YB-1 binds the 3' splice site of EGFR exon 3in glioblastoma cells, but a YB-1 decoycannot act on EGFR alone
RNA oligonucleotides directed at a splice site can suppress a gene without touching its protein, and a companion preprint reports that oligonucleotides against the 3' splice site of EGFR exon 3 reduce proliferation of A172 glioblastoma cells. That work rests on a SpliceAid prediction of a Y-box binding protein 1 (YB-1, YBX1) site at the targeted sequence, which has never been tested. This study tests three premises of that design against deposited CLIP, RIP, knockdown RNA-seq and splice-junction data. First, whether exon 3 is a defensible target among the 26 internal exons of EGFR. Applying four criteria in sequence - encoding sequence upstream of the kinase domain, frameshifting when skipped, lying within the first 200 codons, and carrying the YB-1 CA(U/C)C core at its 3' splice site - leaves exactly one exon, exon 3. The frameshift requirement is not cosmetic: EGFRvIII, the in-frame deletion of exons 2 to 7 (801 nt, 267 codons), is constitutively active, so inducing an in-frame skip in EGFR risks producing a signalling-competent receptor. Second, whether YB-1 binds there. A deposited YBX1 eCLIP peak from U251 glioblastoma cells (GSE213046) maps to chr7:55,143,305-55,143,308, the first four nucleotides of exon 3, immediately 3' of the splice acceptor AG (p = 1.3 x 10^-3, log2FC = 3.29). Six YBX1 CLIP datasets from non-glial cells also detect EGFR binding, but not at this position: across the four with mapped site coordinates, none of 103 sites overlaps exon 2, intron 2 or exon 3, and exonic binding is concentrated in exon 28, the 3'UTR. The U251 dataset differs in that its two most 5' exonic peaks lie in exon 1 and in exon 3. Third, whether a decoy carrying the YB-1 consensus can act specifically on EGFR. It cannot. In U251, YB-1 binding sites fall in 8,801 genes, 7,840 of which are expressed there - 42.8% of the expressed transcriptome - and the CA(U/C)C core occurs 48 times in the EGFR mRNA alone. Across four glioblastoma YBX1-knockdown RNA-seq datasets, EGFR mRNA changes by |log2FC| 0.23 to 0.59 and rises in three of the four, ranking 2,418th to 16,274th by effect size. Public U2AF65 iCLIP adds a constraint on mechanism. The 30 nt upstream of the exon 3 acceptor carry 92 crosslinks in pooled HeLa data, ranking tenth among the 26 internal exons of EGFR, whereas the CD44 variable exon v5 tract, on which the YB-1/U2AF65 mechanism was established, carries 15. The EGFR tract is not short of U2AF65, and the two proteins do not share it: the U2AF65 footprint runs from -21 to -3 while the YB-1 peak sits at +1 to +4, across the acceptor and inside the exon. The route by which YB-1 is known to act at a 3' splice site therefore does not obviously apply here. Separately, skipping of exon 3 shifts the reading frame and generates a premature stop 2,771 nt upstream of the final exon junction, making the transcript a canonical nonsense-mediated decay substrate. Across 43 public RNA-seq samples the skipping junction carried 4 reads against approximately 1,836 inclusion reads (PSI ~ 0.998), all 4 in a glioblastoma line. Together these results support YB-1 contact at the exon 3 3' splice site in glioblastoma specifically, provide a reproducible rationale for that target, and set a quantitative bound on what a YB-1 decoy can be claimed to do.
Authors
- Kenji Ohe (ORCID: https://orcid.org/0000-0002-1958-7572)
- Kyungseung Kwon
Institutions
- Fukuoka University (JP)
Publication Details
- Journal
- Research Square
- Published
- 2026-11-19
- DOI
- https://doi.org/10.21203/rs.3.rs-11720196/v1
- Primary Topic
- RNA Research and Splicing
- Type
- preprint