Programmed Delayed Splicing: A Mechanism for Timed Inflammatory Gene Expression 2267628

Abstract

Abstract Introduction Inflammation involves timed gene expression, suggesting that the fine-tuned onset, amplitude, and termination of expression of hundreds of genes is of critical importance to organismal homeostasis. Recent study of post-transcriptional regulation of inflammatory gene expression led to the suggestion of a regulatory role for pre-mRNA splicing. Methods Here, using a hybrid capture approach to purify incompletely spliced, chromatin-associated pre-mRNAs, we use deep sequencing to study pre-mRNA splicing of the NF-kB transcriptome. To broaden this mechanistic view, we applied deep learning sequence-to-function models (BORZOI) with feature attribution to identify additional regulatory sequences–both intronic and exonic–that may contribute to delayed splicing through mechanisms independent of donor site strength. Results By freezing transcription and examining subsequent splicing of complete transcripts, we find many introns splice tens to hundreds of times slower than average. Investigating the basis of these delays, we focused on evolutionarily conserved introns with suboptimal splice donor sequences and found that strengthening these donor sites by as few as two nucleotides in minigene reporter assays markedly increased gene expression for several targets. This suggests that such sites can act as timing elements that both delay mRNA production and limit expression amplitude. Deep learning approaches with a foundational genomic model revealed non-canonical motifs enriched in slow-splicing introns, pointing to a broader repertoire of cis-elements that can fine-tune transcript maturation during inflammation. Conclusion Together, these findings support a model in which the temporal regulation of pre-mRNA splicing serves as a layer of control in inflammatory gene expression, and raise the possibility that similar timing mechanisms operate in other rapid-response transcriptional programs. Funding Source n/a Topic Categories Computational and Systems Immunology (COMP)