Frontline Science: Targeting the TLR7 signalosome assembly

Abstract

TLRs sense a broad range of microbial molecules and initiate antimicrobial immune response. The members of the TLR family use cytoplasmic Toll/interleukin-1R homology (TIR) domain to initiate intracellular signaling. The activated TLRs dimerize their TIRs and recruit adapter proteins to the dimer, through multiple interactions of receptor and adapter TIR domains. Although TLRs play an essential role in innate immunity, the aberrant TLR signaling may cause pathogenic inflammation. This study has screened a library of cell-permeable decoy peptides (CPDPs) derived from the TLR7 TIR for interference with TLR7 signaling and identified new CPDPs that target the TLR7 signalosome assembly. Peptides 7R1, 7R6, 7R9, and 7R11 inhibited the TLR7-induced signaling in murine and human macrophages. The most potent inhibitory peptide of the four, 7R11, significantly reduced the systemic cytokine levels elicited by administration of a TLR7 agonist to mice. TLR7 TIR surface regions that correspond to inhibitory peptides generally corresponded to four TIR sites that mediate signalosome assembly for other TLRs. The cell-based Förster resonance energy transfer/fluorescence lifetime imaging confirmed that 7R9 and 7R11 interact with adapter TIRs. These findings clarify the molecular mechanisms that trigger the adapter recruitment to activated TLR7 and suggest that 7R9 and 7R11 have a significant translational potential as candidate or lead therapeutics for treatment of TLR7-related inflammatory diseases.

1 |. INTRODUCTION

TLRs are a family of pattern recognition receptors (PRRs) that sense conserved microbial molecules and initiate the immune response to a broad range of pathogens.¹ Although TLRs play an important role in innate immunity, excessive TLR signaling often causes uncontrolled inflammation.² The members of TLR family share a common domain architecture, comprising three main regions: the ligand-binding domain, transmembrane helix, and cytoplasmic Toll/Interleukin-1R homology domain (TIR).³ Agonists cause TLRs to dimerize and bring their TIR domains into direct contact.^3,4 Dimerization of TIRs creates a composite binding site, which recruits TIR-containing mediators of downstream signaling.^5–8 Four adapter proteins mediate the formation of primary TLR signalosomes.^5,9 MyD88 and TIRAP participate in the formation of MyDDosome complex that activates nuclear factor𝜅B (NF-𝜅B).^8,10–12 TRIF and TRAM initiate the formation of Triffosome complex, leading to activation of IFN regulatory factors (IRFs),^5,13,14 although TRIF and TRAM do not participate in TLR7 signaling.⁹

Nonetheless TIR domains share fold, their sequence similarity is limited to only 20–30%.^15–17 The typical TIR structure comprises five 𝛽-strands alternating with five 𝛼-helices and includes loop regions that connect strands and helices of the domain.^18,19 Recent studies suggested that four distinct regions of TIR domains mediate protein interactions required for MyDDosome formation.^8,12,20 Blockage of any of these sites prevents TLR signalosome assembly and impedes consequent signaling.^6,8,21–24

TLR7 is activated by single-stranded RNA (ssRNA).²⁵ This receptor also can be activated by a small imidazoquinoline compound, R848.²⁶ TLR7 activation induces production of TNF𝛼, IL-12p40, and other inflammatory cytokines via MyD88-dependent signaling pathway.^26,27 Recent data have suggested that TIRAP facilitates the MyD88-dependent TLR7 signaling,²² although TLR7 can signal in the absence of TIRAP in some cell lineages.^28,29 Multiple reports have suggested that TLR7 is involved in the development of rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), and other autoimmune diseases.^2,30–32

The cell-permeable decoy peptides (CPDPs) is an effective tool to study the mechanisms of intracellular signaling and develop signaling inhibitors.²⁴ CPDPs comprise two functional parts: a peptide vector that renders peptides cell-permeable and a “decoy” sequence responsible for the CPDP binding specificity.^33,34 The decoys are often derived from protein interaction surfaces.^20,24,33 CPDPs act by blocking the protein interactions mediated by the targeted site.^6,33,34 Several TIR-specific CPDP libraries have been tested in the last decade^{6,8,21–23,35–37} A typical TIR-specific library consists of 11–12 peptides. Each peptide corresponds to a short, nonfragmented segment of a TIR sequence, whereas all peptides of a library represent the entire surface of the TIR domain. The CPDPs most often use the segment of Antennapedia homeodomain as the cell-permeable vector.³⁸ The completed screenings of CPDP libraries have identified a number of decoy peptides capable of blocking TIR-TIR interactions and the ensuing TLR signaling.^{6,8,21–23,35–37,39}

This study examines the CPDP library derived from TLR7 TIR (Table 1) and identifies new TLR inhibitors. The most potent inhibitory peptides, 7R9 and 7R11, correspond to 𝛼-helices D and E. The CPDPs from 𝛼-helices B and C, 7R5 and 7R6, and the N-terminal peptide 7R1 also inhibited TLR7, but were less potent. The peptide 7R11 reduced TLR7 signaling in vivo. Binding experiments have demonstrated that 7R9 and 7R11 bind adapter proteins. The peptide 7R11 bound both adapters, TIRAP and MyD88, whereas 7R9 preferentially bound TIRAP.

TABLE 1.

Sequences of TLR7 TIR-derived decoy peptides

Peptide	Sequence	Predominant structural region
7R1	GYQRLISPDSCYD	N-terminal segment
7R2	DTKDPAVTEWVLAE	AA loop, 𝛼-helix Aa
7R3	VAKLEDPREKHFNL	𝛼-helix A, AB loop, 𝛽-strand B
7R34	EDPREKHFNLCLEER	AB loop, 𝛽-strand B, N-terminal residues of BB loop
7R4	LEERDWLPGQPVLE	𝛽-strand B, BB loop, 𝛼-helix B
7R5	NLSQSIQLSKKT	𝛼-helix B, BC loop, 𝛽-strand C
7R6	DKYAKTENFKIAFY	𝛼-helix C, CD loop
7R7	LSHQRLMDEKVDV	CD loop
7R8	EKPFQKSKF	DD loop
7R9	FLQLRKRLCGSS	DD loop, 𝛼-helix D, DE loop
7R10	VLEWPTNPQAH	DE loop, 𝛽-strand E, EE loop, 𝛼-helix E
7R11	PYFWQCLKNALATD	𝛼-helix E
7R11-ΔN	QCLKNALATD	𝛼-helix E
7R11-ΔC	PYFWQCLKNA	𝛼-helix E
7R11-C/S	PYFWQSLKNALATD	𝛼-helix E
7R12	NHVAYSQVFKETV	C-terminal segment

^aSecondary structure elements of the TIR domain are indicated by capital letters, with letter “A” indicating the most N-terminal element.

2 |. MATERIALS AND METHODS

2.1 |. Animals and cells

C57BL/6J mice were obtained from the Jackson Laboratory (Bar Harbor, ME, USA). Primary peritoneal macrophages were obtained by peritoneal lavage 4 d after i.p. injection of 3 ml sterile 3% thioglycolate (Remel; San Diego, CA, USA). The peritoneal macrophages were cultured in RPMI 1640 medium (Quality Biological; Gaithersburg, MD, USA) containing 2% FBS (Sigma; St. Louis, MO, USA), 1% penicillin/ streptomycin (Sigma), and 2 mM L-glutamine (Sigma). THP-1 macrophages differentiated by incubation with 100 nM phorbol myristate acetate for 3 d were cultured in RPMI 1640 medium containing 10% FBS.

2.2 |. TLR agonists

R848, ssRNA40/LyoVec, ssRNApolyU/LyoVec, ODN1668, S-(2,3-bis(palmitoyloxy)-(2R,2S)-propyl)-(R)-Cys-Ser-Lys4-OH (Pam2C), and S-(2,3-bis (palmitoyloxy)-(2R, 2S)-propyl)-N-palmitoyl-(R)-Cys-SerLys4-OH (Pam3C) were purchased from InvivoGen (San Diego, CA). Phenol purified, Escherichia coli K235 LPS⁴⁰ was a kind gift of Dr. Stefanie N. Vogel (UMB SOM, Baltimore, MD, USA).

2.3 |. Peptide design, synthesis, and reconstitution

The TLR7 TIR decoy peptides were purchased from Aapptec (Louisville, KY, USA) or GenScript (Piscataway, NJ, USA). All CPDPs included the N-terminal Antennapedia homeodomain sequence RQIKIWFQNRRMKWKK.³⁸ The Cy3-labeled peptides were produced by CPC Scientific (Sunnyvale, CA, USA). The Cy3 label was placed at the peptide N-terminus. The purity of all CPDPs was ≥95%. The lyophilized peptides were reconstituted in UltraPure water from Millipore (Burlington, MA, USA) to the concentrations of 2 mM. Concentrations of reconstituted peptides were determined spectrophotometrically.⁴¹ Then, the peptide solutions were either directly added to cell medium to final concentrations of 10, 20, or 40 μM for the in vitro experiments, or diluted with PBS to the final concentration of 0.4 mM (200 nmol in 0.5 mL) for administration to mice.

2.4 |. Evaluation of cytokine expression by quantitative real-time PCR

Mouse peritoneal macrophages, bone marrow-derived macrophages (BMDMs), or differentiated human THP-1 macrophages (2 × 10⁶ per well) were plated in 12-well plates and incubated 24 h at 37°C. cDNA was synthesized from 1 _μg of RNA isolated with Trizol (Life Technologies; Carlsbad, CA, USA) and reverse transcribed with RevertAid RT Kit (ThermoFisher Scientific; Waltham, MA, USA). The cDNA obtained was amplified with gene-specific primers for mouse HPRT, TNF𝛼, IL-1𝛽, or IL-12p40 and SYBR Green PCR mix (Applied Biosystems; Foster City, CA, USA) as described earlier.³⁵ Cytokine mRNA expression was measured 1 h after challenging the cells with a TLR agonist and normalized to the expression of hypoxanthine phosphoribosyltransferase (HPRT).

2.5 |. ELISA evaluation of cytokine secretion

Murine peritoneal macrophages (1 × 10⁶ per well) were plated in 24-well plates and incubated 24 h at 37°C. TNF𝛼, IL-12p40, and IFN𝛾 concentrations in cell supernatants were evaluated 24 h after cell stimulation with a TLR agonist using the ELISA kits from Biolegend (San Diego, CA, USA).

2.6 |. Expression vectors

MyD88-Cerulean (MyD88-Cer), TIRAP-Cer, TLR2-Cer, and Cer expression vectors were described previously.^22,36

2.7 |. Fluorescence lifetime imaging (FLIM)

A total of 1 × 10⁶ HeLa cells were transfected with MyD88-Cer (0.5 μg), TIRAP-Cer (1 μg), TLR2-Cer (2 μg), or Cer not fused to a TIR domain (2 _μg) expression vectors using Lipofectamine 3000 (Invitrogen; Carlsbad, CA, USA). The next day, the transfected cells were trypsinized and reseeded into the 50-well gasket (Grace BioLabs; Bend, OR) mounted on a microscope slide (8000 cells/well). Twenty-four hours later, cells were treated with Cy3–7R9, Cy3–7R11, or Cy3–2R9 for 1 h and fixed on the slides using 4% paraformaldehyde. Fluorescence lifetime images were acquired using the Alba V FLIM system (ISS; Champagne, IL, USA). FLIM data were analyzed using the VistaVision Suite software (Vista v.218 from ISS). Measurements were performed using the time domain (TD) modality of the FLIM system, as described earlier.^8,42 Förster resonance energy transfer (FRET) efficiency (E) was calculated from the average amplitude-weighted lifetimes determined from images that contained 10–40 fluorescent cells. To calculate (E), average lifetimes of the donor-only images (𝜏D) and donor-acceptor images (𝜏DA) were used: E = 1 − (𝜏DA/𝜏D).⁴²

2.8 |. Animal experiments

The mice were i.p. injected with 45 nmol of R848 diluted in 0.5 mL of PBS.²² CPDPs were administered as 0.5 ml PBS aliquots i.p. at the dose of 200 nmol per mouse 1 h before administration of the TLR7 agonist. The plasma samples were obtained 1, 3, and 5 h after mice treatment with R848. Data were collected in 3 independent experiments. ELISA kits from Biolegend were used to measure systemic cytokine levels.

2.9 |. TLR7 TIR domain modeling

Template search was performed using BLAST (basic local alignment search tool), HMM-HMM-based lightning-fast iterative sequence search (HHBlits) and the SWISS-MODEL template library (SMTL).^43,44 The target sequence was searched with BLAST against the primary amino acid sequence contained in SMTL.⁴³ An initial HHblits profile has been built using the procedure outlined in Remmert et al.,⁴⁴ followed by 1 iteration of HHblits against NR20 database. The model was built based on the target-template alignment using ProMod3.⁴⁵ UCSF Chimera viewer was used for image production.⁴⁶

2.10 |. Statistical analysis

Results were statistically analyzed using ANOVA and GraphPad Prism 5 software. Numeric data in text and on graphs are shown as means ± SEM for at least 3 independent experiments. The statistical significance of changes in cytokine mRNA or cytokine concentration in cell supernatants was determined by 1-way ANOVA test; *P < 0.001. The statistical significance of changes in cytokine levels in blood plasma samples was determined by a 2-way ANOVA test; *P < 0.001.

3 |. RESULTS

3.1 |. Identification of TLR7 inhibitors

The library of TLR7-derived CPDPs (Table 1) was screened first for the ability of individual peptides to blunt the R848-induced cytokine expression in murine macrophages. Peptide 2R9 was added to the library as a positive control for TLR7 inhibition.²² Peptides 7R1, 7R5, 7R6, 7R8, 7R9, 7R11, and 2R9 at 40 μM statistically significantly decreased the expression of all cytokines measured (TNF𝛼, IL-1𝛽, and IL-12p40) (Fig. 1A). Inhibitory peptides 7R1, 7R6, 7R9, and 7R11 were further tested for the ability to suppress the secretion of inflammatory cytokines. All selected peptides and 2R9 inhibited the production of TNF𝛼, IL-12p40, and IFN𝛾, when used at 40 μM (Fig. 1B). The most effective TLR7 inhibitor, 7R11, reduced both expression and secretion of inflammatory cytokines at 20 μM (Fig. 1A, B). R848 does not activate murine TLR8.⁴⁷ R848, however, is a potent agonist for human TLR8.⁴⁷ TLR7 inhibitory peptides inhibited the R848-induced inflammatory response in human macrophages similar to their effect in mouse macrophages (Fig. 1C). All TLR7 inhibitory peptides suppressed the expression of TNF𝛼 and IL-12p40 in differentiated THP-1 macrophages at a concentration of 40 μM; 7R11 was also effective at a lower concentration (20 μM) (Fig. 1C). TLR7 inhibitory peptides, with exception to 7R9, did not affect cell viability evaluated by MTT test (Fig. S2). Prolonged cell incubation with 7R9 (5 h at 40 μM) affected cells to a greater extent, yet cell viability was ∼80% at the end of the testing (Fig. S2).

FIGURE 1. Effects of TLR7-derived cell-permeable decoy peptides on R848-induced cytokine activation.

Mouse peritoneal macrophages or differentiated human THP-1 cells were incubated in the presence of 20 or 40 μM CPDPs (Table 1) for 30 min prior to stimulation with 0.285 μM (peritoneal macrophages) or 28.5 μM (THP-1) of R848. (A) mRNA expression in mouse macrophages. (B) Protein concentrations in mouse macrophage supernatants. (C) mRNA expression in human THP-1 cells. (D) The structural model of TLR7 TIR domain demonstrates segments used as decoy peptides

3.2 |. Specificity of TLR inhibition by TLR7-derived CPDPs

Inhibitory peptides 7R1, 7R6, 7R9, and 7R11 were examined for the ability to suppress the expression of TNF𝛼 induced by TLR2, TLR4, and TLR9 agonists. Mouse macrophage treatment with 40 _μM of any selected peptide significantly reduced TNF𝛼 expression in cells stimulated with TLR2/1, TLR4, or TLR9 agonist (Fig. 2). Inhibitory peptides 7R9 and 7R11 (40 μM) also inhibited the TLR2/6 agonist-induced TNF𝛼 expression, whereas 7R1 and 7R6 did not (Fig. 2). The peptide 7R11, apparently the most potent inhibitor of the tested peptides, was effective against both TLR2 and TLR9 signaling when used at 20 μM. These results suggest multispecific action of 7R1, 7R6, 7R9, and 7R11, which might be caused by peptide binding to adapter proteins.^8,22,48

FIGURE 2. The specificity of TLR inhibition by TLR7-derived cell-permeable decoy peptides.

mRNA expression in mouse bone marrow-derived macrophages (BMDMs; TLR9) or peritoneal macrophages (TLR2, TLR4, and TLR7) treated with 20 or 40 μM of CPDPs (Table 1) for 30 min prior to stimulation with E.coli LPS (0.1 μg/mL), ODN1668 (1 μM), Pam2C (39 nM), Pam3C (330 nM), or ssRNAs (5 μg/mL)

The synthetic compound R848 was used as a TLR7 agonist for TLR7 CPDP library screening (Fig. 1).²⁶ We also tested the ability of 7R1, 7R6, 7R9, and 7R11 to inhibit TLR7 signaling induced by two ssRNA agonists. All four peptides inhibited ssRNA40 signaling at 20 and 40 μM (Fig. 2 and Fig. S1). The ssRNApolyU-induced TLR7 signaling was also inhibited by all peptide tested at the concentration of 40 μM (Fig. 2 and Fig. S1). The 7R9 and 7R11 also suppressed the ssRNApolyU-induced cytokine response when used at 20 μM (Fig. 2 and Fig. S1). These data confirmed the ability of 7R1, 7R6, 7R9, and 7R11 peptides to inhibit TLR7 signaling without reference to the particular agonist.

3.3 |. Inhibitory potency of modified 7R11 peptides

We tested three modifications of the most potent TLR7 inhibitory peptide, 7R11. Two modified 7R11 peptides, 7R11-ΔN and 7R11-ΔC, were produced by four amino acid deletions at either N or C terminus (Table 1). The third modified peptide, 7R11-C/S, was a single amino acid replacement variant, in which cysteine was substituted for serine (Table 1). The modified peptides were evaluated based on their effect on R848-induced cytokine expression (Fig. 3). 7R11-ΔC inhibited TLR7 comparably to the parent peptide (Fig. 3). 7R11-ΔN and 7R11-C/S were less effective (Fig. 3). These data suggest that the N-terminal part and the central cysteine of 7R11 are more important for inhibitory activity.

FIGURE 3. Inhibition of TLR7 signaling by 7R11 and its modifications.

mRNA expression in mouse macrophages, treated with 10, 20, or 40 μM of CPDPs (Table 1) for 30 min prior to stimulation with R848 (0.285 μM)

3.4 |. 7R9 and 7R11 bind adapter TIR domains

We used a cell-based FRET-FLIM assay to study the binding of 7R9 and 7R11 to the TIR domains of adapter proteins.^6,8,22,42 We were unable to test the binding of peptides to TLR7 TIR because of extremely low ectopic expression of TLR7.²² The HeLa cells transfected with a TIR-Cer fusion expression vector were incubated in the presence of 2.5, 10, or 40 μM of a Cy3-labeled 7R9 or 7R11 for 1 h. The Cy3-peptides quench Cer fluorescence and decrease Cer fluorescence lifetime only if the Cy-3-labeled CPDP directly binds a Cer-fused TIR.⁴² We previously reported a strong binding of the TLR2-derived peptide 2R9 to TIRAP TIR.^8,22 Therefore, cells expressing TIRAP TIR and treated with Cy3-labeled 2R9 (Cy3–2R9) were used as a positive control for FRET. Cer not fused to a TIR was used as a negative FRET control. Figure 4A demonstrates FLIM images of cells expressing the MyD88-Cer, TIRAP-Cer, and Cer and treated with different concentrations of Cy3-peptides. The images show the average fluorescence lifetime (ns) of Cer. Figure 4B shows the lifetime distribution of the Cer-component in images shown in Figure 4A. Figure 4C demonstrates FRET efficiency for different Cy3-labeled CPDP/Cer-labeled TIR pairs, calculated from the average lifetimes of multiple images.

FIGURE 4. Effect of Cy3-labeled CPDPs on fluorescence lifetime of Cer-component.

(A) FLIM images of HeLa cells expressing Cerulean not fused to a TIR (upper row), Cer fused with the MyD88 (middle row), or TIRAP (bottom row) and treated with Cy3-labeled CPDPs for 1 h. (B) The histograms show frequencies of pixels with average lifetimes in images of panel A. (C) FRET efficiency for different TIR-peptide pairs. Data represent means ±SEM of 2 independent experiments; NF—no FRET

The strongest quenching was observed for the Cy3–7R11 and TIRAP-Cer pair. The Cy3–7R11 peptide decreased Cer fluorescence lifetime in the concentration-dependent manner, with FRET varying from 15% (2.5 μM of Cy3–7R11) to 47% (40 μM; Fig. 4C). Cy3–7R11 also quenched MyD88-Cer fluorescence (Fig. 4A). The FRET efficiency for Cy3–7R11 and MyD88-Cer varied in the range of 17% (2.5 μM of Cy3–7R11) to 36% (40 μM) (Fig. 4C). The quenching of both TIRAP- and MyD88-Cer by Cy3–7R11 suggest peptide binding to both adapter proteins.

Cy3–7R9 dose-dependently decreased the fluorescence lifetime of TIRAP-Cer. However, Cy3–7R9 failed to quench MyD88-Cer. Thus, FRET observed for the Cy3–7R9 and TIRAP-Cer pair varied from 8% (2.5 μM of Cy3–7R9) to 41% (40 μM) (Fig. 4C), whereas no FRET was detected for Cy3–7R9 and MyD88-Cer pair at all three peptide concentrations. Obtained results indicate the selective 7R9 binding to TIRAP, but not to MyD88 TIR.

The FRET in Cy3–7R9-TIRAP and Cy3–7R11–TIRAP pairs was slightly lower than that for the control Cy3–2R9–TIRAP pair (18–56%) (Fig. 4C). Cy3–7R9 and Cy3–7R11 apparently did not interact with TLR2 TIR at lower concentrations of 2.5 and 10 μM (Fig. 4C), but the 40 μM dose of either peptide, however, quenched the fluorescence of TLR2-Cer fusion protein with FRET efficiency of ∼16% for Cy3–7R9 and ∼25% for Cy3–7R11 (Fig. 4C). These FRET values are substantially lower than that in the positive control system (Cy3–2R9 - TIRAP) (Fig. 4C), indicating the binding affinities for peptides to TIRs vary. Both Cy3–7R9 and Cy3–7R11 did not quench fluorescence of Cer not fused to a TIR (Fig. 4C).

3.5 |. 7R11 inhibits R848-induced cytokine response in vivo

Intraperitoneal administration of R848 to mice induces a transient increase in circulating TNF𝛼 and IL-12p40 (Fig. 5). Kinetics of activation is different for cytokines measured. The plasma concentration of TNF𝛼 reached the maximum of ∼0.6 ng/mL one hour after R848 treatment and then decreased rapidly to nearly basal levels in 3 h samples (Fig. 5A). IL-12p40 concentrations peaked at ∼60 ng/mL three hours after R848 administration, and then decreased to ∼20 ng/mL in 5 h samples (Fig. 5B). Pretreatment with 7R11 reduced the plasma TNF𝛼 concentrations by ∼4 times (Fig. 5A), whereas IL-12p40 level was decreased by ∼6 times (Fig. 5B). The control peptide, 7R2, which did not inhibit the R848-induced signaling in cell culture experiments (Fig. 1A), did not affect the systemic TNF𝛼 and IL-12p40 levels as significantly as 7R11 (Fig. 5A, B).

FIGURE 5. 7R11 inhibits TLR7-induced systemic cytokines in vivo.

Mice were i.p. injected with 45 nmol of R848²²; CPDPs were administered as PBS aliquots i.p. at the dose of 200 nmol per mouse 1 h before administration of the TLR7 agonist. Plasma (A) TNF𝛼 and (B) IL-12p40 levels in mice following administration of R848. Data were collected in 3 independent experiments

4 |. DISCUSSION

Screening of TLR7 TIR peptide library has identified several TLR inhibitors. New inhibitory peptides derive from regions that correspond to four TIR-TIR interaction sites identified previously based on the screenings of other TIR peptide libraries (Fig. 6A, B).⁸ The first site, site 1 (S1), is located near 𝛽-strand B (𝛽B) and, in addition to 𝛽B, includes AB and BB loops and the segment N-terminal to 𝛼A.^8,20 S1 mediates the first step of the formation of an intracellular signaling complex, that is, dimerization of TLR TIRs through the asymmetric interaction with S4 of the second TIR of the dimer (Fig. 6C). S4 is located on the opposite to S1 surface of the TIR, near 𝛽-strand E, the second edge-forming strand of the 𝛽-sheet.^8,20 S4 is formed by 𝛼-helix E, 𝛽-strand E, and the EE loop. TLR S1 and S4, which are not involved in the dimerization of TLR TIRs, mediate the second important function in the TLR signalosome assembly, as each site may also interact with an adapter TIR through a lateral interaction, leading to formation of extended oligomers, called “protofilaments” (Fig. 6B and D).^8,12,20 S1 in the TLR7 library is represented by inhibitory peptides 7R1 and 7R3, whereas peptide 7R11 represents S4 (Figs. 1A and 6A).^8,20 The other two interaction sites of TIR domains, S2 and S3 form a composite binding site in the dimerization-dependent manner; this site is responsible for the initiation of adapter recruitment (Fig. 6C). Reciprocal S2-S3 and S3-S2 receptor-adapter and adapter-adapter interactions stabilize the receptor dimer and allow for recruitment of additional adapter molecules, eventually leading to nucleation of Myddosome assembly (Fig. 6B, D, and E).^8,12,20 In the TLR7 library, S2 is represented by peptides 7R5 and 7R6; S3 corresponds to 7R9.

FIGURE 6. Mechanisms of CPDP interference with TLR7 signalosome assembly.

(A) Four TIR sites that mediate recruitment of TLR adapters to activated TLRs. Peptides 7R1 and 7R3 correspond to site 1 (S1, highlighted in red), whereas S2 (highlighted in green) is represented by 7R5 and 7R6 CPDPs. S3 (highlighted in blue), and S4 (highlighted in purple) correspond to peptides 7R9, and 7R11, respectively. The image presents the TLR7 TIR model, which was built using the TLR2 TIR coordinate file (1o77) as the template. (B) TIR interactions that trigger cytoplasmic TLR signaling. S1 and S4 of TLR TIRs mutually interact to form a TLR TIR dimer (represented by two upper TIR domains of the panel). S1 and S4 sites, which are not involved in receptor TIR dimerization may be involved in the recruitment of adapter TIR domains, leading to elongation of the complex. S2 and S3 mediate the second type of TIR-TIR interactions within TLR7 signalosome. Reciprocal S2-S3 and S3-S2 interactions stabilize the initial complex and lead to the formation of the second strand of the filament. The assembly of TLR7 signalosome. (C) TLR7 dimerizes through the interaction between S1 and S4. Receptor dimerization creates a bipartite site for the recruitment of the adapter TIR (TIRAP or MyD88) through TLR7 (1) S2-adapter S3 and TLR7 (2) S3-adapter S2 interactions. (D), the recruitment of TIRAP creates two extra binding sites for additional adapter protein (1), whereas its recruitment makes possible further addition of new adapter (2). (E), the elongation of initial TLR7 signaling complex through simultaneous intrastrand and interstrand interactions

Binding of two most potent inhibitory peptides of the TLR7 library to the adapter TIR domains was studied using the FRET assay coupled with FLIM. Both peptides bound the TIR domains of TLR adapters. Peptide 7R9 (corresponds to S3) demonstrated strong binding to the TIRAP TIR domain; however, to our surprise, 7R9 did not bind the MyD88 TIR. Notably, peptides from the structurally homologous region of other TLRs demonstrated a similar binding specificity. Thus, peptides 2R9, 4R9, and 6R9 (these peptides correspond to 𝛼-helices D of TLR2, TLR4, and TLR6 TIR, respectively), similar to 7R9, preferentially bound TIRAP.^6,22,48 Notably, the sequence similarity of these TIRAP-binding peptides is quite limited. These peptides have following sequences: (2R9) Antp-PQRFCK LRK IMNT; (4R9) Antp-LRQQVE LY R LLSR; (6R9) - Antp-PSRYHK LR ALMAQ; and (7R9) Antp-FLQ LRKR LCGSS. One common feature of these peptides is the central leucine followed by a segment that contains one or several positively charged amino acids (this segment, underlined in the sequences of the previous sentence, correspond to the central round of 𝛼D). Interestingly, each inhibitory peptide from S2 (S2 is the interaction counterpart for S3) identified to date contains either a polar or negatively charged amino acid,²⁰ suggesting the importance of electrostatic interactions in the S2-S3 interactions in the TLR signalosome formation. Binding of 7R9 to TIRAP confirms the recently proposed notion that TIRAP plays a broader role in TLR signaling, than previously recognized. In addition to the previously recognized, obligatory role in the TLR2 and MyD88-dependent TLR4 signaling,¹ TIRAP was found to facilitate the signaling from endosomal TLRs, TLR7, and TLR9.^22,49 The proposition that TIRAP enhances TLR7 signaling was first made by Piao et al. based on the observation that the TIRAP-targeting peptide 2R9 inhibits TLR7 signaling.²² A new finding that 7R9 binds TIRAP suggests that TLR7 and TLR2 interact with TIRAP in a structurally similar manner, shown in Figure 6E.

In addition to TLR7, 7R9, and 7R11 inhibited TLR2, TLR4, and TLR9 (Fig. 2). This finding is consistent with the proposition that 7R9 and 7R11 block TLRs by sequestering the adapter proteins, that is, TIRAP in the case of 7R9 or both TIRAP and MyD88 in the case of 7R11. Notably, not only binding (Fig. 4), but also inhibitory specificity demonstrated by 7R9 (Fig. 2) is similar to that demonstrated by 2R9, a TIRAP-targeting peptide from the TLR2 TIR library.²²

7R11 interacted with TIR domains of both adapters, TIRAP and MyD88 (Fig. 4), suggesting that TLR7 signalosome can elongate in the S4 direction through either TIRAP or MyD88 (Fig. 6E). Peptides derived from S1 of TLR7, 7R1, and 7R3, weakly inhibited TLR7. This observation likely indicates that the actual S1 binding site is composite, so that neither peptide alone sufficiently represents the recognition surface. Available data, however, do not exclude the possibility that TLR7 signalosome can elongate bidirectionally, in both S1 and S4 direction (Fig. 6E), similar to elongation of TLR4 and TLR9 signalosomes.^8,20 S2, similar to S1, produced two inhibitory peptides, each of which demonstrated a weaker inhibitory activity (Fig. 1A). The S2 peptides, 7R5 and 7R6, represent juxtaposed regions of TLR7 TIR surface. Because S3 peptide, 7R9, bound TIRAP, not MyD88, the current models of TLR signalosome assembly^8,12,20 imply that the composite S2 site should be predominantly responsible for the recruitment of MyD88 in lieu of TIRAP to the TLR7 TIR dimer, and for the ability of TLR7 to signal in the TIRAP absence.^28,29

In conclusion, the screening of decoy peptides derived from TLR7 TIR has identified new inhibitory peptides, one of which is sufficiently potent to block the systemic TLR7 activation in mice. Obtained results generally confirm the recent model of primary TLR signalosome that implies that the assembly is through the formation of filamentous, open-ended, double-stranded oligomers that can elongate bidirectionally.^8,12,20 New data highlight the importance of electrostatic interactions, particularly positive charges present in the 𝛼-helix D of TLRs for initiation of intracellular TLR signaling.

The therapy based on TLR7 targeting could be beneficial in the treatment of autoimmune diseases. However, despite multiple circumstantial evidence for TLR7 involvement of pathogenesis of autoimmune diseases, none suggests that TLR7 can be regarded as the only driver of the disease. In addition, no direct model for evaluation of TLR7 contribution in pathogenesis is available at the present time. Therefore, the evaluation of TLR7 inhibitory peptides in advanced animal model of human diseases is not included in this study and should be a matter of future research.

Abbreviations:

BMDM

bone marrow-derived macrophage

Cer

Cerulean

CPDP

cell-permeable decoy peptide

FLIM

fluorescence lifetime imaging

FRET

Förster resonance energy transfer

HHBlits

HMM-HMM-based lightning-fast iterative sequence search

Pam2C

S-(2,3-bis(palmitoyloxy)-(2R,2S)-propyl)-(R)-Cys-Ser-Lys4-OH

Pam3C

S-(2,3-bis(palmitoyloxy)-(2R,2S)-propyl)-N-palmitoyl-(R)-Cys-Ser-Lys4-OH

PRR

pattern recognition receptor

RA

rheumatoid arthritis

S1

Site 1

S2

Site 2

S3

Site 3

S4

Site 4

SLE

systemic lupus erythematosus

ssRNA

single-stranded RNA

TIR

Toll/interleukin-1R homology (domain)