Dynamic Analysis of Proviral Induction and De Novo Methylation: Implications for a Histone Deacetylase-Independent, Methylation Density-Dependent Mechanism of Transcriptional Repression

Abstract

Methylation of cytosines in the CpG dinucleotide is generally associated with transcriptional repression in mammalian cells, and recent findings implicate histone deacetylation in methylation-mediated repression. Analyses of histone acetylation in in vitro-methylated transfected plasmids support this model; however, little is known about the relationships among de novo DNA methylation, transcriptional repression, and histone acetylation state. To examine these relationships in vivo, we have developed a novel approach that permits the isolation and expansion of cells harboring expressing or silent retroviruses. MEL cells were infected with a Moloney murine leukemia virus encoding the green fluorescent protein (GFP), and single-copy, silent proviral clones were treated weekly with the histone deacetylase inhibitor trichostatin A or the DNA methylation inhibitor 5-azacytidine. Expression was monitored concurrently by flow cytometry, allowing for repeated phenotypic analysis over time, and proviral methylation was determined by Southern blotting and bisulfite methylation mapping. Shortly after infection, proviral expression was inducible and the reporter gene and proviral enhancer showed a low density of methylation. Over time, the efficacy of drug induction diminished, coincident with the accumulation of methyl-CpGs across the provirus. Bisulfite analysis of cells in which 5-azacytidine treatment induced GFP expression revealed measurable but incomplete demethylation of the provirus. Repression could be overcome in late-passage clones only by pretreatment with 5-azacytidine followed by trichostatin A, suggesting that partial demethylation reestablishes the trichostatin-inducible state. These experiments reveal the presence of a silencing mechanism which acts on densely methylated DNA and appears to function independently of histone deacetylase activity.

ACKNOWLEDGMENTS

The GFP-Bex1 plasmid and Phoenix A retroviral producer cells were kind gifts of M. Anderson and G. Nolan, respectively. We thank Claire Francastel, Robert Eisenman, Steve Fiering, Reinhard Stoeger, Dan Cimbora, and the Martin and Groudine laboratories for helpful suggestions. We also thank the FHCRC Biotechnology and Flow Cytometry Shared Resource facilities for technical assistance and Kristy Seidel for advice on statistical analysis.

This work was supported by NIH grants to M.G. and D.I.K.M., who is a Scholar of the Leukemia Society of America, and fellowships from the NIH to M.C.L. and Deutsche Forschungsgemeinschaft to D.S.