Bacillus phage phi29 gene 2 (DNA polymerase, p2 / gp2) - curation notes

UniProt: P03680 (DPOL_BPPH2), 575 aa, PE 1: Evidence at protein level.
Taxon: NCBITaxon:2884424, Bacillus phage phi29. Host: Bacillus subtilis.

Provenance caveat. No deep-research API keys exist in this container, so
just deep-research was not run and there is no 2-deep-research-*.md file.
The literature below was assembled directly: from the cached publications in
publications/, from the curated UniProt record (which cites ~50 primary
papers on this protein), and from PubMed/EuropePMC lookups. Every cached
publication used here is abstract-only (full_text_available: false)
except PMID:16511564.

What the protein is

phi29 DNA polymerase is the single-subunit, 575-residue family-B
("alpha-like") replicase that carries out the whole of phi29 genome
replication. UniProt gives it two EC numbers, each with experimental
evidence: EC 2.7.7.7 from PMID:3863101 and EC 3.1.11.- from PMID:2790959
[file:BPPH2/2/2-uniprot.txt "EC=2.7.7.7 {ECO:0000269|PubMed:3863101};"].

Three catalytic activities in one chain, per the curated FUNCTION line
[file:BPPH2/2/2-uniprot.txt "This polymerase possesses three enzymatic
activities: DNA synthesis (polymerase), primer terminal protein (TP)
deoxynucleotidylation, which is the formation of a covalent linkage
(phosphoester) between the hydroxyl group of a specific serine residue in TP
and 5'-dAMP, a reaction directed by the second T at the 3' end, and 3' to 5'
exonuclease activity"]:

  1. DNA-templated DNA polymerisation. Highly processive and strand
    displacing with no accessory proteins: >70 kb products on primed M13
    PMID:2498321, and
    PMID:2498321. The
    crystal structure explains both: a tunnel too narrow for duplex DNA forces
    strand separation ahead of the active site, and the enzyme encircles both
    downstream template and upstream duplex
    PMID:15546620,
    PMID:15546620.
  2. Terminal-protein deoxynucleotidylation (protein-primed initiation).
    The polymerase forms a heterodimer with free terminal protein p3 and
    covalently attaches the first dAMP to the hydroxyl of p3 Ser-232. This is
    catalysis performed by the polymerase, on a protein acceptor, and it
    happens even without DNA
    PMID:1730646, with the same linkage chemistry and the same
    Ser232/YCDTD-motif requirements as genuine initiation
    PMID:1730646. Initiation is directed by an
    internal template thymine, with sliding-back recovery of the terminal
    nucleotides PMID:19011105. The polymerase:TP structure shows the TP
    priming domain occupying the cleft duplex DNA will later use, which is
    also why initiation is confined to the genome ends
    PMID:16511564,
    PMID:16511564.
    GO has no molecular function term for this activity. Searches of
    QuickGO found only GO:0018175 protein nucleotidylation (a BP) and
    GO:0018117 protein adenylylation; neither is a nucleotidyltransferase MF
    and neither is template-directed. Filed as proposed_new_terms.
  3. 3'-5' exonuclease / proofreading. A separate N-terminal active site,
    identified by site-directed mutagenesis of the conserved Exo motifs
    PMID:2790959. UniProt localises it to
    residues 1-191 [file:BPPH2/2/2-uniprot.txt "The N-terminus contains the
    3'-5' exonuclease activity and strand displacement ability"], and editing
    is intramolecular, the primer travelling between the two active sites
    without dissociating [file:BPPH2/2/2-uniprot.txt "DNA polymerase edits the
    polymerization errors using an intramolecular pathway as the primer
    terminus travels from one active site to the other without dissociation
    from the DNA"]. The substrate is a free 3' terminus of single-stranded
    DNA, and separate metal sites serve the two activities
    PMID:1310035.

Mg2+ is the cofactor, with eight curated FT BINDING sites (145, 169, 249,
250, 456, 458), several marked catalytic and backed by PDB 1XI1/2PYJ and by
PMID:9784372. Protein-primed initiation specifically prefers Mn2+
PMID:1310035.

The two CACAO papers

Both are abstract-only in the cache, and both are multi-gene phi29 survey
papers shared with other numerically-named phi29 entries (grep over
genes/BPPH2/*/*-goa.tsv: PMID:820555 is cited on genes 1, 2, 3, 4, 6 and
15; PMID:6402761 on genes 2 and 3).

The third CACAO row re-uses PMID:6402761 for the same term with a different
ECO refinement: ECO:0007161 nonsense mutation phenotypic
evidence (2025-02-08) and ECO:0006099 temperature-sensitive mutant
phenotypic evidence (2025-02-17). Both collapse to IMP in the GAF. phi29 has both sus (nonsense)
and ts mutant collections, and PMID:2118623 is specifically about ts mutants
of this polymerase, so the two rows may record two mutant classes in the same
paper rather than a duplicate; the abstract does not say which mutants were
used, and the full text is unavailable (EuropePMC returns a 500 for PMC325798;
the 1983 PMC copy is scanned images).

GONUTS is now retired, not offline. gowiki.tamu.edu returns HTTP 410
with "GONUTS, the GO wiki, has been retired" and "The GO wiki that was at
gowiki.tamu.edu is no longer online." The Internet Archive has no snapshot of
either the protein page or the per-paper pages (archive.org/wayback/available
returns {"archived_snapshots": {}} for both
.../Bacillus_phage_phi29:2 and .../PMID:6402761). So the figure-level
evidence behind every abstract-only CACAO row in this project is now
permanently unverifiable, and the CACAO.md note that the check should be
re-tried "when GONUTS returns" needs updating.

Qualifiers

All three CACAO rows carry acts_upstream_of_or_within_positive_effect. This
gene product performs replication initiation and genome replication: it is
the enzyme that makes the phosphodiester bonds and the TP-dAMP linkage, and
UniProt's own rows for the same process use involved_in
(2-goa.tsv: involved_in GO:0039693 ... IDA ... UniProt). involved_in is
the correct qualifier; recorded as a WRONG_QUALIFIER source defect on each
of the three rows, exactly as for AcrIIA5 in batch 2.

An InterPro2GO defect

IPR014416 "DNA-directed DNA polymerase, family B, phi29-like virus" maps to
GO:0001882 nucleoside binding (checked against the InterPro API). In GO,
nucleoside binding is not an ancestor of nucleotide binding: GO:0000166
sits under GO:1901265 nucleoside phosphate binding, while GO:0001882
descends from GO:0097367 carbohydrate derivative binding and its children
are all unphosphorylated nucleosides (GO:0032546 deoxyribonucleoside binding etc., checked via QuickGO ancestors/descendants). The polymerase
binds dNTPs, not nucleosides; nothing in the literature or in the curated
FT BINDING features (5-methyl-UTP at 254, 371, 383, 458) shows nucleoside
binding. The mapping looks like a slip for nucleotide binding and propagates
to every phi29-like family-B polymerase. Filed as a source_defects entry
against InterPro.

Separately, GO offers no generic "deoxyribonucleoside triphosphate binding"
term, so GO:0000166 nucleotide binding is the most specific statement
available for the dNTP site; raised in suggested_questions.

Relation to the rest of the phage

Protein-primed replication needs only p2 and p3 in vitro
PMID:3863101, and the same
paper shows p2 is responsible for both the initiation complex and elongation
to full length PMID:3863101. Host DNA polymerase III, gyrase and RNA polymerase inhibitors do
not affect initiation PMID:6402761, so the phage replicase is genuinely autonomous.
In vivo, p5 (SSB) and p6 (nucleoid protein) stimulate replication, but
neither is needed for initiation (PMID:6402761, above).

Term choices made in the review