A recently published report shows that eukaryotic genomes encode preferences for nucleosome positioning, which can be modulated to facilitate specific chromosomal functions.

Nucleosomes — comprising 146 bp of DNA wrapped around a histone protein octamer — form the first level of genome compaction in eukaryotes. This organization has important implications for gene expression, because nucleosome-bound DNA is less accessible to various regulatory factors than nucleosome-free DNA. But how is the positioning of the nucleosomes regulated? Could the genome itself encode a preference for their distribution?

Segal et al. addressed this question in a series of experiments that combined experimental and computational approaches. Working in yeast, the authors isolated regions of DNA that were stably wrapped around nucleosomes on a genome-wide scale. On the basis of the isolated sequences, they used a probabilistic model to predict the sequences that preferentially wrapped around the histone octamer. It is known that certain nucleotide combinations that recur with a certain frequency facilitate DNA bending; Segal et al. not only found such motifs and periodicities using their model, but, by experimentally manipulating the motif sequence and their patterns, they could enhance or reduce the binding affinity of DNA to the nucleosome in accordance with the predictions of the model. The results also showed that the preferences of DNA for nucleosome occupancy vary from one genomic region to another. For example, centromeric sequences have the highest predicted nucleosomal occupancy; conversely, highly expressed genes such as ribosomal or transfer RNA genes show a preference for low occupancy.

Genome-wide scans showed that the high-affinity sequences occur more frequently than predicted by chance. Using a computational approach, the authors showed that various genomic regions favoured one or more nucleosome positioning patterns. Such a preference can have important functional effects; for example, a preference for a strong affinity at a transcription site would make it less accessible to transcription factors, whereas a preference for weak affinity would have the opposite effect. Using various approaches to compare in vivo nucleosome positioning data to their predictions, Segal et al. confirmed the validity of their models and showed that ∼50% of nucleosomal organization in vivo can be explained by these sequence preferences alone.

On the basis of their data, the authors suggest that by encoding weak nucleosomal affinity eukaryotic genomes might guide transcriptional machinery to functional sites. If correct, this hypothesis could also account for why some transcription factor binding sites remain unoccupied in vivo.

Although Segal et al. admit that more accurate DNA–nucleosome interaction models are needed, this report represents the first step towards integrating the effects of chromatin structure into models of gene regulation.