PRKDC (DNA-PKcs) — Phylogenetic Distribution & the Efferocytosis Macrophage-Proliferation Route

Gene: PRKDC / DNA-PKcs (human) · UniProt: P78527
Family: PANTHER PTHR11139:SF68 ("DNA-dependent protein kinase catalytic subunit"); OrthoDB 8724497at2759 (Eukaryota), 5996771at33208 (Metazoa)
Focus: free_text hypothesis prkdc-animal-distribution · Taxon: Homo sapiens (NCBITaxon:9606)
Investigation: 3 iterations · 4 confirmed findings · 2 primary papers

Summary

Verdict: Partially supported — the specific loss claim is verified, but the framing over-reaches, and the result is curation-irrelevant for human PRKDC.

The seed hypothesis bundles an evolutionary claim with a mechanistic premise. The mechanistic premise — that the efferocytosis-induced macrophage-proliferation route involves DNA-PKcs sensing apoptotic-cell DNA — is supported by the primary source (Gerlach et al. 2021, PMID: 34784501), with one refinement: the activating ligand is DNase2a-derived nucleotides, not intact double-strand DNA ends, and DNA-PKcs acts upstream of an mTORC2/Rictor→Myc axis in resolving macrophages. The evolutionary claim — that DNA-PKcs is absent from Drosophila melanogaster and Caenorhabditis elegans — is directly verified. However, the implied narrative that absence characterizes "some animal lineages" broadly is overstated: DNA-PKcs is an ancient, metazoan-ancestral gene retained across nearly all animal clades, with genuine losses confined narrowly to Drosophila, Caenorhabditis, and (outside animals) budding yeast.

Across three iterations, two independent orthology resources (OrthoDB and a per-organism UniProt/PANTHER PTHR11139 family census) converged on the same picture. The OrthoDB orthogroup places DNA-PKcs in 287 of 5,768 sampled eukaryotes, overwhelmingly single-copy, including non-vertebrate metazoans. The UniProt/PANTHER census then confirmed that fly and worm — both complete, gold-standard genomes — retain the sister kinases of the PIKK family (TRRAP, mTOR, ATR, SMG1) but have specifically lost the DNA-PKcs subfamily (SF68), while the mosquito Anopheles (same insect order as Drosophila) retains it. A phylum-broad census further showed SF68 present from sponges and cnidarians through most insects, and revealed that earlier OrthoDB insect "absences" (beetle, bee, silkmoth) were sampling artifacts, not losses.

Bottom line for the curator: this is a genuine and interesting evolutionary observation, but it provides no basis to add, remove, generalize, or specialize any GO term on human PRKDC. Ortholog presence/absence in flies and worms is a property of those genomes, not evidence about the human gene product's molecular function, biological process, or cellular component. The item should be logged as non-core background/evolutionary context, and the efferocytosis route (Gerlach 2021) treated as a context-specific, downstream role rather than a core human function.

Key Findings

Finding 1 — DNA-PKcs is an ancient eukaryotic gene lost independently in specific lineages, not vertebrate-restricted

The historical assumption that DNA-PKcs is a vertebrate-specific PIKK arose largely because it is absent from the two most-studied invertebrate models. The orthology data refute this reading. The OrthoDB orthogroup matching PANTHER PTHR11139:SF68 and UniProt P78527 (Eukaryota node 8724497at2759; Metazoa node 5996771at33208) has a phyletic profile of present in 287 of 5,768 sampled eukaryotic species, and is overwhelmingly single-copy (283/287). Among the metazoan carriers are non-vertebrate representatives such as the cephalochordate Branchiostoma lanceolatum, the echinoderm Acanthaster planci, and the dipteran insects Anopheles merus and Bradysia odoriphaga. Drosophila melanogaster, Caenorhabditis elegans, and Saccharomyces cerevisiae are confirmed absent.

This broad-but-patchy, single-copy pattern is the hallmark of an ancient gene subject to independent secondary losses, not of a late vertebrate innovation. It is directly corroborated by Lees-Miller et al. 2021 (PMID: 33035590), who report that DNA-PKcs "is widely distributed in invertebrates, fungi, plants, and protists." The fly/worm absences are therefore best read as lineage-specific losses against an ancient background — not as evidence of a narrow, derived distribution.

Finding 2 — The efferocytosis route uses DNA-PKcs to sense DNase2a-derived nucleotides (context-specific, not a new GO assignment)

The mechanistic premise is accurate but requires precision about the stimulus. Gerlach et al. 2021 (PMID: 34784501) showed that, after engulfment of apoptotic cells, nucleotides derived from the hydrolysis of apoptotic-cell DNA by phagolysosomal DNase2a activate a DNA-PKcs–mTORC2/Rictor pathway that increases Myc to promote non-inflammatory macrophage proliferation. Silencing macrophage DNase2a or Rictor blocked efferocyte proliferation in both a dexamethasone-thymus model and an atherosclerosis-regression model.

Two nuances matter for curation. First, the proximal stimulus is degraded DNA/nucleotides, not intact double-strand DNA ends — mechanistically distinct from DNA-PKcs's canonical role binding double-strand breaks during non-homologous end joining (NHEJ). Second, the role is cell-type- and context-specific (resolving/efferocytic macrophages), demonstrated in mouse and in vitro. It is a downstream signaling function layered on a specialized cell state, not a constitutive core activity of the gene product.

Finding 3 — Drosophila and C. elegans absence is a true subfamily-specific loss; mosquito retains it

To exclude the possibility that fly/worm absences were database gaps, an independent census of the ATM-related parent family PTHR11139 was run per organism via UniProt+PANTHER. Human carries the DNA-PKcs subfamily PTHR11139:SF68 (P78527, 4,128 aa) alongside mTOR (SF9), ATR (SF69), SMG1 (SF71), and TRRAP (SF1). Drosophila melanogaster has 8 PTHR11139 members but zero SF68 (present: TRRAP/SF1, mTor/SF9, Smg1/SF119, ATR/SF69). C. elegans has 5 PTHR11139 members but zero SF68 (TRRAP, smg-1, ATR, TOR). Because both genomes are complete and the sister kinases are present, SF68 absence is a genuine gene loss, not an unsequenced/unannotated genome.

Critically, Anopheles gambiae retains one SF68 member (Q7Q293, 4,191 aa). The DNA-PKcs loss is therefore lineage-specific within Diptera — Drosophila lost it while mosquitoes kept it — rather than a general feature of flies or insects.

Finding 4 — DNA-PKcs is ancestral to Metazoa and retained by most insects; verified losses are narrowly restricted

A phylum-broad UniProt+PANTHER PTHR11139 census across 14 additional metazoan proteomes places the origin of DNA-PKcs at the base of the animal tree. SF68 is present in the sponge Amphimedon queenslandica, the placozoan Trichoplax adhaerens, the cnidarians Nematostella vectensis and Hydra vulgaris, the mollusk Crassostrea gigas, the annelid Capitella teleta, the echinoderm Strongylocentrotus purpuratus, the chelicerate tick Ixodes scapularis, the crustacean Daphnia pulex, and the insects Tribolium castaneum (beetle), Apis mellifera (bee), Bombyx mori (silkmoth), and Aedes aegypti (mosquito). Several of these insects had been absent from the OrthoDB orthogroup, confirming those absences were sampling artifacts rather than losses. Octopus bimaculoides returned zero PTHR11139 members total (incomplete UniProt proteome), so it is uninformative and must not be scored as a loss.

Taken together, the genuine, verified losses among animals are confined to Drosophila melanogaster and Caenorhabditis elegans (plus Saccharomyces cerevisiae outside the animals) — precisely the historical model organisms that seeded the now-refuted "vertebrate-specific" myth.

Mechanistic Model / Interpretation

The evidence separates cleanly into two threads the seed bundles together: an evolutionary thread and a mechanistic/cell-biological thread. Keeping them apart clarifies the curation implication.

     ANCIENT EUKARYOTIC / METAZOAN-ANCESTRAL ORIGIN
                        │
 DNA-PKcs (PTHR11139:SF68) present at the base of Metazoa
 (sponge, placozoan, cnidarian) and broadly retained
                        │
     ┌──────────────────────────────────┼──────────────────────────────────┐
     │                                   │                                  │
  Deuterostomes                     Protostomes                        Non-bilaterians
  (vertebrates, amphioxus,      (most retain SF68: beetle, bee,        (retain SF68)
   echinoderms — retain)         silkmoth, mosquito, tick,
                 Daphnia, mollusk, annelid)
                        │
              INDEPENDENT SECONDARY LOSSES
                        │
        ┌───────────────┴───────────────┐
 Drosophila melanogaster          Caenorhabditis elegans
   (0 SF68; 8 family)                (0 SF68; 5 family)
 [sister Anopheles KEEPS it]

Evolutionary thread. DNA-PKcs is old, single-copy, and pan-metazoan. The losses in Drosophila and C. elegans are real but narrow — two convergent, lineage-specific deletions coinciding with the two most popular invertebrate models, which created the false impression of vertebrate specificity. The mosquito/fly contrast within Diptera is the single cleanest demonstration that these are lineage-specific events, not clade-wide absences.

Mechanistic thread. In mammals, within efferocytic macrophages, DNA-PKcs senses not intact DNA ends but DNase2a-derived nucleotides produced by phagolysosomal degradation of ingested apoptotic-cell DNA, signaling through mTORC2/Rictor to raise Myc and drive non-inflammatory proliferation that resolves tissue injury. This is a specialized, context-dependent signaling role distinct from the canonical NHEJ/DNA-repair function that anchors human PRKDC's core GO annotations.

The seed's logical step — "no DNA-PKcs ⇒ route cannot operate" — is valid for Drosophila and C. elegans. But as a general statement about "some animal lineages," it is misleadingly broad: the route's molecular hardware is present almost everywhere in animals.

Evidence Matrix

PTHR11139 family census (provenance)

GO Decision Table

GO Curation Implications

Lead (requires curator verification): treat prkdc-animal-distribution as non-core background context, not a GO-term change.

• The phylogenetic-distribution observation does not support adding, removing, generalizing, or specializing any MF/BP/CC term on human PRKDC. Ortholog loss in flies/worms is a property of those genomes, not evidence about the human gene product's function.
• The canonical, well-evidenced molecular function — GO:0004677 DNA-dependent protein kinase activity — is retained and unaffected by this hypothesis.
• If the efferocytosis route were annotated at all, defensible terms (e.g., GO:0043277 apoptotic cell clearance or a regulation-of-macrophage-proliferation BP) must carry experimental evidence codes and a cell-type/species qualifier, and must not be promoted to a core PRKDC function. The stimulus should be described as DNase2a-derived nucleotides, not intact DNA.
• Do not use "protein binding" as the outcome; it is uninformative and not what the evidence addresses.

Net: log the seed as orientation/mechanistic color; it is not a curation-actionable change to the PRKDC GO set.

Mechanistic Scope

• Immediate molecular function tested: DNA-PKcs serine/threonine kinase activity within a signaling cascade (→ mTORC2/Rictor → Myc). The direct gene-product activity is kinase signaling; the actual sensed species are DNase2a-generated nucleotides, placing DNA-PKcs downstream of DNA degradation rather than as a primary intact-DNA sensor.
• Downstream phenotypes (not direct activity): macrophage proliferation, apoptotic-cell clearance, tissue resolution, plaque stabilization — pathway/organismal outcomes partly inferred from loss-of-function (DNase2a/Rictor knockdown).
• Evolutionary distribution: a property of genomes, fully separable from the human gene product's biochemical function.

Evidence Base (Literature)

• Efferocytosis induces macrophage proliferation to help resolve tissue injury. PMID: 34784501 (Gerlach et al., 2021). Establishes the efferocytosis-induced macrophage proliferation route and DNA-PKcs's role; clarifies the proximal stimulus is DNase2a-derived nucleotides. Verified snippet: "nucleotides derived from the hydrolysis of apoptotic cell DNA by phagolysosomal DNase2a activate a DNA-PKcs-mTORC2/Rictor pathway that increases Myc to promote non-inflammatory macrophage proliferation." Supports Finding 2.
• Uncovering DNA-PKcs ancient phylogeny, unique sequence motifs and insights for human disease. PMID: 33035590 (Lees-Miller et al., 2021). Shows DNA-PKcs is ancient and broadly distributed, refuting vertebrate specificity. Verified snippet: "the absence of DNA-PKcs from many model organisms has led to the assumption that DNA-PKcs is a vertebrate-specific PIKK. Here, we find that DNA-PKcs is widely distributed in invertebrates, fungi, plants, and protists." Supports Findings 1, 3, 4.

Public bioinformatics resources used for direct presence/absence calls: OrthoDB (orthogroup 8724497at2759 / 5996771at33208), UniProt (P78527 and per-organism proteomes), PANTHER family PTHR11139 / subfamily SF68.

Limitations and Knowledge Gaps

1. Absence calls depend on proteome completeness. True losses were confirmed only where the genome is complete and sister kinases are present (fly, worm, yeast). Incomplete proteomes (e.g., octopus, 0 family members) are uninformative, not evidence of loss. An exhaustive insect-wide loss map would require a per-proteome HMM sweep against assembled genomes.
2. Sampling bias in OrthoDB. The carrier set is bird-dominated; raw counts and clade-level absences must not be over-read. Iteration 3 demonstrated that beetle/bee/silkmoth OrthoDB absences were artifacts corrected by UniProt.
3. No from-scratch phylogeny. Conclusions rest on subfamily membership (PANTHER) and orthogroups (OrthoDB), not a maximum-likelihood tree with branch support; headline taxa are robust, but loss-event polarity would be strengthened by a formal reconstruction.
4. Human-relevance of the efferocytosis route. Gerlach 2021 is mouse/in-vitro; whether the DNase2a→DNA-PKcs→mTORC2→Myc axis operates identically in human macrophages, and merits a human-gene annotation, is unestablished.
5. Mechanistic directness. The proliferation phenotype relies substantially on knockdown of upstream (DNase2a) and parallel (Rictor) nodes; a catalytic-dead vs. kinase-active rescue would establish non-redundant DNA-PKcs requirement.

Proposed Follow-up Experiments / Actions

1. Genome-level loss confirmation. Run an HMMER/tBLASTn sweep with the PTHR11139:SF68 HMM against assembled Drosophila and C. elegans genomes (not just proteomes), and against "absent" taxa like octopus, to separate true loss from annotation gaps. Expected: no SF68 in fly/worm genomic DNA; SF68 recovered in octopus genome. This is the single most informative confirmatory analysis.
2. Formal phylogeny + ancestral-state reconstruction. Build a PTHR11139 maximum-likelihood tree highlighting SF68 to confirm metazoan-ancestral origin and polarity of the two independent loss events.
3. Human efferocytosis validation. Inhibit (NU7441) or knock out DNA-PKcs in primary human efferocytic macrophages; measure Myc induction and proliferation — distinguishes conserved-in-human from mouse-specific.
4. Stimulus-specificity reconstitution. Compare DNA-PKcs activation by DNase2a-digested nucleotides vs. intact dsDNA ends to confirm the nucleotide-sensing mechanism.
5. Curation action. Mark prkdc-animal-distribution as non-core/background; make no GO change on its basis. Candidate references to verify: PMID 34784501 and PMID 33035590 (snippets above). Pose the curator question: is ortholog distribution intended to scope a BP annotation (valid as a taxon-constraint note) or to assert a function (not valid)? Only the former is defensible.

Provenance

Computed via public OrthoDB REST (data.orthodb.org/current) and UniProt REST (rest.uniprot.org). Verified facts: PTHR11139:SF68 ↔ P78527; DNA-PKcs orthogroup present in 287 eukaryotic species (283 single-copy); Drosophila melanogaster, C. elegans, S. cerevisiae absent from the orthogroup; invertebrate carriers include Branchiostoma lanceolatum, Acanthaster planci, Anopheles merus, Bradysia odoriphaga. UniProt/PANTHER family census confirmed fly (8 members) and worm (5 members) retain TRRAP/mTOR/ATR/SMG1 but zero SF68 (true subfamily-specific loss); Anopheles gambiae retains SF68 (Q7Q293); phylum-broad census confirmed SF68 presence from sponges/cnidarians through most insects, with OrthoDB insect absences shown to be sampling artifacts.

Artifacts

• OpenScientist final report
• OpenScientist final report
• OpenScientist prkdc metazoan phylum census
• OpenScientist prkdc pthr11139 family census