The bulk of the actions carried out in the MASCP project can be summarized as follows:
a) Building transcriptome-wide splicing regulatory networks. This effort represents the core of the MASPC project, as it serves as the basis for reconstructing maps of functional interactions between components of the spliceosome. RNAs corresponding to the knock down of each individual component of the spliceosome (both Major and Minor spliceosomes), splicing regulatory factors, RNA modifying enzymes and chromatin remodelling factors influencing RNA processing in a cancer cell line (HeLa, 304 knock downs in total) were isolated and subject to paired-ended RNA sequencing. Sequencing (100 M reads/sample) was completed and datasets were organized, merged, aligned and, after quality-control, alternative splicing was quantified using novel computational pipelines. Special efforts were made to optimize the datasets to be able to later on integrate RNA-seq data from a collection of knock downs corresponding to every protein-coding gene in the human genome, thus facilitating the investigation of links between splicing regulation and virtually any other cellular pathway. Regulatory networks were generated for different types of alternative splicing events and for subclasses of these depending on architectural features of the genomic regions, regulatory factors, cancer associations, etc. An unexpected conceptual complication emerged when a global picture of the effects of knock down of splicing factors on other components of the spliceosome was obtained: 50 - 60% of the knock downs affected the expression or alternative splicing of other splicing factors; while this introduces intrinsic complexity in the interpretation of the results of these genetic perturbations -making it necessary to adopt a novel conceptual framework for data analysis-, at the same time the results reveal a naturally existing, exquisitely interwoven, regulatory network which is one first major insight of our systems biology approach to alternative splicing.
b) Detailed analysis of regulatory mechanisms, including the establishment of methods for high-throughput analysis of regulatory sequences and factors. In parallel with the systematic sequencing efforts above, mechanistic studies were carried out to understand the function of sequences and regulatory factors identified in our previous functional network studies, including:
b.1. Regulation of splice site recognition. We have developed methods to systematically assess the effects of nucleotide substitutions in alternatively spliced exons. The results reveal high dense regulatory content of alternative exons and extensive epistatic interactions in exon definition, as well as a general Scaling Law that explains the impact of nucleotide substitutions depending on the initial level of exon inclusion. Work on particular subnetworks involved in 3' splice site recognition led to the characterization of a complex essential for cell cycle progression and another complex of proteins interacting with the factor ZRSR1 that plays a role in mouse spermatogenesis.
b.2. Modulation of splicing by core components of the splicing machinery. Structure/function analyses revealed that the splicing regulatory factor RBM5 contacts the core splicing factor SmB/B'/N to modulate alternative splicing, revealing the potential of core components of the Sm complex for influencing splice site selection. Work on small molecules targeting the key core splicing factor SF3B1 provided compounds of improved stability and activity, as well as dissected 3' splice site sequences that modulate drug response, revealing also significantly different splicing regulation effects by structurally different drugs.
b.3. Integration of alternative splicing with cellular programs of cell proliferation and pluripotency. Work in vitro and in xenograft tumors demonstrated that cancer-associated mutations in the splicing regulator RBM10 switch its activity from tumor suppressor to an oncogene. Based upon these mechanistic insights, antisense oligonucleotides (AONs) chemically modified to increase their selectivity and stability in vivo have been designed and validated in mouse models of lung adenocarcinoma with encouraging results. More generally, bioinformatic analyses and in vitro validation uncovered extensive networks of alternative splicing regulation in cancer, the impact of the SPF45 complex on cell cycle progression and insights into the mechanisms by which responses to cellular stress relies on alternative splicing changes mediated by factors phosphorylation by MAPK signaling cascades. Work on reprogramming of B cells in mouse revealed key alternative splicing events and regulatory factors contributing to efficient reprogramming by the transcription factor CEBPa.
