Seek and ye shall find

Seek and Ye Shall Find

Imagine searching for a specific apartment in an unfamiliar city with no map, making your way blindly from door to door. This is the problem confronted by DNA-binding proteins, which must promptly assemble at precise genomic locations concealed within vast expanses of inappropriate nucleotide sequences.

This challenge has led biologists to theorize that such proteins must be able to slide quickly along exposed stretches of DNA to scan for binding sites, and then latch on tightly upon finding the right spot. This model would require at least two protein-DNA interaction states, one for search (S) and one for recognition (R), and Leith et al. have obtained compelling evidence supporting this two-state model.

They used total internal reflection fluoroscopy (TIRF) to track diffusion of fluorescently-labeled p53 dimers as they moved along tethered strands of lambda bacteriophage DNA. Strikingly, they observed that p53 exhibited different diffusion rates within different segments of DNA, varying by a factor of up to 1.6. Computational modeling confirmed that the S/R model accurately predicted this diffusion landscape more effectively than a single binding state model. p53 generally forms a complex comprising two homodimers that binds to a 20-bp response element (RE), and although the lambda genome does not contain canonical REs, it has many similar sequences, and these regions of low diffusivity probably correspond to RE-like regions where p53 transitions from the S to the R state.

The researchers also found evidence for frequent instances of ‘hemi-specific’ binding, wherein individual p53 dimers exhibit distinct binding behavior that allows the tetramer to effectively recognize 10-bp partial REs as well as full REs, adding an additional layer of complexity to the binding site-recognition process for this critical transcription factor. 1

Illustration 1.

Leith, J.S. et al. Proc. Natl. Acad. Sci. USA Published online 25 September 2012, doi: 10.1073/pnas.1120452109

In Search of Mismatched Marks

Each of the nucleosomes that compose chromatin contains a pair of H3-H4 histone dimers. Although it is well known that these proteins can undergo diverse chemical modifications as a component of epigenetic regulation, it remains unclear whether the two dimers perfectly mirror each other with regard to individual epigenetic marks.

This has proven a difficult question to address, but a method devised by Voigt et al. now reveals a striking degree of asymmetry in H3/H4 methylation. They used immunoprecipitation to islate nucleosomes, and then performed mass spectrometry to determine the proportion of H3 or H4 proteins that contain a given modification of interest in both undifferentiated and differentiated cells. Remarkably, they found that both H3K27me2/3 (typically a repressive mark) and H4K20me1 (associated with both activation and repression) were often distributed asymmetrically regardless of differentiation status.

At some gene loci – particularly in embryonic stem cells (ESCs) – H3 contains seemingly contradictory ‘bivalent’ modifications: H3K4me3 and H3K27me3, which are respectively activating and repressive. The researchers confirmed such bivalency at the single-nucleosome level, and demonstrated that it occurs exclusively in an asymmetric fashion. They propose that this state arises when the Polycomb repressor complex 2 (PRC2) protein, which adds the H3K27me3 modification, encounters a nucleosome displaying asymmetric H3K4me3. As ESCs differentiate, these loci can remain bivalent or resolve into symmetric sites with purely repressive or active epigenetic marks.

The apparent abundance of such asymmetric modification calls into question the hypothesis that chromatin modification is passed along semiconservatively during cell division, with each daughter strand receiving one of the parental H3/H4 dimers. It also suggests an additional regulatory mode for epigenetic marks, and further research will be required to untangle the in vivo impact of this asymmetry. 2

Illustration 2.

Voigt, P. et al. Cell, 151, 181–193 (2012).