Mechanism of Endogenous Regulation of the Type I Interferon Response by Suppressor of IκB Kinase ϵ (SIKE), a Novel Substrate of TANK-binding Kinase 1 (TBK1)*

Background: Suppressor of IκB kinase ϵ (SIKE) inhibits a key innate immune effector molecule, TANK-binding kinase 1 (TBK1), through an undefined mechanism. Results: SIKE is a TBK1 substrate. Conclusion: SIKE controls TBK1 activity by acting as a high affinity substrate. Significance: SIKE attenuates phosphorylation of interferon regulatory factor 3 (IRF3) by serving as an alternative, high affinity substrate for TBK1. TANK-binding kinase 1 (TBK1) serves as a key convergence point in multiple innate immune signaling pathways. In response to receptor-mediated pathogen detection, TBK1 phosphorylation promotes production of pro-inflammatory cytokines and type I interferons. Increasingly, TBK1 dysregulation has been linked to autoimmune disorders and cancers, heightening the need to understand the regulatory controls of TBK1 activity. Here, we describe the mechanism by which suppressor of IKKϵ (SIKE) inhibits TBK1-mediated phosphorylation of interferon regulatory factor 3 (IRF3), which is essential to type I interferon production. Kinetic analyses showed that SIKE not only inhibits IRF3 phosphorylation but is also a high affinity TBK1 substrate. With respect to IRF3 phosphorylation, SIKE functioned as a mixed-type inhibitor (Ki, app = 350 nm) rather than, given its status as a TBK1 substrate, as a competitive inhibitor. TBK1 phosphorylation of IRF3 and SIKE displayed negative cooperativity. Both substrates shared a similar Km value at low substrate concentrations (∼50 nm) but deviated >8-fold at higher substrate concentrations (IRF3 = 3.5 μm; SIKE = 0.4 μm). TBK1-SIKE interactions were modulated by SIKE phosphorylation, clustered in the C-terminal portion of SIKE (Ser-133, -185, -187, -188, -190, and -198). These sites exhibited striking homology to the phosphorylation motif of IRF3. Mutagenic probing revealed that phosphorylation of Ser-185 controlled TBK1-SIKE interactions. Taken together, our studies demonstrate for the first time that SIKE functions as a TBK1 substrate and inhibits TBK1-mediated IRF3 phosphorylation by forming a high affinity TBK1-SIKE complex. These findings provide key insights into the endogenous control of a critical catalytic hub that is achieved not by direct repression of activity but by redirection of catalysis through substrate affinity.

TANK-binding kinase 1 (TBK1) serves as a key convergence point in multiple innate immune signaling pathways. In response to receptor-mediated pathogen detection, TBK1 phosphorylation promotes production of pro-inflammatory cytokines and type I interferons. Increasingly, TBK1 dysregulation has been linked to autoimmune disorders and cancers, heightening the need to understand the regulatory controls of TBK1 activity. Here, we describe the mechanism by which suppressor of IKK⑀ (SIKE) inhibits TBK1-mediated phosphorylation of interferon regulatory factor 3 (IRF3), which is essential to type I interferon production. Kinetic analyses showed that SIKE not only inhibits IRF3 phosphorylation but is also a high affinity TBK1 substrate. With respect to IRF3 phosphorylation, SIKE functioned as a mixed-type inhibitor (K i, app ‫؍‬ 350 nM) rather than, given its status as a TBK1 substrate, as a competitive inhibitor. TBK1 phosphorylation of IRF3 and SIKE displayed negative cooperativity. Both substrates shared a similar K m value at low substrate concentrations (ϳ50 nM) but deviated >8-fold at higher substrate concentrations (IRF3 ‫؍‬ 3.5 M; SIKE ‫؍‬ 0.4 M). TBK1-SIKE interactions were modulated by SIKE phosphorylation, clustered in the C-terminal portion of SIKE (Ser-133, -185, -187, -188, -190, and -198). These sites exhibited striking homology to the phosphorylation motif of IRF3. Mutagenic probing revealed that phosphorylation of Ser-185 controlled TBK1-SIKE interactions. Taken together, our studies demonstrate for the first time that SIKE functions as a TBK1 substrate and inhibits TBK1-mediated IRF3 phosphorylation by forming a high affinity TBK1-SIKE complex. These findings provide key insights into the endogenous control of a critical catalytic hub that is achieved not by direct repression of activity but by redirection of catalysis through substrate affinity.
To become active, TBK1 requires phosphorylation at Ser-172 within the kinase domain activation loop (28). IKK␤ (29) and TBK1, via a trans-phosphorylation event (30), provide this modification. Phosphorylated Ser-172 induces reorganization of the activation loop, making the active site competent to bind substrate (30). Alignment of TBK1 substrates revealed a weak consensus sequence for TBK1 phosphorylation consisting of a hydrophobic residue immediately proceeding the modified serine that is mirrored by several hydrophobic residues lining the P ϩ 1 site (30). Kinetic analysis of TBK1 phosphorylation of Ser-36 within the IB␣ peptide (residues 19 -41) revealed that TBK1 functions through a rapid equilibrium, random order mechanism (31), consistent in all IB kinases (31)(32)(33).
To regulate TBK1 activity, endogenous inhibitors that alter TBK1 ubiquitination (A20 (34)) or mask tyrosine phosphorylation sites (SHP-2 (35)) have been characterized. In addition, a yeast two-hybrid screen for interaction partners of the closely related, noncanonical IB kinase, ⌱⌲⌲⑀, identified a 207-amino acid protein dubbed suppressor of IKK⑀ (SIKE) (36). When examined in the context of TLR3 signaling, SIKE associated with IKK⑀ and TBK1 only prior to dsRNA stimulation or viral infection (36), but SIKE mRNA was increased after viral challenge suggesting SIKE formed a negative feedback inhibition loop (37). Moreover, SIKE was shown to disrupt IKK⑀-TBK1 interactions with the upstream signaling component, TIR domain-containing adaptor-inducing interferon-␤, and the downstream substrate IRF3 via an undefined mechanism (36).
Because SIKE acts as a physiological suppressor of TBK1 at a key convergence point in the innate anti-viral response, we investigated the mechanism by which SIKE inhibits the type I interferon response and the regulatory release mechanism of SIKE from TBK1 upon viral activation of TBK1. Our studies show that SIKE acts as a mixed-type inhibitor of TBK1-mediated phosphorylation of IRF3. Surprisingly, TBK1 directly phosphorylated SIKE at six sites that partially mimic IRF3's phosphorylation motif sequence. Analysis of the initial rate data revealed potential negative cooperativity in IRF3 and SIKE reactions. Using site-directed mutagenesis and cell-based assays, we showed that SIKE phosphorylation correlates with activation of the anti-viral response and modulates the TBK1-SIKE interaction. From these findings, we propose that SIKE is not only an inhibitor of TBK1-mediated phosphorylation of IRF3 but is also a TBK1 substrate.

EXPERIMENTAL PROCEDURES
Reagents-Chemicals were purchased from Sigma unless otherwise specified. [␥-32 P]ATP was purchased from Perkin-Elmer Life Sciences. Recombinant TBK1 was purchased from Invitrogen. Anti-phosphoserine antibody was purchased from Qiagen. Anti-FLAG-HRP was purchased from Sigma. Anti-HA antibody was purchased from Covance. Anti-full-length IRF3 and anti-actin-HRP antibodies were purchased from Santa Cruz Biotechnology. Anti-phospho-Ser-396 IRF3 antibody was purchased from Cell Signaling. Anti-rabbit-IgG-HRP and antimouse-IgG-HRP antibodies were purchased from Southern Biotechnology Associates, Inc.
Cell Culture-HEK293 cells were purchased from American Type Culture Collection. Cells were cultured in complete medium (RPMI 1640 supplemented with 10% low endotoxin FBS, 20 mM L-glutamine, 100 mM HEPES, 10 mM sodium pyruvate, and 1ϫ nonessential amino acid solution) at 37°C in 5% CO 2 .
Expression and Purification of Proteins-For pET15b SIKE72 construct expression, the vector was transformed into chemically competent BL21-CodonPlus (DE3)-RIPL following the manufacturer's protocol (Agilent Technologies). A single colony was used to inoculate an overnight culture of Luria broth plus 100 g/ml ampicillin. The overnight culture was subcultured 1:100 into two 1-liter LB/ampicillin flasks grown at 37°C until the cell density reached an A 600 Ϸ0.6. Cultures were induced with 1 mM isopropyl ␤-D-galactopyranoside and allowed to grow for an additional 4 h at 37°C. Cells were harvested by centrifugation at 7,000 ϫ g, and the pellet was solubilized in guanidine hydrochloride (GdnHCl) buffer (6 M guanidine hydrochloride, 100 mM sodium phosphate, pH 8.0, 500 mM NaCl, and 1 mM 2-mercaptoethanol), 5 ml of buffer/g of cell pellet. Lysate was clarified by centrifugation at 14,000 ϫ g. The supernatant was mixed with 5 ml of nickel-nitrilotriacetic acidagarose resin (Qiagen), pre-equilibrated in GdnHCl buffer. The lysate/resin mixture was loaded into a column (Bio-Rad) and washed by gravity with 50 column volumes (CV) of GdnHCl buffer. Bound protein was refolded on the column using a 40-CV reverse gradient of GdnHCl buffer to 100 mM sodium phosphate, pH 8.0, 500 mM NaCl, and 1 mM 2-mercaptoethanol (Buffer 1) and eluted with 5 CV of Buffer 1 plus 500 mM imidazole. Elution fractions were separated on a HiLoad 16/60 Superdex 200 column (GE Healthcare), and peak fractions were screened for target protein by SDS-PAGE analysis. CD and thermal denaturation of WT and mutant SIKE72 proteins indicated primarily helical secondary structure consistent with a predicted coiled coil domain structure and a Ն2-state unfolding curve typical of globular, oligomeric proteins (data not shown).
In For the ATP-varied reactions, excess cold ATP from pre-activation of TBK1 was removed by desalting column (Zeba spin, 7 molecular weight cutoff, 2 ml, Pierce) prior to initiating reactions. Reactions were incubated for 120 min at 30°C (determined in supplemental Fig. S1B) and terminated by boiling in reducing SDS sample buffer (20 l, Invitrogen). Duplicate samples (10 l each) were separated by SDS-PAGE on a 4 -12% Tris-BisTris polyacrylamide gel (Invitrogen). Gels were mounted on filter paper and exposed to a storage phosphor screen (Imaging Screen K, Bio-Rad). The screen was read by a Molecular Imager FX system (Bio-Rad). Band densitometry was measured and converted to 32 P micromolars via a [␥-32 P]ATP standard curve. Inhibition experiments were carried out using the assay methodology described for IRF3-varied and the ATP-varied assay with the addition of SIKE72 at 20.8 or 83.3 nM final concentration. All reactions were analyzed by SDS-PAGE/autoradiography as described above. A minimum of quadruplicate repeats was completed for each reaction. Data were plotted as Michaelis-Menten or double-reciprocal plots and fit to single rectangular hyperbola or linear polynomial equations, respectively, in SigmaPlot (Systat Software, Inc.). Errors were reported as standard deviation. From the intercepts and slopes derived from the double-reciprocal plot fit to a linear polynomial, inhibition constants were calculated for the SIKE ϩ ATP or IRF3 varied data using Equations 1 and 2, To examine inhibition of IRF3 phosphorylation by SIKE72, 4.17 M IRF3(173-427), 100 M ATP, 0.1 mCi of [␥-32 P]ATP, and SIKE72 or SIKE72 mutants at indicated concentrations were prepared in assay buffer (50 l volume). TBK1, 10 l of 29.6 nM and 4.93 nM final concentration, pre-activated as above, was added to initiate the reaction. Reactions were incubated at 30°C for 120 min and analyzed as described above to derive initial rates for IRF3(173-427) phosphorylation. Data were plotted as the percentage of uninhibited reaction rate versus SIKE72 concentration. Errors were reported as standard deviation. To derive the K i, app parameter for SIKE mutants, inhibited reaction rate versus SIKE72 concentration was fit to a twoparameter rectangular hyperbola.
Phosphopeptide Mapping-Recombinantly expressed SIKE72 (500 ng) was incubated with 4.93 nM TBK1 and 100 M ATP in assay buffer (100 l volume) for 24 h at 30°C. The reaction was terminated by diluting the reaction 1:1 with 6 M GdnHCl buffer. SIKE72 was isolated from the reaction by incubation with 40 l of nickel-nitrilotriacetic acid resin, washed with GdnHCl buffer, and bound SIKE eluted with 50 l of GdnHCl buffer plus 0.5 M imidazole. Eluent was separated by SDS-PAGE (10% Tris/ glycine), and the target band was excised for analysis. The gel slice was washed and destained in 200 l of 50% methanol overnight. The gel pieces were dehydrated in acetonitrile, rehydrated in 30 l of 10 mM dithiothreitol in 0.1 M ammonium bicarbonate, and reduced at room temperature for 0.5 h. The DTT solution was removed, and the samples were alkylated in 30 l of 50 mM iodoacetamide in 0.1 M ammonium bicarbonate at room temperature for 0.5 h. The reagent was removed, and the gel pieces were dehydrated in 100 l of acetonitrile. The acetonitrile was removed, and the gel pieces were rehydrated in 100 l of 0.1 M ammonium bicarbonate. The pieces were dehydrated in 100 l of acetonitrile, and the acetonitrile was removed, and the pieces were completely dried by vacuum cen-trifugation. The gel pieces were rehydrated in 1 g of trypsin in 50 mM ammonium bicarbonate on ice for 10 min. Any excess trypsin solution was removed, and 20 l of 50 mM ammonium bicarbonate was added. The samples were digested overnight at 37°C, and the peptides formed were extracted from the polyacrylamide in two 30-l aliquots of 50% acetonitrile, 5% formic acid. Extracts, evaporated to 15 l, were separated by C18 reversed-phase capillary column (Waters NanoAcquity) coupled with nanospray tandem mass spectrometry (LTQ-Orbitrap hybrid, Thermo Electron). The digest was analyzed using the double play capability of the instrument acquiring full scan mass spectra to determine peptide molecular weights and product ion spectra to determine amino acid sequence in sequential scans. The data were analyzed by database searching using the Sequest search algorithm against custom databases generated from the SIKE72 sequence. All potential phosphopeptides were manually examined for correct identification of the modified site.

SIKE Acts as a Mixed Type Inhibitor of TBK1-mediated
Phosphorylation of IRF3-SIKE was originally classified as a physiological suppressor of TBK1 (36). Our studies into the mechanism of SIKE's inhibitory activity focused on the TBK1mediated phosphorylation of IRF3. The IRF3 construct used in these studies included residues 173-427, encompassing the IRF activation domain and TBK1 phosphorylation sites (385, 386, 396, 398, 402, 404, and 405). The Michaelis-Menten plot of IRF3 as the varied substrate gave initial kinetic parameters for K m (Michaelis constant) and V max (maximum velocity) as 2.69 M and 10.5 nM/min reaction, respectively ( Fig. 1A and Table   Table  1). To examine the effect of SIKE72 on IRF3 phosphorylation, SIKE72 was added at ϳ20 and ϳ80 nM to the IR-varied or ATPvaried (Fig. 1, C and D) assays. The primary effect of SIKE72 for IRF3-varied reactions occurred on V max (ϳ1.8-fold decrease), indicative of a noncompetitive inhibitor (Table 1). Interestingly, the K m and V max values calculated from double-reciprocal plots of the ATP-varied reactions indicated a 2.2-3-fold change in both K m and V max values consistent with a mixedtype inhibitor ( Table 1). The K i, app values were calculated from the double-reciprocal plot slope and intercept parameters using Equations 1 and 2, as described under "Experimental Procedures" ( Table 2).
TBK1 Phosphorylates SIKE on Six Serine Residues That Mimic the IRF3 Phosphorylation Motif-To determine the number of SIKE phosphorylation sites, we assessed TBK1-mediated phosphorylation by MALDI-TOF mass spectrometry. Over a 96-h time course, we observed a mixture of singly to triply phosphorylated SIKE72 (supplemental Fig. S2A). Under the same conditions, a single to 7 out of 7 known IRF3 phosphorylation sites were modified, whereas glyoxosomal malate dehydrogenase, not known to be phosphorylated but containing 16 preferred Ser/Thr motif sites out of 36 total Ser/Thr residues, remained unmodified (supplemental Fig. S2, B and C).
SIKE Phosphorylation Status Alters Its Inhibitory Activity-To determine the relationship between the multiple SIKE phosphorylation sites and SIKE-mediated inhibition of IRF3 phosphorylation by TBK1, we probed these sites via truncated mutants or site-directed mutagenesis and assessed the effect of these SIKE mutations on TBK1-mediated IRF3 phosphorylation. The broad effect of SIKE phosphorylation was analyzed by the phosphomimetic mutant, Ser to Glu mutation, at the six  Fig. 1, C and D) to a linear polynomial. f IRF3 construct included residues 173-427 and parameters derived from data at listed concentrations.  identified sites (S6E), and phospho-knock-out mutant, Ser to Ala mutations, at the six identified sites (S6A). Subsets of phospho-knock-out mutants were created to further probe sites essential for SIKE's inhibitory activity: S4A (S185A, S187A, S188A, and S190A), S2A (S133A and S198A), and S185A. SIKE's phosphorylation sites cluster in the C-terminal portion of SIKE. To examine the role of the N-terminal sequence in SIKE inhibition, SIKE truncation mutants containing one (SIKE(72-184)) or retaining all phosphorylation sites (SIKE(113-207)) were constructed. The panel of constructs is summarized in Fig. 4A. Inhibition curves for the phosphomimetic and phosphoknock-out mutants showed that phosphorylation of SIKE72 reduced SIKE's ability to inhibit TBK1-mediated IRF3 phosphorylation, whereas unmodified SIKE72 had greatly enhanced inhibitory activity (supplemental Fig. S4A). Using four concentrations that defined SIKE's inhibitory effect, kinase assays were completed with the full panel of mutants (Fig. 4B). Apparent inhibition constants (K i, app ) for each mutant were derived ( Table 3). The K i, app for the phosphomimetic mutant increased ϳ3-fold over WT SIKE72, suggesting that phosphorylation may reduce the affinity between TBK1-SIKE and/or enhance release of SIKE from TBK1. The K i, app for the C-terminal 113-207 construct, retaining all phosphorylation sites, was similar to WT SIKE72. Surprisingly, the K i, app for the N-terminal 72-184 construct was 196 nM, even though maximal inhibition was onlyϳ25%ofthereaction.Lossofthetwoperipheralphosphorylation sites, Ser-133 and Ser-198, in the S2A construct also did not alter the K i, app value, whereas phospho-knock-out of the four clustered serines, Ser-185, Ser-187, Ser-188, and Ser-190, reduced the K i, app parameter by 2.5-fold. Within this cluster, the K i, app for the point mutation S185A was similar to the S6A mutant (Table 3).
SIKE Is a TBK1 Substrate-To establish kinetic parameters for SIKE as a TBK1 substrate, TBK1 assays with saturating ATP (100 M) and SIKE72 varied from 0.043 to 8.3 M were completed. The K m and V max parameters were 0.41 M and 7.4 nM/min reaction, respectively ( Fig. 6A and Table 1). Strikingly, the double-reciprocal plot revealed a downward concave curve, similar to the double-reciprocal plot of IRF3-varied, TBK1 assays (Fig. 6, B and C, and supplemental Fig. S1C). . TBK1 phosphorylates SIKE on six serine residues that mimic the IRF3 phosphorylation sequence. A, mass spectrum resulting from analysis of indicated phosphopeptide sequence. Two separate ion series were recorded simultaneously, the b-and y-ion series, which represent sequencing inward from the N and C termini, respectively. Lowercase s indicates phosphorylated residue. B, summary of phosphopeptide sequences identified using a Sequest search algorithm against custom databases generated from the His 6 -SIKE72 sequence. ? indicates that a phosphorylation site exists in peptide but cannot be uniquely assigned to a single serine residue. C, sequence alignment of known TBK1-mediated phosphorylation sites on IRF3, required for dimerization and activation, with TBK1-mediated phosphorylation sites on SIKE found through phosphopeptide mapping. Gray letters, phosphorylated residues; *, conserved site; :, functionally conserved site.

DISCUSSION
Multiple innate immune receptors recognize pathogen-derived ligands and initiate an immediate inflammatory response (38,39). Several of these signaling pathways converge to activate TANK-binding kinase 1 (TBK1) (9, 22, 24, 40, 41). Although a central player in the innate immune system's defenses that shapes the downstream innate immune response, mechanisms by which the host controls TBK1 activity are not well understood. To address this issue, we sought to define mechanistically how an endogenous inhibitor of TBK1, SIKE, blocks the TBK1-mediated anti-viral response.
The major discovery of this work is that SIKE is not only an endogenous inhibitor of TBK1-mediated phosphorylation of IRF3 but is also a TBK1 substrate. SIKE has a 6.5-fold lower K m value (0.41 M), as compared with IRF3, with a comparable V max value suggesting that changes in SIKE concentration could effectively manipulate TBK1 function through altered substrate selection. Using kinetic analyses and protein interactions assays, we characterized how this novel TBK1 substrate effectively inhibits type I interferon production.
TBK1 Displays Negative Cooperativity for Substrates IRF3 and SIKE-In our studies, we examined the effect of SIKE on TBK1-mediated IRF3 phosphorylation using a macromolecular IRF3 substrate (residues 173-427). Previous studies on the TBK1 kinetic mechanism, using an inhibitor of B␣ peptide as substrate, showed that the enzyme functioned via a RERO  mechanism (31). Our data obtained with macromolecular substrates are consistent with this mechanism but reveal evidence suggesting the existence of negative cooperativity in substrate binding or the presence of a regulatory site for protein substrate. The equilibria for this mechanism with respect to IRF3 phosphorylation are given in Fig. 7A, and are described by Dalziel's generalized rate Equation 3 (42), where e is TBK1 concentration in the reaction; V o is the initial velocity of the reaction, and ⌽ are Dalziel coefficients that, for an RERO mechanism, pertain to the concentrations of specific complexes in the reaction mechanism (⌽ 1 ϭ TBK1-IRF3, ⌽ 2 ϭ TBK1-ATP, ⌽ 12 ϭ TBK1, and ⌽ 0 ϭ TBK1-ATP-IRF3), labeled in Fig. 7A. We first determined the base-line kinetic parameters for IRF3 and ATP as summarized in Table 1. The double-reciprocal plot of IRF3-varied substrate was not linear but a downward concave curve (Fig. 6C). This type of plot is indicative of one of three phenomena as follows: (i) substrate activation; (ii) nonidentical active sites, or (iii) negative cooperativity. In examining these three explanations, we assumed from both the related IKK␤ (43) and TBK1 (30,44,45) crystallographic structure and our own size exclusion chromatography (data not shown) that TBK1 functions as a dimeric species (required for explanations ii and iii above). Substrate activation entails a substrate-binding site (with lower affinity than the active site) that when bound leads to activation of the reaction. For nonidentical active sites, the dimeric TBK1 would have a high and low affinity active site for substrate a priori. Finally, negative cooperativity between TBK1 active sites would occur if binding of substrate at the first active site lowered the affinity for substrate at the second active site by allosteric interactions between the subunits. Alone, the IRF3 varied assay data did not allow us to discern which mechanism is at play. Luckily, SIKE also functioned as a macromolecular TBK1 substrate. The double-reciprocal plot of the SIKE-varied assays also exhibited this characteristic downward concave curvature (Fig. 6B). As SIKE and IRF3 share no sequence homology and are distinct macromolecular substrates, the substrate activation or nonidentical active site mechanism seemed improbable as these macromolecular substrates must be similarly recognized by the kinase at a site independent of the active site or with equally reduced affinity at a nonidentical active site. Rather, negative cooperativity induced by recognition of IRF3's and SIKE's shared phosphorylation motif could explain the nonlinear double-reciprocal plots. Negative cooperativity requires communication between active sites. Interestingly, in IKK␤ the two active sites appear to be independent of one another with dimerization mediated by the scaffold dimerization domain. In contrast, the TBK1 structure revealed three additional dimerization interfaces between the kinase and ubiquitin-like domains of one subunit to the scaffold dimerization domain of the contralateral subunit (Fig. 7B). These additional subunit interfaces identified in the TBK1 structures may provide a means of communicating substrate binding between active sites. Structural study of a kinase-substrate complex to complement the reported kinase-ATP analog inhibitor structures (30,44,45) would directly address this hypothesis.
In a RERO mechanism, the K m value of IRF3 is the K d value for IRF3 binding to the E-ATP complex. From the two linear regions of the IRF3 double-reciprocal plot, we calculated the K m value at low (0.042-0.42) versus high (0.42-20.8) substrate concentrations as 50 nM and 3.5 M, respectively. Similarly, the two linear regions of SIKE's double-reciprocal plot were 55.9 and 395 nM at low (42-208 nM) and high (208 -8,000 nM) substrate concentrations, respectively. These data suggest that binding of substrate, either IRF3 or SIKE, to the first active site occurs with near equal affinity. The ϳ10-fold difference in K m values at high substrate concentration favoring the TBK1-SIKE complex would suggest that SIKE functions primarily as a competitive inhibitor of IRF3 phosphorylation.
SIKE's Inhibitory Mechanism-For a simple RERO mechanism, an alternative substrate (as we have shown SIKE to be) should give competitive inhibition with respect to IRF3 and mixed (or noncompetitive) inhibition with respect to ATP (46). Analysis of SIKE's inhibitory effects on TBK1-mediated phosphorylation of IRF3 revealed a mixed-type inhibitor ( Fig. 1 and Table 1) with respect to ATP with ϳ2-3-fold decreases in K m and V max values ( Fig. 1D and Table 1), as expected for a RERO mechanism. Surprisingly, SIKE inhibition in IRF3-varied experiments did not exhibit a competitive inhibition pattern but rather displayed primarily as a noncompetitive inhibitor, reducing V max by 3-fold ( Fig. 1C and Table 1). To further interpret these results, we examined SIKE inhibition of the TBK1 reaction in terms of the generalized linear rate equation for a RERO mechanism developed by Dalziel (42). In the RERO mechanism, SIKE could in theory interact with any TBK1 complex and would affect the appropriate ⌽ parameter. The rearranged generalized linear rate Equation 4 with ATP as the varied substrate is as follows: In  sites rather than the existence of an IRF3/SIKE-binding site separate from the active sites in the dimer.
Effect of SIKE Phosphorylation on Inhibitory Function-Four of the six phosphorylated serines of SIKE align well with the phosphorylation sites of IRF3 (Fig. 3C) (15,16). The strongest homology was evident at the SIKE sequence 184 LSISSE 189 composed of a cluster of three phosphorylation sites. When TBK1 substrates are compared, the phosphorylation pattern of TBK1 substrates has three forms as follows: singly as in Akt (22), insulin receptor (21), and optineurin (23), multiple sites but not clustered as in DDX3 (47), or multiple clustered sites as in IRF3 (15,16), IRF7 (48,49) and, as we show here, SIKE. In IRF3/7, these clustered phosphorylation sites, when modified, alter protein function, activating the transcription factor (17)(18)(19)(20). We explored the role of these clustered SIKE phosphorylation sites in the context of inhibition of TBK1-mediated IRF3 phosphorylation. When replaced with the phosphomimetic mutations, Ser to Glu, SIKE's inhibitory activity (K i, app ϳ900 nM versus WT K i, app ϭ 350 nM) and interaction with TBK1 are greatly diminished (Figs. 4B and 5A). As the introduced negative charge grossly mimics SIKE's product form, these results are consistent with product release from an enzyme. Initial studies of SIKE held that SIKE maintained TBK1 in an inactive state prior to TBK1 activation at which point SIKE was released from the TBK1 complex. Here, we attribute the TBK1-SIKE interaction to SIKE's high affinity for the enzyme and SIKE's release to its post-translational modifications.
Similar to IRF3 where individual phosphorylation sites have been attributed to specific function, we probed the six SIKE phosphorylation sites to determine whether a subset of residues was essential to TBK1's preferential binding to SIKE over IRF3. With a complete phospho-knock-out mutant (Ser to Ala), SIKE's inhibitory activity increased 4.5-fold, consistent with retained TBK1 recognition. The S2A versus S4A mutants further defined the cluster of phosphorylation sites centered at the 185 SXSSXS 190 motif as essential to phosphorylation-dependent release of SIKE from TBK1, whereas the phosphorylation state of the S2A sites did not contribute to SIKE's release from TBK1. The single mutation, S185A, alone could enhance inhibition of TBK1-mediated IRF3 phosphorylation. In fact, the K i, app values for the S6A (77 nM) and the single S185A (41 nM) mutants are in line with the Michaelis constant at low SIKE concentrations (55.9 nM, Fig. 6B). Additional interactions between TBK1 and SIKE or the structural context of the phosphorylation motif may also contribute to the high affinity TBK1-SIKE interaction. This is supported by the ability of SIKE(72-184) to inhibit TBK1 function and SIKE  to form an interaction with TBK1 (Figs. 4B and 5A). SIKE(72-184) retains a single phosphorylation site (Ser-133) that when mutated to Ala (S2A mutant) does not alter SIKE inhibitory activity. Therefore, inhibition by this mutant is attributed to not only Ser-133 docking to TBK1's active site, but also interactions between SIKE and TBK1 outside of the active site. Similarly, SIKE(1-112), as it contains no phosphorylation sites, must form a TBK1-SIKE interaction exclusive of the active site. These interactions independent of the active site are consistent with the noncompetitive/mixed type inhibition mediated by SIKE.
SIKE's Role in Innate Immunity-SIKE's original characterization as an endogenous inhibitor of the TBK1-mediated antiviral response only partially explains the SIKE function. Comparison of the pseudo-K d value for the E-ATP complex showed that SIKE is preferred as a TBK1 substrate 6.5-fold over IRF3. At equivalent concentrations, SIKE would appear to inhibit IRF3 phosphorylation. This would not be the first time a TBK1 substrate had initially been defined as an inhibitor. Optineurin, an autophagy receptor, was classified as an endogenous TBK1 inhibitor (50). Subsequent studies defined optineurin as a TBK1 substrate (23,24). Phospho-optineurin has increased affinity for the microtubule-associated protein light chain 3, thereby promoting selective autophagy of ubiquitin-coated cytosolic bacteria and providing a mechanism by which TBK1 functions in maintaining xenophagosomes (24). If not just an endogenous regulator of TBK1-mediated phosphorylation of IRF3, what is the primary function of SIKE? Two pieces of data suggest a SIKE function. First, a related protein, fibroblast growth factor receptor 1 oncoprotein partner 2 (51), shares 56% sequence identity with SIKE but is expressed from chromosome 12 instead of chromosome 1 and lacks all six TBK1 phosphorylation sites. Fibroblast growth factor receptor 1 oncoprotein partner 2 (FGFR1OP2) associates with the actin cytoskeleton and has been linked to wound healing pathways (52). Second, we have shown by fluorescence microscopy that SIKE fused to GFP localizes to the cytosol with accumulation at actin/cytoskeleton-like structures at the cell periphery and within the lamellipodia (data not shown). We hypothesize that SIKE functions in cytoskeletal rearrangement in response to pathogen challenge through its phosphorylation by TBK1. Current studies in the laboratory are focused on testing this hypothesis.
In summary, we have revealed through kinetic analysis that SIKE is a novel, high affinity TBK1 substrate. TBK1-SIKE interactions are modulated by SIKE phosphorylation but are not limited to the typical kinase-substrate interactions within the active site. TBK1 phosphorylation of increasing IRF3 or SIKE concentrations displays negative cooperativity. The mechanism by which substrate binding is allosterically conveyed between subunits remains to be determined. With respect to IRF3 phosphorylation, SIKE effectively functioned as a mixedtype inhibitor of TBK1-mediated IRF3 phosphorylation rather than, given its status as a TBK1 substrate, as a competitive inhibitor. Together, these studies suggest that endogenous control of a critical catalytic hub is not only achieved by direct repression of activity but also by redirection of catalysis through substrate affinity. Importantly, these studies identify a novel TBK1 substrate modified in response to pathogen challenge. Further investigation into SIKE's function in the innate immune response will uncover unique pathways and provide insight into novel mechanisms of host immune defenses.