Neurogenesis relies on a complex network of transcriptional activators and repressors that control the initiation and maintenance of neural traits. The molecular mechanisms involved in lineage commitments — from stem cells to terminally differentiated neurons — are not clear. Now, a study by Ballas and colleagues provides fascinating insight into how the progressive release of transcriptional repressors from promoters regulates the orderly expression of target genes during development.

In culture, embryonic stem (ES) cells can form embryoid bodies, which, when exposed to retinoic acid, differentiate first into neural progenitors and then into neurons. The authors found that the transcriptional repressor REST was expressed in abundance by ES cells, but was downregulated post-translationally to minimal levels in neural progenitors. Interestingly, the residual REST still binds to the chromatin of neuronal genes, thereby repressing their expression, at the specific recognition site RE1 in neural progenitors; however, it is absent from the chromatin of the same genes in postmitotic neurons. This indicates that differentiation of progenitors into mature neurons requires the removal of REST-mediated transcriptional repression.

The silencing of neuronal genes by REST in differentiated non-neuronal cells involves a cofactor, CoREST, which recruits additional silencing machinery. This machinery includes specific histone methyltransferases, as well as methyl-CpG-binding protein (MECP2), which binds to DNA regions with a methylated cytosine-guanidine stretch (mCpG). In ES cells, a complex of REST, CoREST and MECP2 is present on the RE1 and mCpG sites in promoters of neuronal genes such as brain-derived neurotrophic factor (Bdnf) and Calbindin. The release of REST from the complex is sufficient for default expression of some neuronal genes; whereas other genes continue to be expressed at low levels owing to the presence of CoREST and MECP2 on the mCpG sites. The authors showed that membrane depolarization of differentiated neurons resulted in the release of MECP2 from the repressor complex on the Bdnf promoter and increased expression of the gene. Intriguingly, neuronal activity did not have the same effect on the repressor complex on the Calbindin promoter, which indicates that other factors might be necessary for the activation of Calbindin expression.

This elegant study indicates that specific transcriptional recognition sites and epigenetically-modified DNA sequences might provide a platform for the dynamic assembly and disassembly of repressor complexes that are required for lineage commitments and plasticity in mature neurons. How transcriptional activators might fit into the model remains to be seen.