Physics of Life Reviews 8 (2011) 39���50 www.elsevier.com/locate/plrev Review Cracking the chromatin code: Precise rule of nucleosome positioning Edward N. Trifonov a,b,��� a Genome Diversity Center, Institute of Evolution, University of Haifa, Mount Carmel, Haifa 31905, Israel b Department of Functional Genomics and Proteomics, Faculty of Science, Masaryk University, Kotlarska 2, CZ-61137 Brno, Czech Republic Received 20 December 2010 received in revised form 12 January 2011 accepted 14 January 2011 Available online 19 January 2011 Communicated by M. Frank-Kamenetskii Abstract Various aspects of packaging DNA in eukaryotic cells are outlined in physical rather than biological terms. The informational and physical nature of packaging instructions encoded in DNA sequences is discussed with the emphasis on signal processing difficulties ��� very low signal-to-noise ratio and high degeneracy of the nucleosome positioning signal. As the author has been contributing to the field from its very onset in 1980, the review is mostly focused at the works of the author and his colleagues. The leading concept of the overview is the role of deformational properties of DNA in the nucleosome positioning. The target of the studies is to derive the DNA bendability matrix describing where along the DNA various dinucleotide elements should be positioned, to facilitate its bending in the nucleosome. Three different approaches are described leading to derivation of the DNA deformability sequence pattern, which is a simplified linear presentation of the bendability matrix. All three approaches converge to the same unique sequence motif CGRAAATTTYCG or, in binary form, YRRRRRYYYYYR, both representing the chromatin code. �� 2011 Elsevier B.V. All rights reserved. Keywords: Ducleotide periodicity DNA bendability Matrix of bendability Nucleosome sequence pattern Nucleosome mapping Signal processing 1. Basics of chromatin structure 1.1. Introduction Centimeters long DNA molecules of higher organisms are squeezed in the chromosomes of the size of microns, with 105���106-fold compaction. If the semi-rigid chain of 10 cm long DNA (one chromosome) is freely suspended in water solution, it would make a random coil of the size 103.5 times smaller than its length, due to its natural flexibility (statistical segment 100 nanometers [1]). Nature, thus, should have taken care of only additional compaction of about two orders. It is, however, not only matter of compaction. The problem is to fold it in such a way that in the course of cell division rather quick unfolding would be allowed, without entanglement. It then folds back to the original compact state. The most compact metaphase chromosomes, right before cell division, appear in cytological slides as * Address for correspondence: Genome Diversity Center, Institute of Evolution, University of Haifa, Mount Carmel, Haifa 31905, Israel. Tel.: +972 4 828 8096 fax: +972 4 824 6554. E-mail address: trifonov@research.haifa.ac.il. 1571-0645/$ ��� see front matter �� 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.plrev.2011.01.004

40 E.N. Trifonov / Physics of Life Reviews 8 (2011) 39���50 sets identical for all cells of given species, with bands of stain making very specific and different patterns along each of the 5 to 100 chromosomes, depending on species. That is, DNA of, say, human chromosome 9 is folded in a unique way characteristic for that chromosome. The instructions about the details of the folding are somehow stored in DNA itself, in its nucleotide sequence. The code that translates the linear sequence of DNA bases into its 3D trajectory, that is, physical and geometrical nature of the DNA folding, and how it is expressed in the sequence form, is the subject of this review. 1.2. Two obvious modes of organized folding One solution of the problem would be the folding scheme used for shroud lines of a parachute. Here the lines are folded in a zig-zag (side-by-side) manner, quick to unfold to full length without twisting or knotting. One could also wind the rope in a cylindrical spring, solenoid. In this case pulling the ends would not cause an entanglement but, obviously, a significant twist will be inflicted to the fully extended rope. One also could imagine some designs with compensating windings of opposite sign. As long as DNA is concerned Nature utilizes both folding schemes, zig-zag (e.g. [2]) and winding modes [3], although it is not clear yet, what is the actual large scale arrangement (higher order structure) that guarantees quick reversible folding/unfolding. Remarkably, DNA does suffer some degree of the entanglement and twisting, as is evidenced by existence of special enzymes that reduce superhelicity of DNA and even unlock catenanes of circular DNA molecules by cutting and reconnecting the ends [4]. 1.3. Get started DNA, of course, is itself a double helix of two interwound sugar���phosphate chains completing one helical turn every 10.55 base pairs (bp) [5,6]. We are talking, thus, about folding of this duplex molecule. The DNA folding is assisted by proteins, in various degrees for different living kingdoms. In simple organisms, bacteria and viruses, DNA is, essentially, naked. In higher organisms, plants and animals, the folded state involves special proteins, in amount roughly equal to the DNA component. In these organisms the double-helical DNA winds locally into compact cylindri- cal (super)helices of only about 1.5 turns of the DNA, making remarkably standard tight packing units ��� nucleosome core particles, of 11 nm in diameter [7] (note that diameter of the DNA double helix is 2 nm). Their universal geometry is secured by a protein component, histone octamers, apparently nearly identical for the nucleosomes of all species, while DNA sections involved are all sequence-wise different. The particles are standard to such a degree that even their mixture, with different DNA sequences within the particles, can be crystallized [7]. The credit for discovery of the nucleosomes has to be given to four laboratories. First, Hewish and Burgoyne [8] discovered that enzymatic digestion of chromosomes results in fragmentation of DNA to about 200 bp long pieces and multiples of that size. This suggested that the chromosomes are built of units containing DNA of this size, and DNA connecting the units was somehow more accessible to the enzymes. The units themselves, in form of compact �����-bodies���, were then visualized by electron microscopy [9]. Physico-chemical disassembly���assembly studies by van Holde [10] and R. Kornberg [11] revealed that the particles consisted of four types of special proteins, histones, two molecules of each, and DNA wrapped around the histone octamers. The particles got the name ���nucleosomes��� [11]. One could think of random involvement of DNA in the nucleosomes, so that the sequence positions of the nucleo- somes along any given gene would be uncertain and different for two identical copies of the gene. However, already in the 1970s it became clear that at least some nucleosomes do take specific positions along the sequences [12,13]. The latter work of Ponder and Crawford gave a strong clue as to what would be the sequence feature to which the histone octamers are attracted. It appeared that some nucleosomes could occupy several discrete positions on the same DNA fragment. Remarkably, these alternative positions were shifted one from another by 10���11 bp. Here the reader is suggested to exercise her (his) geometrical imagination. What the 10���11 bp shift would mean? 1.4. Degenerate sequence pattern versus unique pattern The histone octamers, thus, seem to like special DNA fragments (sequences) with some common property. In other words, every such privileged fragment carries certain sequence signal expressed in some combination of letters (base pairs). Let us assume that the signal is unique and universal, sort of a label that appears in every 200 bp long