CLYBL (Q8N0X4) — curation journal

Human CLYBL, "citrate lyase beta-like protein". Nuclear-encoded, mitochondrial matrix,
homotrimeric Mg2+-dependent enzyme of the HpcH/HpaI aldolase family (citrate lyase beta
subunit-like subfamily). Humans lack the other citrate-lyase subunits, so CLYBL is not part
of an ATP-independent citrate lyase (UniProt CAUTION on Q8N0X4).

Ubiquitously expressed:
PMID:24334609

A common premature-stop polymorphism (rs41281112, p.Arg259*) ablates the protein in a few
percent of people and is the gene's only consistent human phenotype — reduced circulating
vitamin B12:
PMID:24334609
(UniProt gives the carrier frequency of the null protein as 2.7%.)

Successive function assignments

1. Malate / beta-methylmalate synthase (Strittmatter 2014)

PMID:24334609
This is the Claisen-condensation direction. Kinetics are poor: kcat 0.12 s-1, KM(glyoxylate)
3.6 mM (UniProt Q8N0X4 BIOPHYSICOCHEMICAL PROPERTIES, from this paper). A millimolar KM for
glyoxylate in a mitochondrion is not a physiological operating point.

2. (S)-Citramalyl-CoA lyase; the itaconate model (Shen 2017)

PMID:29056341
PMID:29056341
PMID:29056341
PMID:29056341

This paper already settles the malate-synthase question against the 2014 assignment:
PMID:29056341

and it already saw the thioesterase that would later be named the physiological reaction:
PMID:29056341

3. Malyl-CoA metabolite repair (Griffith 2025)

PMID:40108300
PMID:40108300
PMID:40108300
PMID:40108300

Reaction (UniProt CATALYTIC ACTIVITY, RHEA:38291, EC 3.1.2.30):
(S)-malyl-CoA + H2O = (S)-malate + CoA + H+, PhysiologicalDirection left-to-right.
Kinetics from Griffith 2025 via UniProt: KM 11 uM, kcat 9.4 s-1, Vmax 15.1 umol/min/mg —
comparable to the citramalyl-CoA lyase reaction (KM 22 uM, kcat 1.6 s-1 in the same paper)
and ~100x better than malate synthase.

The argument, and where the B12 variant fits

The itaconate/citramalyl-CoA model and the malyl-CoA model are not mutually exclusive
chemistry — both are Mg2+-dependent reactions on a C4/C5 acyl-CoA at the same active site,
and both converge on protecting MMUT's adenosylcobalamin. They differ on which one explains
the human phenotype.

The decisive observation is where the two substrates come from. Itaconate is made by
ACOD1/IRG1 in activated macrophages — inducible and spatially restricted. Malyl-CoA arises
continuously as a side product of promiscuous TCA-cycle enzymes in every cell. CLYBL is
expressed in every tissue, and the rs41281112 null allele lowers circulating B12 in
unselected, non-inflamed populations. A B12 phenotype that tracks the genotype rather than
inflammatory state fits the constitutive malyl-CoA source, not the inducible itaconate one.
Griffith's KO cells make the point directly: they accumulate malyl-CoA but not itaconyl-CoA,
and still lose adenosylcobalamin.

So the B12 variant does bear on the question, and it argues for malyl-CoA repair as the
housekeeping physiological reaction, with citramalyl-CoA lyase as the same active site doing
real and important work in a specific inflammatory context.

Position taken in this review

• Core MF 1: (S)-malyl-CoA thioesterase. No dedicated GO term exists; the closest is
  GO:0016289 acyl-CoA hydrolase activity, which GOA already carries as IDA from
  PMID:40108300 and as an IEA from the Rhea mapping. Accepted as core; a specific child term
  (S)-malyl-CoA hydrolase activity is proposed.
• Core MF 2: GO:0047777 (S)-citramalyl-CoA lyase activity. Accepted as core. Both
  activities serve the same biological process, GO:0110052 toxic metabolite repair.
• GO:0004474 malate synthase activity → MODIFY. The reaction is genuinely catalysed in
  vitro and the IDA annotations are correct as measurements; but the specificity constant is

  1000-fold below the lyase reaction and the KM for glyoxylate is millimolar, so this is a
  promiscuous reverse activity of the lyase/thioesterase active site, not the enzyme's
  function. Replacement proposed: GO:0016289 acyl-CoA hydrolase activity /
  GO:0047777. Note also that human mitochondria have no glyoxylate shunt, so a malate
  synthase has no pathway to belong to.

• GO:0106121 positive regulation of cobalamin catabolic process → MODIFY, direction
  appears inverted. CLYBL removes the CoA esters that destroy adenosylcobalamin; loss of
  CLYBL is what increases B12 destruction and lowers circulating B12. The gene product
  therefore reduces cobalamin catabolism. GO:0106122 negative regulation of cobalamin catabolic process exists and currently has zero annotations in QuickGO, while GO:0106121
  is annotated essentially only to CLYBL and its orthologues — consistent with the term pair
  having been created for this gene and the wrong sign having been picked. Flagged for a GO
  curator rather than asserted as settled; recorded in suggested_questions.
• GO:0106064 regulation of cobalamin catabolic process (IBA, IEA) → ACCEPT. The parent
  is directionally neutral and therefore not wrong. Not modified, because the IBA was placed
  by a PAINT curator whose tree I have not inspected.
• GO:0000287 magnesium ion binding (two IDAs) — accepted, genuine catalytic cofactor.
• GO:0070207 protein homotrimerization — accepted as non-core (real, from the crystal
  structure, but an assembly property).

Open questions

• Is GO:0106121 the intended direction? If yes, what is the reasoning?
• What is the source of mitochondrial malyl-CoA in vivo (which TCA enzyme, at what flux)?
• Does the rs41281112 null allele show any B12 phenotype modulation by inflammatory state,
  which would be the direct test of itaconate vs malyl-CoA as the dominant driver?
• Should EC 2.3.3.9 (malate synthase) be retained on the UniProt entry at all?

2026-10-18 ORGANELLES pilot: mitochondrial sub-compartment

Applied the ORGANELLES policy (projects/ORGANELLES.md, "Policy") to the three GO:0005739
mitochondrion rows. Lines of evidence for the matrix, all agreeing:

1. Targeting: UniProt annotates a mitochondrial transit peptide, residues 1-22 (ECO:0000255,
   predicted), with the mature chain at 23-340, and the entry has no TRANSMEM/INTRAMEM feature
   [file:human/CLYBL/CLYBL-uniprot.txt "FT TRANSIT 1..22"]. Cleavable presequence plus
   no membrane anchor is the matrix signature.
2. Structure: the mature, MTS-less chain is a soluble Mg(2+)-bound homotrimer
   PMID:29056341.
3. Literature statement (secondary, citing MitoCarta2.0/Calvo 2016)
   PMID:29056341.
4. Function: substrates are matrix CoA esters - malyl-CoA from citric acid cycle side reactions
   PMID:40108300 and itaconate-derived
   citramalyl-CoA - and the protected B12 enzyme MUT is a matrix enzyme
   PMID:29056341; CLYBL loss causes a
   cell-autonomous defect in mitochondrial B12 metabolism PMID:29056341.

Nothing argues for another sub-compartment (no TM segment, no IMS/outer-membrane report). No
direct sub-mitochondrial fractionation / protease-protection experiment was found (PubMed
search "CLYBL mitochondrial matrix" returned nothing), so the new row is IC, not IDA.

Changes: IEA GO_REF:0000044 mitochondrion ACCEPT -> MODIFY (to GO:0005759); IDA (PMID:24334609)
and HTP (PMID:34800366) mitochondrion rows left ACCEPT (organelle-level assays); NEW IC row
GO:0005759 mitochondrial matrix; core_functions locations mitochondrion -> mitochondrial matrix.

2026-10-18 (ORGANELLES revision)

After the PR #5208 review, the IBA/IEA mitochondrion rows are kept (not MODIFYed to the matrix): their sources correctly assert the organelle and could not resolve further. The mitochondrial matrix stays as the NEW IC row, added alongside the root rows as GO does for DARS2 and MMUT.