Neutron diffraction studies on crystals of nucleosome cores using contrast variation

Neutron diffraction data have been collected from crystals of intact nucleosome core particles to a resolution of 25 Å. By varying the proportion of D₂O, the scattering of the mother liquor relative to the protein and DNA can be altered. At 39% D₂O, the solvent scattering matches that of the protein and so only the DNA is scattering, and similarly at 65% D₂0 only the protein scatters. Using this approach the neutron scattering of the two components and of the complete particle (0% D₂O) have been measured. The data corresponding to the principal projections are consistent with a model in which 1.8 turns of a DNA superhelix of pitch 27·5 Å and radius 42 Å are wound around a protein core.     

An investigation of the structure of Alfalfa mosaic virus by small-angle neutron scattering

Small-angle neutron scattering experiments have been performed on the tubular bottom component of Alfalfa mosaic virus (AMV) and the “30 S” particle (a quasispherical reassembled AMV coat protein particle) with the aim of determining the internal structure of the virus. Scattering curves were obtained out to a resolution of $\text{1}\text{50}\text{A}\text{?}{}^{\mathrm{?1}}$ at a number of H₂O/²H₂O ratios and were analysed using a model fitting technique. This involves calculating the scattering intensity due to a parameterised distribution of scattering density representing the particle and comparing this to the experimental data after taking into account the effect of instrumental smearing. The use of the contrast variation method enables the internal consistency of the model to be well tested.Three models are used in an attempt to explain the scattering curve of the 30 S particle. A single homogeneous shell is shown to be inadequate and two other models introducing the presumed T = 1 icosahedral symmetry of the particle are presented and discussed. The most satisfactory of these consists of 60 spherical monomers of radius 19 Å symmetrically placed in pairs about the 2-fold icosahedral positions.The analysis of the bottom component data has yielded a low resolution model for the virus, which is shown to be consistent with its composition as given by earlier physico-chemical measurements. In the model the RNA is uniformly packed throughout the interior of the capsid (which is cylindrical with hemispherical ends) out to a radius of about 65 Å and with a packing fraction of 20%. Within the limitations of an homogeneous shell model, the protein capsid has an outer radius of 94 Å and thickness of 23 Å, but arguments are presented based on the marked lattice structure of the cylindrical capsid and the analysis of the scattering data of the 30 S particle, that this model underestimates the thickness of the protein shell and that it in fact makes contact with the RNA at about 65 Å.     

Study of the internal structure of Escherichia coli ribosomes by neutron and x-ray scattering.

Neutron scattering curves of the small and large subparticles of Escherichia coli ribosomes are presented over a wide range of scattering angles and for several contrasts. It was verified that the native ribosome structure was not affected by ²H₂O in the buffer. The reliability of the neutron scattering curves, obtained in H₂O buffer, was established by X-ray scattering experiments on the same material.The non-homogeneous distribution of RNA and protein in the subparticles of E. coli ribosomes is confirmed, with RNA predominantly within the particle and protein predominantly on its periphery. The distances between the centres of gravity of the RNA and protein components do not exceed 25 Å and 30 Å, in the large and small subparticles, respectively.The volume occupied by the RNA within the large and small subparticles is determined. The ratio of the “dry” volume of the RNA to the occupied volume is found to be 0.56; it is the same in both subparticles. Such packing of RNA is characteristic of single helices of ribosomal RNA at their crystallization and of the helices in transfer RNA crystals. A conclusion is drawn that RNA in ribosomes is in a similar state.Experimental scattering curves for the small subparticle depend significantly on the contrast in the angular region in which the scattering is mainly determined by the particle shape. The scattering curve, as infinite contrast is approached, is similar to that calculated for the particle as observed by electron microscopy. Thus, the long-existing contradiction between electron microscopy data (an elonggated particle with an axial ratio 2:1) and X-ray data (an oblate particle with an axial ratio 1:3.5), concerning the overall shape of the 30 S subparticle, is settled in favour of electron microscopy. The experimental neutron scattering curve of RNA within the small subparticle is well-described by the V-like RNA model proposed recently by Vasiliev et al. (1978).Experimental data are given to support the hypothesis that the maxima in the X-ray scattering curves, in the region of scattering angles corresponding to Bragg distances of 90 to 20 Å, arise from the ribosomal RNA component alone. It is shown that the prominence of the peaks in this region of the scattering curve depends only on the scattering fraction of the RNA component. The scattering fraction can be changed both by using the “native contrast” (ribosomal particles containing different amounts of protein) and by varying the solvent composition. The maxima are most pronounced where the RNA scattering fraction is highest or in solvents where the protein density is matched by the solvent. The scattering vectors of the maxima in the X-ray and neutron scattering curves, however, remain unchanged. This allows us to propose the tight packing of RNA as a common principle for the structural arrangement of RNA in ribosomes.     

Neutron scattering studies of escherichia coli tyrosyl-trna synthetase and of its interaction with trna tyr

Small-angle neutron scattering studies of Escherichia coli tyrosyl-tRNA synthetase indicate that in solution this enzyme is a dimer of M_r, 91 (±6) × 10³ with a radius of gyration R_G of 37.8 ± 1.1 Å.The increase in the scattering mass of the enzyme upon binding tRNA^Tyr has been followed in 20 mm-imidazole · HCl (pH 7.6), 10 mm-MgCl₂, 0.1 mm-EDTA, 10 mm-2-mercaptoethanol, 150 mm-KCl. A stoichiometry of one bound tRNA per dimeric enzyme molecule was found. The R_G of the complex is equal to 41 ± 1 Å. Titration experiments in 74% ²H₂O, close to the matching point of tRNA, show an R_G of 38.5 ± 1 Å for the enzyme moiety in the complex. From these values, a minimum distance of 49 Å between the centre of mass of the bound tRNA and that of the enzyme was calculated.In low ionic strength conditions (20 mm-imidazole-HCl (pH 7.6), 10 mm-MgCl₂, 0.1 mm-EDTA, 10 mm-2-mercaptoethanol) and at limiting tRNA concentrations with respect to the enzyme, titrations of the enzyme by tRNA^Tyr are characterized by the appearance of aggregates, with a maximum scattered intensity at a stoichiometry of one tRNA per two enzyme molecules. At this point, the measured M_r and R_G values are compatible with a compact 1:2, tRNA: enzyme complex. This complex forms with a remarkably high stability constant: (enzyme:tRNA:enzyme)/(enzyme:tRNA)(enzyme) of 0.1 to 0.3(× 10⁶) m^?1 (at 20 °C). Upon addition of more tRNA, the complex dissociates in favour of the 1:1, enzyme:tRNA complex, which has a higher stability constant (1 to 3 (× 10⁶) m^?1).     

Yeast tRNAAsp-Aspartyl-tRNA Synthetase Complex: Low Resolution Crystal Structure

Abstract

Yeast aspartyl-tRNA synthetase, a dimer of molecular weight 125000, and two molecules of its cognate tRNA (M_r = 24160) cocrystallize in the cubic space group 1432 (a = 354 Å). The crystal structure was solved to low resolution using neutron and X-ray diffraction data. Neutron single crystal diffraction data were collected in five solvents differing by their D₂O content in order to use the contrast variation method to distinguish between the protein and tRNA The synthetase was first located at 40 Å resolution using the 65% D₂O neutron data (tRNA matched). tRNA molecules were found at 20 Å resolution using both neutron and X-ray data. The resulting model was refined against 10 Å resolution X-ray data, using density modification and least-squares refinement of the tRNA positions. The crystal structure, solved without a priori phase knowledge, was confirmed later by isomorphous replacement. The molecular model of the complex is in good agreement with results obtained in solution by probing the protected part of the tRNA by chemical reagents.     

Neutron small angle scattering of the Mo-Fe protein (nitrogenase) from Clostridium pasteurianum

Neutron small angle scattering measurements of solutions of the Mo-Fe protein from $\text{C. pasteurianum}$ have yielded the following results. The molecular weight of the protein is 208,000 ± 10,000, in agreement with figures obtained by other methods. The radius of gyration is 39.8 ± 0.7 Å in H₂O, and 37.6 ± 0.3 Å in D₂O. The experimental scattering curves have been compared with the calculated scattering curves of simple homogeneous bodies. It is concluded that the MoFe protein from $\text{C. pasteurianum}$ is a non spherical particle having an axial ratio of 2:1, and that it probably has little, if any, solvent containing cavities.     

Structure of the tomato bushy stunt virus: a model for protein-RNA interaction

The structure of tomato bushy stunt virus has been investigated by neutron small-angle scattering. Data were collected in aqueous solutions containing various amounts of D₂O, and the radial distribution of both protein and RNA could be computed. The main feature of that distribution is the clustering of protein into two concentric shells separated by about 30 Å, where the density is so low that the polypeptide chain must be extended. Most of the RNA is located between these two protein shells. The implications of that structure for protein-RNA interactions are discussed.     

Neutron scattering study of the 13 S fragment of 16 S RNA and its complex with ribosomal protein S4

A fragment with a molecular weight of 170,000 and a sedimentation coefficient of 13 S which is capable of specifically binding ribosomal protein S4 has been obtained by digestion of Escherichia coli 16 S RNA with ribonuclease A. The 13 S fragment of 16 S RNA and its complex with protein S4 have been studied by different physical methods; in the first place, by neutron scattering. It has been shown that this fragment is very compact in solution. The radii of gyration of this fragment (50 ± 3 Å) and of protein S4 within the complex (17 ± 3 Å) coincide, within the limits of experimental error, with the radii of gyration for the free RNA fragment (47 ± 2 Å) and the free ribosomal protein S4 in solution (18 ± 2 Å). Hence the conclusion is drawn that the compactness of the RNA fragment and the ribosomal protein does not change on complex formation. The compact 13 S fragment of 16 S RNA is shown to be contrast-matched in solvent containing 70% ²H₂O which corresponds to a value for the partial specific volume of RNA of 0.537 cm³/g.     

Structural investigations of influenza B virus

Purified preparations of influenza B/Hong Kong/5/72 have been characterized by hydrodynamical measurements, electron microscopy and small-angle neutron scattering. As judged by these techniques the preparations are highly monodisperse, the virus particles being spherical and of molecular weight about 200 × 10⁶. The lipid bilayer is located at a radius of 425 Å and its molecular weight is estimated to be 60 × 10⁶, constituting about 30% of the total virus mass. The external radius is about 580 Å.     

The structure of lipid bilayers and the effects of general anaesthetics: An X-ray and neutron diffraction study

We have looked for the effects of three clinically used inhalational anaesthetics (nitrous oxide, halothane and cyclopropane) on the structure of lecithin/ cholesterol bilayers. The anaesthetics were delivered to the membranes in the gaseous phase, so that effects at clinical concentrations could be determined.High-resolution X-ray diffraction patterns were recorded out to 4 Å and analyzed using swelling experiments. Parallel neutron diffraction experiments were performed and analyzed using H₂O-²H₂O exchange. Methods were developed which enabled us to obtain confidence limits for the X-ray and neutron structure factors.The resultant X-ray and neutron scattering density profiles clearly define the positions of the principal molecular groups in the unperturbed bilayer. In particular, the high-resolution electron density profiles reveal features directly attributable to the cholesterol molecule. A comparison with the neutron scattering density profiles shows that cholesterol is anchored with its hydroxyl group at the water/hydrocarbon interface, aligned with the fatty acid ester groups of the lecithin molecule. We suggest that this positioning of the cholesterol molecule allows it to act as a thickness buffer for plasma membranes.In the presence of very high concentrations of general anaesthetics, the bilayers show increased disorder while maintaining constant membrane thickness. At surgical concentrations, however, there are no significant changes in bilayer structure at 95% confidence levels. We briefly review the literature previously used to support lipid bilayer hypotheses of general anaesthesia. We conclude that the lipid bilayer per se is not the primary site of action of general anaesthetics.     

Comparison of neutron and X-ray scattering of dilute myoglobin solutions.

Experimental results obtained by neutron scattering of dilute solutions of myoglobin are compared with those obtained by X-ray scattering. X-ray scattering remains the more powerful technique at wider angles above 0.3 Å⁻¹, where neutron experiments are less accurate because of low coherent scattering probability and high incoherent background. Neutron scattering is preferable at momentum transfers below 0.2 Å⁻¹; the conditions for applying the contrast variation method for the evaluation of the three basic scattering functions, which are due to shape and internal structure, equation (3), are ideally fulfilled in this region. Furthermore, neutrons allow observation of the hydrogen-deuterium exchange within the protein.     

Contrast variation studies of clathrin coated vesicles by small-angle neutron scattering

Structural information on clathrin coated vesicles has been obtained by small angle neutron scattering using contrast variation. A characteristic peak in the neutron scattering profile, which is apparent in 75 % D₂O, as well as in H₂O, disappears when contrast matching the protein component of the coated vesicles in 42% D₂O. Neutron, as well as dynamic, light scattering give a coated vesicle size of about 900 Å in H₂O and D₂O, but for neutron scattering the diameter decreases when matching out the protein coat of the clathrin coated vesicles. From the match point for the clathrin coated vesicles it is demonstrated that the clathrin cages do contain internal membrane. The mass of 34 MD and composition of 75% protein and 25% lipid found from the analysis of the small-angle scattering data are both in good agreement with the values reported in the literature. Electron microscopy gives an average outer diameter of 880 Å for the coated vesicles and an average diameter of 460 Å for the vesicle itself. Offprint requests to: Correspondence to: R. Bauer     

The crystal structure of 2-deoxy-β-d-arabino-hexopyranose at −150°

2-Deoxy-β-d-arabino-hexopyranose, C₆H₁₂O₅, is orthorhombic, P2₁2₁2₁, with cell dimensions at ?150° [20°], a = 6.484(2) [6.510(3)], b = 10.364(2) [10.427(4)], c = 11.134(3) [11.153(5)] Å, V = 748.2 [757.1] ų, Z = 4, D_x = 1.457 [1.440], and D_m = [1.455] g.cm^?3. The intensities of 1269 reflections were measured by using MoKα radiation. The structure was solved by direct methods, and refined by full-matrix least-squares, with anisotropic, thermal parameters for the carbon and oxygen atoms, and isotropic parameters for the hydrogen atoms. The pyranose has the ⁴C₁(d) conformation, with puckering parameters Q = 0.563 Å, θ = 3.9°, and ? = 350.3°. The departure from ideality is very small, and less than that in β-d-glucopyranose, Q = 0.584 Å and θ = 6.9°. The β-glycosidic, CO bond is short, 1.383(4) Å, and the OCOH torsion angle is ?87°, consistent with the anomeric effect. The hydrogen-bonding scheme consists of infinite chains, with side chains terminating at a ring-oxygen atom.     

Topography of cyclodextrin inclusion complexes : Part II. The iodine-cyclohexa-amylose tetrahydrate complex; its molecular geometry and cage-type crystal structure

Iodine-cyclohexa-amylose tetrahydrate [(C₆H₁₀O₅)₆ ·I₂·d4H₂O] crystallizes in the orthorhombic space-group P2₁2₁2₁, a  14.240 Å, b  36.014 Å, c  9.558 Å. The structure was solved by heavy-atom techniques and refined by least-squares methods to a conventional discrepancy index R  0.148 for the 2872 observed data. The six d-glucose residues are in the C1 chair conformation; the conformational angles vary in magnitude from 45 to 66°, the angles O(5)-C(5)-C(6)-O(6) are close to · 70°, and the six O(4) atoms are almost coplanar (r.m. s. displacement 0.13 Å). Only four of the six O(2) ?O(3) intramolecular hydrogen bonds have formed, which renders the molecule less symmetrical and more conical-shaped than in the previously determined α-cyclodextrin-potassium acetate complex. The iodine molecule is coaxial with the cyclohexa-amylose molecule. The I-I distance is a conventional 2.677 Å. Close interactions between the iodine atoms and the host molecule comprise carbon atoms C(5) and C(6) and oxygen atoms O(4), with interatomic distances all equal to or greater than van der Waals contacts. Intermolecular, almost-linear, short contacts O ? I-I?O with I?O distances of 3.22 and 3.07 Å indicate attractive interaction.The molecules are arranged in herring-bone “cage-type” fashion, with the four water molecules as space-filling mediators; the structure is held together by an intricate network of hydrogen bonds.     

Impact of lyoprotectors on protein-protein separation in the solid state: Neutron- and X-ray-scattering investigation

BackgroundPolyhydroxycompounds (PHC) are used as lyoprotectors to minimize aggregation of pharmaceutical proteins during freeze-drying and storage.MethodsLysozyme/PHC mixtures with 1:1 and 1:3 (w/w) ratios are freeze-dried from either H₂O or D₂O solutions. Disaccharides (sucrose and trehalose), monosaccharide (glucose), and sugar alcohol (sorbitol) are used in the study. Small-angle neutron and X-ray scattering (SANS and SAXS) are applied to study protein-protein interaction in the freeze-dried samples.ResultsProtein interaction peak in the freeze-dried mixtures has been detected by both SANS (D₂O-based samples only) and SAXS (both D₂O- and H₂O-based). In the 1:1 mixtures, protein separation distances are similar (center-of-mass distance of approx. 31 Å) between all lyoprotectors studied. Mixtures with a higher content of the disaccharides (1:3 ratio) have a higher separation distance of approx 40 Å. The higher separation could reduce protein-protein contacts and therefore be associated with less favourable aggregation conditions. In the 1:3 mixtures with glucose and sorbitol, complex SANS and SAXS/WAXS patterns are observed. The pattern for the glucose sample indicate two populations of lysozyme molecules, while the origin of multiple SAXS peaks in the lysozyme/sorbitol 1:3 mixture is uncertain.ConclusionsProtein-protein separation distance is determined predominantly by the lyoprotector/protein weight ratio.General significanceUse of SANS and SAXS improves understanding of mechanisms of protein stabilization by sugars in freeze-dried formulations, and provide a tool to verify hypothesis on relationship between protein/protein separation and aggregation propensity in the dried state.     

Hydrated {Mo72Fe30} clusters: Low-frequency hydrogen modes and self-aggregation

Using incoherent quasi-elastic and inelastic neutron scattering, we have investigated the hydrogen relaxational dynamics and hydrogen vibrational modes in the polyoxomolybdate specie [Mo₇₂Fe₃₀O₂₅₂(CH₃COO)₁₂[Mo₂O₇(H₂O)]₂[H₂Mo₂O₈(H₂O)](H₂O)₉₁]· ≈ 150 H₂O. The translational dynamics of the water molecules in the compound is profoundly different from that of bulk water at the same temperature showing a non-Debye relaxation behavior. The temperature dependence of the relaxation time can be described in terms of an Arrhenius law, indicating that the dynamics is triggered by the breaking of the bonds connecting the crystal water molecules with the hydrophilic nanocapsule surfaces. Inelastic neutron scattering spectra confirm the attenuation of water translational modes with respect to the bulk water case due to the strong destructuring effect imposed by the nanocage interface and the enhancement of the highest frequency librational mode as already found in hydrated Vycor or Gelsil matrix. Small angle X-ray scattering on freshly prepared aqueous solution evidences the presence of nanocapsule structures proper of the monomer (2.6 nm in diameter) that coexist with a small amount of oligomers. After 1 month the polyoxomolibdate specie self-assembles in a supramolecular structure with a polydisperse distribution of dimensions spanning from the monomer to the “blackberry” vesicular structure already reported in literature.     

Comparative structural studies of the first row early transition metal(III) chloride tetrahydrofuran solvates

Two compounds of empirical formula MCl₃- (THF)₃, M = V and Cr, have been characterized by single crystal X-ray studies. The VCl₃(THF)₃ molecule, which has a mer octahedral stereochemistry, crystallizes in the monoclinic space group P2₁/c with a= 8.847(2),b= 12.861(5),c= 15.134(3) Å, β = 91.94(2)°, V = 1721(1) ų and Z = 4. The V-Ci(1) and V-CI(2) distances have a mean value of 2.330 [3] Å while V-CI(3) = 2.297(2) Å, The VO(1) and VO(2) distances have a mean value of 2.061[8] Å while V-O(3) = 2.102(3) Å cis ClVCl angles average 92.0[5]° and cis OVO angles average 86.2[2]° . The isostmctural complex, CrCl₃(THF)₃, has a crystal structure made up of discrete octahedral mer-CrCl₃(THF)₃ molecules with the following unit cell dimensions (space group P2₁/c): a = 8.715(1), b= 12.786(3), c = 15.122(3) Å, β = 92.15(1)°, V = 1684(1) ų and Z = 4. The CrCl(1) and CrCl(2) distances have a mean value of 2.310131 Å while CrCl(3) = 2.283(2) Å. The CrO(1) and CrO(2) distances have a mean value of 2.0101171 Å while CrO(3) = 2.077(4) Å. cis ClCrCl angles average 90.9[4]° and cis OCrO angles average 86.1 [2]°. The structures of these two octahedral complexes and those previously reported for ScCl₃(THF)₃ and TiCl₃(THF)₃ are compared and certain general trends are discussed.     

Metal—phenoxyalkanoic acid interactions. Part 14. The crystal and molecular structures of diaquabis(4-fluorophenoxyacetato)copper(II) bis(4-fluorophenoxyacetic acid)dihydrate and tetra-μ-(4-fluorophenoxyacetato-0,0′)-bis[(2-aminopyrimidine)copper(II)]

The crystal structures of two copper(II) complexes of 4-fluorophenoxyacetic acid (4-FPAH) have been determined by X-ray diffraction. [Cu(4-FPA)₂(H₂O)₂]·2(4-FPAH)·2H₂O (1) is triclinic, space group P1 with Z = 1 in a cell of dimensions a = 14.808(2), b = 9.832(2), c = 6.847(2) Å, α = 87.77(2), β = 98.41(2), γ = 112.33(2)° and was refined to a residual of 0.038 for 1697 ‘observed’ reflections. The coordination sphere in this complex is tetragonally distorted octahedral comprising two waters [CuO, 1.940(3) Å], two unidentate carboxylate oxygens [CuO, 1.942(2) Å] and two ether oxygens [CuO, 2.471(2) Å]. Two adducted [4-FPAH] acid molecules are linked to the un-coordinated oxygens of the acid ligands by hydrogen bonds [2.547(4) Å]. [Cu₂(4-FPA)₄(2-aminopyrimidine)₂] (2) is triclinic, space group P1 with Z = 1 in a cell of dimensions a = 12.688(2), b = 11.422(2), c = 7.951(1) Å, α = 78.74(1), β = 107.51(1), γ = 75.78(1)°, and was refined to a residual of 0.042 for 2683 ‘observed’ reflections. (2) is a centrosymmetric tetracarboxylate bridged dimer with four similar CuO (equatorial) distances [1.967–1.987 Å; 1.977(3) Å mean] and the axial position occupied by the hetero nitrogen of the 2-aminopyrimidine ligand [CuN, 2.176(3) Å]. The Cu---Cu separation is 2.710(1) Å. Crystal data are also presented which confirm the isostructurality of complex (2) with [Cu₂(phenoxyacetate)₄(2-aminopyrimidine)₂], the Co^II, Mg^II and Mn^II4-fluorophenoxyacetate complexes with their phenoxyacetic and 4-chlorophenoxyacetic acid analogues, and of Cd^II4-fluorophenoxyacetate with Cd^II and Zn^II phenoxyacetates.     

Synthesis and x-ray structure of a Hg(II) complex with adenine N(1)-oxide. A model for mercury binding to DNA

The synthesis and crystal structure of the adenine N(1)-oxide complex with mercury(II) chloride, (C₅H₅N₅O)HgCl₂ are reported. Crystals of the coordination compound belong to the monoclinic system, space group P2₁/n with the following primary crystallographic data: a = 6.685(1) Å, b = 11.798(2) Å, c = 10.155(1) Å, β = 100.22(1)°, V = 906.04 ų, Z = 4. The structure was elucidated by conventional Patterson and Fourier methods and refined by the full matrix least-squares technique on the basis of 1977 observed reflections to an R value of 0.074. The basic unit of the structure is a dimer, with a centre of symmetry, consisting of two HgCl₂ moieties and two adenine N(1)-oxide ligands. A polymeric structure results from the bridging interactions of chloride ions. Adenine N(1)-oxide acts as a bidentate bridging ligand, coordinating through N(7) and O(1). The coordination geometry around the mercury ion is a distorted square pyramid with N(7) and three chlorines (two of which are centro-symmetrically related) forming the square plane and O(1) occupying the axial position. Hg also interacts indirectly with N(6) through a Cl

HN hydrogen bond. Principal intracomplex geometrical parameters are as follows: HgN(7) = 2.61(1) Å, HgO(1) = 2.55(1) Å, HgCl(1) = 2.330(3) Å, HgCl(2) = 2.318(3) Å, HgCl(2′) = 3.347(3) Å. The cis angles range from 77.5° to 107.9° and the two trans angles are 155.5° and 163.1°. The centro-symmetrically related bases overlap partially and pack at a distance of 3.2 Å. The glide-related bases are linked by a hydrogen bond, N(9)H

O(1) and are inclined to one another by 109.7°. The results are compared with those derived from spectroscopic and other physicochemical studies on metal interaction with adenine N(1)-oxide. Based on the present structural observations and earlier experimental results a possible mechanism is proposed for mercury interaction with DNA.