Sequence-dependent bending of DNA and phasing of nucleosomes

Conformational analysis has revealed anisotropic flexibility of the B-DNA double helix: it bends most easily into the grooves, being the most rigid when bent in a perpendicular direction. This result implies that DNA in a nucleosome is curved by means of relatively sharp bends ("mini-kinks") which are directed into the major and minor grooves alternatively and separated by 5-6 base pairs. The "mini-kink" model proved to be in keeping with the x-ray structure of the B-DNA dodecamer resolved later, which exhibits two "annealed kinks", also directed into the grooves. The anisotropy of B DNA is sequence-dependent: the pyrimidine-purine dimers (YR) favor bending into the minor groove, and the purine-pyrimidine dinucleotides (RY), into the minor one. The RR and YY dimers appear to be the most rigid dinucleotides. Thus, a DNA fragment consisting of the interchanging oligopurine and oligopyrimidine blocks 5-6 base pairs long should manifest a spectacular curvature in solution. Similarly, a nucleotide sequence containing the RY and YR dimers separated by a half-pitch of the double helix is the most suitable for wrapping around the nucleosomal core. Analysis of the numerous examples demonstrating the specific alignment of nucleosomes on DNA confirms this concept. So, the sequence-dependent "mechanical" properties of the double helix influence the spatial arrangement of DNA in chromatin.     

Sequence-Dependent Bending of DNA and Phasing of Nucleosomes

Abstract

Conformational analysis has revealed anisotropic flexibility of the B-DNA double helix: it bends most easily into the grooves, being the most rigid when bent in a perpendicular direction. This result implies that DNA in a nucleosome is curved by means of relatively sharp bends (“mini-kinks”) which are directed into the major and minor grooves alternatively and separated by 5–6 base pairs. The “mini-kink” model proved to be in keeping with the x-ray structure of the B-DNA dodecamer resolved later, which exhibits two “annealed kinks”, also directed into the grooves.

The anisotropy of B DNA is sequence-dependent: the pyrimidine-purine dimers (YR) favor bending into the minor groove, and the purine-pyrimidine dinucleotides (RY), into the minor one. The RR and YY dimers appear to be the most rigid dinucleotides. Thus, a DNA fragment consisting of the interchanging oligopurine and oligopyrmidine blocks 5–6 base pairs long should manifest a spectacular curvature in solution.

Similarly, a nucleotide sequence containing the RY and YR dimers separated by a half-pitch of the double helix is the most suitable for wrapping around the nucleosomal core. Analysis of the numerous examples demonstrating the specific alignment of nucleosomes on DNA confirms this concept. So, the sequence-dependent “mechanical” properties of the double helix influence the spatial arrangement of DNA in chromatin.     

Analysis of local helix bending in crystal structures of DNA oligonucleotides and DNA-protein complexes.

Sequence-dependent bending of the helical axes in 112 oligonucleotide duplex crystal structures resident in the Nucleic Acid Database have been analyzed and compared with the use of bending dials, a computer graphics tool. Our analysis includes structures of both A and B forms of DNA and considers both uncomplexed forms of the double helix as well as those bound to drugs and proteins. The patterns in bending preferences in the crystal structures are analyzed by base pair steps, and emerging trends are noted. Analysis of the 66 B-form structures in the Nucleic Acid Database indicates that uniform trends within all pyrimidine-purine and purine-pyrimidine steps are not necessarily observed but are found particularly at CG and GC steps of dodecamers. The results support the idea that AA steps are relatively straight and that larger roll bends occur at or near the junctions of these A-tracts with their flanking sequences. The data on 16 available crystal structures of protein-DNA complexes indicate that the majority of the DNA bends induced via protein binding are sharp localized kinks. The analysis of the 30 available A-form DNA structures indicates that these structures are also bent and show a definitive preference for bending into the deep major groove over the shallow minor groove.     

DNA bending: the prevalence of kinkiness and the virtues of normality.

DNA bending in 86 complexes with sequence-specific proteins has been examined using normal vector plots, matrices of normal vector angles between all base pairs in the helix, and one-digit roll/slide/twist tables. FREEHELIX, a new program especially designed to analyze severely bent and kinked duplexes, generates the foregoing quantities plus local roll, tilt, twist, slide, shift and rise parameters that are completely free of any assumptions about an overall helix axis. In nearly every case, bending results from positive roll at pyrimidine-purine base pair steps: C-A (= T-G), T-A, or less frequently C-G, in a direction that compresses the major groove. Normal vector plots reveal three well-defined types of bending among the 86 examples: (i) localized kinks produced by positive roll at one or two discrete base pairs steps, (ii) three-dimensional writhe resulting from positive roll at a series of adjacent base pairs steps, or (iii) continuous curvature produced by alternations of positive and negative roll every 5 bp, with side-to-side zig-zag roll at intermediate position. In no case is tilt a significant component of the bending process. In sequences with two localized kinks, such as CAP and IHF, the dihedral angle formed by the three helix segments is a linear function of the number of base pair steps between kinks: dihedral angle = 36 degrees x kink separation. Twenty-eight of the 86 examples can be described as major bends, and significant elements in the recognition of a given base sequence by protein. But even the minor bends play a role in fine-tuning protein/DNA interactions. Sequence-dependent helix deformability is an important component of protein/DNA recognition, alongside the more generally recognized patterns of hydrogen bonding. The combination of FREEHELIX, normal vector plots, full vector angle matrices, and one-digit roll/slide/twist tables affords a rapid and convenient method for assessing bending in DNA.     

Atomic force microscopy study of DNA flexibility on short length scales: smooth bending versus kinking

The apparently anomalous flexibility of DNA on short length scales has attracted a lot of attention in recent years. We use atomic force microscopy (AFM) in solution to directly study the DNA bending statistics for small lengths down to one helical turn. The accuracy of experimental estimates could be improved due to a large data volume and a refined algorithm for image processing and measuring bend angles. It is found that, at length scales beyond two helical turns (7 nm), DNA is well described by the harmonic worm-like chain (WLC) model with the bending persistence length of 56 nm. Below this threshold, the AFM data are also described by the WLC model assuming that the accuracy of measured bend angles is limited by the physical width of the double helix. We conclude that the double helical DNA behaves as a uniform elastic rod even at very short length scales. Strong bends due to kinks, melting bubbles and other deviations from the WLC model are statistically negligible.     

Anisotropic flexibility of DNA depends on the base sequence. Conformation calculations of double-stranded tetranucleotides AAAA:TTTT, (AATT)2, (TTAA)2, GGGG:CCCC, (GGCC)2, (CCGG)2

The bending flexibility of six tetramers was studied in an assumption that they were extended in the both directions by regular double helices. The bends of B-DNA in different directions were considered. The stiffness of the B-DNA double helix when bent into the both grooves proved to be less pronounced than in the perpendicular direction by the order of magnitude. Such an anisotropy is a feature of the sugar-phosphate backbone structure. The calculated fluctuations of the DNA bending along the dyad axis, 5-7 degrees, are in agreement with the experimental value of DNA persistence length. Anisotropy of the double helix is sequence-dependent: most easily bent into the minor groove are the tetramers with purine-pyrimidine dimer (RY) in the middle. In contrast, YR dinucleotides prefer bending into the major groove, moreover, they have an equilibrium bend of 6-12 degrees into this groove. The above inequality is caused by the stacking interaction of the bases. The bend in the central dimers is distributed to some extent between the adjacent links, though the main fraction of the bend remains within the central link. Variation of the sugar-phosphate geometry in the bent helix is unessential, so that DNA remains within the limits of the B-family of forms: namely, when the helical axis is bent by 20 degrees the backbone dihedral angles vary by no more than 15 degrees. The obtained results are in accord with the X-ray structure of B-DNA dodecamer; they further substantiate our earlier model of DNA wrapping in the nucleosome by means of "mini-kinks" separated by a half-pitch of the double helix, i.e. by 5-6 b. p. Sequence-dependent anisotropy of DNA presumably dictates the three-dimensional structure of DNA in solution as well. We have found that nonrandom allocation of YR dimers leads to the systematic bends in the equilibrium structure of certain DNA fragments. To the four "Calladine rules" two more can be added: the minor-groove steric clash of purines in the YR sequences are avoided by: (1) bending of the helix into the major groove; (2) increasing the distance between the base pairs (stretching the double helix).     

DNA-bending on ligand binding; change of DNA persistence length.

This paper simulates the helix-characteristic changes of apparent DNA persistence length caused by randomly distributed helix bends as induced, e.g., by DNA-bound ligand molecules. The parameters varied are the constant angle gamma of helix bending and the size alpha of the DNA drug binding site, but also the degree of DNA-ligand binding cooperativity and the helix-unwinding angle. If the size of the binding site is comparable with the helix pitch, the influence of phasing between helix bends and helix screw upon the apparent persistence length is obvious. In the accompanying paper experimental data are analyzed in terms of this theoretical background.     

Mapping the transition state for DNA bending by IHF

How DNA-bending proteins recognize their specific sites on DNA remains elusive, particularly for proteins that use indirect readout, which relies on sequence-dependent variations in DNA flexibility/bendability. The question remains as to whether the protein bends the DNA (protein-induced bending) or, alternatively, "prebent" DNA conformations are thermally accessible, which the protein captures to form the specific complex (conformational capture). To distinguish between these mechanisms requires characterization of reaction intermediates and, in particular, snapshots of the transition state along the recognition pathway. We present such a snapshot, from measurements of DNA bending dynamics in complex with Escherichia coli integration host factor (IHF), an architectural protein that bends specific sites on λ-DNA in a U-turn by creating two sharp kinks in DNA. Fluorescence resonance energy transfer measurements in response to laser temperature-jump perturbation monitor DNA bending. We find that nicks or mismatches that enhance DNA flexibility at the site of the kinks show 3- to 4-fold increase in DNA bending rates that reflect a 4- to 11-fold increase in binding affinities, while sequence modifications away from the kink sites, as well as mutations in IHF designed to destabilize the complex, have negligible effect on DNA bending rates despite >250-fold decrease in binding affinities. These results support the scenario that the bottleneck in the recognition step for IHF is spontaneous kinking of cognate DNA to adopt a partially prebent conformation and point to conformational capture as the underlying mechanism of initial recognition, with additional protein-induced bending occurring after the transition state.     

HMGB binding to DNA: single and double box motifs

High mobility group (HMG) proteins are nuclear proteins believed to significantly affect DNA interactions by altering nucleic acid flexibility. Group B (HMGB) proteins contain HMG box domains known to bind to the DNA minor groove without sequence specificity, slightly intercalating base pairs and inducing a strong bend in the DNA helical axis. A dual-beam optical tweezers system is used to extend double-stranded DNA (dsDNA) in the absence as well as presence of a single box derivative of human HMGB2 [HMGB2(box A)] and a double box derivative of rat HMGB1 [HMGB1(box A+box B)]. The single box domain is observed to reduce the persistence length of the double helix, generating sharp DNA bends with an average bending angle of 99 ± 9° and, at very high concentrations, stabilizing dsDNA against denaturation. The double box protein contains two consecutive HMG box domains joined by a flexible tether. This protein also reduces the DNA persistence length, induces an average bending angle of 77 ± 7°, and stabilizes dsDNA at significantly lower concentrations. These results suggest that single and double box proteins increase DNA flexibility and stability, albeit both effects are achieved at much lower protein concentrations for the double box. In addition, at low concentrations, the single box protein can alter DNA flexibility without stabilizing dsDNA, whereas stabilization at higher concentrations is likely achieved through a cooperative binding mode.     

Theoretical Analysis of Disruptions in DNA Minicircles

Under sufficient bending stress, which appears in DNA minicircles and small DNA loops, the double helix experiences local disruptions of its regular structure. We developed a statistical-mechanical treatment of the disruptions in DNA minicircles, studied experimentally by Du et al. The model of disruptions used in our Monte Carlo simulation of minicircle conformations specifies these conformations by three parameters: DNA bend angle at the disruption, θ_d; local DNA unwinding caused by the disruption; and the free energy associated with the disruption in the unstressed double helix, G_d. The model is applicable to any structural type of disruption, kinks or opening of single basepairs. The simulation shows that accounting for both torsional and bending deformation associated with the disruptions is very important for proper analysis. We obtained a relationship between values of G_d and θ_d under which the simulation results are compatible with the experimental data. The relationship suggests that the free energy of basepair opening, which includes flipping out both bases, is significantly higher than the generally accepted value. The model is also applied to the analysis of j-factors of very short DNA fragments.     

Phased psoralen cross-links do not bend the DNA double helix

Although the chemical reaction of psoralens with nucleic acids is well understood, the structure of psoralen-DNA cross-linked products is still not clear. Model building studies base on the crystal structure of the psoralen-thymine monoadduct suggest that each cross-link bends the DNA double helix by 46.5 degrees [Pearlman, D. A., Holbrook, S. R., Pirkle, D. H., & Kim, S.-H. (1985) Science (Washington, D.C.) 227, 1304-1308]. On the other hand, Sinden and Hagerman [Sinden, R. R., & Hagerman, P. J. (1984) Biochemistry 23, 6299-6303] find that, in solution, psoralen cross-linked DNA is not bent. Here we use gel electrophoresis to test the validity of the current models. We have synthesized a series of DNA fragments (21-24 base pairs in length), each containing one unique T-A site for 4'-(hydroxymethyl)-4,5',8-trimethylpsoralen (HMT) cross-linking. Because of an estimated 28 degrees unwinding of the helix by HMT [Wiesehahn, G., & Hearst, J. E. (1978) Proc. Natl. Acad. Sci. U.S.A. 75, 2703-2707], one expects that the 22-bp cross-linked fragment will be repeated nearly in phase with the average helical screw when multimerized. In that sequence ligation will maximally amplify any deformation to the double helix. We find that the ligated multimers of cross-linked DNA migrate close to the multimers of non-cross-linked DNA on polyacrylamide gels. Our observations place an upper limit of 10 degrees on DNA bending induced by psoralen cross-linking and indicate unwinding by about 1 bp, as well as stiffening of the double helix. These properties are not unexpected for classical intercalators.     

Sequence-dependent anisotropic flexibility of B-DNA. A conformational study

Bending flexibility of the six tetrameric duplexes was investigated d(AAAA):d(TTTT), d(AATT)2, d(TTAA)2, d(GGGG):d(CCCC), d(GGCC)2 and d(CCGG)2,. The tetramers were extended in the both directions by regular double helices. The stiffness of the B-DNA double helix when bent into the both grooves proved to be less than that in the perpendicular direction by an order of magnitude. Such an anisotropy is a property of the sugar-phosphate backbone structure. The calculated fluctuations of the DNA bending along the dyad axis, 5-7 degree, are in agreement with experimental value of the DNA persistence length. Anisotropy of the double helix is sequence-dependent: most easily bent into the minor groove are the tetramers with purine-pyrimidine dimer (RY) in the middle. In contrast, YR dinucleotides prefer bending into the major groove. Moreover, they have an equilibrium bend of 6-12 degree into this groove. The above inequality is caused by stacking interaction of the bases. The bend in the central dimer is distributed to some extent between the adjacent links, though the main fraction of the bend remains within the central link. Variation of the sugar-phosphate geometry in the bent helix is inessential, so that DNA remains within the B-family of forms: namely, when the helical axis is bent by 20 degree. the backbone dihedral angles vary by no more than 15 degree. The obtained results are in accord with x-ray structure of the B-DNA dodecamer; they further substantiate our early model of DNA wrapping in the nucleosome by means of "mini-kinks" separated by a half-pitch of the double helix, i.e. by 5-6 b.p. Sequence-dependent anisotropy of DNA presumably dictates the three-dimensional structure of DNA in solution as well. We have found that nonrandom allocation of YR dimers leads to the systematic bends in equilibrium structure of certain DNA fragments.     

Methylation strongly enhances DNA bending in the replication origin region of the Escherichia coli chromosome

Summary Two-dimensional gel electrophoresis, at high and low temperatures, and gel mobilities of circularly permuted DNA segments showed a large bending locus about 50 bp downstream from the right border of the 245 by oriC box, a minimal essential region of autonomous replication on the Escherichia coli chromosome. Bending was strongly enhanced by Dam methylation. In DNA from a Dam^– strain, the mobility anomaly arising from altered conformation was much reduced, but was raised to the original level by methylation in vivo or in vitro. Enhancement of the mobility anomaly was also observed by hybrid formation of the Dam^– strand with the Dam⁺ strand. Near the bending center, GATC, the target of Dam methylase, occurs seven times arranged essentially on the same face of the helix with 10.5 by per turn. We concluded that small bends at each Dam site added up to the large bending detectable by gel electrophoresis.     

The solution structure of the Sac7d/DNA complex: a small-angle X-ray scattering study.

Small-angle X-ray scattering has been used to study the structure of the multimeric complexes that form between double-stranded DNA and the archaeal chromatin protein Sac7d from Sulfolobus acidocaldarius. Scattering data from complexes of Sac7d with a defined 32-mer oligonucleotide, with poly[d(GC)], and with E. coli DNA indicate that the protein binds along the surface of an extended DNA structure. Molecular models of fully saturated Sac7d/DNA complexes were constructed using constraints from crystal structure and solution binding data. Conformational space was searched systematically by varying the parameters of the models within the constrained set to find the best fits between the X-ray scattering data and simulated scattering curves. The best fits were obtained for models composed of repeating segments of B-DNA with sharp kinks at contiguous protein binding sites. The results are consistent with extrapolation of the X-ray crystal structure of a 1:1 Sac7d/octanucleotide complex [Robinson, H., et al. (1998) Nature 392, 202-205] to polymeric DNA. The DNA conformation in our multimeric Sac7d/DNA model has the base pairs tilted by about 35 degrees and displaced 3 A from the helix axis. There is a large roll between two base pairs at the protein-induced kink site, resulting in an overall bending angle of about 70 degrees for Sac7d binding. Regularly repeating bends in the fully saturated complex result in a zigzag structure with negligible compaction of DNA. The Sac7d molecules in the model form a unique structure with two left-handed helical ribbons winding around the outside of the right-handed duplex DNA.     

Analysis of the nucleotide sequence and electrostatic potential distribution in the <Emphasis Type="Italic">Escherichia coli</Emphasis> genome

Analysis of the oligonucleotide composition of the complete E. coli genome and its σ⁷⁰-specific promoters shows that promoter DNA mainly contains AT-rich hexanucleotides that have functionally important physical properties (propensity to form ‘low-melting’ regions and helix bends). Comparative analysis of the electrostatic characteristics reveals that hexanucleotides corresponding to more electronegative surroundings are mostly found in promoter DNA.     

Turn of Promotor DNA by cAMP Receptor Protein Characterized by Bead Model Simulation of Rotational Diffusion

Abstract

The rotation diffusion of DNA double helices and their complexes with the cAMP receptor protein (CRP) has been simulated by bead models, in order to derive information on their structure in solution by comparison with results obtained from dichroism decay measurements. Straight DNA double helices are simulated by linear, rigid strings of overlapping beads. The radius of the beads and the length of the string are increased simultaneously by the same increments from initial outer dimensions derived from crystallographic data to final values, which are fitted to experimental rotation time constants observed for short DNA fragments (< 100 bp). The final values reflect the solvated structure with the same ‘solvation layer’ added in all three dimensions. The protein is simulated by overlapping beads, which are assembled to a structure very similar to that found by x-ray crystallography. Complexes of the protein with DNA are formed with the centres of palindromic DNA sites at the centre of the two helix- turn-helix-motifs of the protein with some overlap of the two components. Simulation of the experimental data obtained for CRP complexes with specific DNA in the presence of cAMP requires strong bending of the double helices. According to our simulation the DNA is almost completely wrapped around the protein both in the complexes with a 62 bp fragment containing the standard CRP site and with a 80 bp fragment containing the second binding site of the lac operon. Simulations of the data obtained for a 203 bp fragment with both binding sites suggest that the two bound CRP proteins are in contact with each other and that the DNA is wrapped around the two protein dimers. A stereochemical model is suggested with a tetrahedral arrangement of the four protein subunits, which provides the advantage that two binding sites of the protein formed by two subunits each are located favorable for tight contacts to two binding sites on bent DNA provided that the DNA sites are separated by an integer number of helix turns. In summary, the simulations demonstrate strong bending, which can be reflected by an arc radius in the range around 50 Å. According to these data the overall bending angle of our longest DNA fragment is approximately 180°, and thus the protruding ends are sufficiently close to each other such that RNA polymerase, for example, could contact both helical segments.     

Bending modes of DNA directly addressed by cryo-electron microscopy of DNA minicircles

We use cryo-electron microscopy (cryo-EM) to study the 3D shapes of 94-bp-long DNA minicircles and address the question of whether cyclization of such short DNA molecules necessitates the formation of sharp, localized kinks in DNA or whether the necessary bending can be redistributed and accomplished within the limits of the elastic, standard model of DNA flexibility. By comparing the shapes of covalently closed, nicked and gapped DNA minicircles, we conclude that 94-bp-long covalently closed and nicked DNA minicircles do not show sharp kinks while gapped DNA molecules, containing very flexible single-stranded regions, do show sharp kinks. We corroborate the results of cryo-EM studies by using Bal31 nuclease to probe for the existence of kinks in 94-bp-long minicircles.     

Sequence-Dependent Anisotropic Flexibility of B-DNA A Conformational Study

Abstract

Bending flexibility of the six tetrameric duplexes was investigated d(AAAA):d(TTTT), d(AATT)₂, d(TTAA) ₂, d(GGGG):d(CCCC), d(GGCC) ₂ and d(CCGG) ₂. The tetramers were extended in the both directions by regular double helices. The stiffness of the B-DNA double helix when bent into the both grooves proved to be less than that in the perpendicular direction by an order of magnitude. Such an anisotropy is a property of the sugar-phosphate backbone structure. The calculated fluctuations of the DNA bending along the dyad axis, 5–7°, are in agreement with experimental value of the DNA persistence length.

Anisotropy of the double helix is sequence-dependent: most easily bent into the minor groove are the tetramers with purine-pyrimidine dimer (RY) in the middle. In contrast, YR dinucleotides prefer bending into the major groove. Moreover, they have an equilibrium bend of 6–12° into this groove. The above inequality is caused by stacking interaction of the bases.

The bend in the central dimer is distributed to some extent between the adjacent links, though the main fraction of the bend remains within the central link. Variation of the sugar-phosphate geometry in the bent helix is inessential, so that DNA remains within the B-family of forms: namely, when the helical axis is bent by 20°, the backbone dihedral angles vary by no more than 15°.

The obtained results are in accord with x-ray structure of the B-DNA dodecamer; they further substantiate our early model of DNA wrapping in the nucleosome by means of “mini-kinks” separated by a half-pitch of the double helix, i.e. by 5–6 b.p. Sequence-dependent anisotropy of DNA presumably dictates the three-dimentional structure of DNA in solution as well. We have found that nonrandom allocation of YR dimers leads to the systematic bends in equilibrium structure of certain DNA fragments.     

Determination of the extent of DNA bending by an adenine-thymine tract

We determined the magnitude of the bend induced in DNA by an adenine-thymine tract by measuring the rate of cyclization of DNA oligonucleotides containing phased A tracts. A series of linear multimers with 2-bp single-stranded ends, in which the (A.T)6 tracts are separated by CG2-3C sequences and are positioned 10 and 11 bp apart alternately, were prepared from 21 bp long synthetic duplexed deoxyoligonucleotides. The cyclization rates of the multimers (105-210 bp) and the bimolecular association rate of the 84 bp long multimer were measured in the presence of DNA ligase. From the rate constants of the cyclization and bimolecular association reactions, ring closure probabilities were obtained for the multimers. The systematically bent molecules were simulated by Monte Carlo methods, and the ring closure probabilities were calculated for a given set of junction bend angles. By comparing the calculated values of ring closure probabilities to experimental values and adjusting the junction bend angles to fit experimental values, the extent of bending at the junctions (or the extent of bending for an adenine tract) was determined. We conclude that an A6 tract bends the DNA helix by 17-21 degrees.