The myeloid transcription factor cebpa functions as a sequence-specific RNA binding protein via a novel n-terminal domain essential for macrophage differentiation Free

Introduction: CCAAT/Enhancer Binding Protein Alpha (CEBPa) is a master regulator indispensable for neutrophil development, yet its precise role in macrophage maturation has remained elusive. This functionality has been historically attributed exclusively to its DNA-binding bZIP domain in the C-terminus. CEBPa's critical hematopoietic role is extensively highlighted in developmental hematology and is disrupted by recurrent inactivating mutations in Acute Myeloid Leukemia. We challenged CEBPa's canonical binding model and here report CEBPa is a bona fide sequence-specific RNA-binding protein (RBP) and demonstrated this non-classical function is mediated by distinct N-terminus RNA binding domains which are essential for executing the terminal macrophage differentiation program.

Methods: We performed unbiased RNA Immunoprecipitation sequencing (RIP-Seq) in HL-60 and THP-1 human myeloid cells on isolated nuclei. The mechanisms of interaction were dissected using assays for direct binding (RNA EMSA (REMSA) and Isothermal Titration Calorimetry (ITC); in-cell validation (CLIP-qRT-PCR); and structural characterization (NMR). The in vivo requirement for RNA binding was tested in a murine CEBPa conditional knock-out bone marrow transplantation model in which Cebpa-null hematopoietic stem and progenitor cells were rescued with retrovirally-expressed wild-type or RNA-binding-deficient CEBPa. Reconstitution was assessed by flow cytometry and population profiling was performed using 10x single-cell RNA/CITE-sequencing.

Results: RIP-Seq revealed CEBPa associates with RNA primarily within introns of key hematopoietic regulators, including PU.1 and CTBP1, indicating interaction with un-spliced, CA-rich pre-mRNA. We demonstrate this interaction is independent of the DNA binding bZIP domain, which showed no affinity for RNA. Instead, RNA binding is mediated by novel, previously unreported domains. REMSA and deletion analysis mapped the core RBDs to two short segments situated in the CEBPa N-terminal region: aa. 70-97 and aa. 159-182, which are functionally and spatially distinct from CEBPa's C-terminal bZIP domain. ITC confirmed a direct, micromolar affinity interaction (Kd = 5.8 µM for aa. 70-182), and NMR analysis of chemical shifts upon RNA addition corroborated this direct binding in vitro. In vivo deletion of these RBDs selectively blocked terminal monocyte-to-macrophage differentiation, evidenced by the absence of mature F4/80⁺ macrophages, while leaving neutrophil development largely intact. Single-cell transcriptomics pinpointed this defect to a specific maturation checkpoint, revealing an accumulation of an aberrant CD11b-negative, LPL-high monocytic population in mice reconstituted with RBD deleted CEBPa; a population whose transcriptome mirrors that of lipid-associated macrophages (LAMs), implicating aberrant lipid metabolic programming in a reported aggressive and invasive cell population. Mechanistically, ChIP-qPCR demonstrated that CEBPa's RNA-binding activity is essential for its efficient recruitment to bind to DNA sites in the PU.1 upstream regulatory enhancer. We propose this is likely mediated by binding to the enhancer-associated lncRNA, LOUP, which is implicated by competition assays and motif analysis. Disruption of this CEBPa-LOUP RNA axis impairs PU.1 activation leading to low PU.1 levels, hindering subsequent macrophage development and differentiation.

Conclusion: Our findings redefine CEBPa as a dual-specificity transcription factor that interacts with both DNA and RNA to orchestrate gene expression via distinct functional domains. This differs from reports of other transcription factors binding to RNA through their DNA binding domains. Furthermore, Alphafold3 structural modelling predicts CEBPa binds RNA through its canonical bZIP domain. Therefore, the discovery of novel, functional, N-terminal RBDs in CEBPa that are essential for macrophage differentiation provides a new paradigm for understanding myeloid development and identifies a class of transcription factors with distinct RNA binding domains that structural modelling cannot yet accurately predict. This work transforms our fundamental knowledge of a key hematopoietic regulator and uncovers a new layer of gene regulation that may be disrupted in AML, in which high expression of the CEBPa p30 isoform, lacking one of the amino terminal RNA binding domains, is observed.