Spontaneous sharp bending of double-stranded DNA

Sharply bent DNA is essential for gene regulation in prokaryotes and is a major feature of eukaryotic nucleosomes and viruses. The explanation normally given for these phenomena is that specific proteins sharply bend DNA by application of large forces, while the DNA follows despite its intrinsic inflexibility. Here we show that DNAs that are 94 bp in length-comparable to sharply looped DNAs in vivo-spontaneously bend into circles. Proteins can enhance the stability of such loops, but the loops occur spontaneously even in naked DNA. Random DNA sequences cyclize 10(2)-10(4) times more easily than predicted from current theories of DNA bending, while DNA sequences that position nucleosomes cyclize up to 10(5) times more easily. These unexpected results establish DNA as an active participant in the formation of looped regulatory complexes in vivo, and they point to a need for new theories of DNA bending.     

A stitch in time: Efficient computation of genomic DNA melting bubbles

Background  

It is of biological interest to make genome-wide predictions of the locations of DNA melting bubbles using statistical mechanics models. Computationally, this poses the challenge that a generic search through all combinations of bubble starts and ends is quadratic.     

The role of DNA bending in type IIA topoisomerase function

Type IIA topoisomerases control DNA supercoiling and separate newly replicated chromosomes using a complex DNA strand cleavage and passage mechanism. Structural and biochemical studies have shown that these enzymes sharply bend DNA by as much as 150°; an invariant isoleucine, which has been seen structurally to intercalate between two base pairs outside of the DNA cleavage site, has been suggested to promote deformation. To test this assumption, we examined the role of isoleucine on DNA binding, bending and catalytic activity for a bacterial type IIA topoisomerase, Escherichia coli topoisomerase IV (topo IV), using a combination of site-directed mutagenesis and biochemical assays. Our data show that alteration of the isoleucine (Ile₁₇₂) did not affect the basal ATPase activity of topo IV or its affinity for DNA. However, the amino acid was important for DNA bending, DNA cleavage and supercoil relaxation. Moreover, an ability to bend DNA correlated with efficacy with which nucleic acid substrates stimulate ATP hydrolysis. These data show that DNA binding and bending by topo IV can be uncoupled, and indicate that the stabilization of a highly curved DNA geometry is critical to the type IIA topoisomerase catalytic cycle.     

Supercoiling-induced DNA bending

Local DNA bending is a critical factor for numerous DNA functions including recognition of DNA by sequence-specific regulatory binding proteins. Negative DNA supercoiling increases both local and global DNA dynamics, and this dynamic flexibility can facilitate the formation of DNA-protein complexes. We have recently shown that apexes of supercoiled DNA molecules are sites that can promote the formation of an alternative DNA structure, a cruciform, suggesting that these positions in supercoiled DNA are under additional stress and perhaps have a distorted DNA geometry. To test this hypothesis, we used atomic force microscopy to directly measure the curvature of apical positions in supercoiled DNA. The measurements were performed for an inherently curved sequence formed by phased A tracts and a region of mixed sequence DNA. For this, we used plasmids in which an inverted repeat and A tract were placed at precise locations relative to each other. Under specific conditions, the inverted repeat formed a cruciform that was used as a marker for the unambiguous identification of the A tract location. When the A tract and cruciform were placed diametrically opposite, this yielded predominantly nonbranched plectonemic molecules with an extruded cruciform and A tract localized in the terminal loops. For both the curved A tract and mixed sequence nonbent DNA, their localization to an apex increased the angle of bending compared to that expected for DNA unconstrained in solution. This is consistent with increased helical distortion at an apical bend.     

Mismatches and bubbles in DNA

Single mismatches in the DNA double helix form nucleation sites for bubbles. Although the overall melting temperature of the duplex is affected to different degrees depending on the probe length, the statistical weights of the bubble states around the defect are always strongly affected. Here we show experimentally that a single mismatch has indeed a dramatic effect on the distribution of intermediate (bubble) states in the melting transition of DNA oligomers. For probe lengths in the range 20-40 bases, the mismatch transforms a transition with many intermediates into a nearly two-state transition. One surprising consequence is the existence of a regime where the sensitivity of a mismatch detection assay based on monitoring intermediate states would increase with probe length. Our results provide experimental constraints on how mismatches should be implemented in models of DNA melting, such as the widely used thermodynamic nearest neighbor model, to which we compare our data.     

Site-specific topoisomerase I-mediated DNA cleavage induced by nogalamycin: a potential role of ligand-induced DNA bending at a distal site

Many DNA binding ligands (e.g., nogalamycin, actinomycin D, terbenzimidazoles, indolocarbazoles, nitidine, and coralyne) and various types of DNA lesions (e.g., UV dimers, DNA mismatches, and abasic sites) are known to stimulate topoisomerase I-mediated DNA cleavage. However, the mechanism(s) by which these covalent and noncovalent DNA interactions stimulate topoisomerase I-mediated DNA cleavage remains unclear. Using nogalamycin as a model, we have studied the mechanism of ligand-induced topoisomerase I-mediated DNA cleavage. We show by both mutational and DNA footprinting analyses that the binding of nogalamycin to an upstream site (from position -6 to -3) can induce highly specific topoisomerase I-mediated DNA cleavage. Substitution of this nogalamycin binding site with a DNA bending sequence (A(5)) stimulated topoisomerase I-mediated DNA at the same site in the absence of nogalamycin. Replacement of the A(5) sequence with a disrupted DNA bending sequence (A(2)TA(2)) significantly reduced the level of topoisomerase I-mediated DNA cleavage. These results, together with the known DNA bending property of nogalamycin, suggest that the nogalamycin-DNA complex may provide a DNA structural bend to stimulate topoisomerase I-mediated DNA cleavage.     

Dynamic bending rigidity of DNA

Rapidly relaxing components in the decay of the transient electric dichroism of DNA restriction fragments were reported by Diekmann et al. [(1982) Biophys. Chem. 15, 263-270] and P?rschke et al. [(1987) Biopolymers 26, 1971-1974]. These are analyzed using a new normal mode theory for weakly bending rods and assigned to bending. The longest bending relaxation times for fragments with 95-250 base pairs coincide with the theoretical curve calculated for a dynamic bending rigidity corresponding to a dynamic persistence length Pd = 2100 A. Analysis of the relative amplitudes of fast and slow components following weak orienting pulses is also consistent with a rather large dynamic persistence length. The enhancement of the relative amplitude of the fast component in large electric fields is attributed to steady-state bending of initially perpendicular DNAs by the field. Several reasons are proposed why the dynamic bending rigidity is 4 times larger than the apparent static bending rigidity inferred from equilibrium persistence length measurements on the same fragments.     

Theory of DNA melting curves

Exact algorithms for the calculation of melting curves of heterogeneous DNA with N base pairs apparently require computer time proportional to N². However, it is shown that a decomposition of the loop entropy factor into a sum of I exponential functions (1) gives an extremely accurate approximation to the loop entropy factor for small values of I and (2) makes the computer time for the exact algorithms proportional to I·N. In effect, exact results for melting curves and lengths of helix or coil stretches are obtained with computer time comparable to that required for the Frank-Kamenetskii approximation. The remarkable accuracy of the latter for the fraction of helical content (errors of 0.01–0.05) is confirmed, but appreciably larger errors are found for the lengths of helix or coil stretches (typical errors of 30–100%).     

DNA sequence determinant for Flp-induced DNA bending

The Flp site-specific recombinase from Saccharomyces cerevisiae induces DNA bending upon interaction with the Flp recognition target (FRT) site. The minimal FRT site comprises the inverted a and b binding elements, which flank a central 8 bp core region. The DNA bend in a complex of two Flp monomers bound to the FRT site is located in the middle of the core region. When the central AT basepair was replaced with a CG, the DNA bend was positioned at the outside end of the core region adjacent to the a binding element. The other basepairs surrounding the central AT basepair were not important to the position of Flp-induced bends. The change also decreased Flp-mediated cleavage of the top strand of the FRT site and increased Flp-mediated cleavage of the bottom strand. The overall recombination proficiency of the site was impaired. We conclude that the central AT basepair provides a point of flexure in the FRT site, which Flp uses to position the bend in dimeric Flp–DNA complexes, and that the structure of the core DNA influences the functionality of the site.     

Intramolecular DNA melting between stable helical segments: melting theory and metastable states.

Melting of DNA in a segment bounded at both ends by regions of greater stability during electrophoresis in denaturing gradient gels show complex properties, not accommodated with standard melting theory. Compact bands of some DNA molecules become anomalously broadened at the retardation level in a denaturing gradient, or double bands may appear in a uniform denaturant concentration. These properties are associated only with molecules for which the distribution of stability calculated by the Poland-Fixman-Freire algorithms indicates that the region of lowest stability does not extend to an end of the molecule. Retention of helicity at the ends is shown by the difference in the effect of base substitution in the end domains and in the least stable domain. Both the appearance of double bands and band broadening can be explained by invoking a hypothetical metastable intermediate in melting, which is converted into the equilibrium melted form at a relatively slow rate, depending on both denaturant concentration and field strength. A kinetic model permits plausible rate constants to be inferred from the patterns. Despite the increased band width, sequence variants with base changes in the least stable domain result in readily detectable band shifts in the gradient.     

Determination of melting temperature and temperature melting range for DNA with multi-peak differential melting curves

Many factors that change the temperature position and interval of the DNA helix–coil transition often also alter the shape of multi-peak differential melting curves (DMCs). For DNAs with a multi-peak DMC, there is no agreement on the most useful definition for the melting temperature, T_m, and temperature melting width, ΔT, of the entire DNA transition. Changes in T_m and ΔT can reflect unstable variation of the shape of the DMC as well as alterations in DNA thermal stability and heterogeneity. Here, experiments and computer modeling for DNA multi-peak DMCs varying under different factors allowed testing of several methods of defining T_m and ΔT. Indeed, some of the methods give unreasonable “jagged” T_m and ΔT dependences on varying relative concentration of DNA chemical modifications (r_b), [Na⁺], and GC content. At the same time, T_m determined as the helix–coil transition average temperature, and ΔT, which is proportional to the average absolute temperature deviation from this temperature, are suitable to characterize multi-peak DMCs. They give smoothly varying theoretical and experimental dependences of T_m and ΔT on r_b, [Na⁺], and GC content. For multi-peak DMCs, T_m value determined in this way is the closest to the thermodynamic melting temperature (the helix–coil transition enthalpy/entropy ratio).     

Cooperative lengths of DNA during melting

The mean cooperative length of domains of DNA, determined from the variance in (G + C) content in derivative melting curves of large bacterial DNAs, varies from 230 base pairs (bp) for (A ? T)-rich domains to 580 bp for (G ? C) domains. These values correspond to values for the cooperativity parameter of 2(±2) × 10^?5 and 3(±2) × 10^?6, respectively, and to +7.2 and +9.6 kcal for the free energy of helix interruption in those regions.     

Anisotropic elastic bending models of DNA

Simplified elastic rod models of DNA were developed in which the rigidity of DNA is sequence dependent and asymmetrical, i.e. the bending is facilitated towards the major groove. By subjecting the models to bending load in various directions perpendicular to the longitudinal axis of DNA, the bending deformation and the average conformation of the models can be estimated using finite element methods. Intrinsically curved sequence motifs [(aaaattttgc)_n, (tctctaaaaaatatataaaaa)_n] are found to be curved by this modelling procedure whereas the average conformation of homopolymers and straight motifs [(a)_n, (atctaatctaacacaacaca)_n] show negligible or no curvature. This suggests that sequence dependent asymmetric rigidity of DNA can provide an explanation in itself for intrinsic DNA curvature. The average rigidity of various DNA sequences was calculated and a good correlation was found with such quantities as the free energy change upon the binding of the Cro repressor, the base stacking energy and the thermal fluctuations at room temperature.     

Predicted role of particular phonons in the strand separation melting of the DNA double helix

The frequency-dependent vibrational fluctuation of the hydrogen bonds around a nucleation defect for the strand separation melting of a DNA polymer poly(dG)-poly(dC) is studied using a modified self-consistent phonon theory. There are two critical frequency bands around the defect at 340 K which is near the temperature at which the hydrogen bonds in neighboring cells melt. The first band is between 60 cm-1 and 120 cm-1 which is essential for the melting proceeding in +z direction(3'----5' in the G backbone). The second is the band under 20 cm-1 which is important for the melting proceeding in -z direction.