Negative regulation of DNMT3A de novo DNA methylation by frequently overexpressed UHRF family proteins as a mechanism for widespread DNA hypomethylation in cancer.

Yuanhui Jia,Jiemin Wong,Jiwen Li,Ruiping Wang,Yan Feng,Jialun Li,Lan Fang,Hongbin Ji,Junying Zhang,Qihan Wu,Pishun Li,Taiping Chen,Hao Cheng,Liangliang Xu,Yan Li,Jackie Han,James X Du ,Fĕi Li ,Wenjin Yu ,Hongxia Zhu

doi:10.1038/celldisc.2016.7

Abstract

Global DNA hypomethylation is a most common epigenetic alteration in cancer, but the mechanism remains elusive. Previous studies demonstrate that UHRF1 but not UHRF2 is required for mediating DNA maintenance methylation by DNMT1. Here we report unexpectedly a conserved function for UHRF1 and UHRF2: inhibiting de novo DNA methylation by functioning as E3 ligases promoting DNMT3A degradation. UHRF1/2 are frequently overexpressed in cancers and we present evidence that UHRF1/2 overexpression downregulates DNMT3A proteins and consequently leads to DNA hypomethylation. Abrogating this negative regulation on DNMT3A or overexpression of DNMT3A leads to increased DNA methylation and impaired tumor growth. We propose a working model that UHRF1/2 safeguards the fidelity of DNA methylation and suggests that UHRF1/2 overexpression is likely a causal factor for widespread DNA hypomethylation in cancer via suppressing DNMT3A.

Highlights

In mammals, DNA methylation at cytosine-C5 in the context of CpG dinucleotides is a key epigenetic modification required for embryonic development, transcriptional regulation, heterochromatin formation, X-inactivation, imprinting and genome stability [1,2,3]
To examine if UHRF2 plays a role in DNA methylation, we knocked down UHRF2 in the human lung cancer cell line A549 using shRNA and examined the DNA methylation status by immunofluorescent staining using an anti-5meC antibody
Cells transfected with shUHRF1 exhibited reduced levels of DNA methylation (Figure 1b), in agreement with that UHRF1 is required for DNA maintenance methylation catalyzed by DNMT1 [14, 15]

Summary

Introduction

DNA methylation at cytosine-C5 in the context of CpG dinucleotides is a key epigenetic modification required for embryonic development, transcriptional regulation, heterochromatin formation, X-inactivation, imprinting and genome stability [1,2,3]. UHRF1/2 negatively controls DNA methylation by DNMT3A evidence indicates that DNMT3A and DNMT3B contribute to maintaining DNA methylation patterns in embryonic stem (ES) and somatic cells, as inactivation of these enzymes leads to gradual loss of DNA methylation, both at single copy genes and repetitive DNA sequences [8,9,10,11,12]. This is likely explained by the fact that DNMT1 is unable to replicate DNA methylation patterns with 100% accuracy [11, 12]. This working model raises a new question as to how the maintenance and de novo methylation activities are coordinated to permit the faithful inheritance of DNA methylation, because too much de novo methylation would tip the balance of DNA methylation inheritance to increased DNA methylation

Methods

Results

Conclusion