Regulatory consequences of variation in transcription factor-DNA binding

Abstract

Transcription factors bind to their target DNA sites in a sequence-specific manner, modulating gene expression, and playing important roles in development, body patterning, cellular differentiation, and stimuli response. Missense mutations in transcription factor DNA-binding domains have been linked with a myriad of Mendelian diseases, while most variants identified through genome-wide association studies of human disease risk and trait variation lie within non-coding loci, and potentially overlap and perturb transcription factor binding sites. The goal of my dissertation is to improve variant interpretation and our understanding of disease mechanisms by investigating the regulatory consequences of both coding and non-coding variation in transcription factor-DNA binding.

I first discuss our study of homeodomain variants. Every single position within the homeodomain DNA-binding domain overlaps pathogenic variants and variants of uncertain significance, motivating our broad overview of the DNA-binding consequences of missense variation across the DNA-binding domain. In addition to elucidating stark alterations in DNA-binding properties linked with disease-associated variants, I also identified variants with subtler changes in DNA-binding specificity. I found 9 novel DNA-binding specificity-determining positions, several distal from homeodomain-DNA interfaces, and suggest possible mechanisms for such alterations in DNA-binding specificity.

I then discuss our generation and profiling of a biochemical and cellular allelic series of KLF1 transcription factor variants, which have been associated with a spectrum of red blood cell disorders, and span a range of DNA-binding affinity and / or specificity differences. I integrate protein binding microarray data with CUT&RUN, ATAC-seq, and RNA-seq data from genome-edited isogenic cell lines to highlight how changes in the in vitro DNA-binding properties of a transcription factor influence transcription factor genomic occupancy and transcription factor-mediated gene regulation.

Finally, I present a framework for using protein binding microarray data to identify possible perturbations of transcription factor binding sites by non-coding variants, and describe how I apply this framework in analysing the roles of signalling transcription factors in red blood cell trait variation.

Our findings enhance our understanding of fundamental transcription factor-DNA recognition rules, and provide insights into the contributions of coding and non-coding transcriptional regulatory variation towards gene regulatory mechanisms underlying human phenotypic variation and disease pathogenesis.