Saccharomyces cerevisiae Apn1: the major apurinic/apyrimidinic (AP) endonuclease and
DNA 3'-repair diesterase, the eukaryotic counterpart of E. coli endonuclease IV (Nfo).
Reviewed for the CACAO project (four GOA rows carry ASSIGNED BY = CACAO, all from
PMID:12767225).
Provenance note on this file. No deep-research provider keys were available in this
container, so just deep-research was not run and no -deep-research-<provider>.md file
exists for APN1. Everything below was assembled by hand from the cached publications in
publications/, the UniProt flat file, PubMed, and QuickGO/OLS term lookups. Two cached
papers have full text (PMID:12397185, PMID:16024777); the rest are abstract-only, which is
flagged where it limits a judgement.
nfo, HAMAP MF_00152, PANTHER PTHR21445:SF0AP endonuclease / 3'-repair diesterase, and it is the major one. Gene disruption
showed the protein accounts for essentially all of both activities in extracts
PMID:1693433. UniProt's FUNCTION states the same
[file:yeast/APN1/APN1-uniprot.txt "APN1 accounts for > 97% of both apurinic/apyrimidinic (AP)"].
3'-diesterase was the activity the protein was first purified by. Johnson & Demple
purified the enzyme on a 3'-phosphoglycolaldehyde substrate
PMID:3056935. The same paper reports that
purification stripped a tightly bound metal and that activity was restored by Co2+, Mn2+,
Ni2+ or Zn2+ PMID:3056935. That the enzyme also hydrolyses a
3'-phosphoglycolate terminus is stated in the Nash-lab full text
PMID:12397185, and APN1/APN2
loss is what sensitises cells to bleomycin, a 3'-PG-generating agent
PMID:12397185. This argues for the specific GO term
GO:0090580 (phosphodiesterase activity, acting on 3'-phosphoglycolate-terminated DNA strands),
which human APEX1 carries, rather than the top-level GO:0008081.
Active-site zinc cluster; two substitutions abolish function. The Endo IV active site
coordinates three Zn2+ via nine residues, and the corresponding Apn1 residues Glu158 and
Asp192 are essential
PMID:12767225,
PMID:12767225. The E158G variant still binds DNA while D192G is catalytically dead
PMID:12767225,
PMID:12767225.
Progressive 3'→5' exonuclease on duplex (not single-stranded) DNA. Ishchenko et al.
characterised this directly: purified Apn1 degraded duplex but not ssDNA
PMID:16024777, with kinetics comparable to its AP endonuclease
activity PMID:16024777. It removes 3'-terminal
8oxoG, mismatches and 3'-blocking groups
PMID:16024777, and genetically this constitutes an Ogg1-independent 8oxoG
repair route PMID:16024777. The duplex-specific
reading matters for term choice: GO:0008311 (double-stranded DNA 3'-5' DNA exonuclease
activity) rather than its parent GO:0008296.
Tdp1-like 3'-tyrosyl-DNA phosphodiesterase activity — real in vitro, incidental in vivo.
Purified Apn1 removes a 3'-tyrosine from nicked/tailed duplexes
PMID:12397185,
and extracts lose the activity in apn1 cells
PMID:12397185. The authors'
own genetics argue the activity does not contribute to Top1-lesion repair
PMID:12397185. So the MF is well supported but should not be read as a core biological role.
Apn1 is required for BER in yeast extracts; the 5'-incision at AP sites is undetectable
without it PMID:9295360. Apn2 is a minor, partly redundant second AP endonuclease
PMID:12397185.
SUBCELLULAR LOCATION: Nucleus.Nucleus).GO:0008270 has an InterPro IEA row and an RCA row from the yeast zinc-proteome survey
PMID:30358795. Both are consistent with the structurally characterised trinuclear cluster
quoted above and with UniProt's curated COFACTOR note, so both are kept.
GOA has four CACAO rows on APN1 — GO:0003906 (also independently made by SGD from the same
paper), GO:0005634 nucleus, GO:0008296, GO:0017005 — all IDA from Jilani et al. 2003
Biochemistry. The cached record is abstract-only (full_text_available: false) and the
paper is paywalled (pubs.acs.org returned 403). The GONUTS wiki, where the student's evidence
statement would be recorded, is offline: https://gowiki.tamu.edu/wiki/index.php/PMID:12767225
returns only "temporarily offline while we perform upgrades", and the Internet Archive is not
reachable from this container. So the student's figure-level justification could not be read.
What the abstract does and does not show, taken row by row:
GO:0003906 AP endonuclease: directly supported — the paper is a structure–functionGO:0005634 nucleus: not mentioned in the abstract at all. The claim is nonetheless trueGO:0008296 3'-5'-DNA exonuclease: not mentioned in the abstract. The activity is real butGO:0008311 is the term the data supportGO:0017005 3'-tyrosyl-DNA phosphodiesterase: not mentioned in the abstract; the activityPattern for the CACAO taxonomy: these are not phenotype-as-process errors and not the
wrong protein. They look like over-harvesting one paper — a single highly-cited
structure–function paper annotated with the protein's whole known activity repertoire,
including activities established in other papers, and at the family-level granularity
(GO:0008296) rather than the granularity the yeast data support (GO:0008311). Compare SGD,
which curated exactly one term from this paper and sourced each other activity to the paper
that demonstrated it.
NEW decision)GO:0000012 single-strand break repair looked like a candidate NEW term, since Apn1 cleans
3'-blocking groups at single-strand breaks
PMID:16024777. Comparator query (QuickGO, 2026-10-29): neither human
APEX1 (UniProtKB:P27695) nor E. coli Nfo (UniProtKB:P0A6C1) carries GO:0000012; both are
annotated to GO:0006284 and GO:0006281 only. The absence across species and MODs reads as a
convention, not a gap, so no NEW row was proposed. No other process term survived the
participation test either (the elevated mitochondrial mutation rate in apn1Δ is a phenotype,
not a process Apn1 performs).