Deciphering the transcriptional regulation code in colorectal cancer genome

摘要

Colorectal cancer (CRC) is the third most common cancer type to threaten life for both men and women in the developed world. The molecular mechanism of CRC is very complicated and involves changes in different categories of biological processes, among which a large number of mutations affecting either coding sequence of transcription factors (TFs) or their binding sites in genome are included. This indicated the importance of fine-tuned transcriptional regulation in normal colon function and the significance of its disruption in tumorigenesis and metastasis. However, the comprehensive knowledge of transcriptional regulation in CRC is still inadequate, leaving obstacles for clinical prognosis and therapy.In order to better understand the transcriptional regulation network in CRC, we designed a study including three individual projects to systematically investigate the mammalian transcriptional regulation in vitro, ex vivo and in vivo.In the in vitro study, we performed the High Throughput Systematic Evolution of Ligands by Exponential Enrichment (HT-SELEX) for the vast majority of mammalian TFs in order to profile their DNA sequence binding specificities. Eventually, we obtained binding profiles for 303 human DNA binding domains (DBDs), 84 mouse DBDs and 151 human full length TFs, representing 411 different TFs in total, which exceeds any existing database for mammalian TF DNA binding motifs and provides rich information for research on transcriptional regulation. By analyzing this data, we also determined some factors affecting TF-DNA binding such as adjacent base stacking and DNA shape, and suggested two advanced models to improve the computational prediction of TF binding.In the ex vivo study, we carried out chromatin immunoprecipitation followed by massively parallel sequencing (ChIP-seq) for over 500 different TFs in a single CRC cell line LoVo, and a relatively smaller scale in another CRC cell line GP5d. We observed that most TFs tended to bind to DNA forming highly dense clusters around cohesin and occupying an unexpectedly small fraction of human genome. Our data suggested that cohesin binding could function as a cellular memory to mark the TF binding sites and facilitate the quick re-establishment of TF binding within the limited time during each cycle of cell division.To test the function of some TF clusters, we generated a conditional knock out mouse strain lacking a 1.3kb TF cluster fragment (Myc-335) 335 kb upstream of Myc gene transcription starting site (TSS). We discovered that Myc-335 was a tumor specific enhancer for Myc gene and it was dispensable for normal intestinal development and function.The study greatly improved our knowledge of TF-DNA interaction and its biological function in the relevant fields of transcriptional regulation, cell cycle, epigenetics, epigenomics and cancer biology, and also provided the whole scientific community with enormous data sets for further analyses.