Book of Abstracts: Albany 2007
June 19-23 2007
Chicken and yeast nucleosomal DNA sequences differ at the ends: A possible relation to the linker histone binding
The linker histones H1/H5 bind to the entry and exit points of nucleosomal DNA and protect additional ~20 bp at one end. However, it is unclear whether there is any sequence feature at the ends of nucleosomal DNA facilitating the linker histone binding. This information cannot be acquired directly from the known chicken  and yeast  nucleosomal core particle (NCP) sequences obtained by MNase digestion. Nevertheless, availability of the chicken and yeast genome sequences provides a unique opportunity to extract the ?missing? information.
Here, we aligned the chicken and yeast NCP sequences (147±5 bp in length) to the corresponding genomes and extracted ~20 bp flanking sequences at both ends. No additional realignment was made ? that is, we used the default ?central? alignment. Next, we analyzed the combined occurrence of dinucleotides AA, TT and AT along the extended NCP sequences. The AA:TT dimer reveals the strongest 10-11 bp periodicity as shown earlier ; in addition, we included the dinucleotide AT because in the free DNA it has the same minor-groove bending propensity as the AA:TT. The combined occurrence of the three dinucleotides (denoted AT2 below) periodically oscillates along the chicken and yeast NCP sequences [1, 2], with the peaks at positions 6, 18, 27, ? 129, 141.
For the yeast NCP, we found that this ~10-bp oscillation goes beyond the end points of nucleosomes, so that the strongest peaks of the AT2 occurrence are observed at positions ?2 and 149. Note that these positions are approximately in phase with the peaks mentioned above. By contrast, for the chicken NCP, the AT2 oscillation does not ?pass? through the ends. Rather, pronounced peaks appear at positions 1 and 146, which are clearly out of phase with the above peaks. Furthermore, the ?out-of-phase? peaks are distributed asymmetrically between the two halves of the chicken NCP: the stronger peaks at the ends are coupled with the stronger peaks at positions 6, 18, etc. Thus, we observed a novel sequence pattern characteristic of the chicken, but not of the yeast NCP: several strong periodically positioned AT2 peaks in the terminal 40 base pairs of nucleosomal DNA, accompanied by the out-of-phase peaks at the very end of the NCP fragment. Importantly, the chicken chromatosome sequences  that are ~170-bp long reveal a similar pattern: the ?out-of-phase? peak appears ~20 bp away from one of the chromatosome ends, followed by a series of four ?in-phase? peaks. The same results were obtained for the AT3 trimers distribution (i.e., when the combined occurrence of the trimers AAA:TTT and AAT:ATT was analyzed).
In our interpretation, the different sequence periodicities at the ends of chicken and yeast nucleosomal DNA reflect different spatial trajectories of DNA at the entry/exit points. The presence or absence of the out-of-phase signal is associated with the linker histone binding. Because the globular domain of linker histone H1/H5 cannot fit between the DNA superhelical turns in nucleosome, we speculate that DNA adopts an additional ?out-of-phase? bend at the chicken NCP end (not visible in the NCP X-ray structures), thereby making enough space to accommodate the linker histone. By contrast, in most of the yeast nucleosomes, the DNA at the ends of NCP fragments follows its ?natural extension? trajectory, leaving no space for the linker histones. This structural difference is consistent with the linker histone stoichiometry in the two species: every chicken nucleosome has one linker histone on average, while in yeast, the H1-analog protein Hho1p is found in one of ~40 nucleosomes .
References and Footnotes
Feng Cui* and Victor B. Zhurkin
Laboratory of Cell Biology