SLC10A2 (ASBT / IBAT / ISBT) curation notes

Provenance of this review

Automated deep research is not available in this container, so the literature
synthesis below was assembled by hand from the cached publications in publications/,
the cached Reactome entries, the UniProt record, PANTHER/PAINT donor tracing and
QuickGO comparator queries:

• Falcon/Edison deep-research API returns 402 Payment Required.
• OpenAI deep-research API returns 401 invalid_api_key.
• perplexity is not a registered provider in this checkout.

No -deep-research-*.md file was written, per CLAUDE.md (hand-written content must
never be named as a deep-research provider output).

Gene identity

• UniProt: Q12908 (Ileal sodium/bile acid cotransporter; ASBT; IBAT; ISBT; NTCP2)
• 348 aa, 7 predicted transmembrane helices, glycoprotein; bile acid:sodium symporter
  (BASS, TC 2.A.28) family; InterPro IPR002657; PANTHER PTHR10361 (same family as
  SLC10A1/NTCP)
• Seven-TM topology was experimentally supported by glycosylation-scanning mutagenesis
  PMID:15350125
• UniProt lists the subunit state as "Monomer and homodimer" and the subcellular
  location only as "Membrane; Multi-pass membrane protein" — i.e. UniProt itself does
  not record the apical restriction, which has to come from the literature (below).

Core transport function

• Reabsorption of bile acids from the intestinal lumen is the defining role
  PMID:7592981
• Strictly Na+-dependent, Cl--independent taurocholate uptake on heterologous
  expression PMID:9458785
• Both conjugated and unconjugated bile acids are substrates PMID:9458785; UniProt curates Km values
  of 12–17 uM for taurocholate, 33–37 uM for cholate, and 2–6 uM for the glyco-
  conjugates.
• Stoichiometry is 2 Na+ : 1 bile acid, and transport is electrogenic and
  bidirectional — measured directly in voltage-clamped CHO cells expressing the human
  protein [PMID:9856990 "These results indicate that the cotransport of bile acids and
  Na+ by human apical sodium-bile acid transporter is electrogenic and bidirectional and
  is best explained by a 2:1 Na+:bile acid coupling stoichiometry."; "A 3-fold reduction
  in extracellular Na+ produced a negative 52 mV shift of the flux-voltage relationship,
  consistent with a 2:1 Na+:bile acid coupling stoichiometry."]. UniProt's curated RHEA
  reactions (RHEA:71875 etc.) likewise use 2 Na+. Note that the cached Reactome reaction
  R-HSA-194187 writes the reaction with a single sodium ion; the experimental literature
  and UniProt both say two.

Apical / microvillar localization (the key family contrast)

This is the fact that distinguishes SLC10A2 from its sister SLC10A1 (NTCP), which is
basolateral in hepatocytes.

• PMID:9856990
• PMID:33222321
• [Reactome:R-HSA-194187 "In the body, ASBT is expressed on the apical surfaces of
  enterocytes, and this reaction is the first step in the process by which bile salts and
  acids are reaborbed from the intestinal lumen and returned to the liver"]
• The apical membrane of an ileal enterocyte is the microvillar (brush-border)
  membrane, which is why GO:0005902 microvillus and GO:0016324 apical plasma membrane
  are both annotated. The experimental grounding for both is on mouse Slc10a2
  (UniProtKB:P70172), which carries GO:0005902 and GO:0016324 by IDA (QuickGO);
  the human rows are Ensembl Compara transfers (GO_REF:0000107) from exactly that
  orthologue. No human-specific immunolocalization paper is present in the cache.

PAINT confirms the apical/basolateral split at the node level

Donor tracing of the IBA rows (SLC10A2-goa.tsv; donor ids resolved via MGI, RGD and
UniProt):

So PAINT places plasma-membrane residence and Na+/bile-acid symport at ancestral SLC10
nodes shared by both branches, but places the apical location at a sub-node whose
donors are exclusively ASBT orthologues. The mirror-image node in the SLC10A1 review is
PTN002570905, which carries GO:0016323 basolateral plasma membrane with NTCP donors.
The apical/basolateral contrast asserted in the SLC10A1 review is therefore confirmed
from the SLC10A2 side, and it is explicitly represented in the phylogeny rather than
being an inference.

Tissue distribution

• Ileum and kidney PMID:9458785; UniProt adds lower expression in cecum. The renal
  proximal-tubule role is a bile-acid salvage mechanism and is flagged "Probable" by
  UniProt.
• Also in biliary and colonic epithelium at lower levels PMID:19498215

Narrow substrate specificity (contrast with NTCP)

• PMID:9458785

This is the reason the SLC10A1 review's GO:0042908 xenobiotic transport replacement
term is not appropriate for SLC10A2: the same paper that justifies it for the liver
carrier explicitly denies the breadth for the ileal/renal one.

Human loss of function

• P290S abolishes transport without affecting synthesis or trafficking
  PMID:7592981 — a clean separation of transport activity from
  localization, and the model in PMID:15350125 rationalizes it as loss of bile-acid
  binding.
• Primary bile acid malabsorption (PBAM1, MIM:613291) is caused by SLC10A2 mutations
  PMID:9109432; the L243P and T262M alleles reach the plasma membrane but cannot
  transport PMID:9109432
• The disease phenotype is itself the enterohepatic-circulation phenotype
  PMID:9109432
• ASBT is a validated drug target; ASBT inhibitors are in clinical use/trials for
  cholestatic pruritus and metabolic disease (PMID:33222321).

Pathway position

• PMID:19498215
• [Reactome:R-HSA-159418 "This recycling involves a series of transport processes: uptake
  by enterocytes mediated by ASBT (SLC10A2), traversal of the enterocyte cytosol mediated
  by ileal bile acid binding protein (I-BABP - FABP6), efflux from enterocytes mediated
  by MRP3 (ABCC3)"]
• UniProt: works with NTCP, OST-alpha/beta and BSEP for enterohepatic recycling
  (PubMed:33222321).

Curation decisions and their reasoning

1. 12 GO:0005515 protein binding rows → REMOVE. Ten are HuRI yeast-two-hybrid
   (PMID:32296183), one is BioPlex 3.0 affinity-purification MS (PMID:33961781 — the
   Huttlin Cell 2021 dual proteome-scale network paper, not a focused study of
   SLC10A2), one is the SLC-superfamily interactome survey (PMID:40355756). None assigns
   a molecular function. Per repo policy generic protein binding is removed as
   uninformative rather than marked over-annotated, and removal does not dispute the
   interactions. The two partners that recur are UPK1B, EFNA5, IFITM3, CTXN3, TTMP,
   NRM, CCL4L2, TMEM222, PSENEN, TEX264 (HuRI) and CLPTM1 (O96005) twice, from two
   independent methods; CLPTM1 is a putative lipid scramblase of the ER/Golgi membrane
   and is a frequent companion of membrane proteins in AP-MS, so the reproducibility
   does not by itself license a functional term. Raised in suggested_questions.
2. Transport and localization rows → ACCEPT. GO:0008508, GO:0015721, GO:0005886,
   GO:0016324, GO:0005902, GO:0016020. The PAINT node placements are all sound (table
   above) and the apical sub-node is correctly ASBT-restricted. GO:0016020 is a broad
   IEA parent of correctly annotated children and is kept.
3. No MODIFY anywhere. Every term in GOA for this gene is at the right
   granularity. In particular GO:0008508 is not obsolete (checked via OLS), and
   GO:0015125 bile acid transmembrane transporter activity would be an ancestor of
   GO:0008508, so adding it would be redundancy, not coverage.
4. One NEW term: GO:0098719 sodium ion import across plasma membrane.
5. Participation. ASBT itself translocates the Na+ ions. This is not a
   substrate/necessity argument: the human protein was voltage-clamped and shown to
   carry charge with a 2:1 Na+:bile acid stoichiometry PMID:9856990, so the sodium
   leg of the symport is work the protein performs, in the same molecular event as the
   bile acid leg. Direction is import into the enterocyte, matching the term
   definition ("from outside of a cell, across the plasma membrane and into the
   cytosol", OLS).
6. Comparator check (QuickGO, 2026-10-04). Na+-coupled symporters conventionally
   carry a sodium-transport process term alongside their solute term: SLC5A1/SGLT1
   (P13866) has GO:0098719 by IDA; SLC34A1/NaPi-IIa (Q06495) has GO:0098719;
   SLC5A5/NIS (Q92911) has GO:0006814 sodium ion transport by IBA and GO:0035725 by
   IEA. So the term is conventionally applied to exactly this role. Within SLC10 it is
   absent from SLC10A2, SLC10A1 and mouse Slc10a2 alike — a family-level gap rather
   than a cross-family convention, which is the pattern that distinguishes a real
   omission from a convention I had not identified.
7. Non-redundancy. GO:0098719 is neither an ancestor nor a descendant of GO:0015721
   bile acid and bile salt transport.
8. GO-CAM check. gocams/index.tsv contains no model for SLC10A2 or Q12908, so no
   curator has already modelled this gene in a different role.
9. Chose the plasma-membrane-specific import term over the broader GO:0035725 because
   the location and direction are both experimentally established.
10. No NEW process term for bile acid reclamation / enterohepatic circulation.
    GO has no "enterohepatic circulation" term (OLS search returns nothing in GO), and
    GO:0015721 bile acid and bile salt transport already carries the transport claim at
    the right granularity. Inventing an intestinal-reabsorption process term was not
    pursued: the existing term plus the apical location already encode the reclamation
    step, and a more specific term would be a new-term request without a comparator set
    to anchor it. Raised in suggested_questions instead.

Open items

• No GO-CAM model for SLC10A2 in gocams/index.tsv at review time; the Reactome
  reaction R-HSA-194187 is the only pathway-level representation, and its sodium
  stoichiometry (1 Na+) disagrees with the experimental 2:1.
• The renal proximal-tubule salvage role is "Probable" in UniProt and rests on mRNA
  detection in kidney PMID:9458785 rather than on a transport assay in renal cells; no
  GO annotation asserts it and none was proposed.
• No human-specific microvillus/brush-border immunolocalization reference is cached; the
  localization rows rest on mouse IDA plus the human apical statements above.