Toll-Like Receptors and Signaling

A Toll-like receptor is a single-pass transmembrane glycoprotein built from three parts: an extracellular (or endosomal-luminal) domain made of curved leucine-rich repeats (LRRs) that physically binds the ligand, a single transmembrane helix, and a cytoplasmic Toll/interleukin-1 receptor (TIR) domain that does the signaling. Humans express ten functional TLRs (TLR1–TLR10); mice have twelve. Each TLR, or TLR heterodimer, recognizes a structurally conserved class of molecule — a pathogen-associated molecular pattern, or PAMP — rather than a specific microbial species. This is the crucial conceptual link back to Pattern Recognition and PAMPs: TLRs are germline-encoded, do not rearrange like antibodies or T cell receptors, and the entire human population inherits essentially the same ten receptors, tuned by evolution to a finite list of molecular shapes that are indispensable to microbial physiology and thus hard for pathogens to discard. The specificity list is short and worth memorizing precisely, because exam and clinical reasoning both hinge on matching ligand to receptor to location.

Two organizing principles fall out of this table. First, TLRs that detect lipid- and protein-based bacterial surface components (TLR1/2, 2/6, 4, 5) sit on the plasma membrane, positioned to sample the extracellular milieu where intact bacteria and their shed components are found. Second, TLRs that detect nucleic acids (TLR3, 7, 8, 9) are sequestered inside endosomes, facing the endosomal lumen rather than the cytosol. This compartmentalization is not incidental — it is the primary mechanism by which the innate immune system avoids constantly sensing self DNA and RNA, which are normally confined to the nucleus and mitochondria and never enter the endosomal pathway in undamaged cells. Only when nucleic acid arrives inside an endosome, typically because it was ingested along with a virus or a phagocytosed bacterium, does it become accessible to TLR7/8/9. TLR4 elegantly straddles both worlds: it first recognizes LPS at the cell surface, then is trafficked into endosomes where it switches signaling output — a mechanism described below.

Step 1 — Ligand capture and dimerization. TLR ectodomains are horseshoe-shaped solenoids built from 19–25 leucine-rich repeats. Ligand binding is often assisted by accessory proteins: LPS recognition by TLR4 requires LPS-binding protein (LBP) to extract LPS monomers from bacterial membranes or micelles, hand them to CD14 (a GPI-anchored surface protein), which in turn delivers LPS to MD-2, a small secreted protein that sits in a pocket on TLR4's ectodomain. Only the LPS·MD-2·TLR4 complex, formed as a 2:2:2 dimer of dimers, is competent to signal. This multi-component relay explains why LPS is such a potent immunostimulant even at picomolar concentrations — CD14 concentrates ligand from a large extracellular volume onto a small number of receptors.

Step 2 — TIR domain juxtaposition. Ligand-induced dimerization of the ectodomains forces the two transmembrane helices, and therefore the two cytoplasmic TIR domains, into a specific side-by-side geometry. This is the actual signal: an otherwise "resting" cytoplasmic domain is now presented as a paired, structurally defined surface that adaptor proteins can dock onto through homotypic TIR-TIR interactions.

Step 3 — Adaptor recruitment: the MyD88 branch versus the TRIF branch. Four TIR-domain-containing adaptors matter: MyD88, TIRAP (also called Mal), TRIF, and TRAM.

• Every TLR except TLR3 recruits MyD88 (myeloid differentiation primary response 88), either directly (TLR5, TLR7/8, TLR9) or via the bridging adaptor TIRAP (TLR1/2, 2/6, and cell-surface TLR4).
• TLR3, and TLR4 after it is endocytosed, instead recruit TRIF (TIR-domain-containing adaptor inducing IFN-β), with TRAM serving as the bridging adaptor specifically for endosomal TLR4.

This branching is the single most important piece of wiring in the whole pathway, because the two adaptors drive different transcription factors and therefore different biological outputs.

Step 4a — The MyD88 pathway to NF-κB and AP-1. MyD88 has a death domain that recruits IRAK4 (IL-1 receptor-associated kinase 4), which phosphorylates and activates IRAK1 and IRAK2. Activated IRAK1/2 recruits and activates TRAF6, an E3 ubiquitin ligase. TRAF6, together with the ubiquitin-conjugating enzyme complex Ubc13/Uev1A, synthesizes K63-linked (non-degradative) polyubiquitin chains on itself and on target proteins including the kinase TAK1. TAK1 activation triggers two downstream cascades in parallel: 1. TAK1 activates the IKK complex (IKKα/IKKβ/NEMO), which phosphorylates IκB, marking it for K48-linked polyubiquitination and proteasomal degradation. This liberates NF-κB (p50/p65), which translocates into the nucleus and activates genes for TNF-α, IL-1β, IL-6, IL-8, and other pro-inflammatory cytokines. 2. TAK1 also activates MAPK cascades (p38, JNK, ERK), which activate the transcription factor AP-1 (Fos/Jun), reinforcing cytokine gene induction and driving additional genes involved in cell survival and further inflammatory signaling. In plasmacytoid dendritic cells specifically, endosomal TLR7 and TLR9 signaling through MyD88 can also recruit IRAK1 and IRF7 directly, driving massive type I interferon production — this is why plasmacytoid dendritic cells are the body's dominant interferon-α producing cell during viral infection, a theme picked up under Type I Interferons and Antiviral Defense.

Step 4b — The TRIF pathway to IRF3 and delayed NF-κB. TRIF recruits TBK1 and IKKε, two related kinases that phosphorylate IRF3 (interferon regulatory factor 3), causing it to dimerize and translocate to the nucleus, where it drives transcription of IFN-β and IFN-stimulated genes — the frontline antiviral transcriptional program. TRIF also recruits TRAF6 and RIP1 to activate a second, delayed wave of NF-κB, which is why TLR4 signaling produces a biphasic cytokine response: an early MyD88-dependent burst of TNF-α from the plasma membrane, followed by a later TRIF-dependent wave of type I interferon and additional NF-κB activity once the receptor has been internalized into endosomes.

Step 5 — Endosomal trafficking as a signaling switch. TLR4 is unusual in signaling from two locations sequentially. At the plasma membrane, LPS-bound TLR4/MD-2 recruits TIRAP/MyD88 and drives the early NF-κB/inflammatory response. The complex is then endocytosed in a CD14-dependent, dynamin-dependent process; once in the endosome, it recruits TRAM/TRIF and drives the IRF3/interferon response. This two-phase, two-location signaling from one receptor is a clean example of how subcellular geography, not just receptor identity, dictates transcriptional output — the same logic that keeps TLR7/8/9 from firing on self nucleic acids applies here to sculpt the kinetics of the LPS response.

Drosophila Toll and antifungal immunity — Lemaitre, Nüsslein-Volhard, and Hoffmann (1996). The Toll gene was originally discovered by Christiane Nüsslein-Volhard's group for its role in establishing dorsal-ventral polarity in the fly embryo — the gene's name (German for "great" or "weird") reflects her exclamation on seeing the mutant phenotype. Bruno Lemaitre, working in Jules Hoffmann's laboratory, later asked whether this same receptor had a second job in adult immune defense. His team infected wild-type flies and flies carrying loss-of-function mutations in Toll or in downstream pathway components (spätzle, cactus, the fly IκB homolog) with the fungus Aspergillus fumigatus. Wild-type flies mounted an antifungal response, upregulating the antifungal peptide gene drosomycin, and survived. Flies lacking functional Toll signaling failed to induce drosomycin and died rapidly of fungal overgrowth, essentially defenseless. This was the first direct demonstration that the Toll pathway, previously known only as a developmental patterning gene, is required for innate antimicrobial defense in an adult animal — establishing the conceptual template (a germline-encoded receptor triggering an inducible antimicrobial transcriptional program) that mammalian TLR biology would later be shown to follow almost exactly.

Mammalian TLR homologs and LPS unresponsiveness — Medzhitov, Janeway, and colleagues (1997); Poltorak, Beutler, and colleagues (1998). Ruslan Medzhitov and Charles Janeway Jr. cloned a human Toll homolog (later named TLR4) and showed that a constitutively active, engineered version of it could induce NF-κB activation and expression of costimulatory molecules and inflammatory cytokines in human cell lines — the first evidence that a mammalian Toll-family receptor could drive innate immune gene induction, published in 1997. The decisive genetic proof of TLR4's specific ligand came the following year from Bruce Beutler's laboratory. Two mouse strains, C3H/HeJ and C57BL/10ScCr, had been known for decades to be uniquely resistant to the lethal effects of LPS (endotoxin) — they simply failed to mount a cytokine storm or go into shock when injected with doses of LPS that killed normal mice. Alexander Poltorak, working with Beutler, used a positional cloning strategy across a genetic cross of LPS-hyporesponsive and LPS-responsive mice to map the causal locus. They found that C3H/HeJ mice carry a point mutation in the Tlr4 gene (a proline-to-histidine substitution in the cytoplasmic TIR domain, abolishing signaling), while C57BL/10ScCr mice carry a complete deletion of the Tlr4 gene. Both independent mutations converged on the same gene, providing airtight genetic proof that TLR4 is the LPS receptor, and that its signaling is both necessary and sufficient to explain the pathological consequences of endotoxin recognition, including septic shock.

MyD88 knockout mice — Akira and colleagues (1998–2000). Shizuo Akira's group generated MyD88-deficient mice and showed that macrophages from these mice failed to produce TNF-α, IL-6, and IL-12 in response to IL-1, IL-18, and most TLR ligands (LPS, peptidoglycan, CpG DNA), while residual NF-κB activation and cytokine production to LPS persisted with delayed kinetics — this residual signal was the first clue that a MyD88-independent pathway existed, which was subsequently traced to TRIF and TRAM, cloned and functionally characterized by Yamamoto, Akira, and colleagues in the early 2000s. Together, these knockout studies established the branching logic (MyD88-dependent early NF-κB versus TRIF-dependent IRF3/interferon) that defines the modern signaling diagram.

Example 1 — Tracing a signaling defect from phenotype to mutation. A researcher has three mouse macrophage cell lines, X, Y, and Z. All three fail to produce TNF-α in response to LPS stimulation. Line X still produces IFN-β normally after LPS stimulation. Line Y produces neither TNF-α nor IFN-β after LPS, but responds normally to poly(I:C) (a TLR3 ligand) with strong IFN-β induction. Line Z fails to respond to LPS, poly(I:C), CpG DNA, and IL-1 alike, with zero cytokine or interferon output to any of these stimuli. What is defective in each? Reasoning: TNF-α production after LPS is driven predominantly through MyD88 (the early, surface TLR4 signal); IFN-β after LPS requires TRIF (the later, endosomal TLR4 signal). Line X loses TNF-α but keeps IFN-β, so its defect must be specific to the MyD88 branch downstream of TLR4 — most consistent with a MyD88 or TIRAP deletion, or an IRAK4/TRAF6 defect that spares TRIF signaling. Line Y loses both TNF-α and IFN-β after LPS but retains a normal TLR3 → IFN-β response; since poly(I:C) signals exclusively through TRIF and this is intact, the shared defect for LPS must lie specifically in TLR4's own signaling apparatus or its adaptors that are unique to TLR4 (e.g., TRAM, the TLR4-specific bridging adaptor for TRIF, or the TLR4/MD-2 receptor complex itself) — not in TRIF, IRF3, or TBK1, since those are shared with the intact TLR3 pathway. Line Z fails every stimulus, including IL-1 (which signals through the IL-1 receptor, not a TLR, but converges on the identical MyD88→IRAK→TRAF6→NF-κB cascade) and CpG (TLR9, obligately MyD88-dependent) as well as poly(I:C) (obligately TRIF-dependent) — the only node shared by all of these pathways is downstream of adaptor recruitment, at TAK1 or the IKK complex itself, since a defect there would collapse both MyD88- and TRIF-driven NF-κB output while an adaptor-specific defect (MyD88 alone, or TRIF alone) could not explain total pan-pathway failure.

Example 2 — Quantitative estimate of LPS sensitivity via CD14. Human monocytes can respond detectably to LPS concentrations as low as roughly 10 picograms per milliliter when LPS-binding protein and soluble or membrane CD14 are present, but require roughly 1,000-fold higher concentrations (around 10 nanograms per milliliter) to respond at comparable magnitude in the absence of functional CD14. Suppose a clinical microbiology lab measures a patient's plasma LPS at 50 pg/mL during early suspected Gram-negative bacteremia. Using the CD14-replete threshold, is this level physiologically significant? Reasoning: 50 pg/mL is five-fold above the roughly 10 pg/mL threshold at which CD14-assisted TLR4 signaling becomes detectable, so this concentration is well within the range capable of triggering a measurable innate immune response — consistent with early cytokine release (TNF-α, IL-6) even before overt clinical signs of shock appear, since TLR4/MD-2/CD14 sensing operates at concentrations roughly a thousand-fold below what would be required if CD14 were absent or non-functional. This calculation illustrates why endotoxin is dangerous even at trace plasma concentrations undetectable by older, less sensitive assays, and why CD14 is described as increasing TLR4 ligand sensitivity by roughly three orders of magnitude.

Bruce Beutler trained as a physician-scientist and, in the early 1980s at Rockefeller University, identified TNF-α as the principal mediator of endotoxin-induced shock — showing that anti-TNF antibodies could rescue mice from otherwise lethal LPS challenge. This early work convinced him that a specific, genetically tractable receptor for LPS must exist upstream of TNF release. Rather than searching biochemically for the receptor, Beutler's laboratory took a forward-genetics approach: they used the long-known LPS-hyporesponsive mouse strains C3H/HeJ and C57BL/10ScCr as a natural genetic experiment, reasoning that whatever gene was broken in both strains, independently, must be the LPS sensor. Positional cloning across large backcross populations, completed with Alexander Poltorak in 1998, pinned the defect to Tlr4 — a point mutation in one strain, a full gene deletion in the other. This work, alongside Jules Hoffmann's Drosophila Toll studies and Ralph Steinman's discovery of the dendritic cell, earned the 2011 Nobel Prize and firmly established pattern recognition, rather than the adaptive immune system alone, as the necessary first line of vertebrate host defense.

TLR biology now sits directly in the clinic. Vaccine adjuvants exploit TLR agonism deliberately: monophosphoryl lipid A (MPL), a detoxified derivative of LPS that retains TLR4 agonist activity but has far lower toxicity, is used in licensed vaccines (including a hepatitis B vaccine and an HPV vaccine) to boost adaptive immune priming by activating dendritic cells through the same MyD88/TRIF machinery described above. Imiquimod, a topical TLR7 agonist, is FDA-approved to treat genital warts and superficial basal cell carcinoma by locally triggering antiviral/antitumor interferon and cytokine responses. Conversely, TLR4 antagonists (such as eritoran, a lipid A mimetic that blocks MD-2) have been trialed, though not yet successfully, as therapies to blunt the cytokine storm of Gram-negative sepsis, underscoring how the same pathway that protects can also be pharmacologically dampened when it turns pathological. In autoimmunity, endosomal TLR7 and TLR9 sensing of self-derived nucleic acid–containing immune complexes is now understood to be a central driver of lupus pathogenesis, motivating TLR7/9 antagonists currently in clinical development. In cancer immunotherapy, TLR9 agonists (CpG oligonucleotides) are used as intratumoral injections to convert immunologically "cold" tumors into inflamed, T cell-infiltrated ones. And in basic hematopoiesis and stem cell biology, chronic TLR signaling in bone marrow niches has been shown to influence hematopoietic stem cell fate, linking innate sensing to blood cell production during chronic infection. Understanding TLR ligand specificity and the MyD88/TRIF branch point is therefore not just foundational immunology — it is the mechanistic basis of a growing pharmacopeia, and it sets up the next layers of the innate immune story: how the cytokines induced downstream (covered under Cytokines and Chemokines: Signaling Molecules of Immunity) recruit and instruct other immune cells, and how the TRIF/IRF3 branch specifically funnels into the dedicated antiviral defense system covered under Type I Interferons and Antiviral Defense.

Ask the tutor: How does the two-phase (MyD88-then-TRIF) signaling of endocytosed TLR4 actually shape the timing of fever, TNF-α release, and interferon production during a real Gram-negative infection?