SMIM26 review notes

Identity and why this gene is interesting

SMIM26 (HGNC:43430, UniProt A0A096LP01) is a 95-amino-acid microprotein. Its locus was
originally annotated as the long non-coding RNA LINC00493 (still the HGNC previous
symbol), so this is a reannotated lncRNA that turned out to encode a protein. UniProt
records a single predicted transmembrane helix (residues 13-35) and calls it a
single-pass mitochondrial outer membrane protein. Two isoforms are recorded
(A0A096LP01-1 and -2); isoform 2 simply lacks residue 40 (VSP_062288) and no
isoform-specific function has been reported, so none of the annotations below are
isoform-scoped.

Translation of the LINC00493 ORF and mitochondrial localisation of its product were
established first: PMID:33386847

Three papers, three interactomes

2023, EMBO Rep (ccRCC). SMIM26 binds AGK and SLC25A11:
PMID:37009826 and
PMID:37009826
Both partners are inner-membrane-associated: SLC25A11 is the mitochondrial
2-oxoglutarate/malate carrier of the inner membrane, and AGK (UniProt Q53H12) is a
lipid kinase (EC 2.7.1.94/2.7.1.107) that is also a subunit of the TIM22 inner-membrane
translocase. This matters for GOA: SMIM26 carries GO:0019901 protein kinase binding with
AGK as the WITH/FROM partner, but AGK is not a protein kinase.

2025, Mol Cell (one-carbon metabolism). A different interactome — sideroflexins and
the mitoribosome:
PMID:40578345 with the loss-of-function phenotype
PMID:40578345
This is the paper behind GOA's GO:0005743 mitochondrial inner membrane (IDA) and
GO:0140978 mitochondrial large ribosomal subunit binding (IDA).

2026, Genes Dev (respiratory chain). A third interactome, and an explicit topology:
PMID:41991342, PMID:41991342,
PMID:41991342, PMID:41991342, and the summary claim
PMID:41991342

SLC25A6 is ADP/ATP translocase 3 (ANT3), the inner-membrane adenine nucleotide carrier —
see this repository's genes/human/SLC25A6 review, which describes it as an
alternating-access ADP/ATP antiporter of the inner membrane. An OMM-anchored protein
reaching SLC25A6 must do so from the intermembrane-space face, which is exactly what the
Genes Dev topology asserts.

The tension GOA now carries

GOA holds, all as IDA/IPI on the same 95-residue protein:

• GO:0005741 mitochondrial outer membrane (three independent papers)
• GO:0005743 mitochondrial inner membrane (one paper)
• GO:0030674 protein-macromolecule adaptor activity (two papers)
• GO:0044325 transmembrane transporter binding (three papers, four different partners:
  SLC25A11, SFXN1, SLC25A6, VDAC1)
• GO:0140978 mitochondrial large ribosomal subunit binding (one paper)

The adaptor and transporter-binding calls are mutually reinforcing: every paper finds
SMIM26 bridging metabolite carriers, and the partner set differs between papers rather than
contradicting between them. A microprotein with one TM helix and ~60 residues of soluble
sequence is a plausible tether for several carriers, and different cell backgrounds and
baits will pull out different ones.

The mitoribosome claim is the one that does not sit comfortably. The mitochondrial large
ribosomal subunit is on the matrix face of the inner membrane. A single-pass protein
anchored in the outer membrane and spanning the intermembrane space cannot reach it. Either

1. there is a second, inner-membrane-integral pool of SMIM26 whose soluble portion faces the
   matrix (which would also explain the GO:0005743 IDA), or
2. the mitoribosome association is indirect — captured through SFXN1/2 or through the
   co-translational insertion machinery — and the direct-binding reading of GO:0140978 is
   too strong.

I cannot distinguish these from the abstracts available (both PMID:40578345 and
PMID:41991342 are abstract-only in this repository's cache). Per project rules an IDA is not
overruled from an abstract, so GO:0005743 and GO:0140978 are marked UNDECIDED rather
than removed or demoted, with the topological objection recorded. These two annotations
stand or fall together; they are the two claims of the inner-membrane/mitoribosome model.

Curation position taken

• GO:0030674 protein-macromolecule adaptor activity → core. Two independent IDA
  papers, and it is the one molecular description all three papers agree on.
• GO:0044325 transmembrane transporter binding → ACCEPT, core-adjacent. It is an
  informative binding term (unlike bare protein binding) and names the actual mechanism.
• GO:0005741 mitochondrial outer membrane → ACCEPT (core location, three papers).
• GO:0005739 mitochondrion → ACCEPT but less informative than the membrane terms.
• GO:0005743 mitochondrial inner membrane and GO:0140978 mitochondrial large ribosomal
  subunit binding → UNDECIDED (see above).
• GO:0019901 protein kinase binding → MODIFY to GO:0019900 kinase binding. The
  interaction with AGK is real, but AGK is a lipid kinase, not a protein kinase.
• GO:0005515 protein binding → MARK_AS_OVER_ANNOTATED per project guidance.

Proposed additions (reviewer proposals, single-source)

• GO:0005758 mitochondrial intermembrane space as a location, from the explicit
  IMS-spanning topology in PMID:41991342.
• No molecular function is proposed beyond the adaptor term. SMIM26 has no catalytic or
  transport activity of its own; what it does is hold carriers together.

Open questions

• Is there an inner-membrane pool of SMIM26, and if so what is its topology? This is the
  single question that would reconcile the two 2025/2026 models.
• Are the reported partner sets (AGK/SLC25A11; SFXN1/2 + mitoribosome; VDAC1/2 + SLC25A6)
  alternative states of one bridging complex, or cell-type-specific interactomes?
• What do the reported phosphorylation sites do? PMID:41991342 reports phosphorylation at
  distinct residues but the abstract does not assign a function to them.