CPSF2

CPSF2 (CPSF100) is a catalytically inactive metallo-beta-lactamase (MBL)-fold subunit of the cleavage machinery for both histone and polyadenylated pre-mRNA 3'-end processing (PMID:19450530, PMID:18688255). Although its own histidine motif is degenerate and it cannot cleave RNA, conserved residues within its MBL motifs are required to build the active endonuclease together with CPSF73, contributing structurally in a manner analogous to RNase Z and RNase J (PMID:18688255). Structurally, the C-terminal domains of CPSF100 and CPSF73 form a stable heterodimer through extensive inter-protein contacts, and CPSF73's CTD3 binds Symplekin, defining the trimeric core cleavage complex that, together with histone-specific factors, is required for 3'-end processing and associates co-transcriptionally with histone genes (PMID:19450530, PMID:37989222). Beyond canonical 3'-end formation, CPSF2 participates in alternative cleavage of immediate early genes through a serum-stimulation-dependent interaction with the TREX component THOC5, which recruits CPSF2 to 3'-UTR targets including Myc and Smad7 (PMID:25274738), and acts as a direct RNA-binding cofactor influencing genome-wide alternative splicing (PMID:26697379).

1. 2005 Medium

   Established that CPSF2 has a nuclear-restricted localization and can engage processing partners outside the canonical CPSF complex, raising the possibility of substrate-specific 3'-end processing roles.

   Evidence Co-IP, subcellular fractionation, and RNAi with cell-cycle readout in HeLa cells

   PMID:15684398

   Open questions at the time

   • The reported lack of interaction with CPSF73/CPSF160 conflicts with later heterodimer structures and was not reconciled
   • Functional substrate set of the RC-68 complex not defined
   • Single-lab observation without reciprocal validation
2. 2008 High

   Resolved how a catalytically dead subunit contributes to catalysis by showing CPSF100's MBL motifs are structurally required to assemble the active CPSF73 endonuclease.

   Evidence In vitro MBL active-site mutagenesis of both subunits with reconstituted histone pre-mRNA cleavage assay

   PMID:18688255

   Open questions at the time

   • Does not resolve the atomic geometry of the assembled active site
   • Contribution to polyadenylated substrate cleavage tested less directly
3. 2009 High

   Defined the CPSF100-CPSF73-Symplekin core as a stable subcomplex required for both histone and polyadenylated pre-mRNA processing and showed it associates co-transcriptionally with target genes.

   Evidence Reciprocal co-IP, RNAi knockdown with processing phenotype, and ChIP in Drosophila

   PMID:19450530

   Open questions at the time

   • Mechanism distinguishing histone versus polyadenylated substrate routing not defined
   • Stoichiometry and assembly order of the trimeric core not established
4. 2014 Medium

   Identified a stimulus-dependent route for CPSF2 recruitment, linking it to TREX-mediated 3'-end processing of immediate early genes.

   Evidence Co-IP interactome with THOC5 as bait, RNA/chromatin IP, and THOC5 RNAi with transcriptome analysis

   PMID:25274738

   Open questions at the time

   • Whether the THOC5 interaction is direct is not established
   • How serum stimulation triggers complex formation is unknown
   • Single lab
5. 2014 Low

   Suggested a cellular consequence of CPSF2 loss in cancer, linking depletion to increased invasion and stem cell marker expansion.

   Evidence RNAi knockdown in thyroid cancer cell lines with invasion assay and stem cell marker flow cytometry

   PMID:24654752

   Open questions at the time

   • Phenotypic knockdown without molecular mechanism or pathway placement
   • No connection drawn to 3'-end processing function
   • Single lab, no in vivo validation
6. 2015 Medium

   Extended CPSF2 function beyond 3'-end cleavage by mapping its direct RNA contacts and showing it influences genome-wide alternative splicing as an RBFOX2 cofactor.

   Evidence iCLIP for direct RNA targets paired with RNA-seq after CPSF2 knockdown

   PMID:26697379

   Open questions at the time

   • Direct physical RBFOX2 interaction not biochemically detailed in abstract
   • Mechanism coupling splicing regulation to 3'-end processing role unknown
   • Single lab
7. 2023 High

   Provided the molecular architecture of the core cleavage complex by solving the CPSF73-CPSF100 C-terminal heterodimer and defining the CPSF73 surface that binds Symplekin.

   Evidence NMR solution structure of the minimal C-terminal heterodimer from E. cuniculi with biochemical Symplekin binding assays and comparative modeling

   PMID:37989222

   Open questions at the time

   • Structure of the full-length assembled endonuclease on RNA substrate not determined
   • Solved in a microsporidian ortholog rather than the human complex

• How CPSF2's distinct activities — core endonuclease assembly, THOC5-dependent immediate early gene processing, and RBFOX2-associated splicing regulation — are coordinated and substrate-selected remains unresolved.
• No structure of the catalytically active complex engaged with substrate RNA
• Mechanism selecting histone versus polyadenylated versus alternatively spliced targets unknown
• Direct versus indirect nature of THOC5 and RBFOX2 associations unsettled