DHFR review notes

Provenance

• The first just deep-research-falcon human DHFR attempt timed out; a later run
  completed and produced DHFR-deep-research-falcon.md. The review was drafted
  from the cached GOA references in publications/, the UniProt record, and the
  two founding disease papers (PMID:21310276, PMID:21310277, fetched with
  just fetch-pmid), then checked against the falcon report.

Function

• Catalysis: recombinant human DHFR reduces dihydrofolate to tetrahydrofolate with
  NADPH PMID:2303423.
• Nuclear thymidylate synthesis: DHFR joins SHMT1 and TYMS in a complex at the
  nuclear lamina PMID:22235121.
• Tetrahydrobiopterin salvage: DHFR reduces 7,8-dihydrobiopterin, with low affinity
  in human endothelial cells PMID:23707606.
• Translational autoregulation by binding its own mRNA PMID:8490020.
• Mitochondria: PMID:21876184 attributed mitochondrial reductase activity to the
  paralog DHFR2 (DHFRL1), but the falcon report summarizes a 2024 proteomic study of
  purified HEK293 mitochondria that found canonical DHFR and no DHFR2-specific
  peptide [file:human/DHFR/DHFR-deep-research-falcon.md "P00374-derived DHFR
  contributes mitochondrial reductase activity in the human cells tested"]. The
  mitochondrion annotations were therefore accepted; that 2024 paper is not yet in
  the local publication cache.

Disease

• DHFR deficiency (constitutional megaloblastic anemia with severe neurologic
  disease): homozygous p.Asp153Val reduces DHFR activity and protein but not mRNA
  PMID:21310277;
  cerebral BH4 deficiency is part of the phenotype PMID:21310276.
• This matches the dismech entry
  Constitutional_Megaloblastic_Anemia_With_Severe_Neurologic_Disease, which marks
  GO:0004146 dihydrofolate reductase activity and GO:0046654 tetrahydrofolate
  biosynthetic process DECREASED.

2026-10-20 ORGANELLES batch 1 (mitochondrion rows)

• GO:0005739 rows: HTP, IBA, IEA and ISS all stay ACCEPT. Before this batch the three synthesis rows shared one reason; each now has its own reason recording its check.
• IBA PTN000556119 (PTHR48069, Eukaryota) seeds: mouse Dhfr MGI:94890, rat Dhfr RGD:2500 and human DHFR2 Q86XF0. DHFR2 is a paralog, but the rodent seeds are canonical orthologues with IDA PMID:25980602. The human study assigns the mitochondrial activity to DHFR PMID:38224738. So this is not a paralog over-transfer. The propagation_review is NO_FAILURE_NON_CORE.
• ISS from mouse P00375, which has IDA from PMID:25980602. The IEA (SL-0173) line is itself ECO:0000250 from P00375.
• Stayed at root. DHFR has no presequence and no TM segment. The only sub-mitochondrial result belongs to DHFRL1 PMID:21876188, and the 2024 study revised that attribution, so the result cannot be transferred. That leaves fewer than two agreeing lines.
• New cached papers: PMID:25980602, PMID:21876188.