j ourna l homepage: www.e lsev ie r.com/ locate /ce l l s ig
Regulation of IRAK-1 activation by its C-terminal domain
Thao Nguyen, Dominic De Nardo, Paul Masendycz, John A. Hamilton, Glen M. Scholz ⁎Department of Medicine and Cooperative Research Centre for Chronic Inflammatory Diseases, The University of Melbourne, Royal Melbourne Hospital, Parkville, Victoria 3050, Australia
mediators of the immune response to infection by virtue of their ability to secretecytokines that trigger inflammation. Toll-like receptors (TLRs) are largely responsible formeditating the activationof macrophages by pathogens. IRAK-1 is a proximal protein kinase in TLR signalling pathways and hence itsactivationmust be tightly regulated. However, the mechanisms which control the activation of IRAK-1 are poorlyunderstood. IRAK-1 contains a death domain at its N-terminus that mediates its interaction with other deathdomain containing proteins, a central Ser/Thr kinase domain, and a C-terminal domain that contains bindingmotifs for TRAF6. We show here that deletion of the death domain or the majority of the C-terminal domainmarkedly enhanced the capacity of IRAK-1 to activate NF-κB in a TLR-independent manner in RAW 264.7macrophages. Furthermore, theC-terminal truncationmutant spontaneouslyoligomerisedand formed complexeswith thenegative regulator IRAK-M in the absence of TLR activation. In contrast to thebindingof IRAK-M to IRAK-1,the death domainof IRAK-1wasnot required for the interactionof IRAK-4with IRAK-1. On the basis of these resultswe propose a model inwhich IRAK-1 is held in a closed, inactive conformation via an intramolecular mechanisminvolving its C-terminal domain and possibly the death domain. Phosphorylation of IRAK-1 by IRAK-4 in responseto TLR activation may then release IRAK-1 from the inhibitory constraint exerted by its C-terminal domain.
Macrophages are a central component of the immune systemwhere they function as sentinels to detect pathogens (e.g. bacteria andviruses) [1]. The activation of macrophages by pathogens is largelymediated by Toll-like receptors (TLRs) [2,3]. TLRs selectively recogniseunique components of pathogens; these components are referred toas pathogen-associated molecular patterns (PAMPs). For example,TLR4 recognises lipopolysaccharide from Gram-negative bacteria[4,5], whereas TLR9 detects CpG DNA (e.g. bacterial DNA) [6]. Thedetection of PAMPs by TLRs triggers the activation of intracellularnetworks of signalling pathways that control the inflammatoryresponses of macrophages to pathogens [7–9].
Formost TLRs, the initial signallingevent is the recruitmentof theTIRdomain-containing adaptor protein MyD88 to the activated TLRcomplex [10–12]. This is then followed by the recruitment of theserine/threonine kinases IRAK-1 and IRAK-4. Phosphorylation of IRAK-1by IRAK-4 likely triggers its autophosphorylation, leading to the fullactivation of IRAK-1 and its dissociation from MyD88 and the receptorcomplex [13–15]. The subsequent binding of TRAF6 to phosphorylatedIRAK-1 [16,17] ultimately leads to the activation of additional proteinkinases (e.g. IκBkinase, JNKandp38MAPkinase) aswell as transcriptionfactors (e.g. NF-κB) [17–20]. Activation of these signalling networksculminates in the highly coordinated expression of inflammatory
61 3 9347 1863.lz).
l rights reserved.
cytokines (e.g. TNF, IL-1β and IL-6) and chemokines (e.g. CCL2) thathelp orchestrate the ensuing innate and adaptive immune responses[1,3,21,22]. While the secretion of these inflammatory mediators bymacrophages is important for effective immune responses to infection,their excessive and prolonged systemic release can be harmful to thehost and in some cases fatal (e.g. septic shock) [23].
Due to its proximal position to TLRs, IRAK-1 is a key molecule ininflammatory signalling. IRAK-1 contains four major structuralcomponents, firstly an N-terminal death domain, which allows it tobind other death domain-containing proteins (e.g. MyD88) and tooligomerise, secondly a so-called ProST region that is followed by aserine/threonine kinase catalytic domain, and finally a C-terminaldomain that contains binding motifs for TRAF6 [24]. IRAK-1 is rapidlyubiquitinated following its activation and subsequently degraded bythe proteasome [25–27]; this likely represents an important negativefeedback mechanism to terminate IRAK-1 signalling. Other mechan-isms of IRAK-1 regulation have also been identified. IRAK-M, astructurally related but catalytically inactive protein kinase whoseexpression can be induced by TLR ligands, suppresses IRAK-1 activityby blocking its phosphorylation by IRAK-4 and/or preventing itsrelease from activated receptor complexes [28]. The adaptor proteinTollip is reported to keep IRAK-1 in an inactive conformation in theabsence of TLR/IL-1 receptor stimulation [29,30].
The C-terminal domain of IRAK-1 is substantially longer than thoseof the other three members of the IRAK family (IRAK-2, IRAK-M andIRAK-4). Here we report that deletion of themajority of the C-terminaldomain of IRAK-1markedly enhanced its TLR-independent capacity tooligomerise and to activate NF-κB. Deletion of the C-terminal domain
720 T. Nguyen et al. / Cellular Signalling 21 (2009) 719–726
also resulted in the TLR-independent binding of IRAK-M to IRAK-1, aninteraction that required the death domain in IRAK-1. Thus, the C-terminal domain of IRAK-1 appears to play a role in negativelyregulating the activation state of IRAK-1.
2. Experimental procedures
2.1. Reagents and antibodies
Pfu DNA polymerase, restriction enzymes and the Dual-Glo™luciferase assay system were purchased from Promega. PCR primerswere synthesised byGeneWorks. Cell culturemediumand supplements,foetal calf serum, anti-V5 antibodies, and pre-cast 10% SDS-PAGE gelswere from Invitrogen. FuGENE HD™ and FuGENE 6™ transfectionreagents, and Complete™ protease inhibitors were supplied by RocheBiochemicals. Horse-radish peroxidase-coupled anti-FLAG antibodies(M2) were obtained from Sigma-Aldrich. Protein G-Sepharose waspurchased from GE Healthcare.
2.2. Expression vectors
The IRAK-1 vector pEF-V5-IRAK-1, which expresses an N-terminalV5-tagged version ofmouse IRAK-1, has been described previously [31].The vector pEF-FLAG-IRAK-1 was created by excising the cDNA insertfrom pEF-V5-IRAK-1 with MluI and cloning it into the MluI site in pEF-FLAG. A vector that expresses V5-IRAK-1 KD, a kinase-dead version ofV5-IRAK-1 in which Lys-239 is replaced by Ala, was generated by PCRusing the primer pair: F1 (5′-ACG CGT GCC GGG GGG CCG GGC CCCGGG-3′) and R1 (5′-ACG CGT TCA GCT CTG GAA TTC ATC ACT TTC TTCAGGTCC-3′), and pIC-FLAG-IRAK-1(K239A) [30] as the template. The PCRproduct was digested with MluI and cloned into the MluI site in pEF-V5.The vector pEF-V5-IRAK-1Δ700–710, which expresses a V5-taggedversion of IRAK-1 lacking Gln-700 to Ser-710, was created by PCR usingthe primer pair: F1 and R2 (5′-ACG CGT TCA GCT CTT TTC AGG TTC TAAATC CAA GCC-3′). The expression vectors pEF-V5-IRAK-1Δ599–710(lacks Trp-599 to Ser-710), pEF-V5-IRAK-1Δ579–710 (lacks Pro-579 toSer-710), and V5-IRAK-1ΔCTD (lacks Ala-551 to Ser-710) were createdsimilarly using the PCR primer pairs: F1 and R3 (5′-ACG CGT TCA GGAATG CAG GGT AGC AGA GAG GCC GGG-3′), F1 and R4 (5′-ACG CGT TCAACTTCTCTGGAGTTGCTGGGCAGCCTG-3′), and F1 andR5 (5′-ACGCGTTCA ACT GCC AGT GGTAGA CAT GTAGGAGTT CTC-3′), respectively. Theexpression vector pEF-FLAG-IRAK-1ΔCTD was created by excising thecDNA insert from pEF-V5-IRAK-1ΔCTDwithMluI and cloning it into theMluI site in pEF-FLAG. A vector expressing V5-IRAK-1ΔDD, which lacksthe N-terminal death domain (i.e. Ala-2 to Pro-111) of IRAK-1, wascreated by PCR using the primer pair: F2 (5′-ACG CGTAGC ACCGCTGCCCCA AGG CCC AGC AGC ATC-3′) and R1. A vector that expresses FLAG-tagged IRAK-M was creating via reverse-transcription PCR using totalRNA from LPS-stimulated mouse bone marrow-derived macrophagesand the primer pair: F3 (5′-ACG CGT GCC GGC CGG TGC GGG GCC CGTGGC GCG CTG-3′) and R6 (5′-ACG CGT TCA CTG CTT TTT GGA CTG TTCATG CTC ACT-3′). The PCR product was digested with MluI and clonedinto the MluI site in pEF-FLAG. The pEF-FLAG-IRAK-4 KD vector, whichexpresses a kinase-dead version of FLAG-tagged IRAK-4 in which bothLys-213 and Lys-214 were replaced with Ala, has been describedpreviously [31]. The FLAG-tagged TRAF6 expression vector was from Dr.Ashley Mansell (Monash Institute of Medical Research, MonashUniversity). The NF-κB firefly luciferase reporter plasmid pELAM-FLwas a generous gift from Dr. Matthew Sweet (Institute of MolecularBiosciences, University of Queensland), while the Renilla luciferasereporter plasmid pRL-TK was from Promega.
2.3. Cell culture and transfections
Mouse RAW 264.7 macrophages and human HEK293T cells werecultured in Dulbecco's-modified Eagle's medium (DMEM) supple-
mented with 10% FCS, 100 units/mL Penicillin, 100 μg/mL Streptomy-cin, and 2 mM GlutaMax-1™ at 37 °C in a humidified atmosphere of5% CO2. RAW 264.7 cells were transiently transfected using FuGENEHD™ transfection reagent, while FuGENE 6™ transfection reagent wasused in the case of HEK293T cells. The cells were typically analysed24 h post-transfection.
2.4. NF-κB gene reporter assays
RAW 264.7 cells were seeded in 6-well tissue culture plates at adensity of 8×105 cells per well and transfected the next day usingFuGENE HD™ transfection reagent. Cells were co-transfected with100 ng pELAM-FL, 10 ng pRL-TK, and 100 ng of the IRAK-1 expressionvector being tested. The total amount of DNA transfected was keptconstant using empty vector where required. The cells were lysed 24 hpost-transfection using Passive lysis buffer and assayed for firefly andRenilla luciferase activity using the Dual-Glo™ luciferase assay system(Promega). The level of Renilla luciferase activity was used tonormalise the transfection efficiencies.
2.5. Cell lysis and Western blotting
Following washing with ice-cold PBS, cells were scraped into NP-40lysis buffer (20 mM Tris–HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1%Nonidet P-40, 10% glycerol, 1 mM sodium orthovanadate, 10 mM NaF,10 mM β-glycerophosphate, and Complete™ protease inhibitors) andincubated on ice for 30 min. The lysates were then clarified bycentrifugation at 13,000 ×g for 10 min at 4 °C and the concentrationsof the cell lysates measured using a Bio-Rad protein assay kit. Lysateswere subjected to electrophoresis on 10% SDS-PAGE gels followed byWestern blotting according to standard techniques. Immunoreactivebands were visualised using ECL reagents (Amersham Bioscience) andexposure to X-ray film (Fuji). Filmswere scanned using either a GS710 orGS800 Calibrated Imaging Densitometer (Bio-Rad) and the imagesexported as TIFF files.
2.6. Immunoprecipitation assays
V5-tagged IRAK-1 proteins were immunoprecipitated by incubat-ing cell lysates (typically containing 250–1000 μg of protein) with 1 μganti-V5 monoclonal antibody for 6 h at 4 °C with continual mixing.Immune complexes were then captured by the addition of Protein G-Sepharose. Following four washes with lysis buffer, the immunopre-cipitates were subjected to electrophoresis on 10% SDS-PAGE gels andWestern blotting.
3. Results
3.1. Deletion of the C-terminal domain or death domain of IRAK-1enhances its TLR-independent activation of NF-κB in macrophages
Currentmodels of TLR signalling propose that IRAK-1 is recruited toactivated TLR complexes via an interaction between its death domainand the death domain in MyD88 [13–15]; the subsequent binding ofTRAF6 to the C-terminal domain of IRAK-1 then mediates NF-κB acti-vation [16–20]. However, themechanisms underlying the regulation ofthese protein–protein interactions have not been elucidated. One pos-sibility is that IRAK-1 adopts an inactive conformation, potentially viaintramolecular interactions, inwhich the death domain andC-terminaldomain are orientated in such a way that their binding to MyD88 andTRAF6, for example, is prevented in the absence of TLR stimulation. TheC-terminal domain of IRAK-1 is substantially longer than those of theother three members of the IRAK family [32]. Thus we hypothesisedthat, in addition to binding TRAF6, the C-terminal domain of IRAK-1mayalso function to keep IRAK-1 in an inactive conformation in unstimu-lated cells. This hypothesis was tested by assessing the effect of deleting
Fig. 1. Activation of NF-κB by IRAK-1 deletion mutants in macrophages. (A) Schematicrepresentation of V5-tagged wildtype IRAK-1 (V5-IRAK-1) and mutants of IRAK-1 lackingeither the majority of the C-terminal domain (V5-IRAK-1ΔCTD) or the N-terminal deathdomain (V5-IRAK-1ΔDD). (B) RAW 264.7 cells were transfected with a NF-κB luciferasereporterplasmidalongwith the indicatedV5-IRAK-1expressionvectors. Luciferase activitywas measured 24 h post-transfection using the Dual-Glo™ luciferase assay system. NF-κBreporter activity is given as the fold increase over levels in cells transfected with emptyvector. Data from three independent experiments (duplicate transfections) are presentedas the mean value±SEM.
Fig. 2. Binding of TRAF6 to IRAK-1 deletion mutants in macrophages. (A) Schematicrepresentation of V5-tagged IRAK-1 in which the three TRAF6-binding motifs in theC-terminal domain of IRAK-1 are indicated by asterisks. The sequences of the TRAF6binding motifs as well as the consensus TRAF6 binding motif are shown (Ar/Ac:aromatic or acidic amino acid). (B) RAW 264.7 cells were co-transfected withplasmids expressing FLAG-TRAF6 and V5-IRAK-1, V5-IRAK-1ΔCTD or V5-IRAK-1ΔDD.Twenty-four h later the V5-tagged IRAK-1 proteins were immunoprecipitated fromlysates of the cells using anti-V5 antibodies. The immunoprecipitates were thensubjected to Western blotting with anti-FLAG antibodies to detect the co-immunoprecipitation of FLAG-TRAF6. The whole cell lysates (WCL) were alsosubjected to Western blotting with anti-FLAG antibodies to confirm comparablelevels of FLAG-TRAF6 expression. The data are representative of three independentexperiments.
721T. Nguyen et al. / Cellular Signalling 21 (2009) 719–726
the C-terminal domain on the ability of IRAK-1 to activate NF-κB in theabsence of TLR activation (Fig. 1).
IRAK-1 undergoes ligand-independent autoactivation when ecto-pically expressed in HEK293T cells as a result of the high levels ofIRAK-1 expression achieved [30,33–35]. In order to minimise ligand-independent activation of IRAK-1, we primarily expressed IRAK-1 (andmutants thereof) in mouse RAW 264.7 macrophages. Expressionlevels of transfected IRAK-1 in RAW 264.7 cells were at least 20-foldlower than those in similarly transfected HEK293T cells (data notshown). Consistent with the idea that IRAK-1 adopts a largely inactiveconformation in unstimulated cells, its expression in RAW 264.7 cellsresulted in only a small increase (~1.5-fold) in the activity of the co-transfected NF-κB reporter plasmid (Fig. 1B). By contrast, ectopicexpression of an IRAK-1 mutant lacking most of the C-terminaldomain (i.e. IRAK-1ΔCTD) resulted in an approximately 4-foldincrease in NF-κB reporter activity (Fig. 1B). Deletion of the N-terminal death domain in IRAK-1 (IRAK-1ΔDD) also resulted inincreased activation of the NF-κB reporter (Fig. 1B). These findingssuggest that the C-terminal domain, as well as the death domain, maykeep IRAK-1 in an inactive conformation in the absence of TLRactivation.
3.2. Deletion of the death domain results in the TLR-independent bindingof IRAK-1 to TRAF6 in macrophages
In view of the above results, the ability of IRAK-1ΔCTD and IRAK-1ΔDD to bind TRAF6 was tested. IRAK-1ΔCTD contains one of the threeTRAF6-bindingmotifs located in the C-terminal domain of IRAK-1,whileIRAK-1ΔDD contains all three TRAF6-binding motifs (Fig. 2A) [36].Although IRAK-1ΔCTD activated NF-κB, it did not form detectablecomplexes with FLAG-TRAF6 in RAW 264.7 cells. It has recently beenreported that Lys-63 ubiquitination of IRAK-1 (on Lys-134 and Lys-180)by TRAF6 can facilitate the recruitment of the IKK complex regulatory
subunit NEMO to IRAK-1 [37]. Consequently, the transient interaction ofTRAF6 with IRAK-1ΔCTD may be sufficient for NF-κB activation. Incontrast to IRAK-1ΔCTD, the IRAK-1 death domain deletion mutant,IRAK-1ΔDD, formed stable complexes with TRAF6 (Fig. 2B). Thisparticular finding suggests that the death domain of IRAK-1 maynegatively regulate the bindingof TRAF6 to IRAK-1 in unstimulated cells.
3.3. Deletion of the C-terminal domain of IRAK-1 results in itsTLR-independent oligomerisation in macrophages
Activation of IRAK-1 is accompanied by its death domain-mediatedoligomerisation [38–40]. Therefore we tested the ability of the IRAK-1C-terminal deletion mutant, IRAK-1ΔCTD, to oligomerise in theabsence of TLR activation. FLAG-tagged IRAK-1ΔCTD formed oligo-mers with V5-IRAK-1ΔCTD when the two proteins were co-expressedin RAW 264.7 cells (Fig. 3A). By contrast, full-length FLAG-IRAK-1failed to oligomerise with V5-tagged IRAK-1 under the sameconditions (Fig. 3A). However, when full-length IRAK-1 was expressedin HEK293T cells, which allows IRAK-1 to spontaneously autoactivate,its oligomerisation was observed (Fig. 3B). Furthermore, the extent towhich full-length IRAK-1 oligomerised was similar to the oligomer-isation of IRAK-1ΔCTD when their different levels of expression weretaken into account (Fig. 3B). Thus, the differences seen in IRAK-1oligomerisation in RAW 264.7 cells are unlikely to have been due todifferences in the intrinsic capacities of full-length IRAK-1 and IRAK-1ΔCTD to oligomerise but rather reflected their different activationstates.
Fig. 3. Oligomerisation of IRAK-1. (A) RAW 264.7 cells and (B) HEK293T cells weretransfectedwith plasmids expressing FLAG-IRAK-1and/or V5-IRAK-1, or FLAG-IRAK-1ΔCTDand/or V5-IRAK-1ΔCTD. Twenty-four h later V5-IRAK-1 and V5-IRAK-1ΔCTD wereimmunoprecipitated from lysates of the cells using anti-V5 antibodies. The immunopre-cipitates were then subjected to Western blotting with anti-FLAG antibodies to detect theco-immunoprecipitation of FLAG-IRAK-1 or FLAG-IRAK-1ΔCTD. Thedata are representativeof three independent experiments.
Fig. 4. Binding of IRAK-M to IRAK-1 deletion mutants. (A) HEK293T cells were transfected witlater V5-IRAK-1 and V5-IRAK-1 KD were immunoprecipitated from lysates of the cells usingwith anti-FLAG antibodies to detect the co-immunoprecipitation of FLAG-IRAK-M. (B) RAW1, V5-IRAK-1ΔCTD or V5-IRAK-1ΔDD and/or FLAG-IRAK-M. Twenty-four h later V5-IRAK-cells using anti-V5 antibodies. The immunoprecipitates were then subjected to Western blM. (A–C) The whole cell lysates (WCL) were also subjected to Western blotting with anti-FLrepresentative of three independent experiments.
722 T. Nguyen et al. / Cellular Signalling 21 (2009) 719–726
3.4. Deletion of the C-terminal domain of IRAK-1 results in itsTLR-independent binding to IRAK-M in macrophages
IRAK-M is a catalytically-inactive protein kinase that inhibits TLRsignalling by binding to activated IRAK-1 and/or IRAK-4 [28]. Toconfirm that IRAK-M only binds to IRAK-1 once IRAK-1 has adopted anactive conformation, wildtype IRAK-1 and a kinase-inactive IRAK-1mutant were co-expressed with IRAK-M in HEK293T cells. IRAK-M co-immunoprecipitated with wildtype IRAK-1 but failed to co-immuno-precipitate with the kinase-inactive IRAK-1 mutant, V5-IRAK-1 KD(Fig. 4A). Significantly, IRAK-M did not co-immunoprecipitate withfull-length IRAK-1 when they were co-expressed in RAW 264.7 cells(Fig. 4B), suggesting that IRAK-1 had adopted a largely inactiveconformation. By contrast, complex formation between IRAK-M andIRAK-1ΔCTDwas detected following their co-expression in RAW264.7cells (Fig. 4B). It should be noted that IRAK-M co-immunoprecipitatedwith V5-IRAK-1 following the stimulation of transfected RAW 264.7cells with the TLR9 ligand CpG DNA (data not shown). IRAK-M also co-immunoprecipitated with V5-IRAK-1 (and IRAK-1ΔCTD) when theywere co-expressed in HEK293T cells (Fig. 4C). Notably, IRAK-M failedto co-immunoprecipitate with IRAK-1ΔDD, which suggests that thebinding of IRAK-M to IRAK-1 is mediated by interactions betweentheir respective death domains. Because IRAK-M only binds to IRAK-1once IRAK-1 has become activated, our finding that IRAK-M forms astable complex with IRAK-1ΔCTD, but not with full-length IRAK-1, inRAW 264.7 cells further supports the idea that the C-terminal domainof IRAK-1 may suppress the activation of IRAK-1 in the absence ofappropriate stimulation (e.g. TLR stimulation).
3.5. The C1 subdomain negatively regulates the activation state of IRAK-1
The C-terminal domain of IRAK-1 can be subdivided into the C1(residues 523–618) and C2 (residues 619–710) subdomains [24]. In orderto define which region of the C-terminal domain of IRAK-1 potentiallyfunctions to keep the kinase in a repressed conformation, a series of C-terminal truncationmutants were generated and their relative ability toactivate NF-κB in RAW 264.7 cells tested (Fig. 5). IRAK-1Δ700–710, aswell as IRAK-1Δ599–710,which lacks the C2 subdomain of IRAK-1 along
h plasmids expressing V5-IRAK-1 or V5-IRAK-1 KD and/or FLAG-IRAK-M. Twenty-four hanti-V5 antibodies. The immunoprecipitates were then subjected to Western blotting264.7 cells and (C) HEK293Tcells were transfectedwith plasmids expressing V5-IRAK-1, V5-IRAK-1ΔCTD and V5-IRAK-1ΔDD were immunoprecipitated from lysates of theotting with anti-FLAG antibodies to detect the co-immunoprecipitation of FLAG-IRAK-AG antibodies to confirm comparable levels of FLAG-IRAK-M expression. The data are
Fig. 5. NF-κB activation and TRAF6 binding by C-terminal truncation mutants of IRAK-1in macrophages. (A) Schematic representation of V5-tagged C-terminal truncationmutants of IRAK-1. The TRAF6 binding motifs in the C-terminal domain of IRAK-1 areindicated by asterisks. (B) RAW 264.7 cells were transfected with a NF-κB luciferasereporter plasmid along with the indicated V5-IRAK-1 expression vectors. Luciferaseactivity was measured 24 h post-transfection using the Dual-Glo™ luciferase assaysystem. NF-κB reporter activity is given as the fold increase over levels in cellstransfected with empty vector. Data from three independent experiments (duplicatetransfections) are presented as the mean value±SEM. (C) RAW 264.7 cells were co-transfected with plasmids expressing the indicated proteins. Twenty-four h later theV5-tagged IRAK-1 proteins were immunoprecipitated from lysates of the cells usinganti-V5 antibodies. The immunoprecipitates were then subjected to Western blottingwith anti-FLAG antibodies to detect the co-immunoprecipitation of FLAG-TRAF6. Thewhole cell lysates (WCL) were also subjected to Western blotting with anti-FLAGantibodies to confirm comparable levels of FLAG-TRAF6 expression. The data arerepresentative of three independent experiments.
Fig. 6. Binding of IRAK-M to C-terminal truncation mutants of IRAK-1 in macrophages.(A) RAW 264.7 cells were co-transfected with plasmids expressing the indicated IRAKproteins. Twenty-four h later the V5-tagged IRAK-1 proteins were immunoprecipitatedfrom lysates of the cells using anti-V5 antibodies. The immunoprecipitates were thensubjected to Western blotting with anti-FLAG antibodies to detect the co-immunopre-cipitation of FLAG-IRAK-M. The whole cell lysates (WCL) were also subjected toWesternblotting with anti-FLAG antibodies to confirm comparable levels of FLAG-IRAK-Mexpression. (B) Quantified data from two independent experiments are presented as themean value±SEM. The level of IRAK-M binding to V5-IRAK-1ΔCTD was arbitrarily givena value of 100%.
723T. Nguyen et al. / Cellular Signalling 21 (2009) 719–726
with the last 20 amino acids of the C1 subdomain, did not have anincreased ability to activate NF-κB (Fig. 5B). However, IRAK-1Δ579–710,an IRAK-1 mutant lacking the C2 subdomain and the last ~40 aminoacids of the C1 subdomain, was capable of activating NF-κB in a TLR-
independent manner, with the levels of NF-κB activation beingcomparable with those achieved by IRAK-1ΔCTD (Fig. 5B).
Because of the role played by TRAF6 in the activation of NF-κB byIRAK-1, the ability of the IRAK-1 truncation mutants to bind TRAF6was also tested. The mutants contain either one or two of the threeTRAF6-binding motifs in the C-terminal domain of IRAK-1 (Fig. 5A). Asshown in Fig. 5C, TRAF6 did not co-immunoprecipitate with the IRAK-1 truncation mutants when they were co-expressed in RAW 264.7cells, although TRAF6 did co-immunoprecipitate with the IRAK-1death domain deletion mutant, V5-IRAK-1ΔDD. In the case of IRAK-1Δ700–710 and IRAK-1Δ599–710, the lack of detectable complexformation with TRAF6 is most likely explained by the truncation oftheir C-terminal domains being insufficient to bring about theactivation of IRAK-1 and the positioning of the C-terminal domainsuch that it can bind TRAF6. Like IRAK-1ΔCTD, IRAK-1Δ579–710 only
724 T. Nguyen et al. / Cellular Signalling 21 (2009) 719–726
contains a single TRAF6-binding motif, which may be insufficient for astable interaction between IRAK-1 and TRAF6.
Given the effects of progressively truncating the C-terminal domainof IRAK-1 on its ability to activate NF-κB in the absence of TLRactivation, we also tested the ability of the truncation mutants to bindIRAK-M. The effects of truncating IRAK-1 on IRAK-M binding werefound to largely mirror those of NF-κB activation by IRAK-1 (Fig. 6).Only very low levels of IRAK-M co-immunoprecipitated with IRAK-1Δ700–710 from transfected RAW 264.7 cells (Fig. 6). While the levelsof IRAK-M that co-immunoprecipitated with IRAK-1Δ599–710 were
Fig. 7. Differential binding of IRAK-4 and IRAK-M to IRAK-1 deletion mutants. (A) RAW264.7 cells and (B) HEK293T cells were co-transfected with plasmids expressing theindicated IRAK proteins. Twenty-four h later the V5-tagged IRAK-1 proteins wereimmunoprecipitated from lysates of the cells using anti-V5 antibodies. The immuno-precipitates were then subjected to Western blotting with anti-FLAG antibodies todetect the co-immunoprecipitation of FLAG-IRAK-4 or FLAG-IRAK-M. The whole celllysates (WCL) were also subjected to Western blotting with anti-FLAG antibodies toconfirm comparable levels of FLAG-IRAK-4 and FLAG-IRAK-M expression. The data arerepresentative of four independent experiments.
higher, they were still significantly lower than those that had co-immunoprecipitated with IRAK-1ΔCTD (Fig. 6). By contrast, the levelsof IRAK-M that co-immunoprecipitated with IRAK-1Δ579–710approached those that bound to IRAK-1ΔCTD (Fig. 6). All four C-terminal truncationmutants had a similar ability to bind IRAK-Mwhenthey were overexpressed in HEK293T cells (data not shown). Taken alltogether, the above findings suggest the amino acid sequence betweenPro-579 and Ser-598 in the C1 subdomain of IRAK-1 is important forkeeping IRAK-1 in a signalling-inactive conformation in unstimulatedcells.
3.6. Death domain-independent binding of IRAK-4 to IRAK-1
Because the phosphorylation of IRAK-1 by IRAK-4 represents anearly event in the TLR-triggered activation of IRAK-1 [13,14,35,41],and deletion of the C-terminal domain of IRAK-1 resulted in theTLR-independent binding of IRAK-M to IRAK-1, we sought to deter-mine the effects of deleting the C-terminal domain and the deathdomain of IRAK-1 on the binding of IRAK-4 to IRAK-1. When co-expressed in RAW 264.7 cells, kinase-dead IRAK-4 did not co-immunoprecipitate with IRAK-1; nor did it co-immunoprecipitatewith IRAK-1ΔDD (Fig. 7A). In contrast to IRAK-M, TLR-independentcomplex formation between kinase-dead IRAK-4 and the C-terminalIRAK-1 deletion mutant was not detected (Fig. 7A). IRAK-4 wascapable of binding to the kinase-dead IRAK-1 mutant, IRAK-1 KD,when they were overexpressed in HEK293T cells; however, thelevels of IRAK-4 that co-immunoprecipitated with wildtype IRAK-1and IRAK-1ΔCTD were much lower (Fig. 7B). Significantly, IRAK-4co-immunoprecipitated with IRAK-1ΔDD from transfected HEK293Tcells (Fig. 7B). Thus, in contrast to the binding of IRAK-M to IRAK-1,the death domain of IRAK-1 was not required for the interaction ofIRAK-4 with IRAK-1; IRAK-4 also appeared to preferentially bind tokinase-inactive IRAK-1.
4. Discussion
IRAK-1 plays a central role in TLR signalling and hence its activationstate must be tightly regulated in order to prevent excessiveinflammation. Mechanisms which terminate signalling by IRAK-1include the binding of IRAK-M to IRAK-1 as well as the proteasomaldegradation of IRAK-1 [25–28]. Molecular chaperones (e.g. Hsp90)may also impact on the signalling capacity of IRAK-1 by directlyregulating its stability [34]. Mechanisms that prevent IRAK-1 frombecoming activated in the absence of TLR ligands are also likely to beimportant for maintaining immune cells in a quiescent state.
IRAK-1 has a relatively long C-terminal domain that contains threeTRAF6 binding motifs [36]. However, additional functions for the C-terminal domain of IRAK-1 have so far not been described. We haveshown here that deletion of the majority of the C-terminal domain ofIRAK-1 markedly enhanced the capacity of IRAK-1 to activate NF-κB ina TLR-independent manner in RAW 264.7 macrophages. Additionally,unlike wildtype IRAK-1, the C-terminal truncation mutant, IRAK-1ΔCTD, spontaneously oligomerised when it was expressed in RAW264.7 cells, an event associated with the activation of IRAK-1 [38–40].Further truncation analysis of IRAK-1 identified a region spanningaround 20 amino acids in the second half of the C-terminal C1subdomain as being important for keeping IRAK-1 in an inactiveconformation. TLR-independent activation of NF-κB by IRAK-1 wasalso increased following deletion of the death domain at the N-terminus of IRAK-1. Together, these findings suggest that a novelfunction of the C-terminal domain of IRAK-1, possibly in concert withthe death domain, is to negatively regulate IRAK-1 in order to preventits inappropriate activation.
The mechanism underlying the potential negative regulation ofIRAK-1 activation by its C-terminal domain is unclear at this point. Thesmall adaptor protein Tollip has been implicated in keeping IRAK-1 in
Fig. 8. A proposedmodel for the regulation of IRAK-1 activation by an intramolecular mechanism. In the absence of TLR activation, IRAK-1 is held in a closed, inactive conformation viaan intramolecular mechanism involving the C-terminal domain and the death domain. Phosphorylation of IRAK-1 by IRAK-4 in response to TLR activation may then release IRAK-1from the inhibitory constraint exerted by its C-terminal domain. Concomitant with this activation of the kinase domain of IRAK-1 is the re-orientation of both the death domain andC-terminal domain in order to facilitate their interaction with MyD88 and TRAF6, respectively, leading to NF-κB activation. In this open, active conformation the death domain inIRAK-1 would also be capable of interacting with IRAK-M in order to terminate IRAK-1 signalling.
725T. Nguyen et al. / Cellular Signalling 21 (2009) 719–726
an inactive state in quiescent cells [13,29,30]. Thus, one possibleexplanation for how deletion of the C-terminal domain of IRAK-1could promote the TLR-independent activation of IRAK-1 is that thedomain is required for the binding of Tollip to IRAK-1. However, we didnot detect complex formation between Tollip and IRAK-1 when theywere co-expressed in RAW 264.7 cells (data not shown). The ability ofthe C-terminal domain of IRAK-1 to regulate the activation state ofIRAK-1 may therefore occur independently of Tollip. Indeed, findingsfrom studies with cells from Tollip knockout mice have raised doubtsas to whether Tollip functions as a negative regulator of IRAK-1 [42].Both basal IRAK-1 activity and ligand-induced activation of IRAK-1were normal in Tollip knockout cells; signalling downstream of IRAK-1(e.g. activation of IKK and MAP kinases) was also unaffected. It hassubsequently been suggested that Tollip regulates inflammatorysignalling by functioning as an endosomal adaptor protein duringreceptor trafficking [43]. Nonetheless, a role for Tollip in keeping IRAK-1in an inactive conformation, potentially via an interaction with the C-terminal domain of IRAK-1, cannot be excluded.
The data presented here are consistentwith amodel inwhich IRAK-1is held in a closed conformation in the absence of TLR activation via anintramolecular mechanism involving the C-terminal domain (Fig. 8). Inthis closed conformation, the catalytic activity of the kinase domainmaybe repressed and the C-terminal domain orientated such that it cannotinteractwith TRAF6. Further support for thismodel comes from analysisof the effect of deleting the C-terminal domain on the binding of IRAK-Mto IRAK-1. At least one function of IRAK-M is to inhibit signalling byactivated IRAK-1 [28]. The finding that deletion of the C-terminaldomain resulted in the TLR-independent binding of IRAK-M to IRAK-1 inRAW 264.7 cells is consistent with the C-terminal domain exerting aninhibitory effect on IRAK-1 activation. Additionally, because the bindingof IRAK-M to IRAK-1 required the death domain in IRAK-1, we wouldargue that the C-terminal domainmayalso regulate thedeath domain attheN-terminus of IRAK-1. Specifically, a conformational change in the C-terminal domain of IRAK-1, which brings about the activation of thecatalytic activity of IRAK-1,may simultaneously lead to a conformationalchange in the death domain that enables it to interact with other deathdomains, such as those in MyD88 and other IRAKs (e.g. IRAK-M).
How could the apparent inhibitory constraint imposed on IRAK-1by its C-terminal domain be released? Phosphorylation of IRAK-1 byIRAK-4 is likely to be at least part of the answer. An in vitro studyrevealed that autophosphorylation of IRAK-1 at Thr-209 partiallyactivated its kinase activity, while the subsequent phosphorylation ofThr-387 in the activation loop led to its full activation [35]. IRAK-4 canphosphorylate peptides containing Thr-209 and Thr-387 [14,35] andhencemay phosphorylate these sites in IRAK-1 in vivo in order to bring
about a conformational change in IRAK-1 and its activation (Fig. 8).Indeed, Kollewe et al. have reported that, whilemutation of Thr-209 tothe phosphomimetics, aspartic acid and glutamic acid, was notsufficient to activate IRAK-1, mutation of Thr-209 caused a conforma-tional change in the kinase domain of IRAK-1 [35]. Whether IRAK-4phosphorylates IRAK-1 at additional sites including within the C-terminal domain is still to be established, as are the effects of IRAK-1phosphorylation on the conformation of the C-terminal domain. It hasbeen reported that IRAK-4 binds to the death domain of IRAK-1 [39].However, we found that IRAK-4 was capable of co-immunoprecipitat-ing with an IRAK-1 mutant lacking the N-terminal death domain (i.e.IRAK-1ΔDD). IRAK-4 also exhibited greater binding to a kinase-inactiveIRAK-1 mutant than to wildtype IRAK-1. Therefore, the structuralfeatures of IRAK-4 and IRAK-1 that regulate their interaction wouldpossibly allow IRAK-4 to phosphorylate IRAK-1 that has adopted aclosed, inactive conformation and in doing so release IRAK-1 from apotential negative-regulatory constraint exerted by its C-terminaldomain. An ensuing conformational change in IRAK-1 may subse-quently allow IRAK-1 to transiently interact with MyD88 and TRAF6,but ultimately bind to IRAK-M in order to terminate its signallingactivity (Fig. 8).
5. Conclusions
In summary, deletion of the C-terminal domain of IRAK-1significantly enhanced the ability of IRAK-1 to oligomerise andactivate NF-κB in the absence of TLR activation; it also resulted inthe increased binding of the negative regulator IRAK-M to IRAK-1.While the binding of IRAK-M to IRAK-1 appeared to be mediated bythe death domain at the N-terminus of IRAK-1, the IRAK-1 activator,IRAK-4, preferentially bound inactive IRAK-1 in a death domain-independent manner. These findings provide the basis for a novelregulatory model in which the activation state of IRAK-1 is regulatedby its C-terminal domain.
Acknowledgements
The authors wish to thank Drs. S. Ghosh (Yale University), A.Mansell (Monash University, Australia) and M. Sweet (University ofQueensland, Australia) for generously providing plasmids. This workwas supported in part by the Cooperative Research Centre for ChronicInflammatory Diseases and the National Health and Medical ResearchCouncil (NHMRC). T.N. was supported by a NHMRC Biomedical (DoraLush) Postgraduate scholarship, while J.A.H. is supported by a NHMRCSenior Principal Research Fellowship.
726 T. Nguyen et al. / Cellular Signalling 21 (2009) 719–726
References
[1] S. Gordon, Res. Immunol. 149 (1998) 685.[2] S. Akira, S. Uematsu, O. Takeuchi, Cell 124 (2006) 783.[3] C.A. Janeway Jr., R. Medzhitov, Annu. Rev. Immunol. 20 (2002) 197.[4] A. Poltorak, X. He, I. Smirnova, M.Y. Liu, C. Van Huffel, X. Du, D. Birdwell, E. Alejos,
M. Silva, C. Galanos, M. Freudenberg, P. Ricciardi-Castagnoli, B. Layton, B. Beutler,Science 282 (1998) 2085.
[5] K. Hoshino, O. Takeuchi, T. Kawai, H. Sanjo, T. Ogawa, Y. Takeda, K. Takeda, S. Akira,J. Immunol. 162 (1999) 3749.
[6] T.H. Chuang, J. Lee, L. Kline, J.C. Mathison, R.J. Ulevitch, J. Leukoc. Biol. 71 (2002)538.
[7] S. Akira, K. Takeda, Nat. Rev. Immunol. 4 (2004) 499.[8] B. Beutler, K. Hoebe, X. Du, R.J. Ulevitch, J. Leukoc. Biol. 74 (2003) 479.[9] S.M. Miggin, L.A. O'Neill, J. Leukoc. Biol. 80 (2006) 220.[10] K. Burns, F. Martinon, C. Esslinger, H. Pahl, P. Schneider, J.L. Bodmer, F. Di Marco, L.
French, J. Tschopp, J. Biol. Chem. 273 (1998) 12203.[11] R. Medzhitov, P. Preston-Hurlburt, E. Kopp, A. Stadlen, C. Chen, S. Ghosh, C.A.
Janeway Jr., Mol. Cell 2 (1998) 253.[12] H. Wesche, W.J. Henzel, W. Shillinglaw, S. Li, Z. Cao, Immunity 7 (1997) 837.[13] K. Burns, S. Janssens, B. Brissoni, N. Olivos, R. Beyaert, J. Tschopp, J. Exp. Med. 197
(2003) 263.[14] S. Li, A. Strelow, E.J. Fontana, H. Wesche, Proc. Natl. Acad. Sci. U.S.A. 99 (2002) 5567.[15] N. Suzuki, S. Suzuki, G.S. Duncan, D.G. Millar, T. Wada, C. Mirtsos, H. Takada, A.
Wakeham, A. Itie, S. Li, J.M. Penninger, H. Wesche, P.S. Ohashi, T.W. Mak, W.C. Yeh,Nature 416 (2002) 750.
[16] Z. Cao, J. Xiong, M. Takeuchi, T. Kurama, D.V. Goeddel, Nature 383 (1996) 443.[17] Z. Jiang, J. Ninomiya-Tsuji, Y. Qian, K. Matsumoto, X. Li, Mol. Cell. Biol. 22 (2002)
7158.[18] L. Deng, C. Wang, E. Spencer, L. Yang, A. Braun, J. You, C. Slaughter, C. Pickart, Z.J.
Chen, Cell 103 (2000) 351.[19] Y. Qian, M. Commane, J. Ninomiya-Tsuji, K. Matsumoto, X. Li, J. Biol. Chem. 276
(2001) 41661.[20] C.Wang, L. Deng, M. Hong, G.R. Akkaraju, J. Inoue, Z.J. Chen, Nature 412 (2001) 346.[21] R. Medzhitov, C.A. Janeway Jr, Semin. Immunol. 10 (1998) 351.
[22] D.T. Fearon, R.M. Locksley, Science 272 (1996) 50.[23] J. Cohen, Nature 420 (2002) 885.[24] S. Gottipati, N.L. Rao, W.P. Fung-Leung, Cell. Signal. 20 (2008) 269.[25] J. Hu, R. Jacinto, C. McCall, L. Li, J. Immunol. 168 (2002) 3910.[26] L. Li, S. Cousart, J. Hu, C.E. McCall, J. Biol. Chem. 275 (2000) 23340.[27] T.T. Yamin, D.K. Miller, J. Biol. Chem. 272 (1997) 21540.[28] K. Kobayashi, L.D. Hernandez, J.E. Galan, C.A. Janeway Jr, R. Medzhitov, R.A. Flavell,
Cell 110 (2002) 191.[29] K. Burns, J. Clatworthy, L. Martin, F. Martinon, C. Plumpton, B. Maschera, A. Lewis,
K. Ray, J. Tschopp, F. Volpe, Nat. Cell. Biol. 2 (2000) 346.[30] G. Zhang, S. Ghosh, J. Biol. Chem. 277 (2002) 7059.[31] D. De Nardo, T. Nguyen, J.A. Hamilton, G.M. Scholz, Cell. Signal. 21 (2009) 246.[32] S. Janssens, R. Beyaert, Mol. Cell. 11 (2003) 293.[33] Z. Cao, W.J. Henzel, X. Gao, Science 271 (1996) 1128.[34] D. De Nardo, P. Masendycz, S. Ho, M. Cross, A.J. Fleetwood, E.C. Reynolds, J.A.
Hamilton, G.M. Scholz, J. Biol. Chem. 280 (2005) 9813.[35] C. Kollewe, A.C. Mackensen, D. Neumann, J. Knop, P. Cao, S. Li, H. Wesche, M.U.
Martin, J. Biol. Chem. 279 (2004) 5227.[36] H. Ye, J.R. Arron, B. Lamothe, M. Cirilli, T. Kobayashi, N.K. Shevde, D. Segal, O.K.
Dzivenu, M. Vologodskaia, M. Yim, K. Du, S. Singh, J.W. Pike, B.G. Darnay, Y. Choi, H.Wu, Nature 418 (2002) 443.
[37] D.B. Conze, C.J. Wu, J.A. Thomas, A. Landstrom, J.D. Ashwell, Mol. Cell. Biol. 28(2008) 3538.
[38] W. Dong, Y. Liu, J. Peng, L. Chen, T. Zou, H. Xiao, Z. Liu,W. Li, Y. Bu, Y. Qi, J. Biol. Chem.281 (2006) 26029.
[39] D. Neumann, C. Kollewe, K. Resch, M.U. Martin, Biochem. Biophys. Res. Commun.354 (2007) 1089.
[40] K. Ross, L. Yang, S. Dower, F. Volpe, F. Guesdon, J. Biol. Chem. 277 (2002) 37414.[41] N. Suzuki, S. Suzuki, W.C. Yeh, Trends Immunol. 23 (2002) 503.[42] A. Didierlaurent, B. Brissoni, D. Velin, N. Aebi, A. Tardivel, E. Kaslin, J.C. Sirard, G.
Angelov, J. Tschopp, K. Burns, Mol. Cell. Biol. 26 (2006) 735.[43] B. Brissoni, L. Agostini, M. Kropf, F. Martinon, V. Swoboda, S. Lippens, H. Everett, N.
Aebi, S. Janssens, E. Meylan, M. Felberbaum-Corti, H. Hirling, J. Gruenberg, J.Tschopp, K. Burns, Curr. Biol. 16 (2006) 2265.
