Automated deep research was attempted with just deep-research-falcon human SPI1 --fallback perplexity-lite, but the run timed out before producing a deep-research file. This review therefore uses the cached GOA publications, UniProt record, Reactome-derived entries, and PANTHER family fetch.
SPI1 encodes PU.1, an ETS-family transcription factor expressed in hematopoietic and immune lineages. The original PU.1 cloning paper describes it as a tissue-specific DNA-binding transcription factor: PU.1 recognized the purine-rich PU-box motif 5'-GAGGAA-3' and acted as a transcriptional activator expressed in macrophages and B cells [PMID:2180582 "the PU.1 protein recognized a purine-rich sequence, 5'-GAGGAA-3' (PU box)"; PMID:2180582 "The PU.1 protein was shown to be a transcriptional activator that is expressed in macrophages and B cells"].
The core molecular function is ETS-domain sequence-specific cis-regulatory DNA binding coupled to transcriptional activation or repression by RNA polymerase II. In human macrophage differentiation, PU.1 binding-site selection depends on sequence affinity, PU.1 concentration, cooperative transcription-factor interactions, and chromatin accessibility [PMID:23658224 "PU.1 selects its binding sites primarily based on sequence affinity"; PMID:23658224 "PU.1 binding control that involves motif-binding affinity, PU.1 concentration, cooperativeness with neighboring transcription factor sites and chromatin domain accessibility"]. PU.1 occupied sites are enriched in active chromatin domains with nearby cooperative motifs PMID:23658224.
PU.1 has pioneer/master-regulator biology in hematopoiesis. A human immunodeficiency study states that PU.1 controls hematopoietic cell fate by decompacting stem-cell heterochromatin and opening otherwise inaccessible genomic sites PMID:33951726. Patient and genome-editing evidence showed SPI1-dependent B-cell, dendritic-cell, and myeloid development: affected patients lacked B cells and had few conventional dendritic cells, disease-similar SPI1 mutations impaired early B-cell and myeloid differentiation, and mutant proteins failed to nuclear-localize or bind target DNA [PMID:33951726 "Affected patients lacked circulating B cells and possessed few conventional dendritic cells"; PMID:33951726 "mutations into human hematopoietic stem and progenitor cells impaired early in vitro B cell and myeloid cell differentiation"; PMID:33951726 "Patient SPI1 mutations encoded destabilized PU.1 proteins unable to nuclear localize or bind target DNA"].
SPI1 is relevant to microglia/macrophage biology through its lineage role rather than a narrow Alzheimer-specific effector mechanism. Human iPSC-derived monocyte/macrophage experiments showed SPI1 dependence and related those cells to tissue-resident macrophages including microglia [PMID:28111278 "human iPSC-derived monocytes/macrophages develop in an MYB-independent, RUNX1-, and SPI1 (PU.1)-dependent fashion"; PMID:28111278 "a good model for MYB-independent tissue-resident macrophages, such as alveolar and kidney macrophages, microglia, Kupffer cells, and Langerhans cells"]. This supports retaining myeloid/macrophage lineage annotations as important but generally non-core relative to the defining transcription-factor and chromatin-regulatory function.
For curation, keep DNA-binding transcription-factor activity, cis-regulatory region binding, chromatin binding, nuclear/chromatin/nucleoplasm localization, RNA polymerase II transcription regulation, and transcription initiation-coupled chromatin remodeling as core. Keep hematopoietic lineage outcomes, macrophage/myeloid differentiation, B-cell differentiation, endothelial-to-hematopoietic transition, dendritic-cell chemotaxis, and microglial/immune effector terms as non-core developmental or cell-type outcomes unless they directly describe PU.1's transcriptional control mechanism. Generic protein binding should be modified to specific transcription-factor binding, STAT-family protein binding, histone deacetylase binding, NFAT binding, or marked over-annotated when the original evidence is a generic interaction screen.
The second-pass audit left the two remaining generic protein binding IPI annotations as UNDECIDED rather than removing them. PMID:21575865 is cached as abstract-only and foregrounds TMPRSS2:ERG interaction with PARP1/DNA-PKcs in ETS fusion-positive prostate cancer, not a directly visible SPI1 assay PMID:21575865. PMID:10207087 is likewise abstract-only in the cache and describes MEF/AML1 interactions, not an exposed SPI1 interaction PMID:10207087. Under the project rule not to overrule experimental annotations from incomplete abstract-only evidence, both calls should remain unresolved pending full-text or supplementary-data review.