SF3B1 Mutations Lead to Changes in Fine-Scale Chromatin Organization through Impaired Transcription

Introduction: Traditionally studied for their role in regulating core and alternative splicing, splicing factors (SFs) also regulate transcription kinetics (Chen et al, Mol Cell 2020, Boddu et al, Mol Cell 2024). Clonogenic mutations in SF3B1 reduce promoter-proximal (p-p) RNA Polymerase II (Pol II) density, leading to a closed promoter chromatin architecture (Boddu et al, Mol Cell, 2024). Transcription is integral to fine-scale genomic organization, with p-p RNA Pol II density stabilizing enhancer-promoter (EP) interactions (Barshad et al., Nat Genet, 2023). The role of transcription perturbations on higher order genome organization in driving pathologic states remains unexplored.

Methods: We optimized a low-input Hi-C protocol (tagHiC; Zhang et al, Cell Rep 2020) in murine hematopoietic progenitors. Approximately 200,000 committed progenitor cells (lin- c-Kit⁺ Sca-1^low FcγR^high/low) were isolated from bone marrow of Sf3b1^+/+ (WT) or Sf3b1^+/K700E (Sf3b1^K700E) mice, by FACS-sorting. TagHi-C libraries were prepared and sequenced to ~700 million reads per sample, generating over 100 million valid junction reads, sufficient for 20-40 kb resolution. Topology-associated domains (TADs) and EP interactions were determined using HiC-Pro and HiC Explorer packages. TAD-separation scores (TAD-SSs), a measure of integrity of TAD boundaries was determined. ChIP-seq datasets against Pol II in lineage-depleted bone marrow cells were used to determine concurrent changes to p-p Pol II in Sf3b1^K700E. Traveling ratio ((TR), calculated as the ratio of Pol II reads in promoter to gene body) was used to determine changes to Pol II distribution in Sf3b1^K700E.

Putative enhancer regions were identified from ENCODE H3K27ac ChIP-seq data. Enhancers within 50 kbp of their neighboring promoters were linked as EP pairs. EP interaction frequencies were calculated from ICE-balanced Hi-C matrices by converting E & P genomic coordinates to bin indices, extracting interaction frequencies, and normalizing to the total contact frequency count in each sample. For promoters with multiple EP pairs, the frequencies of all EP pairs were summed to generate cumulative frequencies for each promoter.

Results: We identified 4,957 gene promoter regions, each linked to at least one enhancer, with a median of 3 putative enhancers per promoter (range 1-25). SF3B1 mutant cells were noted to have significantly reduced normalized EP contact frequencies (median of 8.2 versus 14.3 in WT, p<0.00001). Genes with high p-p Pol II pausing (noted by higher TR) at baseline were more susceptible to such loss of EP interactions in mutant cells (KS test, D=0.03, p<0.01). This suggests that these genes rely on a paused p-p Pol II state to stabilize enhancer-promoter interactions. Correlation analysis between the degree of TR loss (WT - mutant) and the degree of EP interaction loss (WT - mutant) revealed a left-skewed distribution, indicating that as the loss of p-p Pol II density increases, there is a corresponding increase in the loss of EP interactions. 700 genes (14% of the 4957 genes) exhibited a concurrent loss of EP interactions and reduction in p-p Pol II density in the Sf3b1^K700E. This suggests that EP interactions of these genes are critically dependent on paused Pol II at their promoters. Chromatin compartments often house genes in close proximity than can change their behavior as a group due to shared regulatory elements (Lucini et al., NAR 2024). We, therefore, analyzed the pattern of gene density distribution across 1 mega base pair sized bins across the genome. This showed 19 genomic bins enriched for such genes (5 or more genes per bin). We found no significant difference in TAD-SSs between WT and Sf3b1^K700E samples across 19 genomic bins. These findings indicate that the loss of EP interactions in these closely clustered genes are fine-scale and not associated with changes in overall TAD compartmentalization.

Conclusions: Here, we extend our previous findings that SF3B1 mutations alter chromatin landscape through impaired Pol II transcription, showing that higher-order chromatin organization changes in response to such mutations. Specifically, EP interactions are reduced without disrupting TAD domains, indicating that changes in p-p Pol II density affect gene regulation at a fine-scale. Ongoing work will explore the functional consequences of EP interaction loss on gene expression and clonal behavior to guide therapeutic strategies.