AT base pairs are less stable than GC base pairs in Z-DNA: The crystal structure of d(m5CGTAm5CG)

Two hexanucleoside pentaphosphates, 5-methyl and 5-bromo cytosine derivatives of d(CpGpTp-ApCpG) have been synthesized, crystallized, and their three-dimensional structure solved. They both form left-handed Z-DNA and the methylated derivative has been refined to 1.2 Å resolution. These are the first crystal Z-DNA structures that contain AT base pairs. The overall form of the molecule is very similar to that of the unmethylated or the fully methylated (dC-dG)₃ hexamer although there are slight changes in base stacking. However, significant differences are found in the hydration of the helical groove. When GC base pairs are present, the helical groove is systematically filled with two water molecules per base pair hydrogen bonded to the bases. Both of these water molecules are not seen in the electron density map in the segments of the helix containing AT base pairs, probably because of solvent disorder. This could be one of the features that makes AT base pairs form Z-DNA less readily than GC base pairs.     

Parallel-stranded DNA with mixed AT/GC composition: role of trans G.C base pairs in sequence dependent helical stability

Parallel-stranded (ps) DNAs with mixed AT/GC content comprising G.C pairs in a varying sequence context have been investigated. Oligonucleotides were devised consisting of two 10-nt strands complementary either in a parallel or in an antiparallel orientation and joined via nonnucleotide linkers so as to form 10-bp ps or aps hairpins. A predominance of intramolecular hairpins over intermolecular duplexes was achieved by choice of experimental conditions and verified by fluorescence determinations yielding estimations of rotational relaxation times and fractional base pairing. A multistate mode of ps hairpin melting was revealed by temperature gradient gel electrophoresis (TGGE). The thermal stability of the ps hairpins with mixed AT/GC content depends strongly on the specific sequence in a manner peculiar to the ps double helix. The thermodynamic effects of incorporating trans G.C base pairs into an AT sequence are context-dependent: an isolated G. C base pair destabilizes the duplex whereas a block of > or =2 consecutive G.C base pairs exerts a stabilizing effect. A multistate heterogeneous zipper model for the thermal denaturation of the hairpins was derived and used in a global minimization procedure to compute the thermodynamic parameters of the ps hairpins from experimental melting data. In 0.1 M LiCl at 3 degrees C, the formation of a trans G.C pair in a GG/CC sequence context is approximately 3 kJ mol(-)(1) more favorable than the formation of a trans A.T pair in an AT/TA sequence context. However, GC/AT contacts contribute a substantial unfavorable free energy difference of approximately 2 kJ mol(-)(1). As a consequence, the base composition and fractional distribution of isolated and clustered G.C base pairs determine the overall stability of ps-DNA with mixed AT/GC sequences. Thus, the stability of ps-DNA comprising successive > or =2 G.C base pairs is greater than that of ps-DNA with an alternating AT sequence, whereas increasing the number of AT/GC contacts by isolating G.C base pairs exerts a destabilizing effect on the ps duplex. Molecular modeling of the various helices by force field techniques provides insight into the structural basis for these distinctions.     

Monte Carlo simulation of hydration of the nucleic acid fragments

Systems containing a base or a base pair and 25 water molecules, as well as a helical stack and 30 water molecules per base pair, have been simulated. Changes in the base hydration shell structure, after the bases have been included into the pair and then into the base pair stack, are discussed. Hydration shells of several configurations of the base pair stacks are discussed. Probabilities of formation of the hydrogen-bonded bridges of 1, 2 and 3 water molecules between hydrophilic centres have been estimated. The hydration shell structure was shown to depend on the nature of the base pair and on the stack configuration, while dependence of the global hydration shell characteristics on the stack configuration has been proved to be rather slight. The most typical structural elements of hydration shells, in the glycosidic (minor in B-like conformation) and non-glycosidic (major) grooves, for different configurations of AU and GC stacks, have been found and discussed. The number of hydrogen bonds between water molecules and bases per water molecule was shown to change upon transformation of the stack from A to B configuration. This result is discussed in connection with the reasons for B to A conformational transition and the concept of "water economy". Hydration shell patterns of NH2-groups of AU and GC helical stacks differ significantly.     

Molecular dynamics study of solvation effect on diffusivity changes of DNA fragments

DNA sequence analyzing and base pair separation techniques have attracted much attention, such as denaturing gradient gel electrophoresis, temperature gradient gel electrophoresis, and capillary electrophoresis. However, details of sequence separation mechanisms in electrophoresis are not clarified enough. Understanding and controlling flow characteristics of DNA are important not only for fundamental research but also for further developments of bio-nano technologies. In the present study, we theoretically discuss the relationship between diffusivity and hydrated structures of DNA fragments in water solvent using molecular dynamics methods. In particular, influence of base pair substitutions on the diffusivity is investigated, focusing on an adenine-thymine (AT) rich B–DNA decamer 5’-dCGTATATATA-3’. Consequently, it is found that water molecules that concentrate on dissociated base pairs form hydrated structures and change the diffusivity of DNA decamers. The diffusion coefficients are affected by the substitution of GC for AT because of the different manner of interactions between the base molecules and water solvent. This result predicts a possibility of base pair separation according to differences in the diffusivity.     

Effect of a mechanical tension on the hydration of DNA in fibres.

Fiber X-ray diffraction and measurement of fibre dimensions yield information about the effects of a mechanical tension on hydration of DNA in fibres. At a given relative humidity, the mechanical tension changes the DNA conformation but does not modify the number of water molecules associated to a nucleotide. The number of water molecules per nucleotide necessary to maintain B form decreases for increasing tensions applied to the DNA fibre. Form transitions can be opposed by mechanical tensions; an energy of 1 Kcal per mole of nucleotide pairs is sufficient to prevent the B to A transition.     

Salt effects on the denaturation of DNA

When DNA's of differing GC:AT base ratios, e.g. synthetic poly dAT, T4 DNA,calf thymus DNA, E. coli DNA, and M. lysodeikticus DNA, are heat-denatured at neutral pH in increasing concentrations of N(a)(2)SO(4) or C(s)(2)SO(4) as supporting electrolytes,the variation of melting temperature with average base composition, dT(m)/dX(G)(C), changes from 45°C (in 0.002M Na) to ll°C (in 4.5M Na) and from 42°C (in 0.002M Cs) to 3°C(in 4.5M Cs). The decrease of dT(m)/dX(G)(C) is a monotonic function of decreasing water activity in the salt solutions. We interpret this decreased composition dependence of the thermal stability of the various DNA's as being due to a destabilization of the GC base pairs relative to the AT base pairs by the concentrated salt media. A simple quantitative treatment shows that k = 8GC/SAT decreases from a value of 4.14 (in 0.01MN(a)) to 1.86 (in 3M Na) and from 4.18 (in 0.01M Cs) to 1.42 (in 3M Cs). SAT is the equilibrium constant for the formation of a hydrogen-bonded AT base pair from a pair of unbonded bases at the junction between a helical region and a denatured region and SGC is the like constant for the formation of a GC base pair. These results corroborate our previous findings of a strongly reduced composition dependence of the negative logarithm of the methylmercuric hydroxide concentration necessary to produce 50% denaturation when the helix-coil transition of DNA is studied in concentrated Cs(s)SO(4)(ultracentrifugation) instead of in dilute N(a)(2)SO(4) (ultraviolet spectrophotometry).     

Base pair opening dynamics of a 2-aminopurine substituted Eco RI restriction sequence and its unsubstituted counterpart in oligonucleotides.

Studies of 1H NMR selective saturation recovery were performed to determine the imino proton exchange with solvent water of the base pairs in the Eco RI endonuclease recognition sequence GAATTC, placed at the center of self-complementary decamer and dodecamer oligonucleotides. In one oligonucleotide the innermost adenine was replaced by the fluorescent base analogue 2-aminopurine (2AP). From the measurements at different concentrations of TRIS buffer acting as proton exchange catalyst, base pair lifetimes were evaluated. The results at 25 degrees show that the AT base pairs have lifetimes of the order of a few ms, whereas the surrounding GC base pairs in a dodecamer have lifetimes of about 100 ms. The (2AP)T base pair has a shorter lifetime than the corresponding AT base pair. The temperature dependent optical absorption, and for the 2AP containing oligonucleotide fluorescence, were used to study the single strand-duplex equilibrium of the decamers. The results indicate that NMR and the optical techniques, although applied at very different concentrations, monitor the same conformational transition of the oligonucleotide.     

Point mutations change the thermal denaturation profile of a short DNA fragment containing the lactose control elements. Comparison between experiment and theory.

To understand the denaturation process of short DNA segments we have chosen a 203-base pair (bp) restriction fragment containing the lactose control region. A steady decrease in GC content exists between its i proximal and z proximal ends. We confirm that this fragment melts at low salt in two subtransitions. A GC to AT mutation in the AT-rich region (mutation UV5) increases the number of denatured base pairs in the first subtransition and decreases the cooperativity of the melting process. A GC to AT mutation in the GC-rich region (mutation L8) decreases the number of denatured base pairs in the first subtransition and increases the cooperativity. These mutations induce the same shift in the temperature of half denaturation. The effects of both mutations are additive. A short deletion at the z end of the fragment affects only the first subtransition. When four GC pairs are added to both end, the fragment melts in one transition. Comparison with the results obtained with a larger 789-bp lac fragment reveals strong end effects on base pair stability and suggests that denaturation of the 203-bp fragment proceeds unidirectionally from the z end. Good agreement is shown with the predictions made with the "z ipper model" of Crothers et al. (1965).     

Interactions of 4'', 6-diamidine-2-phenylindole with synthetic polynucleotides.

4', 6-Diamidine-2-phenylindole forms fluorescent complexes with synthetic DNA duplexes containing AT, AU and IC base pairs; no fluorescent complexes were observed with duplexes containing GC base pairs or with duplexes containing a single AT base pair sandwiched between GC pairs. The binding site size is one molecule of dye per 3 base pairs. The intrinsic binding constants are higher for alternating sequence duplexes than for the corresponding homopolymer pairs. With the exception of the four-stranded helical poly rI which exhibits considerable fluorescence enhancement upon binding of the ligand, none of the single- or multi- stranded polyribonucleotides and ribo-deoxyribonucleotide hybrid structures form fluorescent complexes with the dye. Poly rI is the only RNA which forms a DNA B-like structure (Arnott et al. (1974) Biochem. J. 141, 537). The B conformation of the helix and the absence of guanine appear to be the major determinants of the specificity of the fluorescent binding mode of the dye. Nonfluorescent interactions of the dye with polynucleotides are nonspecific; UV absorption and circular dichroic spectra demonstrate binding to synthetic single- and double-stranded DNA and RNA analogs, including those containing GC base pairs.     

Structural principles of B-DNA grooves hydration in fibers as revealed by Monte Carlo simulations and X-ray diffraction

Being interested in possible effects of sequence-dependent hydration of B-DNA with mixed sequence in fibers, we performed a series of Monte Carlo calculations of hydration of polydeoxyribonucleotides in B form, considering all sequences with dinucleotide repeat. The computational results allow the ten base-stacking types to be classified in accordance with their primary hydration in the minor groove. As a rule, the minor groove is occupied by two water molecules per base pair in the depth of the groove, which are located nearly midway between the planes of successive base pairs and symmetrically according to the dyad there. The primary hydration of the major groove depends on the type of the given base pair. The coordinates of 3 water molecules per base pair in the depth of the major groove are determined by the type of this pair together with its position and orientation in the helix, and are practically independent on the adjacent base pairs. A/T-homopolymer tracts do not fit into this hydration pattern; the base pair edges are hydrated autonomously in both grooves. Analysis of the Li-B-DNA x-ray diffraction intensities reveals those two water positions in the minor groove. In the major groove, no electronic density peaks in sufficient distance from the base edges were found, thus confirming the absence of any helical invariance of primary hydration in this region. With the help of the rules proposed in this paper it is possible to position the water molecules of the first hydration shell in the grooves of canonical B-DNA for any given sequence.     

The B-Z transition in supercoiled DNA depends on sequence beyond nearest-neighbors.

In order to examine sequence-dependent structural effects in DNA, the ability of alternating purine-pyrimidine fragments to undergo a B-Z transition when cloned in a supercoiled plasmid was determined solely as a function of sequence, with base and nearest-neighbor composition held constant. Sequences of 22 GC and 2 AT base pairs were synthesized such that the AT base pairs varied between contiguous placement and separation by eight GC base pairs. Results show, surprisingly, that the ease of the B-Z transition varies with the position of the two AT base pairs, occurring at lower superhelical densities when AT base pairs are contiguous, and at higher torsional strain when the AT base pairs are moved further apart.     

Mediation of the A/B-DNA helix transition by G-tracts in the crystal structure of duplex CATGGGCCCATG

The crystal structure of the DNA dodecamer duplex CATGGGCCCATG lies on a structural continuum along the transition between A- and B-DNA. The dodecamer possesses the normal vector plot and inclination values typical of B-DNA, but has the crystal packing, helical twist, groove width, sugar pucker, slide and x-displacement values typical of A-DNA. The structure shows highly ordered water structures, such as a double spine of water molecules against each side of the major groove, stabilizing the GC base pairs in an A-like conformation. The different hydration of GC and AT base pairs provides a physical basis for solvent-dependent facilitation of the A↔B helix transition by GC base pairs. Crystal structures of CATGGGCCCATG and other A/B-DNA intermediates support a ‘slide first, roll later’ mechanism for the B→A helix transition. In the distribution of helical parameters in protein–DNA crystal structures, GpG base steps show A-like properties, reflecting their innate predisposition for the A conformation.     

Proton nuclear magnetic resonance investigations of the nucleation and propagation reactions associated with the helix-coil transition of d-ApTpGpCpApT in H2O solution.

The chemical shifts and line widths of the Watson-Crick ring NH resonances of the self-complementary duplex of d-ApTpGpCpApT have been monitored in the presence of 0.1 M phosphate at neutral pH in aqueous solution. While the resonance positions of the terminal and internal AT base pairs shift upfield and broaden as average resonances with increasing temperature (helix and coil exchange several times prior to exchange with water from the coil form), those of the central GC base pairs broaden in the absence of upfield shifts (exchange with water occurs each time helix converts to coil). The line-width changes at the AT base pairs monitor the lifetime of the coil state at these positions prior to exchange with water while the line-width changes at the GC base pairs monitor the lifetime of the helix prior to dissociation to strands. This permits the separation of the propagation reaction at the AT base pairs from the nucleation reaction at the GC base pairs during helix formation. The experimental data have been quantitatively analyzed to yield (at 20 degrees) a nucleation formation rate of approximately 10(3) sec-1 for the GC base pairs (bimolecular rate constant of approximately 6 times 10(6) l. mol-1 sec-1) and a dissociation rate of 6 times 10(2) sec-1 at these same base pairs (unimolecular dissociation to strands). The unimolecular propagation reactions at the terminal and terminal base pairs are associated with reaction rates greater than 10(4) sec-1. These values are consistent with a slow formation of a stable nucleus at the GC base pairs followed by a rapid propagation reaction at the AT base pairs. The line width of the (GC) central base pairs in the presence of phosphate is a direct measure of the lifetime of the total helix and yields an activation energy of 45 kcal for helix to coil conversion measured over a narrow temperature range. The exchange from the coil form with water is catalyzed by 0.1 M phosphate with a rate constant kHPO2-/4 = 0.2 times 10(6) 1. mol-1 sec-1.     

Molecular nature of the mutations in gene ADE2 in Saccharomyces cerevisiae yeasts. III. Determination of the correlation of the GC AT pairs in genes ADE1 and ADE2

Lethal and mutagenic effects and the nature of mutations induced by decay of 32P incorporated into yeast cell DNA as 32P-deoxyguanosine monophosphate (32PdGMP) and 32P-thymidine monophosphate (32P-TMP), were studied. The lethal efficiency per 32P decay is independent of a labelled nucleotide incorporated into DNA. However, the mutagenic efficiency in ADE1, ADE2 genes per 32P decay is approximately 3 times greater for 32PdGMP than for 32P-TMP. This suggests that ADE1, ADE2 genes contain about 3 times more GC base pairs than AT pairs. Variations in a relative frequencies of GC leads to AT and AT leads to GC transitions were obtained depending upon a nucleotide labelled.     

Changes of hydration during conformational transitions of DNA

Fiber X-ray diffraction and measurement of fiber dimensions yields information about the hydration of DNA in fibers. The results obtained give us the fraction of nucleotides in the B form for the A-B transition or the rate of progression for the B-C transition as functions of the number of water molecules per nucleotide. The present experimental results confirm the importance of cooperativity in the A-B transition and the progressive change of the DNA double helix conformation during the C-B transition. At least twenty additional water molecules per nucleotide are necessary to stabilize the B form for DNA molecules in fibers following the A to B transition whereas only ten are sufficient when the B conformation is obtained starting from the C form. Offprint requests to: S. Premilat     

Monte Carlo Simulation of Hydration of the Nucleic Acid Fragments

Abstract

Systems containing a base or a base pair and 25 water molecules, as well as a helical stack and 30 water molecules per base pair, have been simulated. Changes in the base hydration shell structure, after the bases have been included into the pair and then into the base pair stack are discussed. Hydration shells of several configurations of the base pair stacks are discussed. Probabilities of formation of the hydrogen-bonded bridges of 1, 2 and 3 water molecules between hydrophilic centres have been estimated. The hydration shell structure was shown to depend on the nature of the base pair and on the stack configuration, while dependence of the global hydration shell characteristics on the stack configuration has been proved to be rather slight. The most typical structural elements of hydration shells, in the glycosidic (minor in B-like conformation) and non-glycosidic (major) grooves, for different configurations of AU and GC stacks, have been found and discussed. The number of hydrogen bonds between water molecules and bases per water molecule was shown to change upon transformation of the stack from A to B configuration. This result is discussed in connection with the reasons for B to A conformational transition and the concept of “water economy”. Hydration shell patterns of NH₂-groups of AU and GC helical stacks differ significantly.     

Structure of DNA hydration shells studied by Raman spectroscopy

We have used Raman scattering to study the water O-H stretching modes at approximately 3450 and approximately 3220 cm-1 in DNA films as a function of relative humidity (r.h.). The intensity of the 3220-cm-1 band vanishes as the r.h. is decreased from 98% to around 80%, which indicates that the hydrogen-bond network of water is disrupted in the primary hydration shell (which therefore cannot have an "ice-like" structure). The number of water molecules in the primary hydration shell was determined from the intensity of the approximately 3200-cm-1 band as about 30 water molecules per nucleotide pair. The approximately 3400-cm-1 O-H stretch band was used for determining the total water content, and this band persists at 0% r.h., implying that 5-6 tightly bound water molecules per nucleotide pair remain. The frequency of the approximately 3400-cm-1 O-H stretch mode is lower by 30 to 45 cm-1 in the primary hydration shell compared to free water. The water content as a function of r.h. obtained from these experiments agrees with gravimetric measurements. The disappearance of the approximately 3200-cm-1 band and the shift of the approximately 3400-cm-1 O-H stretch band provide a reliable way of measuring the hydration number of DNA.     

Thermodynamics of DNA containing very stable chemically modified base pairs

DNA chemical modifications caused by the binding of some antitumor drugs give rise to a very strong local stabilization of the double helix. These sites melt at a temperature that is well above the melting temperatures of ordinary AT and GC base pairs. In this work we have examined the melting behavior of DNA containing very stable sites. Analytical expressions were derived and used to evaluate the thermodynamic properties of homopolymer DNA with several different distributions of stable sites. The results were extended to DNA with a heterogeneous sequence of AT and GC base pairs. The results were compared to the melting properties of DNA with ordinary covalent interstrand cross-links. It was found that, as with an ordinary interstrand cross-link, a single strongly stabilized site makes a DNA's melting temperature (T(m)) independent of strand concentration. However in contrast to a DNA with an interstrand cross-link, a strongly stabilized site makes the DNA's T(m) independent of DNA length and equal to T(infinity), the melting temperature of an infinite length DNA with the same GC-content and without a stabilized site. Moreover, at a temperature where more than 80% of base pairs are melted, the number of ordinary (non-modified) helical base pairs (n) is independent of both the DNA length and the location of the stabilized sites. For this condition, n(T) = (2 omega-a)S/(1-S) and S = exp[DeltaS(T(infinity)-T)/(RT)] where omega is the number of strongly stabilized sites in the DNA chain, a is the number of DNA ends that contain a stabilized site, and DeltaS, T, and R are the base pair entropy change, the temperature, and the universal gas constant per mole. The above expression is valid for a temperature interval that corresponds to n<0.2N for omega=1, and n<0.1N for omega>1, where N is the number of ordinary base pairs in the DNA chain.