The induction of cytokine synthesis by flagellin is mediated by a Toll-like receptor 5 (TLR5) signaling pathway. Although flagellin activation of the IL-1R-associated kinase and induction of TNF-α synthesis are dependent on TLR5 and not TLR4, we have found that flagellin stimulates NO in macrophages via a pathway that requires TLR5 and TLR4. Flagellin induced NO synthesis in HeNC2 cells, a murine macrophage cell line that expresses wild-type TLR4, but not in TLR4-mutant or -deficient GG2EE and 10ScNCr/23 cells. Flagellin stimulated an increase in inducible NO synthase (iNOS) mRNA and activation of the iNOS promoter. TLR5 forms heteromeric complexes with TLR4 as well as homomeric complexes. IFN-γ permitted GG2EE and 10ScNCr/23 cells to produce NO in response to flagellin. Flagellin stimulated IFN-β synthesis and Stat1 activation. The effect of flagellin on iNOS gene expression was inhibited by a Stat1 mutant protein. Taken together, these results support the conclusions that flagellin induces distinct patterns of inflammatory mediators depending on the nature of the TLR5 signaling complex and that the induction of NO by flagellin involves signaling via TLR5/TLR4 complexes.

Innate immunity serves as an essential first-line defense against microbial pathogens and may also influence the nature of the subsequent adaptive immune response (1, 2, 3). The accumulated evidence indicates that host cells involved in the innate immune response use the members of a relatively small family of structurally related membrane proteins termed Toll-like receptors (TLR)3 to recognize and respond to products from diverse groups of bacterial, viral, and fungal pathogens. Signaling via TLRs results in the production of an array of proinflammatory mediators, including cytokines (for example, IL-1, IL-8, and TNF-α), NO, leukotrienes, and platelet-activating factor. Presently, 10 human and nine murine TLRs have been identified that are characterized by the presence of a leucine-rich extracellular domain and cytoplasmic regions termed Toll/IL-1R homology (TIR) domains (see Ref. 4 for a review). Although TLRs do not, as a rule, exhibit specificity for a single microbial product, they are individually responsive to a limited group of molecules. For example, LPS (5, 6, 7), heat shock protein 60 (8), and Escherichia coli P fimbriae (9) signal via TLR4.

The results of several recent studies demonstrate that flagellin from Gram-negative bacteria signals via TLR5 (10, 11, 12). As with other TLRs, flagellin signaling via TLR5 results in activation of the IL-1R-associated kinase (IRAK) (12, 13). Flagellin stimulates a variety of TLR5-positive cell types, including monocytes, fibroblasts, and epithelial cells to produce cytokines such as TNF-α, IL-1, and IL-8 (3, 10, 11, 13, 14, 15, 16, 17, 18, 19) and NO in monocytes (13).

Although inducers such as LPS and flagellin have been linked with individual TLRs, there is mounting evidence that different TLRs may interact to produce distinct signaling events. For example, the activation of NF-κB-dependent gene expression in CHO cells and the induction of TNF-α in murine macrophages in response to Gram-positive bacterial peptidoglycan are dependent on the coexpression and physical interaction of TLR2 and TLR6 (20). Although the expression of TLR2 is sufficient for a response to phenol-soluble modulin from Staphylococcus epidermidis, the magnitude of the response is enhanced in the presence of TLR6 (21). In a prior study (13), we found that flagellin stimulated IRAK activation and TNF-α production by GG2EE cells, a C3H/HeJ mouse-derived macrophage cell line (22) that expresses a mutant form of TLR4. These observations indicate that at least one form of flagellin signaling is not dependent on the expression of a wild-type form of TLR4. However, as will be detailed in this report, the production of NO in response to flagellin is dependent on the coexpression and interaction of wild-type forms of TLR5 and TLR4.

The C3H/HeN-derived macrophage cell line HeNC2 and the LPS-hyporesponsive C3H/HeJ-derived macrophage cell line GG2EE were maintained in RPMI 1640 containing 10% FBS and 20 μg/ml gentamicin. 10ScNCr/23 cells, a clonal derivative of primary C57BL/10ScNCr macrophages (23), were provided by Dr. E. Lorenz. These cells do not express TLR4 mRNA or protein (5, 6). COS-1 cells and the murine macrophage cell line RAW 264.7 were maintained in DMEM with 10% FBS and gentamicin. Purified, endotoxin-free recombinant His-tagged Salmonella enteritidis flagellin was prepared as previously described (16). As detailed in our prior studies (12, 13, 14, 15, 16), at the concentrations used in our studies the flagellin contained insufficient residual LPS (usually <1 ng/ml) to produce a significant TLR4-dependent response (also see Fig. 1). Recombinant murine IFN-γ was obtained from Life Technologies (Gaithersburg, MD). Anti-FLAG Ab was obtained from Sigma-Aldrich (St. Louis, MO), high affinity anti-hemagglutinin (anti-HA) Ab was obtained from Roche (Indianapolis, IN), anti-phospho-Stat1 Ab was purchased from Zymed (San Francisco, CA), and anti-Stat1 was obtained from Santa Cruz Biotechnology (Santa Cruz, CA).

FIGURE 1.

Activation of the iNOS promoter by flagellin is TLR5 dependent. A, RAW264.7 cells were transiently transfected with the p3XFlag-CMV-14 vector (0) or TLR5-Flag (TLR5), rested overnight, and then incubated in the absence (0) or the presence of 10−9 M flagellin (F) or 100 ng/ml LPS (L) for 6 h. The cells were harvested and assessed for the level of inducible luciferase activity. The values have been normalized using the constitutive Renilla luciferase activity in each sample. B, RAW264.7 cells were transiently transfected with a TLR5 expression plasmid and the iNOS promoter and Renilla luciferase reporter constructs. After resting the cells overnight, the cultures were stimulated with increasing concentrations of flagellin for 6 h before analysis of luciferase activity.

FIGURE 1.

Activation of the iNOS promoter by flagellin is TLR5 dependent. A, RAW264.7 cells were transiently transfected with the p3XFlag-CMV-14 vector (0) or TLR5-Flag (TLR5), rested overnight, and then incubated in the absence (0) or the presence of 10−9 M flagellin (F) or 100 ng/ml LPS (L) for 6 h. The cells were harvested and assessed for the level of inducible luciferase activity. The values have been normalized using the constitutive Renilla luciferase activity in each sample. B, RAW264.7 cells were transiently transfected with a TLR5 expression plasmid and the iNOS promoter and Renilla luciferase reporter constructs. After resting the cells overnight, the cultures were stimulated with increasing concentrations of flagellin for 6 h before analysis of luciferase activity.

Close modal

The p3XFLAG-CMV-14/TLR5 plasmid encoding a FLAG epitope-tagged form of human TLR5 (12) was used as a template to generate a TLR5 PCR product for cloning into the pMH vector (Roche). The protein product encoded by this construct contains a C-terminal HA tag. The human TLR4-FLAG expression plasmid was provided by Tularik (South San Francisco CA). The inducible NO synthase (iNOS) promoter/luciferase reporter construct was provided by Dr. C. J. Lowenstein (The Johns Hopkins University School of Medicine, Baltimore, MD) (24). The Renilla luciferase control vector, pRL-TK, was purchased from Promega (Madison, WI). Stat1 Y701F cDNA was provided by Dr. F. Öberg (University of Uppsala, Uppsala, Sweden).

The production of NO was determined by measuring nitrite formation as previously described (13).

To assess the effect of flagellin on iNOS mRNA levels in HeNC2 and GG2EE cells, the cells (106 cells/culture) were stimulated for 6 h in the absence or the presence of 2.8 × 10−11 M flagellin or 10 ng/ml LPS, and then the cultures were harvested, and total RNA was isolated using the RNAqueous kit from Ambion (Houston, TX). RT-PCR was performed using the RETROscript first-strand synthesis kit (Ambion) and the following forward and reverse primers for murine iNOS: 5′-gaaaaccccttgtgctgttctc-3′ (forward) and 5′-ccgatgtggccttgtggtgaag-3′ (reverse). The following conditions were used for the PCR reaction: denaturing at 94°C for 4 min, 30 cycles of denaturing at 94°C for 30 s, annealing 55°C for 30 s, extension at 72°C for 50 s, and final extension at 72°C for 5 min. Using these primers, a product of 215 bp was generated that included sequences from two exons, thus reducing any problem associated with contamination by genomic DNA. Nonetheless, control reactions were performed with samples lacking reverse transcriptase. Cyclophilin primers were used to control for the quality of the RNA preparations and the efficiency of the RT-PCR reactions. The RT-PCR products were electrophoresed in 3% agarose (NuSieve/agarose, 3/1; BioWhittaker, Walkersville, MD). IFN-β mRNA levels in HeNC2 cells were evaluated in a similar manner using the primers described by Toschchakov et al. (25). The relative levels of IFN-β mRNA were determined by quantitating the intensity of the RT-PCR-generated bands and normalizing the values using internal TLR4 values for each sample.

To assess the effect of flagellin on iNOS promoter activation, RAW 264.7 cells were transiently transfected with an iNOS promoter/luciferase reporter (24) and Renilla luciferase control plasmid with or without TLR expression plasmids. The Renilla luciferase control plasmid provided a means to control for differences in transfection efficiency. The cells were transfected using the FuGene 6 reagent (Roche). After resting the cells overnight, the cultures were incubated in the presence or the absence of flagellin or LPS for 6 h before harvesting the cells. Inducible NOS promoter-dependent and Renilla luciferase activities were measured using the Promega dual-luciferase reporter assay system according to the manufacturer’s instructions.

To assess the association of TLR5 with TLR4, COS-1 cells were transiently transfected with TLR5-HA, TLR5-HA and TLR5-FLAG, or TLR5-HA and TLR4-FLAG expression plasmids using the Effectene transfection reagent (Qiagen, Valencia, CA) as previously described (12). After resting the cells for 48 h, cell lysates were prepared and incubated with anti-FLAG Ab and recombinant protein G-agarose (Invitrogen, San Diego, CA) for 2 h at 4°C (12). The immunoprecipitates were washed, eluted in sample buffer, and electrophoresed in a 7.5% SDS-PAGE, and the proteins were transferred to a polyvinylidene difluoride membrane. The membrane was then probed for TLR5-HA using the anti-HA high affinity Ab. The blot was stripped with Restore Western blot stripping buffer (Pierce, Rockford, IL) and reprobed for TLR5-FLAG and TLR4-FLAG with the anti-FLAG Ab.

The induction of IRAK activation (12, 13) and cytokine synthesis by Gram-negative flagellin is dependent on signaling via TLR5 (10, 11, 12, 13), but not TLR4 (13). In addition to inducing cytokine synthesis, flagellin stimulates macrophages to produce NO (13). Based on the results of a number of studies (24, 25, 26, 27, 28, 29) demonstrating that NO production in macrophages is dependent on the induction of iNOS gene transcription, we evaluated the TLR5 dependence of flagellin activation of the iNOS promoter. We took advantage of a preliminary observation (S.B. Mizel, unpublished observations) that the murine macrophage cell line RAW264.7 does not express TLR5 and therefore does not respond to flagellin. RAW 264.7 cells were transiently transfected with an iNOS promoter/luciferase reporter plasmid and a Renilla luciferase control plasmid (to control for transfection efficiency) with or without a TLR5 expression plasmid. The cells were rested overnight and then incubated for 6 h in the presence or the absence of flagellin or LPS, and the level of iNOS promoter-dependent luciferase activity was determined for each culture (Fig. 1,A). As expected (24), RAW 264.7 cells exhibited a strong response to LPS, since these cells are TLR4 positive. In contrast, flagellin had no significant effect on iNOS promoter-dependent gene expression in cells that had not received the TLR5 expression plasmid. However, flagellin did activate iNOS promoter-dependent gene expression in cells that had been transfected with a TLR5 expression plasmid. The effect of flagellin was concentration dependent; 50% of the maximal response was obtained with ∼5 × 10−11 M flagellin (Fig. 1 B). Based on these results, it is evident that TLR5 is required for flagellin-mediated activation of NO production as well as cytokine synthesis.

As noted previously, flagellin stimulation of IRAK activation and cytokine production is dependent on TLR5, but not TLR4 (13). This conclusion is based on the observation that the TLR4 mutant cell line, GG2EE, produces active IRAK and TNF-α in response to flagellin. In view of this finding, we assumed that the induction of NO production by flagellin would also be TLR4 independent. To evaluate the validity of this assumption, we tested the ability of flagellin to induce NO production in GG2EE cells. The GG2EE cell line was derived from a C3H/HeJ mouse (22) and thus expresses a TLR4 containing a proline to histidine mutation at amino acid residue 714. The HeNC2 cell line was used as a positive control, since these cells produce NO in response to flagellin and express functional TLR4 (13).

The GG2EE and HeNC2 cells were incubated for 48 h in the presence of increasing concentrations of flagellin, and the culture medium was then analyzed for the level of nitrite using the Griess reagent. In confirmation of our earlier studies (13), flagellin induced NO production in HeNC2 cells in a concentration-dependent manner; 50% of the maximal response was achieved with ∼2 × 10−11 M flagellin (Fig. 2). In marked contrast to the results with HeNC2 cells, flagellin did not induce NO production in GG2EE cells, even at concentrations that were several 1000-fold higher than the concentration required for 50% of the maximal response in HeNC2 cells. These results are consistent with the idea that flagellin stimulation of NO synthesis, unlike the activation of IRAK and the induction of TNF-α, is dependent on functional TLR4 as well as TLR5.

FIGURE 2.

Flagellin does not induce NO production in GG2EE cells. HeNC2 (•) and GG2EE (▴) cells were incubated with increasing concentrations of flagellin for 48 h. The culture medium from each sample was harvested and analyzed for the level of nitrite using Griess reagent.

FIGURE 2.

Flagellin does not induce NO production in GG2EE cells. HeNC2 (•) and GG2EE (▴) cells were incubated with increasing concentrations of flagellin for 48 h. The culture medium from each sample was harvested and analyzed for the level of nitrite using Griess reagent.

Close modal

Having found that flagellin does not induce NO production in TLR4-mutant GG2EE cells, we determined whether the lack of NO production was associated with an absence of iNOS mRNA. HeNC2 and GG2EE cells were incubated in the presence or the absence of 2.8 × 10−11 M flagellin or 10 ng/ml LPS for 6 h, and then total RNA was isolated from the individual cultures and assayed for iNOS mRNA using iNOS primers and RT-PCR. Cyclophilin primers were also used to control for sample variation in the quality of RNA or the RT-PCR reaction. As shown in Fig. 3, flagellin and LPS increased HeNC2 cell iNOS mRNA levels over those observed with unstimulated cells. However, neither stimulant triggered an increase in iNOS mRNA levels in GG2EE cells, a finding that is consistent with the lack of NO production in response to flagellin (Fig. 2).

FIGURE 3.

Flagellin does not elicit an increase in iNOS mRNA in GG2EE cells. HeNC2 and GG2EE cells were incubated in the absence (0) or the presence of 2.8 × 10−11 M flagellin (F) or 10 ng/ml LPS (L) for 6 h. Total RNA was isolated from each sample, and the presence or the absence of iNOS mRNA was assessed using RT-PCR. Cyclophilin (CyP) levels were measured to control for the quality of mRNA and the efficiency of the individual RT-PCR reactions.

FIGURE 3.

Flagellin does not elicit an increase in iNOS mRNA in GG2EE cells. HeNC2 and GG2EE cells were incubated in the absence (0) or the presence of 2.8 × 10−11 M flagellin (F) or 10 ng/ml LPS (L) for 6 h. Total RNA was isolated from each sample, and the presence or the absence of iNOS mRNA was assessed using RT-PCR. Cyclophilin (CyP) levels were measured to control for the quality of mRNA and the efficiency of the individual RT-PCR reactions.

Close modal

Although the preceding results are consistent with a requirement for a wild-type form of TLR4 in flagellin-induced NO production, it was also possible that the GG2EE cells possess a defect unrelated to TLR4 that prevents NO synthesis or flagellin signaling leading to NO production.

Previous studies on LPS hyporesponsive C3H/HeJ mice or macrophages derived from these mice established that IFN-γ permits LPS responsiveness (30, 31, 32, 33). In view of these findings, we explored the possibility that IFN-γ might permit GG2EE cells to produce NO in response to flagellin. The cells were incubated for 48 h in the presence of increasing concentrations of flagellin. One set of cultures also received 100 U/ml recombinant murine IFN-γ. As shown in Fig. 4, IFN-γ restored the ability of flagellin to induce NO production in GG2EE cells. As with TNF-α production in GG2EE cells (13), 50% of the maximal NO response was achieved with 3–5 × 10−11 M flagellin. To insure that this response was not unique to GG2EE cells, we conducted the same experiment using a cell line that does not express TLR4. Like the C57BL/10ScNCr mice from which they were derived, 10ScNCr/23 cells do not express TLR4 protein or mRNA. As with GG2EE cells, flagellin by itself had a minimal effect on NO production in 10ScNCr/23 cells (Fig. 5). However, the addition of flagellin and IFN-γ resulted in a marked induction of NO synthesis. IFN-γ by itself had no significant effect on NO production. The results with GG2EE and 10ScNCr/23 cells are consistent with the idea that the lack of flagellin induction of NO in these cell lines is due to the absence of a TLR4-associated signal and not to some component unrelated to TLR signaling. IFN-γ overcomes this condition by providing a compensatory second signal that is functionally equivalent to a signal elicited in the presence of TLR5 and TLR4.

FIGURE 4.

IFN-γ permits NO production in flagellin-stimulated GG2EE cells. GG2EE cells were incubated with increasing concentrations of flagellin (○) or flagellin and IFN-γ (•) for 48 h before the measurement of nitrite in the culture medium.

FIGURE 4.

IFN-γ permits NO production in flagellin-stimulated GG2EE cells. GG2EE cells were incubated with increasing concentrations of flagellin (○) or flagellin and IFN-γ (•) for 48 h before the measurement of nitrite in the culture medium.

Close modal
FIGURE 5.

IFN-γ permits NO production in flagellin-stimulated 10ScNCr/23 cells. 10ScNCr/23 cells were incubated in the presence or the absence of flagellin (10−10 M) or flagellin and IFN-γ (100 U/ml) for 48 h before measurement of nitrite in culture medium.

FIGURE 5.

IFN-γ permits NO production in flagellin-stimulated 10ScNCr/23 cells. 10ScNCr/23 cells were incubated in the presence or the absence of flagellin (10−10 M) or flagellin and IFN-γ (100 U/ml) for 48 h before measurement of nitrite in culture medium.

Close modal

Our results demonstrate that TLR5 by itself is not sufficient for flagellin stimulation of NO production. The responding cells must also express a functional form of TLR4. These results raised the possibility that TLR5 and TLR4 may interact to produce a signaling complex with activity distinct from that of TLR5 alone. The hypothesis that TLR5 and TLR4 may form a heteromeric complex was tested using coimmunoprecipitation analysis. For these experiments we used transfected COS-1 cells that express epitope-tagged forms of TLR4 and TLR5. In a previous study (12) we reported that transiently transfected COS-1 cells expressed high levels of ectopic full-length TLR5. Therefore, we used this cell line to examine the interaction between TLR5 and TLR4. In contrast to COS-1 cells, transiently transfected monocyte and T cell lines such as RAW 264.7, Jurkat, and EL4 do not express high levels of ectopic TLR4 or TLR5, because the majority of the expressed proteins are cleaved near the junction of the extracellular and transmembrane domains (S. B. Mizel, unpublished observations). Furthermore, the transfection efficiency with GG2EE cells is very low (regardless of the transfection reagent or protocol), and thus these cells were not a useful alternative.

To determine whether TLR5 forms homomeric complexes, COS-1 cells were transiently transfected with expression plasmids encoding TLR5-HA and TLR5-FLAG. The formation of heteromeric complexes between TLR5 and TLR4 was evaluated using TLR5-HA and TLR4-FLAG. Detergent cell lysates were prepared and incubated with anti-FLAG Ab to immunoprecipitate TLR5-FLAG or TLR4-FLAG and any associated TLR5-HA. Following electrophoresis and transfer to polyvinylidene difluoride membranes, the blots were probed for the presence of TLR5-HA using an anti-HA Ab. Subsequently, the blots were stripped and reprobed for TLR5-FLAG and TLR4-FLAG. As expected, TLR5-HA was not found in anti-FLAG immunoprecipitates from cells transfected with only TLR5-HA (Fig. 6, lane 1). However, TLR5-HA was present in immunoprecipitates from cells transfected with FLAG-tagged TLR5 (lane 2), a finding that establishes the ability of TLR5 to form homomeric complexes. For reasons that are not apparent at this time, FLAG-tagged TLR5, but not TLR5-HA, migrated as two bands during electrophoresis. The presence of two bands is not due to phosphorylation (S. B. Mizel, unpublished observations). However, it may be due to the use of an alternative translational initiation site. Since COS-1 cells were not exposed to flagellin, we also conclude that TLR5 self-association is not dependent on flagellin binding.

FIGURE 6.

TLR5 forms homomeric complexes as well as heteromeric complexes with TLR4. COS-1 cells were transiently transfected with TLR5-HA; TLR5-HA, TLR5, and TLR5-FLAG; or TLR5-HA and TLR4-FLAG. Lysates were prepared and incubated with anti-FLAG Ab and recombinant protein G to precipitate immune complexes. Following electrophoresis of the immune complexes, the proteins were transferred to polyvinylidene difluoride membrane and probed with anti-HA Ab to detect TLR5-HA (upper panel). The blot was then stripped and reprobed with anti-FLAG Ab to detect TLR5-FLAG or TLR4-FLAG (lower panel).

FIGURE 6.

TLR5 forms homomeric complexes as well as heteromeric complexes with TLR4. COS-1 cells were transiently transfected with TLR5-HA; TLR5-HA, TLR5, and TLR5-FLAG; or TLR5-HA and TLR4-FLAG. Lysates were prepared and incubated with anti-FLAG Ab and recombinant protein G to precipitate immune complexes. Following electrophoresis of the immune complexes, the proteins were transferred to polyvinylidene difluoride membrane and probed with anti-HA Ab to detect TLR5-HA (upper panel). The blot was then stripped and reprobed with anti-FLAG Ab to detect TLR5-FLAG or TLR4-FLAG (lower panel).

Close modal

Consistent with the postulated requirement for TLR4 and TLR5 collaboration in the NO response to flagellin (Figs. 2 and 3), we observed the formation of TLR5/TLR4 heteromeric complexes (Fig. 6, lane 3). However, the level of these heteromeric complexes consistently appeared to be less than the level of TLR5 homomeric complexes. This was the case even though the two proteins appeared to be present at comparable levels. The data in Fig. 6 not only demonstrate that TLR5 forms homomeric complexes, but also provide additional support for the involvement of TLR5/TLR4 heteromeric complexes in some forms of flagellin signaling.

LPS signaling via TLR4 results in activation of IFN regulatory factor-3 (IRF-3), induction of IFN-β, and subsequent activation of the transcription factor Stat1 that, in turn, binds to an IFN-γ-activated site in the iNOS gene (25, 27, 34, 35, 36). Like TLR4 agonists, TLR3 agonists can also induce IRF-3 and IFN-β expression (34) by a process that involves a protein termed TIR domain-containing adaptor-inducing IFN-β (TRIF) (37). In contrast to TLR4 and TLR3 agonists, TLR2 agonists do not induce IFN-β synthesis or Stat1 activation (25). This finding indicates that the ability to transduce signals culminating in IFN-β synthesis and Stat1 activation is limited to a restricted subset of TLRs. To determine whether the presence of TLR4 in the flagellin/TLR5 complex promotes IFN-β synthesis and Stat1 activation, we tested the ability of flagellin to induce IFN-β mRNA and protein as well as Stat1 activation in HeNC2 cells. The cells were incubated in the presence or the absence of flagellin or LPS for 3 h before analysis of IFN-β mRNA content (Fig. 7). As a control we also evaluated IFN-β mRNA expression in LPS-stimulated RAW264.7 cells. Identical amounts of RNA (based on OD260) were loaded in each lane. As an additional control, we measured the relative level of TLR4 mRNA in each sample. Although other investigators have found that LPS can reduce TLR4 expression to varying degrees (5, 38), we found the clones of RAW264.7 and HeNC2 cells we used did not exhibit a significant reduction in TLR4 expression in response to flagellin or LPS during the short time course of our experiments. LPS, as expected, induced an increase in the level of IFN-β mRNA (∼4.5-fold) in HeNC2 and RAW cells. Although flagellin reproducibly induced an increase in IFN-β mRNA in HeNC2 cells, the effect was relatively modest (∼2-fold). Since the smaller effect of flagellin might be due to a difference in the kinetics of induction between flagellin and LPS, we measured the effect of flagellin on IFN-β mRNA levels over a 6-h period. As shown in Fig. 8, the level of IFN-β was maximal after 2 h (∼2.5-fold increase) and declined to background levels after 6 h.

FIGURE 7.

Flagellin induces low level IFN-β mRNA expression in HeNC2 cells. HeNC2 cells were incubated in the presence or the absence of flagellin or LPS for 3 h, and the presence of IFN-β mRNA was assessed by RT-PCR. RAW cells were incubated in the presence or the absence of LPS for 3 h. The relative levels of TLR4 mRNA were also assessed and used to normalize the results with the individual samples.

FIGURE 7.

Flagellin induces low level IFN-β mRNA expression in HeNC2 cells. HeNC2 cells were incubated in the presence or the absence of flagellin or LPS for 3 h, and the presence of IFN-β mRNA was assessed by RT-PCR. RAW cells were incubated in the presence or the absence of LPS for 3 h. The relative levels of TLR4 mRNA were also assessed and used to normalize the results with the individual samples.

Close modal
FIGURE 8.

Time course for the induction of IFN-β mRNA in flagellin-treated HeNC2 cells. The cells were incubated in the presence of flagellin or 0, 2, 4, and 6 h before the determination of IFN-β and TLR4 mRNA contents.

FIGURE 8.

Time course for the induction of IFN-β mRNA in flagellin-treated HeNC2 cells. The cells were incubated in the presence of flagellin or 0, 2, 4, and 6 h before the determination of IFN-β and TLR4 mRNA contents.

Close modal

Given the small, but reproducible, effect of flagellin on IFN-β expression, we assessed the ability of flagellin to activate Stat1. To assess the effect of flagellin on Stat1 activation (a phosphorylation-dependent process), HeNC2 cells were incubated in the presence or the absence of flagellin for 4 h before Western blot analysis using an anti-phospho-Stat1-specific Ab. As a control, cells were also incubated with recombinant IFN-γ. As expected, IFN-γ by itself induced substantial Stat1 activation (Fig. 9,A). In contrast, flagellin had little observable effect on Stat1 activation in HeNC2 cells. Although not as potent as IFN-γ, LPS also stimulated significant Stat1 phosphorylation in HeNC2 cells (Fig. 9 B). Due to the lower level of Stat1 phosphorylation in the presence of LPS, it was necessary to develop the Western blots for 5–10 min as opposed to 30 s to 1 min for IFN-γ. Using the longer development time, we detected a very low level of Stat1-P in the flagellin-stimulated samples. Quantitative analysis revealed that the level of Stat1-P in response to LPS was 20-fold greater than with flagellin.

FIGURE 9.

Flagellin induces low level Stat1 phosphorylation in HeNC2 cells. A, Cells were incubated in the presence or the absence of flagellin (10−10 M) or flagellin and IFN-γ (100 U/ml) for 4 h. Lysates were prepared for Western blot analysis. The blot was probed with a Stat1-P-specific Ab, then stripped and reprobed with a Stat1-specific Ab. B, Cells were incubated for 4 h in the presence or the absence of flagellin (10−10 M) or LPS (10 ng/ml) for 4 h and analyzed for Stat1-P and total Stat1 expression.

FIGURE 9.

Flagellin induces low level Stat1 phosphorylation in HeNC2 cells. A, Cells were incubated in the presence or the absence of flagellin (10−10 M) or flagellin and IFN-γ (100 U/ml) for 4 h. Lysates were prepared for Western blot analysis. The blot was probed with a Stat1-P-specific Ab, then stripped and reprobed with a Stat1-specific Ab. B, Cells were incubated for 4 h in the presence or the absence of flagellin (10−10 M) or LPS (10 ng/ml) for 4 h and analyzed for Stat1-P and total Stat1 expression.

Close modal

In view of the low level of Stat1 activation in response to flagellin, it was possible that flagellin induction of iNOS gene expression may involve a non-Stat1-dependent pathway. To assess this possibility, we tested the ability of a mutant form of Stat1 (Y701F) to block flagellin activation of the iNOS promoter. Previous studies (39, 40, 41) have demonstrated that this mutant Stat1 can act in a dominant negative fashion to block IFN-γ and epidermal growth factor signaling, since the phosphorylation of Y701F is important for dimerization, translocation to the nucleus, DNA binding, and activation of transcription (42). RAW264.7 cells were transiently transfected with TLR5, the iNOS promoter/luciferase reporter, and the Renilla luciferase construct in the presence of an expression plasmid encoding the Stat1 Y701F or empty vector. The next day, the cells were incubated in the presence or the absence of flagellin or LPS for 6 h and then assayed for inducible luciferase activity. As expected, expression of the Stat1 Y701F markedly reduced the response to LPS (Fig. 10). The response to flagellin was also reproducibly reduced in cells expressing the Stat1 mutant protein. These results support the conclusion that TLR5/TLR4 signaling, like TLR4/TLR4 signaling, involves the activation of Stat1.

FIGURE 10.

Inhibition of flagellin and LPS signaling by a Stat1 mutant protein. RAW 264.7 cells were transiently transfected with TLR5 and an iNOS/luciferase reporter in the presence or the absence of Stat1 Y701F or empty vector. The cells were rested overnight and then stimulated with 10−10 M flagellin or 10 ng/ml LPS for 6 h before measuring inducible luciferase activity. Samples were normalized using Renilla luciferase activity.

FIGURE 10.

Inhibition of flagellin and LPS signaling by a Stat1 mutant protein. RAW 264.7 cells were transiently transfected with TLR5 and an iNOS/luciferase reporter in the presence or the absence of Stat1 Y701F or empty vector. The cells were rested overnight and then stimulated with 10−10 M flagellin or 10 ng/ml LPS for 6 h before measuring inducible luciferase activity. Samples were normalized using Renilla luciferase activity.

Close modal

The results presented in this study support the conclusion that the induction of NO synthesis by Gram-negative flagellin is dependent on signaling via TLR5/TLR4 complexes. Although TLR5-positive, TLR4-positive HeNC2 cells produce NO in response to flagellin, TLR5-positive, TLR4-mutant (GG2EE), and deficient (10ScNCr/23) cells are incapable of generating significant levels of NO even at very high concentrations of flagellin (Figs. 2, 4, and 5). The absence of a functional TLR4 is associated with an inability of the cells to initiate iNOS gene expression in the presence of flagellin (Fig. 3). An interaction of TLR5 with TLR4 was demonstrated by the coimmunoprecipitation of these proteins from lysates of COS-1 cells (Fig. 4). However, it is formally possible that this interaction may be limited in nontransfected cells expressing both proteins. Although we have focused our analysis on the role of TLR4 in flagellin-induced NO production, it is likely that TLR5/TLR4 complexes may be required for flagellin-induced expression of a number of other proinflammatory mediators.

The observed formation of heteromeric TLR5-containing complexes and the resultant ability to initiate alternate signaling leading to iNOS gene expression are consistent with the results obtained with other TLRs. As noted earlier, the response to several stimuli that signal via TLR2 can be markedly enhanced by the presence of TLR1 or TLR6 (20). However, in some cases the formation of heteromeric TLR complexes may be inhibitory. For example, TLR1 inhibits TLR4-dependent signaling in human microvascular endothelial cells in response to LPS (43) as well as TLR2-dependent signaling in HEK 293 cells stimulated with S. epidermidis phenol-soluble modulin (21). These types of observations are consistent with the idea that changes in the composition of TLR signaling complexes may result in the recruitment of different cofactors and thus the initiation of distinct second messenger pathways.

As noted earlier, signaling via TLR4 is associated with the induction of IFN-β, an event followed by IFN-β-mediated activation of Stat-1 (25, 27, 34, 35, 36). The addition of IFN-γ to TLR4 mutant or deficient GG2EE and 10ScNCr/23 cells circumvents the need for TLR4 in flagellin signaling leading to NO production, because IFN-γ, like IFN-β, can promote Stat-1 activation and thus complements the signals provided by TLR5/TLR5 complexes. Thus, our findings are consistent with the idea that the role of TLR4 in the TLR5/TLR4 complex is to provide the necessary signaling that culminates in IFN-β expression.

There are a number of possible mechanisms by which TLR4 may contribute to flagellin signaling that culminates in the expression of iNOS and NO production. TLR4 may participate in the binding of flagellin to TLR5/TLR4 complexes. In view of the inability of neutralizing anti-TLR4 Ab to block flagellin induction of NO production in HeNC2 cells (13) and the lack of flagellin responsiveness in TLR5-negative RAW 264.7 cells, this possibility is unlikely. Alternatively, the major contribution of TLR4 in flagellin signaling may be to contribute to the formation of a docking site for one or more factors that initiate a signaling pathway distinct from that used by TLR5 homomeric complexes. An increasing body of evidence is consistent with a model for LPS/TLR4 activation of NO production that involves an autocrine loop that begins with LPS-induced IRF-3 activation and IFN-β synthesis and results in the subsequent IFN-β-mediated activation of Stat1α, a transcription factor that promotes iNOS gene transcription (24, 25, 26, 27, 29, 35). Using this model, one would predict that the inability of a given TLR/agonist complex to promote IFN-β expression would prevent that TLR/agonist complex from triggering the induction of iNOS gene expression. The observations that TLR2 agonists do not induce NO production (44) and that TLR2 agonists are very poor inducers of IFN-β expression and Stat1α activation (25) provide important evidence in support of this model. Our observations indicate that TLR5 by itself cannot mediate a signaling pathway that is required for IFN-β expression and ultimately iNOS expression. Since TLR5 and TLR4 signaling are individually associated with the activation of IRAK, it is likely that the contribution of TLR4 must involve a mechanism unrelated to the binding of MyD88/IRAK complexes. The results of several recent studies indicate that TLR4 signaling may be MyD88 dependent or independent (25, 45, 46, 47, 48). Although LPS induction of macrophage IL-6 and TNF-α synthesis is MyD88 dependent (45), the expression of IFN-β is MyD88 independent and involves TIR domain-containing adaptor protein (TIRAP)/MyD88 adapter-like (Mal) (25, 49) or TRIF (37). The action of TIRAP/Mal appears to be selective, as evidenced by the lack of involvement of this protein in TLR9 (48), TLR2 (25), or TLR5 (50) signaling. If TIRAP/Mal or TRIF are involved in TLR5/TLR4 signaling, then their activity must be more limited, relative to their activity in TLR4/TLR4 signaling. This conclusion is based on our observations that the levels of NO, IFN-β mRNA, and Stat1 activation in response to flagellin are lower than those observed with LPS. It may be that the affinity of TIRAP/Mal or TRIF for TLR5/TLR4 complexes is markedly reduced relative to that for TLR4/TLR4 complexes. Alternatively, a molecule other than TIRAP/Mal or TRIF may be recruited to TLR5/TLR4 complexes that subserves the same function, but does so in a more limited manner. These possibilities are currently under study.

1

This work was supported by National Institutes of Health Grants AI38670 and AI51319 (to S. B. M.).

3

Abbreviations used in this paper: TLR, Toll-like receptor; HI, hemagglutinin; iNOS, inducible NO synthase; IRAK, IL-1R-associated kinase; IRF-3, IFN regulatory factor-3; Mal, MyD88 adapter-like protein; TIR, Toll/IL-1 receptor homology; TIRAP, TIR domain-containing adaptor protein; TRIF, TIR domain-containing adaptor-inducing IFN-β.

1
Kaisho, T., S. Akira.
2002
. Toll-like receptors as adjuvant receptors.
Biochim. Biophys. Acta
1589
:
1
.
2
Re, F., J. L. Strominger.
2001
. Toll-like receptor 2 (TLR2) and TLR4 differentially activate human dendritic cells.
J. Biol. Chem.
276
:
37692
.
3
Sierro, F., B. Dubois, A. Coste, D. Kaiserlian, J.-P. Kraehenbuhl, J.-C. Sirard.
2002
. Flagellin stimulation of intestinal epithelial cells triggers CCL20-mediated migration of dendritic cells.
Proc. Natl. Acad. Sci. USA
98
:
13722
.
4
Aderem, A., R. J. Ulevitch.
2000
. Toll-like receptors in the induction of the innate immune response.
Nature
406
:
782
.
5
Poltorak, A., X. He, I. Smirnova, M. Y. Liu, C. V. Huffel, X. Du, D. Birdwell, E. Alejos, M. Silva, C. Galanos, et al
1998
. Defective LPS signaling in C3H/HeJ and C57BL/10ScCr mice: mutations in Tlr4 gene.
Science
282
:
2085
.
6
Qureshi, S. T., L. Lariviere, G. Leveque, S. Clermont, K. J. Moore, P. Gros, Malo.
1999
. Endotoxin-tolerant mice have mutations in Toll-like receptor 4 (Tlr4).
J. Exp. Med.
189
:
615
.
7
Hoshino, K., O. Takeuchi, T. Kawai, H. Sanjo, T. Ogawa, Y. Takeda, K. Takeda, S. Akira.
1999
. Cutting edge: Toll-like receptor 4 (TLR4)-deficient mice are hyporesponsive to lipopolysaccharide: evidence for TLR4 as the Lps gene product.
J. Immunol.
162
:
3749
.
8
Ohashi, K., V. Burkart, S. Flohe, H. Kolb.
2000
. Heat shock protein 60 is a putative endogenous ligand of the Toll-like receptor-4 complex.
J. Immunol.
164
:
558
.
9
Frendeus, B., C. Wachtler, M. Hedlund, H. Fischer, P. Samuelsson, M. Svensson, C. Svanborg.
2001
. Escherichia coli P fimbriae utilize the Toll-like receptor 4 pathway for cell activation.
Mol. Microbiol.
40
:
37
.
10
Hayashi, F., K. D. Smith, A. Ozinsky, T. R. Hawn, E. C. Yi, D. R. Goodlett, J. K. Eng, S. Akira, D. M. Underhill, A. Aderem.
2001
. The innate immune response to bacterial flagellin is mediated by Toll-like receptor 5.
Nature
410
:
1099
.
11
Gewirtz, A. T., T. A. Navas, S. Lyons, P. J. Godowski, J. L. Madara.
2001
. Bacterial flagellin activates basolaterally expressed TLR5 to induce epithelial proinflammatory gene expression.
J. Immunol.
167
:
1882
.
12
Mizel, S. B., J. A. Snipes.
2002
. Gram-negative flagellin-induced self-tolerance is associated with a block in interleukin-1 receptor-associated kinase release from Toll-like receptor 5.
J. Biol. Chem.
277
:
22414
.
13
Moors, M. A., L. Li, S. B. Mizel.
2001
. Activation of interleukin-1 receptor-associated kinase by Gram-negative flagellin.
Infect. Immun.
69
:
4424
.
14
Ciacci-Woolwine, F., I. C. Blomfield, S. H. Richardson, S. B. Mizel.
1998
. Salmonella flagellin induces tumor necrosis factor α in a human promonocytic cell line.
Infect. Immun.
66
:
1127
.
15
Ciacci-Woolwine, F., P. F. McDermott, S. B. Mizel.
1999
. Induction of cytokine synthesis by flagella from Gram-negative bacteria may be dependent on the activation or differentiation state of human monocytes.
Infect. Immun.
67
:
5176
.
16
McDermott, P. F., F. Ciacci-Woolwine, J. A. Snipes, S. B. Mizel.
2000
. High-affinity interaction between Gram-negative flagellin and a cell surface polypeptide results in human monocyte activation.
Infect. Immun.
68
:
5525
.
17
Eaves-Pyles, T., K. Murthy, L. Liaudet, L. Virag, G. Ross, F. G. Soriano, C. Szabo, A. L. Salzman.
2001
. Flagellin, a novel mediator of Salmonella-induced epithelial activation and systemic inflammation: IκBα degradation, induction of nitric oxide synthase, induction of proinflammatory mediators, and cardiovascular dysfunction.
J. Immunol.
166
:
1248
.
18
Gewirtz, A. T., P. O. Jr, C. K. Simon, L. J. Schmitt, C. H. Taylor, A. D. Hagedorn, A. S. O’Brien, A. S. Neish, J. L. Madara.
2001
. Salmonella typhimurium translocates flagellin across intestinal epithelia, inducing a proinflammatory response.
J. Clin. Invest.
107
:
99
.
19
Eaves-Pyles, T., H. R. Wong, K. Odoms, R. B. Pyles.
2002
. Salmonella flagellin-dependent proinflammatory responses are localized to the conserved amino and carboxyl regions of the protein.
J. Immunol.
167
:
7009
.
20
Ozinsky, A., D. M. Underhill, J. D. Fontenot, A. M. Hajjar, K. D. Smith, C. B. Wilson, L. Schroeder, A. Aderem.
2000
. The repertoire for pattern recognition of pathogens by the innate immune system is defined by cooperation between Toll-like receptors.
Proc. Natl. Acad. Sci. USA
97
:
13766
.
21
Hajjar, A. M., D. S. O’Mahony, A. Ozinsky, D. M. Underhill, A. Aderem, S. J. Klebanoff, C. B. Wilson.
2001
. Functional interactions between Toll-like receptor (TLR) 2 and TLR1 or TLR6 in response to phenol-soluble modulin.
J. Immunol.
166
:
15
.
22
Blasi, E., D. Radzioch, S. K. Durum, L. Varesio.
1987
. A murine macrophage cell line, immortalized by v-raf and v-myc oncogenes, exhibits normal macrophage functions.
Eur. J. Immunol.
17
:
1491
.
23
Lorenz, E., D. D. Patel, T. Hartung, D. A. Schwartz.
2002
. Toll-like receptor 4 (TLR4)-deficient murine macrophage cell line as an in vitro assay system to show TLR4-independent signaling of Bacteroides fragilis lipopolysaccharide.
Infect. Immun.
70
:
4892
.
24
Lowenstein, C. J., E. W. Alley, P. Raval, A. M. Snowman, S. H. Snyder, S. W. Russell, W. J. Murphy.
1993
. Macrophage nitric oxide synthase gene: two upstream regions mediate induction by interferon γ and lipopolysaccharide.
Proc. Natl. Acad. Sci. USA
90
:
9730
.-9734.
25
Toschchakov, V., B. W. Jones, P. Y. Perera, K. Thomas, M. J. Cody, S. Zhang, B. R. G. Williams, J. Major, T. A. Hamilton, M. J. Fenton, et al
2002
. TLR4, but not TLR2, mediates IFN-β-induced STAT1αβ-dependent gene expression in macrophages.
Nat. Immunol.
3
:
392
.
26
Gao, J., D. C. Morrison, T. J. Parmely, S. W. Russell, W. J. Murphy.
1997
. An interferon-γ-activated site (GAS) is necessary for full expression of the mouse iNOS gene in response to interferon-γ and lipopolysaccharide.
J. Biol. Chem.
272
:
1226
.
27
Gao, J. J., M. B. Filla, M. J. Fultz, S. N. Vogel, S. W. Russell, W. J. Murphy.
1998
. Autocrine/paracrine IFN-αβ mediates the lipopolysaccharide-induced activation of transcription factor Stat1α in mouse macrophages: pivotal role of Stat1α in induction of the inducible nitric oxide synthase gene.
J. Immunol.
161
:
4803
.
28
Chan, E. D., D. W. H. Riches.
1998
. Potential role of the JNK/SAPK signal transduction pathway in the induction of iNOS by TNF-α.
Biochem. Biophys. Res. Commun.
253
:
790
.
29
Jacobs, A. T., L. J. Ignarro.
2001
. Lipopolysaccharide-induced expression of interferon-β mediates the timing of inducible nitric-oxide synthase induction in RAW 264.7 macrophages.
J. Biol. Chem.
276
:
47950
.
30
Beutler, B., V. Tkacenko, I. Milsark, N. Krochin, A. Cerami.
1986
. Effect of γ interferon on cachectin expression by mononuclear phagocytes: reversal of the lpsd (endotoxin resistance) phenotype.
J. Exp. Med.
164
:
1791
.
31
Akagawa, K. S., K. Kamoshita, S. Onodera, T. Tokunaga.
1987
. Restoration of lipopolysaccharide-mediated cytotoxic macrophage induction in C3H/HeJ mice by interferon-γ or a calcium ionophore.
Jpn. J. Cancer Res.
78
:
279
.
32
Hogan, M. M., S. N. Vogel.
1987
. Lipid A-associated proteins provide an alternative “second signal” in the activation of recombinant interferon-γ-primed, C3H/HeJ macrophages to a fully tumoricidal state.
J. Immunol.
139
:
3697
.
33
Chapes, S. K., J. W. Killion, D. C. Morrison.
1988
. Tumor cell killing and cytostasis by C3H/HeJ macrophages activated in vitro by lipid A-associated protein and interferon γ.
J. Leukocyte Biol.
43
:
232
.
34
Doyle, S., S. Vaidya, R. O’Connell, H. Dadgostar, P. Dempsey, T. Wu, G. Rao, M. Haberland, R. Modlin, G. Cheng.
2002
. IRF3 mediates a TLR3/TLR4-specific antiviral gene program.
Immunity
17
:
251
.
35
Weinstein, S. L., A. J. Finn, S. H. Dave, F. Meng, C. A. Lowell, J. S. Sanghera, A. L. DeFranco.
2000
. Phosphatidylinositol 3-kinase and mTOR mediate lipopolysaccharide-stimulated nitric oxide production in macrophages via interferon-β.
J. Leukocyte Biol.
67
:
405
.
36
Hoshino, K., T. Kaisho, T. Iwabe, O. Takeuchi, S. Akira.
2002
. Differential involvement of IFN-β in Toll-like receptor-stimulated dendritic cell activation.
Int. Immunol.
14
:
1225
.
37
Yamamoto, M., S. Sato, K. Mori, K. Hoshino, O. Takeuchi, K. Takeda, S. Akira.
2002
. A novel Toll/IL-1 receptor domain-containing adapter that preferentially activates the IFN-β promoter in the Toll-like receptor signaling.
J. Immunol.
169
:
6668
.
38
Medvedev, A. E., K. M. Kopydlowski, S. N. Vogel.
2000
. Inhibition of lipopolysaccharide-induced signal transduction in endotoxin-tolerized mouse macrophages: dysregulation of cytokine, chemokine, and Toll-like receptor 2 and 4 gene expression.
J. Immunol.
164
:
5564
.
39
Nakajima, K., Y. Yamanaka, K. Nakae, H. Kojima, M. Ichiba, N. Kiuchi, T. Kitaoka, T. Fukada, M. Hibi, T. Hirano.
1996
. A central role for Stat3 in IL-6-induced regulation of growth and differentiation in M1 leukemia cells.
EMBO J.
15
:
3651
.
40
Bromberg, J. F., Z. Fan, C. Brown, J. Mendelsohn, J. E. Darnell, Jr.
1998
. Epidermal growth factor-induced growth inhibition requires Stat1 activation.
Cell Growth Differ.
9
:
505
.
41
Walter, M. J., D. C. Look, R. M. Tidwell, W. T. Roswit, M. J. Holtzman.
1997
. Targeted inhibition of interferon-γ-dependent intercellular adhesion molecule-1 (ICAM-1) expression using dominant-negative Stat1.
J. Biol. Chem.
272
:
28582
.
42
Levy, D. E., J. E. Darnell, Jr.
2002
. Stats: transcriptional control and biological impact.
Mol. Cell. Biol.
3
:
651
.
43
Spitzer, J. H., A. Visintin, A. Mazzoni, M. N. Kennedy, D. M. Segal.
2002
. Toll-like receptor 1 inhibits Toll-like receptor 4 signaling in endothelial cells.
Eur. J. Immunol.
32
:
1182
.
44
Jones, B. W., T. K. Means, K. A. Heldwein, M. A. Keen, P. J. Hill, J. T. Belisle, M. J. Fenton.
2001
. Different Toll-like receptor agonists induce distinct macrophage responses.
J. Leukocyte Biol.
69
:
1036
.
45
Kawai, T., O. Adachi, T. Ogawa, K. Takeda, S. Akira.
1999
. Unresponsiveness of MyD88-deficient mice to endotoxin.
Immunity
11
:
115
.
46
Schnare, M., A. C. Holt, K. Takeda, S. Akira, R. Medzhitov.
2000
. Recognition of CpG DNA is mediated by signal pathways dependent on the adaptor protein MyD88.
Curr. Biol.
10
:
1139
.
47
Kaisho, T., O. Takeuchi, T. Kawai, K. Hoshino, S. Akira.
2001
. Endotoxin-induced maturation of MyD88-deficient dendritic cells.
J. Immunol.
166
:
5688
.
48
Horng, T., G. M. Barton, R. Medzhitov.
2001
. TIRAP: an adapter molecule in the Toll signaling pathway.
Nature Immunol.
2
:
835
.
49
Fitzgerald, K. A., E. M. Palsson-McDermott, A. G. Bowie, C. A. Jefferies, A. S. Mansell, G. Brady, E. Brint, A. Dunne, P. Gray, M. T. Harte, D. McMurray, D. E. Smith, J. E. Sims, T. A. Bird, L. A. O’Neill.
2001
. Mal (MyD88-adapter-like) is required for Toll-like receptor-4 signal transduction.
Nature
413
:
78
.
50
Horng, T., G. M. Barton, R. Flavell, R. Medzhitov.
2002
. The adaptor molecule TIRAP provides signaling specificity for Toll-like receptors.
Nature
420
:
329
.