Identification of interleukin-1 receptor-associated kinase 1 as a critical component that induces post-transcriptional activation of IκB-ζ

Authors

  • Tomoyuki Ohba,

    1.  Laboratory of Cell Recognition and Response, Department of Developmental Biology and Neurosciences, Graduate School of Life Sciences, Tohoku University, Sendai, Japan
    2.  Department of Biology, Faculty of Science, Tohoku University, Sendai, Japan
    Search for more papers by this author
  • Yujiro Ariga,

    1.  Laboratory of Cell Recognition and Response, Department of Developmental Biology and Neurosciences, Graduate School of Life Sciences, Tohoku University, Sendai, Japan
    Search for more papers by this author
  • Takashi Maruyama,

    1.  Laboratory of Cell Recognition and Response, Department of Developmental Biology and Neurosciences, Graduate School of Life Sciences, Tohoku University, Sendai, Japan
    2.  Department of Biology, Faculty of Science, Tohoku University, Sendai, Japan
    Search for more papers by this author
  • Nha K. Truong,

    1.  Department of Biology, Faculty of Science, Tohoku University, Sendai, Japan
    Search for more papers by this author
  • Jun-ichiro Inoue,

    1.  Division of Cellular and Molecular Biology, Department of Cancer Biology, Institute of Medical Science, University of Tokyo, Japan
    Search for more papers by this author
  • Tatsushi Muta

    1.  Laboratory of Cell Recognition and Response, Department of Developmental Biology and Neurosciences, Graduate School of Life Sciences, Tohoku University, Sendai, Japan
    2.  Department of Biology, Faculty of Science, Tohoku University, Sendai, Japan
    Search for more papers by this author

T. Muta, 6-3 Aoba, Aramaki, Aoba-ku, Sendai 980-8578, Japan
Fax: +81 22 795 6709
Tel: +81 22 795 6709
E-mail: tmuta@biology.tohoku.ac.jp

Abstract

IκB-ζ, an essential inflammatory regulator, is specifically induced by Toll-like receptor ligands or interleukin (IL)-1β by post-transcriptional activation mediated via a 165-nucleotide element in IκB-ζ mRNA. Here, we analyzed the Toll-like receptor–IL-1 receptor signaling components involved in the post-transcriptional regulation of IκB-ζ with mutated estrogen receptor [ER(T2)] fusion proteins. Upon 4-hydroxytamoxifen treatment, the ER(T2) fusion proteins with IL-1 receptor-associated kinase (IRAK)1 and IRAK4 elicited specific activation of a reporter gene for the post-transcriptional regulation of IκB-ζ. The tumor necrosis factor receptor-associated factor (TRAF)6–ER(T2) protein activated nuclear factor-κB, but not post-transcriptional regulation, indicating that activation of IRAK1/4, but not of TRAF6, is sufficient to activate the 165-nucleotide element-mediated post-transcriptional mechanism. Interestingly, the post-transcriptional mechanism was not activated in TRAF6-deficient cells, indicating an essential role for TRAF6. Thus, the signaling pathway leading to nuclear factor-κB activation and the post-transcriptional activation bifurcates at IRAK1, suggesting a new pathway activated by IRAK1.

Abbreviations
ARE

AU-rich element

EAE

experimental autoimmune encephalomyelitis

ER(T2)

a mutated hormone-binding domain of human estrogen receptor

ER

estrogen receptor

IL

interleukin

IL-1R

inteleukin-1 receptor

IRAK

interleukin-1 receptor-associated kinase

KO

knockout

LPS

lipopolysaccharide

MAPK

mitogen-activated protein kinase

MEF

mouse embryonic fibroblast

MyD

myeloid differentiation factor

NF

nuclear factor

OHT

4-hydroxytamoxifen

siRNA

small interfering RNA

TLR

Toll-like receptor

TNF

tumor necrosis factor

TNFR

tumor necrosis factor receptor

TRAF

tumor necrosis factor receptor-associated factor

Introduction

The regulation of gene expression by transcriptional activation of specific transcription factors has been studied extensively in the past. However, accumulating evidence indicates that post-transcriptional regulation is another important mechanism that can control gene expression [1,2]. In addition to the selection of genes required for activation, the regulation of strength and duration of gene expression is critical to elicit appropriate cellular responses. Such precise tuning is achieved by controlling the stability and translational efficiency of mRNA via specific cis-elements in mRNA in addition to regulating the activities of transcription factors [1,2].

Dynamic alterations in gene expression patterns are observed following activation of immune cells upon microbial infection. A series of immune reactions are initiated by recognition of microbial constituents by pattern recognition receptors, including Toll-like receptors (TLRs), NOD-like receptors, RIG-I-like receptors, and C-type lectin-like receptors [3]. Triggering of any of the pattern recognition receptors induces rapid activation of the transcription factors nuclear factor (NF)-κB and activator protein-1 without de novo synthesis. The activation of these transcription factors elicits the transcription of inflammatory genes, such as that encoding tumor necrosis factor (TNF)-α, within 1 h after the stimulation. In contrast to the situation with these primary response genes, the transcriptional activation of another group of inflammatory genes requires primary response gene products, as it is sensitive to protein synthesis inhibitors such as cycloheximide. These genes, collectively called the secondary response genes, are typically induced 2–6 h after the stimulation via chromosomal remodeling [4]. Thus, the gene expression profiles upon inflammation are programmed in a multistep cascade-like manner [5].

One of the primary gene products that allows the subsequent induction of the secondary response genes is IκB-ζ, which we identified as a gene product that is induced following lipopolysaccharide (LPS) stimulation of macrophages [6]. In cells lacking IκB-ζ, expression of a subset of the secondary response genes, represented by those encoding interleukin (IL)-6 or the IL-12 p40 subunit, is abolished or severely impaired [7]. IκB-ζ is a nuclear IκB family protein harboring the ankyrin repeats that preferentially interact with the NF-κB p50 subunit [8–10]. In addition to LPS, various TLR ligands, except those for TLR3, stimulate the induction of the gene encoding IκB-ζ [7,11]. IκB-ζ activates the transcription of the inflammatory genes via its association with the NF-κB p50 subunit [12,13], and is involved in the initiation of chromosomal remodeling, which is essential for the induction of secondary response genes [14,15]. In addition to these innate immune responses, recent studies have indicated that IκB-ζ plays essential roles in differentiation of Th17 cells, a helper T-cell subset that is critically involved in the pathogenesis of autoimmune diseases [16], as well as natural killer cell activation [17]. In fact, IκB-ζ mice are resistant to experimental autoimmune encephalomyelitis (EAE), an animal model for human multiple sclerosis [16]. IκB-ζ is also induced following B-cell receptor stimulation [18]. Despite these proinflammatory roles for IκB-ζ, IκB-ζ-deficient mice spontaneously develop chronic inflammation of the periocular skin [7,19], suggesting the presence of ‘homeostatic inflammation’ in the absence of infection.

Although NF-κB-mediated transcriptional activation is essential for its expression, robust induction of IκB-ζ requires activation of post-transcriptional mechanisms, including stabilization of its mRNA and translational activation [20,21]. The stabilization of IκB-ζ mRNA is specifically induced by TLR–IL-1 receptor (IL-1R) signaling, but not by another inflammatory signal activated by TNF-α. Thus, IκB-ζ is robustly induced by TLR–IL-1R stimulation, but not by TNF receptor (TNFR) stimulation [20]. This stimulus specificity provides a molecular mechanism that eventually allows stimulus-specific induction of secondary response genes.

Among various post-transcriptional mechanisms, one of the most extensively analyzed is that mediated via AU-rich elements (AREs). AREs are found in 3′-UTRs of mRNAs for many cytokines and growth factors, including granulocyte–macrophage colony-stimulating factor [22]. These mRNAs are unstable in unstimulated cells, but are stabilized following inflammatory stimulation. Stabilization of ARE-containing mRNAs has been shown to be mediated by p38 mitogen-activated protein kinase (MAPK) and MAPK-activated protein kinase 2 [23], and several trans-acting factors that bind to the element were shown to be involved in the regulation of its stability [1]. Although the 3′-UTR of IκB-ζ mRNA contains four ARE-like sequences, ARE-mediated post-transcriptional regulation does not show the same stimulus specificity as that observed with IκB-ζ mRNA. Several known trans-acting factors that act on AREs, including the RNA-binding protein HuR and tristetraprolin, do not appear to be involved in the regulation of IκB-ζ mRNA, and the induction of IκB-ζ mRNA was not affected by inhibitors of p38 MAPK [20]. Instead, a 165-nucleotide sequence in the 3′-UTR of IκB-ζ mRNA was identified as a cis-element that is essential and sufficient for the post-transcriptional regulation of IκB-ζ [24]. Mutations in the ARE-like sequence in the 165-nucleotide element did not affect the activity [24]. Therefore, the post-transcriptional regulatory mechanism that leads to IκB-ζ induction remains largely elusive.

Stimulation of the TLR–IL-1R signaling pathway elicits the recruitment of the adaptor protein myeloid differentiation factor (MyD)88 and IL-1R-associated kinase (IRAK)4 and IRAK1 to the receptors [3,25,26]. This recruitment activates IRAK4, which then phosphorylates IRAK1 to release it from the receptor complex [25,27–29]. IRAK1 activates TNFR-associated factor (TRAF)6, which acts as an E3 ubiquitin ligase that catalyzes Lys63-linked polyubiquitin chain formation on IRAK1 and TRAF6 itself [30,31]. The binding of TAK1-binding protein 1 and TAK1-binding protein 2 to polyubiquitin chains results in the activation of TAK1 [32]. TAK1 phosphorylates and activates IκB kinase and MAPK kinase, which culminates in the activation of NF-κB and MAPKs, respectively. TNFR stimulation also induces the activation of TAK1 via the E3 ubiquitin ligase TRAF2 and cellular inhibitor of apoptosis proteins, thus leading to the activation of NF-κB and MAPKs [33], but does not induce the stabilization of IκB-ζ mRNA [20].

In the present study, the TLR–IL-1R signaling pathway was analyzed to identify a signaling component(s) that induces the activation of the post-transcriptional mechanism for IκB-ζ induction. Because IκB-ζ is specifically induced by the TLR–IL-1R signaling pathway, but not that for TNFR, we focused on the TLR–IL-1R-specific signaling components MyD88, IRAK4, IRAK1, and TRAF6. In order to control the activity of components, they were expressed as mutated hormone binding domain of human estrogen receptor [ER(T2)] fusion proteins, whose activity is regulated by the addition of a small synthetic compound, 4-hydroxytamoxifen (OHT) [34].

Results

Robust induction of IκB-ζ requires transcriptional activation by NF-κB and activation of post-transcriptional regulation mechanisms [20,24]. In order to independently quantitate the two signals, we utilized two distinct reporters, the NF-κB reporter pELAM1-Luc [35], and the reporter for post-transcriptional regulation pGL4.12-SV40-Luc-165nt (the 165-nucleotide reporter) [24], which expresses a fusion mRNA encoding the firefly luciferase ORF followed by the 165-nucleotide element derived from the 3′-UTR of IκB-ζ mRNA via the SV40 promoter. The NF-κB reporter is activated following stimulation of either TNFR or TLR–IL-1R signaling, but the specific activation of the 165-nucleotide reporter is induced only by stimulation of the latter [24]. Post-transcriptional activation mediated by the 165-nucleotide element involves stabilization of IκB-ζ mRNA and the upregulation of translational efficiency [20,21,24], both of which result in activation of the 165-nucleotide reporter.

Both the NF-κB reporter and the 165-nucleotide reporter were activated following overexpression of MyD88, IRAK1, and TRAF6 (Fig. S1). However, activation was not observed following overexpression of IRAK4, which reportedly mediates the activation of IRAK1 downstream of MyD88 [3,27]. In view of the known signaling pathway from MyD88 to TRAF6, these results could possibly include artefacts caused by overexpression of the signaling components and/or sustained activation of the signaling pathway. Thus, in order to evaluate the activities of signaling components, we required a system in which the activity of reporters can be measured shortly after specific activation of the component.

Cells expressing the ER fusion protein of IRAK1, IRAK4 or TRAF6 exhibit NF-κB activation following OHT treatment

We noted that the activity of ER fusion proteins was suppressed in cells and that this suppression could be alleviated by the addition of ER agonists, such as OHT [34]. We thus created expression plasmid constructs for fusion proteins of the MyD88-dependent signaling component and a mutant of ER, ER(T2), which responds to OHT, but not estrogen (Fig. 1A). Four ER(T2) fusion proteins with MyD88, IRAK4, IRAK1 and TRAF6 were successively expressed in HEK293 cells, although the expression levels of IRAK4 and IRAK1 were lower than those of MyD88 and TRAF6 (Fig. 1B). Reporter analysis with pELAM1-Luc showed that NF-κB was activated even in the absence of OHT and slightly upregulated by the addition of OHT (Fig. 1C). The addition of OHT itself did not affect reporter activity.

Figure 1.

 The ER(T2) fusion proteins activate NF-κB. (A) A schematic illustration of ER(T2) fusion protein constructs. cDNA of interest C-terminally fused with ER(T2) together with a FLAG tag was cloned into the pEF6/V5-His A expression vector. (B) Expression of MyD88-ER(T2), IRAK4-ER(T2) and IRAK1-ER(T2) fusion proteins and TRAF6 in HEK293 cells. (C) NF-κB reporter analysis of HEK293 cells expressing the ER(T2) fusion proteins treated with or without OHT. Cells were transfected with an expression vector for the indicated ER(T2) fusion protein (B) together with the NF-κB reporter pELAM1-Luc and the internal control reporter phRL-TK (C). Forty-eight hours after transfection, the cells were lysed and analyzed by western blotting with antibody against FLAG (B) or treated with or without 0.2 μg·mL−1 OHT for 5 h and then lysed for the luciferase assay (C). Data show the mean ± standard error of duplicate transfection samples.

We suspected that activation of NF-κB following expression of the fusion proteins in the absence of OHT might be caused by overexpression of the fusion proteins. Therefore, we examined the effects of expression levels by reducing the amounts of transfected plasmids (Fig. 2). When serially diluted plasmids were transfected into cells, the NF-κB activation levels in the absence of OHT were significantly suppressed to basal levels in cells with reduced amounts of plasmids. Under these conditions, OHT treatment of cells with IRAK4, IRAK1 and TRAF6 fusion proteins significantly stimulated NF-κB activation. On the other hand, cells with MyD88 fusion proteins failed to exhibit OHT-dependent NF-κB activation under the same conditions. Essentially the same results were obtained with NIH3T3 cells (data not shown). We thus focused on IRAK4, IRAK1 and TRAF6 in the following analyses.

Figure 2.

 Cells transfected with reduced amounts of ER(T2) fusion proteins of IRAK4, IRAK1 or TRAF6 exhibit NF-κB activation in an OHT-dependent manner. NF-κB reporter analysis of HEK293 cells transfected with different amounts of the ER(T2) fusion proteins treated with or without OHT. The cells were transfected with serially diluted expression vectors for the indicated ER(T2) fusion proteins together with the NF-κB reporter pELAM1-Luc and the internal control reporter phRL-TK. Forty-eight hours after transfection, the cells were treated with or without 0.2 μg·mL−1 OHT for 5 h, and then lysed for the luciferase assay. Data show the mean ± standard error of duplicate transfection samples.

Activation of IRAK1 or IRAK4, but not TRAF6, is sufficient for post-transcriptional activation via the 165-nucleotide element of IκB-ζ

Upon stimulation of HEK293 cells, the NF-κB reporter (pELAM1-Luc) was activated by both IL-1β and TNF-α (Fig. 3A), whereas activation of the 165-nucleotide reporter was specifically induced by IL-1β (Fig. 3B). OHT treatment affected activation only marginally. The activity of the control reporter pGL4.12-SV40-Luc without the 165-nucleotide element showed weak induction by IL-1β and TNF-α, possibly caused by nonspecific augmentation of translational efficiency (Fig. 3C). Transfection with the fusion proteins IRAK4, IRAK1 and TRAF6 under the conditions in Fig. 2 resulted in the activation of NF-κB in an OHT-dependent manner (Fig. 3A). Significant activation of the 165-nucleotide reporter was also induced in cells expressing IRAK4 or IRAK1 fusion proteins. However, the TRAF6 fusion protein did not elicit activation of the 165-nucleotide reporter following OHT treatment under the same conditions (Fig. 3B). Thus, activation of either IRAK4 or IRAK1 is sufficient for activation of the 165-nucleotide element-mediated post-transcriptional mechanism as well as NF-κB activation. However, activation of TRAF6 alone is not sufficient for post-transcriptional activation, whereas it is sufficient for NF-κB activation.

Figure 3.

 OHT-mediated activation of ER(T2) fusion proteins of IRAK4 or IRAK1, but not of TRAF6, elicits 165-nucleotide element-mediated post-transcriptional activation. (A–C) Reporter analysis of HEK293 cells treated with IL-1β or TNF-α (upper panels) and cells expressing the ER(T2) fusion proteins treated with or without OHT (lower panels). The cells were transfected with pELAM1-Luc (A), pGL4.12-SV40-[luc2CP]-165nt (B), or pGL4.12-SV40-[luc2CP] (C), and the internal control reporter phRL-TK (upper), together with the indicated diluted expression plasmid for the ER(T2) fusion proteins (lower). Forty-eight hours after transfection, the cells were treated with or without 10 ng·mL−1 IL-1β or 10 ng·mL−1 TNF-α (upper), or 0.2 μg·mL−1 OHT (lower), for 5 h, and then lysed for the luciferase assay. Data show the mean ± standard error of duplicate transfection samples.

IRAK1 and TRAF6 are essential for IL-1β-mediated post-transcriptional activation of IκB-ζ

In order to examine whether IRAK1 is required for post-transcriptional activation, we knocked down IRAK1 by introducing small interfering RNA (siRNA) into HEK293 cells. Two siRNAs designed for IRAK1 (IRAK1-#1 and IRAK1-#2) suppressed the expression levels of endogenous IRAK1 protein to approximately 40% and 30%, respectively (data not shown). Upon stimulation with IL-1β, one of the two siRNAs reduced the activation of NF-κB (Fig. 4A). The active siRNA (IRAK1-#2), but not the other one, similarly inhibited IL-1β-mediated activation of the 165-nucleotide element-mediated post-transcriptional mechanism, indicating a critical role for IRAK1 (Fig. 4B).

Figure 4.

 IL-1β-mediated post-transcriptional activation via the 165-nucleotide element is inhibited by knockdown of IRAK1. HEK293 cells were transfected with siRNA for IRAK1 (IRAK1-#1 and IRAK1-#2) or control siRNA together with pELAM1-Luc (A) or pGL4.12-SV40-[luc2CP]-165nt (B), and the internal control reporter phRL-TK. Forty-eight hours after transfection, the cells were treated with or without 10 ng·mL−1 IL-1β for 5 h, and then lysed for the luciferase assay. Data show the mean ± standard error of duplicate transfection samples.

We next investigated whether TRAF6 is dispensable for post-transcriptional activation. We carried out the reporter analyses in wild-type and TRAF6-deficient [TRAF6 knockout (KO)] MEFs [36]. As in the literature, whereas both IL-1β and TNF-α stimulation activated NF-κB in wild-type MEFs, IL-1β-mediated NF-κB activation was abolished in the TRAF6 KO MEFs [37] (Fig. 5A). The activity of the 165-nucleotide reporter, but not that of the control reporter, was specifically upregulated following treatment with IL-1β, but not with TNF-α, in wild-type MEFs. However, IL-1β-mediated activation of the 165-nucleotide reporter was not observed in TRAF6-deficient MEFs, clearly indicating the requirement of TRAF6 for IL-1-induced 165-nucleotide element-mediated post-transcriptional activation (Fig. 5B,C). These results, together with those of Fig. 3, indicate that, although activation of TRAF6 alone does not elicit 165-nucleotide element-mediated post-transcriptional activation, it is an essential component for activation. When overexpressed in the mouse macrophage RAW264.7 cells, TRAF6 failed to activate the 165-nucleotide reporter, whereas it robustly activated the NF-κB reporter (Fig. S2), supporting this conclusion. Overexpressed IRAK1 activated both NF-κB and the 165-nucleotide reporters in RAW264.7 cells.

Figure 5.

 IL-1β-mediated post-transcriptional activation via the 165-nucleotide element is not observed in TRAF6-deficient cells. MEFs from wild-type (WT) or TRAF6 KO mice were transfected with pELAM1-Luc (A), pGL4.12-SV40-[luc2CP]-165nt (B), or pGL4.12-SV40-[luc2CP]-165nt (C), and the internal control reporter phRL-TK. Forty-eight hours after transfection, the cells were treated with or without 10 ng·mL−1 IL-1β or 10 ng·mL−1 TNF-α for 5 h, and then lysed for the luciferase assay. Data show the mean ± standard error of duplicate transfection samples.

Kinase activity of IRAK1, but not of IRAK4, is dispensable for post-transcriptional activation via the 165-nucleotide element IκB-ζ

Because both IRAK4 and IRAK1 are death domain-containing serine/threonine kinases, we examined whether the kinase activity of these proteins was required for activation. Kinase-dead mutants of IRAK1 (K239A/D340A) and IRAK4 (K213A/K214A) were expressed as ER(T2) fusion proteins at higher levels when transfected into HEK293 cells (Fig. 6A). The electrophoretic mobility of wild-type IRAK1 was slightly lower than that of the mutants, which probably reflects autophosphorylation of the wild type followed by degradation [38,39]. Cells expressing the fusion protein with the kinase-dead mutant of IRAK4 did not show NF-κB activation even after the addition of OHT, in agreement with previous studies using the mutant IRAK4 knock-in mouse [29,40] (Fig. 6B). In contrast, the IRAK1 mutant fusion protein strongly activated NF-κB following OHT treatment. The 165-nucleotide reporter, but not the control reporter, was similarly activated by the IRAK1 mutant following OHT treatment (Fig. 6C,D). On the other hand, the IRAK4 mutant did not stimulate 165-nucleotide reporter activity in the presence and absence of OHT. These results indicate that kinase activity of IRAK4 is required for activation of the 165-nucleotide element-mediated post-transcriptional mechanism, but that of IRAK1 is dispensable, suggesting a role for IRAK1 as an adaptor. We further mutated the identified ubiquitination sites on IRAK1, Lys134 and/or Lys180 [41], to arginine, and examined the effects of this with the same system. It was found that mutants of IRAK1 lacking either or both of Lys134 and Lys180 were capable of activating both the NF-κB reporter and the 165-nucleotide reporter in an OHT-dependent manner, indicating that ubiquitination of IRAK1 on these residues is dispensable for the activation of either pathway (Fig. S3).

Figure 6.

 OHT-mediated activation of the ER(T2) fusion protein of a kinase-dead mutant IRAK1, but not that of IRAK4, elicits 165-nucleotide element-mediated post-transcriptional activation. (A) Expression of IRAK1–ER(T2) fusion proteins, a kinase-dead mutant IRAK1 (IRAK1-KD), IRAK4, and a kinase-dead mutant IRAK4 (IRAK4-KD) in HEK293 cells. (B–D) Reporter analysis of HEK293 cells expressing the ER(T2) fusion proteins treated with or without OHT. The cells were transfected with an expression vector for the indicated ER(T2) fusion protein (A) together with pELAM1-Luc (B), pGL4.12-SV40-[luc2CP]-165nt (C), or pGL4.12-SV40-[luc2CP] (D), and the internal control reporter phRL-TK. Forty-eight hours after transfection, the cells were lysed and analyzed by western blotting with antibody against FLAG (A) or treated with or without 0.2 μg·mL−1 OHT for 5 h, and then lysed for the luciferase assay (B–D). Data show the mean ± standard error of duplicate transfection samples.

Discussion

In the present study, we found that the activation of IRAK1/4–ER (T2) fusion proteins is sufficient to trigger post-transcriptional regulation mediated by the 165-nucleotide element of IκB-ζ mRNA. This finding is consistent with the specific induction of IκB-ζ by stimulation of TLR–IL-1R but not by that of TNFR [6,20]. Although overexpression of a specific protein is a useful technique to analyze its molecular function, the results obtained with such experiments should be carefully interpreted, because they may include experimental artefacts. In particular, overexpression of a protein that activates a signaling pathway will result in the sustained activation of the pathway, which may induce secondary effects caused by positive and negative feedback mechanisms. In our analysis, overexpression of IRAK4 did not cause significant activation of NF-κB, which probably reflected the activation of a negative feedback mechanism. Therefore, in the current study, we utilized ER (T2) fusion proteins whose activity is suppressed until the addition of the activator OHT. With this system, we could chase the activation of the pathways by the target protein immediately after intracellular activation by OHT without affecting upstream signaling molecules. Careful titration of the expression levels minimized artefacts possibly caused by overexpression. As a result, the IRAK4 fusion protein activated NF-κB in an OHT-dependent manner, in contrast to what was found in the overexpression experiment.

Many previous studies have indicated that IRAK1 is phosphorylated and activated by IRAK4 [27–29,42]. Thus, IRAK1 is more likely to be closely associated with activation of the post-transcriptional regulation of IκB-ζ. Recent studies have also shown that IRAK1 is important in IL-1α-mediated activation of post-transcriptional regulation [21,43]. Because IL-1-mediated stabilization of CXCL1 (KC) mRNA requires p38 MAPK [44,45], but the induction of IκB-ζ does not [11,20], IRAK1 may activate two distinct post-transcriptional mechanisms on different genes.

Upon stimulation of TLR–IL-1R, IRAK1 is recruited, via its death domain, to the receptor complex containing MyD88 and IRAK4 together with TRAF6 [3,25]. IRAK1 then becomes a target of multiple modifications, such as phosphorylation, ubiquitination, and sumoylation [42]. IRAK4 phosphorylates Thr209 and then Thr387 in the activation loop of IRAK1 [27,28]. The phosphorylation activates IRAK1 and induces its autophosphorylation, which allows dissociation from the receptor complex followed by activation of downstream signaling components, and eventually induces its ubiquitin-mediated degradation [25,28,39]. We found that kinase activity of IRAK4 was essential for both NF-κB activation and 165-nucleotide element-mediated post-transcriptional regulation. This finding supports our hypothesis that IRAK1 phosphorylation by IRAK4 is critical for post-transcriptional activation in the signaling cascade. In macrophages from kinase-dead IRAK4 knock-in mice, the stabilization of various cytokine and chemokine mRNAs following TLR–IL-1R ligation is severely impaired [40]. On the other hand, our observation that the kinase-dead form of the IRAK1 fusion protein efficiently activated the 165-nucleotide reporter indicates that the kinase activity of IRAK1 is not required, and hence it acts as an adaptor for activation of the post-transcriptional mechanism, as in the case of NF-κB activation [46,47]. The kinase activity of IRAK1 has been shown to be essential for induction of type I interferon upon TLR7/9 stimulation via activation of interferon regulatory factor 7, but not for activation of NF-κB and MAPKs [48].

It is a notable finding that the activation of TRAF6, a well-known target of IRAK1, is sufficient for NF-κB activation but not for activation of 165-nucleotide element-mediated post-transcriptional regulation. Because IRAK1 activation, but not TRAF6 activation, is sufficient for activation of 165-nucleotide element-mediated post-transcriptional regulation, the signal bifurcates at IRAK1 in the signaling cascade following TLR–IL-1R ligation. Unexpectedly, however, the post-transcriptional mechanism was not activated in TRAF6-deficient cells following IL-1β stimulation, indicating that TRAF6 is one of the essential components for 165-nucleotide element-mediated post-transcriptional activation in TLR–IL-1R signaling. This is in sharp contrast to IL-1α-induced or IL-17-induced post-transcriptional activation of CXCL1 mRNA, which does not require TRAF6 [43,49]. A recent report indicated that IL-17 stimulates post-transcriptional activation of CXCL1 mRNA via TRAF5, TRAF2, and the splicing regulatory factor SF2 (alternative splicing factor), and suggested that a similar mechanism acts on mRNAs for colony-stimulating factor 3 and IκB-ζ [50]. Consistent with our observation, although TRAF6 is activated by IL-17, overexpression of TRAF6 did not stabilize CXCL1 mRNA [50]. As TRAF6 is essential for IL-1β-mediated post-transcriptional activation of IκB-ζ, different TRAFs appear to be used to activate the post-transcriptional mechanism in response to different stimuli or for different mRNAs.

Because the kinase activity of IRAK1 is dispensable, IRAK1 act as an adaptor in post-transcriptional activation. Although one of the targets of IRAK1 is TRAF6, activation of TRAF6 is not sufficient for post-transcriptional activation, indicating that both IRAK1 and TRAF6 constitute the machinery needed to activate 165-nucleotide element-mediated post-transcriptional regulation. The complex of IRAK1 and TRAF6 may recruit another molecule that is involved in the regulation of mRNA degradation, such as SF2. In addition, because TRAF6 acts as an E3 ubiquitin ligase, it is possible that ubiquitination of IRAK1 by TRAF6 is critically involved in activation. However, our analysis with mutants modified at the identified ubiquitination sites on IRAK1, Lys134 and Lys180 [41] failed to support the critical role of these amino acids in either NF-κB or post-transcriptional activation: the IRAK1 mutants that lacked either or both of Lys134 and Lys180 exhibited robust activation of both NF-κB and the 165-nucleotide reporter. Alternatively, ubiquitination of other amino acids may be critical for activation of 165-nucleotide element-mediated post-transcriptional regulation, or there may be an unknown ubiquitination-independent mechanism that operates through the IRAK1–TRAF6 complex. A recent report has shown that viperin-mediated IRAK1 ubiquitination is critical for type I interferon production in plasmacytoid dendritic cells, but that ubiquitination is dispensable for the induction of cytokine and interferon in non-plasmacytoid dendritic cells [51].

Because this study was mainly carried out by activating specific molecules, we could not precisely address the quantitative significance of IRAK1/4 in 165-nucleotide element-mediated post-transcriptional activation. In fact, NF-κB is activated in response to a TLR2 ligand in cells lacking both IRAK4 and IRAK1, and IκB-ζ induction is attenuated but is still observed in IRAK4-deficient cells, suggesting the involvement of other molecules [29]. Nevertheless, this study demonstrates the need for TRAF6 in the pathway, and indicates the critical roles of molecules that activate TRAF6. IRAK1 and IRAK2 have been reported to act redundantly in early and late TLR–IL-1R signaling, respectively, and both of them are necessary for robust cytokine production [52,53]. Thus, IRAK2 may also play a role in regulating the post-transcriptional mechanism, although the knockdown experiment of Fig. 4 indicated an important role for IRAK1 in post-transcriptional regulation. It has been reported that, upon LPS stimulation, CXCL1 and IL-6 mRNAs exhibit reduced stability and are more abundant in the translation-inactive state in IRAK2-deficient macrophages than in the wild type [53]. However, the stability of IκB-ζ mRNA following stimulation of TLR2 was not affected in IRAK2-deficient macrophages, possibly because of the presence of IRAK1 [52].

IκB-ζ is essential for TLR–IL-1R-mediated induction of IL-6 and IL-12, and IL-17 expression in Th17 cells [7,16]. These cytokines are not only critical for normal immune responses, but are also involved in the development of various autoimmune diseases, such as rheumatoid arthritis. IRAK1-deficient mice are resistant to EAE, a mouse model for human multiple sclerosis [54], and this may be explained at least in part by defective induction of IκB-ζ. IRAK1 is activated not only by TLR–IL-1R stimulation but also by IL-18 and IL-33 signaling. Moreover, mice deficient in IL-1R and IL-18 are also resistant to EAE, demonstrating critical roles of these signaling processes in disease development [55,56]. Further investigations on the post-transcriptional regulation of IκB-ζ would provide more insights into sophisticated mechanisms for immune regulation, and may lead to the development of new therapeutic approaches for various diseases caused by disorders of inflammatory reactions.

Experimental procedures

Cells and reagents

MEFs from C57BL/6 and TRAF6−/− mice [36] were prepared from embryonic day 14.5 embryos. HEK293 cells and MEFs were cultured in DMEM containing 10% (v/v) heat-inactivated fetal bovine serum, 100 U·mL−1 penicillin, and 100 μg·mL−1 streptomycin. OHT was purchased from Sigma-Aldrich (St Louis, MO, USA). Recombinant human IL-1β and TNF-α were obtained from Jena Bioscience GmbH (Jena, Germany). Duplexed modified RNA oligonucleotides (Stealth RNAi) for IRAK1 (IRAK1-#1, Catalog no. HSS105489; and IRAK1-#2, Catalog no. HSS179957) and control (medium GC) were purchased from Invitrogen (Carlsbad, CA, USA).

Plasmids

A cDNA fragment encoding ER(T2) [34] with a C-terminal FLAG-tag was generated by PCR and inserted into the expression vector pEF6/V5-His A (Invitrogen). cDNAs for human MyD88, IRAK1, IRAK4 and TRAF6 were obtained by PCR and subcloned into the 5′-end of ER(T2) in the vector. Kinase-dead forms of mutants IRAK1 and IRAK4 were created by introducing the mutations K239A/D340A for IRAK1 [46,47] and K213A/K214A for IRAK4 [29] by PCR. The sequences of all PCR products were verified by sequencing. The NF-κB reporter plasmid pELAM1-Luc [35] and pGL4.12-SV40-[luc2CP]-165nt, harboring the 165-nucleotide element of IκB-ζ mRNA, have been described previously [24].

Transfection

Transfection of HEK293 cells with plasmids was carried out essentially by the calcium phosphate method [57]. Briefly, plasmid DNA was added to 0.25 m CaCl2 and then mixed with an equal volume of 2× Bes-buffered saline (50 mm Bes, 280 mm NaCl, and 1.5 mm Na2HPO4, pH 6.95). The amount of plasmid DNA was kept constant by addition of an empty vector. The mixture was incubated for 15 min at room temperature, and then added in a dropwise fashion to the cells. Following incubation at 37 °C under 5% CO2 for 24 h, the medium was changed, and the cells were incubated for a further 24 h before analysis. siRNA (2.4 pmol) was transfected into HEK293 cells in a 48-well plate with Lipofectamine RNAiMAX (Invitrogen). Following incubation at 37 °C under 5% CO2 for 48 h, the cells were transfected with 54 ng of pELAM1-Luc or pGL4.12-SV40-[luc2CP]-165nt together with 6 ng of phRL-TK plasmid as an internal control, with the calcium phosphate precipitation method, and the cells were incubated for a further 24 h before analysis. MEFs were transfected with Lipofectamine LTX (Invitrogen).

Western blotting

Cells were lysed with 2× Laemmli’s sample buffer containing 2% (v/v) β-mercaptoethanol [58], subjected to 7.5% (w/v) SDS/PAGE, and transferred to an Immobilon-P Transfer Membrane (Millipore, Billerica, MA, USA). The membrane was blocked with Blocking One (Nacalai Tesque, Kyoto, Japan), and then incubated with anti-FLAG M2–horseradish peroxidase (Sigma-Aldrich). After extensive washing, reacting bands were visualized by chemiluminescence after incubation with Immobilon Western Chemiluminescent horseradish peroxidase Substrate (Millipore) and exposure of the filter to an X-ray film. The data shown are representative of at least two independent experiments.

Reporter assay

Two days after transfection, the cells were stimulated with 10 ng·mL−1 IL-1β or TNF-α, or 0.2 μg·mL−1 OHT, at 37 °C for 5 h. Luciferase activities were measured with the Dual-Luciferase Reporter Assay System, according to the manufacturer’s instructions (Promega, Madison, WI, USA). The reporter activities were normalized by using Renilla luciferase activities derived from the internal control phRL-TK. The data shown are representative of at least two independent experiments.

Acknowledgements

We thank to T. Kawasaki and Y. Kawasaki for technical assistance. This work was supported in part by KAKENHI (21117004 and 21390088) (to T. Muta) and (21200040) (to T. Ohba), and grants (to T. Muta) from the Takeda Science Foundation, the Naito Foundation, the Asahi Grass Foundation, the Suzuken Memorial Foundation, the Novartis Foundation for the Promotion of Science, and the Japan Foundation for Applied Enzymology.

Ancillary