MICU1 (Q9BPX6) — review notes

Journal for the AI review of human MICU1 (mitochondrial calcium uptake 1; originally CBARA1,
"calcium-binding atopy-related autoantigen 1"). 114 GOA rows, 26 distinct GO terms.

1. The standing model

MICU1 is an EF-hand Ca2+-binding protein that sits on the intermembrane-space face of the inner
mitochondrial membrane, where it associates with the MCU/EMRE pore to form the uniporter
holocomplex (uniplex) and gates Ca2+ entry into the matrix.

Founding identification by integrative genomics
PMID:20693986
and already then assigned a sensing role
PMID:20693986.
Gatekeeping (threshold setting)
PMID:23101630,
PMID:23101630.
Heterodimer as the Ca2+ switch
PMID:28615291,
with MICU1 and MICU2 acting in opposite directions
PMID:24560927.
Confirmed structurally in the holocomplex
PMID:32494073.
Tissue-specific dimer composition is real: MICU1 homodimers operate in skeletal muscle and kidney
PMID:36206740.

An internal dispute that GOA already carries

Occlusion vs potentiation. GOA holds three IDA annotations to GO:0019855 calcium channel inhibitor
activity (PMID:32494073, PMID:37036971, PMID:37126688), and the 2023 pair were written explicitly
to settle a "direct clash"
PMID:37036971,
PMID:37126688.
Both sides agree MICU1 sets the threshold; they disagree on whether the low-Ca2+ state is physical
occlusion or allosteric potentiation. The occlusion side now has direct patch-clamp support, so the
inhibitor-activity annotations are kept.

2. MCU-independent functions — the part that is NOT new

MICU1 already has two published uniporter-independent roles, both in GOA:

Cristae junction stabilisation and spatial anchoring of the uniplex at the inner boundary membrane
PMID:31427612,
PMID:31427612.
MICOS-complex formation, explicitly independent of the uniporter
PMID:37098122,
PMID:37098122.

There is also a hard genetic argument that predates all of this: Micu1 null lethality is not rescued
by removing the uniporter
PMID:42129466.

So "MICU1 does things without MCU" is established. The 2026 claim is a different and larger one.

3. The 2026 metabolon claim — what it actually says

Cohen et al., Nat Metab 2026 (PMID:42129466), from the Elrod lab (the same lab as the 2023 MICOS
paper).

What it demonstrates:
PMID:42129466
MICU1-specific evidence includes a co-IP with the relevant dehydrogenase
PMID:42129466.

What it proposes:
PMID:42129466
Note the hedging in the body: "Our results suggest that this may be the primary physiological
calcium signaling mechanism regulating cellular energetics."

What it does NOT say

It does not claim MICU1 has stopped being a gatekeeper. The paper's own framing keeps that intact:
PMID:42129466
The model it displaces is the downstream dogma that matrix [Ca2+], delivered through the uniporter,
is what activates the matrix dehydrogenases and thereby tunes basal energetics. That dogma is not
itself a MICU1 GO annotation. So the mapping from this paper onto MICU1's annotation set is narrower
than the abstract's rhetoric suggests:

Claim in PMID:42129466	Effect on MICU1 GO annotations
MICU proteins scaffold a Ca2+-sensitive GPD2/SDH metabolon	New function; no existing GO term captures it
MICU heterodimerisation is mtCU-independent	Supports existing GO:0046982, no change
MICU1 still gates mtCU	Supports GO:0019855, GO:0061891, GO:1990246, GO:0036444
Matrix Ca2+ is not the master regulator of basal metabolism	Not a MICU1 annotation; nothing to retract

Scepticism applied

One paper, one lab, no independent replication and no published rebuttal. I checked: a PubMed
search for MICU metabolon returns exactly two records, PMID:42129466 and PMID:40678212 — and the
second is the Research Square preprint of the same study, not an independent confirmation. A search
for MICU1 GPD2 returns nothing else at all. Meanwhile the field has gone on using the gatekeeper
model after publication, e.g. an August 2026 Exp Mol Med study describes the uniporter as one that
PMID:42557311.
That is not a rebuttal either, but it does show the standing model is still in working use. That asymmetry means the claim is untested, not that it is
accepted. The primacy assertion ("the primary physiological calcium signaling mechanism") is a
proposal about relative importance, which is not the sort of thing a GO annotation should encode
even if it were settled. Accordingly:

No existing annotation is downgraded, removed or modified on the basis of this paper.
The metabolon scaffolding function is recorded as a proposed_new_terms candidate and in
suggested_questions, not folded into core_functions.
The gatekeeper function stays core.

If replication follows, the right GO representation is probably an MF in the enzyme-regulator branch
(regulation of GPD2/SDH activity) plus a part_of a named metabolon complex, neither of which exists
yet.

4. Legacy and mis-mapped annotations

GO:0006952 defense response, TAS, PMID:9806765. MICU1/CBARA1 was one of four cDNAs pulled from
an expression library screened with IgE from atopic dermatitis patients
PMID:9806765.
Being recognised by autoreactive IgE is a property of the patients' immune systems, not a function
of MICU1. Nothing in that paper assays a defense response performed by the protein. REMOVE.
UniProt already records this correctly, and in the right slot: CC -!- ALLERGEN: Causes an allergic reaction in human. Binds to IgE from atopic dermatitis (AD) patients. (MICU1-uniprot.txt line 690).
An allergen/autoantigen statement is the right representation; "defense response" is not.
GO:1900069 regulation of cellular hyperosmotic salinity response, IMP, PMID:26975899. MICU1 was
one of five hits in a targeted siRNA screen whose readout was mitoflash frequency
PMID:26975899.
Hyperosmotic stress was the stimulus used to evoke mitoflashes; the paper is about EFHD1, and
"salinity" is not what was manipulated. Over-annotation rather than a wrong gene — kept as such,
since it is an experimental annotation whose full text I did read but whose term mapping
over-reaches.
GO:0072732 cellular response to calcium ion starvation, IDA, PMID:32494073. This appears to be a
reading of the Ca2+-free ("inhibited") cryo-EM state as a cellular response to calcium starvation.
A conformational state of a purified complex in a low-Ca2+ buffer is not a cellular response to
calcium ion starvation. MARK_AS_OVER_ANNOTATED.
GO:0031966 mitochondrial membrane and GO:0070509 calcium ion import and GO:0034704 calcium
channel complex are each the correct-but-general parent of a specific term the gene already
carries (GO:0005743, GO:0036444, GO:1990246 respectively). MODIFY with the specific replacement.
GO:0005515 protein binding × 13. All IPI; no functional content. MARK_AS_OVER_ANNOTATED per
project guidance. The interactors themselves (MCU, EMRE, MICU2, MCUR1, SLC25A23, MICU3, UCP2/3)
are better captured by the complex and heterodimerisation terms that the gene already has.

5. Decisions

Disease context (not itself annotated, but it constrains how essential the gene is): biallelic
loss-of-function causes myopathy with extrapyramidal signs, MPXPS, MIM:615673 — early-onset proximal
muscle weakness, raised creatine kinase, learning difficulties and progressive involuntary movement
(UniProt DISEASE block, MICU1-uniprot.txt lines 671–685).

ACCEPT: GO:0005509, GO:0005739, GO:0005743, GO:0005758, GO:0006851, GO:0019855, GO:0036444,
GO:0044284, GO:0046982, GO:0051560, GO:0051561, GO:0061891, GO:1903852, GO:1990246
KEEP_AS_NON_CORE: GO:0042802, GO:0051260, GO:0071277
MODIFY: GO:0031966 → GO:0005743; GO:0034704 → GO:1990246; GO:0070509 → GO:0036444
MARK_AS_OVER_ANNOTATED: GO:0005515, GO:0072732, GO:1900069
REMOVE: GO:0006952