Structure-function Analysis Of Histone Demethylase LSD2 And The Role Of UHRF1 In Genomic Imprinting

LSD1 was the first identified lysine-specific histone demethylase belonging to the flavin-dependent amine oxidase family. LSD2, a mammalian protein related to LSD1, catalyzes demethylation of H3K4me1/2 and plays an important role in transcriptional regulation and genomic imprinting. In collaboration with Prof. Zhongzhou Chen’s lab in China Agricultural University, we solved the crystal structure of LSD2 and analyzed the structural and functional relationship. LSD2 has three structural domains:the N-terminal Zinc finger domain, the centrally located SWIRM domain and the C-terminal amine oxidase domain. They closely pack together to form a boot-shaped structure. The N-terminal zinc finger domain is actually composed of composed of a novel C4H2C2-type zinc finger and a CW-type zinc finger. Although the N-terminal zinc-finger domain does not directly interact with the C-terminal amine oxidase domain, we demonstrated that it is required for demethylase activity and the binding of cofactor FAD. In fact a relay of extensive interactions through the zinc finger-SWIRM-oxidase domains is required for LSD2 demethylase activity and the binding of FAD. Thus, the zinc finger domain is not only a module used in prontein-protein and protein-DNA interactions, but can also have a critical scaffolding role through intramolecular interactions. Our study also provides a molecular explanation for the critical role of the zinc finger and SWIRM domains in controlling LSD2 demethylase activity.Genomic imprinting is a phenomenon that the two alleles of a gene are differentially expressed depending on their parent of origin. DNA methylation is believed to play a critical role in genomic imprinting. To study the role of DNA methylation in establishment and maintenance of genomic imprinting, we made use of the UHRF1 knockout ES cells as a model system. UHRF1, a multifunctional protein, is required for DNA maintenance methylation due to its essential role in targeting Dnmt1 to DNA replication forks during S phase of cell cycle. As a result, DNA methylation in Uhrf1-/-ES cells is substantially lower than that in wild-type ES cells. Through microarray analysis we identified a total of 25 imprinting genes whose expression were either up-regulated or down-regulated in the Uhrf1-/-ES cells. Reexpression of UHRF1 in the Uhrf1-/-ES cells restored the overall DNA methylation to the wild-type level. However, DNA methylation was not restored for most of the imprinting genes except for H19, Nnat and Dlk1. We examined the histone modifications including H3ac, H3K4me2, H3K4me3, H3K9me2, H3K9me3 and H4K20me3 at the regulatory regions (DMR) of the imprinting genes by chromatin immunoprecipitation (ChIP) assays. We also analyzed the binding of ZFP57, a transcription factor involved in transcriptional regulation of imprinting genes, using ChIP assay. We observed an relative enrichment for repressive histone markers such as H3K9me2/3 for the imprinting genes (H19, Nnat and Dlk1) whose DNA methylation was restored upon reexpression of UHRF1 and enrichment for active markers such as H3ac and H3K4me2/3 for the genes whose methylation was not restored. Our study thus suggests a critical role for histone modifications in determining the establishment of imprinting-related DNA methylation.