Endo Sugar Conformation Research Articles

Conformational flexibility in RNA: the role of dihydrouridine.

In order to further understand the structural role of the modified nucleoside dihydrouridine in RNA the solution conformations of Dp and ApDpA were analyzed by one- and two-dimensional proton NRM spectroscopy and compared with those of the related uridine-containing compounds. The analyses indicate that dihydrouridine significantly destabilizes the C3'-endo sugar conformation associated with base stacked, ordered, A-type helical RNA. Equilibrium constants (Keq = [C2'-endo]/[C3'-endo]) for C2'-endo-C3'-endo interconversion at 25 degrees C for Dp, the 5'-terminal A of ApDpA and D in ApDpA are 2.08, 1.35 and 10.8 respectively. Stabilization of the C2'-endo form was shown to be enhanced at low temperature, indicating that C2'-endo is the thermodynamically favored conformation for dihydrouridine. DeltaH values show that for Dp the C2'-endo sugar conformation is stabilized by 1.5 kcal/mol compared with Up. This effect is amplified for D in the oligonucleotide ApDpA and propagated to the 5'-neighboring A, with stabilization of the C2'-endo form by 5.3 kcal/mol for D and 3.6 kcal/mol for the 5'-terminal A. Post-transcriptional formation of dihydrouridine therefore represents a biological strategy opposite in effect to ribose methylation, 2-thiolation or pseudouridylation, all of which enhance regional stability through stabilization of the C3'-endo conformer. Dihydrouridine effectively promotes the C2'-endo sugar conformation, allowing for greater conformational flexibility and dynamic motion in regions of RNA where tertiary interactions and loop formation must be simultaneously accommodated.

Structure of a DNA Duplex Containing a Single 2‘-O-Methyl-β-d-araT: Combined Use of NMR, Restrained Molecular Dynamics, and Full Relaxation Matrix Refinement

Two-dimensional 1H NMR spectroscopy was used to determine the solution structure of the double-stranded DNA oligonucleotide d(5'-CGCATATAGCC-3'): d(5'-GGCTAXATGCG-3'), where X is 1-(2-O-methyl-beta-D-arabinofuranosyl)thymine. The structure determination was based on a total relaxation matrix analysis of NOESY cross-peak intensities using the MARDIGRAS program. The improved RANDMARDI procedure was used during the calculations to include the experimental "noise" in the NOESY spectra. The NOE-derived distance restraints were applied in restrained molecular dynamics calculations. Twenty final structures each were generated for the modified DNA duplex from both A-form and B-form DNA starting structures. The root-mean-square deviation of the coordinates for the 40 structures was 0.82 A. The duplex adopts a normal B-DNA-type helix, and the spectra as well as the structure show that the modified nucleotide X adopts a C2'-endo (S) sugar conformation. There are no significant changes in the helix originating from the modified nucleotide. The CH3O group on X is directed toward the major groove, and there seems to be free space for further modifications at this position.

DNA duplexes flanked by hybrid duplexes: the solution structure of chimeric junctions in [r(cgcg)d(TATACGCG)]2.

Hybrid duplexes and chimeric duplexes containing hybrid segments linked to pure DNA (or pure RNA) segments are involved in transcription and replication, as well as reverse transcription. A complete understanding of the mechanism of these processes requires detailed information on such duplexes and the junctions between duplexes of differing structure. Using two-dimensional NMR, restrained molecular dynamics and mechanics, and back-calculation refinement against the nuclear Overhauser effect spectra at various mixing times, we have determined the solution structure of the chimeric duplex [r(cgcg)d-(TATACGCG)]2 containing a pure DNA segment in the center of a hybrid duplex. The solution structure differs from the previously determined X-ray structure of the analogous duplex [r(gcg)d(TATACGC)]2, which was found to be A-form throughout [Wang, A.H.-J., et al. (1982) Nature 299, 601-604]. The basic features of the solution structure are (a) the RNA residues are all A-form with C3'-endo sugar conformations, (b) the central DNA segment is B-form, (c) the transition from A-form RNA sugar conformations to B-form DNA sugar conformations involves only the dT5 base step, and (d) although the sugar conformations of the DNA residues A6-G12 are closer to B-form, the basic helical properties of the peripheral RNA.DNA hybrid segments are closer to typical A-form than to B-form.

Parallel-Stranded DNA with Mixed Sequence. Evidence for Conformational Transition in Solution at Low Water Activity

Parallel-stranded deoxyoligonucleotide 5'd(CTATAGGGAT)3'/5'd(GATATCCCTA)3' (I-II) was shown to be stable in solution at 3 degrees-5 degrees C, 0.1-0.25 M NaCl, 10(-2) M phosphate buffer, pH 7.0 by means of a set of fluorescent techniques as well as of conventional optical methods. A cooperative change in the CD spectra is observed in trifluoroethanol (TFE) solutions at decreased water activity (relative humidity, r.h.). This distinctive change is supposed to stem from a cooperative conformational transition of parallel double helix from a B-like form with C2' endo sugar conformation to an A-like form designated as Ap. The free energy difference between the Ap and B-like conformation for the parallel duplex is 7.35 kcal/mol which is close to the value 7.40 kcal/mol for the antiparallel 5'd(CTATAGGGAT)3'/3'd(GATATCCCTA)5' (I-III). The ability of parallel helix to transit into Ap form is important for DNA-RNA parallel double helix formation.

Synthesis of 3′-fluoromethylthio-, 3′-fluoromethylsulfinyl- and 3′-fluoromethylsulfonyl-substituted 3′-deoxythymidine

3′- Deoxy- 3′-(fluoromethylthio)thymidine 2 is synthesized from 5′-O-benzoyl-3′-deoxy-3′-(methylsulfinyl)thymidine 9 by using DAST. 5′-O- Benzoyl-3′-deoxy-3′-(fluoromethylthio)thymidine is used as starting material for the synthesis of the sulfoxide 3 and the sulfone 4. Both sulfoxides 3A and 3B show a C3′-endo sugar conformation.

High-resolution NMR studies of chimeric DNA-RNA-DNA duplexes, heteronomous base pairing, and continuous base stacking at junctions.

Two symmetrical DNA-RNA-DNA duplex chimeras, d(CGCG)r(AAUU)d(CGCG) (designated rAAUU) and d(CGCG)r(UAUA)d(CGCG) (designated rUAUA), and a nonsymmetrical chimeric duplex, d(CGTT)r(AUAA)d(TGCG)/d(CGCA)r(UUAU)d(A ACG) (designated rAUAA), as well as their pure DNA analogues, containing dU instead of T, have been synthesized by solid-phase phosphoramidite methods and studied by high-resolution NMR techniques. The 1D imino proton NOE spectra of these d-r-d chimeras indicate normal Watson-Crick hydrogen bonding and base stacking at the junction region. Preliminary qualitative NOESY, COSY, and chemical shift data suggest that the internal RNA segment contains C3'-endo (A-type) sugar conformations except for the first RNA residues (position 5 and 17) following the 3' end of the DNA block, which, unlike the other six ribonucleotides, exhibit detectable H1'-H2' J coupling. The nucleosides of the two flanking DNA segments appear to adopt a fairly normal C2'-endo B-DNA conformation except at the junction with the RNA blocks (residues 4 and 16), where the last DNA residue appears to adopt an intermediate sugar conformation. The DNA-RNA junction residues exhibit quite different COSY, chemical shift, and NOE behavior, but these effects do not appear to propagate into the DNA or RNA segments. The circular dichroism spectra of these d-r-d chimeras also display a mixture of characteristic A-type and B-type absorption bands. The data indicate that A-type and B-type conformations can coexist in a single short continuous nucleic acid duplex, but our results differ somewhat from previous theoretical model studies.

A solid-state deuterium NMR investigation of conformation and order in magnetically oriented [d(CGCGAATTCGCG)

Solid-state deuterium NMR has been used to investigate the oriented liquid crystal phase of the hydrated oligonucleotide [d(CGCGAATTCGCG)]2. Previous investigations have shown that the helix axis is aligned perpendicular to the magnetic field, while at reduced temperatures motion about the helix axis is eliminated. Synthetic oligonucleotide samples incorporating different labeled nucleosides, [2"-2H]-2'-deoxyadenosine, [methyl-2H]thymidine, [8-2H]-2'-deoxyguanosine, and [8-2H]-2'-deoxyadenosine, permitted investigation of both the base and sugar conformation and ordering in the aligned phase. From line-shape analysis of the purine-labeled samples the orientation of the base C8 C-D bond with respect to the helix axis was determined to be theta = 90 degrees with a distribution of sigma total = 20 degrees, which is comparable to the orientation of theta = 90 degrees, sigma total = 15 degrees, with an oriented fraction P = 0.7 found for the C3 symmetry axis of the methyl-labeled dodecamer. The orientation of the sugar 2" C-D bond with respect to the helix axis is theta = 22 degrees with a distribution of sigma total = 15 degrees in agreement with the expected C2'-endo sugar conformation. The fraction of C3'-endo was also investigated and from analysis of line shape cannot exceed 20%. These results, though preliminary in nature, illustrate the application of this aligned phase for structural investigations.

Structure calculations for single-stranded DNA complexed with the single-stranded DNA binding protein GP32 of bacteriophage T4: a remarkable DNA structure

In this study it is established by calculation which regular conformations single-stranded DNA and RNA can adopt in the complex with the single-stranded DNA binding protein GP32 of bacteriophage T4. In order to do so, information from previous experiments about base orientations and the length and diameter of the complexes is used together with knowledge about bond lengths and valence angles between chemical bonds. It turns out that there is only a limited set of similar conformations which are in agreement with experimental data. The arrangement of neighboring bases is such that there is ample space for aromatic residues of the protein to partly intercalate between the bases, which is in agreement with a previously proposed model for the binding domain of the protein [Prigodich, R. V., Shamoo, Y., Williams, K. R., Chase, J. W., Konigsberg, W. H., & Coleman, J. E. (1986) Biochemistry 25, 3666-3671]. Both C2'endo and C3'endo sugar conformations lead to calculated DNA conformations that are consistent with experimental data. The orientation of the O2' atoms of the sugars in RNA can explain why the binding affinity of GP32 for polyribonucleotides is lower than for polydeoxyribonucleotides.

Conformational variations in d(TGACGTCA) and its reverse sequence d(ACTGCAGT): A joint circular dichroism and nuclear magnetic resonance study

Circular dichroism (CD) and nuclear magnetic resonance (NMR) techniques have been used to characterize the structural properties of the two self-complementary DNA octamers d(TGACGTCA) (I) and d(ACTGCAGT) (II). These display as distinctive features reverse sequences and central steps CpG and GpC, respectively. CD experiments lead to B-form DNA spectra characterized by larger magnitude signals in the case of octamer (I). NMR COSY spectra indicate that in the two octamers all the residues are predominantly south, S, (2'-endo) sugar conformation. NMR NOESY spectra show most of the glycosidic angles confined in the range predicted for B-form DNA although important heterogeneity is noticed along the chains, more pronounced in the case of octamer (I). Both the increase of north, N, (3'-endo) sugar conformation and P (pseudorotation phase angle) deviation from its standard B-form DNA value (162 degrees) express local sequence dependent structure distortions, remarkably visible in CpG step of octamer (I) and agreeing with NOESY cross-peaks intensities. Results interpreted according to Calladine's rules indicate higher cross-chain strains in octamer (I) than in octamer (II). All together, we find evidence to support for octamer (I) in solution local structures with A-DNA properties likely dictated by the central CpG step. Such structures could be involved in the DNA recognition by proteins and anticancerous drugs.

NMR studies of triple-strand formation from the homopurine-homopyrimidine deoxyribonucleotides d(GA)4 and d(TC)4.

The complexes formed by the homopurine and homopyrimidine deoxyribonucleotides d(GA)4 and d(TC)4 have been investigated by one- and two-dimensional 1H NMR. Under appropriate conditions [low pH, excess d(TC)4 strand] the oligonucleotides form a triplex containing one d(GA)4 and two d(TC)4 strands. The homopurine and one of the homopyrimidine strands are Watson-Crick base paired, and the second homopyrimidine strand is Hoogsteen base paired in the major groove to the d(GA)4 strand. Hoogsteen base pairing in GC base pairs requires hemiprotonation of C; we report direct observation of the C+ imino proton in these base pairs. Both homopyrimidine strands have C3'-endo sugar conformations, but the purine strand does not. The major triplex formed appears to have four TAT and three CGC+ triplets formed by binding of the second d(TC)4 strand parallel to the d(GA)4 strand with a 3' dangling end. In addition to the triplexes formed, at least one other heterocomplex is observed under some conditions.

Conformational properties of poly[d(G-T)].poly[d(C-A)] and poly[d(A-T)] in low- and high-salt solutions: NMR and laser Raman analysis.

Abstract31P‐ and 1H‐nmr and laser Raman spectra have been obtained for poly[d(G‐T)]·[d(C‐A)] and poly[d(A‐T)] as a function of both temperature and salt. The 31P spectrum of poly[d(G‐T)]·[d(C‐A)] appears as a quadruplet whose resonances undergo separation upon addition of CsCl to 5.5M. 1H‐nmr measurements are assigned and reported as a function of temperature and CsCl concentration. One dimensional nuclear Overhauser effect (NOE) difference spectra are also reported for poly[d(G‐T)]·[d(C‐A)] at low salt. NOE enhancements between the H8 protons of the purines and the C5 protons of the pyrimidines, (H and CH3) and between the base and H‐2′,2″ protons indicate a right‐handed B‐DNA conformation for this polymer. The NOE patterns for the TH3 and GH1 protons in H2O indicate a Watson–Crick hydrogen‐bonding scheme. At high CsCl concentrations there are upfield shifts for selected sugar protons and the AH2 proton. In addition, laser Raman spectra for poly[d(A‐T)] and poly[d(G‐T)]·[d(C‐A)] indicate B‐type conformations in low and high CsCl, with predominantly C2′‐endo sugar conformations for both polymers. Also, changes in base‐ring vibrations indicate that Cs+ binds to O2 of thymine and possibly N3 of adenine in poly[d(G‐T)]·[d(C‐A)] but not in poly[d(A‐T)]. Further, 1H measurements are reported for poly[d(A‐T)] as a function of temperature in high CsCl concentrations. On going to high CsCl there are selective upfield shifts, with the most dramatic being observed for TH1′. At high temperature some of the protons undergo severe changes in linewidths. Those protons that undergo the largest upfield shifts also undergo the most dramatic changes in linewidths. In particular TH1′, TCH3, AH1′, AH2, and TH6 all undergo large changes in linewidths, whereas AH8 and all the H‐2′,2″ protons remain essentially constant. The maximum linewidth occurs at the same temperature for all protons (65°C). This transition does not occur for d(G‐T)·d(C‐A) at 65°C or at any other temperature studied. These changes are cooperative in nature and can be rationalized as a temperature‐induced equilibrium between bound and unbound Cs+, with duplex and single‐stranded DNA. NOE measurements for poly[d(A‐T)] indicate that at high Cs+ the polymer is in a right‐handed B‐conformation. Assignments and NOE effects for the low‐salt 1H spectra of poly[d(A‐T)] agree with those of Assa‐Munt and Kearns [(1984) Biochemistry 23, 791–796] and provide a basis for analysis of the high Cs+ spectra. These results indicate that both polymers adopt a B‐type conformation in both low and high salt. However, a significant variation is the ability of the phosphate backbone to adopt a repeat dependent upon the base sequence. This feature is common to poly[d(G‐T)]·[d(C‐A)], poly[d(A‐T)], and some other pyr–pur polymers [J. S. Cohen, J. B. Wouten & C. L Chatterjee (1981) Biochemistry 20, 3049–3055] but not poly[d(G‐C)].

Conformational studies of nucleic acids. II. The conformational energetics of commonly occurring nucleosides.

We have examined the conformational energetics of the eight most commonly occurring nucleosides--A, U, G, C, dA, dT, dG, dC--as monitored by a semi-empirical energy force field. These are the first reported calculations to completely explore the entire conformational spaces available to all eight major nucleosides using experimentally consistent furanose geometries and an appropriate force field. Central to our approach is the ability to model an experimentally reasonable furanose for each nucleoside directly from only one parameter, the phase angle of pseudorotation P, as described in the previous paper (D.A. Pearlman, and S.-H. Kim, preceeding paper in this issue). This allows us to specify the conformation of a nucleoside by three variables: torsion angle gamma (O5'-C5'-C4'-C3'); torsion angle chi (O4'-C1'-N9/N1-C4/C2); and P. In our study each of these parameters was allowed to vary independently and in small increments over the range 0-360 degrees. The empirically observed preferences for C3'-endo and C2'-endo sugar conformations, for anti and syn values of chi and for staggered (g+, t, g-) values of gamma can be explained on the basis of the energy maps so obtained. Finer details, such as the different conformational preferences of ribonucleosides and deoxyribonucleosides and of purines and pyrimidines, can also be extracted from these maps and are consistent with experiment. The calculations support previous descriptions of pseudorotation as hindered. Statistical Boltzmann population factors for different conformational ranges in gamma, chi, and P, as predicted by the calculations, are consistent with factors obtained from crystallographic data. The excellent results here provide additional support for the suitability of the new sugar modeling technique used.

Molecular orbital studies on nucleoside antibiotics, VII. Conformation of 2′-amino- 2′-deoxyguanosine

Conformational properties of the nucleoside antibiotic 2′-amino-2′-deoxyguanosine have been investigated by the PCILO method along with those of its parent nucleoside, guanosine. This antibiotic, formed by replacement of the 2′-hydroxyl group by an amino group in guanosine, shows anti-tumor activity and also inhibits RNA and protein syntheses. Both C(2′)-endo and C(3′)-endo sugar conformations have been considered in the computations. The results indicate striking similarity between the conformations of the antibiotic and the parent nucleoside, particularly in simulated aqueous environment. The biological implication of this result in terms of the antibiotic activity is discussed.

Visualization of drug-nucleic acid interactions at atomic resolution: IV. Structure of an aminoacridine-dinucleoside monophosphate crystalline complex, 9-aminoacridine-5-Iodocytidylyl (3′–5′) guanosine

9-Aminoacridine forms a crystalline complex with the dinucleoside monophosphate, 5-iodocytidylyl(3′–5′)guanosine (iodoCpG). These crystals are monoclinic, space group P2 1 with a = 13.98 A ̊ , b = 30.58 A ̊ , c = 22.47 A ̊ and β = 113.9 °. The structure has been solved to atomic resolution by Patterson and Fourier methods, and refined by a combination of Fourier and sum-function Fourier methods. The asymmetric unit contains four 9-aminoacridine molecules, four iodoCpG molecules and 21 water molecules, a total of 245 atoms. 9-Aminoacridine demonstrates two different intercalative binding modes and, along with these, two slightly different intercalative geometries in this model system. The first of these is very nearly symmetric, the 9-amino group lying in the narrow groove of the intercalated base-paired nucleotide structure. The second shows grossly asymmetric binding to the dinucleotide, the 9-amino group lying in the wide groove of the structure. Associated with these two different intercalative binding modes is a difference in geometries in the structures. Although both structures demonstrate C3′ endo (3′–5′) C2′ endo mixed sugar puckering patterns (i.e. both cytidine residues have C3′ endo sugar conformations, while both guanosine residues have C2′ endo sugar conformations), with corresponding twist angles between base-pairs of about 10 °, they differ in the magnitude of the helical screw axis dislocation accompanying intercalation (Sobell et al., 1977 a,b). In the pseudosymmetric intercalative structure, this value is about +0.5 Å, whereas in the asymmetric intercalative structure this value is about +2.7 Å. These conformational differences can be best described as a “sliding” of base-pairs on the intercalated acridine molecule. Although the pseudosymmetric intercalative structure can be used in 9-aminoacridine-DNA binding, the asymmetric intercalative structure cannot since this poses stereochemical difficulties in connecting neighboring sugar-phosphate chains to the intercalated dinucleotide. It is possible, however, that the asymmetric binding mode is related to the mechanism of 9-aminoacridine-induced frameshift mutagenesis (Sakore et al., 1977), and we discuss this possibility here in further detail.

Visualization of drug-nucleic acid interactions at atomic resolution: V. Structure of two aminoacridine-dinucleoside monophosphate crystalline complexes, proflavine-5-iodocytidylyl (3′–5′) guanosine and acridine orange-5-iodocytidylyl (3′–5′) guanosine

Acridine orange and proflavine form complexes with the dinucleoside monophosphate, 5-iodocytidylyl(3′–5′)guanosine. The acridine orange-iodoCpG † † Abbreviation used: iodoCpG, 5-iodocytidylyl(3′–5′)guanosine. crystals are monoclinic, space group P2 1, with unit cell dimensions a = 14.36 A ̊ , b = 19.64 A ̊ , c = 20.67 A ̊ , β = 102.5 °. The proflavine-iodoCpG crystals are monoclinic, space group C2, with unit cell dimensions a = 32.14 A ̊ , b = 22.23 A ̊ , c = 18.42 A ̊ , β = 123.3 °. Both structures have been solved to atomic resolution by Patterson and Fourier methods, and refined by full matrix least-squares. Acridine orange forms an intercalative structure with iodoCpG in much the same manner as ethidium, ellipticine and 3,5,6,8-tetramethyl- N-methyl phenanthrolinium ( Jain et al., 1977,1979 ), except that the acridine nucleus lies asymmetrically in the intercalation site. This asymmetric intercalation is accompanied by a sliding of base-pairs upon the acridine nucleus and is similar to that observed with the 9-aminoacridine-iodoCpG asymmetric intercalative binding mode described in the previous papers ( Sakore et al., 1977,1979 ). Basepairs above and below the drug are separated by about 6.8 Å and are twisted about 10 °; this reflects the mixed sugar puckering pattern observed in the sugar-phospate chains: C3′ endo (3′–5′) C2′ endo (i.e. each cytidine residue has a C3′ endo sugar comformation, while each guanosine residue has a C2′ endo sugar conformation), alterations in glycosidic torsional angles and other small but significant conformational changes in the sugar-phosphate backbone. Proflavine, on the other hand, demonstrates symmetric intercalation with iodoCpG. Hydrogen bonds connect amino groups on proflavine with phosphate oxygen atoms on the dinucleotide. In contrast to the acridine orange structure, base-pairs above and below the intercalative proflavine molecule are twisted about 36 °. The altered magnitude of this angular twist reflects the sugar puckering pattern that is observed: C3′ endo (3′–5′) C3′ endo. Since proflavine is known to unwind DNA in much the same manner as ethidium and acridine orange ( Waring, 1970), one cannot use the information from this model system to understand how proflavine binds to DNA (it is possible, for example, that hydrogen bonding observed between proflavine and iodoCpG alters the intercalative geometry in this model system). Instead, we propose a model for proflavine-DNA binding in which proflavine lies asymmetrically in the intercalation site (characterized by the C3′ endo (3′–5′) C2′ endo mixed sugar puckering pattern) and forms only one hydrogen bond to a neighboring phosphate oxygen atom. Our model for proflavine-DNA binding, therefore, is very similar to our acridine orange-DNA binding model. We will describe these models in detail in this paper.

Visualization of drug-nucleic acid interactions at atomic resolution: I. Structure of an ethidium/dinucleoside monophosphate crystalline complex, ethidium:5-Iodouridylyl (3′–5′) adenosine

Ethidium forms a crystalline complex with the dinucleoside monophosphate 5-iodouridylyl(3′–5′)adenosine (iodoUpA). These crystals are monoclinic, space group C2, with unit cell dimensions, a = 28.45 A ̊ , b = 13.54 A ̊ , c = 34.13 A ̊ , β = 98.6 ° . The structure has been solved to atomic resolution by Patterson and Fourier methods, and refined by full matrix least-squares to a residual of 0.20 on 2017 observed reflections. The asymmetric unit contains two ethidium molecules, two iodoUpA molecules and 27 water molecules, a total of 155 atoms excluding hydrogens. The two iodoUpA molecules are held together by adenine · uracil Watson-Crick-type base-pairing. Adjacent base-pairs within this paired iodoUpA structure and between neighboring iodoUpA molecules in adjoining unit cells are separated by about 6.7 Å; this separation results from intercalative binding by one ethidium molecule and stacking by the other ethidium molecule above and below the base-pairs. Non-crystallographic 2-fold symmetry is utilized in this model drug-nucleic acid interaction, the intercalated ethidium molecule being oriented such that its phenyl and ethyl groups lie in the narrow groove of the miniature nucleic acid double-helix. Base-pairs within the paired nucleotide units are related by a twist of 8 °. The magnitude of this angular twist is related to conformational changes in the sugar-phosphate chains that accompany drug intercalation. These changes partly reflect the differences in ribose sugar ring puckering that are observed (both iodouridine residues have C3′ endo sugar conformations, whereas both adenosine residues have C2′ endo sugar conformations), and alterations in the glycosidic torsional angles describing the base-sugar orientations. Additional small but systematic changes occur in torsional angles that involve the phosphodiester linkages and the C4′C5′ bond. Solution studies have indicated a marked sequence-specific binding preference in ethidium-dinucleotide interactions, and a probable structural explanation for this is provided by this study. This structure and the accompanying one described in the second paper [ethidium:5-idocytidylyl(3′–5′)guanosine] are examples of model drug-nucleic acid intercalative complexes, and the information provided by their structure analyses has led to a general understanding of intercalative drug binding to DNA. This is described in the third paper of this series.

Visualization of drug-nucleic acid interactions at atomic resolution: II. Structure of an ethidium/dinucleoside monophosphate crystalline complex, ethidium:5-iodocytidylyl (3′–5′) guanosine

Ethidium forms a second crystalline complex with the dinucleoside monophosphate 5-iodocytidylyl(3′–5′)guanosine (iodoCpG). These crystals are monoclinic, P2 1, with a = 14.06 A ̊ , b = 32.34 A ̊ , c = 16.53 A ̊ , β = 117.8 ° . The structure has been solved to atomic resolution using rigid-body Patterson vector search and Fourier methods, and refined by full matrix least-squares to a residual of 0.16 on 3180 observed reflections. The structure consists of two ethidium molecules, two iodoCpG molecules, 27 water molecules and four methanol molecules, a total of 165 atoms (excluding hydrogens) in the asymmetric unit. Both iodoCpG molecules are hydrogen-bonded together by guanine · cytosine Watson-Crick base-pairing. Adjacent base-pairs within this paired iodoCpG structure and between neighboring iodoCpG molecules in adjoining unit cells are separated by 6.7 Å. This distance reflects the presence of an ethidium molecule intercalated between base-paired iodoCpG molecules and another ethidium molecule stacked above (and below) the dinucleotide. Approximate 2-fold symmetry is used in the interaction; this reflects the pseudo-2-fold symmetry axis of the phenanthridinium ring system in ethidium coinciding with the approximate 2-fold axis relating base-paired iodoCpG molecules. The phenyl and ethyl groups of the intercalated ethidium molecule lie in the narrow groove of the miniature iodoCpG double-helix. The stacked ethidium, however, lies in the opposite direction, its phenyl and ethyl groups neighboring iodine atoms on cytosine residues. Base-pairs within the paired nucleotide units are related by a twist of about 8 °. The magnitude of this angular twist reflects conformational changes in the sugar-phosphate chains accompanying intercalation. These primarily reflect the differences in ribose sugar ring puckering that are observed (i.e. both iodocytidine residues have C3′ endo sugar conformations, while both guanosine residues have C2′ endo sugar conformations), and alterations in the glycosidic torsional angles that describe the base-sugar orientation. The information provided by this structure analysis (along with the accompanying one (ethidium:iodoUpA), described in the previous paper) has led to an understanding of the general nature of intercalative drug binding to DNA. This is described in the third paper of this series.