How do transcription factors find their targets?

Abstract

Transcription factors (TFs) are proteins that bind to DNA in a sequence-specific manner and regulate gene transcription. ChIP-seq and other techniques have allowed a study of TF binding genome-wide but have shown that TFs bind to only a subset of in vitro predicted sites. This highlights that it is difficult to accurately predict which sites TF bind in vivo. In this current study, we have uncovered how the functional domains (FD, i.e., non-DNA binding domains) of TFs mediate protein-protein interactions that aid in TF genome localisation and investigate how DNA methylation affects the genomic localisation of TFs. This is important not only to better understand gene regulation but also to be able to develop next-generation artificial TFs. Analysis of ChIP-seq and RNA-seq data showed that disrupting the interaction between TF Krüppel -like factor 3 (KLF3) and its newly discovered FD co-partner WDR5 significantly impacted both KLF3 and WDR5 genomic localisation and gene activation. This demonstrated that protein-protein interactions with the FD can influence genome-wide TF binding. Next, we investigated how this finding translated to a family of TFs, the KLF family which all share a similar DNA binding domain (DBD) but bind and regulate different target genes across different tissues. Using publicly available ChIP-seq data, we showed that KLF family members have vastly different in vivo genome-wide binding profiles in HEK293 cells despite having similar consensus binding motifs. We then showed using ChIP-seq that replacing the KLF3 FD with the KLF1 FD reduces the number of binding sites and impacts genomic localisation. Taken together, these results demonstrate the importance of the FD in genome-wide binding and how FDs, by mediating specific protein-protein interactions, may allow TF families to achieve functional diversity despite their similar DBDs. We also investigated DNA methylation at the β-globin locus which is a model locus for studying transcriptional regulation where methylation has been identified to affect gene expression, but the underlying mechanisms are yet to be described. By measuring DNA methylation levels in human erythroid cell lines that show differential expression of HBG1/2 and HBB, we identified CpG sites at or near these genes that were differentially methylated and were close to regulatory regions and well-known TF binding sites. Future experiments will investigate whether these differences may directly affect TF binding leading to these gene-expression changes. Overall, my project has illuminated novel ways in which TF binding can be regulated in vivo to allow precise patterns of cellular gene expression, both by FD recruitment of partner proteins that regulate genomic localisation and by target site methylation that may alter TF binding.