SLC8B1 / NCLX (Q6J4K2) — review notes

1. What the protein is

SLC8B1 encodes NCLX, a 584-residue, ~13-TM member of the Ca2+/cation antiporter (CaCA) superfamily,
resident in the mitochondrial inner membrane and enriched in cristae. It is the long-accepted route
for Ca2+ efflux from the mitochondrial matrix.

• Localisation and the founding functional identification:
  PMID:20018762
  PMID:20018762
  PMID:20018762
• Independent replication that NCLX (not LETM1) carries matrix Ca2+ extrusion:
  PMID:24898248
• Physiological consequences in beta cells:
  PMID:23056385
• Coupling to store-operated Ca2+ entry:
  PMID:28219928

The Ca2+ efflux role is about as well replicated as anything in mitochondrial Ca2+ biology. The
part that is now in question is the counterion.

2. The counterion problem

An anomaly has been on the record since NCLX was first cloned: unlike canonical NCX proteins, it
lacks several of the serines that coordinate Na+, and it uniquely accepts Li+ (PMID:15060069,
"Lithium-calcium exchange is mediated by a distinct potassium-independent sodium-calcium exchanger",
the paper behind the GO:0005432 IDA of 2004). Two 2025/2026 cryo-EM studies turned this anomaly
into an explicit challenge.

Fan et al., Nature 2025 (PMID:40931067) — H+/Ca2+, not Na+/Ca2+

PMID:40931067
PMID:40931067
PMID:40931067
PMID:40931067

They also report the most direct challenge to the functional literature — mito-NCX survives loss of
NCLX, in four independently derived KO lines including two from another laboratory:
PMID:40931067

Zhang et al., Nat Commun 2026 (PMID:42431881) — broader selectivity, not H+-specificity

Same structural premise, different functional conclusion:
PMID:42431881
PMID:42431881

Crucially, this study does not refute Na+ as a counterion — it demotes it from the counterion
to one of several. And it explicitly declines to adjudicate the TMEM65 question:
PMID:42431881

Counter-evidence that should not be forgotten

Before the structures, mutagenesis mapped distinct Na+- and Li+-selective residues in the NCLX
transport site — i.e. positive functional evidence that NCLX does have counterion-specific sites:
PMID:28130126
PMID:28130126

And the Li+ anomaly itself was the founding observation:
PMID:15060069

The two structure papers also flatly contradict each other on the counterion question. Fan et al.
find monovalent cations do nothing:
PMID:40931067
while Zhang et al. find Na+, K+ and Li+ all act as counterions. Both used heterologous expression;
they disagree about the result, not merely about its interpretation.

What the two structures agree on

1. NCLX is a bona fide Ca2+ exchanger with a conserved central Ca2+ site.
2. It lacks the canonical Na+-coordinating residues.
3. It assembles as a trimer, not a dimer:
   PMID:40931067
   PMID:42431881

Point 3 is a clean, independently replicated contradiction of the GO:0042803 protein homodimerization activity annotation (IPI, PMID:20018762). Self-association is right; the
stoichiometry is not.

3. Provenance caveat — this is not three independent votes

PMID:40931067 (NCLX = H+/Ca2+ exchanger) and PMID:40691517 (TMEM65 = mito-NCX) share senior authors
(Feng L., Tsai M.-F.). They form a single, internally coherent model in which TMEM65 does mito-NCX
and NCLX does something else. The independent structural study (PMID:42431881) reproduces the
structural observation but reaches a different functional conclusion and keeps Na+ in the list of
counterions. So the state of the evidence is:

• Na+-exclusivity: challenged by two structures, one of which nonetheless still finds Na+ works.
• Ca2+ exchange by NCLX: unchallenged by anyone.
• NCLX as the essential mediator of mitochondrial Ca2+ efflux: challenged by one lab's KO data,
  against ~15 years of loss-of-function work from several labs.

4. Curation decisions

GO:0005432 calcium:sodium antiporter activity → MODIFY to GO:0015368 calcium:monoatomic cation antiporter activity (its direct is_a parent; verified via QuickGO, and already annotated to this
gene by Reactome TAS). Rationale:

• This is not a removal. Every experimental annotation is retained, just at the term the current
  evidence actually supports. The brief's rule against discarding experimental annotations on the
  strength of one structure paper is respected.
• The over-specific part of GO:0005432 is exactly the part under challenge — the identity of the
  counterion as Na+. The generalised parent is compatible with all three readings (Na+/Ca2+,
  H+/Ca2+, and non-selective cation/Ca2+) and with the Li+ exchange that has been on the record
  since 2004.
• The supporting argument is not a single paper: it is a structural fact reported independently
  twice, plus a long-standing sequence anomaly, plus NCLX's anomalous Li+ tolerance.
• All seven GO:0005432 rows (IBA, Ensembl IEA, IMP, IDA ×4, Reactome TAS) take the same action, per
  the same-term-same-action rule.

GO:0086038 calcium:sodium antiporter activity involved in regulation of cardiac muscle cell membrane potential → MARK_AS_OVER_ANNOTATED (IEA/ISS only). The term's definition ties the
exchange to "regulating the membrane potential of a cardiac muscle cell" — a sarcolemmal NCX1
(SLC8A1) role. A cristae-resident exchanger does not set the cardiomyocyte membrane potential
directly; this is an orthology projection that carries the Na+ claim and a wrong membrane.

GO:0042803 protein homodimerization activity → MODIFY to GO:0042802 identical protein binding (verified direct is_a parent). Self-association is real (co-IP, PMID:20018762; both
cryo-EM structures), the dimer stoichiometry is contradicted by both structures.

GO:0035725 sodium ion transmembrane transport → MARK_AS_OVER_ANNOTATED. Derived
automatically from GO:0005432 by inter-ontology inference (GO_REF:0000108); it inherits the
counterion claim that is under challenge, and no direct evidence for NCLX-mediated Na+ flux exists
independent of it.

Plasma membrane / sarcolemma (GO:0005886, GO:0042383) → MARK_AS_OVER_ANNOTATED. Residue of
the pre-2010 literature, when NCKX6/NCLX was characterised at the plasma membrane; Reactome still
models a plasma-membrane Li+/Ca2+ exchanger (R-HSA-425822). Everything since PMID:20018762 places
NCLX in the IMM/cristae.

Everything in the Ca2+-efflux cluster is accepted as core: GO:0099093, GO:0006851,
GO:0051560, GO:0006874, GO:0051480, GO:0070588, GO:0015368, and the IMM/crista locations.
Tissue-specific downstream processes (insulin secretion, glucose homeostasis, SOCE regulation,
lymphocyte chemotaxis, cardiac membrane potential) are kept as non-core.

5. Open questions recorded in the review

1. What is the physiological counterion in situ — Na+, H+, or several?
2. Why does mito-NCX persist in NCLX KO cells (PMID:40931067) when NCLX knockdown/knockout reduces
   mitochondrial Ca2+ efflux in so many other systems?
3. Does the trimeric assembly imply cooperativity, and what is the exchange stoichiometry?
4. What is the physical relationship between NCLX and TMEM65 (see TMEM65-notes.md)?