During the past decade, the genomic DNA sequence of humans and many other organisms has been accurately determined, but profound questions remain about the influence of nucleotide sequence on genome function. For example, much genomic DNA does not encode protein – does it influence genome activity? To understand how genetic information is computed to create a cellular phenotype, we need better functional annotation of the genome. Our laboratory is part of this international effort.

Specifically, we study the dinucleotide sequence CpG, which is distributed genome-wide and has several properties expected of a genomic signalling module:

1) It exists in three major chemically modified forms – non-methylated, methylated and hydroxymethylated; 
2) Its frequency varies widely, being most dense at so-called CpG islands – clusters of dense CpG that suuround the control regions of most genes; 
3) Its different chemically modified forms are recognised by distinct proteins that can recruit complexes to alter chromatin structure.

The extent to which CpG signalling is involved in development, differentiation and disease is only just beginning to emerge. Our work so far indicates that methyl-CpG binding proteins recruit chromatin modifying enzymes to reinforce gene silencing, whereas CpG binding proteins promote the formation of potentially active chromatin. The results suggest that CpG acts as a global modulator of genome activity.

One of the best-studied methyl-CpG binding proteins is MeCP2, which in neurons is almost as abundant as the proteins that ubiquitously package DNA, the histones. MeCP2-deficient children acquire serious neurological disorders, in particular the autism spectrum disorder Rett Syndrome. Due to its monogenic origin, Rett Syndrome has become one of the most experimentally accessible of such disorders and studies of MeCP2 offer a golden opportunity to understand its complex pathology at a molecular level. Our mouse model of Rett Syndrome has accelerated our understanding of this disorder, most notably by demonstrating that advanced Rett-like symptoms in mice can be “cured” by putting back a functional MeCP2 gene.

To gain further insight into our mouse model for the reversal of Rett Syndrome, please watch the videos below. The first shows a mouse deficient in the protein MeCP2 exhibiting Rett-like symptoms. The second shows the recovery of the mouse from the first video after MeCP2 has been replaced.


This result raises the prospect that Rett Syndrome in humans might be treatable and has stimulated an international search for potential therapies. We are engaged in this effort while also improving our knowledge of MeCP2 function in the brain.