DPY19L2 has a well-supported role in spermatid development and a conserved inner nuclear membrane location. Its broad mannosyltransferase activity is a curated phylogenetic inference. The human protein's substrate and the causal link between a possible enzymatic reaction and acrosome attachment remain unresolved. One integrated core retains these distinctions; no specific C-mannosyltransferase reaction, molecular adaptor activity, LINC-complex membership or additional annotation is asserted.
The nine normal-seeded source assertions and their qualifiers are preserved. Seven are accepted; the broad nucleus and membrane assertions are refined to inner nuclear membrane using independent localization evidence. Ten raw GOA records map to nine seeded assertions because two ISS records differ only in identifier/date. Neither raw record is lost. The evidence guide is a locator and access record, not an independent experimental source.
The literal PAINT IBD places broad mannosyltransferase activity at PTN001266774. The human target lies below that ancestor. Experimental donor records join to mouse Dpy19l1/A6X919 and Dpy19l3/Q71B07 (PMID:28202721), human DPY19L3/Q6ZPD9 (PMID:29405629) and worm dpy-19/P34413 (PMID:23562325). The location/development node PTN001266831 is also on the target path. Its separate IRD removes the ancestral ER-membrane inference, not the enzyme MF. The target's own human developmental evidence among the IBD descendants is legitimate grounding, not circularity.
The recovered 2017 original tests complete mouse DPY19L2 ORF expression in CHO cells. Results and Figure 4C report no restoration of UNC5A surface export. The Discussion describes no demonstrated C-mannosylation of that tested substrate by L2, while explicitly allowing other substrate requirements, expression conditions or cellular contexts. The linked supplementary Figure S6 was not accessible: the PMC endpoint returned download-challenge HTML, Europe PMC returned a non-open-access error and the publisher returned HTTP 403. No supplement controls or figure pixels are claimed as inspected.
The 2023 structural original provides experimental worm CeDPY19 structures and assays. Human L2 appears in the AlphaFold comparison in Figure 3b and its associated Results; it is not an experimentally solved human structure or human enzyme assay. This supports retention of the broad inherited function with explicit uncertainty, not a more specific substrate assertion. The short exact limitation appears once in the review's reference finding and is not repeated here.
The full human deletion study supports globozoospermia and a spermiogenesis phenotype. Its immature germ-cell observations include absent proacrosomal vesicles despite appropriate polarity/chromatoid-body positioning. The independent human cohort also supports failed head elongation/acrosome formation; its testis and mature-sperm immunoblots are separate observations rather than proof of an identified molecular bridge.
In the 2012 localization and knockout study, endogenous mouse ultrastructure and immunogold support the inner nuclear membrane location and later acrosome spreading/attachment role. Methods explicitly use a human DPY19L2-GFP construct in HEK293 and NIH3T3 hosts. Host identity is not construct identity. Initial acrosome binding can occur before later defects; displaced manchette and transport phenotypes do not establish DPY19L2 motor activity. Proposed SUN5/LINC interactions are not direct binding evidence. The 2021 FAM209 paper reports endogenous mouse Fam209-FLAG IP/MS and DPY19L2 co-IP: an association, not a purified interface or human assay.
The normal PMID:21630459 cache remains abstract-only. A genuine original-author thesis contains the study-specific Annex 8.2. Printed page 198/PDF page 199 lists DPY19L2/Q6NUT2 with 8.7% coverage and five peptides. The table row was checked visually as well as in extracted text. The thesis Methods describe CTAB-based isolation of human sperm nuclei and LC-MS analysis; the linked results identify the four-donor nuclear proteome. This is positive nuclear-fraction detection, without inner-membrane resolution or authenticated isoform-specific peptides. A separate Annex 8.3 target row belongs to another fractionation experiment and is not substituted for the 2011 result.
The inner-membrane refinements are nonexclusive and do not assign the original proteomics assay a resolution it did not measure. HPA cell-line records report mitochondrial and nucleoplasmic pools; only the underlying XML records, not microscopy images, were inspected. These non-germ-cell observations and the 2011 mature-sperm immunoblot non-detection remain visible. Their differences are not resolved by assuming contamination, identical developmental stages or interchangeable assay sensitivity.
The official products are 758 and 205 residues. The shorter product deletes residues 1–553, leaving only residue 554 of the final predicted 534–554 membrane segment and no complete predicted membrane-spanning helix. The ten segments of the full-length protein are predictions. Neither the historical tagged construct nor the proteomic peptides were authenticated to one natural isoform; loss of predicted helices does not itself establish inactivity. The family/transcript study PMID:16526957 also cautions that related pseudogene transcripts complicate expression assignment.
Machine inputs and publication caches are preserved byte-for-byte, including their literal availability flags. Separately recovered original bodies and the thesis supplement are distinguished from those flags. The configured Falcon request failed authentication and its fallback reported insufficient quota; no provider report was fabricated. Independent original review supplied the three reserved development/localization/association sources, while the owner assessed the remaining source assertions, donor phylogeny and relevant experiments. Final repository validation, status normalization and publication remain separate from this scientific assessment.