EML1 has two experimentally separable tubulin-binding modes. Its N-terminal self-association region together with the adjacent basic region supports microtubule-polymer binding (PMID:25740311, Results on self-association and microtubule binding; Discussion). The concave TAPE surface associates with soluble alpha/beta-tubulin even after removal of the N-terminal region used for self-association (PMID:24706829, Results on soluble tubulin and Methods on constructs and copurification). Both modes now appear explicitly in the core-function description. The structured molecular-function term remains GO:0008017 microtubule binding; its ancestor GO:0015631 does not separately encode the soluble substrate state. This ontology limitation does not make the two experimentally distinct modes biologically redundant.
Purified human EML1 and its 1–174 region bind pig-brain microtubules in PMID:25740311. The 80–815 Δ-TD truncation loses polymer binding; the isolated 23–78 construct is a separate experiment, not the deleted interval of an internal deletion. The L59A/D61A double mutant reduces binding. PMID:24706829 instead measures soluble alpha/beta-tubulin copurification with the human TAPE region under cold, microtubule-depolymerizing conditions. Neither experiment establishes a catalytic reaction, polymerization rate or affinity constant.
Mouse cortical progenitor experiments in PMID:24859200 connect Eml1 to spindle orientation, cycling and positioning, while human variants support disease relevance. The original Discussion explicitly reports tagged protein at metaphase spindle poles; endogenous localization was not analyzed. Methods identify mouse Eml1 for FLAG/EGFP vectors, but a separate YFP gift lacks an accession in the inspected passage. These observations support a conserved, contextual localization, without assigning every construct to a particular natural human product.
The review preserves 39 source annotations: 11 ACCEPT, 4 MODIFY, 23 KEEP_AS_NON_CORE and 1 UNDECIDED. Cytoplasm is refined to cytosol using the mouse donor localization and the independent soluble protein evidence; this does not claim that historical imaging alone resolved cytosol. Cytoskeleton terms are refined to microtubule while retaining the original qualifiers. Eight source-specific interaction pairs remain generic non-core binding; partner identities do not establish EML1 RNA-binding, chaperone or protease activity. BioGRID 3018756 and 3018857 corroborate the FXR1 pair, and the selected BioPlex edge corroborates NUDC association.
The calcium-ion-binding NAS annotation from PMID:9226380 remains unresolved. Only its authentic abstract was recovered; homology and the historical 717-residue clone do not establish or refute a calcium-binding assay. Two current official products, O00423-1 (815 residues) and O00423-3 (834 residues), are retained exactly. Historical clones, tags and experimental truncations are not additional natural products or proof of equal product activity.
All three PAINT assertions were checked against their literal ancestral nodes and experimental donor records. The human and mouse displayed sequences match the captured MSA. Donor count and self-donor status are not grounds for rejection. ARBA outputs and full condition sets are available in the portable evidence, but target-specific FunFam premise matching was not independently established. An initial HPA lookup used an incorrect identifier and was excluded; the correct EML1 XML did not provide a subcellular immunofluorescence section, so HPA is not a localization anchor here.
The normal cache flags remain unchanged. Separately recovered original text and the one inspected figure are documented in the source guide and original locator inventory. The portable replay reproduces database joins, not paper experiments or curation judgments. Provider attempts failed and supplied no research report; the review rests on the documented primary and database evidence. No publication quotations or new annotations were added.
The three PAINT rows, two ARBA rows and eight source-specific interaction rows now cite the existing reproducible evidence report where their reasons depend on its ancestry, rule or pair-record projections. The core description cites PMID:24706829 for soluble-tubulin association separately from the polymer-binding sources. Calcium binding and the physiological relationship between the two tubulin-binding modes are recorded as explicit questions. All 39 imported source objects, actions and both natural products are unchanged; supported generic binding remains non-core under the project instruction. No publication quotations or source-cache edits were introduced.