Blocking TIR Domain Interactions in TLR9 Signaling Artur Javmen, Henryk Szmacinski, Joseph R. Lakowicz and Vladimir Y. Toshchakov This information is current as of October 1, 2021. J Immunol published online 18 June 2018 http://www.jimmunol.org/content/early/2018/06/15/jimmun ol.1800194 Downloaded from Supplementary http://www.jimmunol.org/content/suppl/2018/06/18/jimmunol.180019 Material 4.DCSupplemental

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on October 1, 2021 Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2018 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published June 18, 2018, doi:10.4049/jimmunol.1800194 The Journal of Immunology

Blocking TIR Domain Interactions in TLR9 Signaling

Artur Javmen,* Henryk Szmacinski,† Joseph R. Lakowicz,† and Vladimir Y. Toshchakov*

Interaction of TLR9 with ligands activates NF-kB, leading to proinflammatory production. Excessive TLR activation is a pathogenic factor for inflammatory diseases. This study has examined cell-permeating decoy peptides (CPDPs) derived from the TLR9 Toll/IL-1R resistance (TIR) domain. CPDP 9R34, which included AB loop, b-strand B, and N-terminal BB loop residues, inhibited TLR9 signaling most potently. CPDPs derived from a-helices C, D, and E (i.e., 9R6, 9R9, and 9R11) also inhibited TLR9-induced but were less potent than 9R34. 9R34 did not inhibit TLR2/1, TLR4, or TLR7 signaling. The N-terminal deletion modification of 9R34, 9R34-DN, inhibited TLR9 as potently as the full length 9R34. Binding of 9R34-DN to TIR domains was studied using cell-based Fo¨rster resonance energy transfer/fluorescence lifetime imaging approach. Cy3-labeled 9R34-DN dose-dependently decreased fluorescence lifetime of TLR9 TIR–Cerulean (Cer) fusion . Cy3–9R34-DN also bound TIRAP TIR, albeit with a lesser affinity, but not MyD88 TIR, whereas CPDP from the opposite TIR surface, 9R11, bound both adapters and TLR9. i.p. administration of 9R34-DN suppressed oligonucleotide-induced systemic cytokines and lethality in mice. This study identifies a potent, TLR9-specific CPDP that targets both dimerization and adapter recruitment. Location of TIR Downloaded from segments that represent inhibitory CPDPs suggests that TIR domains of TLRs and TLR adapters interact through structurally homologous surfaces within primary receptor complex, leading to formation of a double-stranded, filamentous structure. In the presence of TIRAP and MyD88, primary complex can elongate bidirectionally, from two opposite ends, whereas in TIRAP- deficient cells, elongation is unidirectional, only through the aE side. The Journal of Immunology, 2018, 201: 000–000. http://www.jimmunol.org/ oll-like receptors, a family of pattern recognition re- unmethylated CpG motifs, often present in bacterial and viral ceptors, are evolutionarily conserved and present in most DNA (1, 6, 13, 14). T animal phyla. TLRs function as a component of the TLRs are type I transmembrane receptors that initiate intracellular responsible for recognition of various signaling through Toll/IL-1R resistance domains (TIR). Activated microbial molecules (1, 2). These receptors play an important TLRs dimerize and bring their TIR domains to direct physical contact role in the initiation and maintenance of inflammatory response (15). The formed TIR dimers recruit TIR domain-containing adapter to pathogens. Excessive or prolonged TLR signaling is a path- through TIR–TIR interactions (16–18). There are four TIR- ogenic factor in many inflammatory and autoimmune diseases containing TLR adapter proteins that are involved in signaling (1, 3–8). TLRs have different cellular localization. TLR1, TLR2, (i.e., MyD88, TIRAP/Mal, , and TRAM) (5, 19, 20). Recruit- by guest on October 1, 2021 TLR4, TLR5, and TLR6 are located at the plasma membrane. ment of MyD88 to TLRs promotes the assembly of MyDDosome, the These TLRs recognize microbial lipids, lipopeptides, and pro- signaling complex formed by death domains of MyD88, IRAK4, and teins. TLR3, TLR7, TLR8, TLR9, and TLR13 are localized IRAK2 (21, 22). MyDDosome formation activates IRAKs, leading to in intracellular compartments and recognize nucleic acids, the activation of NF-kB (19, 22, 23). TLR4 and TLR3 can also en- whereas TLR11 and TLR12, also endosomally expressed, rec- gage TRIF, leading to the formation of triffosome and activation of ognize bacterial proteins (9–11). Agonists of TLR10 remain IFN regulatory factor–3 (IRF-3) (20, 24, 25). TIRAP and TRAM, unknown (12). TLR9 is activated by DNA sequences that contain often referred to as bridging adapters, stabilize TLR TIR dimers and facilitate the recruitment of MyD88 and TRIF, respectively (26–29). TIRs are a/b protein domains, the core of which is formed by *Department of Microbiology and Immunology, University of Maryland School the central, typically 5-stranded, parallel b-sheet, surrounded by of Medicine, Baltimore, MD 21201; and †Center for Fluorescence Spectroscopy, Department of and Molecular Biology, University of Maryland School five a-helices (30–32). Sequence similarity of TIR domains is of Medicine, Baltimore, MD 21201 typically limited to only 20–30% (29, 33, 34). TIR domains can ORCID: 0000-0002-7942-2294 (V.Y.T.). interact through structurally diverse regions (18, 35–39). For in- Received for publication February 12, 2018. Accepted for publication May 24, 2018. stance, TLR2 TIR a-helix D interacts with TIRAP, whereas its AB This work was supported by National Institutes of Health Grants AI-082299 (to V.Y.T.), loop is involved in receptor dimerization (37). RR 26370 (to J.R.L.), and GM125976 and OD019975 (to J.R.L.). Involvement of TLRs in development of rheumatoid arthritis, Address correspondence and reprint requests to Dr. Vladimir Y. Toshchakov, Depart- atherosclerosis, systemic lupus erythematosus, and systemic ment of Microbiology and Immunology, University of Maryland School of Medicine, sclerosis suggests a significant potential of TLR antagonists in the 685 W. Baltimore Street, HSF1 Room 380, Baltimore, MD 21201. E-mail address: [email protected] treatment of inflammatory diseases (6, 40–50). Our previous The online version of this article contains supplemental material. studies screened libraries of cell-permeating decoy peptides Abbreviations used in this article: BMDM, bone marrow–derived macrophage; Cer, (CPDPs) derived from TIR domains of TLR2 and TLR4, their Cerulean; CPDP, cell-permeating decoy peptide; D-Gal, D-galactosamine; FLIM, adapters, and their coreceptors. The CPDPs tested in this and our fluorescence lifetime imaging; FRET, Fo¨rster resonance energy transfer; HHBlits, HMM-HMM–based lightning-fast iterative sequence search; ODN, oligonucleotide; previous studies are composed of two functional parts. The Pam3C, S-(2,3-bis(palmitoyloxy)-(2R,2S)-propyl)-N-palmitoyl-(R)-Cys-Ser-Lys4-OH; S1, common, N-terminal part of a CPDP is the cell-permeating surface 1; S2, surface 2; S3, surface 3; S4, surface 4; TIR, Toll/IL-1R resistance peptide vector derived from Antennapedia homeodomain domain. (RQIKIWFQNRRMKWKK) (51). Many studies, including our Copyright Ó 2018 by The American Association of Immunologists, Inc. 0022-1767/18/$35.00 previous CPDP screenings, have demonstrated the efficacy of

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1800194 2 MECHANISMS OF ADAPTER RECRUITMENT TO ACTIVATED TLR9 this vector for intracellular delivery of cargoes of diverse size Evaluation of cytokine expression by quantitative and chemical nature in cell culture and small animal models (37, real time PCR 52–56). The C-terminal half of a CPDP is a segment of TIR BMDM (2 3 106 per well) were plated in 12-well plates, incubated domain primary sequence that represents a particular, non- overnight, and treated with CPDPs for 30 min. cDNA was synthesized fragmented patch of TIR domain surface that might serve as a from 1 mg of total RNA isolated using Trizol (Life Technologies, Carlsbad, TIR–TIR interface. The TIR libraries were designed in such a CA) and reverse transcribed using RevertAid RT Reverse Transcription Kit way that all peptides in a library collectively represent the entire (Thermo Fisher Scientific, Waltham, MA). The cDNA obtained was am- plified with -specific primers for mouse HPRT, TNF-a, IL-1b, IL- surface of the TIR. Screenings of TIR peptide libraries have 12p40,orIL-6 and Fast SYBR Green Master Mix (Applied Biosystems, resulted in identification of several TLR inhibitors that have Foster City, CA) as previously described (53). distinct specificities and dissimilar sequences (18, 37, 38, 53, 55, ELISA evaluation of cytokine secretion 57). The anti-inflammatory properties of several peptides that resulted from these studies have been confirmed by several in- BMDM (1 3 106 per well) were plated in 24-well plates and treated with dependent groups (58–60). Thus, Hu et al. (58) observed a potent CPDP for 30 min prior to stimulation with a TLR agonist for 5 h. Mouse a anti-inflammatory effect of the TIRAP-derived peptide TR6 (53) TNF- and IL-12p40 concentrations were measured in supernatants using ELISA kits from BioLegend (San Diego, CA). in a mouse model of LPS-induced mastitis. The same research group demonstrated the protective action of TRAM-derived Immunoprecipitation of MyD88-containing signaling peptide TM6 (55) in LPS-induced acute lung injury (59). complexes and immunoblotting TLR4-derived peptide 4BB (36) decreased effects of LPS on BMDM (4 3 106 per well) were plated in six-well plates and treated with calcium fluxes and neuronal excitability in studies conducted by CPDP for 30 min prior to stimulation with ODN 1668 for 90 min. Cells Allette et al. (60). Although several independent research groups were lysed in 600 ml of buffer containing 20 mM HEPES (pH 7.4), Downloaded from have reported analogous TLR inhibitory peptides from bacterial 150 mM NaCl, 10 mM NaF, 2 mM Na3VO4, 1 mM EDTA, 1 mM EGTA, 0.5% Triton X-100, 0.1 mM DTT, and protease inhibitor mixture (Roche, and viral proteins (61–63), no inhibitor that targets TLR9 has Indianapolis, IN). Cell extracts were incubated overnight with 1 mg of anti- been reported to date. mouse MyD88 Ab (AF3109) (R&D Systems, Minneapolis, MN), followed The present study identifies new CPDP inhibitors of TLR9. The by 4 h incubation with 20 ml of Protein G Sepharose beads (Thermo Fisher most potent inhibitory peptide of the TLR9 TIR library, 9R34, and Scientific). The beads were then washed three times with 500 ml of lysis

m http://www.jimmunol.org/ its modification, 9R34-DN, are derived from the surface-exposed buffer and boiled in 60 l of Laemmli Sample Buffer (Bio-Rad, Hercules, b CA). Cell extracts were electrophoresed on 10% acrylamide gel by SDS- segment of TLR9 TIR that includes AB loop, -strand B, and PAGE and transferred to PVDF membrane (Bio-Rad). Rabbit anti-MyD88 N-terminal residues of BB loop. Peptides from a-helices E, D, and IgG was purchased from Cell Signaling Technology (Danvers, MA). to a lesser extent, C (i.e., 9R11, 9R9, and 9R6) also inhibited Mouse anti-IRAK4 IgG was obtained from Abcam (Cambridge, U.K.). TLR9 but were less potent. Using a cell-based binding assay, we Alkaline phosphatase–conjugated secondary Abs against mouse and rabbit D IgG were purchased from Cell Signaling Technology. The stabilized demonstrated that 9R34- N binds the TLR9 TIR domain and, substrate for alkaline phosphatase was from Promega (Madison, WI). with a slightly lower affinity, TIRAP TIR. 9R34-DN inhibited systemic cytokine activation induced in mice by oligonucleotide Expression vectors (ODN) 1668 and protected D-galactosamine (D-Gal)–pretreated

MyD88–Cer, TIRAP–Cer, and TLR2–Cer expression vectors were de- by guest on October 1, 2021 mice against TLR9-induced lethality. The results presented iden- scribed previously (37, 55). Because of low expression level, full-length tify novel TLR9 inhibitors and provide new insights into the TLR4–Cer and TLR9–Cer vectors were replaced with TLR4 TIR–Cer and molecular mechanisms of TLR9 signaling. TLR9 TIR–Cer, respectively. These vectors do not include TLR4 and TLR9 ectodomains but encode full transmembrane sections and TIR domains (37). Materials and Methods Fluorescence lifetime imaging Animals and cells HeLa cells (3 3 106) were transfected with 10 mg of MyD88–Cer, TIRAP– Eight-week-old female C57BL/6J mice were obtained from The Jackson Cer, TRAM-Cer, TLR2-Cer, TLR4 TIR–Cer, or TLR9 TIR–Cer expression Laboratory (Bar Harbor, ME). Bone marrow–derived macrophages vectors using a Lipofectamine 3000 from Invitrogen (Carlsbad, CA) per (BMDM) were isolated and cultivated according to Weischenfeldt and manufacturer recommendations. Twenty-four hours later, the transfected Porse (64). Mouse tibias and femurs were flushed with ice-cold PBS, and cells were trypsinized and reseeded into a 50-well gasket (Grace Bio-Labs, the cells obtained were transferred to RPMI 1640 supplemented with Bend, OR) mounted on a microscope slide at the density 8000 cells per 10% L929 cell supernatant. The cells were cultured for 10–14 d prior to well. The next day, cells were treated with Cy3-labeled CPDPs for 1 h and experiments. fixed on the slides with 4% paraformaldehyde solution. Fluorescence lifetime images were acquired using the Alba V fluores- TLR agonists cence lifetime imaging (FLIM) system (ISS, Champagne, IL). The exci- ODN 1668, S-(2,3-bis(palmitoyloxy)-(2R,2S)-propyl)-N-palmitoyl-(R)-Cys- tation was from the laser diode 443 nm coupled with scanning module (ISS) through multiband dichroic filter 443/532/635 nm (Semrock) to Olympus Ser-Lys4-OH (Pam3C), and R848 were purchased from InvivoGen (San 3 Diego, CA). Phenol-purified Escherichia coli K235 LPS (65) was a kind gift IX71S microscope with objective 20 0.45 numeric aperture (UPlan; of Dr. Stefanie N. Vogel (University of Maryland School of Medicine). Olympus, Center Valley, PA). Emission was observed through bandpass filter 480/30 nm (Chroma Technology, Bellows Falls, VT) and detected by Peptide design, synthesis, and reconstitution a photomultiplier H7422-40 (Hamamatsu Photonics). Measurements were performed using frequency domain and time domain modalities of the Decoy peptides from TLR9 TIR domain were synthesized in tan- FLIM system. Frequency domain FLIM data were acquired using ISS dem with the cell-permeating Antennapedia homeodomain sequence A320 FastFLIM electronics with n harmonics of 20 MHz laser repetition (RQIKIWFQNRRMKWKK) (51) and placed at the N terminus, as in our frequency (n = 1, 2, 3, 4, 5, 6). FastFLIM was calibrated using fluorescein prior studies that demonstrated that this vector is effective for intracel- in buffer (pH 8) as a standard with a single lifetime of 4 ns. Time domain lular delivery of inhibitory decoy sequences in vitro and in vivo (18, 37, FLIM images were acquired using time-correlated single photon counting 38, 53, 55, 57). Peptide sequences are shown in Table I. CPDPs were TCSPC Model SPC-830 (Becker & Hickl). Images (256 3 256 pixels) synthesized by AAPPTec (Louisville, KY) or GenScript (Piscataway, were acquired with resolution varied in the range ∼0.4–0.7 mm/pixel, using NJ). The Cy3-labeled peptides were produced by CPC Scientific scan speed 1 ms/pixel. To avoid pixel intensity saturation in bright cells (Sunnyvale, CA) and LifeTein (Franklin Township, NJ). The Cy3 label and to improve the signal from dim cells, two to five overlapping scans was placed at the peptide N terminus. The purity of all CPDPs was were used for acquisition of FLIM images. Image sizes varied from 100 3 $95%. Concentrations of reconstituted peptides were determined spec- 100 to 180 3 180 mm to accommodate multiple cells. FLIM data trophotometrically (66). were analyzed with VistaVision Suite software (Vista, version 212, ISS). The Journal of Immunology 3

Fluorescence lifetime was determined using single- and biexponential intensity Each peptide in the library corresponds to a short segment of TLR decay models. Average lifetime images were generated based on pixel- sequence that presumably forms a nonfragmented patch of TIR by-pixel analysis using the combined signal of two neighboring pixels in all surface. All peptides of the library together encompass the entire directions (bin 2 = 25 pixels). The binning was used to improve the signal for analysis. The average lifetimes were calculated from fitted parameters, decay surface of TLR9 TIR. TLR9 sequence fragments were synthesized times ti and amplitudes ai, ta = Saiti. The biexponential fitting procedure was in tandem with the cell-permeating, 16-aa–long fragment of performed with one-lifetime component fixed at 3.1 ns, which was found Antennapedia homeodomain positioned N-terminally (51). Se- common for most images. To calculate the effective Fo¨rster resonance energy quences of TLR9 TIR peptides are presented in Table I. CPDPs transfer (FRET) efficiency (E), average lifetimes of the donor-only images were first evaluated at 40 mM based on their effect on cytokine (tD) and donor-acceptor images (tDA)wereused:E=12 tDA/tD. It should be noted that nonquenched donor molecules also contribute to the value of tDA, expression in BMDM stimulated with ODN 1668 for 1 h. Peptide and do so more significantly at low acceptor concentrations as was discussed 9R3, 9R34, 9R9, and 9R11 significantly suppressed TNF-a, previously (67). IL-1b, and IL-12p40 mRNA expression measured1hafterstimu- Animal experiments lation, whereas the effect on IL-6 mRNA expression was statistically significant only for 9R34 (Fig. 1A). Peptides 9R6 and 9R8 also Systemic cytokine levels were measured in plasma samples obtained 1, 3, and demonstrated some inhibitory activity but were overall less potent. 5 h after i.p. administration of 7.86 nmol of ODN 1668 or 67 nmol of Pam3C per mouse, using ELISA kits from BioLegend. CPDPs were dissolved in PBS and Peptides 9R3 and 9R34 have overlapping sequences (Table I). 9R34 administered i.p. at the dose of 200 nmol/mouse 1 h before administration of a inhibited expression of all tested cytokines when used at lower TLR agonist (37). For survival experiments, mice were sensitized to TLR9 concentrations of 10 or 20 mM (Fig. 1A). agonists by i.p. injection of 20 mg of D-Gal 30 min before administration of The three most potent inhibitory peptides identified in mRNA peptides or PBS. ODN 1668 was used at the doses of 3.93 or 7.86 nmol/mouse screens, 9R34, 9R9, and 9R11, were examined further based on

1.5 h after D-Gal. Animal survival was monitored for 50 h. All agents were Downloaded from administered i.p. in 500 ml volume of PBS. Animals had free access to food their effect on cytokine secretion assessed in 5 h supernatants of and water during the observation period. All animal experiments were con- ODN-stimulated BMDM (Fig. 1B). CPDPs 9R34, 9R9, and 9R11 ducted in accordance with the national guidelines for the care and use of reduced secretion of IL-12p40 when used at 40 mM. CPDPs 9R34 laboratory animals under a protocol that was approved by the University of and 9R11 were also effective at 20 mM. Both peptides also de- Maryland, Baltimore Institutional Animal Care and Use Committee. creased the production of TNF-a when used at 40 mM. CPDP 9R9 TLR9 TIR domain modeling did not significantly inhibit the secretion of TNF-a (Fig. 1B). a http://www.jimmunol.org/ Template search was performed using BLAST (Basic Local Alignment Apparent discrepancy between effects of 9R9 on TNF- expres- Search Tool), HMM-HMM–based lightning-fast iterative sequence search sion (Fig. 1A) and secretion (Fig. 1B) might be due to a weaker (HHBlits), and the SWISS-MODEL template library (SMTL) (68, 69). The inhibitory potency of this peptide and the transitory nature of target sequence was searched with BLAST against the primary amino acid TNF-a mRNA expression (Supplemental Fig. 1). Thus, the ab- sequence contained in the SWISS-MODEL template library (68). An a initial HHblits profile has been built using the procedure outlined in (69), sence of 9R9 effects on TNF- contents in 5-h supernatants in- followed by one iteration of HHblits against NR20. The model was built dicates that the initial reduction of mRNA expression observed in based on the target-template alignment using ProMod3 (70). All images 1-h samples (Fig. 1A) is compensated by elevated expression at were produced using the UCSF Chimera viewer (71). the later time points (Supplemental Fig. 1). Based on these results, CPDPs 9R34 and 9R11 were selected for further evaluation. Data representation by guest on October 1, 2021 To exclude the possibility that the reduction in cytokine ex- Numerical data were statistically analyzed by the one-way ANOVA using pression caused by peptides is due to general toxicity of inhibitory GraphPad Prism 5 software. CPDPs, we conducted the MTT cell viability assay. Inhibitory CPDP did not affect BMDM viability in a measurable way with or Results without concurrent TLR9 stimulation in any experimental condi- Identification of TLR9 inhibitory peptides tions tested (Supplemental Fig. 2). The TLR9 peptide library was designed similarly to the libraries TLR stimulation causes formation of MyDDosome, a signaling used in our previous studies of TLR-derived peptides (18, 37, 38). complex composed through interactions of IRAK and MyD88

Table I. Sequences of TLR9 TIR-derived decoy peptides

Peptide Sequence Predominant Structural Region 9R1 GRQSGRDEDALPYD N-terminal segment 9R2 DKTQSAVADWVYNE AA loop, a-helix Aa 9R3 RGQLEECRGRWALR a-Helix A, AB loop 9R34 GRWALRLCLEERD AB loop, b-strand B, N-terminal residues of BB loop 9R34-DN ALRLCLEERD AB loop, b-strand B, BB loop 9R34-DC GRWALRLCLE AB loop, b-strand B 9R34-C/S GRWALRLSLEERD AB loop, b-strand B, BB loop 9R34-DL1 GRWALRACLEERD AB loop, b-strand B, BB loop 9R34-DL2 GRWALRACAEERD AB loop, b-strand B, BB loop 9R4 LEERDWLPGKTLFE b-Strand B, BB loop, a-helix B 9R5 NLWASVYGSRKT a-Helix B, BC loop, b-strand C 9R6 HTDRVSGLLRASFL a-Helix C 9R7 LAQQRLLEDRKD a-Helix C, CD loop, b-strand D 9R8 SPDGRRSRYVR DD loop, a-helix D 9R9 RYVRLRQRLCRQS a-Helix D, DE loop 9R10 QSVLLWPHQPSGQ DE loop, b-strand E, EE loop 9R11 RSFWAQLGMALTRD a-Helix E 9R12 NHHFYNRNFCQGPT C-terminal segment aSecondary structure elements of the TIR domain are consecutively indicated by letters, with A indicating the most N-terminal element. 4 MECHANISMS OF ADAPTER RECRUITMENT TO ACTIVATED TLR9 Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 1. Effects of TLR9-derived CPDP on ODN 1668–induced cytokine expression and secretion. Mouse BMDM were incubated in the presence of a 10, 20, or 40 mM decoy peptide for 30 min prior to stimulation with ODN 1668 (1 mM). (A) Cytokine mRNA expression measured 1 h after ODN 1668 challenge and normalized to the expression of hypoxanthine phosphoribosyltransferase (HPRT). Data represent the mean 6 SEM of at least three inde- pendent experiments. The statistical significance of changes in cytokine mRNA was determined by a one-way ANOVA test. *p , 0.001. (B) Cytokine concentrations in supernatants were evaluated by ELISA 5 h after cell stimulation. Data represent mean 6 SEM of three independent experiments. The statistical significance of changes in cytokine levels was determined by the one-way ANOVA test. *p , 0.01. (C) BMDM were lysed 3 h after cell treatment with ODN 1668, lysates were immunoprecipitated with anti-MyD88 Ab, and immune complexes were assessed with anti-IRAK4 Ab. Data show a representative blot of two separate experiments. (D) Segments of TIR domains represented by CPDPs. (E) Positions of inhibitory peptides 9R34, 9R9, and 9R11. death domains (21, 22). To test if peptides inhibit MyDDosome suggesting that the 9R34 effects on the LPS-induced TNF-a formation, MyD88 was immunoprecipitated from lysates of ODN- mRNA is transient. stimulated BMDM and the precipitates analyzed for IRAK4 Inhibitory properties of modified 9R34 peptides presence (72). Pretreatment of BMDM with CPDP 9R34 or 9R11, but not with control peptide 9R2, prevented ODN 1668–induced The experiments described above suggested that 9R34 is the most MyDDosome formation (Fig. 1C). potent inhibitor from the peptides screened (Fig. 1). We noted, however, that millimolar stock solutions of this peptide were un- Specificity of signaling inhibition by TLR9-derived peptides stable; the peptide solutions were also unstable in the presence of CPDPs 9R34 and 9R11 were examined for inhibition of TLR1/2, high protein concentrations. Therefore, several modifications of TLR4, and TLR7 signaling. BMDM were stimulated with Pam3C, 9R34 were generated to determine more precisely the epitopes E. coli LPS, or R848, and TNF-a mRNA was measured 1 h after responsible for signaling inhibition and to improve peptide solu- stimulation. TLR9 inhibitory peptides did not inhibit signaling bility. The inhibitory potency of modified peptides was compared induced by TLR1/2 or TLR7 agonist. Peptide treatment reduced with that of the parent peptide. Five modifications of 9R34 were the TNF-a mRNA expression in LPS-stimulated cells by ∼50– tested. Two modifications, 9R34-DN and 9R34-DC, are 9R34 70% (Fig. 2). This level of inhibition of the LPS-stimulated ex- deletion variants that were shortened by three amino acids at either pression was significantly less than that of the ODN-stimulated N or C terminus. Three other modifications contained single expression, in which more than 99% of expression was inhibited amino acid replacements. 9R34-C/S had cysteine substituted for (Fig. 1A). Accordingly, 9R34 did not reduce the LPS-induced serine. 9R34-DL1 and 9R34-DL2 had one or two middle sequence TNF-a secretion measured in 5-h BMDM supernatants (Fig. 2B), leucines replaced by alanines (Table I). The results suggested that The Journal of Immunology 5 Downloaded from

FIGURE 2. Specificity of TLR inhibition by TLR9-derived CPDP. (A) BMDM were treated with 40 mM of indicated peptides for 30 min prior to stimulation with ODN 1668 (1 mM), E.coli LPS (0.1 mg/ml), R848 (2.85 mM), or Pam3C (0.33 mM). TNF-a mRNA expression was measured 1 h after

TLR stimulation and normalized to the expression of HPRT. Data show mean 6 SEM of at least three independent experiments. The statistical significance http://www.jimmunol.org/ of changes was determined by a one-way ANOVA test. *p , 0.01. (B) BMDM were treated with 20 or 40 mM of 9R34 for 30 min prior to stimulation with E.coli LPS (0.1 mg/ml). TNF-a concentration in supernatants was evaluated by ELISA 5 h after cell stimulation. Data represent mean 6 SEM of three independent experiments.

9R34-DN, 9R34-DC, and 9R34C/S were as potent as the parent the binding between donor and acceptor moieties. A larger decrease in peptide, whereas 9R34-DL1 and 9R34-DL2 inhibited TLR9 sig- average lifetime corresponds to a larger number of donor–acceptor naling less potently (Fig. 3). The 9R34-DN stock solutions were pairs present within the sample, indicating an effective FRET stable at concentrations up to 10 mM. This peptide remained (Fig. 4C). FRET efficiencies presented in Fig. 4C were calculated from by guest on October 1, 2021 soluble in bovine serum at high micromolar concentrations. These the average lifetimes corresponding to entire images that contained observations show that 9R34 inhibits TLR9 signaling sequence- multiple cells as in images shown in Fig. 4A, using data obtained in specifically and suggest that 9R34-DN is better suited for in vivo three independent experiments. applications. The strongest peptide-induced quenching was detected for the TLR9 TIR–Cer-expressing cells treated with Cy3–9R34-DN. Selectivity of TIR binding by TLR9-derived CPDPs This peptide quenched the TLR9 TIR–Cer fluorescence dose- To identify potential binding partners of inhibitory peptides, we dependently with FRET efficiency varying in the range of 9–23% used FRET approach coupled with FLIM (18, 37, 67). HeLa cells (Fig. 4C). Fluorescence lifetime of TLR9 TIR–Cer construct was were transiently transfected with an expression vector that encodes decreased even in cells treated with the minimal peptide con- a TIR domain fused with Cerulean (Cer) fluorescent protein. Cells centration used (2.5 mM) (Fig. 4). Cy3–9R34-DN also notably were treated with various concentrations of a Cy3-labeled inhib- quenched the fluorescence of cells that expressed the TIRAP–Cer itory peptide 48 h after transfection (18). Cy3 is a suitable FRET (Fig. 4C). The Cy3–9R34-DN effect on TIRAP–Cer fluorescence acceptor for Cer fluorescence. A direct molecular interaction of a was weaker than on that of the TLR9 TIR–Cer, as suggested by Cy3-labeled peptide with Cer-fused TIR domain should quench the absence of the effect of the lowest peptide concentration Cer fluorescence, leading to a decreased fluorescence lifetime. (Fig. 4C). The quenching observed for the control TIR-peptide TLR9, MyD88, and TIRAP were selected as potential primary pair, TIRAP–2R9, was slightly higher than for the TLR9–9R34- binding partners for Cy3–9R34-DN and Cy3–9R11. Cer-fused DN pair (Fig. 4). The average fluorescence lifetime of Cer-labeled TLR2 and TLR4 TIR domains, and Cer not fused to a TIR, MyD88 TIR was not affected by Cy3–9R34-DN at lower peptide were used as controls. We previously demonstrated that the TLR2- concentrations (2.5 and 10 mM) (Fig. 4C). The quenching of derived peptide 2R9 binds TIRAP TIR (37). Therefore, the MyD88 TIR–Cer was notable at 40 mM peptide concentration but TIRAP–Cer with Cy3–2R9 was used as a positive binding control. considerably less in comparison with the effect on TLR9 and Fig. 4A shows FLIM images of HeLa cells transfected with dif- TIRAP TIR–Cer (Fig. 4C). Cy3–9R34-DNat40mM also slightly ferent Cer–TIR constructs and treated with inhibitory peptides at affected the fluorescence lifetime of control TLR2 TIR–Cer, various concentrations. The images show the average fluorescence TLR4 TIR–Cer, TRAM–Cer, and Cer not fused to a TIR domain lifetime of Cer component calculated on the pixel-by-pixel basis (Fig. 4, Supplemental Fig. 3). The effects on these constructs show using the biexponential model with one component having a fixed the level of unspecific quenching of Cer by Cy3-labeled peptides lifetime (67). Histograms of Fig. 4B demonstrate the distribution at high peptide concentrations. of average fluorescence lifetime in corresponding images of Cy3-9R11 dose-dependently quenched the fluorescence of TLR9 Fig. 4A. A progressive shift of average fluorescence lifetime toward TIR–Cer with FRET efficiency varying in the range of 3–27% lower values observed with increasing peptide concentrations indicates (Fig. 4C). In addition, Cy3–9R11 induced FRET from TIRAP–Cer 6 MECHANISMS OF ADAPTER RECRUITMENT TO ACTIVATED TLR9 Downloaded from http://www.jimmunol.org/

FIGURE 3. Inhibition of TLR9 signaling by 9R34 analogs. Mouse BMDM were incubated in the presence of a 10, 20, or 40 mM decoy peptide for 30 min prior to stimulation with ODN 1668 (1 mM). TNF-a (A), IL-1b (B), and IL-12p40 (C)mRNAexpressionwasmeasured1hafterODN 1668 challenge and normalized to the expression of hypoxanthine phosphoribosyltransferase (HPRT). CPDP sequences are shown in Table I. Data represent mean 6 SEM of at least three independent experiments. by guest on October 1, 2021 but only when used at higher concentrations of 10 or 40 mM effect on the ODN-induced cytokines, 9R34-DN did not affect (Fig. 4C). The quenching of MyD88-Cer fluorescence lifetime by cytokine levels induced by TLR2 agonist Pam3C (Fig. 5B), Cy3–9R11 was also detectable but less effective (Fig. 4C). Un- thereby confirming the cell culture data presented in Fig. 2A. Our specific Cer quenching by Cy3–9R11 was observed at a peptide previous studies have identified 2R9, a TLR2-derived peptide that concentration of 40 mM, similarly to the effects of Cy3–9R34-DN inhibits TLR2, TLR4, TLR7, and TLR9, primarily because it (Fig. 4). targets an adapter shared by these TLRs (37). As expected, 2R9 In summary, the FLIM experiments suggest that both 9R34-DN suppressed cytokine induction elicited by either ODN 1668 or and 9R11 bind TLR9 and TIRAP TIR. 9R11 also interacted with Pam3C, whereas 9R34-DN selectively inhibited only the ODN the MyD88 TIR but with a lower affinity, whereas 9R34-DN did 1668–induced cytokines (Fig. 5A, 5B). Negative control peptide not interact with MyD88. 9R2 [this peptide did not inhibit TLR9 in cell culture experiments (Fig. 1)] was tested in vivo only with respect to ODN-induced D 9R34- N blunts ODN 1668–induced systemic cytokine cytokines; as expected, 9R2 did not significantly reduce the response in vivo and protects against TLR9-induced lethality ODN-induced cytokine levels in mice (Fig. 5A). These data To test the in vivo efficacy and specificity of 9R34-DN, C57BL/6J clearly demonstrate that TLR9 inhibition by 9R34-DN is selective. mice were mock-treated or treated i.p. with a single dose of in- With the exception of LPS, administration of a purified TLR hibitory or control peptide (200 nmol/mouse) and challenged with agonist is not lethal to mice. Mice, however, can be sensitized to TLR TLR9 agonist ODN 1668 (7.86 nmol/mouse) or TLR2/1 agonist agonists by pretreatment with D-Gal (73). We tested if 9R34-DN Pam3C (67 nmol/mouse) 1 h later. Plasma concentrations of TNF- protects D-Gal–pretreated mice from TLR9-induced lethality. Mice a and IL-12p40 were measured 1, 3, and 5 h after the challenge. were first i.p. administered with D-Gal at a dose of 20 mg/mouse. Both TLR agonists increased concentrations of TNF-a and IL- Thirty minutes later, mice were mock-treated with PBS or treated 12p40 in 1-h samples from below the limit of detection to ∼1.2 with a CPDP at a dose of 200 nmol/mouse. ODN 1668 was ad- and ∼6 ng/ml in the case of ODN 1668, and to 0.12 and 1 ng/ml ministered 1 h after peptide treatment at the dose of 3.93 or 7.86 after treatment with Pam3C (Fig. 5). The concentration of TNF-a nmol/mouse. Either ODN 1668 dose used induced 100% lethality in dropped rapidly to near basal levels in 3-h samples, whereas the D-Gal–pretreated mice (Fig. 5). 9R34- or 9R34-DN–adminis- circulating IL-12p40 continued to grow and increased more than tered i.p. provided full protection to mice that received ODN 1668 10-fold in the next 2 h, peaking at ∼110 ng/ml in the ODN 1668– at the lower dose of 3.93 nmol/mouse (Fig. 5C). TM4, a potent treated mice and at ∼11 ng/ml in mice challenged with Pam3C inhibitor of TLR4 derived from TRAM, fully protected mice against (Fig. 5). ODN 1668–treated mice pretreated with 9R34-DN mortality caused by LPS (55). TM4 was less effective than 9R34- eliminated the 1-h TNF-a peak and decreased the IL-12p40 peak DN in protection against ODN-induced lethality (Fig. 5C). Partial by ∼25-fold in the 3-h samples (Fig. 5A). In a sharp contrast to the protection observed after TM4 treatment apparently is due to The Journal of Immunology 7 Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 4. Effect of Cy3-labeled CPDPs on Cer fluorescence lifetime. (A) FLIM images of HeLa cells expressing Cer not fused with a TIR domain (upper row) or fused with TLR9 TIR or TIRAP–Cer (middle and bottom rows) and treated with a Cy3-labeled decoy peptide for 1 h. FLIM images show the average fluorescence lifetime of Cer component. Images in different rows were taken at different digital magnifications as described in Materials and Methods.(B) Histograms show frequencies of occurrence of particular average lifetimes in images shown in (A). (C) FRET efficiency for different TIR–peptide pairs. Data represent mean 6 SEM of two independent experiments. suppression of TLR4-mediated inflammatory response and conse- Comparison of the inhibition pattern within TLR9 peptide quent injury caused by endogenous danger-associated molecules that library (Fig. 1) with that of previously screened TLR libraries may be produced in the body in response to mechanical, chemical, (18, 37, 38) suggests that the location of TIR segments capable of or biological injury (74). Both 9R34-DN and TM4 were less pro- blocking TIR–TIR interactions is generally conserved in TLRs, tective when tested in mice treated with a higher ODN 1668 dose yet there are some differences. Thus, peptides derived from the (7.86 nmol/mouse). Nevertheless, 9R34-DN rescued more than 50% AB loop (i.e., region 3 peptides) were inhibitory in TLR4, TLR2, of mice challenged with this higher dose of ODN 1668 (Fig. 5C). and TLR9 sets (Fig. 1) (18, 37). Inhibitory ability of BB loop peptides (these peptides were designated as region 4 peptides), Discussion however, differed in these three peptide sets. The BB loop pep- This study examines the peptide library derived from the TLR9 TIR tides of the TLR4 and TLR2 sets inhibited (18, 36), whereas the for a dominant negative effect on TLR9 signaling and so extends BB loop peptide of TLR9 did not (Fig. 1) [the TLR2 BB loop our previous research that screened analogous peptides from TIR peptide inhibited TLR2 only when leucine located at the border of domains of TLR4, TLR2, their adapters, and their coreceptors (18, TLR2 regions 3 and 4 was included (36, 37)]. It should be noted, 37, 38, 53, 55, 56). Each peptide of the TLR9 library represents a however, that peptides from AB or BB loop of TLR2 coreceptors patch of TIR surface that might mediate a TIR–TIR interaction did not inhibit the TLR2-mediated signaling (38, 57). In the TLR9 within the primary receptor complex (Fig. 1, Table I). Screening set, peptides derived from the region centered on the region 3 and results have suggested that peptides 9R34, 9R34 modifications, 4 border, peptides 9R34 and 9R34-DN, inhibited TLR9 signaling 9R9, and 9R11 inhibit TLR9 signaling. These inhibitory peptides more potently than did 9R3 (Figs. 1, 6). Notably, all inhibitory represent three noncontiguous surface regions of TLR9 TIR, each peptides derived from region 3 or 4 of a TLR targeted the TLR of which might represent a separate TIR–TIR interface (Fig. 1). TIR of corresponding dimerization partner (18, 37, 67), suggesting 8 MECHANISMS OF ADAPTER RECRUITMENT TO ACTIVATED TLR9 Downloaded from http://www.jimmunol.org/

FIGURE 5. 9R34-DN inhibits ODN 1668–induced systemic cytokines in vivo and protects D-Gal–pretreated mice from ODN 1668–induced lethality. (A) Plasma TNF-a and IL-12p40 levels in mice following administration of ODN 1668. (B) Plasma TNF-a and IL-12p40 levels in mice following admin- istration of Pam3C. Data shown in (A) and (B) represent the mean 6 SEM obtained in at least three independent experiments. The statistical significance of changes in cytokine concentrations was determined by a two-way ANOVA test. *p , 0.001. (C) Survival of D-Gal–pretreated mice after treatment with by guest on October 1, 2021 ODN 1668. The statistical significance of changes in mortality and survival time was determined by the Gehan–Breslow–Wilcoxon test. *p , 0.01. that this region may serve as the receptor dimerization surface. peptides inhibited cognate receptors and bound MyD88 and TIRAP, The TLR9 inhibitory peptide 9R34-DN followed the same pattern respectively (38). These findings suggest that a-helix D of TLR TIR and bound the receptor TIR domain (Fig. 4). These data suggest domains is an essential part of the adapter recruitment site. The that the position of TLR dimerization interface is generally con- TLR9 peptide 9R9 inhibited signaling weaker than did 9R34 or 9R11 served in the TLR family, with one TLR TIR dimerization inter- (Fig. 1A, 1B). For this reason, we could not directly confirm if this face being formed by a broad area that may include residues of peptide indeed targets a TLR adapter. The area that corresponds to AB and/or BB loops and b-strand B (Figs. 1D, 1E, 6A, 6B). 9R9 and includes helix D and neighboring loops is shown as surface Because the position of this dimerization interface slightly varies 3(S3)inFig.6. in different TIRs and is formed by nonconserved amino acid se- The third inhibitory peptide of the TLR9 library, 9R11, corre- quences, this broad area is shown as surface 1 (S1) in Fig. 6. sponds to a-helix E (Fig. 1D). Interestingly, peptide 4RaE from Binding studies have suggested that 9R34-DN binds not only a-helix E of TLR4 also inhibited cognate signaling (18). Peptides TLR9 TIR but also TIRAP TIR (yet with apparently lower af- from a-helices E of TLR1, 2, or 6, however, did not affect TLR2 finity) but does not interact with MyD88 TIR (Fig. 4). Such a signaling, although TLR1 and TLR6 peptides from a neighboring binding profile of 9R34-DN is consistent with models of primary region that included DE and EE loops and b-strand E inhibited receptor complex in which TLR TIR dimer is formed by an (37, 38). A cell-based binding assay has demonstrated that 4RaE asymmetric interaction in which S1 of one TLR TIR domain binds the TLR4 TIR (18). Consistently, new data show that the mediates TLR TIR dimerization, whereas the same region of the structurally homologous TLR9 peptide, 9R11, also binds to the second TLR of the dimer is available for TIRAP recruitment receptor TIR (Fig. 4). 9R11, however, demonstrated a multi- (Fig. 6, Ref. 18, 75, 76). specific binding, as this peptide also bound TIRAP and, to a lesser Peptides from region 9 (this region includes a-helix D and DE degree, MyD88 (Fig. 4C). Thus, inhibition profiles obtained for loop) were inhibitory in all TLR TIR libraries tested to date (18, TLR2, TLR4, and TLR9 peptide libraries collectively reveal a 37, 38). Notably, all previously identified region 9 TLR peptides significant similarity and suggest a common mode of TLR TIR preferentially bound adapter TIR domains. Thus, 4R9, a TLR4 dimerization in which one TIR of the dimer interacts with the peptide, did not bind TLR4 TIR (18) but coimmunoprecipitated other through S1, a large area that is centered on the b-strand B with TIRAP (56). 2R9, a peptide from region 9 of TLR2, was and may include either BB or AB loop or parts of both (Fig. 6). S1 TIRAP-selective in a cell-based FRET assay and bound recombinant appears to interact asymmetrically with the region generally lo- TIRAP with nanomolar affinity (37). TLR1 and TLR6 region 9 cated on the opposite surface of the TIR in the vicinity of a-helix The Journal of Immunology 9 Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 6. Putative TIR interaction sites and the modes of TIR domain interactions in TLR signaling complexes. (A) Four TIR surfaces (S1–S4) that mediate the assembly of primary TLR signaling complexes. S1 (yellow highlight) and S4 (orange highlight) are located on opposite TIR sides near b-strands B and E, the strands that form lateral, surface-exposed edges of the b-sheet. S2 (shown in green) is formed by a-helices B and C, whereas S3 (dark blue) is formed by a-helix D and may include adjacent loops. In the TLR9 peptide set, S1, S3, and S4 are represented by peptides 9R34, 9R9, and 9R11, respectively, whereas of the peptides that should jointly form S2 (i.e., 9R5 and 9R6), only 9R6 inhibited weakly (Fig. 1). The images show a homology model of TLR9 TIR computationally built using TLR2 TIR protein database file (pdb) (pdb identifier: 1o77) as a template. (B) TIR domain interactions that initiate intracellular TLR signaling. Two upper TIR domains of this panel represent dimerized receptor TIR domains, whereas two lower TIRs represent adapters. The upper TIRs show the same TLR9 TIR model as in (A) but shown in the ribbon style. Two lower TIRs are representations of TIRAP TIR model (pdb identifier: 5UZB, chain A). S1 and S4 mutually interact to mediate the intrastrand TIR–TIR interactions of the complex. S1 and S4 of TLR TIRs form receptor dimers. These surfaces may also recruit adapters through lateral, intrastrand interactions, thereby leading to elongation of the complex in either one or two directions. Color coding of TIR segments in (B) correspond to that in (A). S2 interacts with S3, forming the interstrand interactions. S2 and S3 are located on the same TIR hemisphere; these surfaces mediate interstrand receptor–adapter and adapter–adapter interactions [only receptor–adapter S2–S3 interactions are shown in (B)]. S2 and S3 of one TIR domain interact with cognate surfaces of two separate TIR domains of the opposite strand in the cooperative mode (18). (C) Schematics of adapter recruitment to activated TLR9. (Figure legend continues) 10 MECHANISMS OF ADAPTER RECRUITMENT TO ACTIVATED TLR9 and b-strand E (surface 4; S4). It should be noted that the surface interact through slightly different structural regions. Particularly, area near helix E is highly fragmented; therefore, in addition to the our data suggest that the AB loop of TLRs plays a more im- helix E residues, the second dimerization interface may include portant role in TIR–TIR recognition than the BB loop. The isolated residues from neighboring regions (i.e., b-strand E and finding that TIR–TIR interface positions are structurally con- the adjacent loops), the N-terminal half of b-strand D, and resi- served is remarkable considering sequence dissimilarity of cor- dues of the most C-terminal round of a-helix D. The TLR TIR responding segments. An example of sequence dissimilarity interaction mechanism suggested by TIR peptide screenings is of corresponding interface regions is sequences of TIRAP, generally in line with several earlier proposed, asymmetric TIR– TLR4, TLR2, and TLR9 inhibitory peptides from region 9 and TIR interaction models, in which a broad b-strand B region of one 11 (TR9: AAYPPELRFMYYVD; 4R9: LRQQVELYRLLSR; TIR interacts with the opposite surface of the second TIR through 2R9: PQRFCKLRKIMNT; TR11: GGFYQVKEAVIHY; 4aE: a-helix E and/or neighboring regions (18, 75, 77). HIFWRRLKNALLD; 9R11: RSFWAQLGMALTRD). The most recent and detailed model for the assembly of TIR The surfaces that mediate TIR–TIR interactions are highlighted signaling complexes that underlie the adapter recruitment to ac- in Fig. 6 as S1–4. In TLR9, S1 and S4 correspond to peptides tivated TLRs was proposed by Ve et al. (77). This model resulted 9R34 and 9R11. These surfaces mediate dimerization of TLR from the study of filamentous structures spontaneously formed in TIRs and adapter recruitment through lateral intrafilament inter- solution by TIR domains of TIRAP, MyD88, or a mixture of these actions (Fig. 6B). S2 and S3 in TLR9 correspond to peptides 9R6 proteins. Ve et al. (77) postulated that interactions of TIR domains and 9R9 [9R6 exhibited partial activity in respect to TNF-a and in the primary TLR signaling complexes resemble interactions in IL-12p40 mRNA (Fig. 1A)]. These surfaces recruit adapter TIRs the double-stranded protofilaments that comprise larger TIRAP that initiate the formation of the second strand of the primary TIR filaments (Fig. 6C). Two types of TIR–TIR interactions mediate complex (Fig. 6B). The second strand stabilizes the complex and Downloaded from the assembly of the double-stranded TIR filaments. TIR interac- provides sufficient number of MyD88 molecules to initiate tions within the strands occur through two regions that are located MyDDosome formation (Fig. 6). near b-strands that form two opposite edges of the central b-sheet, Experimental data support this model of TIR domain interactions strands B and E [these areas correspond remarkably well to S1 and in several ways. One prediction from the filamentous, double- S4 suggested by decoy peptide screenings (Fig. 6)]. stranded structure of the signaling-initiating TIR complex is that Interactions of TIR domains that belong to different strands of both surfaces that mediate TLR TIR dimerization may be re- http://www.jimmunol.org/ the double-stranded TIRAP protofilament are mediated by two sponsible for adapter recruitment through the intrastrand interac- separate, mutually interacting surfaces (77) (Fig. 6). One inter- tions (Fig. 6B, 6C). Our data confirm this prediction because both strand interface of TIRAP protofilaments included residues of inhibitory peptides that correspond to two separate TLR9 dimer- a-helix D and CD loop; this interface corresponds to S3 (Fig. 6). ization interfaces, 9R34-DN and 9R11, demonstrate multispecific The second interstrand interface of TIRAP protofilaments was binding. Peptide 9R34-DN, from S1, binds TLR9 and TIRAP TIR jointly formed by helices B and C (77). The corresponding resi- domains, not MyD88, whereas 9R11 (this peptide represents S4) dues of TLR9 are indicated as surface 2 (S2) in Fig. 6. Impor- interacts with all relevant TIR domains (Fig. 4C). tantly, interaction surfaces in the TIRAP protofilament match to Another consequence from the adapter recruitment model suggested by guest on October 1, 2021 the putative TIR–TIR interfaces suggested by previous studies of by TIRAP filament structure and decoy peptide screenings is that TIRAP-derived decoy peptides (53, 78, 79). Thus, early studies adapters are independently recruited to filament ends through different suggested that the BB loop peptide can block important TIRAP surfaces of the receptor dimer that may have different binding speci- functions (78, 79). Couture et al. (53) later found additional ficities. This, in turn, implies that the initial receptor complex can peptides (TR3, TR5, TR6, TR9, and TR11) that inhibit TIRAP- elongate from either or both ends (Fig. 6C–F). In the case of TLR9, the mediated signaling and specified that signaling inhibition by peptide from S1 binds TIRAP, not MyD88, whereas the peptide from TIRAP BB loop peptide critically depends on the N-terminal opposite S4 can bind either TIRAP or MyD88 (Figs. 4, 6C). Such leucine. TIRAP-derived inhibitory peptides TR5 and TR6, first binding preferences suggest that the initial complex can elongate bi- reported by Couture et al. (53), correspond to the second inter- directionally in the presence of TIRAP and MyD88 (Fig. 6C). In the filament interface of TIRAP protofilaments, indicated as S2 in absence of TIRAP, the TLR9 complex can elongate only in one di- Fig. 6. Peptide TR9 represented S3, whereas TR11 and TR3, re- rection, through MyD88 recruitment only to S4 of the TLR9 TIR spectively, corresponded to S4 and S1 (77). dimer (Fig. 6D). Such models reconcile previous discussions on the Thus, screening of TLR9-derived peptides reveals significant role of TIRAP in TLR9 signaling. Although early studies concluded topological similarity of positions of new inhibitory peptides with that TIRAP is not required for TLR9 signaling (80, 81), it was later those seen in previously screened peptide libraries. Moreover, the discovered that TLR9 responses to certain viruses are critically di- inhibitory peptides come from regions that are structurally ho- minished in TIRAP absence and the TIRAP-targeting peptide strongly mologous to TIR–TIR interfaces in the TIRAP protofilaments suppresses TLR9 signaling in wild-type cells (27, 37). The TIRAP- (77). These findings together support the hypothesis of structural dependent bidirectional elongation of protosignalosomes explains both similarity of receptor and adapter TIR interactions in the signal- the ability of TLR9 to signal in TIRAP-deficient models and the initiating complexes; yet particular TIRs of the complex may sensitivity of TLR9 responses to TIRAP targeting in wild-type cells.

TLR9 TIR dimers recruit adapters through intrastrand and interstrand interactions. The TLR9 TIR complex elongates bidirectionally in the presence of TIRAP and MyD88. Bidirectional elongation of TLR9 signaling protofilaments is possible because of the ability of TLR9 S1 to interact with TIRAP and the ability of TLR9 S4 to bind either TIRAP or MyD88. Adapter–adapter S2–S3 interstrand interactions occur after elongation of the receptor filament through intrafilament S1–S4 interactions. (D) Schematics of MyD88 recruitment to activated TLR9 in the absence of TIRAP. In the absence of TIRAP, the TLR9 filaments elongate unidirectionally only through S4 of the dimer. (E and F) TLR2/1 and TLR2/6 signaling filament elongation models. TLR2 filaments elongate unidirectionally because S1 of both TLR2 coreceptors is inept in binding the TLR2 adapters (38, 57). Strong TIRAP-dependence of TLR2 signaling is due to high affinity TLR2-TIRAP interaction through S3 of TLR2 (37). The Journal of Immunology 11

Latty et al. (82) recently applied single-molecule fluorescence 14. Ohto, U., T. Shibata, H. Tanji, H. Ishida, E. Krayukhina, S. Uchiyama, microscopy to study the formation of MyDDosomes in response to K. Miyake, and T. Shimizu. 2015. Structural basis of CpG and inhibitory DNA recognition by Toll-like receptor 9. Nature 520: 702–705. TLR4 stimulation. Intriguingly, the authors observed that a large 15. Kirk, P., and J. F. Bazan. 2005. Pathogen recognition: TLRs throw us a curve. LPS dose caused formation of “super MyDDosomes,” the sig- Immunity 23: 347–350. 16. Huyton, T., J. Rossjohn, and M. Wilce. 2007. Toll-like receptors: structural naling complexes that are composed of twice as many MyD88 pieces of a curve-shaped puzzle. Immunol. Cell Biol. 85: 406–410. molecules as regular MyDDosomes (82). These observations 17. Botos, I., D. M. Segal, and D. R. Davies. 2011. The structural biology of Toll- support the possibility of bidirectional protofilament elongation like receptors. Structure 19: 447–459. 18. Toshchakov, V. Y., H. Szmacinski, L. A. Couture, J. R. Lakowicz, and that may cause formation of two MyDDosomes per one TLR4 S. N. Vogel. 2011. Targeting TLR4 signaling by TLR4 Toll/IL-1 receptor dimer simultaneously, analogously to the schematic shown in domain-derived decoy peptides: identification of the TLR4 Toll/IL-1 receptor Fig. 6C for TLR9. domain dimerization interface. J. Immunol. 186: 4819–4827. 19. Ferrao, R., J. Li, E. Bergamin, and H. Wu. 2012. Structural insights into the Our previous studies demonstrated that peptide derived from S1 assembly of large oligomeric signalosomes in the Toll-like receptor-interleukin-1 of TLR2, 2R3, inhibits TLR2/1 and TLR2/6 signaling and binds receptor superfamily. Sci. Signal. 5: re3. both TLR2 coreceptors, whereas peptides from S1 of TLR1 or 20. Gay, N. J., M. F. Symmons, M. Gangloff, and C. E. Bryant. 2014. Assembly and localization of Toll-like receptor signalling complexes. Nat. Rev. Immunol. 14: TLR6 do not inhibit (37, 38). We also found that S3 peptides from 546–558. TLR1, TLR2, and TLR6 inhibit corresponding signaling and bind 21. Lin, S. C., Y. C. Lo, and H. Wu. 2010. Helical assembly in the MyD88-IRAK4- MyD88 or TIRAP (37, 38). These findings suggest that primary IRAK2 complex in TLR/IL-1R signalling. Nature 465: 885–890. 22. Motshwene, P. G., M. C. Moncrieffe, J. G. Grossmann, C. Kao, M. Ayaluru, TLR2 complexes elongate in the unidirectional mode, as sche- A. M. Sandercock, C. V. Robinson, E. Latz, and N. J. Gay. 2009. An oligomeric matically shown in Fig. 6E and 6F. signaling platform formed by the Toll-like receptor signal transducers MyD88 In summary, this study identifies TLR9-derived inhibitory and IRAK-4. J. Biol. Chem. 284: 25404–25411.

23. Medzhitov, R., and T. Horng. 2009. Transcriptional control of the inflammatory Downloaded from peptides that block TLR9 signaling in vitro and in vivo and suggests response. Nat. Rev. Immunol. 9: 692–703. a common mode of TIR domain interactions in the primary receptor 24. Han, K. J., X. Su, L. G. Xu, L. H. Bin, J. Zhang, and H. B. Shu. 2004. Mech- complex. Interacting in that mode, TIR domains form a double- anisms of the TRIF-induced interferon-stimulated response element and NF- kappaB activation and apoptosis pathways. J. Biol. Chem. 279: 15652–15661. stranded, parallel structure that can elongate from one or both 25. Kim, M. H., D. S. Yoo, S. Y. Lee, S. E. Byeon, Y. G. Lee, T. Min, H. S. Rho, ends. Interactions within each filament are mediated by regions M. H. Rhee, J. Lee, and J. Y. Cho. 2011. The TRIF/TBK1/IRF-3 activation located near b-strands that form the edges of TIR b-sheet, whereas pathway is the primary inhibitory target of resveratrol, contributing to its broad- spectrum anti-inflammatory effects. Pharmazie 66: 293–300. http://www.jimmunol.org/ interfilament interactions occur through two sites, one of which is 26. Kagan, J. C., and R. Medzhitov. 2006. Phosphoinositide-mediated adaptor re- predominantly formed by helix D and another which combines cruitment controls Toll-like receptor signaling. Cell 125: 943–955. 27. Bonham, K. S., M. H. Orzalli, K. Hayashi, A. I. Wolf, C. Glanemann, residues of helices B and C. W. Weninger, A. Iwasaki, D. M. Knipe, and J. C. Kagan. 2014. A promiscuous lipid-binding protein diversifies the subcellular sites of toll-like receptor . Cell 156: 705–716. Disclosures 28. Dunne, A., M. Ejdeback, P. L. Ludidi, L. A. O’Neill, and N. J. Gay. 2003. A.J. and V.Y.T. have a patent pending entitled “TLR9 inhibitors to suppress Structural complementarity of Toll/interleukin-1 receptor domains in Toll-like inflammatory response to pathogens” filed through the University of Mary- receptors and the adaptors Mal and MyD88. J. Biol. Chem. 278: 41443–41451. land, Baltimore Office of Research and Development. The other authors 29. Patterson, N. J., and D. Werling. 2013. To con protection: TIR-domain con- have no financial conflicts of interest. taining proteins (Tcp) and innate immune evasion. Vet. Immunol. Immunopathol.

155: 147–154. by guest on October 1, 2021 30. Narayanan, K. B., and H. H. Park. 2015. Toll/interleukin-1 receptor (TIR) domain-mediated cellular signaling pathways. Apoptosis 20: 196–209. References 31. Rock, F. L., G. Hardiman, J. C. Timans, R. A. Kastelein, and J. F. Bazan. 1998. A 1. Kawai, T., and S. Akira. 2011. Toll-like receptors and their crosstalk with other family of receptors structurally related to Toll. Proc. Natl. innate receptors in infection and immunity. Immunity 34: 637–650. Acad. Sci. USA 95: 588–593 . 2. Kornblit, B., and K. Mu¨ller. 2017. Sensing danger: toll-like receptors and out- 32. Xu, Y., X. Tao, B. Shen, T. Horng, R. Medzhitov, J. L. Manley, and L. Tong. come in allogeneic hematopoietic stem cell transplantation. Bone Marrow 2000. Structural basis for signal transduction by the Toll/interleukin-1 receptor Transplant. 52: 499–505. domains. Nature 408: 111–115. 3. Takeda, K., and S. Akira. 2005. Toll-like receptors in innate immunity. Int. 33. Sims, J. E., C. J. March, D. Cosman, M. B. Widmer, H. R. MacDonald, Immunol. 17: 1–14. C. J. McMahan, C. E. Grubin, J. M. Wignall, J. L. Jackson, S. M. Call, et al. 4. Lai, Y., and R. L. Gallo. 2008. Toll-like receptors in skin infections and in- 1988. cDNA expression cloning of the IL-1 receptor, a member of the immu- flammatory diseases. Infect. Disord. Drug Targets 8: 144–155. noglobulin superfamily. Science 241: 585–589. 5. O’Neill, L. A., and A. G. Bowie. 2007. The family of five: TIR-domain- 34. Schneider, D. S., K. L. Hudson, T. Y. Lin, and K. V. Anderson. 1991. Dominant containing adaptors in Toll-like receptor signalling. Nat. Rev. Immunol. 7: and recessive mutations define functional domains of Toll, a transmembrane 353–364. protein required for dorsal-ventral polarity in the Drosophila embryo. Dev. 6. Zhou, Y., L. Fang, L. Peng, and W. Qiu. 2017. TLR9 and its signaling pathway in 5: 797–807. 35. Nu´n˜ez Miguel, R., J. Wong, J. F. Westoll, H. J. Brooks, L. A. O’Neill, N. J. Gay, multiple sclerosis. J. Neurol. Sci. 373: 95–99. 7. Cook, D. N., D. S. Pisetsky, and D. A. Schwartz. 2004. Toll-like receptors in the C. E. Bryant, and T. P. Monie. 2007. A dimer of the Toll-like receptor 4 cyto- plasmic domain provides a specific scaffold for the recruitment of signalling pathogenesis of human disease. Nat. Immunol. 5: 975–979. adaptor proteins. PLoS One 2: e788. 8. Duffy, L., and S. C. O’Reilly. 2016. Toll-like receptors in the pathogenesis of 36. Toshchakov, V. Y., and S. N. Vogel. 2007. Cell-penetrating TIR BB loop decoy autoimmune diseases: recent and emerging translational developments. Immu- peptides a novel class of TLR signaling inhibitors and a tool to study topology of noTargets Ther. 5: 69–80. TIR-TIR interactions. Expert Opin. Biol. Ther. 7: 1035–1050. 9. Hatai, H., A. Lepelley, W. Zeng, M. S. Hayden, and S. Ghosh. 2016. Toll-like 37. Piao, W., K. A. Shirey, L. W. Ru, W. Lai, H. Szmacinski, G. A. Snyder, receptor 11 (TLR11) interacts with flagellin and profilin through disparate E. J. Sundberg, J. R. Lakowicz, S. N. Vogel, and V. Y. Toshchakov. 2015. A mechanisms. PLoS One 11: e0148987. decoy peptide that disrupts TIRAP recruitment to TLRs is protective in a murine ´ ´ 10. Hedhli, D., N. Moire, H. Akbar, F. Laurent, B. Heraut, I. Dimier-Poisson, and model of influenza. Cell Rep. 11: 1941–1952. M. N. Me´ve´lec. 2016. The antigen-specific response to Toxoplasma gondii 38. Piao, W., L. W. Ru, and V. Y. Toshchakov. 2016. Differential adapter recruitment profilin, a TLR11/12 ligand, depends on its intrinsic adjuvant properties. Med. by TLR2 co-receptors. Pathog. Dis. 74: ftw043. Available at: https://doi.org/10. Microbiol. Immunol. 205: 345–352. 1093/femspd/ftw043. Accessed: June 13, 2018. 11. Koblansky, A. A., D. Jankovic, H. Oh, S. Hieny, W. Sungnak, R. Mathur, 39. Monie, T. P., M. C. Moncrieffe, and N. J. Gay. 2009. Structure and regulation of M. S. Hayden, S. Akira, A. Sher, and S. Ghosh. 2013. Recognition of profilin by cytoplasmic adapter proteins involved in innate immune signaling. Immunol. Toll-like receptor 12 is critical for host resistance to Toxoplasma gondii. Im- Rev. 227: 161–175. munity 38: 119–130. 40. Barrat, F. J., T. Meeker, J. Gregorio, J. H. Chan, S. Uematsu, S. Akira, B. Chang, 12. Oosting, M., S. C. Cheng, J. M. Bolscher, R. Vestering-Stenger, T. S. Plantinga, O. Duramad, and R. L. Coffman. 2005. Nucleic acids of mammalian origin can I. C. Verschueren, P. Arts, A. Garritsen, H. van Eenennaam, P. Sturm, et al. 2014. act as endogenous ligands for Toll-like receptors and may promote systemic Human TLR10 is an anti-inflammatory pattern-recognition receptor. Proc. Natl. lupus erythematosus. J. Exp. Med. 202: 1131–1139. Acad. Sci. USA 111: E4478–E4484. 41. Roelofs, M. F., L. A. Joosten, S. Abdollahi-Roodsaz, A. W. van Lieshout, 13. Ashkar, A. A., and K. L. Rosenthal. 2002. Toll-like receptor 9, CpG DNA and T. Sprong, F. H. van den Hoogen, W. B. van den Berg, and T. R. Radstake. 2005. innate immunity. Curr. Mol. Med. 2: 545–556. The expression of toll-like receptors 3 and 7 in rheumatoid arthritis synovium is 12 MECHANISMS OF ADAPTER RECRUITMENT TO ACTIVATED TLR9

increased and costimulation of toll-like receptors 3, 4, and 7/8 results in syn- 63. Ke, Y., W. Li, Y. Wang, M. Yang, J. Guo, S. Zhan, X. Du, Z. Wang, M. Yang, ergistic cytokine production by dendritic cells. Arthritis Rheum. 52: 2313–2322. J. Li, et al. 2016. Inhibition of TLR4 signaling by Brucella TIR-containing 42. Curtiss, L. K., and P. S. Tobias. 2009. Emerging role of Toll-like receptors in protein TcpB-derived decoy peptides. Int. J. Med. Microbiol. 306: 391–400. atherosclerosis. J. Lipid Res. 50(Suppl.): S340–S345. 64. Weischenfeldt, J., and B. Porse. 2008. Bone marrow-derived macrophages 43. Ciechomska, M., R. Cant, J. Finnigan, J. M. van Laar, and S. O’Reilly. 2013. (BMM): isolation and applications. CSH Protoc. 2008: pdb.prot5080. Role of toll-like receptors in systemic sclerosis. Expert Rev. Mol. Med. 15: e9. 65. Hirschfeld, M., Y. Ma, J. H. Weis, S. N. Vogel, and J. J. Weis. 2000. Cutting 44. Falck-Hansen, M., C. Kassiteridi, and C. Monaco. 2013. Toll-like receptors in edge: repurification of lipopolysaccharide eliminates signaling through both atherosclerosis. Int. J. Mol. Sci. 14: 14008–14023. human and murine toll-like receptor 2. J. Immunol. 165: 618–622. 45. Celhar, T., and A. M. Fairhurst. 2014. Toll-like receptors in systemic lupus 66. Pace, C. N., F. Vajdos, L. Fee, G. Grimsley, and T. Gray. 1995. How to measure erythematosus: potential for personalized treatment. Front. Pharmacol. 5: 265. and predict the molar absorption coefficient of a protein. Protein Sci. 4: 2411– 46. Thwaites, R., G. Chamberlain, and S. Sacre. 2014. Emerging role of endosomal 2423. toll-like receptors in rheumatoid arthritis. Front. Immunol. 5: 1. 67. Szmacinski, H., V. Toshchakov, and J. R. Lakowicz. 2014. Application of phasor 47. Bhattacharyya, S., and J. Varga. 2015. Emerging roles of innate immune sig- plot and autofluorescence correction for study of heterogeneous cell population. naling and toll-like receptors in fibrosis and systemic sclerosis. Curr. Rheumatol. J. Biomed. Opt. 19: 046017. Rep. 17: 474. 68. Altschul, S. F., T. L. Madden, A. A. Scha¨ffer, J. Zhang, Z. Zhang, W. Miller, and 48. Wu, Y. W., W. Tang, and J. P. Zuo. 2015. Toll-like receptors: potential targets for D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation of lupus treatment. Acta Pharmacol. Sin. 36: 1395–1407. protein database search programs. Nucleic Acids Res. 25: 3389–3402. 49. O’Neill, L. A., C. E. Bryant, and S. L. Doyle. 2009. Therapeutic targeting of 69. Remmert, M., A. Biegert, A. Hauser, and J. So¨ding. 2011. HHblits: lightning-fast Toll-like receptors for infectious and inflammatory diseases and cancer. Phar- iterative protein sequence searching by HMM-HMM alignment. Nat. Methods 9: macol. Rev. 61: 177–197. 50. Connolly, D. J., and L. A. O’Neill. 2012. New developments in Toll-like receptor 173–175. targeted therapeutics. Curr. Opin. Pharmacol. 12: 510–518. 70. Guex, N., and M. C. Peitsch. 1997. SWISS-MODEL and the Swiss-PdbViewer: 51. Derossi, D., A. H. Joliot, G. Chassaing, and A. Prochiantz. 1994. The third helix an environment for comparative protein modeling. Electrophoresis 18: 2714– of the Antennapedia homeodomain translocates through biological membranes. 2723. J. Biol. Chem. 269: 10444–10450. 71. Pettersen, E. F., T. D. Goddard, C. C. Huang, G. S. Couch, D. M. Greenblatt, 52. Nakase, I., Y. Konishi, M. Ueda, H. Saji, and S. Futaki. 2012. Accumulation of E. C. Meng, and T. E. Ferrin. 2004. UCSF Chimera–a visualization system for arginine-rich cell-penetrating peptides in tumors and the potential for anticancer exploratory research and analysis. J. Comput. Chem. 25: 1605–1612. Downloaded from drug delivery in vivo. J. Control. Release 159: 181–188. 72. Zanoni, I., Y. Tan, M. Di Gioia, A. Broggi, J. Ruan, J. Shi, C. A. Donado, 53. Couture, L. A., W. Piao, L. W. Ru, S. N. Vogel, and V. Y. Toshchakov. 2012. F. Shao, H. Wu, J. R. Springstead, and J. C. Kagan. 2016. An endogenous Targeting Toll-like receptor (TLR) signaling by Toll/interleukin-1 receptor (TIR) caspase-11 ligand elicits interleukin-1 release from living dendritic cells. Science domain-containing adapter protein/MyD88 adapter-like (TIRAP/Mal)-derived 352: 1232–1236. decoy peptides. J. Biol. Chem. 287: 24641–24648. 73. Silverstein, R. 2004. D-galactosamine lethality model: scope and limitations. J. 54. Sarko, D., B. Beijer, R. Garcia Boy, E. M. Nothelfer, K. Leotta, M. Eisenhut, Endotoxin Res. 10: 147–162. A. Altmann, U. Haberkorn, and W. Mier. 2010. The pharmacokinetics of cell- 74. Shirey, K. A., W. Lai, A. J. Scott, M. Lipsky, P. Mistry, L. M. Pletneva,

penetrating peptides. Mol. Pharm. 7: 2224–2231. C. L. Karp, J. McAlees, T. L. Gioannini, J. Weiss, et al. 2013. The TLR4 an- http://www.jimmunol.org/ 55. Piao, W., S. N. Vogel, and V. Y. Toshchakov. 2013. Inhibition of TLR4 signaling tagonist Eritoran protects mice from lethal influenza infection. Nature 497: 498– by TRAM-derived decoy peptides in vitro and in vivo. J. Immunol. 190: 2263– 502. 2272. 75. Jiang, Z., P. Georgel, C. Li, J. Choe, K. Crozat, S. Rutschmann, X. Du, T. Bigby, 56. Piao, W., L. W. Ru, K. H. Piepenbrink, E. J. Sundberg, S. N. Vogel, and S. Mudd, S. Sovath, et al. 2006. Details of Toll-like receptor:adapter interaction V. Y. Toshchakov. 2013. Recruitment of TLR adapter TRIF to TLR4 signaling revealed by germ-line mutagenesis. Proc. Natl. Acad. Sci. USA 103: 10961– complex is mediated by the second helical region of TRIF TIR domain. Proc. 10966. Natl. Acad. Sci. USA 110: 19036–19041. 76. Ve, T., S. J. Williams, and B. Kobe. 2015. Structure and function of Toll/ 57. Toshchakov, V. Y., M. J. Fenton, and S. N. Vogel. 2007. Cutting edge: differential interleukin-1 receptor/resistance protein (TIR) domains. Apoptosis 20: 250–261. inhibition of TLR signaling pathways by cell-permeable peptides representing 77. Ve, T., P. R. Vajjhala, A. Hedger, T. Croll, F. DiMaio, S. Horsefield, X. Yu, BB loops of TLRs. J. Immunol. 178: 2655–2660. P. Lavrencic, Z. Hassan, G. P. Morgan, et al. 2017. Structural basis of TIR- 58. Hu, X., Y. Fu, Y. Tian, Z. Zhang, W. Zhang, X. Gao, X. Lu, Y. Cao, and domain-assembly formation in MAL- and MyD88-dependent TLR4 signaling.

N. Zhang. 2016. The anti-inflammatory effect of TR6 on LPS-induced mastitis in by guest on October 1, 2021 Nat. Struct. Mol. Biol. 24: 743–751. mice. Int. Immunopharmacol. 30: 150–156. 78. Horng, T., G. M. Barton, and R. Medzhitov. 2001. TIRAP: an adapter molecule 59. Hu, X., Y. Tian, S. Qu, Y. Cao, S. Li, W. Zhang, Z. Zhang, N. Zhang, and Y. Fu. in the Toll signaling pathway. Nat. Immunol. 2: 835–841. 2017. Protective effect of TM6 on LPS-induced acute lung injury in mice. Sci. 79. Toshchakov, V., B. W. Jones, P. Y. Perera, K. Thomas, M. J. Cody, S. Zhang, Rep. 7: 572. B. R. G. Williams, J. Major, T. A. Hamilton, M. J. Fenton, and S. N. Vogel. 2002. 60. Allette, Y. M., Y. Kim, A. L. Randolph, J. A. Smith, M. S. Ripsch, and F. A. White. 2017. Decoy peptide targeted to Toll-IL-1R domain inhibits LPS TLR4, but not TLR2, mediates IFN-beta-induced STAT1alpha/beta-dependent and TLR4-active metabolite morphine-3 glucuronide sensitization of sensory in macrophages. Nat. Immunol. 3: 392–398. neurons. Sci. Rep. 7: 3741. 80. Horng, T., G. M. Barton, R. A. Flavell, and R. Medzhitov. 2002. The adaptor 61. Lysakova-Devine, T., B. Keogh, B. Harrington, K. Nagpal, A. Halle, molecule TIRAP provides signalling specificity for Toll-like receptors. Nature D. T. Golenbock, T. Monie, and A. G. Bowie. 2010. Viral inhibitory peptide of 420: 329–333. TLR4, a peptide derived from vaccinia protein A46, specifically inhibits TLR4 81. Yamamoto, M., S. Sato, H. Hemmi, H. Sanjo, S. Uematsu, T. Kaisho, by directly targeting MyD88 adaptor-like and TRIF-related adaptor molecule. J. K. Hoshino, O. Takeuchi, M. Kobayashi, T. Fujita, et al. 2002. Essential role for Immunol. 185: 4261–4271. TIRAP in activation of the signalling cascade shared by TLR2 and TLR4. Nature 62. Snyder, G. A., D. Deredge, A. Waldhuber, T. Fresquez, D. Z. Wilkins, 420: 324–329. P. T. Smith, S. Durr, C. Cirl, J. Jiang, W. Jennings, et al. 2014. Crystal structures 82. Latty, S. L., J. Sakai, L. Hopkins, B. Verstak, T. Paramo, N. A. Berglund, of the Toll/Interleukin-1 receptor (TIR) domains from the Brucella protein TcpB E. Cammorota, P. Cicuta, N. J. Gay, P. J. Bond, et al. 2018. Activation of Toll- and host adaptor TIRAP reveal mechanisms of molecular mimicry. J. Biol. like receptors nucleates assembly of the MyDDosome signaling hub. Elife 7: Chem. 289: 669–679. e31377.