Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (245K) 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 (16)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Trotta, E.
Right arrow Articles by Paci, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Trotta, E.
Right arrow Articles by Paci, M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Nucleic Acids Research Pages 4706-4713  

Solution structure of DAPI selectively bound in the minor groove of a DNA T·T mismatch-containing site: NMR and molecular dynamics studies
Introduction
Materials And Methods
   Sample preparation and NMR spectroscopy
   Molecular modelling
   Relaxation matrix refinement
Results And Discussion
   Proton resonance assignment
   T·T mismatch
   Chemical shift changes upon binding
   Intermolecular NOEs in the DAPI-[d(GCGATTCGC)]2 complex
   Binding mechanism
   Molecular modelling of the DAPI-[d(GCGATTCGC)]2 complex
   Structure analysis of the DAPI-[d(GCGATTCGC)]2 complex
   Conclusions
Acknowledgements
References

Solution structure of DAPI selectively bound in the minor groove of a DNA T·T mismatch-containing site: NMR and molecular dynamics studies

Solution structure of DAPI selectively bound in the minor groove of a DNA T·T mismatch-containing site: NMR and molecular dynamics studies

Edoardo Trotta* and Maurizio Paci1

Istituto di Medicina Sperimentale, Area della Ricerca di Roma `Tor Vergata', CNR, Via del Fosso del Cavaliere snc, 00133 Roma, Italy and 1Dipartimento di Scienze e Tecnologie Chimiche, Università di Roma `Tor Vergata', 00133 Roma, Italy

Received June 11, 1998; Revised August 18, 1998; Accepted September 1, 1998

ABSTRACT

The solution structure of the complex between 4[prime],6-diamidino-2-phenylindole (DAPI) and DNA oligomer [d(GCGATTCGC)]2, containing a central T·T mismatch, has been characterized by combined use of proton one- and two-dimensional NMR spectroscopy, molecular mechanics and molecular dynamics computations including relaxation matrix refinement. The results show that the DAPI molecule binds in the minor groove of the central region 5[prime]-ATT-3[prime] of the DNA oligomer, which predominantly adopts a duplex structure with a global right-handed B-like conformation. In the final models of the complex, the DAPI molecule is located nearly isohelical with its NH indole proton oriented towards the DNA helix axis and forming a bifurcated hydrogen bond with the carbonyl O2 groups of a mismatched T5 and the T6 residue of the opposite strand. Mismatched thymines adopt a wobble base pair conformation and are found stacked between the flanking base pairs, inducing only minor local conformational changes in global duplex structure. In addition, no other binding mechanisms were observed, showing that minor groove binding of DAPI to the mismatch-containing site is favoured in comparison with any other previously reported interaction with G·C sequences.

INTRODUCTION

The synthetic antibiotic 4[prime],6-diamidino-2-phenylindole (DAPI) (Fig. 1A) is a DNA-binding drug that interferes with the activity of a number of DNA processing enzymes involved in regulatory and structural functions such as DNA ligase, exonuclease III and DNA polymerase I (1-7). Strong inhibitory activity of DAPI has been reported for RNA polymerase II and attributed to its high binding affinity in the minor groove of AT sequences (7). Such a strong DAPI-DNA interaction interferes with the binding of TBP (TATA-binding protein) to its consensus TATA box sequence, preventing formation of the transcription factors-DNA complex recognized and required by RNA polymerase II to initiate gene transcription.

DAPI is also used in biological and cytochemical studies as a fluorescent dye for DNA and chromosomes (8) and its nucleic acids complexes have been extensively studied showing different and sequence-dependent binding mechanisms to nucleic acids (9-26). DAPI preferentially binds B-DNA in the minor groove of two or more consecutive A·T base pairs, while it exhibits a quite different binding geometry and lower affinity for different DNA sequences (9-11,16,17). A single and exclusive binding mechanism of DAPI in the minor groove of consecutive A·T base-pairs of DNA has been reported and characterized by X-ray and NMR spectroscopy (14,20,22). In contrast, experimental evidence for binding mechanisms different from minor groove interaction have been reported for G·C or non-consecutive A·T sequences: intercalation (9-11,21), major groove binding (17) and [pi],[pi] stacking interactions with double helix ends (22). Steric clash and electrostatic repulsion due to the N2 amino group of guanine residues, as well as relatively unfavourable van der Waals interactions, preclude a minor groove interaction of DAPI within G·C sequences. The absence of binding in the shallow minor groove of AU sequences of RNA shows that van der Waals interactions, which are strongly favoured in the narrow minor groove of AT sequences of B-DNA, play an essential role in favouring minor groove complexes. For these reasons, DAPI should be considered a highly sensitive probe for evaluating chemical and structural alterations in the minor groove of consecutive A·T base pairs. Such studies appear fundamental to an understanding of the correlation between binding selectivity and the drug as well as DNA structure. Moreover, they are essential in planning new selective DNA-binding drugs with useful biological applications.


Figure 1. Chemical structure and numbering system for DAPI (A), d(GCGATTCGC) in the duplex form (B) and possible wobble structures of the T·T mismatch (C). The box in the duplex structure of oligomer (B) indicates the mismatch site.

In this paper we report the solution structure of DAPI bound in the minor groove of a T·T mismatch-containing site showing that a Watson-Crick A·T base pair can be replaced by its T·T mutated form without preventing minor groove binding. The results also show that the minor groove of T·T mismatches flanked by A·T base pairs represents a new genomic target for DAPI with a binding affinity higher than G·C base pairs.

MATERIALS AND METHODS

Sample preparation and NMR spectroscopy

The DNA oligomers were synthesized and purified as previously described (21). DAPI was purchased from Sigma and used without further purification. The purity of DAPI and DNA oligomers was checked by 1H NMR and their concentration measured spectrophotometrically using [epsis]342 = 23 000 (28) and [epsis]255(0.1 M KOH) = 81 600/M/cm (27), respectively. The concentration of 1:1 DAPI-[d(GCGATTCGC)]2 drug-duplex complex was 3 mM. NMR samples were suspended in 250 mM NaCl, 10 mM sodium phosphate, pH 7.0, in either 100% D2O or 90% H2O and 10% D2O solution.

NMR spectra were run as previously reported (22) using Bruker AM 400 and AMX 600 instruments operating at 400 and 600 MHz, respectively.

Molecular modelling

Molecular modelling and calculations were carried out using the InsightII v.2.3.0 and Discover v.2.9.5 software packages (Biosym Technologies Inc.). The AMBER force field was employed in all calculations and partial atomic charges for DAPI were used as previously reported (19). Calculations were performed in vacuo and a distance-dependent dielectric constant [epsis] = rij and [epsis] = 4rij was used to simulate solvent effects in molecular mechanics and dynamics calculations, respectively. A cut-off of 18 Å with a switching distance of 2 Å was used for non-bonded interactions and 1,4 electrostatic interactions were scaled by 0.5. Counterions were simulated by reducing to -0.2 the net charge of phosphate groups.

Two different starting structures of duplex were generated in the canonical A and B conformations and a T·T mismatch was incorporated by replacing the adenine base of a Watson-Crick A·T base pair with a thymine base. The starting complexes were created by manually docking a DAPI molecule in the minor groove of A and B duplexes and positioned to meet intermolecular NOEs. The initial structures were firstly subjected to restrained energy minimization (rEM) and then subjected to NOE-restrained Verlet molecular dynamics (rMD) by heating to 300 K for 2 ps followed by 100 ps of constant temperature simulation with time steps of 0.5 fs. Finally, the average structures of the last 5 ps of rMD simulations were restrained energy-minimized until the maximum derivative was <0.5 kcal/mol/Å, followed by 500 cycles of unrestrained energy minimization. Intramolecular DNA distance restraints were calculated from cross-peak volumes of NOESY spectra acquired at 25°C with 50 ms mixing times. The isolated spin pair approximation (ISPA) relationship rij = rref(NOEref/NOEij)1/6 was used to estimate distances, where rij and NOEij are the unknown distance and measured NOE volume between protons i and j and rref and NOEref represent the reference distance and the relevant NOE volume. The best resolved cross-peak among H5-H6 (distance 2.44 Å), H1[prime]-H2[prime][prime] (distance 2.37 Å) and H3[prime]-H2[prime] (distance 2.40 Å) was used as reference for each corresponding residue. Lower and upper bounds of distances in restrained calculations were set to ±0.5 Å and ±0.8 Å around calculated values rij < 3.7 Å and rij [ge] 3.7 Å, respectively. Consistent with experimental NOE data, Watson-Crick hydrogen bonds observed for non-mismatched base pairs were reinforced by incorporating distance restraints during structure calculations. Moreover, for imino-imino sequential NOEs observed in H2O, with the exclusion of those involving mismatched thymines, lower and upper bounds restraints of 1.9 and 4.0 Å, respectively, were applied. For mismatched thymines only intranucleotide H6-H1[prime] and intra-furanose NOE-derived distance restraints were used while no hydrogen bond and internucleotide distance restraints were applied during calculations. Lower bounds for intermolecular distance restraints were set to 1.7 Å and upper bounds up to 3.5 and 4.5 Å, depending on NOE intensities classified into strong and weak, respectively. A flat-bottomed harmonic potential for NOE-derived distances with a force constant of 50 kcal/mol Å2 was used. Calculations for each initial structure (A- and B-DNA) were performed at two different random number seeds giving rise to a total of four structures, which were used as starting models for relaxation matrix refinement. Eventual biases introduced by manual docking of the drug and a limited number of complex structures were safely excluded by the results of three further rMM-rMD simulations using a simulated annealing protocol: two further simulations with DAPI shifted along the minor groove by 1 bp in both directions and one simulation with DAPI inverted with respect to the manually docked B-type starting structure.

Relaxation matrix refinement

The four model structures resulting from ISPA-rMD calculations were subsequently refined by an iterative relaxation matrix procedure using the program FIRM (Full Iterative Relaxation Matrix) (29). An isotropic correlation time of 2 ns gave the best results in the tested range of 1.5-4 ns and was used throughout (30). NOE intensities were obtained from a NOESY spectrum acquired at 400 MHz and 31°C with a mixing time of 250 ms. A flat-bottomed harmonic potential for NOE-derived distances with a force constant of 25 kcal/mol Å2 was added to the AMBER force field. The flat bottom of the NOE restraint potential was initially set to ±20% of the distances calculated from the hybrid relaxation rate matrix and then reduced to ±10% in the last iterative procedures. All lower bounds exceeding 4 Å were set to 4 Å. No internucleotide distance restraints of mismatched thymines were used and hydrogen bond, imino-imino and intermolecular distance restraints were applied as reported above for the ISPA-rMD procedure. In addition to FIRM calculated distance restraints, a restricted number of restraints, concerning NOESY cross-peaks better resolved in spectra acquired at 25 than 31°C, were taken from previous ISPA-rMD simulations. Distance restraints were used for 40 ps of rMD calculations and the average structure of the last 5 ps was subjected to 500 cycles of restrained energy minimization. The resulting models were used for a new FIRM calculation of distance restraints. This procedure was repeated until agreement between back-calculated NOEs from the model structures and experimental NOEs did not improve further. Two R factors were calculated for evaluating the agreement between model structure and experimental NOEs (29):

where NOEexp represents the experimental NOE intensity and NOEcalc is the back-calculated NOE intensity from the model structures. Convergence of the four final structures was evaluated by calculating root mean square deviation (RMSD) of atomic coordinates. Lastly, helical parameters of the final DNA structures were measured using the program CURVES 5.1 (31), according to the definition of the EMBO Workshop on DNA Curvature and Bending (32).

RESULTS AND DISCUSSION

Proton resonance assignment

Analysis of NOESY spectra of DNA oligomer alone and in the complex, acquired at different temperatures and mixing times, showed cross-peak intensities clearly more compatible with a global B-like than A-like duplex conformation. In addition, comparison of NOESY spectra acquired with a 50 ms mixing time did not reveal marked alterations in oligomer structure upon binding with DAPI. Therefore, proton resonances of [d(GCGATTCGC)]2 were assigned by following the standard procedures described in the literature (33). (Supplementary material in NAR on-line gives the proton chemical shift values.) The sequential NOE connectivities of DNA resonances in the complex via (H6/H8)-H2[prime]/H2[prime][prime]/CH3 and (H6/H8)-H1[prime]/H5 protons are shown in Figures 2 and 3, respectively. As observed, the DNA oligomer is symmetrical in the NMR spectra and sequential NOE connectivities are not interrupted at T·T mismatch sites. The resonances of labile protons were assigned by following both sequential connectivities in the imino-imino region of the spectrum (Fig. 4) and interstrand cross-peaks between guanine H1 and cytosine base protons as well as A4H2 and T6H3. All the resonances assigned to imino and amino protons located outside mismatch site exhibited the typical chemical shifts and NOE connectivities of Watson-Crick base pairing. The chemical shift values of remaining DNA imino resonances at 10.24 and10.44 p.p.m. are consistent with those previously reported for hydrogen bonded imino protons of T·T wobble pairing (34-36) and were assigned to mismatched thymines. The observed sequential NOEs of these imino resonances with T6H3 confirm their assignment to mispaired T5, consistent with intercalation of these bases between the flanking A4·T6 base pairs. All these results show that the oligomer is prevalently a right-handed duplex structure, ruling out the predominant presence of other forms, such as monomeric single-stranded and hairpin structures.


Figure 2. Expanded region of NOESY spectrum of a 1:1 DAPI-[d(GCGATTCGC)]2 complex acquired with a mixing time of 250 ms in D2O buffer at 31°C. The broken line indicates the sequential assignment via (H6/H8)-H2[prime]/2[prime][prime] resonances.


Figure 3. Expanded region of NOESY spectrum of a 1:1 DAPI-[d(GCGATTCGC)]2 complex acquired with a mixing time of 250 ms in D2O buffer at 31°C. The broken line indicates the sequential assignment via (H6/H8)-H1[prime] resonances. Intermolecular DAPI-DNA cross-peaks are labeled and assigned as follows: a, H2[prime]/6[prime]-T5H1[prime]; b, H3[prime]/5[prime]-T5H1[prime], H3[prime]/5[prime]-C7H1[prime]; c, H7-T5H1[prime]; d, H2[prime]/6[prime]-T6H1[prime]; e, H7-T6H1[prime].


Figure 4. Expanded region of NOESY spectrum of a 1:1 DAPI-[d(GCGATTCGC)]2 complex acquired with a mixing time of 250 ms in H2O buffer at 8°C. The plot shows sequential NOEs between imino resonances of adjacent base pairs. The assignment of imino resonances belonging to drug and DNA are indicated by solid lines.

Following the procedure previously described (20,21), the drug resonances in the complex at 7.24, 7.49, 7.89, 7.92, 8.03, 8.29 and 11.28 p.p.m. were assigned to H3, H5, H4, H7, H3[prime]/5[prime], H2[prime]/6[prime] and H1 protons, respectively.


Figure 5. Proton chemical shift of [d(GCGATTCGC)]2 and [d(GCGATCGC)]2 plotted versus their position in the sequence. Resonances belonging to nonamer [d(GCGATTCGC)]2 and octamer [d(GCGATCGC)]2 are marked by o and x, respectively.

T·T mismatch

Sequential NOEs expected for a Watson-Crick right-handed structure are observed along the whole DNA molecule without interruption at the T·T mismatch site (Figs 2-4). In particular, mismatched thymines exhibit internucleotide base-sugar and base-base connectivities, including sequential imino NOEs, expected for a thymine in an A·T Watson-Crick structure. This strongly suggests that mismatched thymines are inserted and stacked between flanking base pairs. Such a consideration is confirmed by the results reported in Figure 5. In this figure proton chemical shifts of our nonamer [d(GCGATTCGC)]2 are compared with those previously reported for the B-type octamer [d(GCGATCGC)]2 with the same nucleotide sequence lacking the T·T mismatch (20). As observed, insertion of a T·T mismatch does not significantly change the chemical shift of proton resonances except for the T5H2[prime], T5H2[prime][prime] and T5H6 resonances of [d(GCGATCGC)]2. Considering that T5 of [d(GCGATCGC)]2 is the only residue that changes its flanking bases as a consequence of T·T mismatch insertion (5[prime]-purine-T-pyrimidine-3[prime] to 5[prime]-pyrimidine-T-pyrimidine-3[prime]), its observed chemical shift changes could be simply attributed to a sequence effect and not to a conformational alteration. On the other hand, resonances of A4 appear essentially unperturbed, confirming that mismatched bases are positioned and oriented between flanking base pairs in a similar way to that previously reported for the Watson-Crick paired thymine of octamer [d(GCGATCGC)]2. Lastly, the observation of no chemical shift differences of DNA resonances in the outer G·C regions (Fig. 5) is also consistent with minor perturbations induced by a T·T mismatch on the global double helix structure.

Imino proton resonances of mismatched thymines exhibit chemical shift values consistent with carbonyl hydrogen bonded imino protons (34-36). As expected in such a case and as illustrated in Figure 6, the imino proton resonance of the mismatched T5 is still present at 41°C and its broadening at high temperature is comparable with those observed for the resonances of the Watson-Crick hydrogen bonded G8 and T6 imino protons. In contrast, due to end fraying of the DNA oligomer, the resonance of the G1 imino proton of the terminal base pairs is not observable at temperatures >28°C. This is strong evidence for an easier access of solvent protons to the Watson-Crick paired ends than to the mismatch site in the complex and suggests a relatively more stable duplex structure at mispaired T·T than at the extremities of the oligomer.


Figure 6. Imino proton spectra of [d(GCGATTCGC)]2 and the DAPI-[d(GCGATTCGC)]2 complex. (A) Spectra of the complex acquired at different temperatures. (B) Spectra acquired at 8°C of DNA alone and after addition of DAPI.

Consistent with a rapid equilibrium between the two symmetrical T·T wobble structures (36; Fig. 1C), in spectra of DNA alone we observe a single averaged resonance for the two T5 imino protons with an integrated intensity corresponding to two protons for the duplex (Fig. 6B). The addition of DAPI causes splitting of this resonance (Fig. 6B) as an effect of either a reduced exchange rate between symmetrical wobble structures or destruction of the DNA symmetry by the asymmetrical drug. The absence of splitting of other imino resonances, in particular of imino resonances of T6 residues, which are inside the binding site, suggests that reduced exchange between wobble structures might be the cause of splitting of the T5 resonances. On the other hand, in previously reported NMR studies of DAPI bound in the minor groove of symmetrical DNA binding sites, splitting of imino resonances has never been observed, showing that the asymmetrical DAPI molecule did not alter the 2-fold symmetry of the DNA binding sites on the NMR timescale, even at 2°C (20,22). Consistently, a strong cross-peak, attributable to both NOE and chemical exchange, is observed in NOESY spectra between T5 imino signals (Fig. 4). As expected in such a case, the rapid equilibrium between symmetrical wobble structures is recovered by raising the temperature of the complex (Fig. 6).

The conformation of the furanose ring of the mismatched T5 was evaluated as lying predominantly in the south range of B-DNA by the relative NOE intensities H4[prime]-H3[prime] > H4[prime]-H1[prime] > H4[prime]-H2[prime][prime] observed in spectra acquired with short mixing times. The glycosidic torsion angle adopts an anti conformation, as shown by the weak intranucleotide NOE H6-H1[prime].

In conclusion, the results reported above are consistent with an equilibrium between the two symmetrical wobble T(anti)·T(anti) conformations, which are alternatively stabilized by two imino-carbonyl hydrogen bonds involving base O2 and O4 of the opposite mismatched thymines (Fig. 1C). For this reason, with exclusion of the intranucleotide H6-H1[prime] distance restraint, no sequential base-sugar intranucleotide and hydrogen bonddistance restraints concerning mispaired thymines were applied during structure calculations. Moreover, the results of qualitative analyses of mismatch and global DNA structure allowed us to use standard procedures for molecular dynamics calculations which start from quite different canonical A- and B-type conformations by using two different random seed values for each starting structure (30).

Chemical shift changes upon binding

Binding of DAPI to [d(GCGATTCGC)]2 induces an appreciable shift of a number of DNA proton resonances. The largest shifts are observed for protons exposed in the minor groove and belonging to residues T5, T6 and C7: T6H4[prime] (2.17 p.p.m. upfield shifted), C7H4[prime] (0.84 p.p.m. upfield shifted), T6H1[prime] (0.70 p.p.m. upfield shifted), T5H4[prime] (0.40 p.p.m. upfield shifted) and H5[prime]-5[prime][prime] of T6 and C7. In contrast, no shifts larger than 0.1 p.p.m. are observed for proton resonances belonging to G1, C2, G3, A4, G8 and C9 residues, with the exception of the resonance of minor groove proton A4H2, which is 0.24 p.p.m. downfield shifted upon binding. Appreciable shift of the resonances belonging to DNA imino and amino protons is only observed for T6H3, which is 0.2 p.p.m. downfield shifted. DAPI resonances in the complex are moderately shifted (0.00-0.21 p.p.m.).

Intermolecular NOEs in the DAPI-[d(GCGATTCGC)]2 complex

A total of 11 dipolar contacts between DAPI and [d(GCGATTCGC)]2 protons were observed. In particular, all the observed intermolecular NOEs involved protons exposed in the minor groove (Fig. 3) and no intermolecular NOEs were found for protons belonging to G1, C2, G3, G8 and C9 residues. Strong NOEs were observed for DAPI-DNA dipolar contacts H7-T5H1[prime], H7-T6H1[prime], H2[prime]/6[prime]-T5H1[prime], H3[prime]/5[prime]-T5H1[prime] and H3[prime]/5[prime]-A4H2 and weak NOEs for H7-A4H2, H2[prime]/6[prime]-T6H1[prime], H2[prime]/6[prime]-T6H5[prime], H3[prime]/5[prime]-T6H5[prime], H5-C7H4[prime] and H5-C7H5[prime].

Binding mechanism

Very different perturbations of the proton chemical shifts of DAPI and DNA resonances have been described in the literature for GC-type intercalation and AT-specific minor groove binding (11,20-22). Our results are quite consistent with selective binding of DAPI in the minor groove of the central 5[prime]-ATTC-3[prime] region containing the mismatch site. As expected in such a case, the strongest shifts of DNA proton resonances induced by DAPI binding are observed for protons located in the minor groove of the 5[prime]-ATTC-3[prime] region. Consistent with previous NMR studies of DAPI minor groove complexes (20,22), the drug aromatic rings induce positive shielding effects on the H1[prime] and H4[prime] DNA resonances which exhibit the strongest upfield shifts upon binding, while negative drug shielding effects are observed for adenine H2 and thymine H3 resonances, which are downfield shifted.

Chemical shift changes of drug resonances upon binding to DNA are also consistent with minor groove complexes and in contrast to the strong upfield shifts reported in the literature for DAPI bound to GC sequences (11,21).

The results of chemical shift perturbations upon binding are supported by the observation of intermolecular dipolar contacts involving exclusively DNA protons exposed in the minor groove of the central 5[prime]-ATTC-3[prime] region. Finally, consistent with minor groove complexes of DAPI (20,22), marked changes in DNA structure upon drug binding were not evident in the NOESY spectra.

These unequivocal results allowed us to reasonably use minor groove complexes as starting models for structural calculations.

Molecular modelling of the DAPI-[d(GCGATTCGC)]2 complex

A total of 325 distance restraints were applied during structure refinement by the iterative relaxation matrix procedure: 262 derived from FIRM calculations, 22 from ISPA, 22 Watson-Crick hydrogen bonds, 8 involving imino protons and 11 evaluated from intermolecular NOEs. As a final result, four refined structures (DA1f, DA2f, DB1f and DB2f) were obtained showing NOE R factors RNOE = 0.25 ± 0.01 and Rd = 0.047 ± 0.001 and converging to a pairwise 0.38 Å [le] RMSD [le] 0.94 Å. Lower values of R factors and RMSD were calculated by excluding terminal base pairs: RNOE = 0.22 ± 0.01, Rd = 0.044 ± 0.001 and 0.27 Å [le] RMSD [le] 0.90 Å. Each final structure showed four to six violations of NOE distance restraints >0.2 Å, while no violation exceeded 0.5 Å. A moderate global violation of NOE distance restraints of the final structure also emerges from the low values of the associated energy terms (19.48 ± 1.38 kcal/mol). Figure 7 shows the best fit superimposition of the four refined final models (DA1f, DA2f, DB1f and DB2f) compared with their starting structures DA (A-type) and DB (B-type).


Figure 7. Comparison of starting structures DA (A-DNA) and DB (B-DNA) and best fit superimposed final structures (DA1f, DA2f, DB1f and DB2f) obtained by the relaxation matrix refinement procedure. Mismatched thymines (green) and DAPI (red) are shown.

Structure analysis of the DAPI-[d(GCGATTCGC)]2 complex

In the final models the ligand is located nearly isohelical with its NH indole proton oriented towards the DNA helix axis and forming a bifurcated hydrogen bond with the O2 atom of a mismatched T5 and T6O2 of the opposite strand. If the two DNA strands are defined as [Igr] and [Igr][Igr] on the basis of DAPI orientation (Figs 8 and 9), all rMD and final refined structures give rise to the same wobble configuration of the T·T mismatch, characterized by interhelical hydrogen bonds T5(I)N3-T5(II)O2 and T5(II)N3-T5(I)O4, as well as the same intermolecular bifurcated hydrogen bond involving the NH indole proton of DAPI and T5(I)O2 and T6(II)O2. The complex is further stabilized by two additional hydrogen bonds between the DAPI indole amidine group and T6(I)O2 and between the drug phenyl amidine group and C7(II)O2. In addition to these permanent hydrogen bonds, analysis of the trajectory of the last cycle of rMD simulations shows transient formation of other intermolecular hydrogen bonds, which are indicated in Figure 9. Considering that the final structures only represent an averaged static model of a more complex dynamic state which could include such transient hydrogen bonds, for a more correct analysis of binding and the overall dynamic stability of the complex the contribution of these transient hydrogen bonds should also be considered. In this regard, we must point out that solvent molecules were not included in our simulations, however, analysis of 22 crystal structures has shown that water bridges are not a common feature in most DNA minor groove-binding drug complexes (37).

The final models agree with the reduced exchange with the solvent observed for the NH indole proton of DAPI in the complex (Fig. 6) and with the large upfield shift of the H1[prime] and H4[prime] resonances of T6 residues, as well as the downfield shift of the A4H2 and T6H3 resonances induced by current effects of the ligand aromatic rings.


Figure 8. The view shows the mismatched T5 residues (green) and DAPI molecule (red) of the four best fit superimposed final structures (DA1f, DA2f, DB1f and DB2f). Broken lines indicate interstrand hydrogen bonds between mispaired thymines as well as the intermolecular hydrogen bond between the DAPI NH indole and O2 of T5(I).

The RMSD and R factor values show that the initial A-type structure of the complex (DA) exhibits poor agreement with the experimental NOEs (RNOE = 0.64 and Rd = 0.151 on excluding terminal base pairs) and is very different from the final refined structures DA1f, DA2f, DB1f and DB2f (3.13 Å [ge] RMSD [ge]3.91 Å on excluding terminal base pairs). In contrast, the initial B-type structure (DB) is clearly more consistent with both the NOE results and the final structures (RNOE = 0.38, Rd = 0.072 and 1.32 Å [ge] RMSD [ge] 1.73 Å on excluding terminal base pairs). To better evaluate the conformity of our experimental NOEs with the canonical B-type structure, a further simulation was performed by reducing the bounds of NOE distance restraints from 10 to 5% in the last cycle of the FIRM-rMD-rEM refinement procedure of DA1f. The resulting structure exhibits better agreement with the NOE data and a lower RMSD value with the initial B-type structure than DA1f (excluding terminal base pairs RNOE = 0.20, Rd = 0.037 and RMSD = 1.40 Å). This result clearly shows that the predominant character of the B-like structure of our oligomer is well characterized by our experimental NOEs.

In the final models of the complex, DNA adopts a right-handed double helix structure with all glycosidic torsion angles in the anti conformation. The furanose ring pucker of all residues exhibits a south C2[prime]-endo-C1[prime]exo conformation (B-type), with the exception of C9 and T6, which in the models adopt an O1[prime]-endo geometry midway between south (B-DNA) and north (A-DNA) conformations. The prevalent B-like conformation of our oligomer is also evidenced by measuring the minor groove width of our oligomer, which ranges from 4 to 7 Å (supplementary material in NAR on-line), consistent with a canonical B-DNA (6 Å) and strongly sharper than that expected for an A-DNA structure (11 Å). Furthermore, within the binding site, the minor groove is up to 0.7 Å sharper than in the crystal structure of DAPI bound to [d(CGCGAATTCGCG)]2 (14), while it becomes wider at the outer flanking sequences (supplementary material in NAR on-line).


Figure 9. Schematic representation of the DAPI-[d(GCGATTCGC)]2 complex showing intermolecular hydrogen bonds observed in the final structures (DA1f, DA2f, DB1f and DB2f) (broken line) and during the last rMD simulations (solid lines). The donor hydrogen bonding groups of DAPI and the acceptor groups of DNA are indicated by circles.

Although experimental NOEs (R factors), furanose ring pucker and minor groove width are strongly consistent with a global B-type conformation of our oligomer in the complex, fine analysis of DNA global helical parameters shows some shift from standard values which appear sequence-dependent or related to mismatch or binding site. The wobble configuration of the T5(I)·T5(II) mismatch appears characterized by strong negative values of stretch (-1.66 ± 0.03 Å) and shear (-2.12 ± 0.11 Å) which are selectively observed at the mismatch site. The negative shear also shows a shift of T5(I) towards the minor groove, with its carbonyl O2 group exposed and accessible for hydrogen bonding with the NH indole of DAPI. Some parameters appear to change in response to formation of a bifurcated hydrogen bond, which forces the position and the orientation of T5(I)O2 and T6(II)O2 towards the NH indole proton of DAPI. This is shown by the particular negative values of buckle (-20.7 ± 1.3°) and propeller twist (-32.9 ± 3.1°) as well as positive tip (2.6 ± 0.2°) at the A4(I)·T6(II) base pair and by negative roll (-10.8 ± 6.9°) and tilt (-15.2 ± 1.2°) at the A4-T5 base pair step. Moreover, the A4(I)·T6(II) base pair exhibits a selective strong negative stagger (-0.97 ± 0.06 Å) that is not associated with a corresponding increase in rise in the A4-T5 base pair step and brings T5(I) close to T6(II). Negative roll is observed in the center of the oligomer at the A4-T5 (-10.8 ± 6.9°) and T5-T6 steps (-7.0 ± 1.9°), while positive or null values are found at the ends. This results in a local bending of our oligomer towards the minor groove in the central 5[prime]-ATT-3[prime] sequence and towards the major groove at the outer base pair steps. The negative values of tilt also indicate an added bending toward strand I at A4-T5 steps (-15.2 ± 1.2°). In the inner region of the oligomer 5[prime]-ATTC-3[prime], which is involved in binding and contains the T·T mismatch, positive opening values are observed indicating base pair opening towards the major groove. Rise values do not exceed those of the canonical B structure, consistent with the absence of an intercalation binding mechanism. This parameter varies in the range 2.58-3.54 Å along the whole sequence, with lower values that appear simply associated with the purine-pyrimidine base pair steps. Finally, unwinding of the helix, evidenced by the high value of twist angle (46 ± 2°), is observed at the step T5-T6 and a high value of shift (0.88 ± 0.04 Å) is measured at the G3(I)·C7(II) base pair.

Conclusions

As is well established in the literature, the preferential DNA target of DAPI is the minor groove of two or more consecutive Watson-Crick A·T base pairs. In contrast, no minor groove binding has been observed with DNA sequences containing G·C base pairs or isolated A·T base pairs. It has also been reported that intercalation of DAPI within G·C tracts is favoured in comparison with minor groove binding to non-consecutive A·T base pairs (21). In the oligomer of the present study, although adjacent Watson-Crick A·T base pairs are not present, DAPI binds selectively in the minor groove of the inner 5[prime]-ATT-3[prime] region, while no binding has been observed inside G·C tracts. This indicates that T·T mismatches flanked by A·T base pairs can be considered a more favoured DAPI genomic target than G:C base pairs.

Considering that minor groove binding of DAPI is highly sensitive to chemical and structural features of the DNA double helix, this result is consistent with the observation that a global B-like conformation is not significantly altered by a T·T mismatch and only minor structural differences between our oligomer [d(GCGATTCGC)]2 and [d(GCGATCGC)]2 were observed. If one step of the in vivo mismatch repair mechanism is affected by either recognition of major changes in the minor groove or perturbation of the global Watson-Crick double helix structure, this result appears consistent with the reported low efficiency of the T·T mismatch repair mechanism with respect to other mismatched base pairs (38).

The results also show that the two possible T·T wobble pairings are interchangeable and that one of the two O2 of the mismatched thymines could be indifferently exposed in the minor groove. Therefore, in contrast to Watson-Crick A·T base pairs, the geometry of hydrogen acceptors in the minor groove of T·T mismatches is variable and easily adaptable to the geometry of the hydrogen donors of the ligands.

Finally, the results suggest that DAPI could reduce the exchange rate between the two wobble forms and, from the results of structure calculations, the wobble pairing which allows formation of a bifurcated hydrogen bond involving thymines of opposite strands appears preferred. Moreover, as shown by H2O NMR experiments, base pair stability of the T·T mismatch in the complex appears higher than terminal G·C and comparable with the other inner Watson-Crick base pairs. Such a stable base pairing structure of 5[prime]-ATT-3[prime] in the complex with DAPI appears of biological interest, considering that the AAT·ATT class of trinucleotide repeats is the most abundant and the most frequently polymorphic in the human genome (39) and shows a propensity to adopt a non-hydrogen bonded structure (40). Further studies in our laboratory are in progress to better evaluate thermodynamic data on binding to 5[prime]-ATT-3[prime] by also using (ATT)n oligomers.

ACKNOWLEDGEMENTS

This work was partly supported by the Consiglio Nazionale delle Ricerche, Progetto Strategico `Biologia Strutturale'. We thank Dr Daniela Orru[prime] for helpful discussion and Fabio Bertocchi and Enrico Rossi for technical assistance with the 400 and 600 MHz instruments. The NMR Service of the CNR Research Area of Montelibretti, Roma, courtesy of A.L.Segre, is gratefully acknowledged for running the 600 MHz experiments.

See supplementary material available in NAR Online.

REFERENCES

1. Storl,K., Storl,J., Zimmer,C.H. and Lown,J.W. (1993) FEBS Lett., 317, 157-162. MEDLINE Abstract

2. Palù,G., Valisena,S., Barcellona,M.L., Masotti,L. and Meloni,G.A. (1987) Biochem. Biophys. Res. Commun., 145, 40-45. MEDLINE Abstract

3. Parolin,C., Zanotti,G. and Palù,G. (1995) Biochem. Biophys. Res. Commun., 208, 332-338. MEDLINE Abstract

4. Welch,J.J., Rauscher,F.J. and Beerman,T.A. (1994) J. Biol. Chem., 269, 31051-31058. MEDLINE Abstract

5. Parolin,C., Montecucco,A., Chiarocchi,G., Pedrali-Noy,G., Valisena,S., Palumbo,M. and Palù,G. (1990) FEMS Microbiol. Lett., 68, 341-346. MEDLINE Abstract

6. Straney,D.C. and Crothers,D.M. (1987) Biochemistry, 26, 1987-1995. MEDLINE Abstract

7. Chiang,S.-Y., Welch,J., Rausher,F.J. and Beerman,T.A. (1994) Biochemistry, 33, 7033-7040. MEDLINE Abstract

8. Kapuscinski,J. (1995) Biotech. Histochem., 70, 220-233. MEDLINE Abstract

9. Wilson,W.D., Tanious,F.A., Barton,H.J., Jones,R.L., Fox,K., Wydra,R.L. and Strekowski,L. (1990) Biochemistry, 29, 8452-8461. MEDLINE Abstract

10. Wilson,W.D., Tanious,F.A., Barton,H.J., Wydra,R.L., Jones,R.L., Boykin,D.W. and Strekowski,L. (1990) Anticancer Drug Des., 5, 31-42. MEDLINE Abstract

11. Wilson,W.D., Tanious,F.A., Barton,H.J., Strekowski,L. and Boykin,D.W. (1989) J. Am. Chem. Soc., 111, 5008-5010.

12. Kubista,M., Åkerman,B. and Nordén,B. (1987) Biochemistry, 26, 4545-4553. MEDLINE Abstract

13. Manzini,G., Barcellona,M.L., Avitabile,M. and Quadrifoglio,M. (1983) Nucleic Acids Res., 11, 8861-8876. MEDLINE Abstract

14. Larsen,T.A., Goodsell,D.S., Cascio,D., Grzeskowiak,K. and Dickerson,R.E. (1989) J. Biomol. Struct. Dyn., 7, 477-491. MEDLINE Abstract

15. Loontiens,F.G., McLaughlin,L.W., Diekmann,S. and Clegg,R.M. (1991) Biochemistry, 30, 182-189. MEDLINE Abstract

16. Jansen,K., Nordén,B. and Kubista,M. (1993) J. Am. Chem. Soc., 115, 10527-10530.

17. Kim,S.K., Eriksson,S., Kubista,M. and Nordén,B. (1993) J. Am. Chem. Soc., 115, 3441-3447.

18. Sehlstedt,U., Kim,S.K. and Nordén,B. (1993) J. Am. Chem. Soc., 115, 12258-12263.

19. Mohan,S. and Yathindra,N. (1994) J. Biomol. Struct. Dyn., 11, 849-867. MEDLINE Abstract

20. Trotta,E., D'Ambrosio,E., Del Grosso,N., Ravagnan,G., Cirilli,M. and Paci,M. (1993) J. Biol. Chem., 268, 3944-3951. MEDLINE Abstract

21. Trotta,E., D'Ambrosio,E., Ravagnan,G. and Paci,M. (1995)Nucleic Acids Res., 23, 1333-1340. MEDLINE Abstract

22. Trotta,E., D'Ambrosio,E., Ravagnan,G. and Paci,M. (1996) J. Biol. Chem., 271, 27608-27614. MEDLINE Abstract

23. Eriksson,S., Kim,S.K., Kubista,M. and Nordén,B. (1993) Biochemistry, 32, 2987-2998. MEDLINE Abstract

24. Kapuscinski,J. and Szer,W. (1979) Nucleic Acids Res., 6, 3519-3534. MEDLINE Abstract

25. Tanious,F.A., Spychala,J., Kumar,A., Greene,K., Boykin,D.W. and Wilson,W.D. (1994) J. Biomol. Struct. Dyn., 11, 1063-1083. MEDLINE Abstract

26. Tanious,F.A., Veal,J.M., Buczak,H., Ratmeyer,L.S. and Wilson,W.D. (1992) Biochemistry, 31, 3103-3122. MEDLINE Abstract

27. Kastrup,R.V., Young,M.A. and Krugh,T.R. (1978) Biochemistry, 17, 4855-4865. MEDLINE Abstract

28. Kapuscinski,J. and Skoczylas,B. (1977) Anal. Biochem., 83, 252-257. MEDLINE Abstract

29. Edmondson,S.P. (1992) J. Magn. Reson., 98, 283-298.

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

31. Lavery,R. and Sklenar,H. (1988) J. Biomol. Struct. Dyn., 6, 63-91. MEDLINE Abstract

32. Diekman,S. (1989) EMBO J., 8, 1-4.

33. Feigon,J., Sklenar,V., Wang,E., Gilbert,D.E., Macaya,R.F. and Schultze,P. (1992) Methods Enzymol., 211, 235-253. MEDLINE Abstract

34. Arnold,F.H., Wolk,S., Cruz,P. and Tinoco,I. (1987) Biochemistry, 26, 4068-4075. MEDLINE Abstract

35. Kouchakdjian,M., Li,B.F.L., Swann,P.F. and Patel,D.J. (1988) J. Mol. Biol., 202, 139-155. MEDLINE Abstract

36. Gervais,V., Cognet,J.A.H., Le Bret,M., Sowers,L.C. and Fazakerley,G.V. (1995) Eur. J. Biochem., 228, 279-290. MEDLINE Abstract

37. Tabernero,L., Bella,J. and Alemán,C. (1996) Nucleic Acids Res., 24, 3458-3466. MEDLINE Abstract

38. Kramer,B., Kramer,W. and Fritz,H.-J. (1984) Cell, 38, 879-887. MEDLINE Abstract

39. Gastier,J.M., Pulido,J.C., Sunden,S., Brody,T., Buetow,K.H., Murray,J.C., Weber,J.L., Hudson,T.J., Sheffield,V.C. and Duyk,G.M. (1995)Hum. Mol. Genet., 4, 1829-1836. MEDLINE Abstract

40. Ohshima,K., Kang,S., Larson,J.E. and Wells,R.D. (1996) J. Biol. Chem., 271, 16784-16791. MEDLINE Abstract

*To whom correspondence should be addressed. Tel: +39 0672594447; Fax: +39 0672594328; Email: Edoardo.Trotta@ims.rm.cnr.it

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: 30 Sep 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
E. R. Lacy, K. K. Cox, W. D. Wilson, and M. Lee
Recognition of T{middle dot}G mismatched base pairs in DNA by stacked imidazole-containing polyamides: surface plasmon resonance and circular dichroism studies
Nucleic Acids Res., April 15, 2002; 30(8): 1834 - 1841.
[Abstract] [Full Text] [PDF]

Home page
 Nucleic Acids ResHome page
J.-H. Lee, Y.-J. Choi, and B.-S. Choi
Solution structure of the DNA decamer duplex containing a 3'-T{middle dot}T base pair of the cis-syn cyclobutane pyrimidine dimer: implication for the mutagenic property of the cis-syn dimer
Nucleic Acids Res., April 15, 2000; 28(8): 1794 - 1801.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (245K) 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 (16)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Trotta, E.
Right arrow Articles by Paci, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Trotta, E.
Right arrow Articles by Paci, M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?