PDF, English - main document Restricted access: Repository staff only until 1 March 2025. Login+Download (27MB) | Terms of use

Abstract

All cells in a multicellular organism contain the same genetic information, yet individual cell types express different genes in order to fulfil their specific func- tions. Expression patterns are regulated dynamically throughout the cell cycle and in response to internal and external stimuli. Improper regulation of gene expres- sion patterns is associated with diseases such as cancer, where developmental gene expression profiles are often hijacked. The serine - arginine rich splicing factor (SR) proteins are a conserved family, consisting of 12 members in humans. They were first described as splicing factors, promoting the inclusion of weak splice sites and shown to work as a network, regulat- ing their own and each others expression in a cell type dependent manner. Various SR proteins have also been shown to have additional RNA processing functions, including regulation of transcription, translation initiation and non-sense mediated decay. Other than splicing the precise mechanisms by which SR proteins modulate gene expression were unknown. My study focused on family member SRSF2 and a recurrent mutant form of SRSF2 (P95H) commonly found in cancer. I directly compared the effects of conditional deletion or mutation of SRSF2 in mouse skin. Both perturbations caused a loss of cellularity and cell cycle arrest accompanied by impaired differentiation and increased DNA damage. I showed that SRSF2 acts as an important regulator of transcription and that the P95H mutation results in a loss of protein function. To decipher the molecular mechanisms of SRSF2 action it was knocked down in primary human keratinocytes, an established in vitro model for normal skin, and two squamous cell carcinoma cell lines. In all cell lines analysed SRSF2 depletion caused cell cycle arrest and increased DNA damage. In an attempt to look at the relative contributions of splicing and transcription both processes were analysed. No conserved changes to splicing were observed across cell lines, while analysis of transcription rates via SLAM-sequencing and RNA polmerase II (Pol II) occupancy via Cut&Run experiments revealed that depletion of SRSF2 was associated with reduced global nascent transcription due to enhanced stalling of Pol II at transcrip- tion start sites, suggesting a direct effect on transcription is the main cause of the effects of SRSF2. DNA replication and repair associated genes were most affected by loss of SRSF2 due to their specific genomic organisation as bi-directional gene pairs. These results provide insights into why SRSF2 is an essential gene, but how mu- tant forms of SRSF2 contribute to carcinogenesis remained unexplained. Therefore I generated cancer cell clones expressing reduced levels of SRSF2 using CRISPR-cas9 mediated gene editing approaches. Cancer cells with reduced SRSF2 function can cycle but accumulate mutations faster and thereby enable faster tumour evolution. Introducing the Srsf2 P95H mutation into one allele of the mouse skin inhibited tumour formation following exposure to carcinogens. However, within SRSF2 P95H established clones cell growth was not affected and clone size tended to be larger than that of control cells. In summary, my studies showed for the first time that the P95H mutation causes a loss of function and that SRSF2’s role in transcription, and not splicing, is required for proper expression of DNA replication and repair genes due to their organisation as bi-directional gene pairs.