TLR8-mediated NF-κB and JNK Activation Are TAK1-independent and MEKK3-dependent*

TLR8-mediated NF-κB and IRF7 activation are abolished in human IRAK-deficient 293 cells and IRAK4-deficient fibroblast cells. Both wild-type and kinase-inactive mutants of IRAK and IRAK4, respectively, restored TLR8-mediated NF-κB and IRF7 activation in the IRAK- and IRAK4-deficient cells, indicating that the kinase activity of IRAK and IRAK4 is probably redundant for TLR8-mediated signaling. We recently found that TLR8 mediates a unique NF-κB activation pathway in human 293 cells and mouse embryonic fibroblasts, accompanied only by IκBα phosphorylation and not IκBα degradation, whereas interleukin (IL)-1 stimulation causes both IκBα phosphorylation and degradation. The intermediate signaling events mediated by IL-1 (including IRAK modifications and degradation and TAK1 activation) were not detected in cells stimulated by TLR8 ligands. TLR8 ligands trigger similar levels of IκBα phosphorylation and NF-κB and JNK activation in TAK1-/- mouse embryo fibroblasts (MEFs) as compared with wild-type MEFs, whereas lack of TAK1 results in reduced IL-1-mediated NF-κB activation and abolished IL-1-induced JNK activation. The above results indicate that although TLR8-mediated NF-κB and JNK activation are IRAK-dependent, they do not require IRAK modification and are TAK1-independent. On the other hand, TLR8-mediated IκBα phosphorylation, NF-κB, and JNK activation are completely abolished in MEKK3-/- MEFs, whereas IL-1-mediated signaling was only moderately reduced in these deficient MEFs as compared with wild-type cells. The differences between IL-1R- and TLR8-mediated NF-κB activation are also reflected at the level of IκB kinase (IKK) complex. TLR8 ligands induced IKKγ phosphorylation, whereas IKKα/β phosphorylation and IKKγ ubiquitination that can be induced by IL-1 were not detected in cells treated with TLR8 ligands. We postulate that TLR8-mediated MEKK3-dependent IKKγ phosphorylation might play an important role in the activation of IKK complex, leading to IκBα phosphorylation.

Recent studies have begun to unravel how a subset of TLRs, TLR7, TLR8, and TLR9, employ a novel MyD88-dependent pathway to mediate the activation of transcription factors NF-B, IRF5, and IRF7 and induction of interferon-␣. Whereas TLR9 has been shown to recognize bacterial DNA, a synthetic compound (imidazoquinoline compound R848) with antiviral activity and single-stranded RNA have been described as ligands for TLR7 and TLR8. It has been recently reported that the transcription factor IRF7 interacts with MyD88 to form a complex in the cytoplasm, and this interaction resulted in activation of IFN␣-dependent promoters (26,27). IRAK4 and TRAF6 have also been implicated in this pathway, and ubiquitin ligase activity of TRAF6 has been shown to mediate IRF7 activation. The detailed molecular mechanisms for this novel TLR7-, TLR8-, and TLR9-mediated MyD88-dependent pathway are still unclear.
In this paper, we report that TLR8 mediate a unique NF-B activation pathway. The intermediate signaling events mediated by IL-1 (including IRAK modifications and degradation and TAK1 activation) were not detected in cells stimulated by TLR8 ligands. Using IRAK-and TAK1-deficient cells, we found that although TLR8-mediated NF-B and JNK activation are IRAK-dependent, they do not require IRAK modification and are TAK1-independent. On the other hand, TLR8-mediated IB phosphorylation and NF-B and JNK activation are completely abolished in MEKK3 Ϫ/Ϫ MEFs, whereas IL-1-mediated signaling was only moderately reduced in these deficient MEFs as compared with wild-type cells. The differences between IL-1R-and TLR8-mediated NF-B activation are also reflected at the level of the IKK complex. TLR8 ligands induced IKK␥ phosphorylation, whereas IKK␣/␤ phosphorylation and IKK␥ ubiquitination that can be induced by IL-1 were not detected in cells treated with TLR8 ligands. We postulate that TLR8-mediated MEKK3-dependent IKK␥ phosphorylation might play an important role in the activation of IKK complex, leading to IB␣ phosphorylation and NF-B activation.
Recombinant Plasmids and Transfection-Endothelial cell leukocyteadhesionmolecule-1promoter-derivedluc,anNF-Bdependent luciferase reporter plasmid, was described by Schindler and Baichwal (30). Interferon-␣2 (IFN␣2) or interferon-␣4 (IFN␣4)-dependent-luciferase reporter constructs were generated by PCR and inserted into pBasicLuc. The human TLR7/8 and IKK␥ expression constructs were created by inserting cDNA fragments, amplified by PCR, into the pcDNA3.1 expression vector. Wild-type, kinase-inactive (KK213AA, two lysine residues in the ATP binding pocket were mutated to alanine, abbreviated as IRAK4mt) IRAK-4 cDNA was cloned into the retrovirus vector, pBabe-puro. Mammalian expression vectors encoding wild-type and kinase-inactive (K293A) IRAK (driven by the thymidine kinase promoter) were described elsewhere (12). Transfection of the indicated plasmids by FuGENE 6 transfection reagents was done as recommended by the manufacturer (Roche Applied Science). Transfection solution was prepared by mixing 1 g of plasmid DNA and 3 l of FuGENE 6 transfection reagent in 100 l of serum-free medium. After incubation at room temperature for 15 min, the mixture was added to tissue culture wells containing 1 ϫ 10 6 cells in 2 ml of complete culture medium. MEKK3 siRNA sequence 5Ј-GAT-CCCCGCCTTAGGATATTGCTGTTTTCAAGAGAAACA-GCAATATCCTAAGGCTTTTTGGAAA-3Ј was cloned into vector pSUPER, which was obtained from Dr. Reuven Agami's group (Center for Biomedical Genetics, The Netherlands) (31).
Luciferase Reporter Assays-Cells (2 ϫ 10 5 were transiently transfected using FuGENE 6 (Roche Applied Science), following the manufacturer's protocol. Cells were transfected with the indicated expression vectors plus 100 ng of the luciferase reporter plasmids and 10 ng of ␤-galactosidase plasmid for normalization, with a 1:3 ratio of DNA/FuGENE 6. Transfection of empty vector was used to ensure that all samples received equal amounts of DNA. At 36 h after transfection, cells were stimulated with ligands for 6 h. Cells were lysed, and luciferase activity was assessed using reporter lysis buffer and luciferase assay reagent (Promega). All results reported represent duplicate experiments with at least three independent transfections.
Western Blot Analysis-Cells that were not treated or treated with IL-1 (10 ng/ml), 3M-02 (5 g/ml), or 3M-03 (5 g/ml) were harvested, washed in cold PBS buffer, pelleted, and lysed in ice-cold lysis buffer (30 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% Nonidet P-40, and 1 mM phenylmethylsulfonyl fluoride) for 30 min. Cell debris was pelleted by centrifugation for 10 min at 13,000 ϫ g. Supernatants were separated on 10% SDS-PAGE, transferred to supported nitrocellulose membrane (Millipore), and blocked in a 5% solution of nonfat dry milk prepared in 1ϫ PBS and 0.05% Tween 20. Blots were incubated with primary antibody diluted in PBS overnight at 4°C, washed three or four times for 10 min each with PBS, and then detected with horseradish peroxidase-conjugated secondary antibody diluted 1:5000 in PBS plus 5% nonfat milk and developed using the enhanced chemiluminescence method (ECL Plus; Amersham Biosciences) following the manufacturer's protocol.
To determine whether such differential signaling outcomes are only restricted to 293 cells, we isolated primary mouse kidney epithelial cells. As shown in Fig. 1d, both IL-1 and imidazoquinoline compounds induced the activation of NF-B, JNK, ERK, and p38 in the primary kidney epithelial cells. However, whereas IL-1 stimulation induced both IB␣ phosphorylation and degradation, IB␣ was phosphorylated but not degraded upon imidazoquinoline compounds (3M-002 and 3M-003) in these primary cells. Since the murine TLR8 gene is not functional, the imidazoquinoline-induced signaling in mouse kidney cells is probably through TLR7. Taken together, these results suggest that TLR7/8 and IL-1R might mediate differential signaling pathways to lead to NF-B activation.
TLR8 Do Not Lead to IRAK Modification and Degradation-To further compare intermediate signaling events mediated by TLR8 and IL-1R, we employed MyD88-deficient cells (I3A), IRAK-deficient cells (I1A), and IRAK4-deficient cells. Similar to IL-1 signaling (12,15,32,33), MyD88, IRAK, and IRAK4 are also required for TLR8-mediated activation of NF-B-dependent (E-selectin; Fig. 2, a, c, and d) and IRF5/7-dependent (IFN␣2; Fig. 2, b-d) promoters (data not shown). We previously showed that IRAK and IRAK4 kinase-inactive mutants had the same ability as the wild-type IRAK and IRAK4 in restoring IL-1-mediated signaling in IRAK-and IRAK4-deficient cells, respectively, indicating that the kinase activity of IRAK4 is not necessary for the IL-1 pathway (12,33). The fact that only the impairment of the kinase activity of both IRAK and IRAK4 efficiently abolished the IL-1 pathway suggests that the kinase activity of IRAK and IRAK4 is probably redundant for IL-1-mediated signal-TLR8-mediated Signaling JULY 28, 2006 • VOLUME 281 • NUMBER 30 JOURNAL OF BIOLOGICAL CHEMISTRY 21015 ing (33). As shown in Fig. 2, c and d, both wild-type and kinase-inactive IRAK and IRAK4 can also restore TLR7/8mediated activation of NF-B-dependent (E-selectin) and IRF5/7-dependent (IFN␣2) promoters in IRAK-and IRAK4deficient cells, respectively. These results indicate that whereas IRAK and IRAK4 are necessary for TLR8 signaling, the individual kinase activity of IRAK and IRAK4 is not required, implicating potential redundancy of the kinase activity of these two kinases in TLR8-mediated signaling ( Fig. 2 and data not shown).

TLR8-mediated Signaling
flanked (floxed) Map3k7 flox/flox ), and confirmed by genomic Southern, reverse transcription-PCR, and Western analysis (21). To further investigate the role of TAK1 in TLR7/8-mediated signaling, we transfected human TLR8 into TAK1 ϩ/ϩ (Map3k7 flox/flox ) and TAK1 Ϫ/Ϫ (Map3k7 Ϫ/Ϫ ) MEFs (21). As shown in Fig. 3b, IL-1-induced NF-B and JNK activation are greatly reduced in TAK1 Ϫ/Ϫ MEFs as compared with that in TAK1 ϩ/ϩ MEFs. However, NF-B and JNK activation were FIGURE 2. TLR8 does not lead to IRAK modification and degradation. a and b, MyD88 and IRAK are required for TLR8 signaling. a, 293, I1A, or I3A cells were co-transfected transiently with TLR8 as well as E-selectin-luc with increasing amounts of human MyD88 or IRAK1 expression constructs. Thirty-six hours later, the cells were either untreated or stimulated for 6 h with 3M-002 (5 g/ml) and 3M-003 (5 g/ml), and luciferase levels were measured. b, I1A or I3A cells were transiently transfected as in Fig. 3a, including IFN-␣2-or IFN-␣4-dependent luc. IFN-␣2-or IFN-␣4-induced luciferase activity in cells treated with 3M-02 (5 g/ml) or untreated cells was measured. Data represent mean relative induction of luciferase activity Ϯ S.D. for one of three independent experiments, each done in duplicate. c and d, the kinase activities of IRAK1 and IRAK4 are not required for TLR8 signaling. IRAK1-deficient I1A cells (c) and human IRAK4-deficient fibroblasts (d) were co-transfected transiently with TLR8 as well as E-selectin-luc or IFN-␣2-dependent luc with increasing amounts of human wild-type or kinase-inactive IRAK1 or IRAK4 expression constructs. Thirty-six hours later, the cells were either untreated or stimulated for 6 h with 3M-02 (5 g/ml). Luciferase activities were normalized to ␤-galactosidase. Data are presented as -fold induction of luciferase activity in the treated cells. The experiments were repeated four times. Shown are the data from a typical experiment. e, TLR8 does not lead to IRAK modification and degradation. 293 cells transfected with TLR8 were either left untreated or treated (times shown above blot) with IL-1 (10 ng/ml), 3M-002 (5 g/ml), or 3M-003 (5 g/ml). The cells were then lysed and analyzed by immunoblot with antibodies against IRAK1 and actin. f, extracts of 293 cells with or without IL-1 stimulation were either untreated or treated with calf intestinal phosphatase (CIP). These lysates were then analyzed by Western blots with anti-IRAK, anti-TAK1, and anti-actin. JULY 28, 2006 • VOLUME 281 • NUMBER 30 induced to similar levels by imidazoquinoline compounds in TAK1 ϩ/ϩ and TAK1 Ϫ/Ϫ MEFs (Fig. 3b). These results clearly indicate that TAK1 is not required for TLR8-mediated NF-B and JNK activation.
Imidazoquinoline-induced IB␣ phosphorylation (Fig. 4b), NF-B and JNK activation (Fig. 3b) were completely abolished in MEKK3-deficient MEFs as compared with that in wild-type control cells. The lack of MEKK3 in MEKK3-deficient MEFs was confirmed by Western analysis (data not shown). These results clearly indicate that MEKK3 is required for TLR8-mediated signaling. The TLR8-mediated signaling was diminished in 293-IL-1R-TLR8 cells where MEKK3 was knocked down by siRNA, confirming the role of MEKK3 in TLR8 signaling (Fig. 3c).
The next question is whether the kinase activity of IKK is activated at all upon activation of TLR8. As shown in Fig. 4, e and f, imidazoquinoline compounds can activate IB kinase in 293-IL-1R-TLR8 cells at similar levels as IL-1 stimulation. Consistent with this finding, imidazoquinoline-induced IB phosphorylation was abolished in IKK␣/␤ Ϫ/Ϫ MEFs (29), whereas JNK activation was intact in these cells. The lack of IKK␣ and IKK␤ in these IKK␣/␤-deficient MEFs was confirmed by Western analysis (data not shown). Taken together, these results suggest that imidazoquinoline-induced NF-B activation is IKK-dependent, although the mechanism of FIGURE 3. TLR8-mediated signaling is TAK1-independent and MEKK3-dependent. a, TLR8 does not lead to TAK1 and TAB1 phosphorylation. 293 cells transfected with TLR8 were either left untreated or treated (times shown above blot) with IL-1 (10 ng/ml), 3M-002 (5 g/ml), or 3M-003 (5 g/ml). The cells were then lysed and analyzed by immunoblot with antibodies against TAK1, TAB1, or actin. b and c, TLR8-mediated signaling is TAK1-independent and MEKK3-dependent. b, wild-type (ϩ/ϩ), TAK1-deficient (Ϫ/Ϫ), and MEKK3-deficient (Ϫ/Ϫ) MEFs transiently transfected with human TLR8 were stimulated with IL-1 (10 ng/ml) or 3M-002 (5 g/ml) for the indicated durations. Nuclear extracts were prepared, and NF-B DNA binding activity was determined by an electrophoretic mobility shift assay using a probe specific for NF-B. Whole-cell lysates were prepared and subjected to Western blot analysis using antibodies specific for phospho-JNK, JNK, and FLAG. c, 293-IL-1R-TLR8 cells were co-transfected with E-selectin-Luc with pSuper MEKK3 siRNA or pSuper vector control to knock down the expression of MEKK3. Thirty-six hours later, the cells were either untreated or stimulated for 6 h with 3M-002 (5 g/ml) or 3M-003 (5 g/ml), and luciferase levels were measured. The levels of MEKK3 in the transfected cells were measured by Western analysis with anti-MEKK3 antibody.

DISCUSSION
We here report that TLR8 mediates a novel MyD88-dependent NF-B activation pathway that is significantly different from the classical MyD88-dependent NF-B pathway mediated by IL-1R. One striking difference between these two pathways is IRAK modification. IL-1 stimulation leads to IRAK phosphorylation, ubiquitination, and modification. IRAK modifications and degradation were not detected in cells stimulated by TLR8 ligands, although TLR8-mediated NF-B activation is IRAKdependent. It is possible that additional adaptor molecules are associated with IL-1R and/or TLR8, influencing IRAK modification upon ligand stimulation.

TLR8-mediated Signaling
the fact that TLR8 does not lead to IRAK modification, imidazoquinoline treatment failed to activate TAK1. Using TAK1deficient MEFs, we found that TLR8-mediated NF-B and JNK activation are TAK1-independent. Taken together, our results indicate that the unmodified IRAK in TLR8 pathway probably does not utilize TAK1 to mediate downstream signaling. On the other hand, TLR8-mediated IB␣ phosphorylation and NF-B and JNK activation are completely abolished in MEKK3 Ϫ/Ϫ MEFs, whereas IL-1-mediated signaling was only moderately reduced in these deficient MEFs as compared with wild-type cells. Therefore, instead of TAK1, the unmodified IRAK in the TLR8 pathway probably utilizes MEKK3 to mediate NF-B and JNK activation.
The differences between IL-1R-and TLR8-mediated pathways are also reflected at the level of the IKK complex. TLR8 ligands induced IKK␥ phosphorylation, whereas IKK␣/␤ phosphorylation and IKK␥ ubiquitination that can be induced by IL-1 were not detected in cells treated with TLR8 ligands. Previous studies suggest that ligand-induced phosphorylation of IKK␣/␤ plays a critical role in IKK activation (41). However, although imidazoquinoline treatment does not lead to IKK␣/␤ phosphorylation, TLR8 ligands still activate the kinase activity of IKK, indicating that IKK is activated through a differential mechanism in TLR8 signaling. Using MEKK3-deficient MEFs, MEKK3 is shown to be required for TLR8-induced IKK␥ phosphorylation. We postulate that TLR8-mediated MEKK3dependent IKK␥ phosphorylation might play an important role in the activation of IKK complex, leading to IB phosphorylation and NF-B activation.
The dogma for NF-B activation is that signal-induced phosphorylation of IB␣ targets this inhibitor of NF-B for ubiquiti-nation and subsequent degradation, thus allowing NF-B to enter the nucleus to turn on the target gene (42). One intriguing finding is that whereas IL-1 stimulation leads to IB␣ phosphorylation and degradation, imidazoquinoline treatment only causes IB␣ phosphorylation but not degradation. By NF-B DNA binding assay and NF-B-dependent luciferase reporter assay, it is clear that imidazoquinoline treatment can lead to NF-B activation. It is possible that the TLR8-mediated IB␣ phosphorylation might lead to dissociation of IB␣ from NF-B without IB␣ degradation. The dissociated NF-B migrates to the nucleus to activate gene transcription. Alternately, the phosphorylation of IB␣ by imidazoquinoline treatment changes the confirmation of the NF-B-IB␣ complex, which exposes the nuclear localization signal on NF-B, leading to nuclear localization of NF-B. Previous studies showed that phosphorylation of IB␣ in the C-terminal PEST region is critical for IB␣ degradation (42,43). It is possible that imidazoquinoline treatment might miss such specific phosphorylation on IB␣, which leads to a lack of IB␣ degradation. Future studies are required to investigate the detailed molecular mechanism of TLR8-mediated NF-B activation.
The physiological significance of this novel TLR8-mediated MyD88-dependent pathway still needs to be further investigated. Our unpublished studies showed that this novel MyD88dependent pathway is predominantly present in primary kidney and intestine/colon epithelial cells. 4 Therefore, it is possible that this unique pathway plays a critical in maintaining the homeostasis of epithelium. Recent studies showed that the TLR8-MyD88-IRAK4 pathway is required to reverse the suppressive function of Treg cells (44). It should be very interesting to investigate whether this novel MyD88-dependent pathway plays a role in reversing the suppressive function of Treg cells.