CSTF3

CSTF3 (CstF-77) is the central scaffold subunit of the cleavage stimulation factor (CstF) complex, essential for pre-mRNA 3′ cleavage and polyadenylation. It forms an elongated homodimer through its HAT-C domain and bridges CstF-64 via its C-terminal region and CstF-50 via other interfaces; the C-terminus of CstF-77 stabilizes the CstF-64 RNA recognition motif and enhances RNA binding, and the CstF-64–CstF-77 interaction is required for nuclear import of the preformed CstF complex (PMID:17386263, PMID:17584787, PMID:30257008, PMID:21119002). CstF-77 expression is autoregulated through a negative-feedback intronic polyadenylation mechanism that generates non-coding isoforms, and perturbation of CstF-77 levels causes widespread changes in alternative polyadenylation with consequences for cell proliferation and differentiation (PMID:23874216, PMID:16316725). In specific signaling contexts, such as integrin α3β1–MEK/ERK signaling in keratinocytes or in ovarian cancer, CSTF3 levels control proximal versus distal polyadenylation site selection on target transcripts including Mmp9 and NEAT1, thereby influencing mRNA isoform stability, platinum drug resistance, and wound-healing responses (PMID:38898019, PMID:41628695).

1. 2002 High

   Genetic rescue in Drosophila established that the CstF-77 ortholog Su(f) is required for pre-mRNA 3′ cleavage in vivo, proving that CstF-77 function in mRNA 3′ end formation is conserved across metazoans and that species-specific subunit interactions constrain interchangeability.

   Evidence Chimeric human CstF-77/Su(f) rescue of su(f) lethal mutants in Drosophila

   PMID:12149458

   Open questions at the time

   • Specific residues mediating species-specific CstF subunit interactions were not mapped
   • Whether CstF-77's role extends beyond cleavage to polyadenylation stimulation was not tested in this system
2. 2005 Medium

   Discovery of a conserved intronic polyadenylation site in the CSTF3 gene itself revealed that short, non-functional transcripts are produced across vertebrate tissues, foreshadowing an autoregulatory mechanism.

   Evidence RT-PCR and SAGE data analysis of human and mouse CSTF3 loci

   PMID:16316725

   Open questions at the time

   • No perturbation experiment demonstrated that intronic PAS usage is regulated by CstF-77 protein levels
   • Functional significance of the short isoforms was not addressed
3. 2007 High

   Crystal structures of CstF-77 from mouse and Encephalitozoon cuniculi revealed the HAT-repeat architecture, demonstrated that homodimerization is mediated by the HAT-C subdomain producing a 165 Å elongated dimer, and identified the C-terminal region as the CstF-64 docking site, providing the first structural framework for CstF complex assembly.

   Evidence X-ray crystallography (2 Å resolution), analytical ultracentrifugation, yeast two-hybrid, light scattering

   PMID:17386263 PMID:17584787

   Open questions at the time

   • No structure of the full trimeric CstF complex or the CstF-77–CstF-50 interface
   • Structural basis of how dimerization couples to RNA recognition was not resolved
4. 2009 High

   Demonstration that the CstF-64 hinge domain is essential for CstF-77 binding and that this interaction governs nuclear import of the preformed CstF complex established that CstF assembly and nuclear localization are coupled prerequisites for polyadenylation.

   Evidence In vivo SLAP polyadenylation assay, domain deletion/mutation, co-IP, and nuclear localization assays in mammalian cells

   PMID:19887456

   Open questions at the time

   • The nuclear import receptor or NLS responsible for CstF complex import was not identified
   • Whether CstF-50 participates in import was not tested
5. 2010 High

   Showing that CstF-77 and symplekin bind mutually exclusively to CstF-64's hinge domain clarified that CstF-64 partitions between distinct complexes, with CstF-77 binding being the route to nuclear accumulation and stoichiometric CstF assembly.

   Evidence Mutant interaction studies, co-IP, nuclear localization assays

   PMID:21119002

   Open questions at the time

   • Functional consequence of CstF-64–symplekin complex in 3′ processing versus the CstF-64–CstF-77 complex was not delineated
   • Stoichiometry of CstF subcomplexes in the nucleus versus cytoplasm was not quantified
6. 2013 High

   Establishing that CstF-77 autoregulates its own expression through a negative-feedback intronic polyadenylation loop — and that CstF-77 perturbation causes genome-wide alternative polyadenylation changes affecting proliferation and differentiation — revealed CstF-77 as both a target and a regulator of the APA landscape.

   Evidence Reporter assays, RT-PCR, CstF-77 overexpression/knockdown, U1 snRNP inhibition, cell proliferation assays

   PMID:23874216

   Open questions at the time

   • Identity of CstF-77-dependent APA target genes driving proliferation/differentiation phenotypes was not catalogued genome-wide
   • Whether other CstF subunits participate in the feedback loop was not resolved
7. 2018 High

   NMR and functional studies revealed that the last 30 amino acids of CstF-77 directly stabilize the CstF-64 RRM and modulate its RNA-binding properties, providing a molecular explanation for how CstF-77 allosterically enhances polyadenylation site recognition.

   Evidence NMR spectroscopy of recombinant CstF-77 C-terminal peptide with CstF-64 RRM, reverse genetics, nuclear localization assays

   PMID:30257008

   Open questions at the time

   • How RRM stabilization translates to PAS selectivity in vivo was not tested
   • Whether CstF-77 similarly modulates the paralog CstF-64τ (CSTF2T) was not examined
8. 2024 Medium

   CSTF3 was shown to directly bind downstream of the NEAT1 proximal PAS to control NEAT1_1/NEAT1_2 isoform balance, linking CstF-mediated APA to platinum drug resistance in ovarian cancer via PI3K/AKT/mTOR signaling.

   Evidence CSTF3 knockdown, RNA isoform analysis, luciferase assays, scRNA-seq APA profiling in ovarian cancer cells

   PMID:38898019

   Open questions at the time

   • Direct binding of CSTF3 at NEAT1 PAS was inferred from KD rather than CLIP or biochemical binding assays
   • Whether the PI3K/AKT/mTOR link is direct or mediated by NEAT1_1-dependent gene programs is unresolved
   • Single-lab finding not yet independently replicated
9. 2026 Medium

   An integrin α3β1–MEK/ERK signaling axis was found to induce CSTF3 transcription in keratinocytes, driving proximal PAS usage on Mmp9 (producing a more stable short isoform) and genome-wide APA shifts, thereby connecting extracellular matrix signaling to CSTF3-dependent post-transcriptional regulation in wound healing.

   Evidence Epidermis-specific α3 knockout mice, RNA in situ hybridization, CSTF3 KD, DaPars2 genome-wide APA analysis

   PMID:41628695

   Open questions at the time

   • Whether MEK/ERK directly activates the CSTF3 promoter or acts through intermediate transcription factors is unknown
   • In vivo functional rescue of wound-healing phenotype by CSTF3 re-expression was not performed
   • Single-lab finding; recently published

• No high-resolution structure of the full trimeric CstF complex exists, the import receptor mediating CstF nuclear entry is unidentified, and the rules by which CstF-77 levels selectively reprogram PAS choice on specific transcripts genome-wide remain undefined.
• No full CstF trimer structure or CstF-77–CstF-50 interface structure
• Nuclear import receptor/mechanism for the CstF complex is unknown
• Determinants of target-specific PAS selection by CstF-77 are unresolved