Canonical NLRP3 inflammasome assembly and pyroptotic output

A decomposition of the canonical NLRP3 inflammasome: the cytosolic sensor-adaptor-protease platform that converts a diffuse cellular perturbation into active caspase-1, and the lytic execution arm that caspase-1 then triggers. The chain is linear and each hand-off is a change of molecular currency rather than of chemistry: NLRP3 senses a perturbation indirectly and presents a pyrin-domain filament (GO:0044546), that filament nucleates ASC polymerisation into the one-per-cell speck (GO:0061703), the clustered ASC CARDs recruit procaspase-1 and permit proximity-induced autoprocessing (GO:0061702), the autoprocessed protease cleaves gasdermin D and the interleukin-1-family precursors, the released GSDMD N-terminal fragment oligomerises into wide plasma-membrane pores (GO:0141201), and NINJ1 finally ruptures the membrane. Only one of the seven steps is catalytic - caspase-1 substrate cleavage. NLRP3, ASC and NEK7 supply geometry, and the two terminal effectors (GSDMD and NINJ1) act on the membrane rather than running chemistry, so the module is best read as a geometry cascade with a single proteolytic step in the middle and two independent membrane-acting effectors at the end. Design intent: one ordered chain of parts, like necroptosis and unlike the redundant-suppressor topology of ferroptosis, because there is no second route from the speck to the pore. The two places where the biology is genuinely alternative are handled differently from each other. NEK7 licensing is an optional regulatory part rather than a core step, because human myeloid cells can activate NLRP3 without it when IKKbeta has recruited NLRP3 to PI4P (PMID:36384135) and because raising NLRP3 abundance makes potassium-efflux-driven activation independent of both NEK7 and priming (PMID:42378282). Interleukin-1 beta and interleukin-18 are recorded as caspase-1 substrates inside the cleavage step and are deliberately given no annoton of their own: they are what the protease acts on, they perform none of the work of their own maturation, and their cytokine activity belongs to the receptor pathways downstream of this module.

Boundary decisions. (1) Scope is the canonical NLRP3 inflammasome and the pyroptotic execution it triggers: sensor activation, adaptor polymerisation, caspase-1 autoactivation, caspase-1 substrate cleavage, GSDMD pore formation, NINJ1 membrane rupture. Signal 1 (priming) is explicitly out of the core. NF-kappaB-driven transcriptional upregulation of NLRP3 and IL1B is reached by reference to modules/nfkb_canonical_signaling.yaml, and the receptor biology that generates that signal belongs to modules/tlr4_md2_lps_signaling.yaml and modules/tnf_signaling.yaml. Priming does real mechanistic work - it raises NLRP3 abundance and, through IKKbeta, moves NLRP3 into a PI4P-bound primed state - so it is modelled as context connected to those modules, not as a part here, and PMID:42378282 is cited for abundance substituting for it. (2) Other inflammasomes and other routes to a gasdermin pore are out of scope by the same logic. The non-canonical CASP4/CASP5 route to GSDMD (reviewed in genes/human/CASP4) reaches the same execution arm without NLRP3, ASC or caspase-1; NLRC4, NLRP1, AIM2, CARD8 and NLRP6 are sibling sensors on the same ASC/caspase-1 core; granzyme-driven GSDME cleavage is a third route into GO:0141201. None of them is modelled here. Those shared downstream parts (caspase-1 autoactivation onwards) are the natural factoring point if this module is later generalised to a canonical-inflammasome module with a sensor variant set; modules/nlr_signaling.yaml is the existing, much coarser sketch of the sensor layer and is not superseded by this module. (3) NEK7 is an optional licensing part, not an obligate core step, and this is the module's main judgement call. The case for including it is strong: NEK7 acts downstream of potassium efflux, is a stoichiometric subunit of the active disc, bridges adjacent NLRP3 LRRs, and is required in mouse macrophages in vivo. The case against making it obligatory is specific and human: when IKKbeta recruits NLRP3 to PI4P, NEK7 is redundant for inflammasome formation, and that IKKbeta-mediated route is reported as the predominant priming mechanism in human myeloid cells (PMID:36384135); raising NLRP3 abundance likewise renders potassium-efflux-driven activation independent of both NEK7 and priming (PMID:42378282). A part whose loss can be fully compensated in the organism the module is grounded in is regulation, so NEK7 carries optional: true and a POSITIVELY_REGULATES edge onto the sensor step rather than sitting in the chain. Its activity is asserted as GO:0140677 molecular function activator activity, not as a kinase activity: kinase-dead NEK7 rescues inflammasome activation. The local PAINT row for NEK7 in PTHR44329 propagates GO:0004674 protein serine/threonine kinase activity, which is the mitotic function and not this one, so it is deliberately not cited as support for the licensing role. (4) IL-1 beta and IL-18 are substrates, not participants. They are named in the substrates list of the caspase-1 cleavage annoton and nowhere else: the proteolysis is performed entirely by caspase-1, the precursors contribute no residue, no positioning and no chemistry to their own cleavage, and being required for the output is not the same as doing any of its work. Their cytokine activity (GO:0005125 in the IL-1 and IL-18 receptor pathways, as modelled in the cached GO-CAM) is downstream biology outside this boundary. The one thing the module does assert about them is a property of the pore, not of the cytokines: the GSDMD conduit favours the mature forms over their precursors (PMID:33883744), which is why cytokine release is placed on the pore annoton. Neither gene has a review under genes/human/, which is a further reason not to assert activities for them here. (5) Two membrane-acting effectors, two different acts. GSDMD-NT forms a defined wide pore and is asserted with GO:0022829, following the gene review; NINJ1 is asserted with GO:0140912 membrane destabilizing activity, the same term the necroptosis module uses for MLKL, and its description is phrased to match - both are lytic effectors that destabilise the bilayer rather than forming a gated channel. Note that GSDMD pore formation and NINJ1 rupture are not alternatives: the pore comes first and is survivable, rupture is terminal. (6) No module-level deep research exists for this module. No provider run was available when it was curated (the Falcon endpoint was unavailable), so there is deliberately no nlrp3_inflammasome-deep-research-<provider>.md file. Every structural claim is grounded in the six per-gene reviews under genes/human/, cached publications under publications/, local PAINT rows under interpro/panther/, and the two cached production GO-CAMs named in evidence. If module-level deep research is run later it should be named canonical_nlrp3_inflammasome-deep-research-<provider>.md. (7) Taxon scope is mammalian, represented with human gene products, although the GO terms used carry an in-taxon Vertebrata constraint and the architecture is broader: the cached zebrafish GO-CAM gomodel:60418ffa00001019 models an NLRP3 inflammasome driving pyroptosis through caspase-a/caspase-b and gasdermin Eb, i.e. through non-orthologous caspases and gasdermins. The module is therefore not asserted for vertebrates as a whole, since NEK7 licensing, the GSDMD step and NINJ1 are all characterised in mammals. (8) History. This file began as a seven-part CONCRETE decomposition named nlrp3_inflammasome, built from the NLRP3, PYCARD, CASP1, NEK7, GSDMD and NINJ1 gene reviews. A separate four-part ABSTRACT overview, canonical_nlrp3_inflammasome, was created in #4960 and grounded on GO-CAM gomodel:635b1e3e00001811. The two were folded together under the canonical id and title because they modelled one pathway; the concrete seven-part decomposition below is kept (it carries the step-level gene functions the overview lacked), and the overview's GO-CAM grounding is retained on the matching annotons. The scope decisions in (1)-(7) are unchanged by the merge. module_type was widened from SIGNALING_PATHWAY (the overview's value) to BIOLOGICAL_PROCESS, since the merged module spans both the signalling assembly chain and the lytic execution arm. (9) CANDIDATE ASSESSED, NOT MODELLED - VPS4B at the ASC speck (PMID:42714851, litscan handoff #3920). A 2026 report identifies the ESCRT AAA-ATPase Vps4B as a component of the NLRP3 and AIM2 inflammasomes that interacts endogenously with ASC and caspase-1, is recruited to the speck by Ca2+ signalling downstream of NLRP3 and ASC polymerisation, and forms a ring encircling the ASC filament that disassembles and liberates the used caspase-1 CARD, so that unoccupied ASC CARDs are replenished and caspase-1 recruitment continues. Routing first: the handoff proposed modules/nlr_signaling.yaml, but that file is the coarse sensor-layer sketch and its own notes delegate the NLRP3 branch here, so this module is where the question had to be answered. If the candidate were admitted it would attach to the casp1_autoactivation part (order 4), which is the step that builds the caspase-1 CARD platform this activity would recycle. It is deliberately not added, for three reasons. (i) Weight: a single primary report, cached abstract-only here, with no independent replication. VPS4B carries no inflammasome GO annotation (genes/human/VPS4B/VPS4B-goa.tsv, 94 rows, none inflammasome- or pyroptosis-related) and appears in no cached GO-CAM (gocams/index.tsv), so nothing outside this one abstract places it at the speck. (ii) Participant identity: VPS4B is not known to act alone. UniProt records that it interacts with VPS4A and that the interaction suggests a heteromeric assembly, and ComplexPortal CPX-338 is the VPS4A/B complex; the abstract does not say whether VPS4A is present at the speck or redundant there, so a GENE_PRODUCT annoton on VPS4B alone could mis-state the functional unit, and the choice between that and a PROTEIN_COMPLEX participant is not determinable from the available evidence. (iii) Mechanism and selectivity: the reported geometry is the canonical VPS4 mechanism - ATP-dependent oligomerisation into stacked hexameric or heptameric rings that translocate a substrate polymer through a central pore, as for CHMP2A-CHMP3 helical tubes - which makes the activity plausible rather than anomalous. But the report seats the ring on the ASC filament while making the caspase-1 CARD polymer the substrate, and how that selectivity is achieved, and whether ATP hydrolysis by VPS4B is required for the inflammasome phenotype at all, is not established by the abstract. Asserting an ATP-driven disassembly activity here would therefore read mechanism into a correlation. Re-evaluate and add when all three are met: independent replication of the ASC/caspase-1 interaction and the speck-renewal phenotype; evidence fixing the functional unit (whether VPS4A is required, which decides the selector type); and an ATPase-dead separation of function showing the phenotype needs ATP hydrolysis, which is what would license GO:0140545 ATP-dependent protein disaggregase activity or GO:0051261 protein depolymerization at this step rather than a binding term. The admission path is then a regulatory annoton on casp1_autoactivation with optional: true and a POSITIVELY_REGULATES edge onto that node, by the same reasoning that makes NEK7 optional in (3): a factor that renews a platform the module already builds is regulation of throughput, not a step in the chain. This candidate is also recorded as a structured, queryable entry in the knowledge_gaps block below (status OPEN). The genes/human/VPS4B review is unchanged: no NEW annotation is proposed there from one abstract.