Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (676K) Freely available
Right arrow Supplementary Data
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (11)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Aramini, J. M.
Right arrow Articles by Germann, M. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Aramini, J. M.
Right arrow Articles by Germann, M. W.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Nucleic Acids Research Pages 5644-5654  

NMR solution structures of [d(GCGAAT-3[prime]-3[prime]-[alpha]T-5[prime]-5[prime]-CGC)2] and its unmodified control
Introduction
Materials And Methods
   Materials and sample preparation
   NMR spectroscopy
   Duplex correlation times
   DQF-COSY simulations
   PSEUROT calculations
   Structural restraints
   Starting models
   Restrained molecular dynamics and energy minimization
   Structure determination strategy
   Structure analysis
Results And Discussion
   NMR Restraints
   Pseudorotation analysis
   NMR structure determination
   Analysis of the final structures
Conclusions
Acknowledgements
Supplementary Material
References

NMR solution structures of [d(GCGAAT-3[prime]-3[prime]-[alpha]T-5[prime]-5[prime]-CGC)<sub>2</sub>] and its unmodified control

NMR solution structures of [d(GCGAAT-3[prime]-3[prime]-[alpha]T-5[prime]-5[prime]-CGC)2] and its unmodified control

James M. Aramini, Anwer Mujeeb1 and Markus W. Germann*

Department of Microbiology and Immunology, Kimmel Cancer Institute, Thomas Jefferson University, Philadelphia, PA 19107, USA and 1Department of Pharmaceutical Chemistry, University of California, San Francisco, CA 94143, USA

Received August 21, 1998; Revised and Accepted October 14, 1998

PDB accession nos 1bwt and 1bx5

ABSTRACT

We present the high-resolution solution structures of a self-complementary DNA decamer duplex featuring a single [alpha]-anomeric nucleotide per strand encompassed by a set of 3[prime]-3[prime] and 5[prime]-5[prime] phosphodiester linkages, d(GCGAAT-3[prime]-3[prime]-[alpha]T-5[prime]-5[prime]-CGC)2, alphaT, and its unmodified control, d(GCGAATTCGC)2, obtained by restrained molecular dynamics. Interproton distance and deoxyribose ring torsion angle restraints were deduced from homonuclear NOESY and DQF-COSY data, respectively. For both the control and alphaT duplexes, excellent global convergence was observed from two different (A- and B-) starting models. The final average structures of the two duplexes are highly homologous, and overall possess the traits characteristic of right-handed B-DNA duplexes. However, localized differences between the two structures stem from the enhanced conformational exchange in the deoxyribose ring of the cytidine following the 5[prime]-5[prime] linkage, the C3[prime]-exo pseudorotation phase angle of the[alpha]-nucleotide, and unusual backbone torsions in the 3[prime]-3[prime] and 5[prime]-5[prime] phosphodiester linkages. The structural data reported here are relevant to the design of antisense therapeutics comprised of these modifications.

INTRODUCTION

Antisense therapy is a conceptually simple approach centered on the specific arrest of gene expression at the translational level by synthetic oligodeoxyribonucleotides (ODNs) (1,2). The development of ODNs with improved antisense properties in terms of nuclease resistance, cellular uptake, affinity and selectivity of binding to the target mRNA, and RNase H degradation of the mRNA has been vigorously pursued over the years, leading to the development of a tremendous number of modifications and substitutions to the chemical make-up of nucleic acids (3,4). In this vein, ODNs composed exclusively of [alpha]-anomeric nucleotides exhibit several desirable traits including (i) high nuclease resistance (5); (ii) stable and selective hybridization to complementary [beta]-DNA and [beta]-RNA sequences (6,7), featuring parallel-stranded duplex formation with Watson-Crick base pairing (8-10); and (iii) in vitro inhibition of mRNA translation and viral proliferation (11,12). However, [beta]-RNA strands hybridized to [alpha]-DNA are resistant to cleavage by RNase H (13). Recently, synthetic designs incorporating a combination of [alpha]- and [beta]-anomeric nucleotides and polarity reversals brought about by 3[prime]-3[prime] and 5[prime]-5[prime] phosphodiester linkages (14,15) have led to the development of [alpha]-[beta] chimeric ODNs capable of eliciting RNase H activity, while preserving the properties of [alpha]-DNA that are favorable from an antisense perspective (16,17). Of relevance to this design are the encouraging antisense and antitumor activities recently reported for synthetic anti-HPV type-16 (human papilloma virus) [alpha]-[beta] chimeric 16mers featuring stretches of four contiguous [beta]-nucleotides separated by [alpha]-anomers and polarity reversals in carcinoma cell lines and tumor-bearing mice in a putative RNase H-dependent manner (18).

The promising anticancer potential of nucleic acids containing both [alpha]-anomeric nucleotides and polarity reversals stresses the need to elucidate the structural features of such modifications, which can be ultimately incorporated into the rational design of improved antisense therapeutics. To this end, we have investigated the thermodynamic, enzymatic and spectroscopic properties of a family of self-complementary decamers in which a single [alpha]-anomeric T, C, A or G is inserted into the sequence with the opposite orientation via 3[prime]-3[prime] and 5[prime]-5[prime] phosphodiester linkages (19,20). In this paper we present the complete 3D solution structure determinations of the alphaT duplex and the unmodified control (Fig. 1) by restrained molecular dynamics using NMR restraints derived from relaxation matrix and pseudorotation analyses of 2D nuclear Overhauser (NOESY) and double quantum filtered correlation (DQF-COSY) spectroscopic data, respectively. Our results constitute the first look at the structural features of nucleic acids containing a combination of [alpha]- and [beta]-anomeric nucleotides linked together by 3[prime]-3[prime] and 5[prime]-5[prime] phosphodiester bonds at an atomic level, and the first complete pseudorotation analysis of [alpha]-anomeric nucleotides within an ODN.


Figure 1. Sequences of the control and alphaT decamer duplexes. Arrows indicate the strand polarity in the 3[prime]->5[prime] direction; the 3[prime]-3[prime] and 5[prime]-5[prime] phosphodiester linkages are denoted by tail-to-tail and head-to-head junctions, respectively. [alpha]-anomeric nucleotides are shown in outline type. Black arrows in the control duplex represent the EcoRI restriction enzyme cleavage sites. The sequence numbering, in the 5[prime]->3[prime] direction, is shown for the alphaT sequence.

MATERIALS AND METHODS

Materials and sample preparation

Detailed procedures for the synthesis and purification of the alphaT, d(GCGAAT-3[prime]-3[prime]-[alpha]T-5[prime]-5[prime]-CGC), and control, d(GCGAATTCGC), decamers are described in our earlier work (19). [alpha]-anomeric deoxyribonucleosides were purchased from Sigma Chemical Co. and R.I. Chemicals Inc., and used without further purification. NMR samples of the two duplexes (alphaT, 6.8 mM; control, 2.6 mM) and four nucleosides ([alpha]-d-dA: 75 mM, pD 6.9; [alpha]-d-dC: 51 mM, pD 7.7; [alpha]-d-dG: 25 mM, pD 9.4; [alpha]-d-T: 95 mM, pD 6.4) were prepared in buffered (50 mM NaCl, 10 mM Na2HPO4, 0.1 mM EDTA, pD 6.5) and unbuffered D2O, respectively.

NMR spectroscopy

All NMR experiments were performed at 303 K on a Bruker AMX 600 NMR spectrometer. The acquisition and processing parameters for the 2D experiments conducted on the control and alphaT duplexes are given in our earlier studies (19,20); those most relevant to this discussion are as follows: (i) NOESY, mixing times ([tau]m) of 125 and 250 ms for the control and 75, 150 and 300 ms for alphaT and relaxation delays of 4.0 (control) or 5.0 s (alphaT); (ii) DQF-COSY, strip transformed using 45° shifted sine-bell multiplication in both dimensions (0.56-0.59 Hz/point resolution); and (iii) 31P-1H correlation on alphaT (21), strip transformed using sine-bell multiplication for resolution enhancement (0.74 and 0.18 Hz/point resolution in [omega]2 and [omega]1, respectively). The deoxyribose ring spin-spin (J) coupling constants for the [alpha]-anomeric mononucleosides were determined by a series of 1D selective homonuclear decoupling experiments; 1H chemical shift assignments for these nucleosides were made on the basis of selective 1D-NOE difference spectroscopy and DQF-COSY. 1H and 31P chemical shift scales were calibrated with respect to DSS (2,2-dimethylsilapentane-5-sulphonate) and 85% H3PO4 (capillary), both in D2O and at 303 K.

Duplex correlation times

Overall isotropic molecular correlation times, [tau]c, required for the determination of interproton distances by the relaxation matrix approach, were estimated by the truncated driven 1D NOE method (22). Correlation times ranging from [ap]2.8 to 3.4 ns were obtained for the entire family of [alpha]-containing duplexes comprised of this sequence; such values are consistent with literature data for duplexes of this size (23).

DQF-COSY simulations

Deoxyribose spin-spin (J) coupling constants in the decamer duplexes were deduced by DQF-COSY cross-peak simulation using the programs SPHINX and LINSHA (24). Assuming a six component spin system (H1[prime], H2[prime], H2[prime][prime], H3[prime], H4[prime] and P) and strong coupling between only the geminal protons (H2[prime] and H2[prime][prime]), coupling constants and resonance linewidths were systematically varied to obtain the best match with the multiplet fine structure of DQF-COSY cross-peaks, particularly along [omega]2. All other variables, such as the digital resolution, acquisition time, and window functions, emulated the experimental conditions. Selected cross-peaks from the 31P-1H correlation spectrum of the alphaT duplex were simulated in an analogous fashion.

PSEUROT calculations

For the [alpha]-anomeric mononucleosides, pseudorotation parameters were computed using the PSEUROT 6.2 program (25). In each case, a total of 800 calculations per value of [Phi]N were performed in which PN is systematically varied (but kept fixed in each calculation along with [Phi]N) and [Phi]S, PS and fS optimized as judged by the residual difference between the calculated and experimental coupling constants.

Structural restraints

In the restrained molecular dynamics (rMD) and energy minimization (rEM) protocols described below, all distance, bond angle and torsion angle restraints were introduced into the AMBER force field equations as pseudo-energy constraint terms, in which individual restraints are described by a flat-well potential of the form equation 1:

1

where k is the restraint force constant and r is the model distance or torsion/bond angle. The force constants and bounds (r1 to r4) for the restraints are dependent on their origin as described in the following sub-sections; the force constants employed are based on values optimized in earlier studies (26,27).

Quantitative distance restraints. Using cross-peak volumes and integration uncertainties obtained by the analysis package SPARKY 3.33 (UCSF, NMR laboratory) the RANDMARDI procedure (28), an extension of the complete relaxation matrix approach MARDIGRAS 3.2 (29,30), was used to calculate quantitative distance restraints for non-labile protons. In addition to accounting for spin diffusion at longer mixing times, RANDMARDI randomly varies the experimental cross-peak volumes on the basis of their uncertainties, yielding more accurate estimates of the distance bounds (r2 and r3). Distance bounds were obtained by performing 30 runs of RANDMARDI calculations on each NOESY data set, assuming correlation times of 2.0, 3.0 and 4.0 ns, isotropic molecular motion, and a 3-site jump model for methyl group rotation (31); the normalization factor for the input data was computed by summing over all volumes. Limits defining the distance restraints were derived from the average of the lower and upper bounds from all RANDMARDI runs, and widths of the parabolic portions of the potential were set to 2.0 Å (i.e. r1 = r2 - 2.0; r4 = r3 + 2.0). In general, force constants for the quantitative distance restraints were set to 30 kcal/(mol.Å2). For both duplexes, the force constants for [ap]10% of the quantitative restraints were reduced to 15 due to weak cross-peak intensity or partial overlap.

Qualitative distance restraints. In cases where cross-peak volumes could not be quantitated due to degeneracy, saturation effects or pre-saturation, qualitative distance bounds were assigned based on the intensity and nature of the cross-peak (i.e. strong, 2.0-3.5 Å; weak, 2.0-6.0 Å). Due to the characteristically long T1 relaxation times of the adenosine H2 protons (19), only an upper bound of 4.0 Å was used for the interresidue ds(2;2) contacts. Also, in the case of the control duplex, exchangeable proton distance restraints of 2.0-4.0 Å were included on the basis of 1-1 NOESY data in water (19). All qualitative distance restraints, which constititute [ap]10% of the restraint sets, were given a force constant of 30 kcal/(mol.Å2).

Deoxyribose ring torsion angle restraints. Restraints for the endocyclic deoxyribose ring torsion angles ([nu]0 to [nu]4) were based on the average value for each individual torsion angle, [nu]i, and its uncertainty, [Delta][nu]iave, which were derived from pseudorotation analysis of the vicinal 1H-1H coupling constants as follows:r1 = [nu]iave - [Delta][nu]iave; r2 = [nu]iave - [Delta][nu]iave/2; r3 = [nu]iave + [Delta][nu]iave/2 and r4 = [nu]iave + [Delta][nu]iave. All torsion angle restraints were assigned a force constant of 60 kcal/(mol.rad2).

Watson-Crick restraints. To maintain base pairing during the rMD protocol, the NMR restraints were augmented by Watson-Crick base pairing restraints (32) applied with a force constant of 25 kcal/(mol.Å2). In each case, r1 was defined as 0, and r4 was set 0.5 Å above the upper bound (r3), respectively. The [alpha]T-A base pair restraint was softened by the addition of 0.3 Å to r3 and r4. Weak bond angle restraints [k = 5-10 kcal/(mol.rad2)] ranging from 170 to 190°, with 20° parabolic windows on either side, were employed to encourage linear Watson-Crick hydrogen-bond formation.

Backbone torsion angle restraints. In the rMD runs, the following restraints [k = 5-10 kcal/(mol.rad2)] were applied to all backbone torsion angles with the exception of [delta] (which is related to the endocyclic torsion angle [nu]3) to preserve helical right-handedness during the high temperature phase of the simulated annealing protocol: [alpha] = -90 to -30°, [beta] = 170-230°, [gamma] = 20-90°, [epsis] = 140-220° and [zeta] = 250-330°; the parabolic regions extended 10° beyond these bounds. In the alphaT case, all backbone torsions involving the unusual 3[prime]-3[prime] and 5[prime]-5[prime] linkages were left unrestrained to avoid undue conformational bias.

Starting models

Canonical A- and B- right-handed duplex DNA models for the control sequence were generated by the NUCGEN subroutine within AMBER 4.1 using the 1994 parametrization of the all atom nucleic acid force field (33). To mimic the effects of counterions, the charges on the phosphate moieties were uniformly attenuated to give a net charge of -0.2/nucleotide (26). For the A- and B-form starting models of alphaT, the [alpha]-anomeric thymidine units were generated from the parameters of the [beta]-anomeric unit and incorporated into the appropriate sequence positions using LEaP (34). All four starting models were subjected to 1000 steps of steepest descent plus 4000 steps of conjugate gradient energy minimization in AMBER 4.1 using the protocols described below.

Restrained molecular dynamics and energy minimization

All in vacuo rMD calculations were performed using the SANDER program and the following parameters: a distance-dependent dielectric constant ([epsis] = r), 1 fs time steps, a 30 Å cut-off for non-bonded interactions, charge modification of the end hydrogens, the SHAKE algorithm (35) to fix all bond lengths, removal of translational and rotational motions every 100 fs, non-bonded pair list updates every 25 fs, and tight coupling of the entire system to a temperature bath. Analogous conditions were employed in rEM calculations except that no SHAKE was applied. Typically, 25 ps rMD runs were performed containing a high temperature simulated annealing period to overcome high-energy barriers in which the temperature and restraint force constant weights were linearly varied as follows: 0-2 ps: T = 6 -> 900 K, k = 0.1 -> 5; 2-8 ps: T = 900 K, k = 5; 8-12 ps: T = 900 -> 300 K, k = 5 -> 1; 12-25 ps: T = 300 K, k = 1. After each rMD run coordinates from the final 10 or 20 snap shots (i.e. 2-4 ps) were averaged and restrained energy minimized for 200 steps of steepest descent followed by 1800 steps conjugate gradient.

Structure determination strategy

We employed similar strategies to elucidate the NMR structures for both the control and alphaT duplexes. After the first RANDMARDI cycle to extract distance restraints, four 25 ps rMD runs were performed from A- and B-starting models using different random seed numbers. The ensemble of eight structures was averaged and restrained energy minimized, yielding an average structure for the first cycle. This structure was then used as the new model for a second cycle of RANDMARDI calculations to refine the quantitative distance restraints. A new cycle of three 25 ps rMD calculations using the updated restraints was performed against each starting model (i.e. A, B and the previous average structure). A new structure was thus obtained by superimposing the coordinates from the resulting ensemble followed by 2000 steps of rEM. This protocol was repeated until there was no further improvement in the quality of the final structure based on the R factors obtained from CORMA calculations (vide infra). For both duplexes, all nine structures from the final rMD runs constitute the `final ensembles' presented.

R factor calculations

To assess the quality of structures over the course of the refinement process, crystallographic, R (equation 2), and sixth-root, Rx (equation 3), factors were computed from the observed (ao) and calculated (ac) cross-peak volumes for each NOESY data set assuming a [tau]c of 3 ns using the program CORMA version 5.2 (36,37).

2
3

Structure analysis

All models and restraints were displayed using the MidasPlus 2.1 graphics program (38). Global helical parameters for the final average structures of the control and alphaT duplexes were obtained using the CURVES 5.1 algorithm (39).

RESULTS AND DISCUSSION

NMR Restraints

Using the complete relaxation matrix approach RANDMARDI (28) to determine intra- and interresidue quantitative distance restraints, we observed excellent initial convergence from the A and B starting models, and tight well-widths in each cycle of the refinement for both the control and alphaT duplexes. Moreover, in both cases the entire molecule is well constrained, with an average of [ap]24-27 experimental restraints (i.e. NOE-derived distance plus J-based deoxyribose torsion angle restraints) per nucleotide. The resolved diastereotopic H5[prime] and H5[prime][prime] protons of the [alpha]-nucleotide, stereospecifically assigned on the basis of intraresidue NOE and DQF-COSY data (40), generate substantially more restraints and provide a thorough probe of the 5[prime]-5[prime] phosphodiester linkage. The statistical data for all final restraints used in the two structure determinations are listed in Table 1.


Table 1. Statistical data for the final restraints used in the NMR structure determination of the control and alphaT duplexesa
aThe data given correspond to the total number of restraints per duplex. The origin of the various restraints as well as their respective force constants are described in Materials and Methods. The backbone torsion angles were applied only in rMD calculations.
bAverage well-widths for the quantitative (RANDMARDI-derived) distance restraints.

Pseudorotation analysis

Following the approach of Schmitz and James (23), the programs SPHINX and LINSHA (24) were used to simulate the following six DQF-COSY multiplets of all nucleotides except the 3[prime]-terminus (C10) in the control and alphaT duplexes: H1[prime]([omega]1)-H2[prime]/H2[prime][prime]([omega]2), H2[prime]/H2[prime][prime]([omega]1)-H1[prime]([omega]2) and H3[prime]([omega]1)-H2[prime]/H2[prime][prime]([omega]2). Figure 2 shows the experimental and simulated DQF-COSY cross-peaks for two key residues in alphaT, namely [alpha]T7 and its neighbor C8, along with C8 in the control for comparison. By matching the simulated to experimental multiplet fine structure for each cross-peak, we obtained four vicinal 1H-1H coupling constants (J1[prime]2[prime], J1[prime]2[prime][prime], J2[prime]3[prime] and J2[prime][prime]3[prime]) required for the pseudorotation analysis, along with the two-bond geminal coupling between the diastereotopic H2[prime] and H2[prime][prime] (J2[prime]2[prime][prime]); these are given in Table 2. In most cases only an upper bound for the J2[prime][prime]3[prime] can be estimated due to the complete absence of this cross-peak.


Table 2. Coupling constant and pseudorotation data for the control, alphaT and [alpha]-nucleosidesa
aPseudorotation analyses for the control and alphaT duplexes were performed graphically using contour plots of the individual coupling constants as a function of fS and PS at various [Phi]S values assuming the following: [beta]-anomeric sugars, PN = 18°; [Phi]N = [Phi]S; [alpha]T7, PN = 0°; [Phi]N = 37°. For G1, T6 and G9 in alphaT, the uncertainties in J1[prime]2[prime] and J1[prime]2[prime][prime] were increased to ±0.7 and ±0.5 Hz, respectively, due to significant cross-peak overlap (G1 and G9) or signal broadening (T6). For the [alpha]-anomeric mononucleosides, all J values given are ±0.1 Hz and the pseudorotation results were computed with the PSEUROT 6.2 program.


Figure 2. (Left) Expansions (60 Hz in [omega]1 × 40 Hz in [omega]2) of six DQF-COSY cross-peaks (white boxes) for the [alpha]T7 and C8 residues in the alphaT duplex, as well as C8 for the control, and their corresponding SPHINX/LINSHA simulations (blue boxes). Negative contours are shown in red, positive contours in black. The following coupling constants and resonance line widths (both in Hz) produced the best fits (best estimates for the J3[prime]4[prime] and J3[prime]P, which only affect the [omega]1 axes of the two cross-peaks involving H3[prime] and were not extensively varied in the simulations, are given in parentheses): [alpha]T7 of alphaT: J1[prime]2[prime] = 8.3, J1[prime]2[prime][prime] = 3.0, J2[prime]3[prime] = 6.0, J2[prime][prime]3[prime] = 2.0, J2[prime]2[prime][prime] = -17, (J3[prime]4[prime] = 2.0, J3[prime]P = 8.0), [Delta][nu]1/2 (H1[prime]) = 6.0, [Delta][nu]1/2 (H2[prime]) = 8.0, [Delta][nu]1/2 (H2[prime][prime]) = 10.0, [Delta][nu]1/2 (H3[prime]) = 5.0; C8 of alphaT: J1[prime]2[prime] = 6.0, J1[prime]2[prime][prime] = 6.5, J2[prime]3[prime] = 6.3, J2[prime][prime]3[prime] = 5.3,J2[prime]2[prime][prime] = -14, (J3[prime]4[prime] = 4.5, J3[prime]P = 5.0), [Delta][nu]1/2 (H1[prime]) = 5.5, [Delta][nu]1/2 (H2[prime]) = 7.5, [Delta][nu]1/2 (H2[prime][prime]) = 7.0, [Delta][nu]1/2 (H3[prime]) = 5.5; C8 of control: J1[prime]2[prime] = 9.1, J1[prime]2[prime][prime] = 5.9, J2[prime]3[prime] = 6.2,J2[prime][prime]3[prime] = 3.0, J2[prime]2[prime][prime] = -14, (J3[prime]4[prime] = 2.5, J3[prime]P = 5.0), [Delta][nu]1/2 (H1[prime]) = 5.5, [Delta][nu]1/2 (H2[prime]) = 8.0, [Delta][nu]1/2 (H2[prime][prime]) = 7.0, [Delta][nu]1/2 (H3[prime]) = 5.5. (Right) Contour plots of J1[prime]2[prime] (black),J1[prime]2[prime][prime] (red), J2[prime]3[prime] (blue) and J2[prime][prime]3[prime] (green) as a function of fs and Ps for the values of puckering amplitude, [Phi]S and [Phi]N (PN was fixed), which give in the best convergence (region in black) of the coupling constant data: [alpha]T7 of alphaT: [Phi]S = 33°, [Phi]N = 37° (PN = 0°); C8 of alphaT: [Phi]S = [Phi]N = 37° (PN = 18°); C8 of control: [Phi]S = [Phi]N = 37° (PN = 18°). The upper and lower bounds used for each coupling constant correspond to the experimental uncertainties listed in Table 2.

Using the methodologies developed over the years by the Altona laboratory (reviewed in 41), the coupling constant data obtained in this manner can then be used to deduce the pseudorotation parameters which describe the conformational and dynamic properties of each deoxyribose ring, namely the puckering amplitude, [Phi]m, phase angle of pseudorotation, P, and mole fraction of S-pucker, fS, assuming the sugar ring undergoes a rapid two-state equilibrium bewteen N- and S-conformations. The pseudorotation theory, equations and coefficients pertinent to this study are collected in Appendix I (25,41-43).

A series of contour plots were generated for the four principle coupling constants (J1[prime]2[prime], J1[prime]2[prime][prime], J2[prime]3[prime] and J2[prime][prime]3[prime]) as a function of fS and PS for a range of [Phi]S values, while [Phi]N and PN were kept fixed. For all [beta]-anomeric sugars we assumed [Phi]N = [Phi]S and PN = 18° (41) and found the best fit of the data in the range, 35 [le] [Phi] [le] 38°. On the basis of the results we obtained for [alpha]-anomeric mononucleosides (Table 2) as well as literature data (44), [Phi]N and PN for [alpha]T7 were set to 37 and 0°, respectively. The best fits of the coupling constant data for [alpha]T7 and C8 in alphaT and C8 in the control are shown in Figure 2, and the complete set of pseudorotation parameters for all nucleotides examined in this fashion are listed in Table 2. All [beta]-anomeric sugars fall into the C1[prime]-exo to C2[prime]-endo range in the pseudorotation cycle, and, with the exception of C8 in alphaT, exhibit a high propensity for the S-conformation. Our PS and fS values for these residues are characteristic of deoxyribose ring puckering in B-DNA (41), and are in excellent agreement with several pseudorotation studies of DNA duplexes based on spectral simulations and/or sums of couplings (45-48). However, in the case of C8, which is linked to the [alpha]-anomeric nucleotide by a 5[prime]-5[prime] phosphodiester bond, the coupling constant data describes an approximately equal distribution over the N- and S-puckers. An unusually prominent H3[prime]-H4[prime] DQF-COSY cross-peak observed for C8 adds further credence to this notion, since the value of J3[prime]4[prime] is sensitive to this equilibrium. In accordance with the behavior of the free [alpha]-anomeric nucleosides (Table 2; 49), we observe that [alpha]T7 in alphaT is also highly S-like in character and with a reduced pucker amplitude. However, the phase angle of the S-conformer is much higher, falling in the C3[prime]-exo window. In addition, the geminal coupling constant, J2[prime]2[prime][prime], for [alpha]T7 is unusually large, an observation we have also made for other [alpha]-containing decamers of this series (J.M.Aramini and M.W.Germann, unpublished results).

The pseudorotation parameters and their uncertainties obtained with the graphical approach described here were converted into the final endocyclic torsion angle ([nu]0 to [nu]4) restraints used in the structure determinations as described in Appendix I. We chose to use torsion angle restraints rather than directly employing the J values as restraints in our structure calculations, in order to make use of the improved Donders formalism of the Karplus equation, as opposed to the generic expression built into SANDER; in addition, this approach ensures the proper treatment of the couplings involving H1[prime] in the inverted [alpha]-anomeric nucleotide.

NMR structure determination

In the structure determinations for both the control and alphaT duplexes, we obtained excellent convergence from different starting models within four RANDMARDI/rMD cycles, yielding two tight ensembles of structures with average all atom pairwise root-mean-square deviations (RMSDs) of 0.85Å (control) and 0.76 Å (alphaT). A number of criteria are indicative of the high quality of the final energy minimized average structures derived from these ensembles. The final control and alphaT structures exhibit low Amber energies (EAMBER) of -968.7 and -869.2 kcal/mol, respectively, as well as low distance (ENOE) and torsion angle (Etors) restraint violation energies compared with the starting models (Table 3). Average restraint violations, expressed as the average difference between a given distance ([Delta]dav) or torsion angle ([Delta][nu]av) and the nearest upper or lower bound (r2 or r3), also drop precipitously from their values in the A- and B-models, particularly the former (Table 3). The crystallographic and sixth-root R factors calculated for each NOESY data set using CORMA are very low for the final structures, and constitute a vast improvement compared with those obtained for the starting models (Table 4); the final Rx factors, ranging from 3.7 to 4.2%, are among the lowest observed to date using these methodologies.


Table 3. Energies and restraint violations for the starting models and final NMR structures of the control and alphaT duplexesa
aAll values were obtained from one-step rEM computations using the final restraints (no backbone restraints were included); [Delta]dav and [Delta][nu]av correspond to the average distance and torsion angle deviations from the upper (if r > r3) or lower (if r < r2) bounds.


Figure 3. Stereoview of the superimposed final structures of the control (grey) and alphaT (blue) duplexes; the [alpha]-anomeric thymidine and subsequent cytidine in both strands of the alphaT duplex are depicted in red and green, respectively. In matching the coordinates of the two structures, the terminal and [alpha]-anomeric residues were disregarded.

On the basis of pairwise RMSDs between the final structures and their respective starting models (Table 5), both the control and alphaT duplexes essentially fall into the category of B-DNA. Moreover, the overall structures are highly homologous, as suggested by an all atom RMSD of [ap]1.0 Å between them, excluding the terminal base pairs and the two [alpha]-nucleotides (Fig. 3). In the alphaT duplex, the orientation of the [alpha]-anomeric nucleotide in each strand is reversed, yet base pairing and stacking is preserved.


Table 4. Crystallographic (R) and sixth-root (Rx) R factors for the starting models and final NMR structures of the control and alphaT duplexesa
aFor each NOE data set, the contributions to the total crystallographic and sixth-root R factors (in bold) from intra- and inter-residue NOEs are listed in separate columns.

Finally, in order to place these structures in the context of published coordinates of DNA duplexes containing this sequence, we compared the final average structure for the control with the refined crystal structures of the Dickerson dodecamer [d(CGCGAATTCGCG)2] available in the Nucleic Acid Database (50) at 1.9 Å [accession nos: NDB, BDL020; PDB, 9BNA (51)] and 1.4 Å [accession nos: NDB, BDL084; PDB, 355D (52)] resolution. Considering only heavy atoms, we obtained RMSD values of [ap]0.9 Å for the central 8 bp of the control and either structure (the pairwise RMSD of the crystal structures is 0.7 Å). This is indicative of a global accordance between our structures and the literature data for this well-studied dodecamer.


Table 5. Pairwise all atom root-mean-square-deviations (Å) between initial models and final average structures of the control and alphaT duplexesa
aThe all atom RMSD of the initial A- and B-models for the control duplex prior to energy minimization was 5.66 Å.

Analysis of the final structures

The global helical parameters obtained using CURVES 5.1 show a high degree of similarity between control and alphaT; in the interest of brevity, plots of these data are not included here. Instead, we wish to focus on the differences in the structural data between [alpha]-containing and unmodified duplexes, specifically in P, [chi] and the backbone torsion angles, [alpha]-[zeta] (Fig. 4), and illustrate that the unusual moieties result in perturbations that are highly localized in nature.


Figure 4. Plots of P, [chi] and backbone torsion angles for all residues in the final structures of the control ([open circle]) and alphaT ([closed square]) duplexes.

Deoxyribose ring conformation. With the exception of the 3[prime]-terminal cytidine, essentially all [beta]-anomeric nucleotides reside in the S-conformation, ranging in pseudorotation phase angle from C1[prime]-exo to C2[prime]-endo, as predicted by pseudorotation analysis. Hence, the restraints that were derived from sources that exhibit a unique dependence on conformational averaging (NOE versus J-data) converge to the same result if the equilibrium is favored by one conformer; this has been repeatedly shown for high S sugars (>75%) in DNA duplexes (26,27,45). However, the sugar puckering of two pairs of residues in the alphaT duplex warrant special mention. First, the [alpha]-anomeric thymidines are both shifted to a C3[prime]-exo (P [ap] 207°) conformation, consistent with the J-coupling data. Very similar pseudorotation phase angles were reported for the [alpha]-anomeric nucleotides in the NMR structures of two parallel-stranded [alpha]-DNA/[beta]-DNA duplexes (8,9). Second, in each strand the deoxyribose ring of the cytidine linked to the [alpha]T via a 5[prime]-5[prime] phosphodiester bond falls in the S-hemisphere of the pseudorotation cycle, although the phase angle is below that predicted by the J-coupling data (Table 2). Close inspection of the distance restraints obtained for this nucleotide using RANDMARDI reveals that key intraresidue and sequential interproton contacts that are highly dependent on the sugar ring conformation, such as di(6,8;2[prime]), di(6,8;2[prime][prime]), ds(2[prime];6,8) and ds(2[prime][prime];6,8) (53), are consistent with either an A-form, a B-form, or are intermediate between the two motifs (data not shown). This dichotomy cannot be simultaneously fulfilled in the standard rMD/rEM calculations without increased energy penalties and restraint violations. Lifting these restraints leads to a sharp reduction in energy violations. Moreover, when only the deoxyribose restraints of the C8 and C18 nucleotides are considered, we observe increases of a similar magnitude in CORMA Rx factors when these residues are constrained to either the N- or S-sugar puckers, further substantiating the lack of preference for either deoxyribose conformation in the NOE data.

Glycosidic torsion angle, [chi]. The glycosidic torsion angle, [chi], which reflects the relative orientation of the base and sugar moieties, is anti, specifically (-)anticlinical, for all [beta]-anomeric nucleotides in both duplexes ([chi] [ap] -100 to -140°). For the two [alpha]-nucleotides this torsion angle is also anti, but shifted to (+)anticlinical ([chi] [ap] 140°). However, the origin of this difference lies in the change in chirality at the C1[prime] position, and the resulting 120° shift in base position with respect to the sugar ring (Fig. 5A). To equally compare [chi] with [beta]-anomeric nucleotides one could measure the torsion angle using C2[prime]; this results in a correction of [ap] -240°, and a glycosidic torsion angle, [chi]a [ap] -100°, which is, thus, conformationally analogous to that for [beta]-anomeric nucleotides (8).


Figure 5. Portions of the final alphaT structure relevant to the definitions of [chi] and the unusual backbone torsion angles in the 3[prime]-3[prime] and 5[prime]-5[prime] phosphodiester linkages. Color scheme: C, grey; N, blue; O, red; P, magenta; for clarity all hydrogen atoms were removed. (A) View looking down the N1-C1[prime] bond axis of the T6 and [alpha]T7 residues from the final alphaT structure; the `formal' [chi] torsion angles are shown. To avoid confusion, only the C2 and O2 of the pyrimidine base are included. (B) The C3[prime] (A5) to C5[prime] (C8) segment of alphaT encompassing a normal 3[prime]-5[prime] phosphodiester bond and the 3[prime]-3[prime] and 5[prime]-5[prime] linkages.

Backbone torsion angles. The majority of the structural perturbations due to the single [alpha]-anomeric nucleotide and two polarity reversals per strand are absorbed by the phosphodiester backbone (Fig. 6). Due to the polarity reversals, the atoms involved in defining four backbone torsion angles in the 3[prime]-3[prime] and 5[prime]-5[prime] linkages, namely two [alpha] and two [zeta] torsions, are altered, and we introduce the following definitions (Fig. 5B): [zeta](a-1) = C3[prime]([alpha]-1) - O3[prime]([alpha]-1) - P([alpha]) - O3[prime]([alpha]); [zeta](a) = O3[prime]([alpha]-1) - P([alpha]) - O3[prime]([alpha]) - C3[prime]([alpha]); [alpha](a) = C5[prime]([alpha]) - O5[prime]([alpha]) - P([alpha]+1) - O5[prime]([alpha]+1); [alpha](a+1) = O5[prime]([alpha]) - P([alpha]+1) - O5[prime]([alpha]+1) - C5[prime]([alpha]+1), where ([alpha]-1) and ([alpha]+1) represent the nucleotides preceding and following the [alpha]-nucleotide, respectively, in the 5[prime]->3[prime] direction. Three significant backbone changes are evident for the two [alpha]-anomeric thymidines; specifically, [alpha](a) is shifted from g- to g+, a g+ to trans shift for [gamma], and [epsis] moves from [epsis]t to [epsis]-. Homo- (1H-1H) and heteronuclear coupling (i.e. 1H-31P) information can be used to independently validate the values of the [beta], [gamma] and [epsis] backbone torsion angles found for the [alpha]-nucleotide in the final structure (reviewed in 40).


Figure 6. (A) Close-up of the cross-peak between H3[prime] of T6 and P of [alpha]T7 from the resolution-enhanced 31P-1H correlation spectrum of alphaT. Positive and negative contours are in black and red, respectively. The passive coupling in the [omega]1 dimension corresponds to JH3[prime]([alpha]T7)-P([alpha]T7). (B) Dependence of JH3[prime]-P on the torsion angle [epsis] based on the recently revised Karplus relation: JH3[prime]-P = 15.3cos2([thetas]) - 6.2cos([thetas]) + 1.5, and [epsis] = -[thetas] - 120° (55). One possible solution for this torsion angle on the basis of the coupling in (A) is shown. (C) Experimental (white boxes) and SPHINX/LINSHA simulated (blue boxes) H5[prime]/H5[prime][prime]-31P cross-peaks (30 Hz in [omega]1 × 50 Hz in [omega]2) from the resolution-enhanced 31P-1H correlation spectrum of alphaT. The coupling constants and linewidths used to generate the simulations are as follows: J5[prime]5[prime][prime] = -11.5, J4[prime]5[prime][prime] = 6.1, J5[prime][prime]P = 5.8, J4[prime]5[prime] = 2.0, J5[prime]P = 4.0; [Delta][nu]1/2 (H5[prime]) = [Delta][nu]1/2 (H5[prime][prime]) = 6.0, [Delta][nu]1/2 (31P) = 9.0; an additional generic JH-P of 4 Hz was included to account for small couplings between P and H4[prime]/H5[prime]/H5[prime][prime] of C8 in the less important [omega]1 dimension. The heteronuclear couplings extracted from (A) and (C) are generally consistent with values deduced from the appropriate multiplets in DQF-COSY spectra acquired with and without 31P decoupling (data not shown).

Epsilon torsion angle. Insight into [epsis] of the [alpha]-nucleotide can be obtained from fine structure of the cross-peak between the phosphorus atom of [alpha]T7 and the H3[prime] of T6 from the resolution enhanced 1H-31P correlation spectrum of alphaT (Fig. 6A). Note the large passive coupling in the 31P ([omega]1) dimension corresponding to the JH3[prime]-P between H3[prime] of [alpha]T7 and its phosphorus atom; we have obtained similar values for the analogous couplings in related [alpha]-containing decamers using a selective heteronuclear experiment (54) at reduced field strength (J.M.Aramini and M.W.Germann, unpublished results). This coupling constant is directly related to [epsis], by the Karplus plot shown in Figure 6B (55). One possible solution of [epsis] for such a large JH3[prime]-P coupling is the g- conformation found in the final average structure of alphaT.

Beta and gamma torsion angles. Analysis of the heteronuclear correlation between the phosphorus atom in the 5[prime]-5[prime] junction (i.e. P of C8) and the stereospecifically assigned H5[prime]/H5[prime][prime] of [alpha]T7 (Fig. 6C) sheds light on the conformational dynamics about the [beta] and [gamma], and torsions in the [alpha]-nucleotide. Simulation of the heteronuclear multiplets along [omega]2 (1H) yields JH5[prime][prime]-P, J4[prime]5[prime][prime] and JH5[prime]-P, J4[prime]5[prime], respectively, assuming a geminal coupling, J5[prime]5[prime][prime], of -11.5 Hz (56). With respect to [gamma], for a two-state conformational equilibrium, a J4[prime]5[prime][prime] of 6.1 Hz corresponds to an ~50:50 population of [gamma]- and [gamma] t conformers. For [beta], the heteronuclear couplings (JH5[prime][prime]-P = 5.8 Hz; JH5[prime]-P = 4.0 Hz) are slightly larger than those commonly observed in regular B-DNA, yet remain consistent with a large preference for the trans rotamer [f([beta]t) [ap] 80%; 40], and the slightly larger value of [beta] for the [alpha]-nucleotides in the final average structure of this duplex fall well within the region predicted by the Karplus relation.

Alpha and zeta torsion angles. While independent measurements of [alpha] and [zeta] are inaccessible by NMR, the 31P chemical shift of the phosphodiester group in nucleic acids is well-known to be highly dependent on these torsion angles, as well as several other factors (57). Whereas the normal BI conformation is characterized by an [epsis]t, [zeta]-, [alpha]-, [beta]t configuration, the equivalent torsion angles in the unusual 3[prime]-3[prime] and 5[prime]-5[prime] phosphodiester linkages are [epsis](a-1)t, [zeta](a-1)-, [zeta]a-/t, [epsis]a- and [beta]at, [alpha]a+, [alpha](a-1)-/t, [beta](a+1)t, respectively. Thus, the altered [alpha] and [zeta] values observed for the [alpha]-nucleotides may be at least partially responsible for the significant upfield([Delta][delta] = -1.4 p.p.m.) and downfield ([Delta][delta] = +1.4 p.p.m.) shifts, respectively, for the 3[prime]-3[prime] and 5[prime]-5[prime] phosphodiester linkages of alphaT and related [alpha]-containing decamers (20).

CONCLUSIONS

We have elucidated the high-resolution NMR solution structures of a self-complementary decamer duplex containing one [alpha]-anomeric nucleotide per strand inserted into the sequence in the opposite orientation via 3[prime]-3[prime] and 5[prime]-5[prime] phosphodiester linkages and its unmodified control using a combination of relaxation matrix methods, pseudorotation analysis, and restrained molecular dynamics. Our results conclusively demonstrate that the presence of these unnatural moieties in the alphaT duplex results in local structural variations within the segment encompassing the [alpha]-nucleotide and polarity reversals. These include perturbations to the phosphodiester backbone and an increased pseudorotation phase angle (C3[prime]-exo) for the [alpha]-anomeric sugar compared with free nucleosides. In addition, the conformational equilibrium of the deoxyribose ring in the nucleotide following the 5[prime]-5[prime] linkage (C8) is disrupted, resulting in enhanced sugar puckering dynamics enabling this sugar to adopt both N- and S-puckers with approx-imately equal probability. Overall, however, the [alpha]-containing duplex exhibits minimal global deviations from the B-DNA motif of the control sequence. The dynamic `hot spot' at C8 is quite interesting on a number of levels. First, for duplex DNA it has been shown that structural perturbations arising from conformational variations at a single nucleotide, such as the C3[prime]-endo [harr] C2[prime]-endo equilibrium, are highly localized in nature (58); this is exactly the case for the alphaT duplex, which further exemplifies DNA's remarkable plasticity. Next, all of the results presented in this article pertain to structures determined with the empirical (NMR) restraints enforced at all times. Clearly, the local conformational heterogeneity at C8 cannot be treated in this manner. In such cases one must resort to the use of other refinement techniques such as time-averaged rMD (59) and PARSE (probability assessment via relaxation rates of a structural ensemble; 60); these methods have been successfully applied to the investigation of deoxyribose repuckering events in nucleic acids (60,61). In practice, refined average structures, such as those reported here, are a necessary prerequisite for these procedures (23). The alphaT example accentuates the importance of conducting pseudorotation analysis a priori as a means of establishing the presence of deoxyribose ring conformational averaging in the molecule. Finally, decreases in the fraction of S-pucker have been reported in recent NMR (61-64) and molecular dynamics (65) studies of DNA-RNA hybrids. Moreover, modifications that promote N-type sugar puckering, along with other factors, have been correlated with increased thermostability of the ODN-RNA hybrids (66). Hence, this phenomenon is potentially important from the standpoint of the hybridization of chimeric [alpha],[beta]-ODNs with biologically relevant mRNA targets (18), a crucial step in the mechanism of these prospective antisense drugs.

ACKNOWLEDGEMENTS

We thank Dr R. T. Pon (University of Calgary) for providing us with the ODNs used in this study. We are also indebted to Prof. C. Altona (University of Leiden) for helpful insights pertaining to the pseudorotation analysis presented here, and Dr T. Goddard (UCSF) for developing the front-end conversion program `bruk2ucsf' for SPARKY. This work was supported by a grant from the National Institutes of Health GM OD55404-01.

SUPPLEMENTARY MATERIAL

I.
Plots of RANDMARDI distances derived from the A- versus B-starting models for both the control and alphaT.
II.
Plots of the number of quantitative RANDMARDI-derived intraresidue and sequential (5[prime]->3[prime]) distance restraints per residue for the control and alphaT.
III.
Table of torsion angle restraints used in structure refinements of the control and alphaT.
IV.
Final ensembles of structures for the control and alphaT duplexes.
V.
Plots of the global helical parameters of the control and alphaT final average structures obtained by CURVES 5.1.
VI.
Appendix I: pseudorotation theory and torsion angle restraints.

See supplementary material available in NAR Online.

REFERENCES

1. Agrawal,S. (1996) Trends Biotechnol., 14, 376-387. MEDLINE Abstract

2. Matteucci,M.D. and Wagner,R.W. (1996) Nature, 384, S20-S22. MEDLINE Abstract

3. De Mesmaeker,A., Häner,R., Martin,P. and Moser,H.E. (1995) Acc. Chem. Res., 28, 366-374.

4. Matteucci,M. (1997) In Chadwick,D.J. and Cardew,G. (eds), CIBA Foundation Symposium 209: Oligonucleotides as Therapeutic Agents. J. Wiley & Sons, NY, pp. 5-14.

5. Vichier-Guerre,S., Pompon,A., Lefebvre,I. and Imbach,J.-L. (1994) Antisense Res. Dev., 4, 9-18. MEDLINE Abstract

6. Thuong,N.T., Asseline,U., Roig,V., Takasugi,M. and Hélène,C. (1987) Proc. Natl Acad. Sci. USA, 84, 5129-5133. MEDLINE Abstract

7. Paoletti,J., Bazile,D., Morvan,F., Imbach,J.-L. and Paoletti,C. (1989) Nucleic Acids Res., 17, 2693-2704. MEDLINE Abstract

8. Lancelot,G., Guesnet,J.-L. and Vovelle,F. (1989) Biochemistry, 28, 7871-7878. MEDLINE Abstract

9. Guesnet,J.-L., Vovelle,F., Thuong,N.T. and Lancelot,G. (1990) Biochemistry, 29, 4982-4991. MEDLINE Abstract

10. Gmeiner,W.H., Rayner,B., Morvan,F., Imbach,J.-L. and Lown,J.W. (1992) J. Biomol. NMR, 2, 275-288. MEDLINE Abstract

11. Bertrand,J.-R., Imbach,J.-L., Paoletti,C. and Malvy,C. (1989) Biochem. Biophys. Res. Commun., 164, 311-318. MEDLINE Abstract

12. Lavignon,M., Tounekti,N., Rayner,B., Imbach,J.-L., Keith,G., Paoletti,J. and Malvy,C. (1992) Antisense Res. Dev., 2, 315-324. MEDLINE Abstract

13. Gagnor,C., Rayner,B., Leonetti,J.-P., Imbach,J.-L. and Lebleu,B. (1989) Nucleic Acids Res., 17, 5107-5114. MEDLINE Abstract

14. Debart,F., Tosquellas,G., Rayner,B. and Imbach,J.-L. (1994) Bioorg. Med. Chem. Lett., 4, 1041-1046.

15. Koga,M., Wilk,A., Moore,M.F., Scremin,C.L., Zhou,L. and Beaucage,S.L. (1995) J. Org. Chem., 60, 1520-1530.

16. Boiziau,C., Debart,F., Rayner,B., Imbach,J.-L. and Toulmé,J.-J. (1995) FEBS Lett., 361, 41-45. MEDLINE Abstract

17. Aramini,J.M., van de Sande,J.H. and Germann,M.W. (1998) In Leontis,N.B. and SantaLucia,J.,Jr (eds), ACS Symposium Series 682: Molecular Modeling of Nucleic Acids. American Chemical Society, Washington, DC, pp. 92-105.

18. Tan,T.M.C., Kalisch,B.W., van de Sande,J.H., Ting,R.C.Y. and Tan,Y.H. (1998) Antisense Nucleic Acid Drug Dev., 8, 95-101.

19. Aramini,J.M., Kalisch,B.W., Pon,R.T., van de Sande,J.H. and Germann,M.W. (1996) Biochemistry, 35, 9355-9365. MEDLINE Abstract

20. Aramini,J.M., van de Sande,J.H. and Germann,M.W. (1997) Biochemistry, 36, 9715-9725. MEDLINE Abstract

21. Sklenár,V., Miyashiro,H., Zon,G., Miles,T. and Bax,A. (1986) FEBS Lett., 208, 94-98. MEDLINE Abstract

22. Lane,A.N. (1995) Methods Enzymol., 261, 413-435. MEDLINE Abstract

23. Schmitz,U. and James,T.L. (1995) Methods Enzymol., 261, 3-44. MEDLINE Abstract

24. Widmer,H. and Wüthrich,K. (1987) J. Magn. Reson., 74, 316-336.

25. van Wijk,J., Haasnoot,K., de Leeuw,F., Huckreide,D. and Altona,C. (1995) PSEUROT 6.2. A Program for the Conformational Analysis of Five Membered Rings. University of Leiden, The Netherlands.

26. Mujeeb,A., Kerwin,S.M., Kenyon,G.L. and James,T.L. (1993) Biochemistry, 32, 13419-13431. MEDLINE Abstract

27. Weisz,K., Shafer,R.H., Egan,W. and James,T.L. (1994) Biochemistry, 33, 354-366. MEDLINE Abstract

28. Liu,H., Spielmann,H.P., Ulyanov,N.B., Wemmer,D.E. and James,T.L. (1995) J. Biomol. NMR, 6, 390-402. MEDLINE Abstract

29. Borgias,B.A. and James,T.L. (1989) Methods Enzymol., 176, 169-183. MEDLINE Abstract

30. Borgias,B.A. and James,T.L. (1990) J. Magn. Reson., 87, 475-487.

31. Liu,H., Thomas,P.D. and James,T.L. (1992) J. Magn. Reson., 98, 163-175.

32. Saenger,W. (1984) Principles of Nucleic Acid Structure, Springer-Verlag, NY.

33. Pearlman,D.A., Case,D.A., Caldwell,J.W., Ross,W.S., Cheatham,T.E.,III, Ferguson,D.,M., Seibel,G.L., Singh,U.C., Weiner,P.K. and Kollman,P.A. (1995) Amber 4.1. University of California, San Francisco.

34. Schafmeister,C.E.A.F., Ross,W.S. and Romanovski,V. (1995) LEaP (Version 1.0), University of California, San Francisco.

35. Ryckaert,J.P., Ciccotti,G. and Berendsen,H.J.C. (1977) J. Comput. Phys., 23, 327-341.

36. Keepers,J.W. and James,T.L. (1984) J. Magn. Reson., 57, 404-426.

37. Borgias,B.A. and James,T.L. (1988) J. Magn. Reson., 79, 493-512.

38. Ferrin,T.E., Huang,C.C., Jarvis,L.E. and Langridge,R. (1988) J. Mol. Graphics, 6, 13-27.

39. Lavery,R. and Sklenar,H. (1996) Curves 5.1. Helical Analysis of Irregular Nucleic Acids. Laboratoire de Biochimie Theorique CNRS, Paris, France.

40. Wijmenga,S.S., Mooren,M.M.W. and Hilbers,C.W. (1993) In Roberts,G.C.K. (ed.), NMR of Macromolecules; A Practical Approach. Oxford University Press, NY, pp. 217-288.

41. van Wijk,J., Huckriede,B.D., Ippel,J.H. and Altona,C. (1992) Methods Enzymol., 211, 286-306. MEDLINE Abstract

42. Donders,L.A., de Leeuw,F.A.A.M. and Altona,C. (1989) Magn. Reson. Chem., 27, 556-563.

43. Altona,C., Francke,R., de Haan,R., Ippel,J.H., Daalmans,G.J., Westra Hoekzema,A.J.A. and van Wijk,J. (1994) Magn. Reson. Chem., 32, 670-678.

44. Latha,Y.S. and Yathindra,N. (1992) Biopolymers, 32, 249-269. MEDLINE Abstract

45. Schmitz,U., Sethson,I., Egan,W.M. and James,T.L. (1992) J. Mol. Biol., 227, 510-531. MEDLINE Abstract

46. Weisz,K., Shafer,R.H., Egan,W. and James,T.L. (1992) Biochemistry, 31, 7477-7487. MEDLINE Abstract

47. Mujeeb,A., Kerwin,S.M., Egan,W., Kenyon,G.L. and James,T.L. (1992) Biochemistry, 31, 9325-9338. MEDLINE Abstract

48. Conte,M.R., Bauer,C.J. and Lane,A.N. (1996) J. Biomol. NMR, 7, 190-206. MEDLINE Abstract

49. Thibaudeau,C., Földesi,A. and Chattopadhyaya,J. (1998) Tetrahedron, 54, 1867-1900.

50. Berman,H.L., Olson,W.K., Beveridge,D.L., Westbrook,J., Gelbin,A., Demeny,T., Hsieh,S.-H., Srinivasan,A.R. and Schneider,B. (1992) Biophys. J., 63, 751-759.

51. Westhof,E. (1987) J. Biomol. Struct. Dyn., 5, 581-600. MEDLINE Abstract

52. Shui,X., McFail-Isom,L., Hu,G.G. and Williams,L.D. (1998) Biochemistry, 37, 8341-8355. MEDLINE Abstract

53. Wüthrich,K. (1986) NMR of Proteins and Nucleic Acids, John Wiley & Sons, NY.

54. Sklenár,V. and Bax,A. (1987) J. Am. Chem. Soc., 109, 7525-7526.

55. Mooren,M.M.W., Wijmenga,S.S., van der Marel,G.A., van Boom,J.H. and Hilbers,C.W. (1994) Nucleic Acids Res., 22, 2658-2666.

56. Pikkemaat,J.A. and Altona,C. (1996) Magn. Reson. Chem., 34, S33-S39.

57. Gorenstein,D.G. (1994) Chem. Rev., 94, 1315-1338.

58. Lane,A.N. (1996) Magn. Reson. Chem., 34, S3-S10.

59. Torda,A.E., Scheek,R.M. and van Gunsteren,W.F. (1990) J. Mol. Biol., 214, 223-235. MEDLINE Abstract

60. Ulyanov,N.B., Schmitz,U., Kumar,A. and James,T.L. (1995) Biophys. J., 68, 13-24. MEDLINE Abstract

61. González,C., Stec,W., Reynolds,M.A. and James,T.L. (1995) Biochemistry, 34, 4969-4982. MEDLINE Abstract

62. González,C., Stec,W., Kobylanska,A., Hogrefe,R.I., Reynolds,M. and James,T.L. (1994) Biochemistry, 33, 11062-11072. MEDLINE Abstract

63. Gyi,J.I., Conn,G.L., Lane,A.N. and Brown,T. (1996) Biochemistry, 35, 12538-12548. MEDLINE Abstract

64. Gyi,J.I., Lane,A.N., Conn,G.L. and Brown,T. (1998) Biochemistry, 37, 73-80. MEDLINE Abstract

65. Cheatham,T.E.,III and Kollman,P.A. (1997) J. Am. Chem. Soc., 119, 4805-4825.

66. Freier,S.M. and Altmann,K.-H. (1997) Nucleic Acids Res., 25, 4429-4443. MEDLINE Abstract

*To whom correspondence should be addressed. Tel: +1 215 503 4581; Fax: +1 215 923 2117; Email: mwg@lac.jci.tju.edu

This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 3 Dec 1998
Copyright©Oxford University Press, 1998.
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Nucleic Acids ResHome page
G. L. Olsen, E. A. Louie, G. P. Drobny, and S. Th. Sigurdsson
Determination of DNA minor groove width in distamycin-DNA complexes by solid-state NMR
Nucleic Acids Res., September 1, 2003; 31(17): 5084 - 5089.
[Abstract] [Full Text] [PDF]

Home page
 Nucleic Acids ResHome page
D. A. Klewer, A. Hoskins, P. Zhang, V. J. Davisson, D. E. Bergstrom, and A. C. LiWang
NMR structure of a DNA duplex containing nucleoside analog 1-(2'-deoxy-{beta}-D-ribofuranosyl)-3-nitropyrrole and the structure of the unmodified control
Nucleic Acids Res., November 15, 2000; 28(22): 4514 - 4522.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (676K) Freely available
Right arrow Supplementary Data
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (11)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Aramini, J. M.
Right arrow Articles by Germann, M. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Aramini, J. M.
Right arrow Articles by Germann, M. W.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?