Silver(I) Coordination in Silver(I)-Mediated Homo Base Pairs of 6-Pyrazolylpurine in DNA Duplexes Involves the Watson-Crick Edge.

DNA duplexes comprising 6-(1H-pyrazol-1-yl)-9H-purine (6PP), 1-deaza-6PP (1D 6PP), 7-deaza-6PP (7D 6PP) and 1,7-dideaza-6PP (1,7D 6PP) 2'-deoxyribonucleosides, respectively, were investigated towards their ability to form metal-mediated base pairs in the presence of AgI . In 6PP and 7D 6PP, the AgI ion can coordinate to the nucleobase via the endocyclic N1 nitrogen atom, that is, via the Watson-Crick edge. In contrast, this nitrogen atom is not available in 1D 6PP and 1,7D 6PP, so that in 1D 6PP an AgI coordination is only possible via the Hoogsteen edge (N7). Reference duplexes with either adenine:adenine mispairs or canonical adenine:thymine base pairs were used to investigate the impact of the pyrazolyl moiety on the AgI -binding properties. To determine the thermal and structural duplex stabilities in the absence or presence of AgI , all duplexes were examined by UV and circular dichroism spectroscopic studies. These investigations shed light on the question of whether N1- or N7-coordination is preferred in purine-based metal-mediated base pairs.

Introduction

In 1963, a base pairing pattern for canonical nucleobases in double‐stranded DNA was proposed that is different from the one originally put forward by Watson and Crick. While Watson and Crick suggested the formation of base pairs via the N1 nitrogen atom of the purine bases in DNA, Hoogsteen crystallized a complex of 9‐methyladenine and 1‐methylthymine in which the N7 nitrogen atom of the adenine derivative was involved in the formation of the A:T base pair. Even though it is nowadays accepted that canonical base pairing occurs almost exclusively via the Watson–Crick edge in antiparallel‐stranded DNA duplexes, the Hoogsteen edge can be of relevance when artificial nucleosides are involved. As artificial nucleobases are not restricted to the geometry and binding sites of their natural counterparts, formation of base pairs may involve hydrogen bonding, hydrophobic interactions in combination with shape complementarity or coordination of metal ions. In the latter case, hydrogen bonds are formally replaced by coordinate bonds to metal ions resulting in the formation of metal‐mediated base pairs. DNA duplexes comprising artificial ligand‐derived nucleosides are typically destabilized in the absence of suitable transition metal ions compared to canonical DNA duplexes, because their arrangement of Lewis‐basic donor sites is optimized for the formation of metal complexes rather than for hydrogen bonds.[ , ] However, upon the addition of suitable metal ions, the melting temperature of these DNA duplexes increases, indicating a higher thermal stability due to the formation of metal‐mediated base pairs. The respective degree of stabilization depends on the chosen nucleobase and hence on the type of coordinating metal ion. Several applications have been put forward for DNA with metal‐mediated base pairs, including metal‐responsive structural and catalytic transformations, sensors for metal ions or oligonucleotides, modulation of the charge‐transfer capabilities of DNA, generation of metal nanoclusters, and several more. While even the naturally occurring nucleobases are capable of forming metal‐mediated base pairs, the majority of published metal‐mediated base pairs contains artificial nucleosides, based on monodentate, bidentate,[ 15] tridentate or even tetradentate ligands. Depending on the identity of the artificial nucleobase, up to three metal ions can be introduced into a single metal‐mediated base pair. Many of the artificial nucleobases are derivatives of purine and pyrimidine. In the context of this work, purine‐derived ligands are of particular interest. Most of these derivatives bear an additional metal‐binding entity at their C2 and/ or their C6 position,[ , ] while others are guanine and adenine derivatives with formally exchanged N/C−H positions.[ 20]

6‐(1H‐Pyrazol‐1‐yl)‐9H‐purine (6PP) is an adenine‐derived artificial nucleobase comprising a pyrazolyl moiety instead of the exocyclic amine group (Scheme 1), that is, bearing an additional metal‐binding moiety at its N6 position. Because of the free rotatability around the C6‐N1* bond, the nitrogen atom of the pyrazolyl substituent can either face the Watson–Crick or the Hoogsteen edge. Prior studies of metal complexes involving the N9‐methylated 6PP model nucleobase Me6PP and either AgI or CuII have shown that coordination via the Hoogsteen edge (N7) and the pyrazole N2* nitrogen atom is favored in the solid state. This is in good agreement with earlier studies on the canonical purine bases, indicating that metal ions preferentially bind to the N7 atom. Similarly, the N7 position was proposed as the most likely CuII‐binding site in oligonucleotides carrying an artificial nucleoside with a 6‐(1H‐3,5‐dimethylpyrazol‐1‐yl)‐9H‐purine nucleobase. On the other hand, taking into consideration the reports on different geometries of the A:T pair in a DNA duplex (by Watson and Crick) and in a model nucleobase complex (by Hoogsteen), it is possible that the preferred binding mode of 6PP in a DNA duplex differs from that found in the Me6PP‐M‐Me6PP model nucleobase complexes (M=AgI, CuII).

Scheme 1

The artificial nucleobase 6‐(1H‐pyrazol‐1‐yl)‐9H‐purine (6PP) including an atom‐numbering scheme using the purine nomenclature. Metal‐binding sites for the formation of metal‐mediated base pairs are indicated in bold.

To identify the AgI‐binding pattern in a 6PP‐AgI‐6PP base pair inside a DNA duplex, a selection of 6PP‐derived 2′‐deoxyribonucleosides was incorporated into antiparallel‐stranded DNA duplexes bearing three adjacent artificial homo base pairs in the center of the duplex. In addition to the 6PP deoxyribonucleoside, this set includes the 1‐deaza‐6PP (1D6PP), 7‐deaza‐6PP (7D6PP) and 1,7‐dideaza‐6PP (1,7D6PP) deoxyribonucleosides. While the absence of the N7 nitrogen atom in 7D6PP should favor AgI coordination via the Watson–Crick edge (i.e. via N1) (Scheme 2 a), 1D6PP lacks the N1 nitrogen atom but possesses an N7 nitrogen atom, allowing an AgI‐mediated base pair formation via the Hoogsteen edge (Scheme 2 b). It is important to note that for metal‐mediated base pair formation via the Watson–Crick edge the nucleosides retain their normal anti orientation of nucleobase and deoxyribose moiety (Scheme 2 a). In contrast, the less common syn orientation must be adopted to enable metal‐mediated base pairing via the Hoogsteen edge inside an antiparallel‐stranded double helix (Scheme 2 b). Nevertheless, several examples have been reported in which such a syn orientation is present in metal‐mediated base pairs.[ , , ] In addition to duplexes bearing 6PP, 1D6PP or 7D6PP, DNA duplexes comprising either 1, 7D6PP or A deoxyribonucleosides were studied to determine the relevance of the pyrazolyl moiety on AgI‐mediated base pair formation. All DNA duplexes were investigated by temperature‐dependent UV spectroscopy and circular dichroism (CD) spectroscopy to compare their AgI‐binding behavior and its influence on thermal stability and duplex structure.

Scheme 2

Possible binding patterns of 6PP‐AgI‐6PP base pairs via the Watson–Crick edge (a) or the Hoogsteen edge (b). The nitrogen atoms not involved in AgI binding are depicted in grey to emphasize the expected pattern for a DNA duplex with either 7D6PP or 1D6PP nucleosides. The terms anti and syn refer to the relative orientation of nucleobase and deoxyribose.

Results and Discussion

The general sequence of the DNA duplexes under investigation is depicted in Scheme 3. Six duplexes were selected, including either homo base pairs (X=Y) of 6PP (I), 1D6PP (II), 7D6PP (III), 1, 7D6PP (IV) and adenine (V) or central A:T pairs (VI), serving as a reference. All duplexes contain three consecutive metal‐binding sites because initial tests had indicated that these sequences can be significantly stabilized by the addition of AgI.

Scheme 3

DNA duplexes investigated in this study. For definition of X and Y, see text.

To evaluate the thermal stability of the DNA duplexes, their melting temperatures were determined UV‐spectroscopically in the absence of AgI and after the addition of up to 6 μm AgI, corresponding to two AgI ions per artificial base pair in duplexes I–V. The AgI concentration was increased in steps of 1 μm between 0 μm and 3 μm of AgI to evaluate possible differences in binding affinity. An exemplary melting curve is given in Figure 1, showing the data for duplex I. The melting curves of all other duplexes are shown in the Supporting information (Figure S2). Melting temperatures (T m) were derived from each melting curve and plotted against the respective concentration of AgI (Figure 2). The T m values for 0 equiv (0 μm), 1 equiv (3 μm) and 2 equiv (6 μm) of AgI as well as the increase in T m (ΔT m) between 0 and 1 equiv of AgI are summarized in Table 1.

Figure 1

Melting curves of DNA duplex I with three central 6PP:6PP pairs in the presence of 0 μm (red), 1 μm (orange), 2 μm (yellow), 3 μm (green), 4 μm (blue) and 6 μm (purple) of AgI. 3 μm AgI≙1 equiv of AgI with respect to the number of 6PP:6PP pairs. Experimental conditions: 1 μm DNA duplex, 5 mm MOPS buffer (pH 6.8), 150 mm NaClO4.

Figure 2

Plot of the melting temperatures of duplexes I (a), II (b), III (c), IV (d), V (e) and VI (f) with linear fit (solid line), hypothetical linear fit between 0 μm and 3 μm AgI (broken line) or asymptotic fit (dotted line). The hypothetical linear fit indicates the anticipated course of data points for three high‐affinity binding sites.The asymptotic fit resembles the course of data points for an unspecified number of low‐affinity binding sites. 3 μm AgI≙1 equiv of AgI corresponds to the concentration required to introduce one AgI ion into each designated AgI‐binding site. For clarity, the respective artificial nucleobase is shown, too. In Figure 2 f, no nucleobase is indicated, because this is the reference measurement with three central A:T base pairs. Experimental conditions: 1 μm DNA duplex, 5 mm MOPS buffer (pH 6.8), 150 mm NaClO4.

Table 1

Melting temperatures T m for duplexes I‐VI.[a]

Duplex	T m [°C] 0 μm AgI	T m [°C] 3 μm AgI	T m [°C] 6 μm AgI	ΔT m [°C] 0→3 μm AgI
I	26.7(1)	37.7(4)	40.1(4)	11.0(4)
II	25.1(2)	32.6(3)	33.6(5)	7.5(4)
III	25.9(1)	42.4(5)	46.0(8)	16.5(5)
IV	25.0(2)	32.8(3)	33.7(6)	7.8(4)
V	16.6(2)	22.0(4)	–	5.4(4)
VI	42.4(1)	47.1(1)	47.6(1)	4.7(1)

[a] Standard deviation given in parenthesis. Values obtained by fitting the derivative of the melting curve with a Gauss function, considering a confidence interval of 95 % (for T m), or by using error propagation (for ΔT m).

–

As expected, duplex VI bearing only canonical base pairs reveals the highest melting temperature with a T m of 42.4 °C in the absence of AgI. Interestingly, all duplexes comprising 6PP‐derived nucleosides (I–IV) have almost the same thermal stability in the absence of AgI (T m=25.0–26.7 °C), indicating that the absence or presence of the endocyclic nitrogen atoms N1 and N7 plays a minor role in the stabilization of the DNA duplex. However, the absence of the pyrazolyl moiety has a significant impact on stability. Duplex V with its central A:A mispairs is further destabilized by 10 °C compared to duplex I, even though both 6PP (duplex I) and adenine (duplex V) contain N1 and N7 atoms. Apparently, the aromatic pyrazole ring introduces additional stability. Upon the addition of 3 μm AgI, all duplexes are thermally stabilized, including reference duplex VI. As this duplex does not contain any designated AgI binding sites, AgI may interact with the donor atoms of the natural nucleobases,[ , ] leading to a slightly increased melting temperature (ΔT m=4.7 °C). The same behavior is also observed for duplex V (ΔT m=5.4 °C). In both cases, the plot of melting temperature vs. amount of AgI (Figure 2 e,f) shows an asymptotic behavior, indicating a small binding constant.

Interestingly, it is possible to distinguish between duplexes I–IV based on the plots shown in Figure 2 with respect to the absence (duplexes II (1D6PP, Figure 2 b) and IV (1,7D6PP, Figure 2 d)) or presence (duplexes I (6PP, Figure 2 a) and III (7D6PP, Figure 2 c)) of an endocyclic N1 atom. For the former, not only the differences in thermal stabilization upon AgI‐binding are the same (II: 7.5 °C; IV: 7.8 °C), but also the absolute melting temperature in the absence of AgI (II: 25.1 °C; IV: 25.0 °C) and in the presence of 3 μm AgI (II: 32.6 °C; IV: 32.8 °C). As the 6PP‐derived nucleosides in duplexes II and IV do not contain any N1 nitrogen atoms, possible AgI‐mediated base pairs cannot involve their Watson–Crick edge. Nevertheless, their pyrazolyl moiety is likely to play a role in coordinating the AgI ions, because the increase in T m upon the addition of AgI is slightly higher for II and IV compared to V and VI. For these four duplexes (II, IV, V and VI), the plots of melting temperature vs. amount of AgI can be fitted best in an asymptotic manner, which is representative of a low binding affinity. However, the increase in T m upon binding of AgI is larger for those duplexes containing an artificial nucleobase with the pyrazolyl moiety (II, IV), indicating the relevance of this substituent in AgI binding. Here, a potential low‐affinity binding site could be provided by two consecutive artificial base pairs coordinating one AgI ion in‐between them. This possibility is currently being explored in our laboratory, using optimized oligonucleotide sequences.

The largest increase in thermal stability is observed for duplex III (ΔT m=16.5 °C), and duplex I is also significantly stabilized (ΔT m=11.0 °C). The plots of T m vs. amount of AgI are quite similar for these two duplexes (Figure 2 a, c). Here, the melting temperature increases linearly with increasing concentrations of AgI up to the addition of 3 μm AgI, which corresponds to the presence of one AgI ion per designated AgI‐mediated base pair. Excess AgI leads to a much smaller additional increase in T m. In combination, these observations clearly indicate the formation of stable AgI‐mediated base pairs. Moreover, as duplex III contains 7D6PP moieties lacking the N7 atom, these AgI‐mediated base pairs must involve the Watson–Crick edge with its N1 donor site. Interestingly, the thermal stability of duplex III bearing 7D6PP significantly exceeds that of duplex I with 6PP once the AgI‐mediated base pairs are formed. Different explanations are feasible here: The larger stabilization could be due to different electronic properties of 7D6PP compared to 6PP, leading to a higher affinity towards AgI ions. Alternatively, the 7D6PP nucleosides may fit better into the base stack compared to 6PP. However, a final conclusion cannot be drawn, because a determination of the pK a values by pD‐dependent 1H NMR spectroscopy does not indicate significantly different basicities of the 6PP and 7D6PP deoxyribonucleosides (Figure S1, Supporting information). Similarly, the second possible explanation is unlikely, considering that canonical DNA with A:T pairs was reported to be more stable than DNA with 7DA:T base pairs.

The addition of more than one equivalent of AgI (3 μm) results in a minor additional increase in T m for all six duplexes when compared to the stabilizing effect of the first equivalent of AgI. For duplexes II, IV and VI, this additional increase amounts to ca. 1 °C, while it is slightly higher for duplexes I (2.4 °C) and III (3.6 °C). Similar to what is observed for the interaction of canonical duplexes with AgI, it is likely that once all designated AgI‐binding sites are saturated with AgI, additional interactions with the remaining natural nucleobases occur. This leads to the additional small increase in T m, observed as an asymptotic change of T m between 3 μm and 6 μm AgI (Figure 2). For duplex V, no melting temperature could be determined in the presence of 6 μm AgI because no sigmoid melting curve was observed (Figure S2e, Supporting information).

Besides the thermal stability of the DNA duplexes, a possible change of their structure upon the addition of AgI was investigated by CD spectroscopy. The CD spectra for duplexes I–VI are depicted in Figure 3. At first glance, the spectra for duplexes I and II, III and IV or V and VI look similar in the absence of AgI. This does not only reflect the helical structure but also the different UV absorbance resulting from the presence of different artificial deoxyribonucleosides. The UV spectra for duplexes I–VI (Figure S3, Supporting information) confirm the presence of additional absorbance maxima at 305 nm (for I and II) and ca. 325 nm (for III and IV) that can be assigned to the pyrazolyl moiety as seen before for other C6‐substituted purines.

Figure 3

CD spectra of duplexes I (a), II (b), III (c), IV (d), V (e) and VI (f) in the presence of 0 μm (red), 1 μm (orange), 2 μm (yellow), 3 μm (green), 4 μm (blue) and 6 μm (purple) AgI. Important changes are highlighted by arrows. 3 μm AgI≙1 equiv of AgI with respect to the number of mismatches. For clarity, the respective artificial nucleobase is shown, too. In Figure 2 f, no nucleobase is indicated, because this is the reference measurement with three central A:T base pairs. Experimental conditions: 1 μm DNA duplex, 5 mm MOPS buffer (pH 6.8), 150 mm NaClO4.

In analogy to the trends found for T m, the CD spectrum of duplex I (Figure 3 a) changes significantly up to the addition of 3 μm AgI. Already in the presence of only 1 μm AgI, the small maximum at 265 nm shifts to 272 nm and increases simultaneously. The maximum at 293 nm becomes a shoulder in the presence of 2 μm AgI, while the maximum at 305 nm assigned to the pyrazolyl moiety decreases steadily until 3 μm of AgI are present in solution, only to remain unchanged thereafter. All this indicates that three AgI ions are coordinated per duplex, causing a structural change due to the formation of three AgI‐mediated base pairs. Once all mismatches are saturated, the helical structure does not change any further. For duplex II (Figure 3 b), the same changes of the maximum at 305 nm are observed, but in contrast to duplex I, excess AgI influences the DNA structure further, leading to a continued decrease of this maximum. The same trend is followed by the maximum at 288 nm, but in contrast to the related maximum of duplex I (at 293 nm), it does not disappear. Furthermore, the negative maximum at 260 nm in the CD spectrum of duplex II steadily changes to become a minimum, while the respective maximum of duplex I at 265 nm increases and shifts to higher wavelengths as mentioned before. The fact that the CD spectra of duplexes I and II change differently upon the addition of AgI suggests that different AgI‐binding modes are adopted by the duplexes. The 1D6PP:1D6PP pairs in duplex II can bind the AgI ions only via the Hoogsteen edge, so that a binding of AgI via the Watson–Crick edge appears likely for the 6PP:6PP pairs in duplex I.

CD spectra of duplex III (containing 7D6PP, Figure 3 c) show only minor changes upon the addition of up to one equivalent of AgI and further addition does not affect the duplex structure anymore. Upon the addition of 1 μm AgI, the minimum at 246 nm decreases and the maximum at 276 nm becomes slightly more intense while the local maximum assigned to the presence of the pyrazolyl substituent shifts from 320 nm to 330 nm. As such a shift is not observed for the corresponding maximum in the CD spectrum of duplex IV (containing 1,7D6PP), it can be assumed that the formation of AgI‐mediated base pairs involving the N1 nitrogen atom leads either to a structural change of the duplex or to an electronic change of the nucleobases. Furthermore, the direction of the change of the maximum at 276 nm of duplex III is perfectly in line with that of duplex I, confirming once more that duplexes with either 6PP:6PP or 7D6PP:7D6PP pairs behave similarly upon the addition of AgI.

Even if the melting temperatures of duplexes V and VI are affected only slightly by the addition of AgI, the concomitant structural changes are highly significant. The maximum in the CD spectrum of duplex V at 275 nm decreases continuously with increasing AgI concentration. It is worth mentioning that the T m plot reaches a plateau in the presence of 2 μm AgI, while major changes in the CD spectra continue to take place even upon the addition of more than 2 μm AgI. The significantly affected structure in the presence of 6 μm AgI also explains why no sigmoid melting curve was observed anymore under these conditions. In contrast to all other duplexes, duplex VI bearing only canonical base pairs shows significant yet no continuous changes in the presence of increasing amounts of AgI. Even though the T m plot suggests a stable, almost unaltered duplex structure in the absence and presence of AgI, the CD experiments refute this assumption. After the addition of 1 μm AgI, the negative Cotton effect at 245 nm becomes less intense and the maximum at 278 nm is shifted slightly to higher wavelengths. While the presence of 2 μm AgI does not induce any further changes to the CD spectrum, the addition of 3 μm AgI leads to a further increase of the ellipticity at 245 nm along with a broadening of the minimum and the maximum. Upon the addition of more than 3 μm AgI, massive changes in the CD spectrum are observed, whereas the thermal stability studies indicate almost no change (ΔT m=0.5 °C). The broad minimum now splits into two intensive minima and the positive Cotton effect decreases in intensity. These changes clearly point out that melting temperatures and thus thermal stabilities alone do not allow the prediction of the duplex integrity. A possible explanation for the large CD spectroscopic changes is the coordination of AgI to the N7 position of the purine nucleobases, as previously proposed in other studies. Interestingly, oligonucleotides in which all purine residues are replaced by 7‐deazapurine moieties show a completely different behaviour, supporting the assumption that the CD spectroscopic changes observed for duplexes V and VI involve a coordination of AgI to N7.

Conclusions

By using a set of DNA duplexes with artificial (deaza)‐6‐pyrazolylpurine‐derived 2′‐deoxyribonucleosides, we could show that the affinity towards AgI varies depending on the provided binding sites. Strongly stabilizing AgI‐mediated base pairs are preferentially formed via the Watson–Crick edge (coordination via N1), as observed for the duplexes with 6PP and 7D6PP (I and III). In these cases, metal‐mediated base pair formation is accompanied by a structural change of the duplex, but the presence of excess AgI does not further affect the helical structure. Artificial nucleosides lacking the N1 atom as a potential metal‐binding site (1D6PP and 1,7D6PP, duplexes II and IV) form less stabilizing metal‐mediated base pairs. Moreover, the binding affinity is significantly decreased in these cases. Nevertheless, the presence of a pyrazolyl moiety increases the AgI affinity of the nucleobase compared to that of adenine, as AgI may still interact with the N2* nitrogen atom of the appended pyrazole. These duplexes therefore show a slight increase in their melting temperature in the presence of one equivalent of AgI (3 μm), accompanied by changes in the helical structure. In contrast to duplexes I and III, excess AgI induces further structural changes in duplexes II and IV. Furthermore, duplexes with A:A mispairs or canonical A:T pairs as their central base pairs were investigated. Although their thermal stability is barely affected by the presence of AgI, structural changes of the DNA helices are considerably strong. These observations emphasize that not only the presence of the N1 nitrogen atom but the synergy of the latter and the pyrazolyl moiety leads to the site‐specific incorporation of AgI ions to form DNA duplexes with strongly stabilizing AgI‐mediated base pairs. In contrast to what had previously been suggested based on single‐molecule X‐ray diffraction analysis of AgI complexes of the corresponding model nucleobases, 6PP seems to prefer AgI‐binding via its Watson–Crick edge in AgI‐mediated homo base pairs in a DNA duplex.

Experimental Section

Oligonucleotides used for duplexes I–VI were synthesized and purified as described previously. Phosphoramidites of the natural nucleobases were purchased (Glen Research) and artificial phosphoramidites were synthesized similar to the procedure published before for the 6PP phosphoramidite (see Supporting information for details). The desalted oligonucleotides were characterized by MALDI‐TOF mass spectrometry (see Supporting information, Figure S4, for details; duplex I, strand 1: calcd for [M+H]+: 4258.8 Da; found: 4258.8 Da; duplex I, strand 2: calcd for [M+H]+: 3982.7 Da; found: 3982.6 Da; duplex II, strand 1: calcd for [M+H]+: 4255.8 Da; found: 4256.0 Da; duplex II, strand 2: calcd for [M+H]+: 3981.7 Da; found: 3982.2 Da; duplex III, strand 1: calcd for [M+H]+: 4258.8 Da; found: 4258.4 Da; duplex III, strand 2: calcd for [M+H]+: 3980.7 Da; found: 3981.0 Da; duplex IV, strand 1: calcd for [M+H]+: 4251.8 Da; found: 4251.8 Da; duplex IV, strand 2: calcd for [M+H]+: 3975.7 Da; found: 3975.3 Da; duplex V/VI, strand 1: calcd for [M+H]+: 4106.8 Da; found: 4106.8 Da; duplex V, strand 2: calcd for [M+H]+: 3831.7 Da; found: 3831.1 Da; duplex VI, strand 2: calcd for [M+H]+: 3803.7 Da; found: 3804.1 Da. Oligonucleotide concentrations were determined by UV spectroscopy using the following molar extinction coefficients for the artificial nucleosides: ϵ 260(6PP)=4.8 cm2 μmol−1, ϵ 260(1D6PP)=8.6 cm2 μmol−1, ϵ 260(7D6PP)=7.0 cm2 μmol−1, ϵ 260(1, 7D6PP)=2.5 cm2 μmol−1. UV experiments were carried out on a CARY 100 Bio UV spectrometer (Agilent) and CD studies were performed using a J‐815 CD spectrometer (Jasco). All measurements were done in a quartz cuvette with 1 cm diameter at 5 °C. In case of the melting studies, absorbance at 260 nm was recorded continuously while heating from 5 °C to 70 °C. The melting temperature was determined from a Gaussian fit of the first derivative of the melting curve. The probed DNA solutions consist of 1 μm DNA duplex, 5 mm MOPS buffer (pH 6.8) and 150 mm NaClO4. AgNO3 was used for titration experiments.

Conflict of interest

The authors declare no conflict of interest.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

Click here for additional data file.