Topological polymorphism of nucleosome fibers and folding of chromatin.

We discuss recent observations of polymorphic chromatin packaging at the oligonucleosomal level and compare them with computer simulations. Our computations reveal two topologically different families of two-start 30-nm fiber conformations distinguished by the linker length L; fibers with L ≈ 10n and L ≈ 10n+5 basepairs have DNA linking numbers per nucleosome of ΔLk ≈ -1.5 and -1.0, respectively (where n is a natural number). Although fibers with ΔLk ≈ -1.5 were observed earlier, the topoisomer with ΔLk ≈ -1.0 is novel. These predictions were confirmed experimentally for circular nucleosome arrays with precisely positioned nucleosomes. We suggest that topological polymorphism of chromatin may play a role in transcription, with the {10n+5} fibers producing transcriptionally competent chromatin structures. This hypothesis is consistent with available data for yeast and, partially, for fly. We show that both fiber topoisomers (with ΔLk ≈ -1.5 and -1.0) have to be taken into account to interpret experimental data obtained using new techniques: genome-wide Micro-C, Hi-CO, and RICC-seq, as well as self-association of nucleosome arrays in vitro. The relative stability of these topoisomers is likely to depend on epigenetic histone modifications modulating the strength of internucleosome interactions. Potentially, our findings may reflect a general tendency of functionally distinct parts of the genome to retain topologically different higher-order structures.

CTCF-dependent chromatin boundaries formed by asymmetric nucleosome arrays with decreased linker length.

The CCCTC-binding factor (CTCF) organises the genome in 3D through DNA loops and in 1D by setting boundaries isolating different chromatin states, but these processes are not well understood. Here we investigate chromatin boundaries in mouse embryonic stem cells, defined by the regions with decreased Nucleosome Repeat Length (NRL) for ∼20 nucleosomes near CTCF sites, affecting up to 10% of the genome. We found that the nucleosome-depleted region (NDR) near CTCF is asymmetrically located >40 nucleotides 5'-upstream from the centre of CTCF motif. The strength of CTCF binding to DNA and the presence of cohesin is correlated with the decrease of NRL near CTCF, and anti-correlated with the level of asymmetry of the nucleosome array. Individual chromatin remodellers have different contributions, with Snf2h having the strongest effect on the NRL decrease near CTCF and Chd4 playing a major role in the symmetry breaking. Upon differentiation, a subset of preserved, common CTCF sites maintains asymmetric nucleosome pattern and small NRL. The sites which lost CTCF upon differentiation are characterized by nucleosome rearrangement 3'-downstream, with unchanged NDR 5'-upstream of CTCF motifs. Boundaries of topologically associated chromatin domains frequently contain several inward-oriented CTCF motifs whose effects, described above, add up synergistically.

Nucleosome spacing periodically modulates nucleosome chain folding and DNA topology in circular nucleosome arrays.

The length of linker DNA that separates nucleosomes is highly variable, but its mechanistic role in modulating chromatin structure and functions remains unknown. Here, we established an experimental system using circular arrays of positioned nucleosomes to investigate whether variations in nucleosome linker length could affect nucleosome folding, self-association, and interactions. We conducted EM, DNA topology, native electrophoretic assays, and Mg2+-dependent self-association assays to study intrinsic folding of linear and circular nucleosome arrays with linker DNA length of 36 bp and 41 bp (3.5 turns and 4 turns of DNA double helix, respectively). These experiments revealed that potential artifacts arising from open DNA ends and full DNA relaxation in the linear arrays do not significantly affect overall chromatin compaction and self-association. We observed that the 0.5 DNA helical turn difference between the two DNA linker lengths significantly affects DNA topology and nucleosome interactions. In particular, the 41-bp linkers promoted interactions between any two nucleosome beads separated by one bead as expected for a zigzag fiber, whereas the 36-bp linkers promoted interactions between two nucleosome beads separated by two other beads and also reduced negative superhelicity. Monte Carlo simulations accurately reproduce periodic modulations of chromatin compaction, DNA topology, and internucleosomal interactions with a 10-bp periodicity. We propose that the nucleosome spacing and associated chromatin structure modulations may play an important role in formation of different chromatin epigenetic states, thus suggesting implications for how chromatin accessibility to DNA-binding factors and the RNA transcription machinery is regulated.

DNA-RNA interactions are critical for chromosome condensation in Escherichia coli.

Bacterial chromosome (nucleoid) conformation dictates faithful regulation of gene transcription. The conformation is condition-dependent and is guided by several nucleoid-associated proteins (NAPs) and at least one nucleoid-associated noncoding RNA, naRNA4. Here we investigated the molecular mechanism of how naRNA4 and the major NAP, HU, acting together organize the chromosome structure by establishing multiple DNA-DNA contacts (DNA condensation). We demonstrate that naRNA4 uniquely acts by forming complexes that may not involve long stretches of DNA-RNA hybrid. Also, uncommonly, HU, a chromosome-associated protein that is essential in the DNA-RNA interactions, is not present in the final complex. Thus, HU plays a catalytic (chaperone) role in the naRNA4-mediated DNA condensation process.

Regulation of chromatin folding by conformational variations of nucleosome linker DNA.

Linker DNA conformational variability has been proposed to direct nucleosome array folding into more or less compact chromatin fibers but direct experimental evidence for such models are lacking. Here, we tested this hypothesis by designing nucleosome arrays with A-tracts at specific locations in the nucleosome linkers to induce inward (AT-IN) and outward (AT-OUT) bending of the linker DNA. Using electron microscopy and analytical centrifugation techniques, we observed spontaneous folding of AT-IN nucleosome arrays into highly compact structures, comparable to those induced by linker histone H1. In contrast, AT-OUT nucleosome arrays formed less compact structures with decreased nucleosome interactions similar to wild-type nucleosome arrays. Adding linker histone H1 further increased compaction of the A-tract arrays while maintaining structural differences between them. Furthermore, restriction nuclease digestion revealed a strongly reduced accessibility of nucleosome linkers in the compact AT-IN arrays. Electron microscopy analysis and 3D computational Monte Carlo simulations are consistent with a profound zigzag linker DNA configuration and closer nucleosome proximity in the AT-IN arrays due to inward linker DNA bending. We propose that the evolutionary preferred positioning of A-tracts in DNA linkers may control chromatin higher-order folding and thus influence cellular processes such as gene expression, transcription and DNA repair.

Dynamics of Chromatin Fibers: Comparison of Monte Carlo Simulations with Force Spectroscopy.

To elucidate conformational dynamics of chromatin fibers, we compared available force-spectroscopy measurements with extensive Monte Carlo simulations of nucleosome arrays under external force. Our coarse-grained model of chromatin includes phenomenological energy terms for the DNA-histone adhesion and the internucleosome stacking interactions. We found that the Monte Carlo fiber ensembles simulated with increasing degrees of DNA unwrapping and the stacking energy 8 kT can account for the intricate force-extension response observed experimentally. Our analysis shows that at low external forces (F < 3.0 picoNewtons), the DNA ends in nucleosomes breathe by ∼10 bp. Importantly, under these conditions, the fiber is highly dynamic, exhibiting continuous unstacking-restacking transitions, allowing accessibility of transcription factors to DNA, while maintaining a relatively compact conformation. Of note, changing the stacking interaction by a few kT, an in silico way to mimic histone modifications, is sufficient to transform an open chromatin state into a compact fiber. The fibers are mostly two-start zigzag folds with rare occurrences of three- to five-start morphologies. The internucleosome stacking is lost during the linear response regime. At the higher forces exceeding 4 picoNewtons, the nucleosome unwrapping becomes stochastic and asymmetric, with one DNA arm opened by ∼55 bp and the other arm only by ∼10 bp. Importantly, this asymmetric unwrapping occurs for any kind of sequence, including the symmetric ones. Our analysis brings new, to our knowledge, insights in dynamics of chromatin modulated by histone epigenetic modifications and molecular motors such as RNA polymerase.

Novel nucleosomal particles containing core histones and linker DNA but no histone H1.

Eukaryotic chromosomal DNA is assembled into regularly spaced nucleosomes, which play a central role in gene regulation by determining accessibility of control regions. The nucleosome contains ∼147 bp of DNA wrapped ∼1.7 times around a central core histone octamer. The linker histone, H1, binds both to the nucleosome, sealing the DNA coils, and to the linker DNA between nucleosomes, directing chromatin folding. Micrococcal nuclease (MNase) digests the linker to yield the chromatosome, containing H1 and ∼160 bp, and then converts it to a core particle, containing ∼147 bp and no H1. Sequencing of nucleosomal DNA obtained after MNase digestion (MNase-seq) generates genome-wide nucleosome maps that are important for understanding gene regulation. We present an improved MNase-seq method involving simultaneous digestion with exonuclease III, which removes linker DNA. Remarkably, we discovered two novel intermediate particles containing 154 or 161 bp, corresponding to 7 bp protruding from one or both sides of the nucleosome core. These particles are detected in yeast lacking H1 and in H1-depleted mouse chromatin. They can be reconstituted in vitro using purified core histones and DNA. We propose that these 'proto-chromatosomes' are fundamental chromatin subunits, which include the H1 binding site and influence nucleosome spacing independently of H1.

Organization of DNA in a bacterial nucleoid.

BACKGROUND: It is unclear how DNA is packaged in a bacterial cell in the absence of nucleosomes. To investigate the initial level of DNA condensation in bacterial nucleoid we used in vivo DNA digestion coupled with high-throughput sequencing of the digestion-resistant fragments. To this end, we transformed E. coli cells with a plasmid expressing micrococcal nuclease. The nuclease expression was under the control of AraC repressor, which enabled us to perform an inducible digestion of bacterial nucleoid inside a living cell. RESULTS: Analysis of the genomic localization of the digestion-resistant fragments revealed their non-random distribution. The patterns observed in the distribution of the sequenced fragments indicate the presence of short DNA segments protected from the enzyme digestion, possibly because of interaction with DNA-binding proteins. The average length of such digestion-resistant segments is about 50 bp and the characteristic repeat in their distribution is about 90 bp. The gene starts are depleted of the digestion-resistant fragments, suggesting that these genomic regions are more exposed than genomic sequences on average. Sequence analysis of the digestion-resistant segments showed that while the GC-content of such sequences is close to the genome-wide value, they are depleted of A-tracts as compared to the bulk genomic DNA or to the randomized sequence of the same nucleotide composition. CONCLUSIONS: Our results suggest that DNA is packaged in the bacterial nucleoid in a non-random way that facilitates interaction of the DNA binding factors with regulatory regions of the genome.

Rotational positioning of nucleosomes facilitates selective binding of p53 to response elements associated with cell cycle arrest.

The tumor suppressor protein p53 exhibits high affinity to the response elements regulating cell cycle arrest genes (CCA-sites), but relatively low affinity to the sites associated with apoptosis (Apo-sites). This in vivo tendency cannot be explained solely by the p53-DNA binding constants measured in vitro. Since p53 can bind nucleosomal DNA, we sought to understand if the two groups of p53 sites differ in their accessibility when embedded in nucleosomes. To this aim, we analyzed the sequence-dependent bending anisotropy of human genomic DNA containing p53 sites. For the 20 CCA-sites, we calculated rotational positioning patterns predicting that most of the sites are exposed on the nucleosomal surface. This is consistent with experimentally observed positioning of human nucleosomes. Remarkably, the sequence-dependent DNA anisotropy of both the p53 sites and flanking DNA work in concert producing strong positioning signals. By contrast, both the predicted and observed rotational settings of the 38 Apo-sites in nucleosomes suggest that many of these sites are buried inside, thus preventing immediate p53 recognition and delaying gene induction. The distinct chromatin organization of the CCA response elements appears to be one of the key factors facilitating p53-DNA binding and subsequent activation of genes associated with cell cycle arrest.

Topological polymorphism of the two-start chromatin fiber.

Specific details concerning the spatial organization of nucleosomes in 30 nm fibers remain unknown. To investigate this, we analyzed all stereochemically possible configurations of two-start nucleosome fibers with short DNA linkers L = 13-37 bp (nucleosome repeat length (NRL) = 160-184 bp). Four superhelical parameters-inclination of nucleosomes, twist, rise, and diameter-uniquely describe a regular symmetric fiber. The energy of a fiber is defined as the sum of four terms: elastic energy of the linker DNA, steric repulsion, electrostatics, and a phenomenological (H4 tail-acidic patch) interaction between two stacked nucleosomes. By optimizing the fiber energy with respect to the superhelical parameters, we found two types of topological transition in fibers (associated with the change in inclination angle): one caused by an abrupt 360° change in the linker DNA twisting (change in the DNA linking number, ΔLk = 1), and another caused by overcrossing of the linkers (ΔLk = 2). To the best of our knowledge, this topological polymorphism of the two-start fibers was not reported in the computations published earlier. Importantly, the optimal configurations of the fibers with linkers L = 10n and 10n + 5 bp are characterized by different values of the DNA linking number-that is, they are topologically different. Our results are consistent with experimental observations, such as the inclination 60° to 70° (the angle between the nucleosomal disks and the fiber axis), helical rise, diameter, and left-handedness of the fibers. In addition, we make several testable predictions, among them different degrees of DNA supercoiling in fibers with L = 10n and 10n + 5 bp, different flexibility of the two types of fibers, and a correlation between the local NRL and the level of transcription in different parts of the yeast genome.

Prediction of nucleosome rotational positioning in yeast and human genomes based on sequence-dependent DNA anisotropy.

BACKGROUND: An organism's DNA sequence is one of the key factors guiding the positioning of nucleosomes within a cell's nucleus. Sequence-dependent bending anisotropy dictates how DNA is wrapped around a histone octamer. One of the best established sequence patterns consistent with this anisotropy is the periodic occurrence of AT-containing dinucleotides (WW) and GC-containing dinucleotides (SS) in the nucleosomal locations where DNA is bent in the minor and major grooves, respectively. Although this simple pattern has been observed in nucleosomes across eukaryotic genomes, its use for prediction of nucleosome positioning was not systematically tested. RESULTS: We present a simple computational model, termed the W/S scheme, implementing this pattern, without using any training data. This model accurately predicts the rotational positioning of nucleosomes both in vitro and in vivo, in yeast and human genomes. About 65 - 75% of the experimentally observed nucleosome positions are predicted with the precision of one to two base pairs. The program is freely available at http://people.rit.edu/fxcsbi/WS_scheme/. We also introduce a simple and efficient way to compare the performance of different models predicting the rotational positioning of nucleosomes. CONCLUSIONS: This paper presents the W/S scheme to achieve accurate prediction of rotational positioning of nucleosomes, solely based on the sequence-dependent anisotropic bending of nucleosomal DNA. This method successfully captures DNA features critical for the rotational positioning of nucleosomes, and can be further improved by incorporating additional terms related to the translational positioning of nucleosomes in a species-specific manner.

Transcriptional activation of yeast genes disrupts intragenic nucleosome phasing.

Nucleosomes often undergo extensive rearrangement when genes are activated for transcription. We have shown previously, using paired-end sequencing of yeast nucleosomes, that major changes in chromatin structure occur when genes are activated by 3-aminotriazole (3AT), an inducer of the transcriptional activator Gcn4. Here, we provide a global analysis of these data. At the genomic level, nucleosomes are regularly phased relative to the transcription start site. However, for a subset of 234 strongly induced genes, this phasing is much more irregular after induction, consistent with the loss of some nucleosomes and the re-positioning of the remaining nucleosomes. To address the nature of this rearrangement, we developed the inter-nucleosome distance auto-correlation (DAC) function. At long range, DAC analysis indicates that nucleosomes have an average spacing of 162 bp, consistent with the reported repeat length. At short range, DAC reveals a 10.25-bp periodicity, implying that nucleosomes in overlapping positions are rotationally related. DAC analysis of the 3AT-induced genes suggests that transcription activation coincides with rearrangement of nucleosomes into irregular arrays with longer spacing. Sequence analysis of the +1 nucleosomes belonging to the 45 most strongly activated genes reveals a distinctive periodic oscillation in the A/T-dinucleotide occurrence that is present throughout the nucleosome and extends into the linker. This unusual pattern suggests that the +1 nucleosomes might be prone to sliding, thereby facilitating transcription.

Nucleosome reorganisation in breast cancer tissues.

BACKGROUND: Nucleosome repositioning in cancer is believed to cause many changes in genome organisation and gene expression. Understanding these changes is important to elucidate fundamental aspects of cancer. It is also important for medical diagnostics based on cell-free DNA (cfDNA), which originates from genomic DNA regions protected from digestion by nucleosomes. RESULTS: We have generated high-resolution nucleosome maps in paired tumour and normal tissues from the same breast cancer patients using MNase-assisted histone H3 ChIP-seq and compared them with the corresponding cfDNA from blood plasma. This analysis has detected single-nucleosome repositioning at key regulatory regions in a patient-specific manner and common cancer-specific patterns across patients. The nucleosomes gained in tumour versus normal tissue were particularly informative of cancer pathways, with ~ 20-fold enrichment at CpG islands, a large fraction of which marked promoters of genes encoding DNA-binding proteins. The tumour tissues were characterised by a 5-10 bp decrease in the average distance between nucleosomes (nucleosome repeat length, NRL), which is qualitatively similar to the differences between pluripotent and differentiated cells. This effect was correlated with gene activity, differential DNA methylation and changes in local occupancy of linker histone variants H1.4 and H1X. CONCLUSIONS: Our study offers a novel resource of high-resolution nucleosome maps in breast cancer patients and reports for the first time the effect of systematic decrease of NRL in paired tumour versus normal breast tissues from the same patient. Our findings provide a new mechanistic understanding of nucleosome repositioning in tumour tissues that can be valuable for patient diagnostics, stratification and monitoring.

The Methyltransferases METTL7A and METTL7B Confer Resistance to Thiol-Based Histone Deacetylase Inhibitors.

Histone deacetylase inhibitors (HDACi) are part of a growing class of epigenetic therapies used for the treatment of cancer. Although HDACis are effective in the treatment of T-cell lymphomas, treatment of solid tumors with this class of drugs has not been successful. Overexpression of the multidrug resistance protein P-glycoprotein (P-gp), encoded by ABCB1, is known to confer resistance to the HDACi romidepsin in vitro, yet increased ABCB1 expression has not been associated with resistance in patients, suggesting that other mechanisms of resistance arise in the clinic. To identify alternative mechanisms of resistance to romidepsin, we selected MCF-7 breast cancer cells with romidepsin in the presence of the P-gp inhibitor verapamil to reduce the likelihood of P-gp-mediated resistance. The resulting cell line, MCF-7 DpVp300, does not express P-gp and was found to be selectively resistant to romidepsin but not to other HDACis such as belinostat, panobinostat, or vorinostat. RNA-sequencing analysis revealed upregulation of the mRNA coding for the putative methyltransferase, METTL7A, whose paralog, METTL7B, was previously shown to methylate thiol groups on hydrogen sulfide and captopril. As romidepsin has a thiol as the zinc-binding moiety, we hypothesized that METTL7A could inactivate romidepsin and other thiol-based HDACis via methylation of the thiol group. We demonstrate that expression of METTL7A or METTL7B confers resistance to thiol-based HDACis and that both enzymes are capable of methylating thiol-containing HDACis. We thus propose that METTL7A and METTL7B confer resistance to thiol-based HDACis by methylating and inactivating the zinc-binding thiol.

Histone N-tails modulate sequence-specific positioning of nucleosomes.

The precise mechanisms governing sequence-dependent positioning of nucleosomes on DNA remain unknown in detail. Existing algorithms, taking into account the sequence-dependent deformability of DNA and its interactions with the histone globular domains, predict rotational setting of only 65% of human nucleosomes mapped in vivo. To uncover novel factors responsible for the nucleosome positioning, we analyzed potential involvement of the histone N-tails in this process. To this aim, we reconstituted the H2A/H4 N-tailless nucleosomes on human BRCA1 DNA (~100 kb) and compared their positions and sequences with those of the wild-type nucleosomes. In the case of H2A tailless nucleosomes, the AT content of DNA sequences is changed locally at superhelical location (SHL) ±4, while maintaining the same rotational setting as their wild-type counterparts. Conversely, the H4 tailless nucleosomes display widespread changes of the AT content near SHL ±1 and SHL ±2, where the H4 N-tails interact with DNA. Furthermore, a substantial number of H4 tailless nucleosomes exhibit rotational setting opposite to that of the wild-type nucleosomes. Thus, our findings strongly suggest that the histone N-tails are operative in selection of nucleosome positions, which may have wide-ranging implications for epigenetic modulation of chromatin states.

p53 binding to nucleosomal DNA depends on the rotational positioning of DNA response element.

The sequence-specific binding to DNA is crucial for the p53 tumor suppressor function. To investigate the constraints imposed on p53-DNA recognition by nucleosomal organization, we studied binding of the p53 DNA binding domain (p53DBD) and full-length wild-type p53 protein to a single p53 response element (p53RE) placed near the nucleosomal dyad in six rotational settings. We demonstrate that the strongest p53 binding occurs when the p53RE in the nucleosome is bent in the same direction as observed for the p53-DNA complexes in solution and in co-crystals. The p53RE becomes inaccessible, however, if its orientation in the core particle is changed by approximately 180 degrees. Our observations indicate that the orientation of the binding sites on a nucleosome may play a significant role in the initial p53-DNA recognition and subsequent cofactor recruitment.

Distinctive sequence patterns in metazoan and yeast nucleosomes: implications for linker histone binding to AT-rich and methylated DNA.

Linker histones (LHs) bind to the DNA entry/exit points of nucleosomes and demonstrate preference for AT-rich DNA, although the recognized sequence patterns remain unknown. These patterns are expected to be more pronounced in metazoan nucleosomes with abundant LHs, compared to yeast nucleosomes with few LHs. To test this hypothesis, we compared the nucleosome core particle (NCP) sequences from chicken, Drosophila and yeast, extending them by the flanking sequences extracted from the genomes. We found that the known approximately 10-bp periodic oscillation of AT-rich elements goes beyond the ends of yeast nucleosomes, but is distorted in metazoan sequences where the 'out-of-phase' AT-peaks appear at the NCP ends. The observed difference is likely to be associated with sequence-specific LH binding. We therefore propose a new structural model for LH binding to metazoan nucleosomes, postulating that the highly conserved nonpolar 'wing' region of the LH globular domain (tetrapeptide GVGA) recognizes AT-rich fragments through hydrophobic interactions with the thymine methyl groups. These interactions lead to DNA bending at the NCP ends and formation of a 'stem-like' structure. The same mechanism accounts for the high affinity of LH to methylated DNA-a feature critical for stabilization of the higher-order structure of chromatin and for repression of transcription.

nuScore: a web-interface for nucleosome positioning predictions.

UNLABELLED: Sequence-directed mapping of nucleosome positions is of major biological interest. Here, we present a web-interface for estimation of the affinity of the histone core to DNA and prediction of nucleosome arrangement on a given sequence. Our approach is based on assessment of the energy cost of imposing the deformations required to wrap DNA around the histone surface. The interface allows the user to specify a number of options such as selecting from several structural templates for threading calculations and adding random sequences to the analysis. AVAILABILITY: The nuScore interface is freely available for use at http://compbio.med.harvard.edu/nuScore. CONTACT: peter_park@harvard.edu; tolstorukov@gmail.com SUPPLEMENTARY INFORMATION: The site contains user manual, description of the methodology and examples.

Satb1 integrates DNA binding site geometry and torsional stress to differentially target nucleosome-dense regions.

The Satb1 genome organizer regulates multiple cellular and developmental processes. It is not yet clear how Satb1 selects different sets of targets throughout the genome. Here we have used live-cell single molecule imaging and deep sequencing to assess determinants of Satb1 binding-site selectivity. We have found that Satb1 preferentially targets nucleosome-dense regions and can directly bind consensus motifs within nucleosomes. Some genomic regions harbor multiple, regularly spaced Satb1 binding motifs (typical separation ~1 turn of the DNA helix) characterized by highly cooperative binding. The Satb1 homeodomain is dispensable for high affinity binding but is essential for specificity. Finally, we find that Satb1-DNA interactions are mechanosensitive. Increasing negative torsional stress in DNA enhances Satb1 binding and Satb1 stabilizes base unpairing regions against melting by molecular machines. The ability of Satb1 to control diverse biological programs may reflect its ability to combinatorially use multiple site selection criteria.

A novel roll-and-slide mechanism of DNA folding in chromatin: implications for nucleosome positioning.

How eukaryotic genomes encode the folding of DNA into nucleosomes and how this intrinsic organization of chromatin guides biological function are questions of wide interest. The physical basis of nucleosome positioning lies in the sequence-dependent propensity of DNA to adopt the tightly bent configuration imposed by the binding of the histone proteins. Traditionally, only DNA bending and twisting deformations are considered, while the effects of the lateral displacements of adjacent base pairs are neglected. We demonstrate, however, that these displacements have a much more important structural role than ever imagined. Specifically, the lateral Slide deformations observed at sites of local anisotropic bending of DNA define its superhelical trajectory in chromatin. Furthermore, the computed cost of deforming DNA on the nucleosome is sequence-specific: in optimally positioned sequences the most easily deformed base-pair steps (CA:TG and TA) occur at sites of large positive Slide and negative Roll (where the DNA bends into the minor groove). These conclusions rest upon a treatment of DNA that goes beyond the conventional ribbon model, incorporating all essential degrees of freedom of "real" duplexes in the estimation of DNA deformation energies. Indeed, only after lateral Slide displacements are considered are we able to account for the sequence-specific folding of DNA found in nucleosome structures. The close correspondence between the predicted and observed nucleosome locations demonstrates the potential advantage of our "structural" approach in the computer mapping of nucleosome positioning.