Unreviewed TrEMBL entry, 86 aa, named only SubName: Full=Anti-CRISPR protein (AcrIIC1)
{ECO:0000313|PDB:5VGB} — the name comes from a PDB deposition, not from a curator.
PE 1: Evidence at protein level. Taxon NCBITaxon:487 (Neisseria meningitidis),
assigned to the bacterium rather than to a phage because the gene lies in a resident
prophage / mobile element of N. meningitidis.
The accession is the assayed protein. PDB; 5VGB; X-ray; 1.50 A; B=1-86 is the
NmeCas9 HNH domain bound to AcrIIC1, and RCSB gives its primary citation as PMID:28844692
"A Broad-Spectrum Inhibitor of CRISPR-Cas9" with structure title "Crystal structure of
NmeCas9 HNH domain bound to anti-CRISPR AcrIIC1". Three further structures of this
sequence exist (7X31 NMR, 7X4B 1.61 Å, 8IF0 1.57 Å), 7X31/7X4B from PMID:36400778. CDD
cd22213 AcrIIC1. So experimental evidence codes are appropriate on this accession.
Like L7P7R7, this is an instance of the curation gap recorded in
modules/anti_crispr_suppression.yaml: four deposited structures, a mechanism established
in Cell, and a UniProt name that says only "Anti-CRISPR protein".
Pawluk et al. found the AcrIIC family by looking for Cas9 inhibitors in Neisseria
PMID:27984730,
PMID:27984730.
This is the reference GOA (CACAO) cites for the IDA row.
The module (annoton acriic1_hnh_blocker) places AcrIIC1 in
cas9_hnh_blockade_variant, describing it as binding the HNH catalytic domain and
permitting target binding while preventing cleavage, with breadth across orthologues
because the HNH fold is conserved. Every element of that is directly supported.
Harrington et al. establish, in order:
An additional layer not in the module: AcrIIC1 activity is redox-regulated through an
intersubunit disulphide (Cys17–Cys80, recorded in the UniProt FT DISULFID lines from
7X4B), with the dimer being the inactive form
PMID:36400778,
PMID:36400778.
This is interesting but is not annotated: the physiological relevance of a redox
switch was demonstrated in the context of eukaryotic cellular environments for
gene-editing purposes, and whether the switch operates in the native Neisseria
cytoplasm is not established. It is recorded as a knowledge gap and a question instead.
The module deliberately asserts no GO id here, saying "No GO id asserted: the activity is
occlusion of a catalytic domain within a multidomain nuclease, leaving DNA binding intact,
and no GO molecular-function term captures that."
I disagree in part. GO:0140721 nuclease inhibitor activity is defined as "Binds to
and modulates the activity of a nuclease", and that is exactly and literally what AcrIIC1
does — it binds Cas9, a nuclease, at the active-site interface of one of its two nuclease
domains, and abolishes cleavage. The term is less specific than the mechanism (it does
not say that one domain of a multidomain nuclease is occluded while the other activities,
including DNA binding, survive), but it is not wrong, and GO curation practice is to
annotate the nearest correct existing term and request the specific one, rather than to
assert nothing. Asserting nothing leaves this protein with no molecular function at all,
which is strictly less informative.
So: GO:0140721 is asserted as the core molecular function, and no new term is proposed
for it in this review. The gap — that GO cannot express single-domain occlusion within a
multidomain enzyme, and therefore cannot distinguish AcrIIC1 from an inhibitor that
disables the whole enzyme — is recorded as an ONTOLOGY knowledge gap with the specific
observation that the diagnostic feature (DNA binding survives) is the part GO loses.
I deliberately do not propose a term for it. Writing a term for "occlusion of one
catalytic domain within a multidomain nuclease" would be a term about protein architecture
rather than about activity, and GO molecular-function terms are about what the gene
product does; the better request is probably an extension on the existing term naming the
occluded domain, which is a curation-mechanism question rather than a new class. Recorded
as a question for ontology editors.
I also do not assert GO:0043021 ribonucleoprotein complex binding here, unlike the
other reviews in this round. AcrIIC1 binds the isolated HNH domain in solution — that is
how the chimera and size-exclusion experiments were done, and the crystal structure 5VGB
is of the free HNH domain, not of a Cas9-guide complex. So guide dependence, which is the
basis for the GO:0043021 assertions on AcrIIA4, AcrIIA2 and AcrF8, is not established for
AcrIIC1 and is in fact contraindicated.
One row: acts_upstream_of_or_within_positive_effect GO:0098672, IDA, PMID:27984730,
assigned by CACAO. Same adjudication as for AcrIIA4: the term is correct and maximally
specific, the evidence is on this sequence, so ACCEPT. The qualifier is poorly chosen —
AcrIIC1 does not act upstream of the suppression, its binding to the Cas9 HNH domain is
the suppression, so involved_in is better supported; and the positive_effect component
reads as though the protein up-regulated something. Flagged in review.reason for CACAO
rather than rewritten, since a qualifier change is not a term change.
PMID:28844692 establishes HNH active-site engagement and DNA-bound inactive Cas9; scope is the tested susceptible orthologues. PMID:36400778 supports a redox-sensitive monomer/dimer transition, but native Neisseria regulation is not established. PDB 5VGB maps the HNH-binding structure to this accession. The bacterial source taxon is retained despite mobile-element origin.
Existing GOA assertions were preserved. Earlier reviewer-authored NEW proposals were reassessed independently. Primary references were obtained through the normal publication cache. A normal Falcon attempt with perplexity-lite fallback failed; logs are in .cache/module-audit-logs/anti-crispr-deep-NEIME-acrIIC1.log. These are manual research notes, not provider-generated research.