Splicing factors are frequently altered by mutations and copy-number changes both in cancer and in
germline genetic diseases resulting in multi-system developmental syndromes. Despite the fact that virtually all
genes in humans undergo splicing, spliceosomal genetic alterations tend to exhibit surprisingly specific effects
on subsets of splicing events, leaving most insignificantly changed. These effects can be allele-specific, cell-type
specific, and dependent on the genetic background of the afflicted cell. This makes it especially challenging to
determine which affected splicing events contribute to disease etiology. The fact that a limited set of introns is
responsive to any specific splicing factor alteration indicates that introns and their flanking exons have evolved
in structure and sequence to confer differential sensitivity to the action of different spliceosome components.
This raises a fundamental question: what are the features common to sets of introns that confer this specificity?
Using naturally occurring splicing gene mutations, amplifications, and deletions, these perturbations will be
modelled in a genetically stable, untransformed, isogenic cell system where it is possible to isolate the effect of a
single alteration on the transcriptome and on the binding patterns of the altered protein. These studies will shed
light on the mechanisms of normal spliceosome function, and provide insight into which genes and biological
pathways affected by splicing dysfunction likely contribute to disease states.
The proposed experiments will employ three distinct methods to model spliceosome perturbations
associated with human disease, with a focus on factors that physically or functionally interact with the essential
spliceosome protein SF3B1. (Specific Aim 1) Introduction of an allelic series of cancer-associated SF3B1
missense-mutations into isogenic cell lines using recombinase-mediated cassette exchange (RMCE); (Specific
Aim 2) CRISPRa/i-mediated activation or inhibition of transcription to up- or down- regulate splicing factors
that are amplified in cancers (PUF60, SF3B4, and U2AF2) and lost in developmental syndromes (PUF60,
SF3B4); and (Specific Aim 3) rapid depletion of spliceosomal RNA helicases (DDX39B, DDX46, and DHX16)
and their putative co-factors (SUGP1, RBM17, and GPKOW) at the protein level using auxin-inducible degrons.
Three distinct methods of RNA sequencing will be used to quantify the changes resulting from these
perturbations: poly(A)-selected RNAseq, allele-specific eCLIP, and a novel intron lariat capture sequencing
approach. Lastly, we will integrate these genomic data sets into models using deep learning neural networks to
interrogate our central hypothesis: the sequence and structure of individual mammalian introns have evolved to
confer differential dependence on specific ‘core’ components of the spliceosome, and that mutations,
amplifications, and deletions in these core components causal for human disease will uncover intron-centric
gene expression regulatory circuits that are controlled though modulation of the abundance or activity of the
associated splicing factors in normal cells.
