Identification and validation of transcription factors that regulate chromatin dynamics

Gene expression has to be tightly regulated during all cellular processes. During embryonic
development differentiating cells loose their developmental potential and acquire specific
functions by activating lineage-specific genes. Gene transcription programs are regulated by
transcription factors (TFs) in concert with dynamic changes in local chromatin organisation
of the DNA template. Both pathways are crucial for specific reprogramming of cells. However,
how TFs and chromatin marks exactly contribute to regulate gene expression programs
is not fully understood. For instance, the binding patterns of most mammalian TFs are still
unknown as well as how binding specificity is achieved. Chromatin modifications are highly
dynamic and cell-type specific. By regulating access to the DNA template they might guide
TF binding. As most chromatin modifications have simply been associated with gene activity,
a central remaining question is how chromatin modifications impact on gene expression
and if they are a cause or consequence of the transcriptional state of a gene.
Further it is still an open question how chromatin marks are targeted to specific loci and
how they are dynamically regulated. Trimethylation of histone 3 at lysine 27 (H3K27me3) is
set by the Polycomb group of proteins, which regulate body patterning during development.
Polycomb-mediated H3K27me3 is associated with gene repression and essential for cellular
differentiation. Further work shows that H3K27me3 targets are cell-type specific and highly
dynamic during differentiation. It is unclear how these changes are regulated. Thus, we
hypothesise that TFs, by recognising distinct DNA motifs, could contribute to the required
specificity of chromatin reprogramming. In collaboration with the group of Erik van Nimwegen
we applied an unbiased approach to model changes in H3K27me3 methylation during in vitro
neuronal differentiation in terms of predicted transcription factor binding sites. This
approach predicts many TFs to regulate H3K27me3 at specific stages of cellular differentiation.
We experimentally focus on the validation of the RE-1 silencing transcription factor (REST)
and the family of SNAIL TFs, which are both predicted to regulate a gain of H3K27me3 levels
as stem cells differentiate to neuronal progenitor cells. We determine genome-wide binding
sites of REST at these two cellular stages and show that measured binding sites of REST show
a high overlap with predicted ones. Mapping H3K27me3 in stem cells and progenitor cells of
wild type and REST knock out (RESTko) cells shows a specific loss of H3K27me3 at
promoter-proximal REST binding sites in neuronal progenitors, validating the computational
prediction. Moreover, short promoter fragments containing either REST or SNAIL binding sites
are sufficient to recruit H3K27me3, whereas deletion of the respective binding sites results
in a significant loss of H3K27me3. These results suggest that TFs are important contributors
in the regulation of chromatin dynamics. However, further experiments are required to test if
this is a general feature of TFs or a specialised role for REST and SNAIL proteins. In this
context the extension of TF binding maps is crucial, as binding preferences for only 20-30%
of all TFs are known at present. Extending this list, together with further perturbation
experiments, will elucidate to what extent TF binding patterns can explain both changes in
chromatin state as well as transcription.