Curation journal for the CACAO annotation audit (see projects/CACAO.md).
Deep research was unavailable for this review. No deep-research API keys are
configured in this container, so just deep-research was not run and there is no
doc-deep-research-<provider>.md file. The literature below was assembled by hand from
the cached publications in publications/, the UniProt flat file
(genes/BPP1/doc/doc-uniprot.txt), PubMed, and QuickGO.
GONUTS is offline. https://gowiki.tamu.edu/wiki/index.php/PMID:11222604 returns only
"This site is temporarily offline while we perform upgrades", confirming the status
recorded in projects/CACAO.md as of 2026-10-29. The student-recorded evidence behind
the four CACAO rows could therefore not be inspected, and the abstract-only rows
(PMID:11222604, PMID:18398006) were judged from the published abstracts plus the later
mechanistic literature.
Doc is a 126-residue protein encoded by the second gene of the phd/doc operon of
bacteriophage P1, which lysogenises E. coli as a single-copy plasmid prophage
PMID:8411153. It is the toxin of a type II toxin–antitoxin (addiction) module: loss
of the prophage stops synthesis of the labile antitoxin Phd, ClpXP degrades the residual
Phd, and free Doc kills the plasmid-free segregant [PMID:8411153 "The plasmid withdrawal
response depends on a gene named doc (death on curing), encoding a 126 amino acid
protein."; PMID:9829946 "Degradation of antidote by the host-encoded ClpXP protease
liberates the toxin, which then poisons the plasmid-free segregant"].
Structurally it is a single Fido (Fic/Doc) domain, PROSITE PS51459, Pfam PF02661 Fic,
NCBIfam TIGR01550 DOC_P1, with four crystal structures (3DD7, 3DD9, 3K33, 3KH2)
[file:BPP1/doc/doc-uniprot.txt "DOMAIN 4..120 /note=\"Fido\""]. Its Fic motif
HxFx(D/N)(A/G)NKR is degenerate relative to the canonical AMPylating motif
PMID:24448800, and that degeneracy is what converts the enzyme from an
adenylyltransferase to a kinase.
The mechanism was revised between 2008 and 2014, and this matters for the GO annotations.
2008 model (superseded in its mechanistic part). Liu et al. showed that inducing Doc
in E. coli arrests growth and sharply inhibits translation without perturbing
transcription or replication, and does so without mRNA cleavage PMID:18398006. They interpreted
this as direct occupancy of the 30S subunit at the hygromycin B site in helix 44
PMID:18398006. The phenotype (host translation inhibited) is solid; the mechanism (30S
occlusion) is not.
2013/2014 model (current). Two independent groups showed that Doc is a protein kinase
that phosphorylates a single conserved threonine of elongation factor EF-Tu:
Only a threonine is phosphorylated. The Cruz MS search was run against EF-Tu with
"differential modifications of phosphorylated tyrosine, serine, and threonine residues"
PMID:24448800 at 100% sequence coverage and returned Thr-382 alone. No serine
phosphorylation by Doc has been reported by anyone. This is the basis for the review's
treatment of GO:0106310 protein serine kinase activity (see below).
Residue numbering caveat: the papers write Thr-382 of mature E. coli EF-Tu, while
UniProt's FUNCTION line for Q06259 writes 'Thr-383' (P0CE47 numbering, which retains the
initiator Met). Same residue, two conventions — do not record a bare residue number
without saying which sequence it is on.
Reverse reaction. Doc also removes the phosphate when ADP or GDP is present
PMID:24141193, so the same active
site runs both directions depending on the nucleotide pool. Phd blocks both directions.
Doc is a corepressor of its own operon, and does it through Phd rather than through DNA.
The decisive experiment is with the non-toxic DocH66Y allele: PMID:9829946. In vivo, Doc deepens Phd-mediated repression ~4-fold
PMID:9829946, and at high Doc:Phd ratios the effect inverts
PMID:9829946 — conditional cooperativity. The structural
basis: monomeric Doc engages two Phd dimers and orders Phd's intrinsically disordered
C-terminal domain, which in turn structures Phd's N-terminal DNA-binding domain
PMID:20603017. Doc does join the protein–DNA complex
(supershift) but contributes no DNA contact of its own PMID:9829946.
Comparator check: Phd itself (UniProtKB:Q06253) carries GO:0003677 DNA binding (IDA,
PMID:20603017), GO:0001217 DNA-binding transcription repressor activity and GO:0045892
negative regulation of DNA-templated transcription in GOA; Doc carries none of the
transcription terms but does carry GO:0003677 by UniProt keyword projection. The division
of labour in GOA is therefore exactly inverted from the experimental picture, and this
review corrects it (REMOVE GO:0003677; NEW GO:0003714 + GO:0045892).
Hazan et al. reported that phd/doc post-segregational killing requires the E. coli
chromosomal mazEF module PMID:11222604
and that host protein synthesis is inhibited under PSK conditions PMID:11222604. Separately, Doc overexpression induces the
mRNA interferase activity of host RelE [file:BPP1/doc/doc-uniprot.txt "Overexpression of
doc can induce the mRNA interferase activity of host RelE in vivo"]. Both are downstream
host responses, not activities of Doc, and neither was annotated as such — correctly.
| Term | Evid | Source | Action |
|---|---|---|---|
| GO:0004674 protein Ser/Thr kinase activity | IDA | CACAO (PMID:24448800) | ACCEPT (core) |
| GO:0004674 | IEA EC:2.7.11.1 | UniProt | ACCEPT |
| GO:0106310 protein serine kinase activity | IEA RHEA:17989 | RHEA | MODIFY → GO:0004674 (GO:0106311 protein threonine kinase activity was merged into GO:0004674) |
| GO:0005524 ATP binding | IEA KW | UniProt | ACCEPT |
| GO:0000166 nucleotide binding | IEA KW | UniProt | TOO_GENERAL |
| GO:0016301 kinase activity | IEA InterPro+KW | UniProt | TOO_GENERAL |
| GO:0016740 transferase activity | IEA KW | UniProt | TOO_GENERAL |
| GO:0003677 DNA binding | IEA KW-0238 | UniProt | REMOVE (+ source defect) |
| GO:0005515 protein binding (Phd) | IPI | IntAct | MODIFY → GO:0003714 |
| GO:0005515 protein binding (EF-Tu) | IPI | IntAct | MODIFY → GO:0004674 |
| GO:0039604 symbiont-mediated suppression of host translation | IDA | CACAO (PMID:18398006) | ACCEPT (core) |
| GO:0039604 | IMP | CACAO (PMID:11222604) | ACCEPT |
| GO:0044696 killing by virus of host cell by post-segregational killing | IMP | CACAO (PMID:11222604) | ACCEPT (core) |
Plus three NEW rows: GO:0005525 GTP binding (IDA, PMID:24141193), GO:0003714 transcription
corepressor activity (IDA, PMID:9829946) and GO:0045892 negative regulation of
DNA-templated transcription (IDA, PMID:9829946).
All four CACAO rows survive. This is a well-curated gene product: each row is on the right
protein, from a paper that assays it, with an evidence code matching the assay, and with a
term at an appropriate granularity. None of the batch-1 failure shapes applies:
GO:0039604 is the direct readout of both cited papersGO:0044696 isThe one notable point is a timing observation rather than a defect: the
GO:0039604 IDA row rests on PMID:18398006, whose mechanistic conclusion (30S occlusion)
was overturned five years later by PMID:24141193 and PMID:24448800. The term is
unaffected, because GO:0039604 asserts the outcome and not the mechanism — but it is a
reminder that a CACAO row anchored to a mechanism paper can outlive the mechanism. This is
recorded in the review as a finding_review with finding_status: OVERTURNED on the
30S-binding finding, not as an action change. Worth adding to the taxonomy as
mechanism superseded, term unaffected: a row that remains correct for a reason the
original curator did not intend.
The defects in this gene's GOA are all on the automatic side, from UniProt keyword and
Rhea projection: a DNA-binding term that the 1998 paper explicitly refutes, and a
serine-kinase term for an enzyme with no serine substrate. Both predate and are unrelated
to the student work.