Correspondence: Ken-ichi Tanamoto, Division of Microbiology, National Institute of Health Sciences, 1-18-1 Kamiyoga, Setagaya, Tokyo 158-8501, Japan. Tel.: +81 3 3700 1141, ext. 272; fax: +81 3 3707 6950; e-mail: email@example.com
IRAK-4 plays an essential role in Toll-like receptor (TLR)/IL-1 receptor signaling. However, its signaling and regulation mechanisms have remained elusive. We have reported previously that stimulation of TLR2, TLR4 or TLR9, but not TLR3, leads to downregulation of IRAK-4 protein. Here, we show that expression of MyD88 leads to downregulation of endogenous as well as exogenously expressed IRAK-4 protein in HEK293 cells. Expression of TRIF did not cause IRAK-4 downregulation although it induced NF-κB activation. Expression of either a deletion mutant of MyD88 lacking its death domain or MyD88s, neither of which induced NF-κB activation, did not lead to IRAK-4 downregulation. MyD88-induced downregulation was observed in an IRAK-4 mutant lacking the kinase domain, but not in another mutant lacking the death domain. These results demonstrate that downregulation of IRAK-4 requires activation of the MyD88-dependent pathway and that the death domains of both MyD88 and IRAK-4 are important for this downregulation.
Toll-like receptor (TLR)/IL-1 receptor (IL-1R) family members share common intracellular signaling proteins including MyD88, IL-1R-associated kinase (IRAK) family and TRAF6 (Fujihara et al., 2003; Janssens & Beyaert, 2003). Ligand binding to TLR/IL-1R triggers the recruitment of adaptor proteins, such as MyD88 and TRIF, to the Toll/IL-1 receptor (TIR) domain of TLR/IL-1R via a homophilic TIR–TIR interaction, which in turn recruits IRAK-4 and IRAK-1 into the receptor complex. IRAK-4 does not bind IRAK-1 directly but is recruited into the complex through binding with MyD88. This allows IRAK-1 and IRAK-4 to come in close proximity, which induces IRAK-4 to phosphorylate IRAK-1 (Li et al., 2002). The phosphorylated IRAK-1 interacts with TRAF6, leading to the activation of NF-κB (Cao et al., 1996).
IRAK-4 plays an essential role in TLR/IL-1R signaling. Residual activation of NF-κB in response to IL-1 is still observed in IRAK-1-deficient cells (Kanakaraj et al., 1998; Thomas et al., 1999). In contrast, almost no response to TLR/IL-1R stimulation is observed in IRAK-4-deficient mice (Suzuki et al., 2002) or in patients with IRAK-4 mutations (Picard et al., 2003; Medvedev et al., 2005). Although IRAK-4 is known to phosphorylate IRAK-1 (Li et al., 2002), the requirement of its kinase activity is still controversial (Lye et al., 2004; Qin et al., 2004). In addition, although the internal regions of MyD88 located between its C-terminal TIR and N-terminal death domains have been reported to be necessary for the interaction with IRAK-4 (Burns et al., 2003), the signaling mechanism and the regulation of IRAK-4 have remained elusive. We have reported previously that prolonged stimulation of TLR2, TLR4 or TLR9 leads to downregulation of IRAK-4 protein (Hatao et al., 2004). In this study, we found that expression of MyD88 led to downregulation of IRAK-4 and analyzed the structural requirements of IRAK-4 for this downregulation.
Materials and methods
Cell culture and reagents
The HEK293 cell line (obtained from the Human Science Research Resources Bank, Tokyo, Japan) was grown in Dulbecco's modified Eagle's medium (Invitrogen, Carlsbad, CA) supplemented with 10% (v/v) heat-inactivated fetal calf serum (Invitrogen), penicillin (100 U mL−1) and streptomycin (100 μg mL−1). An antiserum against EIAV-tag epitope (amino acid sequence: ADRRIPGTAEE) was a kind gift from Dr Nancy Rice (NCI-Frederick Cancer Research and Development Center). An antibody against β-actin (AC-15) was obtained from Sigma-Aldrich (St Louis, MO).
The coding region of human MyD88 was amplified by reverse transcriptase (RT)-PCR from total RNA prepared from human spleen (OriGene Technologies, Rockville, MD). Plasmids containing mouse IRAK-4 (IMAGE: 3995220) and human TRIF (IMAGE: 5180098) were obtained from the Mammalian Gene Collection. Mutations found in the IRAK-4 and TRIF plasmids were corrected by PCR-mediated mutagenesis. The coding regions of all constructs described above were subcloned into mammalian expression vectors containing the N-terminal EIAV-tag or the FLAG-tag sequence. NF-κB-dependent luciferase reporter plasmid pELAM-L was described previously (Muroi et al., 2002). All mutant plasmids were created by PCR-mediated mutagenesis and mutations were confirmed by DNA sequencing.
NF-κB reporter assay and immunoblotting
The NF-κB-dependent luciferase reporter assay was performed as described elsewhere (Muroi & Tanamoto, 2002). Briefly, HEK293 cells (2–5 × 105 cells) were plated in six-well plates and on the following day transfected by the calcium phosphate precipitation method with indicated plasmids, together with 0.2 μg of pELAM-L and 5 ng of phRL-TK (Promega, Madison, WI) for normalization. At 24–32 h after transfection, cellular extracts were prepared by adding a lysis buffer (10 mM HEPES–KOH, pH 7.9, 10 mM KCl, 5 mM EDTA, 40 mM β-glycerophosphate, 0.5% NP-40, 30 mM NaF, 1 mM Na3VO4, 1 mM DTT, 100 nM okadaic acid) containing a protease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany). The reporter gene activity was measured using a portion of the cellular extract according to the manufacturer's (Promega) instruction. Another portion of the cellular extract was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Proteins were transferred to a polyvinylidene difluoride membrane (Immobilon-P; Millipore, Bedford, MA) and subjected to immunoblotting with the indicated antibodies. The signals were visualized using an enhanced chemiluminescence system (Amersham Biosciences, Piscataway, NJ).
Expression of MyD88 leads to downregulation of IRAK-4
We have already reported that prolonged stimulation of TLR2, TLR4 or TLR9, but not TLR3, leads to downregulation of IRAK-4 protein (Hatao et al., 2004). It is well known that signaling through these TLRs, with the exception of TLR3, involves the adaptor protein MyD88, and expression of MyD88 leads to the activation of NF-κB. Therefore, we transfected an increasing amount of MyD88 expression plasmid together with an NF-κB-dependent reporter plasmid in HEK293 cells and detected IRAK-4 protein (Fig. 1). The protein level of endogenous IRAK-4 decreased in proportion to the amount of MyD88 plasmid transfected (top panel) although the level of β-actin was not affected (third panel from the top). Expression of either a dominant-negative mutant (MyD88-DN; amino acids 135–296) or a splicing variant (MyD88s; lacks amino acids 110–154) of MyD88, neither of which induced NF-κB activation (bottom panel), did not lead to the decrease in IRAK-4 protein (top panel), although these proteins were expressed properly (second panel from the top). To confirm this, we expressed EIAV-tagged IRAK-4 together with an increasing amount of MyD88, MyD88-DN or MyD88s and detected EIAV-tagged IRAK-4 with an anti-EIAV antibody (Fig. 1b). Exogenously expressed IRAK-4 was also decreased by the expression of MyD88 but not by MyD88-DN or MyD88s (top panel). In this experiment as well, the level of β-actin was not affected (third panel), and MyD88 and all of its mutants were expressed properly (second panel) with the activation of NF-κB by wild-type MyD88 only (bottom panel).
Because stimulation of TLR3 did not lead to downregulation of IRAK-4 (Hatao et al., 2004), and another adaptor protein TRIF is involved in TLR3-mediated signaling, we next asked whether expression of TRIF affects the level of IRAK-4 (Fig. 2). We transfected expression plasmids for EIAV-tagged IRAK4 and either MyD88 or TRIF together with an NF-κB-dependent reporter plasmid in HEK293 cells and detected IRAK-4 protein. Although expression of MyD88 and TRIF both activated NF-κB, only MyD88 induced downregulation of IRAK-4, indicating that downregulation of IRAK-4 requires the activation of the MyD88-dependent pathway. The level of β-actin was not affected by expression of these proteins.
Structural requirement of IRAK-4 for MyD88-induced downregulation
We next asked which domain of IRAK-4 is required to undergo MyD88-induced downregulation. IRAK-4 consists of an N-terminal death domain and a C-terminal kinase domain. We created deletion mutants lacking these domains and examined whether expression of MyD88 affects the level of these mutants (Fig. 3). Expression of MyD88 in HEK293 cells led to downregulation of wild-type IRAK-4 and the deletion mutant (dKD) of the kinase domain, indicating that the kinase domain is not necessary for downregulation. However, the IRAK-4 mutant (dDD) lacking the death domain did not undergo downregulation although the MyD88-induced activation of NF-κB was observed at a level comparable to the case when wild-type IRAK-4 was expressed (Fig. 3).
We have reported previously that prolonged stimulation of TLR2, TLR4 or TLR9 led to downregulation of IRAK-4 protein whereas the stimulation of TLR3 did not affect the IRAK-4 level (Hatao et al., 2004). MyD88 is involved in signaling through all of these TLRs except TLR3 (Fujihara et al., 2003). Thus, we examined the effect of expression of MyD88 on IRAK-4 protein level and found that expression of MyD88 leads to downregulation of endogenous as well as exogenously expressed IRAK-4 protein (Fig. 1). Our result clearly demonstrates that stimulation of TLRs is not necessary but expression of MyD88 is enough for downregulation of IRAK-4. We used three different IRAK-4 expression plasmids, in which expression of IRAK-4 is controlled by cytomegalovirus (CMV), herpes simplex virus (HSV) thymidine kinase and GAPDH promoters, and IRAK-4 levels expressed through these plasmids were all downregulated by MyD88 (data not shown). In addition, expression of endogenous IRAK-4 is regulated by the promoter different from those described above. It is, therefore, unlikely that the downregulation of IRAK-4 was caused by a decrease in IRAK-4 transcription.
We have reported previously that downregulation of IRAK-4 induced by prolonged stimulation of TLR seems to be mediated through cleavage of IRAK-4 by a protease induced by the activation of NF-κB (Hatao et al., 2004). In this study, MyD88-DN and MyD88s, neither of which induces NF-κB activation, did not lead to IRAK-4 downregulation. On the other hand, expression of TRIF induced a strong activation of NF-κB comparable to MyD88, but also did not cause downregulation (Fig. 2). Therefore, it appears that activation of NF-κB is not enough to induce downregulation of IRAK-4.
An IRAK-4 mutant lacking its death domain did not undergo downregulation, although strong activation of NF-κB was observed upon expression of MyD88 (Fig. 3). This may indicate that a protease that recognizes the death domain is responsible for the downregulation of IRAK-4. We have reported previously that downregulation of IRAK-4 in response to TLR stimulation was not inhibited by broad-spectrum caspase inhibitors (Z-VAD-FMK and A-Asp-CH2-DCB), a serine protease inhibitor (E-64) and a cathepsin B inhibitor (CA-074 methyl ester), and that the proteasome was unlikely to be involved because a smaller molecular weight protein (c. 32 kDa), which appears to be a cleavage product of IRAK-4, was detected by an IRAK-4 antibody following the decrease in IRAK-4 protein (Hatao et al., 2004). We also have not been able to find proteases that specifically recognize the death domain of IRAK-4 using a public protease database (PeptideCutter: http://www.expasy.org/tools/peptidecutter/). Thus, a novel protease or a protease whose recognition sequence has not been identified may be involved in the downregulation of IRAK-4.