AIGR Deep Research Report: ABHD8 (Q96I13) — Lipid Catalysis versus Inflammasome Adaptor

Gene: ABHD8 | Organism: Homo sapiens (NCBITaxon:9606) | UniProt: Q96I13
Focus type: function_assignment | Hypothesis slug: lipid-catalysis-versus-inflammasome-adaptor
Source: genes/human/ABHD8/ABHD8-ai-review.yaml (free-text selector)

Summary

The seed hypothesis proposes that human ABHD8 retains one or more specific lipid functions — glycerophospholipase activity, generic ester hydrolase activity, lysophosphatidic acid (LPA) acyltransferase activity, participation in phosphatidic acid biosynthesis, or lipid homeostasis — and asks that each be adjudicated separately from ABHD8's demonstrated NLRP3 adaptor role. It further argues, correctly as a methodological principle, that neither the presence of an adaptor function nor the absence of substrate experiments by itself establishes catalytic or pathway loss. After three iterations of sequence, domain, orthology, and literature analysis, the evidence resolves into a clean split between catalytic potential (supported) and specific lipid activities (unsupported by direct evidence, and in one case mechanistically implausible).

ABHD8 possesses a fully intact, canonical serine-hydrolase catalytic triad — Ser252 (nucleophile), Asp370 (acid), His398 (base) — embedded in a classic GXSXG nucleophile elbow (the motif GHSYG) within a complete α/β-hydrolase (AB_hydrolase-1) domain. This triad is conserved across all reviewed vertebrate orthologs from human to Xenopus, spanning more than 360 million years of divergence, which indicates purifying selection to preserve catalytic competence. ABHD8 is therefore not a dead pseudoenzyme, and a blanket claim of catalytic or pathway loss would be unjustified. This directly vindicates the seed's caution.

At the same time, no ABHD8-specific biochemical assay demonstrates any lipid substrate. The specific lipid GO terms are phylogenetic/family inferences (IBA from GO_Central; IEA from InterPro), not experiments. The LPA-acyltransferase claim is mechanistically inconsistent with a Ser-His-Asp hydrolase fold and almost certainly reflects paralog/family carry-over. The only experimentally demonstrated function is the NLRP3 scaffold/adaptor role: ABHD8 recruits the palmitoyltransferase ZDHHC12 to NLRP3, promoting NLRP3 palmitoylation and chaperone-mediated-autophagy (CMA) degradation (PMID: 39225180). Accordingly, curation should retain the NLRP3 term (GO:1900226, IMP) and cytoplasm (GO:0005737, IDA) as core, keep any lipid-catalysis annotation at most as generic hydrolase potential (GO:0003824, IEA), and treat the specific lipid terms as over-annotations pending direct substrate assays.

Verdict: Partially supported / unresolved — leaning over-annotated for the specific lipid activities, while catalytic potential is genuinely supported.

Key Findings

Finding 1 — ABHD8 has an intact canonical serine-hydrolase catalytic triad but no demonstrated substrate

UniProt Q96I13 annotates an AB_hydrolase-1 domain spanning residues ~177–279, with the nucleophile-elbow GXSXG motif present as G250-H-S252-Y-G254 (GHSYG). The charge-relay / catalytic triad is annotated at Ser252 (nucleophile), Asp370 (acid), His398 (base), and the residue identities S/D/H were confirmed directly from the sequence. The protein maps to InterPro IPR000073 (AB_hydrolase_1), Pfam PF00561, and the ABHD8-specific PANTHER subfamily PTHR42886:SF83.

Critically, UniProt carries no CATALYTIC ACTIVITY comment and no specific reaction for ABHD8. The only FUNCTION comment is the NLRP3 scaffold role citing PMID 39225180. The molecular-function GO annotation is the generic GO:0003824 (catalytic activity) with evidence IEA:InterPro only. In other words, the domain architecture supports catalytic capability, but nothing in the primary record pins down what ABHD8 actually hydrolyzes. The presence of a textbook triad establishes potential; it does not establish any specific reaction. This is the crux of the adjudication: catalytic potential and demonstrated function are, at present, decoupled.

Finding 2 — The specific lipid-pathway GO terms are phylogenetic/family inference, not experiment

The GO annotation profile for Q96I13 separates cleanly into "inferred" versus "experimental":

The "phosphatidic acid biosynthetic process" term is IBA (inferred from biological ancestor) — it derives from the behavior of other members of the phylogenetic tree, not from any ABHD8 assay. The seed-listed activities (glycerophospholipase, ester hydrolase, LPA acyltransferase) have no ABHD8-specific primary assay anywhere in the reviewed literature. Furthermore, LPA acyltransferase is an acyltransfer reaction that is mechanistically incompatible with a Ser-His-Asp hydrolase fold — the α/β-hydrolase triad catalyzes hydrolysis, not the acyl-CoA-dependent acyltransfer of a canonical LPAAT. This annotation is therefore the strongest candidate for paralog/family carry-over error.

The only experimentally demonstrated function — the NLRP3 scaffold (PMID 39225180) — is an adaptor/scaffold mechanism, not a demonstrated hydrolase reaction: ABHD8 recruits palmitoyltransferase ZDHHC12 to NLRP3 to promote NLRP3 palmitoylation and CMA-mediated degradation.

Finding 3 — The catalytic triad is conserved across vertebrate orthologs (purifying selection)

Across all five reviewed UniProt ABHD8 orthologs — Homo sapiens (Q96I13), Mus musculus (Q8R0P8), Bos taurus (Q17QP1), Macaca fascicularis (Q4R584), and Xenopus laevis (Q6AX59) — the GHSYG nucleophile-elbow motif (Ser nucleophile), the catalytic Asp (human context VHG-M-HDK-F), and the catalytic His (human context DEG-S-HM) are all present. The human→Xenopus comparison spans >360 million years of divergence, and retention of the complete Ser-His-Asp triad plus the GXSXG elbow across that span is strong evidence of purifying selection to preserve catalytic competence.

This is the single most important piece of evidence for the seed hypothesis: an enzyme whose catalytic machinery has been conserved for hundreds of millions of years is unlikely to be a mere structural relic. It argues that ABHD8 does have a genuine catalytic activity — we simply do not yet know its substrate. Conservation, however, cannot tell us which substrate, and it does not license the specific lipid terms currently annotated. It is entirely compatible with a hydrolase acting on a substrate class not yet tested.

Finding 4 — Full text confirms the NLRP3 function is scaffold-only, with no catalysis invoked

The full abstract of PMID 39225180 (Yang et al., 2025, Autophagy) states that ABHD8 "acts as a scaffold to recruit palmitoyltransferase ZDHHC12 to NLRP3 for its palmitoylation as well as subsequent CMA-mediated degradation." ABHD8 deficiency stabilizes NLRP3 and promotes inflammasome activation; overexpression ameliorates LPS/alum-triggered activation in vivo; and the SARS-CoV-2 N protein disrupts the ABHD8–NLRP3 association. Crucially, the abstract contains no mention of ABHD8 hydrolase or lipase enzymatic activity, no lipid substrate, no phosphatidic acid or LPAAT reaction, and no catalytic-site (Ser252Ala) mutant experiment. The mechanism is entirely consistent with a scaffold/adaptor that positions an enzyme (ZDHHC12) next to its target (NLRP3) — ABHD8 itself need not be catalytically active to perform this role.

Two verified citation snippets anchor this finding:
- "ABHD8 acts as a scaffold to recruit palmitoyltransferase ZDHHC12 to NLRP3 for its palmitoylation as well as subsequent CMA-mediated degradation." — directly states the demonstrated mechanism is a scaffold/adaptor role, not a catalytic reaction.
- "ABHD8 deficiency results in the stabilization of NLRP3 protein and promotes NLRP3 inflammasome activation" — establishes the loss-of-function phenotype underpinning the GO:1900226 negative-regulation term without invoking catalytic activity.

Mechanistic Model / Interpretation

The evidence supports a two-track model in which catalytic potential and demonstrated function are currently decoupled:

                 ABHD8 (Q96I13)
                      │
┌─────────────────────┴─────────────────────┐
│                                            │
  CATALYTIC POTENTIAL                       DEMONSTRATED FUNCTION
  (structure / evolution)                   (experimental — PMID 39225180)
│                                            │
  AB_hydrolase-1 domain                     NLRP3 inflammasome control
  GHSYG nucleophile elbow                          │
  Triad: Ser252–Asp370–His398          ABHD8 = scaffold ──► recruits
│                              ZDHHC12 (palmitoyltransferase)
  Conserved human→Xenopus                            │
  (>360 Myr, purifying selection)        NLRP3 palmitoylation ──► CMA
│                                degradation of NLRP3
  ⇒ NOT a dead pseudoenzyme                          │
│                              ABHD8 loss ⇒ NLRP3 stabilized ⇒
  ✗ NO known substrate                 inflammasome activation ↑
  ✗ NO ABHD8-specific assay                          │
  ✗ Specific lipid GO = IBA/IEA         ⇒ GO:1900226 (neg. reg. NLRP3), IMP
│                              ⇒ GO:0005737 (cytoplasm), IDA
  LPAAT term mechanistically
  INCONSISTENT with hydrolase fold

The key interpretive point is that the demonstrated NLRP3 role does not require ABHD8 catalysis — a scaffold that co-localizes an acyltransferase (ZDHHC12) with its substrate (NLRP3) can be catalytically silent. Conversely, the conserved triad does not require the NLRP3 role to make sense — it strongly implies a real (but undiscovered) hydrolase activity, possibly acting on a lipid or ester substrate that has simply never been assayed for ABHD8. These two tracks may or may not intersect: it is possible that ABHD8's hydrolase activity contributes to some aspect of the palmitoylation/degradation cycle, but no experiment has tested this.

A useful comparator is the paralog ABHD16A/BAT5 (PMID: 25290914), which was a "poorly characterized" ABHD-family serine hydrolase of unknown natural substrate until activity-based protein profiling and a fluorescent glycerol assay revealed it to be a genuine lipase preferring long-chain unsaturated monoacylglycerols (MAGs), with "only marginal" lysophospholipase, DAG, or TAG activity. This illustrates two things for ABHD8: (a) family membership alone did not predict the real substrate — it turned out to be MAGs, not the phospholipids one might guess — and (b) the substrate had to be found empirically. By analogy, transferring specific lipid-pathway terms to ABHD8 by phylogeny is exactly the kind of inference that direct assay of ABHD16A overturned.

Evidence Base / Evidence Matrix

GO Curation Implications

Lead requiring curator verification. The evidence supports the following curation actions:

Net recommendation: Treat the NLRP3-degradation scaffold role as the primary, experimentally supported function (GO:1900226, IMP), with cytoplasm (GO:0005737, IDA) as the supported localization. Keep any hydrolase MF at the most generic level (GO:0003824, IEA) explicitly flagged as "catalytic potential, substrate unknown." Do not promote glycerophospholipase, ester-hydrolase, LPA-acyltransferase, or phosphatidic-acid biosynthesis to experimental evidence codes. Do not downgrade the final recommendation to bare "protein binding" — if a binding term is added, name the interactant (ZDHHC12 / NLRP3); the informative annotation is GO:1900226, which should lead.

Mechanistic Scope

The immediate molecular function actually demonstrated for ABHD8 is protein scaffolding: it physically bridges the palmitoyltransferase ZDHHC12 to NLRP3, enabling NLRP3 palmitoylation and its subsequent CMA degradation. This is a direct gene-product activity (a binding/adaptor function), distinct from:

• Downstream pathway consequence: reduced NLRP3 protein, attenuated inflammasome assembly, lower IL-1β — consequences of the scaffold activity, not a separate ABHD8 activity.
• Loss-of-function phenotype: ABHD8 deficiency → NLRP3 stabilization → inflammasome hyperactivation — this defines the negative-regulation BP term but is an inference from loss of the scaffold.
• Disease manifestations: GWAS/eQTL associations (breast/ovarian cancer 19p13.1, mtDNA copy number, aerodigestive SqCC) implicate ABHD8 expression but are not evidence of enzymatic function and must not be conflated with molecular activity.

The hypothesized lipid-hydrolase activity, by contrast, is only inferred from domain architecture and family membership. It has a strong structural/evolutionary basis (conserved triad) but zero direct assay evidence in ABHD8. It must therefore be scoped as "catalytic potential, substrate unknown," not as an established molecular function.

Conflicts and Alternatives

1. Paralog/family carry-over (most important). The specific lipid terms most plausibly originate from generic α/β-hydrolase (Pfam PF00561, InterPro IPR000073) and PANTHER (PTHR42886) inheritance rather than from ABHD8 data. The ABHD16A/BAT5 case (PMID 25290914) shows that ABHD family members can be genuine lipases, but their real substrates are often not what family-level inference predicts. This both supports the idea that ABHD8 could be a real hydrolase and undermines confidence in any specific transferred substrate term.

2. LPA acyltransferase is mechanistically inconsistent. An α/β-hydrolase Ser-His-Asp triad performs hydrolysis, not the acyl-CoA-dependent acyltransfer of a canonical LPAAT. This term is the strongest candidate for an outright annotation error.

3. Scaffold ≠ enzyme. The NLRP3 mechanism (PMID 39225180) explicitly frames ABHD8 as a scaffold. A scaffold can be catalytically silent; the demonstrated function does not require or test hydrolase activity. This competes with the interpretation that ABHD8's biological importance flows from catalysis.

4. Disease-genetics confounding. Multiple GWAS/eQTL papers (PMIDs 27601076, 33667223, 30704525, 30559148) associate the ABHD8 locus with cancer risk, mtDNA copy number, and other traits — but these reflect regulatory/expression effects at 19p13.1 (a locus shared with ANKLE1 and BRCA-pathway neighbors), not ABHD8 enzymatic activity. They should not be read as evidence for the lipid-catalysis hypothesis.

5. No isoform-specific catalytic data. No study distinguishes catalytic behavior among ABHD8 isoforms; conservation analysis was at the canonical-sequence level only.

Limitations and Knowledge Gaps

The overarching limitation: this analysis is literature- and sequence/annotation-based (UniProt, GO, PubMed). No wet-lab data were generated; the intact-triad inference is computational. Local repository *-bioinformatics analyses were intentionally withheld and not consulted. Findings are reported conservatively, distinguishing direct results (triad presence; NLRP3 scaffold assay) from inference (specific lipid substrates).

Proposed Follow-up Experiments / Actions

To most efficiently separate "genuine lipid hydrolase" from "catalytically silent scaffold," in rough priority order:

1. Activity-based protein profiling (ABPP) of recombinant human ABHD8 with fluorophosphonate/serine-hydrolase probes — a positive label directly demonstrates an active serine hydrolase, independent of substrate identity. This is exactly how ABHD16A/BAT5 was confirmed catalytically active (PMID 25290914).
2. Substrate panel hydrolysis assay (fluorescent glycerol / LC-MS lipidomics) across MAG, DAG, TAG, lysophospholipids, glycerophospholipids, and LPA to identify the preferred substrate — or to show ABHD8 is inactive on all tested lipids.
3. Catalytic-dead rescue in the NLRP3 pathway: re-express WT vs. Ser252Ala ABHD8 in ABHD8-null cells and measure NLRP3 palmitoylation/degradation and IL-1β. Full rescue by Ser252Ala implies the scaffold role is catalysis-independent; failure implies catalysis is mechanistically coupled.
4. AlphaFold-based active-site pocket analysis + molecular docking of candidate lipids to predict substrate class before wet-lab assay.
5. Dedicated LPAAT assay (acyl-CoA + LPA → PA) to directly test the mechanistically-suspect acyltransferase annotation; a negative result justifies removing the term.
6. Lipidomics of ABHD8-KO vs. WT cells to detect any endogenous lipid-homeostasis role (e.g., PA levels) attributable to ABHD8 and to test the phosphatidic-acid-biosynthesis IBA prediction directly.

Curation Leads (verify before applying)

• Candidate reference: PMID 39225180 — snippet to verify: "ABHD8 acts as a scaffold to recruit palmitoyltransferase ZDHHC12 to NLRP3 for its palmitoylation as well as subsequent CMA-mediated degradation." → supports retaining GO:1900226 (IMP) as core and a scaffold/protein-bridging annotation over lipase.
• Candidate reference (methodological/paralog): PMID 25290914 — establishes the ABPP + substrate-panel assay standard that ABHD8 lacks.
• Action-change leads: (a) do not elevate glycerophospholipase / ester hydrolase / LPAAT beyond IEA; (b) flag LPA-acyltransferase as likely incorrect; (c) mark GO:0006654 (phosphatidic acid biosynthesis) as IBA / non-core / candidate removal; (d) retain GO:1900226 and GO:0005737.
• Suggested curator questions: Is the specific lipid MF an author statement or purely pipeline inference? Was any triad mutant tested for the NLRP3 phenotype? Where did the LPAAT assignment originate?

Conclusion

Human ABHD8 (Q96I13) retains a fully intact, evolutionarily conserved serine-hydrolase catalytic triad, so it is not a dead pseudoenzyme and the seed's caution against inferring catalytic loss from adaptor function is justified. However, no ABHD8-specific biochemical assay demonstrates any lipid substrate: the specific terms (glycerophospholipase, ester hydrolase, LPA acyltransferase, phosphatidic acid biosynthesis, lipid homeostasis) are unproven family/phylogenetic inferences (IEA/IBA), with the LPA-acyltransferase term mechanistically inconsistent with the hydrolase fold. The only experimentally demonstrated function is the NLRP3-degradation scaffold/adaptor role (recruiting ZDHHC12 to NLRP3; GO:1900226 IMP, PMID 39225180). Curation should retain that core term and localization, keep lipid catalysis at most as generic hydrolase potential, and treat the specific lipid annotations as over-annotations pending direct substrate assays.

Artifacts

• OpenScientist final report
• OpenScientist final report