DNAI2 supplies an intermediate-chain component of the outer dynein arm, helping organize the motor that drives motile cilia. Human respiratory-cell immunofluorescence and ultrastructure show that loss of DNAI2 disrupts outer-arm incorporation. Normal human sperm-tail localization is directly observed, but the affected infertile male in the principal 2008 study had no sperm analysis. DNAI2 contributes to the complex's motor activity; the heavy chains contain the ATPase and motor domains PMID:18950741(https://doi.org/10.1016/j.ajhg.2008.10.001).
The 2008 paper contains a spatial inconsistency in its cross-gene comparison: its abstract reverses the proximal/distal pattern in DNAI1-mutant respiratory cilia. The full Results/Discussion and Figure 7 caption support moderate proximal DNAI2 retention and distal loss. A conflicting ODA type-number label in the Results is not used to override the spatial observation. The human genetic cohort in the CCDC114 paper separately includes DNAI2 p.Trp453* with isolated outer-arm defects; the ARMC4 paper directly images DNAI2 in human respiratory axonemes PMID:23261302(https://doi.org/10.1016/j.ajhg.2012.11.003), PMID:23849778(https://doi.org/10.1016/j.ajhg.2013.06.009).
DNAI2 has a cytoplasmic preassembly pool in immature human airway cells, in addition to its mature axonemal pool. Endogenous human HEATR2/DNAAF5 associates with DNAI2 in day-40 bronchial air–liquid interface cultures. Human DNAAF6 and DNAI2 also associate in controlled reciprocal tagged-cell assays. These experiments support the positive interactions but do not make DNAI2 a chaperone or identify purified interfaces PMID:25232951(https://doi.org/10.1371/journal.pgen.1004577), PMID:28176794(https://doi.org/10.1038/ncomms14279), PMID:29916806(https://doi.org/10.7554/eLife.34389).
DynAP localization is conserved evidence from explicitly tagged Xenopus Dnai2. It concentrates in outer-arm-rich subregions and exchanges slowly compared with assembly factors; it is not evidence that the dynein subunit itself performs chaperone chemistry. Human cytoplasmic preassembly observations provide compatible context, but the Xenopus source is not mislabeled as direct human DNAI2 DynAP imaging PMID:30561330(https://doi.org/10.7554/eLife.38497), PMID:33263282(https://doi.org/10.7554/eLife.58662).
The five IBA assertions were traced to their literal ancestral IBD records in PTHR12442. The sequence-matched human leaf descends from both PTN000277823 and the narrower PTN000278033. Target self-inclusion is valid experimental grounding. The deeper light/heavy-chain-binding claims are judged as inherited functions, not as pairwise transfer from a counted set of donors. Direct biochemical binding in other intermediate-chain systems and human/ortholog outer-arm association are distinguished from an unmeasured human DNAI2 interface PMID:11914076(https://doi.org/10.1021/bi011970h), PMID:25830415(https://pubmed.ncbi.nlm.nih.gov/25830415/), PMID:10336435(https://doi.org/10.1074/jbc.274.22.15447), PMID:10601339(https://doi.org/10.1083/jcb.147.6.1261).
The unusual external-plasma-membrane annotation is left unresolved. Its mouse donor is genuinely positive: the accessible abstract reports Dnaic2 on secondary/antral oocyte surfaces. The full localization controls remain inaccessible, so the extracellular-facing topology and transfer to the human oocyte are not established here. This is not a negative result and is not rejected because DNAI2 also functions in cilia PMID:18547164(https://doi.org/10.1089/dna.2008.0752).
The official natural products have 605 and 593 residues. The latter lacks canonical residues 450–461; this 12-residue event is distinct from the disease-associated 49-residue exon-11 deletion. No equivalent function or inactivity is inferred from the natural deletion alone. The source assays have not been assigned to a current natural accession solely from an unversioned clone name UniProt Q9GZS0, PMID:18950741(https://doi.org/10.1016/j.ajhg.2008.10.001).
The source-evidence guide records public locators and the actual access scope. The positive generic-binding annotations follow the scoped project instruction. Genuine Falcon and configured Perplexity fallback attempts failed with authentication/quota errors; no provider research report was manufactured. The manual evidence review is recorded here, independently of those failed attempts.
The two accepted chain-binding activities are now explicit in the structured core: heavy-chain binding and light-chain binding are complementary contributions of the same outer-arm intermediate-chain role. The heavy-chain entry retains the contribution to microtubule motor activity; neither entry assigns autonomous ATP hydrolysis or motor force generation to DNAI2. All 27 annotation decisions and both natural products are unchanged.
The positive evolutionary basis remains the PAINT placement of both binding functions at PTN000277823, ancestral to the sequence-matched human DNAI2 leaf. Human DNAI2–DNAH5 proximity in nasal cells, mouse endogenous outer-arm co-immunoprecipitation, and Xenopus GFP-Dnai2 enrichment of heavy and light chains corroborate that structural role. The first paper's Figure 4A is human proximity ligation; Figures 4C–E show mouse complexes. Its endogenous-IP Methods also describe human airway cultures; that additional protocol is not presented here as a purified interaction assay. The second paper's Figure 3B and AP-MS Methods use Xenopus animal caps and matched GFP-only controls. Neither source measures a purified human DNAI2 binding interface PMID:29916806(https://doi.org/10.7554/eLife.34389), PMID:33263282(https://doi.org/10.7554/eLife.58662).
The light-chain donor experiment tests WDR34/WDR60 with TCTEX1D2 by in-vitro binding, not human DNAI2. Likewise, the mapped Dictyostelium cytoplasmic heavy-chain interface is not transferred to axonemal DNAI2. The retained binding classes follow the curator's ancestral judgment, with the distinct experimental systems stated explicitly PMID:25830415(https://doi.org/10.4161/15384101.2014.985066), PMID:10336435(https://doi.org/10.1074/jbc.274.22.15447).
The cilium-assembly refinement also retains its verified ontology basis: GO:0036158 is_a GO:0070286 part_of GO:0035082 part_of GO:0060271. Cytoplasmic preassembly can therefore be part of cilium assembly in this ontology. The source-evidence guide records the fresh typed-path endpoint. Short quotations occur once in the two core entries, while these notes supply their organism, assay and mechanistic limits.