DPH3P1 notes (Q9H4G8, "Putative DPH3 homolog B")

2026-10-04: Tier 3 audit (MICROPROTEINS project, pseudogene / uncertain products)

UniProt Q9H4G8 is PE 5 (Uncertain), with "CAUTION: Could be the product of a pseudogene". The sequence
comes only from genomic DNA (AL035669, NOT_ANNOTATED_CDS), from the chromosome 20 genome sequencing
project (UniProt RN [1]). There is no cDNA, CCDS, RefSeq protein or Ensembl translation.
HGNC:16136 (REST, 2026-10-04): locus_type pseudogene, "diphthamide biosynthesis 3
pseudogene 1". Previous symbols C20orf143, ZCSL1, DPH3B. No MANE transcript.
Ensembl ENSG00000233838: biotype processed_pseudogene, 20q13.33, a single 249-nt exon. The
transcript ENST00000486648 has no translation.
GTEx v8 returns no expression rows for the gene, while parent DPH3 is expressed in all 54
tissues ([file:human/DPH3P1/DPH3P1-bioinformatics/RESULTS.md]).
PubMed: 0 hits for DPH3P1, DPH3B or ZCSL1 (esearch, 2026-10-04). No literature exists on the locus.
Proteomics: UniProt lists a PeptideAtlas cross-reference, but my automated query to check
for DPH3P1-unique peptides failed. Not established either way.

The parent is DPH3 (Q96FX2; yeast KTI11/Dph3). DPH3 is a CSL-type zinc finger protein that
binds iron and acts as the electron donor that reduces the Dph1-Dph2 Fe-S cluster in the first
step of diphthamide biosynthesis: "yeast Dph3 (also known as KTI11), a CSL-type zinc finger
protein, can bind iron and in the reduced state can serve as an electron donor to reduce the
Fe-S cluster in Dph1-Dph2" PMID:24422557. The metal ligands are four cysteines: "the
Cys25, Cys27, Cys47, and Cys50 are responsible for metal coordination" PMID:24422557.
Dph3 can also donate an iron atom to convert the [3Fe-4S] cluster to [4Fe-4S] PMID:34154323.
Alignment (bioinformatics folder): 70/82 identical, with the frame intact on the genome
(78 aa, matching UniProt). All four DPH3 metal-ligand cysteines (C26, C28, C48, C51) are
retained, and the C-terminal LVKC is missing. The sequence therefore shows no residue-level
loss. It sits inside the DPH3 clade and would probably bind metal if it were made.
PANTHER: the entry is assigned to its own subfamily PTHR21454:SF23 "DPH3 HOMOLOG B-RELATED",
while human DPH3 is in SF31. The IBA rows come from PTN000485452, the DPH3 ortholog node
(with/from includes yeast S000007587 Kti11 and mouse MGI:1922658 Dph3). The node placement is
sound for real DPH3 orthologs. DPH3P1 inherits from it only because a recent processed
retrocopy of DPH3 sits in the tree as a leaf.

The problem is not phylogenetic placement or residue loss. The problem is that the leaf is a
pseudogene with no evidence of transcription into mRNA or of translation. Following the Tier 3
rule ("REMOVE where the product is unlikely to exist"), all five rows (3 IBA, 2 IEA
InterPro) were first set to REMOVE.
Revised after PR #3680 review: the two metal-binding MF rows (GO:0005506, GO:0046872) are
MARK_AS_OVER_ANNOTATED, because their residue basis (the CSL zinc-finger cysteines) is intact
and only the product's existence is in doubt. The diphthamide process and cytosol rows stay
REMOVE.
Propagation routes: PAINT IBA from PTN000485452 (iron ion binding, cytosol, diphthamide
modification), and InterPro2GO from IPR044248 (DPH3/4-like; diphthamide BP and metal ion
binding).
Upstream fix: exclude PE5 or pseudogene-flagged leaves from IBA propagation, or have PANTHER
mark SF23 as a pseudogene leaf. The same applies to the InterPro2GO pipeline for PE5 entries
carrying a "product of a pseudogene" caution.

GOA has no tRNA wobble uridine modification row for DPH3P1. Yeast Dph3/Kti11 does work in
Elongator-dependent wobble uridine modification PMID:27694803, but GOA only gives DPH3P1
the diphthamide term, so there was nothing to review for that process.