TLR Innate Sensing Pathway: Receptors, MyD88/TRIF and Assays

Toll-like receptors are the reason a milligram of endotoxin can kill a mouse. They are the fastest, least discriminating layer of immunity: germline-encoded sensors that read conserved microbial chemistry and, within minutes, commit a cell to an inflammatory or an interferon programme. The pathway looks simple in outline — ligand, receptor, adaptor, transcription factor — but almost everything that matters is decided by two variables: which compartment the receptor sits in, and which adaptor it recruits.

The sensors and why their address matters

The map divides at the membrane. TLR2, heterodimerising with TLR1 or TLR6, reads acylated lipopeptides at the cell surface — triacylated through the 2/1 pair, diacylated through 2/6. TLR4 reads lipid A. TLR5 reads the D1 domain of flagellin, a region held under structural constraint because it forms the filament interface and so cannot mutate away from detection. All three see chemistry a bacterium cannot easily abandon.

The endosomal set is different in kind. TLR3 binds double-stranded RNA, TLR7 single-stranded RNA, TLR9 unmethylated CpG DNA — none of which is inherently foreign. What makes them safe as danger signals is address rather than chemistry: they are chaperoned from the ER to endolysosomes by UNC93B1 and need ectodomain cleavage and acidic pH before they signal. That constraint is the basis of nucleic-acid tolerance, and it fails when immune complexes ferry self-DNA to a compartment where TLR9 is already licensed.

LPS recognition: CD14, MD-2 and the two lives of TLR4

TLR4 is the one receptor here that cannot work alone. Lipopolysaccharide is extracted from the bacterial membrane by LBP, handed to CD14, and delivered as a monomer to MD-2 (LY96), the accessory protein that sits in the TLR4 ectodomain pocket and grips the acyl chains. Only when MD-2 is loaded does TLR4 dimerise. An MD-2-null cell is therefore functionally TLR4-deficient despite normal receptor expression, and tetra-acylated lipid A analogues antagonise because they fill the pocket without forcing the dimer.

The non-obvious part is what CD14 does next. Beyond ligand transfer it drives TLR4 internalisation, and the internalised receptor signals again from the endosome — this time through TRIF rather than MyD88. TLR4 has two sequential lives: an early surface burst of NF-κB-driven TNF-α, then a later endosomal wave of IRF3-driven IFN-β. A CD14-deficient cell makes TNF-α but little IFN-β from the same LPS dose.

The Myddosome: MyD88, IRAK4/1 and TRAF6

Every TLR here except TLR3 signals through MyD88. Dimerisation juxtaposes the cytoplasmic TIR domains, which nucleate MyD88 by TIR–TIR contact; TLR2 and TLR4 additionally need the bridging adaptor TIRAP/Mal. The MyD88 death domains then polymerise with IRAK4 and IRAK1 into a helical Myddosome of roughly six to eight MyD88, four IRAK4 and four IRAK1/2 subunits. Because it nucleates cooperatively, sub-threshold occupancy gives no complex at all — a digital switch with built-in gain, which is why TLR dose–response curves are so steep.

Activated IRAK1 then engages TRAF6, an E3 ligase that tags nothing for destruction. With Ubc13/Uev1A it builds K63-linked polyubiquitin on itself and on NEMO, and those chains act as a scaffold: TAB2 and TAB3 bind them and drag TAK1 (MAP3K7) into position to autophosphorylate. Deubiquitinases such as A20 and CYLD dismantle the same chains, so signal duration is set by a ubiquitin editing contest rather than ligand persistence — which is why loss of A20 produces spontaneous inflammation.

TAK1 to the nucleus: NF-κB, p38/JNK and AP-1

TAK1 is the branch point and does two things at once. It phosphorylates IKKβ within the IKKα/IKKβ/NEMO complex, which marks IκBα on serines 32 and 36 for β-TrCP-directed destruction and frees NF-κB p65/p50. In parallel it phosphorylates MKK3/6 and MKK4/7, activating p38 and JNK, which in turn phosphorylate the Fos and Jun components of AP-1. The two factors co-occupy the promoters of the same inflammatory genes.

That combinatorial requirement is what stops TLR signalling being all-or-nothing at the transcriptional level. TNF-α and IL-6 need both factors plus chromatin remodelling. IL-1β is different again: NF-κB supplies only the pro-form, and a separate inflammasome-dependent caspase-1 cleavage is required before anything is secreted — which is why LPS alone yields abundant TNF-α but little mature IL-1β.

The TRIF arm, type I interferon and adjuvant design

TLR3 binds no MyD88 and signals exclusively through TRIF; endosomal TLR4 recruits TRIF via the bridging adaptor TRAM. TRIF has two outputs: its RHIM domain engages RIP1 for a delayed second phase of NF-κB, while its N-terminus recruits TRAF3 to assemble TBK1 with IKKε. TBK1 phosphorylates the C-terminal serine cluster of IRF3, driving dimerisation and CBP/p300 recruitment at the IFN-β enhanceosome. IRF7 then amplifies: it is itself an interferon-stimulated gene, so the first IFN-β pulse through IFNAR induces it and broadens the response.

Plasmacytoid dendritic cells short-circuit that loop by holding IRF7 constitutively high and coupling it directly to MyD88 downstream of TLR7 and TLR9 — the one context in which the MyD88 arm makes type I interferon rather than inflammatory cytokines. Adjuvant design exploits exactly these branch points: monophosphoryl lipid A is a TRIF-biased TLR4 agonist with much reduced MyD88-driven pyrogenicity, imiquimod targets TLR7, CpG 1018 targets TLR9. Reading IFN-β alongside TNF-α and IL-6 is how you tell a biased agonist from a merely weak one.

Key targets and matching reagents

Studying TLR innate sensing in vivo

TLR experiments are unusually easy to contaminate, because the reagents used to perturb the pathway can trigger it. The work falls into three groups.

1. Neutralising the inflammatory output

The two cytokines carrying most of the acute phenotype are addressable with functional-grade antibodies: anti-mouse TNF-α and anti-mouse IL-1β, both low-endotoxin. In an LPS challenge model they separate the fast TNF-driven component of shock from the slower inflammasome-dependent IL-1β component. Endotoxin content of the antibody is not a formality here: a preparation carrying nanogram quantities of LPS is an agonist of the pathway you are trying to block. The full In Vivo antibody range is formulated accordingly.

2. Quantifying receptor and adaptor abundance

Perturbation experiments need denominators. TLR2, TLR3, TLR4, TLR5, TLR7 and TLR9 kits confirm that a population expresses the sensor you are stimulating. Pair them with CD14 and MD-2 for the LPS co-receptor module, and with MyD88 and TRIF to establish which adaptor arm is available at all — adaptor abundance often explains a discrepant TLR4 result between macrophages and dendritic cells.

3. Reading the signalling module and its transcriptional output

Between receptor and cytokine sit the nodes that tell you where a block acts. IRAK4, TRAF6, TAK1 (MAP3K7), IKKβ and p38 (MAPK14) cover the MyD88 arm; TBK1, IRF3 and IRF7 cover the interferon arm. Finish with the transcription factors and their products — NF-κB p65, AP-1 (Jun), then TNF-α, IL-6, IL-1β and IFN-β in serum or supernatant.

Featured products for this pathway

Frequently asked questions

For research use only. Not for use in diagnostic or therapeutic procedures.