Unreviewed TrEMBL entry, 123 aa, RecName: Full=Anti-CRISPR protein AcrIIA2
{ECO:0008006|Google:ProtNLM} — i.e. the name is a machine prediction (ProtNLM), not a
curator assertion. PE 4: Predicted, no PDB, no GO annotations in GOA at all. The only
non-electronic support for the AcrIIA2 assignment is NCBIfam NF033945 AcrIIA2_fam and
CDD cd22261 AcrIIA2, both HMM family hits. Seeded taxon was wrong
(NCBITaxon:10663); UniProt gives OX NCBI_TaxID=1458856 for Listeria phage LP-101, and
the review has been corrected to that.
The published AcrIIA2 work was done on a different, closely related sequence. Rauch et
al. name their reference sequence explicitly
PMID:28041849.
NCBI efetch of AEO04363.1 returns gp29 [Listeria monocytogenes J0161], 123 aa:
MTLTRAQKKYAEAMHEFINMVDDFEESTPDFAKEVLHDSDYVVITKNEKYAVALCSLSTDECEYDTNLYL
DEKLVDYSTVDVNGVTYYINIVETNDIDDLEIATDEDEMKSGNQEIILKSELK
Pairwise comparison against A0A059T5F6 (same length, no gaps): 114/123 identical =
92.7%, differing only at positions 3, 5, 8, 11, 39, 44, 56, 81, 112. All nine are
conservative or surface substitutions; none is in a position the structural paper
identifies as a Cas9 contact. So this is a genuine orthologue of the characterised
protein, encoded by a free Listeria phage rather than by a L. monocytogenes prophage.
Consequence for evidence codes. The functional claims for this accession are
transferred by sequence similarity, not demonstrated on it. Every annotation proposed here
therefore carries ISS, with AEO04363.1 named as the source. The alternative — writing
IDA because the family is well characterised — would assert an experiment that was not
done on this sequence, and would be exactly the kind of accession-level slippage that
modules/anti_crispr_suppression.yaml records as an open gap for the AcrIII-1 exemplar.
The module (annoton acriia2_dna_mimic) places AcrIIA2 in the
cas9_dna_mimicry_variant as "the convergence case", i.e. a second, sequence-unrelated
Listeria inhibitor that attacks the same Cas9 surface as AcrIIA4. The assignment is
correct, and the structural paper states it in almost those words.
Liu et al. solved the AcrIIA2–SpyCas9–sgRNA complex at 3.3 Å
PMID:30606466,
identified the occluded residues
PMID:30606466,
and drew the mimicry conclusion together with the explicit statement that the binding
motif differs from AcrIIA4's
PMID:30606466.
That last clause is the whole point of the module's convergence claim: same enzyme, same
surface, same mimicry strategy, different binding motif and no sequence relationship.
Cached record is abstract-only, so these are abstract quotes and are marked as such.
AcrIIA2's guide dependence is also on record
PMID:28448066,
as is the in vivo inhibition of both Lmo Cas9 and SpyCas9
PMID:28448066,
PMID:28041849.
One mechanistic subtlety from the genetics is worth keeping: AcrIIA2 was more effective in
the cleavage assay than in the dCas9 binding assay, which the authors read as partial
inhibition of binding plus inhibition of cleavage
PMID:28041849.
Pure PAM-site occlusion predicts equal effects on binding and cleavage, so this is a hint
that AcrIIA2 may be doing something at the catalytic step as well. It is a single
qualitative observation in a heterologous system and is recorded as a knowledge gap rather
than annotated.
Breadth across orthologues is attributed to a conserved target surface
PMID:28041849.
GOA holds zero annotations for A0A059T5F6. Combined with an unreviewed entry whose
protein name is a ProtNLM prediction, this is a concrete instance of the second knowledge
gap recorded in modules/anti_crispr_suppression.yaml ("Most characterised Acr proteins
have no reviewed UniProt entry, and several carry uninformative names"). Here the name is
not uninformative but is machine-generated, which is arguably worse: it reads like a
curator assertion and is not one. The family HMM hits (NF033945, cd22261) do support the
assignment, and the 92.7% identity to the assayed protein supports it independently.
Same gap as AcrIIA4 and AcrF2: no GO molecular-function term for mimicry-based competitive
inhibition. The proposed term nucleic acid mimicry-based competitive inhibitor activity
is restated here with AcrIIA2-specific supporting text, since the three reviews should be
able to cite one proposal independently.
The record is viral LP101_024 (tax1458856), not the directly assayed J0161 prophage protein AEO04363.1. Existing alignment notes report 114/123 identical residues. This supports ISS-level transfer, not IDA or an assertion that every substitution is functionally neutral. PMID:30606466 supports the characterized family mechanism; PMID:28448066 establishes guide-dependent binding of the studied homologue. Descriptions and core function now carry that accession-level boundary.
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-9CAUD-acrIIA2.log. These are manual research notes, not provider-generated research.