Q9VIU4 is the native 427-residue PA product of FBgn0053116 (AAF53821/NP_788074); FlyBase lists one transcript and one polypeptide. The locus is also called dCCS1 and formerly CG10355. FlyBase.
PMID:23449981(https://pmc.ncbi.nlm.nih.gov/articles/PMC3630839/) is cached with full text. Its title emphasizes the Golgi ceramide phosphoethanolamine synthase, but its experiments also explicitly assay CG33116. The comparative family analysis identifies “three proteins from the CEPT subfamily (dCCS1/CG33116, dCCS2/CG6016 and dCCS3/CG7149)”. It tests all four candidate synthases and assigns CPE synthesis to CG4585/dCCS4, not CG33116. The phosphatidylethanolamine activity in the review is a conserved-subfamily/PAINT inference, not a misreading of the CPE assay as a positive CG33116 result.
The decisive localization passage is “Both dCCS1-V5 and dCCS2-V5 localized exclusively to the ER”, with overlapping ER marker staining in S2 cells and consistent results in HeLa. The same experiment places dCCS3/dCCS4 at the Golgi. Thus CG33116 has specific evidence of localization divergence within the ancestral family: the Golgi IBA is removed while ER assignments are accepted. This judgment rests on the actual target comparison, not on the number of descendants supporting the ancestral annotation.
The original ProtNLM name in current UniProt and source predictions is not the proof of substrate specificity. The published subfamily analysis, direct ER localization, and curated ancestral activity assignment supply the inference. All three original GO predictions are LSP relative to those more precise functions and locations.
The completed Falcon investigation was inspected and agrees on target identity, ER localization, the conserved PE-synthesis inference, and the distinction from CG4585. It also identifies pathway-context studies of Pect and Toll activation; these are not imported as CG33116-specific neuronal or immune functions. The decisive local excerpts are taken from the full primary article rather than repeated synthesis prose.