DMS3-30 (A0SMP0) — Pseudomonas phage DMS3 gene 30 / AcrIE3

Provenance of this file

No deep-research provider keys exist in this container, so just deep-research was
not run and no -deep-research-<provider>.md file was created. The research below
is from the cached publications in publications/, PubMed/E-utilities, UniProt REST,
QuickGO and the RCSB PDB.

GONUTS (gowiki.tamu.edu), the CACAO curation wiki, was checked again on 2026-11-01.
It is no longer merely "temporarily offline while we perform upgrades" as recorded in
CACAO batches 1–2: it now returns HTTP 410 with the page "GONUTS, the GO wiki, has
been retired — The GO wiki that was at gowiki.tamu.edu is no longer online." The
Internet Archive has no snapshot of either
gowiki.tamu.edu/wiki/index.php/UniProtKB:A0SMP0 or .../PMID:24736222
(archive.org/wayback/available returns "archived_snapshots": {} for both). The
student-evidence check the CACAO project plan relies on is therefore permanently
unavailable, not deferred.

Identity

• UniProt A0SMP0 (A0SMP0_9CAUD), TrEMBL, unreviewed, 68 aa, PE 4: Predicted,
  DE SubName: Full=Uncharacterized protein {ECO:0000313|EMBL:ABG66684.1},
  GN ORFNames=DMS3-30. Genome reference is the DMS3 complete sequence
  (EMBL DQ631426, ABG66684.1) from PMID:18952788.
• Despite the "Uncharacterized protein" name, the entry already carries the family
  assignment: DR CDD; cd22551; AcrIE3; 1., DR InterPro; IPR049758; AcrIE3-like.,
  DR Pfam; PF28100; AcrIE3-like; 1. So this protein is AcrIE3, the type I-E
  anti-CRISPR of phage DMS3, originally named ACR3-30 ("3" = DMS3, "30" = gene 30).
• Independent confirmation that AcrIE3 == A0SMP0: both deposited AcrIE3 crystal
  structures have exactly this 68-residue sequence. PDB 8HEL chain A is
  MKITNDTTTYEVAELMGSEADELDGRIMMGLLSRECVVDTDDLSEDQWLALIDESQKVRREQFESDEALEHHHHHH
  (A0SMP0 plus a His tag) and is labelled "AcrIE3 | Pseudomonas phage DMS3
  (2913973)"; PDB 8ZEY chain A is the untagged 68-mer, identical to A0SMP0.
• Gene-symbol question: the stub emitted gene_symbol: A0SMP0, i.e. the accession.
  GOA's SYMBOL for this row is DMS3-30, UniProt's GN ORFNames is DMS3-30, and
  the folder is DMS3-30. The review therefore uses gene_symbol: DMS3-30. The
  community-standard protein name is AcrIE3 (used by CDD, Pfam and InterPro and by
  every paper since 2024), but it is not a gene symbol from any nomenclature
  authority, and UniProt has not adopted it for this entry — see the source defect
  below.

What the CACAO-cited paper actually shows for this gene

[PMID:24736222 Pawluk et al. 2014, mBio, full text cached] is the paper that
discovered the type I-E anti-CRISPRs. It assays four anti-CRISPR genes, each from
a different phage, and DMS3 gene 30 is one of them.

Two independent experiments in that paper are about this protein:

1. Gain of function, plasmid expression of the DMS3 gene (Fig. 3C). The gene was
   cloned out of DMS3 and expressed from pHERD30T in P. aeruginosa SMC4386, whose
   native type I-E system they had just shown to be active:
   PMID:24736222 and
   PMID:24736222. The DMS3 member of that set is
   ACR3-30 = gene 30 = this protein.
   Specificity controls in the same figure: five type I-F anti-CRISPR genes and
   JBD30-36 had no effect
   PMID:24736222.
2. Loss of function in the phage's own genome during lytic infection (Fig. 4).
   PMID:24736222;
   PMID:24736222; conclusion
   PMID:24736222.
3. Negative result, host specificity. The four type I-E Acrs did not inhibit the
   E. coli K-12 type I-E system even though all four were expressed
   PMID:24736222. They
   also do not inhibit the P. aeruginosa type I-F system
   PMID:24736222.

Is "DMS3m" a different phage from DMS3? No — it is an isogenic point mutant of DMS3
itself, made so that DMS3 would be CRISPR-targeted. From the cited companion paper
PMID:23242138. The lysogen panel in Fig. 3A used wild-type DMS3 itself
PMID:24736222. So every experiment
behind this annotation is on phage DMS3 or an isogenic derivative of it.

Mechanism (post-2024)

Two 2024 structural papers are about this exact protein (both abstract-only in the
cache; the PDB sequences tie them to A0SMP0, see above).

• [PMID:39541974 Koo et al. 2025, Structure; PDB 8HEL] "Our results reveal that
  AcrIE3 selectively binds to the Cas8e subunit of the Cascade complex." and
  "Through extensive mutational analyses, we show that AcrIE3 interacts with the
  protospacer adjacent motif (PAM) recognition site in Cas8e through its negatively
  charged surface residues." So AcrIE3 is a Cas8e-binding, acidic-surface inhibitor
  that occludes the PAM-recognition site of the type I-E surveillance complex —
  a DNA-mimicry-like competitive block of target recognition, upstream of Cas3
  recruitment.
• [PMID:38797150 Kim et al. 2024, BBRC; PDB 8ZEY] "Our structural examination showed
  that AcrIE3 adopts a helical bundle fold comprising four α-helices, with a notably
  extended loop at the N-terminus." and "surface analysis of AcrIE3 revealed the
  presence of three acidic regions, which potentially play a crucial role in the
  inhibitory function of this protein."

Note the two papers do not agree on the helix count/fold description in their
abstracts (Koo: "all-helical fold with a negatively charged surface"; Kim: "helical
bundle fold comprising four α-helices"), but they agree on the acidic surface, which
is the functionally load-bearing claim and is the one supported by mutagenesis in Koo
et al.

A related 2025 paper on the sibling AcrIE5 states the comparison explicitly
[PMID:41161091 — not cached, abstract only via E-utilities: "AcrIE5 may compete with
DNA binding to the PAM recognition site of type I-E Cascade, similar to AcrIE3 and
AcrIE4"]. Not cited in the review (not in publications/).

Ontology gap: no MF term for anti-CRISPR activity

A QuickGO ontology search for "CRISPR" returns only BP terms (GO:0099048, GO:0043571,
GO:0098672, and two obsoletes). There is no molecular-function term for inhibiting a
CRISPR-Cas surveillance complex. GO:0140721 nuclease inhibitor activity (the term
the AcrIIA5 review used for a Cas9 inhibitor) is a poor fit here: Cascade is not the
nuclease — Cas3 is — and AcrIE3 blocks target DNA recognition by Cas8e, not catalysis.
In GOA, GO:0140721 and its children are in practice used only for ribonuclease
inhibitors (RraA/RraB-type, InterPro and UniProt electronic rows). The review
therefore leaves core_functions.molecular_function unset and files a
proposed_new_terms entry, with proposed_molecular_function pointing at it.

The CACAO row, and the two screened-for shapes

GOA row under review (the gene's only annotation of any kind):

UniProtKB  A0SMP0  DMS3-30  acts_upstream_of_or_within_positive_effect
GO:0098672  symbiont-mediated suppression of host CRISPR-cas system
biological_process  ECO:0007040  IDA  PMID:24736222  taxon 2913973  CACAO  20250503

Shape (a), screen 3 — wrong qualifier: PRESENT, and systematic

acts_upstream_of_or_within_positive_effect says the gene product influences the
process from outside it, and positively. AcrIE3 is the effector of
GO:0098672: it binds host Cas8e and blocks the host surveillance complex. The right
qualifier is involved_in. This replicates the AcrIIA5 finding exactly (same
qualifier, same term, same assigning group).

It is not a one-off. A QuickGO pull of every annotation to GO:0098672 (18 rows,
2026-11-01) shows 9 of 18 carry acts_upstream_of_or_within_positive_effect, and
all 9 are CACAO; every involved_in row on the term is either CACAO or UniProt
keyword-derived. Affected accessions: A0A220GHA5 (AcrIIA5, already filed),
A0A247D711 (acrIIA4), A0A2D0TCG3, A0A425B3G2, A0A6B1YCA6, A0SMP0, B3G1L5, L7P7M1
(JBD30_035, also in batch 3), Q6TM71 (orf31).

The sharpest comparator is on the same paper: Q6TM71 (D3112 orf31 / ACR3112-31)
carries both the CACAO acts_upstream_of_or_within_positive_effect row from
PMID:24736222 and a UniProt involved_in row (GO_REF:0000043) for the identical
term and claim. The correct qualifier for this exact assertion is therefore already
present in GOA, on the sibling gene from the same figure of the same paper.

Shape (b), screen 2 — cross-isolate experimental transfer: ABSENT

The screen flagged PMID:24736222 because CACAO cites it on gene products from two
closely related Mu-like Pseudomonas phages (A0SMP0 from DMS3, Q6TM71 from D3112).
Reading the paper shows this is not the Gag shape. The paper identified four
distinct anti-CRISPR families and assayed one representative gene from each of
four different phages in the same figure — JBD88a gene 32, DMS3 gene 30, D3112 gene
31, JBD5 gene 34 — so each annotation rests on its own phage's gene. DMS3 gene 30 is
additionally assayed in its own genome in Fig. 4, and "DMS3m" there is an isogenic
point mutant of DMS3, not a sibling isolate. IDA is the correct evidence code on both
accessions; no ISS is warranted. The two-taxon span is a genuine comparative study,
which is the failure mode the screen's own design note anticipated ("taxon count alone
is a bad screen").

Source defect: UniProt has not characterized this entry

A0SMP0 is still DE SubName: Full=Uncharacterized protein {ECO:0000313|EMBL:ABG66684.1}
with PE 4: Predicted and no GO annotation of its own, twelve years after
PMID:24736222 established its function and two years after two crystal structures were
solved from it, while its own record already cross-references CDD AcrIE3, InterPro
IPR049758 AcrIE3-like and Pfam PF28100. The directly comparable sibling from the
same figure of the same paper, Q6TM71, is a reviewed Swiss-Prot entry with
RecName: Full=Anti-CRISPR protein 31, a FUNCTION line citing PubMed:24736222, and
three keyword-derived involved_in GO rows (GO:0039504, GO:0052031, GO:0098672).
A0SMP0 has none of these. Recorded as a MISSING_STATEMENT source defect against
UniProt (on the file:9CAUD/DMS3-30/DMS3-30-uniprot.txt reference's
reference_review.source_defects, which is where defects in cited curated record text
belong), requesting AcrIE3 / Anti-CRISPR protein 30 as the protein name,
PE 1: Evidence at protein level, and a FUNCTION line from PubMed:24736222 and
PubMed:39541974.

Terms considered and not proposed

• GO:0039504 symbiont-mediated suppression of host adaptive immune response and
  GO:0052031 symbiont-mediated perturbation of host defense response — these are
  what UniProt's keyword gives the sibling Q6TM71. GO:0052031 is an ancestor of
  GO:0098672 (checked via QuickGO ancestors), so adding it would be a parent-level
  restatement; GO:0039504 is off the GO:0098672 branch and describes adaptive immunity
  in the vertebrate sense. Neither is proposed as NEW.
• No cellular-component term is proposed. The protein acts inside the infected
  bacterial cell, and GO:0030430 host cell cytoplasm would be an inference with no
  localization assay behind it; it is recorded only as the core function's location,
  where the activity necessarily happens (host Cascade is cytoplasmic).

Outcome

One existing annotation row, action ACCEPT. Two source defects filed: the CACAO
qualifier (WRONG_QUALIFIER, resource CACAO, on the annotation's review) and the
uncharacterized UniProt entry (MISSING_STATEMENT, resource UNIPROT, on the
file:…-uniprot.txt reference's reference_review). One proposed_new_terms entry
("CRISPR-Cas surveillance complex inhibitor activity"), referenced from the core
function's proposed_molecular_function because no GO MF term fits. Three knowledge
gaps, five suggested questions, five suggested experiments. just validate 9CAUD DMS3-30 reports valid, 0 errors; the one substantive warning is that the core
function's location GO:0030430 host cell cytoplasm has no matching
existing_annotations row, which is deliberate — no paper localizes the protein, and
the location is an inference from where host Cascade is, so it is recorded only as the
site of the activity and not asserted as a CC annotation. (The AcrIIA5 review takes
the same position.)