Mendizabal Isabel, Zeng Jia, Keller Thomas E, Yi Soojin V
School of Biological Sciences, Georgia Institute of Technology, Atlanta, GA 30332, USA.
Department of Genetics, Physical Anthropology and Animal Physiology, University of the Basque Country UPV/EHU, Barrio Sarriena s/n, 48940 Leioa, Spain.
Nucleic Acids Res. 2017 May 5;45(8):4390-4400. doi: 10.1093/nar/gkx020.
Genomic DNA methylation maps (methylomes) encode genetic and environmental effects as stable chemical modifications of DNA. Variations in DNA methylation, especially in regulatory regions such as promoters and enhancers, are known to affect numerous downstream processes. In contrast, most transcription units (gene bodies) in the human genome are thought to be heavily methylated. However, epigenetic reprogramming in cancer often involves gene body hypomethylation with consequences on gene expression. In this study, we focus on the relatively unexplored phenomenon that some gene bodies are devoid of DNA methylation under normal conditions. Utilizing nucleotide-resolution methylomes of diverse samples, we show that nearly 2000 human genes are commonly hypomethylated. Remarkably, these genes occupy highly specialized genomic, epigenomic, evolutionary and functional niches in our genomes. For example, hypomethylated genes tend to be short yet encode significantly more transcripts than expected based upon their lengths, include many genes involved in nucleosome and chromatin formation, and are extensively and significantly enriched for histone-tail modifications and transcription factor binding with particular relevance for cis-regulation. Furthermore, they are significantly more prone to cancer-associated hypomethylation and mutation. Consequently, gene body hypomethylation represents an additional layer of epigenetic regulatory complexity, with implications on cancer-associated epigenetic reprogramming.
基因组DNA甲基化图谱（甲基化组）将遗传和环境效应编码为DNA的稳定化学修饰。已知DNA甲基化的变化，尤其是在启动子和增强子等调控区域的变化，会影响众多下游过程。相比之下，人类基因组中的大多数转录单元（基因体）被认为高度甲基化。然而，癌症中的表观遗传重编程通常涉及基因体低甲基化，并对基因表达产生影响。在本研究中，我们关注的是一个相对未被探索的现象，即一些基因体在正常条件下没有DNA甲基化。利用不同样本的核苷酸分辨率甲基化组，我们发现近2000个人类基因通常是低甲基化的。值得注意的是，这些基因在我们的基因组中占据高度专业化的基因组、表观基因组、进化和功能生态位。例如，低甲基化基因往往较短，但根据其长度编码的转录本明显多于预期，包括许多参与核小体和染色质形成的基因，并且在组蛋白尾部修饰和转录因子结合方面广泛且显著富集，这与顺式调控特别相关。此外，它们明显更容易发生与癌症相关的低甲基化和突变。因此，基因体低甲基化代表了表观遗传调控复杂性的额外层面，对癌症相关的表观遗传重编程具有影响。
