Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (874K) Freely available
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 (8)
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Ouali, M.
Right arrow Articles by Taillandier, E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ouali, M.
Right arrow Articles by Taillandier, E.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© 1997 Oxford University Press 4816-4824

Hydration of the dTn·dAn×dTn parallel triple helix: a Fourier transform infrared and gravimetric study correlated with molecular dynamics simulations

Hydration of the dT n ·dA n × dT n parallel triple helix: a Fourier transform infrared and gravimetric study correlated with molecular dynamics simulations Mohammed Ouali, Hervé Gousset, Frédéric Geinguenaud, Jean Liquier, Jacques Gabarro-Arpa1, Marc Le Bret1 and Eliane Taillandier*

Laboratoire CSSB, URA CNRS 1430, UFR Santé Médecine et Biologie Humaine, Université Paris XIII, 74 rue Marcel Cachin, 93017 Bobigny, France and 1Laboratoire de Physicochimie de Macromolécules Biologiques URA CNRS 147, Institut Gustave Roussy, 94805 Villejuif, France

Received July 21, 1997; Revised and Accepted October 10, 1997

ABSTRACT

We present a comparative analysis of the water organization around the dTn·dAn×dTn triple helix and the Watson-Crick double helix dTn·dAn respectively by means of gravimetric measurements, infrared spectroscopy and molecular dynamics simulations. The hydration per nucleotide determined by gravimetric and spectroscopic methods correlated with the molecular dynamics simulations shows that at high relative humidity (98% RH) the triple helix is less solvated than the duplex (17 ± 2 water molecules per nucleotide instead of 21 ± 1). The experimental desorption curves are different for both structures and indicate that below 81% RH the triplex becomes more hydrated than the duplex. At this RH the FTIR spectra show the emergence of N-type sugars in the adenosine strand of the triplex. When the third strand is bound in the major groove of the Watson-Crick duplex molecular dynamics simulations show the formation of a spine of water molecules between the two thymidine strands.

INTRODUCTION

During the last 10 years theoretical and experimental studies of triplex structures have been stimulated by possible applications in the control of gene expression through the binding of either a single strand to a specific DNA duplex (the so called triple helix forming oligonucleotide or TFO strategy; 1 -4 ) or of two strands to a single-strand RNA (an extension of the antisense concept; 5-7). Earlier structural studies of triple helices were performed more than 20 years ago by Arnott and co-workers (8 -10 ). They showed that triple helix formation occurs when an oligonucleotide binds inside the major groove of a double helix having a purine strand by forming hydrogen bonds. In the case of the dTn·dAn double helix an oligonucleotide dTn binds to the dAn strand of the DNA Watson-Crick duplex through Hoogsteen hydrogen bonds forming a dTn·dAn×dTn triple helix (Fig. 1 ). Recent studies have revisited the original model of Arnott. In particular, it has been shown that all sugars adopt the S-type conformation (11 -13 ). Important modifications are observed in the Watson-Crick duplex after binding of the third strand forming the triple helix. The helical parameters of the Watson-Crick duplex in the triplex were found to be intermediate between those of the A and B conformations (14 ). Such geometrical changes should give rise to differences between hydration patterns of the double and triple helices. In this paper we examine the question of how the organization of water molecules differs around the dTn·dAn×dTn triple helix and the dTn·dAn Watson-Crick double helix. An aqueous medium is usually a prerequisite for duplex or multistranded helix formation. The extent of hydration, coupled with ionic strengh, determines the polymorphic character of the DNA structure, which is dependant on the surrounding water molecules covering the surface of DNA double helices (15 -17 ). Part of these molecules should be pushed away if an oligonucleotide binds in the major groove of the Watson-Crick double helix. These various considerations led us to suppose that formation of the triple helix induces a global desolvation of the three-stranded complex mainly due to the volume excluded by the third strand oligonucleotide in the major groove, which might be responsible for the structural modifications observed in the Watson-Crick duplex involved in the triplex. Such a phenomenon should be observable from a global point of view via gravimetric measurements (18 ) and vibrational spectroscopy (19 -21 ). Until now little has been known of the hydration of triple helices, mainly due to the lack of experimental data. However, theoretical studies of (dC·dG×dG)7 triple helix hydration have been performed by means of molecular dynamics simulations (22 -24 ). It has notably been shown that a spine of hydration occurs between the third dG strand and the pyrimidine strand of the complex.


Figure 1. Residue numbers in the duplex and triplex sequences. Scheme of the T·A×T base triplet indicating the definitions of the grooves.

In the present paper we compare hydration of the dTn·dAn duplex and of the dTn·dAn×dTn triple helix using gravimetric measurements, Fourier transform infrared spectroscopy (FTIR) and molecular dynamics to determine the desolvation effect occuring as a consequence of triplex formation. We shall first present the gravimetric measurements and the FTIR results, which allow us to study the structure of DNA complexes as well as hydration from a global point of view by comparing the hydration per nucleotide of a double and a triple helix. We shall then compute hydration of the obtained structures at the atomic level.

MATERIALS AND METHODS

Experimental

Sample preparation. dAn·dTn and dTn were purchased from Sigma and used without further purification. Stock solution concentrations were determined by measurement of the absorption around 260 nm with a Kontron Uvikon 933 spectrophotometer using extinction coefficients of 6000 and 8400 mol l-1cm-1 respectively for dAn·dTn and dTn. The double-stranded polymer solution was heated for 5 min at 80°C and cooled down to room temperature before addition of a stoichiometric amount of dTn so as to obtain the dTn·dAn×dTn triple helical structure. For FTIR spectroscopy homogenous films were obtained by slow evaporation of a stock solution droplet deposited on a ZnSe window. The sample was then placed in a cuvette closed by a KRS5 window. At the bottom of the cuvette a small vessel filled with saturated salt solution was introduced. Monitoring of nucleic acid film hydration was achieved by changing the saturated salt solution in the vessel. The salts used were 98% Pb(NO3)2, 93% NH4H2PO4, 84% KCl, 81% KBr, 75% NaCl, 66% NaNO2, 58% NaBr, 47% KSCN, 31% CaCl2 and 11% LiCl. Other points were obtained by gently bubbling nitrogen passed through saturated salt solutions into the FTIR cell. The relative humidities imposed were checked with a capacitive thermohygrometer Testo 610. The RH values are accurate to ±1%.

Gravimetric measurements. Films of dAn·dTn and dTn·dAn×dTn were exposed in polypropylene Eppendorf tubes to atmospheres of constant RH for periods of 24-48 h and weighed to 10-5 g using a Sartorius microbalance. Blank identical vessels were weighed under the same conditions so as to allow correction for water adsorbed on the tube. Correction was found to be important only above 86% RH. The dry weight was determined after the experiment by lyophilizing the samples.

FTIR spectroscopy. FTIR spectra were recorded using a Perkin Elmer series 2000 spectrophotometer. Twenty scans were accumulated. The presented spectra are non-treated data (no baseline correction, no smoothing and no deconvolution). So as to compare the water amount computed for different samples, normalization of the spectra was performed by setting the ratio of the integrated relative intensities of the absorption assigned to the symmetrical stretching vibration of the phosphate groups (1090 cm-1) of the duplex and triplex films equal to 1 in the spectra recorded at 3% RH obtained by circulating dry nitrogen in the cell.

Calculations

Construction of the structures. The dT10·dA10×dT10 triplex was generated from our previous molecular modeling of triple helices correlated with infrared spectroscopy (14 ) using the MORCAD package (25 ). The dT10·dA10 double helix was generated from the data of Arnott et al. (26 ).

Minimization of the structures. Molecular modeling by conformational energy minimization was carried out using the MORMIN program, a rapid and efficient quasi-Newtonian minimizer (25 ). The new AMBER 94 force field was used to describe interactions between atoms; the 1-4 electrostatic interactions were damped by a factor of 1.2 as suggested by the authors (27 ). Neither water nor counterions were explicitly included in the energy minimization process. However, their effects were simulated by using a sigmoidal distance-dependent dielectric function (28 ,29 ). Energy refinements of the structures were terminated when the root mean squares were <0.05 kcal/Å.

Molecular dynamics in aqua with explicit counterions. Starting from the minimized structure, the sodium counterions are set along the phosphate bissector with the EDIT module of AMBER (30 ). The triplex with its counterions was set in a box of Monte Carlo water so that there was a minimum of 8 Å between the box edges and any atom of the DNA-counterion complex. For the dT10·dA10 double helix the constructed system contains 18 sodium ions and 2167 water molecules in a rectangular box of 49.32 × 41.25 × 41.24 Å3. For the dT10·dA10×dT10 triple helix the constructed system contains 27 counterions and 2669 water molecules in a rectangular box of 51.34 × 45.79 × 44.96 Å3. Periodic boundaries were applied on the solvent boxes, with the minimal image model. The elementary boxes were electrically neutral. The TIP3P model (31 ) was used for water-water interaction potential and the dielectric constant was taken as equal to 1. A simple cut-off of 10 Å was used for long range interactions and the non-bonded pairs list was updated every 20 steps.

A hundred steps of steepest descent minimization without any constraints on the two systems were performed in order to remove the strain of the initial structures prior to the molecular dynamics simulation. The molecular dynamic simulations were performed in the (NTV) canonical ensemble. Starting from the minimized structures, the Verlet algorithm was used with a time interval of 2 fs/step (32 ). In order to ensure stability of the Verlet algorithm, bond lengths involving hydrogen atoms were constrained with the SHAKE algorithm (33 ). The protocol of the molecular dynamic simulations and artifact effects due to the use of a simple cut-off of the electrostatic parameters have already been discussed in previous works (34 ,35 ). The phase production was run over 600 ps for each system and the generated conformations were displayed using the MORCAD package.

RESULTS AND DISCUSSION

Gravimetric measurements

The results of our gravimetric study are summarized in Figure 2 , which presents the desorption curves for the duplex (solid squares) and the triplex (open circles). The water content is given as mol water/mol nucleotide. At high relative humidity (98% RH) the duplex was more hydrated than the triplex. The number of water molecules per nucleotide was estimated as equal to 21 ± 0.4 for the duplex and 17 ± 0.4 for the triplex. At 86% RH the figures were respectively 14 ± 0.6 and 12 ± 0.6 water molecules/nucleotide. At 81% RH the hydration is similar for both structures (around 9 ± 1 water molecules/nucleotide). When the RH was decreased we observed a crossover of the desorption curves corresponding to the triplex and duplex. Below 81% RH the triplex was found to be more hydrated than the duplex.


Figure 2.Water desorption curves obtained by gravimetry. Solid squares, dTn·dAn duplex; open circles, dTn·dAn×dTn triplex.

FTIR spectroscopy

It was important to ensure that the duplex and triplex structures were maintained during the dehydration process. The FTIR spectra recorded at different RH show that double and triple helices remained formed until relative humidities of ~58%. Figure 3 , left presents the FTIR spectra of a dTn·dAn film recorded at different relative humidities (between 3 and 98% RH) in the 4000-700 cm-1 spectral domain. Hydration of the sample can be evaluated by measurement of integrated intensity of water absorption located around 3400 cm-1. Under the same conditions of relative humidity, hydration of the duplex and triplex are different, as can be observed on the plot presented in Figure 3 , right. Hydration is computed as the integrated area of absorption between 3800 and 2400 cm-1 corrected for the presence of residual absorption of the nucleic acid itself. This residual intensity can be estimated from the spectrum recorded at 3% RH, bearing in mind that under such conditions no significant structure, neither duplex nor triplex, should be expected. Two conclusions can be derived from this plot. First, under high relative humidity conditions, which mimic as closely as possible the conditions encountered in a highly concentrated gel, the hydration per nucleotide of the duplex is larger than that of the triplex (measured ratio 1.2). Second, water desorption is stronger in the case of the duplex than in the case of the triplex. Below 75% RH the water per nucleotide content in the triplex is larger than that in the duplex (see inset to Fig. 3 , right, which shows variation of the hydration ratio duplex:triplex with decreasing RH). This is in agreement with the gravimetric measurements. Moreover, the FTIR spectra show that a conformational transition of some sugar puckers occurs in the triplex and not in the duplex upon decreasing RH.


Figure 3. (Left) FTIR spectra of a dTn·dAn film recorded at relative humidities between 3 and 98% RH. (Right) Water desorption curves obtained by FTIR (integrated absorption of the water vibration located at 3400 cm-1). Solid squares, dTn·dAn duplex; open circles, dTn·dAn×dTn·triplex. (Inset) Ratio of hydration per nucleotide duplex/triplex.

FTIR spectroscopy allows us to directly characterize the nucleic acid structures and in particular the sugar conformations in double and triple helices using marker bands. Thus S-type sugars are detected by the existence of an absorption band around 841 cm-1, while N-type sugars are characterized by absorption around 860 cm-1. As previously mentioned in the litterature (12 ,36 ), the duplex under high relative humidity conditions adopts a B-form geometry with S-type sugars, characterized in the infrared spectrum by absorption at 841 cm-1. Dehydration of the sample does not induce a B -> A conformational change, which would be reflected in the FTIR spectra by emergence of a band around 860 cm-1 characteristic of N-type sugars. Instead, only a broadening of the 841 cm-1 band is observed at very low RH, previously interpreted as reflecting a heteronomous conformation of the duplex under these low hydration conditions. The spectrum of the dTn·dAn×dTn triplex recorded at high RH shows that the sugars here again are in a S-type conformation (band at 841 cm-1). However, when hydration is decreased a new contribution appears around 860 cm-1, indicative of the emergence of N-type sugars. The plot presented in Figure 4 shows the emergence of N-type sugars in the triplex when hydration is decreased. The existence of N-type sugars is clearly detected below 75% RH. The presence of the band at 1281 cm-1 characteristic of thymidines with S-type sugars in the triplex spectra (not shown) at relative humidities down to 58% RH allows us to propose that under these intermediate hydration conditions the emergence of N-type sugars can be correlated with a geometrical reorientation of the adenosine strand in the triplex. The hydration at which N-type sugars begin to be observed corresponds to the value below which water desorption in the triplex becomes weaker than in the duplex, as can be judged by the crossover of the hydration curves presented in Figure 3 , right. We can notice that all native DNAs as well as double-stranded oligo- and polynucleotides except dTn·dAn sequences undergo a S-type -> N-type sugar geometry transition in films when the relative humidity is decreased (37 ). Such a transition has also been obtained by interaction of duplexes with intercalating drugs (15 ). In the case of triple helices N-type sugars have been observed even in solution for dGn×dGn·dCn (38 ) or induced by intercalating propidium iodide in intermolecular triplexes containing all four bases (39 ).


Figure 4. Emergence of N-type sugars in the dTn·dAn×dTn triplex with decreasing hydration followed by variation of the ratio of absorptions characteristic of N- and S-type sugars.

Calculations

Water molecule distribution around the dT10·dA10 Watson-Crick duplex. Calculation of the cylindrical distribution of water molecules with respect to the global helical axis of the DNA molecule is well adapted to study the hydration pattern of a short DNA fragment (Fig. 5 ). The cylindrical distribution is calculated with respect to the main axis of the phosphate group cylinder over the whole dynamics process. The cylindrical distribution of the water molecules for the double helix exhibits a peak with two shoulders (Fig. 5 a), the first shoulder located at 5.3 Å from the DNA helical axis, corresponding to hydration of the base pairs, the second occurring at ~9.8 Å, corresponding to sugar hydration, and the peak situated at 12.5 Å, reflecting phosphodiester backbone hydration. The integration of this distribution from 0 to 14.1 Å (end of the last peak) gives 20 water molecules/nucleotide, which is consistent with gravimetric measurements. The angular distribution of water molecules inside the DNA cylinder is calculated with helicoidal coordinates (Fig. 5 b). This distribution gives access to hydration of the different DNA grooves. To perform this calculation we have considered nine elementary cells constituted by two successive base pairs. For each cell the center of the local cylinder formed by the phosphate groups is calculated. For each water molecule inside the local cylinder the angle between the oxygen atom of the water molecule, the C1' atom of the closest thymidine and the center of the cell is computed. We then calculate the angular distribution of water molecules, i.e. the number of molecules per unit angle. This procedure performed over all the cells and for all the molecules inside the grooves and averaged over the whole dynamics allows discrimination of the statistical hydration of the different grooves. Figure 5 b shows two peaks of roughly the same intensity. The first thin peak having a maximum located at 60° is representative of minor groove hydration. The second broad peak is assigned to major groove hydration and its maximum is found at 250°. As the minor and major grooves are equally deep, these two peaks are expected to have the same intensity and any difference in the broadness of the two peaks should be directly correlated with the difference observed in the groove widths. The analysis of the water distribution indicates that the major groove is more hydrated than the minor groove.(a) Water molecules were counted up to 4.0 Å from atoms. The average was performed over the whole phase production of the dynamics and the equivalent atoms. Standard deviations are in parentheses.(b) Hydration of phosphate groups were computed as the average of the hydration of the two oxygen atoms.


Table 1 . (a) Hydration per atomic site on the two strands of the free Watson-Crick double helix; (b) hydration per atomic site on the three strands of the triple helix


Figure 5. (a) Cylindrical distribution of the water oxygen atoms with respect to the helical axis averaged over the whole dynamics for the free dT10·dA10 double helix calculated over nine adjacent symmetrical boxes. The water molecules that do not lie between the mean planes of the upper and lower base pairs are discarded. (b) Angular distribution of the water oxygen atoms inside the grooves over the whole dynamics for the free dT10·dA10 double helix. The y-axis represents the number of water molecules. (c) Cylindrical distribution of the water oxygen atoms with respect to the helical axis averaged over the whole dynamics for the dT10·dA10×dT10 triple helix. (d) Angular distribution of water molecules inside the grooves for the triple helix.Intra-strand water bridges: each entry represents a water molecule that forms hydrogen bonds with two atoms belonging to the same strand.Inter-strand water bridges in the minor groove and major groove of the free Watson-Crick double helix: each entry represents a water molecule that forms hydrogen bonds with two atoms of different strands.The names, the numbers of residues and the names of the atoms involved in the water bridge are given. The values represent the probability of occurrence (percent) of the water bridge. When several numbers are given on the same line this indicates that different water molecules occupy the site at different moments of the dynamics. The same water molecule can be involved in different bridges at different times. The cut-off for hydrogen bonds (distance between H and the acceptor atom) is 2.5 Å. Only water bridges having a probability of occurrence >4% are shown.


Table 2 . Water bridges for the dT10·dA10 Watson-Crick duplex

Hydration of the dT10·dA10 Watson-Crick duplex per atomic site. We now examine hydration of the DNA double helix in terms of number of water-nucleic acid contacts quantitatively (Table 1 a). We consider the average number of water molecules per atomic site and per nucleoside inside a sphere of 4.00 Å radius. A water molecule can belong to two different close sites because of overlap of the spheres. N3 atoms of thymines and N1 atoms of adenines are poorly hydrated in spite of their hydrophilicity, because both of them participate in central base pairing hydrogen bonds and are therefore fully screened from contact with water molecules. Adenine N9 and thymine N1 atoms, forming with the C1' sugar atoms the glycosidic bond, are also poorly hydrated, because they are blocked by the adjacent sugar moiety. O2 atoms of thymines and N3 atoms of adenines located inside the minor groove are well hydrated, however, they have less water contacts than methyl carbons of thymines and N6, N7 atoms of adenines belonging to the major groove, even if the methyl group is hydrophobic. The two strands present a sugar phosphate backbone of roughly the same hydration. These conclusions are consistent with analysis of the angular water molecule distributions.

Water bridges in the dT10·dA10 double helix. Water bridges have been reported to have a great influence on the stability of DNA (15 ). Table 2 shows intra-strand water bridges involving mainly the bases. The cut-off for the hydrogen bonds (distance between the hydrogen and the acceptor atom) is taken at 2.5 Å. We have presented only water bridges having an occurrence >4% of the whole 600 ps phase production of the dynamics (>24 ps). In a previous work we have shown that diffusion motion of water molecules in the vicinity of DNA complexes was statistically smaller than that of bulk water molecules (34 ). So most of the water bridges having a large probability of occurrence from our mathematical procedure remain in the vicinity of the DNA core during the whole dynamics. The adenosine strand presents a water bridge between the N7 atom of an adenine, the hydrogen of a water molecule and the H1N6 atom of the next adenine with the oxygen atom of the same water molecule, inside the major groove, at each step. The water bridge having the longest time of occurrence takes place between adenine 13 and adenine 14 and is present during 95.47% of the whole dynamics. Other water bridges link N7 atoms of adjacent adenines. With regard to the minor groove we notice the presence of water bridges between the N3 atom and the O4' atom between two adjacent bases. Concerning the thymidine strand, most intra-strand water bridges involve the O2 atom of the bases, but they present a probability of occurrence smaller than those of the adenosine strand. The most important intra-strand water bridges involving adenines are mainly located inside the major groove, while those involving thymines are mainly situated inside the minor groove.

Inter-strand water bridges are also presented in Table 2 . We show the probability of occurrence of a water bridge between O2 atoms of thymines and N3 atoms of adenines as previously reported for an ordered water pattern in the minor groove involving AT pairs and corresponding to a `spine of hydration' for the dodecamer d(CGCGAATTCGCG)2 (40 ). These water molecules do not only remain between these two sites and the low probabilities of occurrence reflect their fluctuations inside the minor groove between different atomic sites. Few inter-strand water bridges are observed inside the major groove, which mainly take place due to exchange phenomena between different atomic sites. These results show that water bridges are essentially a dynamic phenomenon.


Table 3 . Intra-strand and inter-strand water bridges for the triple helix dT10·dA10×dT10

Top: water bridges in the minor groove of the Watson-Crick duplex; bottom: water bridges involving the third strand. The calculation protocol was as in Table 2.

Water molecule distribution of the dT10·dA10×dT10 triple helix. Figure 5 c shows the cylindrical distribution of water molecules in the dT10·dA10×dT10 triple helix. Slide parameters of the computed duplex in the triplex present significant negative values (data not shown), reflecting a displacement of successive base pairs in the direction of the adenine strand. Negative slide values have two consequences on the double helix structure in the triplex: first, deepening of the major groove, allowing it to accomodate a third strand; second, shift of the helical axis from the center of the Watson-Crick base pairs towards the major groove, thus widening the helix. As a consequence, hydration of the triple helix begins at 0 Å from the helical axis and a first peak appears before 5 Å. A second peak located between 9.7 and 10.4 Å is assigned to sugar solvation and a third peak around 14.2 Å represents mainly phosphodiester backbone hydration. Integration of this curve from 0 to 15.2 Å (end of the last peak) gives a total of 15 molecules/nucleotide. The triple helix is less solvated than the free Watson-Crick double helix by ~5 water molecules/nucleotide and the ratio double helix/triple helix hydration is 1.3, close to the ratio obtained by gravimetric and FTIR measurements. However, the triplex is more hydrated than an A-form double helix, for which a value of 10.5 water molecules/nucleotide is classically assumed (15 ). Figure 5 d shows the angular distribution of water molecules inside the grooves of the triple helix. Two peaks are observed, the first, corresponding to minor groove hydration, occurs at 40°, the second, representing mainly hydration of the major groove, appears at 270°. Furthermore, a residual peak is observed at ~150°, reflecting hydration of sugars and partially of phosphates belonging to the third strand. The shift of the peak representing solvation of the major groove from 250° for the free duplex to 270° for the triple helix clearly shows displacement of water molecules consecutively due to the presence of the third strand inside the major groove (Fig. 5 b and d). Integration of the angular distribution peaks corresponding to hydration of the minor groove in the free duplex and in the duplex involved in the triplex shows that no dehydration effects are observed inside the minor groove.

Hydration of the dT10·dA10×dT10 triple helix per atomic site. N6 and N7 atoms of the adenosine strand are, as expected, less hydrated in the triple helix because they are screened by the third thymidine strand, as they are involved in Hoogsteen hydrogen bonds. N7 atoms are fully desolvated, while the N6 site is only partially dehydrated (compare Table 1 a and b). O2 atoms of the thymidine third strand are less solvated than O2 atoms belonging to the thymidine strand of the Watson-Crick double helix, as they are screened by the adenosine strand atoms and also because the adenosine strand and the thymidine third strand are very close. O2 atoms of the thymidine strand involved in the Watson-Crick double helix and N3 atoms of the adenosine strand are not desolvated, which indicates that hydration of the minor groove is preserved when triplex formation occurs. O4' atoms of the third strand are less hydrated than O4' atoms belonging to the Watson-Crick duplex. Atoms of the sugar-phosphate backbone of the adenosine strand exhibit small but significant dehydration when compared with the adenosine strand of the free Watson-Crick duplex. Water bridges in the dT10·dA10×dT10 triple helix. The thymidine strand of the Watson-Crick duplex presents the same water bridges as in the free double helix. Intra-strand water brigdes still occur between O2 and O4' atoms (Table 3 ). The adenosine strand no longer exhibits water bridges between N7 and H1N6 sites, as the N7 sites of the adenosine strand are involved in Hoogsteen hydrogen bonds. Fewer intra-strand water bridges are observed for the adenosine strand. We have also detected intra-strand water bridges for the thymidine third strand. They mainly involve sugar phosphate atoms.

The probability of occurrence of inter-strand water bridges involving O2 atoms of thymines and N3 atoms of adenines inside the minor groove of the Watson-Crick duplex are also shown in Table 3 . We have checked that the water molecules fluctuate inside the minor groove during the whole dynamics, forming the so called `spine of hydration' of the minor groove (Fig. 6 ).


Figure 6. (Left) Spine of hydration in the minor groove of the Watson-Crick duplex involved in the triple helix. (Right) Spine of hydration in the major groove between the thymidine strands of the triplex. Red, water molecules; pink, adenine strand; light green, thymine first strand; blue, thymine third strand.

We now focus on inter-strand water bridges involving the third strand and the two other strands (Table 3 ). Water bridges between the thymidine third strand and the adenosine strand having a large time of occurrence were observed between the O3' sugar atoms and oxygen atoms of phosphate groups respectively. Such water bridges should contribute to stabilization of the whole triplex. Furthermore, we have found a new line of water molecules bound between the O4 atom of thymine of the third strand and the O4 atom of the thymine of the first strand for each base plane inside the major groove (Fig. 6 , right). The probability of occurrence of these water bridges is high and we have checked that these molecules remain in the vicinity of the O4 atoms during simulation. Only small fluctuations occur for these molecules when compared with those of bulk water molecules. These molecules are highly organized and play a significant role in stablilization of the triplex. They form a new spine of water molecules inside the major groove. The computed probabilities of occurrence of these water molecules correspond to residence times between 0.2 and 0.5 ns, compatible with recent NMR data concerning water molecules in the minor groove of the AATT region of the d(CGCGAATTCGCG) duplex (41 ).

CONCLUSION

We have studied hydration of a dT10·dA10×dT10 triple helix and of the free dT10·dA10 Watson-Crick double helix. We have also examined the desorption phenomena of both double and triple helices with decreasing RH. A good agreement between gravimetric and FTIR measurements has been obtained; a relation between water desorption and structural changes in the triple helix has been proposed to interpret the crossover occuring for desorption curves obtained both by FTIR and gravimetric measurements. We have then correlated the calculated structures with experimental data obtained at 98% RH so as to validate our molecular dynamics simulations. We have shown and quantified dehydration due to binding of the thymidine third strand by coupling molecular simulations, infrared spectroscopy and gravimetric measurements. We have obtained a total of 21 ± 1 water molecules/nucleotide in the case of the free duplex, while only 17 ± 2 water molecules/nucleotide were found for the triple helix. For the triple helix a spine of water molecules bound to the O4 atoms of thymines of both the first and third strands has been determined at each base plane.

REFERENCES

1 Letai,A.G., Palladino,M.A., Fromm,E., Rizzo,V. and Fresco,J.R.(1988) Biochemistry, 27, 9108-91112. MEDLINE Abstract

2 Hélène,C. and Toulmé,J.J. (1990) Biochim. Biophys Acta, 1049, 99-125. MEDLINE Abstract

3 Sun,J.S. and Hélène,C. (1993) Curr. Opin. Struct. Biol., 3, 345-356.

4 Beal,P.A. and Dervan,P.B. (1991) Science, 251, 1360-1363. MEDLINE Abstract

5 Uhlman,E. and Peyman,A. (1990) Chem. Rev., 90, 543-584.

6 Tso,P.O.P., Aurelian,L., Cheng,E. and Miller,P.S.(1992) Annls NY Acad. Sci., 660, 156-177.

7 Giovannageli,C., Thuong,N.T. and Hélène,C. (1993) Proc. Natl. Acad. Sci. USA, 90, 10013-10017.

8 Arnott,S. and Bond,P.J. (1973) Nature, 244, 99-101.

9 Arnott,S. and Selsing,E.J. (1974) J. Mol. Biol., 88, 509-521. MEDLINE Abstract

10 Arnott,S., Bond,P.J., Selsing,E. and Smith,P.J.C. (1976) Nucleic Acids Res., 3, 2459-2470. MEDLINE Abstract

11 Thomas,G.A. and Peticolas,W.L. (1983) J. Am. Chem. Soc., 105, 993-996.

12 Liquier,J., Coffinier,P., Firon,M. and Taillandier,E. (1991) J. Biomol. Struct. Dyn., 3, 671-687.

13 Howard,F.B., Miles,H.T., Liu,K., Frazier,J., Raghunathan,G. and Sasisekharan,V. (1992) Biochemistry, 31, 10671-10677. MEDLINE Abstract

14 Ouali,M., Letellier,R., Adnet,F., Liquier,J., Sun,J.S., Lavery,R. and Taillandier,E. (1993) Biochemistry, 32, 2098-2103. MEDLINE Abstract

15 Saenger,W. (1983) Principles of Nucleic Acid Structure. Springer-Verlag, New York, NY.

16 Saenger,W. (1987) Annu. Rev. Biophys. Chem., 16, 93-114.

17 Berman,H.M. (1991) Curr. Opin. Struct. Biol., 1, 423-427.

18 Falk,M., Hartmann,K.A.,Jr and Lord,R.C. (1962) J. Am. Chem. Soc., 84, 3843-3846.

19 Falk,M., Hartmann,K.A.,Jr and Lord,R.C. (1963) J. Am. Chem. Soc., 85, 387-394.

20 Tao,N.J., Lindsay,S.M. and Rupprecht,A. (1989) Biopolymers, 28, 1019-1030. MEDLINE Abstract

21 White,A.P., Reeves,K.K., Snyder,E., Farrel,J. and Powell,J.W. (1996) Nucleic Acids Res., 24, 3261-3266. MEDLINE Abstract

22 Mohan,V., Smith,P.E. and Pettitt,M.B. (1993) J. Am. Chem. Soc., 115, 9297-9298.

23 Mohan,V., Smith,P.E. and Pettitt,M.B. (1993) J. Phys. Chem. Soc., 97, 12984-12990.

24 Weerasinghe,S., Smith,P.E., Mohan,V., Cheng,Y.K. and Pettitt,B.M. (1995) J. Am. Chem. Soc., 117, 2147-2158.

25 Le Bret,M., Gabarro-Arpa,J., Gilbert,J.C. and Lemaréchal,C. (1991) J. Chim Phys. Phys.-Chim. Biol., 88, 2489-2496.

26 Arnott,S., Chandrasekharan,R., Birdsall,R.D.L., Leslie,A.G.W. and Ratliff,R.L. (1980) Nature, 283, 743-745. MEDLINE Abstract

27 Cornell,W.D., Cieplak,P., Bayly,C.I., Gould,I.R., Merz,K.M.,Jr, Ferguson,D.M., Spellmeyer,D.C., Fox,T., Caldwell,J.W. and Kollman,P.A. (1995) J. Am. Chem. Soc., 117, 5179-5197.

28 Hingerty,B., Richie,R.H., Ferrel,T.L. and Turner,J.E. (1985) Biopolymer, 24, 427-439.

29 Lavery,R., Sklenar,H., Zakrzwska,K. and Pullman,B.J. (1986) J. Biomol. Struct. Dyn., 3, 986-1014.

30 Pearlman,D.A., Case,D.A., Caldwell,J.C., Seibel,J.L., Singh,U.C., Weiner,P. and Kollman,P.A. (1991) AMBER 4.0, University of California, San Francisco, CA. MEDLINE Abstract

31 Jorgensen,W.L., Chandrasekhar,J., Madura,J., Impeyer,R.W. and Klein,M.L. (1983) J. Chem. Phys., 79, 926-935.

32 Verlet,L. (1967) Phys. Rev., 159, 98-103.

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

34 Ouali,M., Bouziane,M., Ketterlé,C., Gabarro-Arpa,J., Auclair,C. and Le Bret,M. (1996) J. Biomol. Struct. Dyn., 13, 835-853. MEDLINE Abstract

35 Ketterlé,C., Gabarro-Arpa,J., Ouali,M., Bouziane,M., Auclair,C., Helissey,P., Giorgi-Renault,S. and Le Bret,M. (1996) J. Biomol. Struct. Dyn., 13, 963-977. MEDLINE Abstract

36 Pilet,J., Blicharski,J. and Brahms,J. (1975) J. Am. Chem. Soc., 14, 1869-1875.

37 Liquier,J. and Taillandier,E. (1996) In Mantsch,H.H. and Chapman,D. (eds), Infrared Spectroscopy of Biomolecules. John Wiley & Sons, Chichester, UK, pp. 131-158.

38 Ouali,M., Letellier,R., Akhebat,A., Liquier,J., Adnet,F. and Taillandier,E. (1993) J. Am. Chem. Soc., 115, 4264-4270.

39 Dagneaux,C., Shchyolkina,A.K., Florentiev,V.L., Liquier,J. and Taillandier,E. (1995) C.R. Acad. Sci. Paris, 318, 559-562.

40 Kopka,M., Fratini,A.V., Drew,H.R. and Dickerson,R.E. (1983) J. Mol. Biol., 163, 129-146. MEDLINE Abstract

41 Denisov,V.P., Carlstrom,G., Venu,K. and Halle,B. (1997) J. Mol. Biol., 268, 118-136. MEDLINE Abstract

*To whom correspondance should be addressed. Tel: +33 1 48 38 76 90; Fax: +33 1 48 37 74 43; Email: etaill@sb.univ-paris13.fr
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow Print PDF (874K) Freely available
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 (8)
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Ouali, M.
Right arrow Articles by Taillandier, E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ouali, M.
Right arrow Articles by Taillandier, E.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?