Structural Studies of DNA Fragments: The G·T Wobble Base Pair in A,B and Z DNA; The G·A Base Pair in B-DNA

Abstract

The crystal structures of five double helical DNA fragments containing non-Watson-Crick complementary base pairs are reviewed. They comprise four fragments containing G·T base pairs: two deoxyoctamers d(GGGGCTCC) and d(GGGGTCCC) which crystallise as A type helices; a deoxydodecamer d(CGCGAATTTGCG) which crystallises in the B-DNA conformation; and the deoxyhexamer d(TGCGCG), which crystallises as a Z-DNA helix. In all four duplexes the G and T bases form wobble base pairs, with bases in the major tautomer forms and hydrogen bonds linking N1 of G with 02 of T and 06 of G with N3 of T. The X-ray analyses establish that the G·T wobble base pair can be accommodated in the A, B or Z double helix with minimal distortion of the global conformation. There are, however, changes in base stacking in the neighbourhood of the mismatched bases. The fifth structure, d(CGCGAATTAGCG), contains the purine purine mismatch G·A where G is in the anti and A in the syn conformation. The results represent the first direct structure determinations of base pair mismatches in DNA fragments and are discussed in relation to the fidelity of replication and mismatch recognition.     

Database proton NMR chemical shifts for RNA signal assignment and validation

The Biological Magnetic Resonance Data Bank contains NMR chemical shift depositions for 132 RNAs and RNA-containing complexes. We have analyzed the ¹H NMR chemical shifts reported for non-exchangeable protons of residues that reside within A-form helical regions of these RNAs. The analysis focused on the central base pair within a stretch of three adjacent base pairs (BP triplets), and included both Watson–Crick (WC; G:C, A:U) and G:U wobble pairs. Chemical shift values were included for all 4³ possible WC-BP triplets, as well as 137 additional triplets that contain one or more G:U wobbles. Sequence-dependent chemical shift correlations were identified, including correlations involving terminating base pairs within the triplets and canonical and non-canonical structures adjacent to the BP triplets (i.e. bulges, loops, WC and non-WC BPs), despite the fact that the NMR data were obtained under different conditions of pH, buffer, ionic strength, and temperature. A computer program (RNAShifts) was developed that enables convenient comparison of RNA ¹H NMR assignments with database predictions, which should facilitate future signal assignment/validation efforts and enable rapid identification of non-canonical RNA structures and RNA-ligand/protein interaction sites.     

Stacking geometry for non‐canonical G:U wobble base pair containing dinucleotide sequences in RNA: dispersion‐corrected DFT‐D study

Emergence of thousands of crystal structures of noncoding RNA molecules indicates its structural and functional diversity. RNA function is based upon a large variety of structural elements which are specifically assembled in the folded molecules. Along with the canonical Watson‐Crick base pairs, different orientations of the bases to form hydrogen‐bonded non‐canonical base pairs have also been observed in the available RNA structures. Frequencies of occurrences of different non‐canonical base pairs in RNA indicate their important role to maintain overall structure and functions of RNA. There are several reports on geometry and energetic stabilities of these non‐canonical base pairs. However, their stacking geometry and stacking stability with the neighboring base pairs are not well studied. Among the different non‐canonical base pairs, the G:U wobble base pair (G:U W:WC) is most frequently observed in the RNA double helices. Using quantum chemical method and available experimental data set we have studied the stacking geometry of G:U W:WC base pair containing dinucleotide sequences in roll‐slide parameters hyperspace for different values of twist. This study indicates that the G:U W:WC base pair can stack well with the canonical base pairs giving rise to large interaction energy. The overall preferred stacking geometry in terms of roll, twist and slide for the eleven possible dinucleotide sequences is seen to be quite dependent on their sequences. © 2015 Wiley Periodicals, Inc. Biopolymers 103: 328–338, 2015.     

An innate twist between Crick’s wobble and Watson-Crick base pairs

Non-Watson-Crick pairs like the G·U wobble are frequent in RNA duplexes. Their geometric dissimilarity (nonisostericity) with the Watson-Crick base pairs and among themselves imparts structural variations decisive for biological functions. Through a novel circular representation of base pairs, a simple and general metric scheme for quantification of base-pair nonisostericity, in terms of residual twist and radial difference that can also envisage its mechanistic effect, is proposed. The scheme is exemplified by G·U and U·G wobble pairs, and their predicable local effects on helical twist angle are validated by MD simulations. New insights into a possible rationale for contextual occurrence of G·U and other non-WC pairs, as well as the influence of a G·U pair on other non-Watson-Crick pair neighborhood and RNA-protein interactions are obtained from analysis of crystal structure data. A few instances of RNA-protein interactions along the major groove are documented in addition to the well-recognized interaction of the G·U pair along the minor groove. The nonisostericity-mediated influence of wobble pairs for facilitating helical packing through long-range interactions in ribosomal RNAs is also reviewed.     

Diversity of base-pair conformations and their occurrence in rRNA structure and RNA structural motifs

In addition to the canonical base-pairs comprising the standard Watson-Crick (C:G and U:A) and wobble U:G conformations, an analysis of the base-pair types and conformations in the rRNAs in the high-resolution crystal structures of the Thermus thermophilus 30S and Haloarcula marismortui 50S ribosomal subunits has identified a wide variety of non-canonical base-pair types and conformations. However, the existing nomenclatures do not describe all of the observed non-canonical conformations or describe them with some ambiguity. Thus, a standardized system is required to classify all of these non-canonical conformations appropriately. Here, we propose a new, simple and systematic nomenclature that unambiguously classifies base-pair conformations occurring in base-pairs, base-triples and base-quadruples that are associated with secondary and tertiary interactions. This system is based on the topological arrangement of the two bases and glycosidic bonds in a given base-pair. Base-pairs in the internal positions of regular secondary structure helices usually form with canonical base-pair groups (C:G, U:A, and U:G) and canonical conformations (C:G WC, U:A WC, and U:G Wb). In contrast, non-helical base-pairs outside of regular structure helices usually have non-canonical base-pair groups and conformations. In addition, many non-helical base-pairs are involved in RNA motifs that form a defined set of non-canonical conformations. Thus, each rare non-canonical conformation may be functionally and structurally important. Finally, the topology-based isostericity of base-pair conformations can rationalize base-pair exchanges in the evolution of RNA molecules.     

Interactions of molecules with nucleic acids. I. An algorithm to generate nucleic acid structures with an application to the B-DNA structure and a counterclockwise helix.

An algorithm is developed that enables the routine determination of backbone conformations of nucleic acids. All atomic positions including hydrogen are specified in accord with experimental bond lengths and angles but with theoretically determined conformational angles. For two Watson-Crick base pairs at a separation of 3.38 Å, and perpendicular to a common helical axis, minimum energy configurations are found for all 10 combinations at helical angles of α ～ 36°–38°, corresponding to the B-DNA structure with C(2′)-endo sugar puckers. Backbone configurations exist only within the range 35.5° ? α ? 42°, which suggests the origin of the 10-fold helix. Calculated stacking energies for the B-DNA structure increases for each of the clustered groups of base pairs: G·C with G·C, G·C with A·T, and A·T with A·T, and they are in approximate agreement with experimental observations. The counter-clockwise helix is examined, and physically meaningful structures are found only when the helical axes of successive base pairs are disjointed.     

Influence of base-pair changes and cooperativity parameters on the melting curves of short DNAs

Theoretical melting curves were calculated for four DNA restriction fragments, 157–257 base pairs (bp), and a series of hypothetical block DNAs with sequences d(C_2xA_xC_2x). d(C_2xT_xG_2x), 5 ? x ? 40. These DNAs provided a mixture of A·T/G·C sequence distributions with which to investigate the effects of parameters and base-pair changes on the melting of short DNAs. The sensitivity of DNA melting curves to changes in internal loop melting parameters σ and κ was examined. As Expected, theoretical melting curves of short DNAs with a quasirandom base-pair sequence vary little with changes in internal loop parameters. End melting dominates the transition behaviour of these moleucles. This was also observed for the block DNAs up to x = 22. Beyond this length, melting curves are highly sensitive to the internal loop parameters. Sensitivity is also predicted for a 157-bp fragment with a block distribution of A·T and G·C pairs. These results indicate that accurate evaluation of internal loop parameters is possible with short DNAs (100–200 bp) containing a G·C/A·T/G·C block distribution with at least 22 bp in each block. Duplex-to-single-strands dissociation parameters were reevaluated form experimental melting curve data of eight DNA fragments using a least squares fit approach. This analysis confirmed parameter values previously found with a simplified dissociation model. A Priori predictions are made on the effects of base-pair changes on the melting curves of three characterized DNA restriction fragments. Single base-pair changes are predicted to induce small but measurable changes in the melting curves. The characteristics of the altered melting curves depend on the location of the base-pair change.     

Hybrid simulation approach incorporating microscopic interaction along with rigid body degrees of freedom for stacking between base pairs

Stacking interaction between the aromatic heterocyclic bases plays an important role in the double helical structures of nucleic acids. Considering the base as rigid body, there are total of 18 degrees of freedom of a dinucleotide step. Some of these parameters show sequence preferences, indicating that the detailed atomic interactions are important in the stacking. Large variants of non‐canonical base pairs have been seen in the crystallographic structures of RNA. However, their stacking preferences are not thoroughly deciphered yet from experimental results. The current theoretical approaches use either the rigid body degrees of freedom where the atomic information are lost or computationally expensive all atom simulations. We have used a hybrid simulation approach incorporating Monte‐Carlo Metropolis sampling in the hyperspace of 18 stacking parameters where the interaction energies using AMBER‐parm99bsc0 and CHARMM‐36 force‐fields were calculated from atomic positions. We have also performed stacking energy calculations for structures from Monte‐Carlo ensemble by Dispersion corrected density functional theory. The available experimental data with Watson–Crick base pairs are compared to establish the validity of the method. Stacking interaction involving A:U and G:C base pairs with non‐canonical G:U base pairs also were calculated and showed that these structures were also sequence dependent. This approach could be useful to generate multiscale modeling of nucleic acids in terms of coarse‐grained parameters where the atomic interactions are preserved. This method would also be useful to predict structure and dynamics of different base pair steps containing non Watson–Crick base pairs, as found often in the non‐coding RNA structures. © 2015 Wiley Periodicals, Inc. Biopolymers 105: 212–226, 2016.     

Metastable secondary structures in ribosomal RNA: a new method for analyzing the titration behavior of rRNA

The pH titration behavior of E. coli rRNA in the acid range has been analyzed by combining spectrophotometric and potentiometric titration data. The “simplest” model for the system, which considers as possible reactions the protonation of adenine (A), cytosine (C), and guanine (G) residues along with the opening of A·U and G·C base pairs, does not adequately account for the titration properties. It is postulated that extra reactions may occur in addition to those in the “simplest” model, and a new analytical method was developed to deal with this situation. Our approach yields the ultraviolet spectral changes which accompany the extra reactions, from which the nature of these reactions can in principle be deduced. The calculations also give, at each pH, the extents of the extra reactions as well as the extents of those reactions which comprise the “simplest” model. We infer that in acidic RNA solutions of 0.1M ionic strength there occur at least two extra reactions, each of which involves G residues. We propose that in the pH range 6.0 ≥ pH ≥ 3.8 triple-stranded helical sequences, presumably protonated G·C·G, are formed. These regions are replaced at lower pH by acid-stable structures involving G·G and A·A base pairs. In solutions of lower ionic strength (I = 0.01M) no triple strands are formed, but G·G and A·A regions seem to develop even at pH values as high as 6.0. At I = 0.1M, an acid–base titration cycle between pH 7 and 2.8 is not reversible; rRNA shows true hysteresis behavior. We conclude that in ribosomal RNA's, which are generally G-rich, guanine residues may participate in hitherto unpredicted conformations, some of which may be metastable while others are equilibrium structures.     

Ring-current shifts in the 300 MHz nuclear magnetic resonance spectra of six purified transfer RNA molecules

We present the 300 MHz high-resolution proton nuclear magnetic resonance spectra of the ring NH hydrogen-bonded protons of six purified tRNAs. Good agreement was obtained between the observed spectra and those computed on the assumption of the suitable cloverleaf models. In the computation it is assumed that the hydrogen-bonded ring NH in each type of base pair has an intrinsic position with respect to 2,2-dimethyl-2-silapentane-5-sulfonate, i.e. in A·U it is at ?14·8 parts per million, in G·C at ?13·7 parts per million and in A·Ψat ?13·5 parts per million. The shifts of these resonances from these positions are calculated by including ring current fields from the nearest neighbors. The agreement is very good, adding support to our earlier findings that there is no evidence for additional Watson-Crick base pairs detected beyond those in the cloverleaf. In general, resolved resonances are fitted by the computed spectra to within ±0·2 part per million showing that there is no need for any additional physical mechanism to explain the nuclear magnetic resonance positions. Hence, the nuclear magnetic resonance spectra can be interpreted in terms of the structure of their neighbors and in a few important cases this has been particularly valuable in understanding the structure beyond the end of a helical region. In the tRNA^Glu_E.coli′ for example, the positions of the resonances in A·U no. 7 and A·U no. 49 at the interior ends of the acceptor and -T-Ψ-C- stems, respectively, strongly suggest that these two stems are in a continuous helix. Other structural effects at the ends of the helical regions are also suggested by the nuclear magnetic resonance spectra.     

RNA structure and dynamics: A base pairing perspective

RNA is now known to possess various structural, regulatory and enzymatic functions for survival of cellular organisms. Functional RNA structures are generally created by three-dimensional organization of small structural motifs, formed by base pairing between self-complementary sequences from different parts of the RNA chain. In addition to the canonical Watson–Crick or wobble base pairs, several non-canonical base pairs are found to be crucial to the structural organization of RNA molecules. They appear within different structural motifs and are found to stabilize the molecule through long-range intra-molecular interactions between basic structural motifs like double helices and loops. These base pairs also impart functional variation to the minor groove of A-form RNA helices, thus forming anchoring site for metabolites and ligands. Non-canonical base pairs are formed by edge-to-edge hydrogen bonding interactions between the bases. A large number of theoretical studies have been done to detect and analyze these non-canonical base pairs within crystal or NMR derived structures of different functional RNA. Theoretical studies of these isolated base pairs using ab initio quantum chemical methods as well as molecular dynamics simulations of larger fragments have also established that many of these non-canonical base pairs are as stable as the canonical Watson–Crick base pairs. This review focuses on the various structural aspects of non-canonical base pairs in the organization of RNA molecules and the possible applications of these base pairs in predicting RNA structures with more accuracy.     

Why the tautomerization of the G·C Watson–Crick base pair via the DPT does not cause point mutations during DNA replication? QM and QTAIM comprehensive analysis

The ground-state tautomerization of the G·C Watson–Crick base pair by the double proton transfer (DPT) was comprehensively studied in vacuo and in the continuum with a low dielectric constant (??=?4), corresponding to a hydrophobic interface of protein–nucleic acid interactions, using DFT and MP2 levels of quantum-mechanical (QM) theory and quantum theory “Atoms in molecules” (QTAIM). Based on the sweeps of the electron-topological, geometric, polar, and energetic parameters, which describe the course of the G·C???G^?·C^? tautomerization (mutagenic tautomers of the G and C bases are marked with an asterisk) through the DPT along the intrinsic reaction coordinate (IRC), it was proved that it is, strictly speaking, a concerted asynchronous process both at the DFT and MP2 levels of theory, in which protons move with a small time gap in vacuum, while this time delay noticeably increases in the continuum with ??=?4. It was demonstrated using the conductor-like polarizable continuum model (CPCM) that the continuum with ??=?4 does not qualitatively affect the course of the tautomerization reaction. The DPT in the G·C Watson–Crick base pair occurs without any intermediates both in vacuum and in the continuum with ??=?4 at the DFT/MP2 levels of theory. The nine key points along the IRC of the G·C base pair tautomerization, which could be considered as electron-topological “fingerprints” of a concerted asynchronous process of the tautomerization via the DPT, have been identified and fully characterized. These key points have been used to define the reactant, transition state, and product regions of the DPT reaction in the G·C base pair. Analysis of the energetic characteristics of the H-bonds allows us to arrive at a definite conclusion that the middle N1H?N3/N3H?N1 and the lower N2H?O2/N2H?O2 parallel H-bonds in the G·C/G^?·C^? base pairs, respectively, are anticooperative, that is, the strengthening of the middle H-bond is accompanied by the weakening of the lower H-bond. At that point, the upper N4H?O6 and O6H?N4 H-bonds in the G·C and G^?·C^? base pairs, respectively, remain constant at the changes of the middle and the lower H-bonds at the beginning and at the ending of the G·C???G^?·C^? tautomerization. Aiming to answer the question posed in the title of the article, we established that the G^?·C^? Löwdin’s base pair satisfies all the requirements necessary to cause point mutations in DNA except its lifetime, which is much less than the period of time required for the replication machinery to forcibly dissociate a base pair into the monomers (several ns) during DNA replication. So, from the physicochemical point of view, the G^?·C^? Löwdin’s base pair cannot be considered as a source of point mutations arising during DNA replication.     

Configuration of wobble base pairs having pyrimidines as anticodon wobble bases: significance for codon degeneracy

Degeneracy of the genetic code was attributed by Crick to imprecise hydrogen-bonded base-pairing at the wobble position during codon–anticodon pairing. The Crick wobble rules define but do not explain the RNA base pair combinations allowed at this position. We select six pyrimidine bases functioning as anticodon wobble bases (AWBs) to study their H-bonded pairing properties with the four major RNA bases using density functional theory at the B3LYP/6-31G(d,p) level. This is done to assess the extent to which the configuration of a solitary RNA wobble base pair may in itself determine specificity and degeneracy of the genetic code by allowing or disallowing the given base pair during codon–anticodon pairing. Calculated values of select configuration markers for the base pairs screen well between allowed and disallowed base pairs for most cases examined here, where the base pair width emerges as an important factor. A few allowed wobble pairs invoke the involvement of RNA nucleoside conformation, as well as involvement of the exocyclic substituent in H-bonding. This study, however, cannot explain the disallowed status of the Ura?Gua wobble pair on the basis of configuration alone. Explanation of the allowed status of the V?Ura pair requires further study on the mediatory role of water molecules. Apart from these two cases, these computational results are sufficient, on the basis of base pair configuration alone, to account for the specificity and degeneracy of the genetic code for all known cases of codon–anticodon pairing which involve the pyrimidine AWBs studied here.     

Interactions of molecules with nucleic acids. VII. Intercalation and T.A specificity of daunomycin in DNA

Intercalation complexes of daunomicin(+1) with tetramer duplexes in DNA are studied with the theoretically determined intercalation sites (I, ?0.4), (II, ?0.4), and (III, ?1.4). These sites occur with base pairs separated by 6.76 Å for helical angles of 26°, 22°, and 8° about the intercalation site. Site I is preferred, and this is in agreement with experimental unwinding angles. Optimum binding positions and conformations are established, and these are in agreement with experimental results from crystal structures. A systematic procedure is devised to study base-pair and base-sequence specificity, which results in the demonstration that the most stable sequences are mainly ↑BP₁, T·A, DAUN, A·T, BP₄↓ and ↑BP₁, T·A, DAUN, G·C, BP₄↓, i.e., with the TpA and CpG (pyrimidine)p(purine) sequences about the intercalation site. These 32 possible sequences are found among the 40 most stable complexes. These theoretical calculations of intercalation complexes with daunomicin(+1) provide the first example in which a drug specifically selects the base pair T·A and prefers it in a particular sequence about the intercalation site. This specificity is in agreement with some experimental results. Problems associated with the interpretation of specificity are discussed in terms of the base, base-pair, and base-sequence resulting from the DNA site and the DNA–drug interactions. T·A specificity is rationalized by noting that the 2′deoxyribo-5′-monophosphate backbone attached to A is slightly more negative than that on the other nucleotides. Hence, a preference exists for binding to the protonated daunosamine (+1) groups. Stereographic projections of daunomycinone and daunomycin(+1) in a bond model and in a space-filling model with steric contours illustrate the results.     

Control of box C/D snoRNP assembly by N6‐methylation of adenine

N⁶‐methyladenine is the most widespread mRNA modification. A subset of human box C/D snoRNA species have target GAC sequences that lead to formation of N⁶‐methyladenine at a key trans Hoogsteen‐sugar A·G base pair, of which half are methylated in vivo. The GAC target is conserved only in those that are methylated. Methylation prevents binding of the 15.5‐kDa protein and the induced folding of the RNA. Thus, the assembly of the box C/D snoRNP could in principle be regulated by RNA methylation at its critical first stage. Crystallography reveals that N⁶‐methylation of adenine prevents the formation of trans Hoogsteen‐sugar A·G base pairs, explaining why the box C/D RNA cannot adopt its kinked conformation. More generally, our data indicate that sheared A·G base pairs (but not Watson–Crick base pairs) are more susceptible to disruption by N⁶mA methylation and are therefore possible regulatory sites. The human signal recognition particle RNA and many related Alu retrotransposon RNA species are also methylated at N6 of an adenine that forms a sheared base pair with guanine and mediates a key tertiary interaction.     

Structure of the hydration shells of oligo(dA-dT).oligo(dA-dT) and oligo(dA).oligo(dT) tracts in B-type conformation on the basis of Monte Carlo calculations

Monte Carlo simulations [(N, V, T)-ensemble] were performed for the hydration shell of poly(dA-dT).poly(dA-dT) in canonical B form and for the hydration shell of poly(dA).poly(dT) in canonical B conformation and in a conformation with narrow minor groove, highly inclined bases, but with a nearly zero-inclined base pair plane (B' conformation). We introduced helical periodic boundary conditions with a rather small unit cell and a limited number of water molecules to reduce the dimensionality of the configuration space. The coordinates of local maxima of water density and the properties of one- and two-membered water bridges between polar groups of the DNA were obtained. The AT-alternating duplex hydration mirrors the dyad symmetry of polar group distribution. At the dApdT step, a water bridge between the two carbonyl oxygens O2 of thymines is formed as in the central base-pair step of Dickerson's dodecamer. In the major groove, 5-membered water chains along the tetranucleotide pattern d(TATA).d(TATA) are observed. The hydration geometry of poly(dA).poly(dT) in canonical B conformation is distinguished by autonomous primary hydration of the base-pair edges in both grooves. When this polymer adopts a conformation with highly inclined bases and narrow minor groove, the water density distribution in the minor groove is in excellent agreement with Dickerson's spine model. One local maximum per base pair of the first layer is located near the dyad axis between adjacent base pairs, and one local maximum per base pair in the second shell lies near the dyad axis of the base pair itself. The water bridge between the two strands formed within the first layer was observed with high probability. But the water molecules of the second layer do not have a statistically favored orientation necessary for bridging first layer waters. In the major groove, the hydration geometry of the (A.T) base-pair edge resembles the main features of the AT-pair hydration derived from other sequences for the canonical B form. The preference of the B' conformation for oligo(dA).oligo(dT) tracts may express the tendency to common hydration of base-pair edges of successive base pairs in the grooves of B-type DNA. The mean potential energy of hydration of canonical B-DNA was estimated to be -60 to -80 kJ/mole nucleotides in dependence on the (G.C) contents. Because of the small system size, this estimation is preliminary.     

Thermodynamics and kinetics for base-pair opening in the P1 duplex of the Tetrahymena group I ribozyme

The thermodynamics and kinetics for base-pair opening of the P1 duplex of the Tetrahymena group I ribozyme were studied by NMR hydrogen exchange experiments. The apparent equilibrium constants for base pair opening were measured for most of the imino protons in the P1 duplex using the base catalysts NH3, HPO4(2-) or TRIS. These equilibrium constants were also measured for several modified P1 duplexes, and the C-2.G23 base pair was the most stable base pair in all the duplexes. The conserved U-1*G22 base pair is required for activity of the ribozyme and the data here show that this wobble base pair destabilizes neighboring base pairs on only one side of the wobble. A 2'-OMe modification on the U-3 residue stabilized its own base pair but had little effect on the neighboring base pairs. Three base pairs, U-1*G22, C-2*G23 and A2*U21 showed unusual equilibrium constants for opening and possible implications of the opening thermodynamics of these base pairs on the undocking rates of the P1 helix with catalytic core are discussed.     

Structural studies of DNA fragments: the G.T wobble base pair in A, B and Z DNA; the G.A base pair in B-DNA

The crystal structures of five double helical DNA fragments containing non-Watson-Crick complementary base pairs are reviewed. They comprise four fragments containing G.T base pairs: two deoxyoctamers d(GGGGCTCC) and d(GGGGTCCC) which crystallise as A type helices; a deoxydodecamer d(CGCGAATTTGCG) which crystallises in the B-DNA conformation; and the deoxyhexamer d(TGCGCG), which crystallises as a Z-DNA helix. In all four duplexes the G and T bases form wobble base pairs, with bases in the major tautomer forms and hydrogen bonds linking N1 of G with O2 of T and O6 of G with N3 of T. The X-ray analyses establish that the G.T wobble base pair can be accommodated in the A, B or Z double helix with minimal distortion of the global conformation. There are, however, changes in base stacking in the neighbourhood of the mismatched bases. The fifth structure, d(CGCGAATTAGCG), contains the purine purine mismatch G.A where G is in the anti and A in the syn conformation. The results represent the first direct structure determinations of base pair mismatches in DNA fragments and are discussed in relation to the fidelity of replication and mismatch recognition.