for the activation of the IFNb promoter in contrast to TBK1. To confirm this observation, we tested whether C-terminal deletion of IKKe affects endogenous IFNb gene expression and found that wt IKKe, but not 1-685 and 1-640 mutants, induces IFNb mRNA (Fig. 1C). Given that deletion of the C-terminal region in TBK1 and IKKe yielded a different effect on downstream signaling, we continued the investigation with a detailed domain analysis of IKKe. A series of IKKe mutants lacking a range of approximately 10 to 100 residues at the C-terminal were generated and the relative induction of IFNb promoter activity was measured ( Fig. 2A). As shown in Fig. 2B, the expression of wt IKKe as well as IKKe/1-705 activated the IFNb promoter, while the promoter activity elicited by 1-685 or further truncated mutants was comparable to that of an empty vector or 1-716/K38A kinase defective mutant (Fig. 2B, top panel). These results indicate that the region between 686 and 705 contains an essential domain required for the induction of IFNb promoter activity. We tested whether these truncated forms of IKKe induce the phosphorylation of endogenous IRF3 at Ser 386 and Ser 396, which is a hallmark of IRF3 transcriptional activity. Unexpectedly, three mutants (1-685, 1-671, and 1-657) that failed to activate the IFNb promoter nevertheless induced both Ser 386 and Ser396 phosphorylation, although the phosphorylation signals were weaker than that of wt IKKe or 1-705. In contrast, 1-640 and 1-620 failed to phosphorylate IRF3, similar to the K38A mutant. These results indicate that in addition to 686-705, the 641-657 region of IKKe contains a novel functional domain involved in IRF3 phosphorylation ( Fig. 2A). We also tested whether corresponding these functional domains in TBK1 are required for phosphorylation of IRF3. As shown in Fig. 2C, TBK1/1-692 as well as TBK1/1-649 normally induced IFNb promoter activity and phosphorylation of IRF3 at Ser 386 and Ser 396 (Fig. 2C, left  panels). Thus, in contrast to IKKe, the C-terminal region of TBK1 does not contain functional domains required for both activation of IRF3 and IFNb promoter activity and is dispensable for activation of downstream signaling.
Putative helical region in IKKe contributes to the IFNb promoter activity The TBK1 and IKKe C-terminal regions are relatively conserved and certain residues are identical in the human and murine forms (Fig 3A). A TBK1 region corresponding to 686-705 of IKKe was previously proposed to form a helical structure, and L693 and K694 in TBK1 are involved in the induction of type I IFN promoter activity under physiological conditions [19]. A helical model of human IKKe and TBK1 prompted us to the observation that there are identical residues for the two kinases and that the hydrophobic residues are concentrated in a particular surface region (Fig. 3B, helix a and e). To investigate whether these residues are involved in IKKe function, IKKe and TBK1 mutants with a single aa substitution were generated ( Fig. 3A and 3B, triangles), and the IFNb promoter activity induced by these mutants was measured (Fig. 3C). Data showed that mutations in M683, L686, M690, and L697, all of which are located in a particular surface of the putative helical structure, failed to activate IFNb promoter (Fig. 3C, the filled triangles in Fig. 3A and 3B). We also confirmed that L686S and L697S mutants of IKKe failed to induce endogenous IFNb gene expression, like as kinase deficient K38A mutant (Fig. 3D). On the other hand, mutations in L693 and L704 of TBK1, corresponding to L686 and L697 of IKKe (Fig. 3A), had little, if any, effect on IFNb promoter activity (Fig. 3E). These results are consistent with the previous observation that mutation in L693 or L704 of TBK1 did not affect the induction of type I IFN promoter activity in the overexpression system [19]. We also investigated whether these IKKe mutants phosphorylate IRF3 at Ser 386 and Ser 396 (Fig. 3C, the lower panels), and found that phosphorylation of endogenous IRF3 was observed not only in wt IKKe and the mutants that induced the IFNb promoter, but also the mutants that failed to activate the IFNb promoter, except for kinase defective mutant K38A. These results indicate that phenotypes of these mutants with single aa substitution are similar to that of C-terminal deletion mutant. The putative helical region in the IKKe C-terminal is required for activation of the IFNb promoter, at least in this overexpression system, in contrast to TBK1 in which the counterpart is dispensable in the same experiment (Fig 3D) [19], demonstrating the functional difference between the C-terminal regions of TBK1 and IKKe in the type I IFN production pathway.
At least three promoter elements, including interferon-sensitive response element (ISRE), kB, and activator protein 1 (AP-1), which is the binding site of IRFs, NFkB, and AP-1 transcription factors, respectively, were identified in IFNb gene and these transcription factors orchestrate IFNb gene expression [21,22,23]. To know lack of IFNb activity in mutant IKKe was caused by abnormal activation of IRF3, we investigated whether IRF specific ISRE-reporter is activated by mutant IKKe. Again, IFNb promoter was activated by wt IKKe, but not by L689S and M697S mutants (Fig 4F, left panel). In the same experimental condition, wt IKKe activated ISRE promoter (Fig 4F, right panel). Similar to IFNb promoter, L689S and M697S mutants failed to activate ISRE. These results indicate that C-terminal mutants of IKKe could not activate IRF3 dependent transcription normally and that is probably the reason why these mutants failed to activate IFNb promoter, although transcriptional activities of NFkB and AP-1 are also abnormal.
Given that C-terminal mutants which failed to induce IFNb promoter activity, still have an ability to phosphorylate IRF3, we next investigated whether IRF3 activated by C-terminal mutants forms of IKKe were normally translocated to nucleus. Wt and mutant IKKe were transfected to 293ET cells and endogenous IRF3 in nucleus was observed by using confocal microscopy. As shown in Fig. 4, nuclear accumulation of IRF3 in L683S and L697S mutant expressing cells was observed, although cells in which IRF3 is accumulated in nucleus were extensively detected in wt IKKe expressing cells than that of these mutants. In this experimental condition, accumulation of IRF3 in nucleus was not observed in K38A mutant expressing cells at all. These results suggest that C-terminal region of IKKe is not essentially required for, at least, phosphorylation and nuclear translocation of IRF3 and therefore, lack of ability to induce IFNb promoter activity may be caused by the deficiency of nuclear events following IRF3 nuclear translocation step.

C-terminal region of IKKe is involved in the solid dimer formation
Mutations of TKB1 at the contact residues of dimerization led to a decrease of the auto-phosphorylation that enhances kinase activity, and thereby to a failure of the Ser 396 phosphorylation of IRF3 and ultimately of IFNb gene expression [18]. We subsequently investigated whether a mutation in the C-terminal region of IKKe would have an effect on its dimer formation. V5tagged wt and FLAG-tagged IKKe mutants were co-transfected into cells and dimer formation was assessed by immunoprecipitation (Fig. 5). As shown in Fig.5A, full length IKKe (1-716) was co-precipitated with full length or 1-705, while the association was greatly reduced in 1-685 or further deletion mutants. In the case of the substituted mutants, both the wt IKKe and the mutants inducing IFNb promoter activity were strongly associated with wt IKKe, while the four mutants that failed to activate the IFNb promoter were at almost undetectable levels (Fig 5B). These results suggest that the 686-705 region of IKKe includes the contact residues for dimer formation.

Discussion
We have shown here that the C-terminal region of IKKe is involved in the induction of IFNb promoter activity, while the corresponding region of TBK1 is not required. Mutant IKKe with a single aa substitution at C-terminal resulted in greatly decreased dimer formation and concomitantly lost the ability to induce IFNb promoter activity, indicating dimerization of IKKe via C-terminal region is required for type I IFN production. In agreement with our observations of IKKe, mutations preventing dimerization of TBK1 resulted in a failure of type I IFN production [17,18]. However, TBK1 forms a dimer without the C-terminal region, since the contact residues directly involved in TBK1 dimer formation are not located in the C-terminal, but rather, widely expanded to the KD, ULD and SSD [17,18]. Like as IKKe, the critical dimer contact of IKKbis located in the C-terminal, and mutation of the critical C-terminal contact residues, these residues are not conserved in TBK1 and IKKe, resulted in a failure of dimerization of IKKb and the loss of ability to phosphorylate its substrate [16]. Thus, even though IKKe is more closely related to TBK1 in terms of their aa sequences or substrates, the dimer organization of IKKe is apparently different from TBK1, and closer to that of IKKb. It will be of interest to study the IKKe structure to elucidate the dimer organization and to compare it with TBK1 and IKKb, especially with regard to the C-terminal. Phospho-IRF3 signals were detected in cells transfected with various IKKe mutants lacking the ability to activate the IFNb promoter. These results are quite unexpected, since it is well established that phosphorylated IRF3 is transcriptionally active and that any stimulation which induces the phosphorylation of IRF3 at Ser 386 and Ser 396 leads to an increase in IFNb promoter activity [10]. Notably, the phospho-IRF3 signals in mutant IKKe expressing cells without any increased IFNb promoter activity were clearly weaker than in the IFNb promoter activated cells. Therefore, it is possible that the kinase activity of   these mutants is insufficient to increase IFNb promoter activity, although the kinase activity of these mutants is increased compared to the K38A mutant. Consistent with this idea, nuclear translocation of IRF3in cells expressing C-terminal mutated forms of IKKe was decreased compared to wt IKKe expressing cells (Fig. 4). Another possibility in loss of type I IFNb promoter activity of IKKe C-terminal mutants is that certain Ser or Thr residues in IRF3, in addition to those in the C-terminal Ser/Thr rich region, including Ser 386 and Ser 396, may have to be phosphorylated for transcriptional activation, and these residues are not phosphorylated by these IKKe mutants [24]. Alternatively, some nuclear factors other than IRF3, which are also required for activation of IFNb promoter activity, may not be activated by mutant IKKe and dimerization of IKKe is critically required for phosphorylation of the nuclear factors. In any, dimerization deficient IKKe still phosphorylates IRF3, indicating that dimerization is not essential in IKKe kinase activity for IRF3.