Study Of A Small Heat Shock Protein(sHSP)ROS5(REPRESSOR OF SILENCING5)on Anti-silencing And Regulation Of DNA Demethylation

DNA methylation is an important epigenetic modification involved in many biological processes, such as regulation of gene expression, maintainence of silencing of transposons as well as genome stability. The establishment of DNA methylation in plants relies on the de novo pathway mediated by RNA-directed DNA methylation (RdDM), While maintainence of DNA methylation in CG, CNG and CNN contexts is conducted by DNA methyltransferase1(MET1), CHROMOMETHYLASE3(CMT3) and RNA-directed DNA methylation (RdDM) pathway, respectively. The active DNA demethylation in plants is accomplished through the base excision repair pathway initiated by REPRESSOR OF SILENCING1(ROS1) family. Many progresses have been achieved in the study of active DNA demethylation pathway recently, E.g., DNA phosphatase ZDP1, DNA repair protein XRCC1, histone acetyltransferase INCREASE DNA METHYLATION1(ROS4/IDM1) and histone chaperone STRUCTURE SPECIFIC RECOGNITION PROTEIN1(SSRP1) were found to be involved in active DNA demethylation. However, there are still many questions about the mechnism of DNA demethylation to be answered.The C24transgenic line carrying RD29A:LUC and35S:NPTII is a powerful screening system for studying the DNA methylation and demethylation pathway. DNA demethylase ROS1was found using this system. By screening the supressors of ROS1in the RD29A:LUC site, the Zhu Jiankang lab found nearly all of the members involved in RdDM pathway. To investigate the DNA demethylation pathway by ROS1further, our lab has screened mutants leading to the silencing of35S:NPTII transgene.Here we isolated a new gene ROS5, which encodes a nucleus-localized small heat shock protein (sHSP). However, the expression of ROS5is not induced upon heat treatment, and its mutation does not impair the thermotolerance of plants. Meanwhile, unlike normal sHSP, ROS5has no chaperone activity. These results suggest that ROS5is an atypical sHSP. Mutation of ROS5leads to DNA hypermethylation in the3’end of NPTII and NOS region as well as thousands of endogenous regions (CG:1179, CNG:84, CNN:309). The hypermethylated DMRs (Differentially methylated regions) in CG and CNG contexts (CG:969, CNG:40) mainly localize in genie regions, while hypermethylated DMRs in CNN context (175) mainly localize in TEs. Lots of DMRs regulated by ROS5are overlapped with that regulated by ROS1and ROS4/IDM1. Importantly, ROS5can physically interact with ROS4/IDM1, and both of them may be in the same complex with ROS1. These results suggest that ROS5, ROS1and ROS4/DDM1may cooperately prevent the DNA hypermethylation in the overlapped regions. ROS4/BDM1is an acetyltransferase, which can acetylate H3K18and H3K23, while mutation of ROS5can lead to obvious decrease in H3K18ac, and weak decrease in H3K23ac, suggesting that mutation of ROS5may partially dispair the activity of ROS4/IDM1. Based on these results, we propose that ROS5regulates DNA demethylation by interacting with ROS4/IDM1, thereby creating a relatively loosened chromatin environment that facilitates the recruitment of ROS1to erase DNA methylation. We also found that the methylome of C24is different from Columbia (Col), further confirming that epigenetic modifications between different ecotypes are largely different. Importantly, DMRs regulated by ROS5are also significantly different between C24and Col.