A unified model for the origin of DNA sequence-directed curvature

The fine structure of the DNA double helix and a number of its physical properties depend upon nucleotide sequence. This includes minor groove width, the propensity to undergo the B-form to A-form transition, sequence-directed curvature, and cation localization. Despite the multitude of studies conducted on DNA, it is still difficult to appreciate how these fundamental properties are linked to each other at the level of nucleotide sequence. We demonstrate that several sequence-dependent properties of DNA can be attributed, at least in part, to the sequence-specific localization of cations in the major and minor grooves. We also show that effects of cation localization on DNA structure are easier to understand if we divide all DNA sequences into three principal groups: A-tracts, G-tracts, and generic DNA. The A-tract group of sequences has a peculiar helical structure (i.e., B*-form) with an unusually narrow minor groove and high base-pair propeller twist. Both experimental and theoretical studies have provided evidence that the B*-form helical structure of A-tracts requires cations to be localized in the minor groove. G-tracts, on the other hand, have a propensity to undergo the B-form to A-form transition with increasing ionic strength. This property of G-tracts is directly connected to the observation that cations are preferentially localized in the major groove of G-tract sequences. Generic DNA, which represents the vast majority of DNA sequences, has a more balanced occupation of the major and minor grooves by cations than A-tracts or G-tracts and is thereby stabilized in the canonical B-form helix. Thus, DNA secondary structure can be viewed as a tug of war between the major and minor grooves for cations, with A-tracts and G-tracts each having one groove that dominates the other for cation localization. Finally, the sequence-directed curvature caused by A-tracts and G-tracts can, in both cases, be explained by the cation-dependent mismatch of A-tract and G-tract helical structures with the canonical B-form helix of generic DNA (i.e., a cation-dependent junction model).     

DNA A-tracts bending: polarization effects on electrostatic interactions across their minor groove

Bending by the DNA A-tracts constitutes a contentious issue, suggesting deficiencies in the physics employed so far. Here, we inquire as to the importance in this bending of many-body polarization effects on the electrostatic interactions across their narrow minor groove. We have done this on the basis of the findings of Jarque and Buckingham who developed a procedure based on a Monte Carlo simulation for two charges of the same sign embedded in a polarizable medium. Remarkably, the present analysis reveals that for compact DNA conformations, which result from dynamic effects, an overall attractive interaction operates between the phosphate charges; this interaction is especially strong for the narrow minor groove of the A-tracts, suggesting a tendency for DNA to bend toward this groove. This tendency is in agreement with the conclusions of electrophoretic and NMR solution studies. The present analysis is also consistent with the experimental observations that the minor groove is much more easily compressible than the major groove and the bending propensity of the A-tracts is greatly reduced at “premelting” temperatures. By contrast, the dielectric screening model predicts a repulsion between the phosphate charges and is not consistent with the aforementioned bending tendency or experimental observations.     

Sequence-dependent variation in DNA minor groove width dictates orientational preference of Hoechst 33258 in A-tract recognition: solution NMR structure of the 2:1 complex with d(CTTTTGCAAAAG)(2)

The solution structure of the dodecamer duplex d(CTTTTGCAAAAG)₂ and its 2:1 complex with the bis-benzimidazole Hoechst 33258 has been investigated by NMR and NOE-restrained molecular dynamics (rMD) simulations. Drug molecules are bound in each of the two A-tracts with the bulky N-methylpiperazine ring of each drug located close to the central TG (CA) step, binding essentially to the narrow minor groove of each A-tract. MD simulations over 1 ns, using an explicit solvation model, reveal time-averaged sequence-dependent narrowing of the minor groove from the 3′-end towards the 5′-end of each TTTT sequence. Distinct junctions at the TpG (CpA) steps, characterised by large positive roll, low helical and propeller twists and rapid AT base pair opening rates, add to the widening of the groove at these sites and appear to account for the bound orientation of the two drug molecules with the N-methylpiperazine ring binding in the wider part of the groove close to the junctions. Comparisons between the free DNA structure and the 2:1 complex (heavy atom RMSD 1.55 Å) reveal that these sequence-dependent features persist in both structures. NMR studies of the sequence d(GAAAAGCTTTTC)₂, in which the A-tracts have been inverted with the elimination of the TpG junctions, results in loss of orientational specificity of Hoechst 33258 and formation of multiple bound species in solution, consistent with the drug binding in a number of different orientations.     

Uranyl photoprobing of nonbent A/T- and bent A-tracts. A difference of flexibility?

In this study, we have systematically compared the uranyl photocleavage of a range of bent A-tracts and nonbent TA-tracts as well as interrupted A-tracts. We demonstrate that uranyl photocleavage of A-tracts and TA-tracts is almost identical, indicating a very similar minor groove conformation. Furthermore, a 10 base pair A-tract is divided into two independent tracts by an intervening TA or GC step. Uranyl probing also clearly distinguishes the bent A4T4 and the nonbent T4A4 sequences as adopting different structures, and our interpretation of the data is consistent with a structure for the bent A4T4 sequence that resembles a continuous A-tract, whereas the nonbent T4A4 sequences are closer to two independent and opposite A-tracts that cancel each other in terms of macroscopic bending. Finally, we also note that even single TA and TAT steps are highly sensitive to uranyl photocleavage and propose that in addition to average minor groove width, uranyl also senses DNA helix flexibility/deformability. Thus, the structural difference of TA-tracts and A-tracts may to a large extent reflect a difference in flexibility, and DNA curvature may consequently require a rigid narrow minor groove conformation that creates distinct A-tract-B-DNA junctions as the predominant cause of the bending.     

Control of DNA minor groove width and Fis protein binding by the purine 2-amino group

The width of the DNA minor groove varies with sequence and can be a major determinant of DNA shape recognition by proteins. For example, the minor groove within the center of the Fis–DNA complex narrows to about half the mean minor groove width of canonical B-form DNA to fit onto the protein surface. G/C base pairs within this segment, which is not contacted by the Fis protein, reduce binding affinities up to 2000-fold over A/T-rich sequences. We show here through multiple X-ray structures and binding properties of Fis–DNA complexes containing base analogs that the 2-amino group on guanine is the primary molecular determinant controlling minor groove widths. Molecular dynamics simulations of free-DNA targets with canonical and modified bases further demonstrate that sequence-dependent narrowing of minor groove widths is modulated almost entirely by the presence of purine 2-amino groups. We also provide evidence that protein-mediated phosphate neutralization facilitates minor groove compression and is particularly important for binding to non-optimally shaped DNA duplexes.     

Molecular dynamics simulations of A. T-rich oligomers: sequence-specific binding of Na+ in the minor groove of B-DNA

Molecular dynamics simulations have been employed to probe the sequence-specific binding of sodium ions to the minor groove of B-DNA of three A. T-rich oligomers having identical compositions but different orders of the base pairs: C(AT)(4)G, CA(4)T(4)G, and CT(4)A(4)G. Recent experimental investigations, either in crystals or in solution, have shown that monovalent cations bind to DNA in a sequence-specific mode, preferentially in the narrow minor groove regions of uninterrupted sequences of four or more adenines (A-tracts), replacing a water molecule of the ordered hydration structure, the hydration spine. Following this evidence, it has been hypothesized that in A-tracts these events may be responsible for structural peculiarities such as a narrow minor groove and a curvature of the helix axis. The present simulations confirm a sequence specificity of the binding of sodium ions: Na(+) intrusions in the first layer of hydration of the minor groove, with long residence times, up to approximately 3 ns, are observed only in the minor groove of A-tracts but not in the alternating sequence. The effects of these intrusions on the structure of DNA depend on the ion coordination: when the ion replaces a water molecule of the spine, the minor groove becomes narrower. Ion intrusions may also disrupt the hydration spine modifying the oligomer structure to a large extent. However, in no case intrusions were observed to locally bend the axis toward the minor groove. The simulations also show that ions may reside for long time periods in the second layer of hydration, particularly in the wider regions of the groove, often leading to an opening of the groove.     

DNA conformational analysis in solution by uranyl mediated photocleavage.

Uranyl mediated photocleavage of double stranded DNA is proposed as a general probing for DNA helix conformation in terms of minor groove width/electronegative potential. Specifically, it is found that A/T-tracts known to constitute strong distamycin binding sites are preferentially photocleaved by uranyl in a way indicating strongest uranyl binding at the center of the minor groove of the AT-region. The A-tracts of kinetoplast DNA show the highest reactivity at the 3'-end of the tract--as opposed to cleavage by EDTA/Fell--in accordance with the minor groove being more narrow at this end. Finally, uranyl photocleavage of the internal control region (ICR) of the 5S-RNA gene yields a cleavage modulation pattern fully compatible with that obtained by DNase I which also--in a more complex way--senses DNA minor groove width.     

Characterization of divalent cation localization in the minor groove of the A(n)T(n) and T(n)A(n) DNA sequence elements by (1)H NMR spectroscopy and manganese(II)

The localization of Mn(2+) in A-tract DNA has been studied by (1)H NMR spectroscopy using a series of self-complementary dodecamer oligonucleotides that contain the sequence motifs A(n)(n) and T(n)A(n), where n = 2, 3, or 4. Mn(2+) localization in the minor groove is observed for all the sequences that have been studied, with the position and degree of localization being highly sequence-dependent. The site most favored for Mn(2+) localization in the minor groove is near the 5'-most ApA step for both the T(n)A(n) and the A(n)T(n) series. For the T(n)A(n) series, this results in two closely spaced symmetry-related Mn(2+) localization sites near the center of each duplex, while for the A(n)T(n) series, the two symmetry-related sites are separated by as much as one half-helical turn. The degree of Mn(2+) localization in the minor groove of the T(n)A(n) series decreases substantially as the AT sequence element is shortened from T(4)A(4) to T(2)A(2). The A(n)T(n) series also exhibits length-dependent Mn(2+) localization; however, the degree of minor groove occupancy by Mn(2+) is significantly less than that observed for the T(n)A(n) series. For both A(n)T(n) and T(n)A(n) sequences, the 3'-most AH2 resonance is the least broadened of the AH2 resonances. This is consistent with the observation that the minor groove of A-tract DNA narrows in the 5' to 3' direction, apparently becoming too narrow after two base pairs for the entry of a fully hydrated divalent cation. The results that are reported illustrate the delicate interplay that exists between DNA nucleotide sequence, minor groove width, and divalent cation localization. The proposed role of cation localization in helical axis bending by A-tracts is also discussed.     

Increased temperature and 2-methyl-2,4-pentanediol change the DNA structure of both curved and uncurved adenine/thymine-rich sequences

DNA curvature is affected by elevated temperature and dehydrating agents such as 2-methyl-2,4-pentanediol (MPD) (used in crystallization). This effect of MPD has been ascribed to a specific distortion of the structure of adenine tracts (A-tracts), probably through a deformation of the characteristic narrow minor groove. Uranyl photoprobing indicates that a narrowed minor groove is present in all A/T regions containing four or more A/T base pairs. Consequently, this technique may be employed to study conformational changes in other A/T-rich sequences than pure A-tracts. In this study we use uranyl photoprobing to demonstrate that the effect of elevated temperature and MPD is analogous on both "normal" and curve-inducing A/T-rich sequences. The results therefore indicate that under these conditions the minor groove is widened in all A/T sequences and not only in pure A-tracts as previously suggested. Thus, the rather subtle structural difference of AT regions and A-tracts in nonbent DNA versus A-tracts in bent DNA may be quantitative rather than qualitative; i.e., the structure is more persistent and/or rigid in bent DNA.     

Structure of hybrid backbone methylphosphonate DNA heteroduplexes: effect of R and S stereochemistry.

Methyl phosphonate oligonucleotides have been used as antisense and antigene agents. Substitution of a methyl group for oxygen in the phosphate ester backbone introduces a new chiral center. Significant differences in physical properties and hybridization abilities are observed between the R(p) and S(p) diastereomers. Chirally pure methylphosphonate deoxyribooligonucleotides were synthesized, and the solution structures of duplexes formed between a single strand heptanucleotide methylphosphonate, d(Cp(Me)Cp(Me)Ap(Me)Ap(Me)Ap(Me)Cp(Me)A), hybridized to a complementary octanucleotide, d(TpGpTpTpTpGpGpC), were studied by NMR spectroscopy. Stereochemistry at the methylphosphonate center for the heptanucleotide was either RpRpRpRpRpRp (R(p) stereoisomer) or RpRpRpSpRpRp (S(p) stereoisomer, although only one of the six methylphosphonate centers has the S(p) stereochemistry). The results show that the methylphosphonate strands in the heteroduplexes exhibit increased dynamics relative to the DNA strand. Substitution of one chiral center from R(p) to S(p) has a profound effect on the hybridization ability of the methylphosphonate strand. Sugars in the phosphodiester strand exhibit C(2)(') endo sugar puckering while the sugars in the methyl phosphonate strand exhibit an intermediate C(4)(') endo puckering. Bases are well stacked on each other throughout the duplex. The hybridization of the methylphosphonate strand does not perturb the structure of the complementary DNA strand in the hetero duplexes. The sugar residue 5' to the S(p) chiral center shows A-form sugar puckering, with a C(3)(')-endo conformation. Minor groove width in the R(p) stereoisomer is considerably wider, particularly at the R(p) vs S(p) site and is attributed to either steric interactions across the minor groove or poorer metal ion coordination within the minor groove.     

Sequence Recognition of DNA by Protein-Induced Conformational Transitions

The binding of proteins to specific sequences of DNA is an important feature of virtually all DNA transactions. Proteins recognize specific DNA sequences using both direct readout (sensing types and positions of DNA functional groups) and indirect readout (sensing DNA conformation and deformability). Previously we showed that the P22 c2 repressor N-terminal domain (P22R NTD) forces the central non-contacted 5'-ATAT-3' sequence of the DNA operator into the B′ state, a state known to affect DNA hydration, rigidity and bending. Usually the B′ state, with a narrow minor groove and a spine of hydration, is reserved for A-tract DNA (TpA steps disrupt A-tracts). Here, we have co-crystallized P22R NTD with an operator containing a central 5′-ACGT-3′ sequence in the non-contacted region. C·G base pairs have not previously been observed in the B′ state and are thought to prevent it. However, P22R NTD induces a narrow minor groove and a spine of hydration to 5'-ACGT-3'. We observe that C·G base pairs have distinctive destabilizing and disordering effects on the spine of hydration. It appears that the reduced stability of the spine results in a higher energy cost for the B to B′ transition. The differential effect of DNA sequence on the barrier to this transition allows the protein to sense the non-contacted DNA sequence.     

DNA sequence-dependent positioning of the linker histone in a nucleosome: A single-pair FRET study

Linker histones (LHs) bind to nucleosomes with their globular domain (gH) positioned in either an on- or an off-dyad binding mode. Here, we study the effect of the linker DNA (L-DNA) sequence on the binding of a full-length LH, Xenopus laevis H1.0b, to a Widom 601 nucleosome core particle (NCP) flanked by two 40 bp long L-DNA arms, by single-pair FRET spectroscopy. We varied the sequence of the 11 bp of L-DNA adjoining the NCP on either side, making the sequence either A-tract, purely GC, or mixed with 64% AT. The labeled gH consistently exhibited higher FRET efficiency with the labeled L-DNA containing the A-tract than that with the pure-GC stretch, even when the stretches were swapped. However, it did not exhibit higher FRET efficiency with the L-DNA containing 64% AT-rich mixed DNA when compared to the pure-GC stretch. We explain our observations with a model that shows that the gH binds on dyad and that two arginines mediate recognition of the A-tract via its characteristically narrow minor groove. To investigate whether this on-dyad minor groove-based recognition was distinct from previously identified off-dyad major groove-based recognition, a nucleosome was designed with A-tracts on both the L-DNA arms. One A-tract was complementary to thymine and the other to deoxyuridine. The major groove of the thymine-tract was lined with methyl groups that were absent from the major groove of the deoxyuridine tract. The gH exhibited similar FRET for both these A-tracts, suggesting that it does not interact with the thymine methyl groups exposed on the major groove. Our observations thus complement previous studies that suggest that different LH isoforms may employ different ways of recognizing AT-rich DNA and A-tracts. This adaptability may enable the LH to universally compact scaffold-associated regions and constitutive heterochromatin, which are rich in such sequences.     

Sequence-dependent effects in drug-DNA interaction: the crystal structure of Hoechst 33258 bound to the d(CGCAAATTTGCG)2 duplex.

The bis-benzimidazole drug Hoechst 33258 has been co-crystallized with the dodecanucleotide sequence d(CGCAAATTTGCG)2. The structure has been solved by molecular replacement and refined to an R factor of 18.5% for 2125 reflections collected on a Xentronics area detector. The drug is bound in the minor groove, at the five base-pair site 5'-ATTTG and is in a unique orientation. This is displaced by one base pair in the 5' direction compared to previously-determined structures of this drug with the sequence d(CGCGAATTCGCG)2. Reasons for this difference in behaviour are discussed in terms of several sequence-dependent structural features of the DNA, with particular reference to differences in propeller twist and minor-groove width.     

A-Tract bending: insights into experimental structures by computational models

While solution structures of adenine tract (A-tract) oligomers have indicated a unique bend direction equivalent to negative global roll (commonly termed "minor-groove bending"), crystallographic data have not unambiguously characterized the bend direction; nevertheless, many features are shared by all A-tract crystal and solution structures (e.g. propeller twisting, narrow minor grooves, and localized water spines). To examine the origin of bending and to relate findings to the crystallographic and solution data, we analyze molecular dynamics trajectories of two solvated A-tract dodecamers: 1D89, d(CGCGA(6)CG), and 1D98, d(CGCA(6)GCG), using a new general global bending framework for analyzing bent DNA and DNA/protein complexes. It is significant that the crystallographically-based initial structures are converted from dissimilar to similar bend directions equivalent to negative global roll, with the average helical-axis bend ranging from 10.5 degrees to 14.1 degrees. The largest bend occurs as positive roll of 12 degrees on the 5' side of the A-tracts (supporting a junction model) and is reinforced by gradual curvature at each A-tract base-pair (bp) step (supporting a wedge model). The precise magnitude of the bend is subtly sequence dependent (consistent with a curved general sequence model). The conversion to negative global roll only requires small local changes at each bp, accumulated over flexible moieties both outside and inside the A-tract. In contrast, the control sequence 1BNA, d(CGCGA(2)TTCGCG), bends marginally (only 6.9 degrees ) with no preferred direction. The molecular features that stabilize the bend direction in the A-tract dodecamers include propeller twisting of AT base-pairs, puckering differences between A and T deoxyriboses, a narrow minor groove, and a stable water spine (that extends slightly beyond the A-tract, with lifetimes approaching 0.2 ns). The sugar conformations, in particular, are proposed as important factors that support bent DNA. It is significant that all these curvature-stabilizing features are also observed in the crystallographic structures, but yield overall different bending paths, largely due to the effects of sequences outside the A-tract. These results merge structural details reported for A-tract structures by experiment and theory and lead to structural and dynamic insights into sequence-dependent DNA flexibility, as highlighted by the effect of an A-tract variant of a TATA-box element on bending and flexibility required for TBP binding.     

DNA structural variations in the E. coli tyrT promoter

X-ray studies have established that the structure of a right-handed, Watson-Crick double helix can change from place to place along its length as a function of base sequence. The base pairs transmit deformations out to the phosphate backbone, where they can then be recognized by proteins and other DNA-binding reagents. Here we have examined at single-bond resolution the interactions of three commonly used nucleases (DNAase I, DNAase II, and copper-phenanthroline) with a DNA of natural origin, the 160 bp tyrT promoter. All three of these reagents seem sensitive to DNA backbone geometry rather than base sequence per se. Their sequence-dependent patterns of cleavage provide evidence for structural polymorphism of several sorts: global variation in helix groove width, global variation in radial asymmetry, and local variation in phosphate accessibility. These findings explain how sequence zones of a certain base composition, or purine-pyrimidine asymmetry, can influence the recognition of DNA by protein molecules.     

DNA structure: what's in charge?

DNA structure is well known to be sensitive to hydration and ionic strength. Recent theoretical predictions and experimental observations have raised the idea of the intrusion of monovalent cations into the minor groove spine of hydration in B-form DNA. To investigate this further, extensions and further analysis of molecular dynamics (MD) simulations on d(CGCCGAATTCGCG), d(ATAGGCAAAAAATAGGCAAAAATGG) and d(G(5)-(GA(4)T(4)C)(2)-C(5)), including counterions and water, have been performed. To examine the effective of minor groove ions on structure, we analyzed the MD snapshots from a 15 ns trajectory on d(CGCGAATTCGCG) as two subsets: those exhibiting a minor groove water spine and those with groove-bound ions. The results indicate that Na(+) at the ApT step of the minor groove of d(CGCCGAATTCGCG) makes only small local changes in the DNA structure, and these changes are well within the thermal fluctuations calculated from the MD. To examine the effect of ions on the differential stability of a B-form helix, further analysis was performed on two longer oligonucleotides, which exhibit A-tract-induced axis bending localized around the CpG step in the major groove. Plots of axis bending and proximity of ions to the bending locus were generated as a function of time and revealed a strong linear correlation, supporting the idea that mobile cations play a key role in local helix deformations of DNA and indicating ion proximity just precedes the bending event. To address the issue of "what's in charge?" of DNA structure more generally, the relative free energy of A and B-form d(CGCGAATTCGCG) structures from MD simulations under various environmental circumstances were estimated using the free energy component method. The results indicate that the dominant effects on conformational stability come from the electrostatic free energy, but not exclusively from groove bound ions per se, but from a balance of competing factors in the electrostatic free energy, including phosphate repulsions internal to the DNA, the electrostatic component of hydration (i.e. solvent polarization), and electrostatic effects of the counterion atmosphere. In summary, free energy calculations indicate that the electrostatic component is dominant, MD shows temporal proximity of mobile counterions to be correlated with A-track-induced bending, and thus the mobile ion component of electrostatics is a significant contributor. However, the MD structure of the dodecamer d(CGCGAATTCGCG) is not highly sensitive to whether there is a sodium ion in the minor groove.