Light-Induced Formation of Thymine-Containing Mercury(II)-Mediated Base Pairs.

By applying caged thymidine residues, DNA duplexes were created in which HgII -mediated base pair formation can be triggered by irradiation with light. When a bidentate ligand was used as the complementary nucleobase, an unprecedented stepwise formation of different metal-mediated base pairs was achieved.

Metal‐mediated base pairs represent a prominent type of nucleic acid functionalization. By using ligand‐containing nucleosides, hydrogen bonds within a base pair can formally be replaced by a centrally located metal ion.1 The T‐HgII‐T (Figure 1 a) and C‐AgI‐C base pairs involving the canonical thymine (T) and cytosine (C) residues are among the best investigated metal‐mediated base pairs.1c In addition, numerous artificial nucleobases have been shown to be useful in the generation of metal‐mediated base pairs designed for particular functionalities.2 In several cases, crystal structures and solution structures have confirmed the proposed base‐pairing patterns.3 As a result of the additional metal‐based functionality, metal‐mediated base pairs have been applied in a variety of research areas, ranging from DNA charge transfer4 and oligonucleotide sensors5 to the generation of DNA‐templated metal clusters6 and switchable devices.7 Whenever a switching functionality has been introduced in the context of metal‐mediated base pairing, the switching process (mostly of DNA topology) was triggered by the addition (or removal) of a suitable metal ion. In this Communication, we report for the first time the light‐triggered formation of a metal‐mediated base pair. Several examples exist for the use of light to regulate DNA function.8 Many of these involve the application of so‐called caged nucleobases, that is, nucleobases carrying a photoremovable protecting group.9 Thymine has been one of the first nucleobases investigated in the context of photocaged nucleobases.10

Figure 1

Schematic representation of a) metal‐mediated T‐HgII‐T base pair, b) thymine residue TNPP bearing a photo‐removable protecting group, c) thymine residue TPP bearing a similar protecting group that is not removed upon irradiation.

As thymine is well‐known to coordinate to HgII ions, we decided to probe the light‐triggered HgII‐mediated base pair formation of caged thymidine. Towards this end, a caged thymidine derivative TNPP with well‐established caging properties (Figure 1 b)10 was introduced into different oligonucleotide sequences (Table 1). The duplex sequence applied in this study has previously been used in many reports on metal‐mediated base pairs, allowing a comparison with other metal‐mediated base pairs.11 Duplexes I–IV bear one central T:T mismatch, with the caged nucleobase being located either in the pyrimidine‐rich strand (duplex I), in the purine‐rich strand (duplex II) or in both strands (duplex III). Duplex IV with unprotected thymine residues serves as a reference. Similarly, duplexes V–VII contain one central T:P base pair. The artificial phenanthroline‐derived nucleoside analogue P has previously been shown to form a stable HgII‐mediated base pair with thymine, but not with cytosine.2k While the central thymine residue in duplex V bears a photocleavable protecting group, duplexes VI and VII serve as references bearing an unprotected thymidine (duplex VI) or a noncleavable substituent (duplex VII).

Table 1

DNA duplexes under investigation in the present study.[a]

Duplex		Sequence
I	ODN1	5′‐d(CTT TCT TNPPTC CCT C)
	ODN2	3‘‐d(GAA AGA TAG GGA G)
II	ODN3	5′‐d(CTT TCT TTC CCT C)
	ODN4	3‘‐d(GAA AGA TNPPAG GGA G)
III	ODN1	5′‐d(CTT TCT TNPPTC CCT C)
	ODN4	3‘‐d(GAA AGA TNPPAG GGA G)
IV	ODN3	5′‐d(CTT TCT TTC CCT C)
	ODN2	3‘‐d(GAA AGA TAG GGA G)
V	ODN1	5′‐d(CTT TCT TNPPTC CCT C)
	ODN5	3‘‐d(GAA AGA PAG GGA G)
VI	ODN3	5′‐d(CTT TCT TTC CCT C)
	ODN5	3‘‐d(GAA AGA PAG GGA G)
VII	ODN6	5′‐d(CTT TCT TPPTC CCT C)
	ODN5	3‘‐d(GAA AGA PAG GGA G)

[a] TNPP=thymidine bearing a (2‐nitrophenyl)propoxy group, TPP=thymidine bearing a phenylpropoxy group, P=(S)‐3‐(1H‐Imidazo[4,5‐f][1,10]‐phenanthrolin‐1‐yl)propane‐1,2‐diol.

The formation of coordinate bonds in a metal‐mediated base pair is typically accompanied by an increase in the DNA duplex melting temperature T m.12 Accordingly, metal‐mediated base‐pair formation was probed by temperature‐dependent UV spectroscopy. Towards this end, the melting temperature of each duplex was determined in the absence of HgII, after the addition of one equivalent of HgII prior to irradiation, in the presence of one equivalent of HgII after irradiation, and in the presence of two equivalents (i.e. excess) of HgII after irradiation. Table 2 lists the melting temperatures as derived from the temperature‐dependent UV spectra. Figure 2 exemplifies the melting curves of duplex I at pH 6.8.

Table 2

Melting temperatures T m of the DNA duplexes.[a]

Duplex	pH	T m [°C]	T m [°C]	T m [°C]	ΔT m [°C]
		0 HgII, no light	1 HgII, no light	1 HgII, irradiated	upon irradiation
I	6.8	29.4(2)	29.3(3)	46.2(4)	+16.9(5)
I	9.0	29.5(5)	28.8(5)	43.4(3)	+14.6(6)
II	6.8	32.1(4)	31.6(4)	45.7(8)	+14.1(9)
II	9.0	30.6(4)	28.6(6)	42.0(6)	+13.4(8)
III	6.8	32.8(6)	32.5(5)	43.8(6)	+11.3(8)[b]
III	9.0	29.9(6)	28.7(5)	39.6(8)	+10.9(9)
IV	6.8	35.5(2)	45.7(2)	45.8(3)	n.a.[c]
IV	9.0	32.4(2)	42.0(2)	41.9(2)	n.a.[c]
V	6.8	34.1(5)	42.0(6)	49.6(9)	+8(1)
VI	6.8	31.7(3)	48.2(4)	48.1(3)	n.a.[c]
VII	6.8	35.5(3)	40.8(8)	40(2)	±0(2)

[a] Given in parenthesis is the standard error (3σ) obtained upon fitting the derivative of the melting curve with a Gauss function (T m) or using error propagation (ΔT m). [b] Data for higher melting point of biphasic transition due to incomplete formation of the HgII‐mediated base pair. [c] Not applicable.

Figure 2

Melting curves of duplex I at pH 6.8 in the absence of HgII (black), in the presence of one equivalent of HgII prior to irradiation (red), in the presence of one equivalent of HgII after irradiation (blue) and in the presence of two equivalents of HgII after irradiation (green). Experimental conditions: 1 μm duplex, 150 mm NaClO4, 2.5 mm Mg(ClO4)2, 5 mm MOPS buffer (pH 6.8).

As can be seen, the addition of HgII does not lead to any change in T m prior to irradiation of the sample. After 1 min of irradiation of a heated sample, a significant increase in T m of ∼17 °C is observed. Initial experiments with irradiation at room temperature had resulted in a biphasic melting behavior (Figure S1a, Supporting Information), with the first melting transition coinciding with the T m of the HgII‐free duplex, suggesting an incomplete photodeprotection and hence an incomplete formation of the T‐HgII‐T pair. Several subsequent attempts to achieve a complete formation of the metal‐mediated base pair failed, including an extended irradiation time, the use of a different buffer, and the addition of HgII after the irradiation rather than prior to it (data not shown). Finally, two conditions were established that lead to a complete T‐HgII‐T formation, namely performing the irradiation at elevated temperature (ca. 50 °C) or investigating the duplex at pH 9.0 rather than pH 6.8 (Figure S1b, Supporting Information). The latter is nicely explained by previous mechanistic studies indicating a deprotonation step during photodeprotection.13 The photodeprotection of the oligonucleotides was confirmed mass spectrometrically, as shown exemplarily for ODN1 (Figure S2, Supporting Information). According to the mass spectrum, a minor amount of ODN1 remains protected even under optimized conditions. It is not clear whether this can be attributed to the absence of buffer under the conditions of mass spectrometry. If it is also present in buffer, then this amount is small enough not to be detected in the DNA melting studies.

Essentially the same behavior is found for duplex II (Figure S3, Supporting Information), in which the TNPP:T pair is formally replaced by a T:TNPP pair. This indicates that the relative position of the caged nucleobase does not significantly influence the outcome of the HgII‐mediated base pair formation. Interestingly, in duplex III bearing a TNPP:TNPP pair, the photo‐deprotection is incomplete even when heating the sample (Figure S4a, Supporting Information) or when irradiating for an extended time, indicating the relevance of steric factors during deprotection, too. This is confirmed by a mass spectrometric study (Figure S5, Supporting Information), which shows a reduced efficiency of photodeprotection of ODN1 when present in duplex III.14 Nonetheless, a complete T‐HgII‐T formation in duplex III is achieved at pH 9.0 (Figure S4b, Supporting Information). An investigation of reference duplex IV bearing a central T:T mispair shows the anticipated T‐HgII‐T base pair formation immediately after the addition of one equivalent of HgII (Figure S6, Supporting Information). Here, an irradiation of the solution is not required. In fact, irradiation does not significantly influence T m any further. For duplexes I–IV, the formation of the T‐HgII‐T pair is accompanied by a decrease in molar ellipticity [θ] at ∼280 nm (Figure S7, Supporting Information). In all four cases, the drop in [θ] occurs under those conditions that evoke an increase in T m, confirming a simultaneous deprotection and base‐pair formation.

As can be seen from Table 2, T m of duplexes I–III in the absence of HgII are decreased by 3–6 °C with respect to that of duplex IV, indicating a destabilizing effect of the bulky protecting group. After formation of the T‐HgII‐T pair, the melting temperatures of duplexes I and II are, within standard error, identical to that of duplex IV. The melting temperature of duplex III is marginally lower, which may indicate the presence of a minor fraction of still protected oligonucleotide even under the optimized photodeprotection conditions in this case.

In a T‐HgII‐T pair, the HgII ion binds both nucleobases in a monodentate fashion (Figure 1 a). Hence, the Hg−N bond involving the first thymine residue must be formed prior to the formation of the N−Hg bond to the other thymine.15 A different scenario is anticipated for the T‐HgII‐P pair (Figure 3).2k It contains the phenanthroline‐derived nucleoside analogue P that had been applied in a series of studies, including the concomitant site‐specific incorporation of AgI and HgII into the same duplex and the first enantiospecific formation of a metal‐mediated base pair.2k, 16 As P is a bidentate ligand, it is expected to be metalated first during the formation of a metal‐mediated base pair,5b irrespective of the identity of the complementary nucleobase. If the complementary nucleobase is a thymine residue, then a T‐HgII‐P base pair is formed.2k The question arises as to what will happen if the protected TNPP acts as complementary nucleobase. Herein, two scenarios are feasible. The steric clash of the metalated P residue and the bulky thymine derivative may result in an extrusion of one base from the duplex,5b so that a destabilization of the duplex would be expected. Alternatively, the formation of a HgII‐mediated base pair involving the TNPP ligand may occur, which should be accompanied by a minor duplex stabilization. To evaluate these possibilities, duplex V with a central TNPP:P pair was investigated.

Figure 3

Proposed structure of the metal‐mediated T‐HgII‐P base pair.

Again, temperature‐dependent UV spectroscopy was applied to probe metal‐mediated base‐pair formation. Figure 4 shows the melting curves obtained for duplex V. An increase in T m of 7.9±0.8 °C is observed after the addition of one HgII per duplex prior to photodeprotection, suggesting that a TNPP‐HgII‐P base pair is indeed formed with the caged nucleobase. Subsequent irradiation at room temperature leads to a further increase in T m of 8±1 °C, indicating photodeprotection and formation of a T‐HgII‐P pair. Hence, the chelating phenanthroline‐derived ligand P binds the HgII ion irrespective of the identity of the complementary nucleobase. Even though the thymine residue bears a bulky protecting group, it is forced to engage in metal‐mediated base pairing. Metal‐mediated base‐pair formation may additionally be facilitated by the more flexible acyclic backbone of P. Finally, photodeprotection relieves the steric strain, accompanied by a further increase in the melting temperature.

Figure 4

Melting curves of duplex V at pH 6.8 in the absence of HgII (black), in the presence of one equivalent of HgII prior to irradiation (red), in the presence of one equivalent of HgII after irradiation (blue) and in the presence of two equivalents of HgII after irradiation (green). Experimental conditions: 1 μm duplex, 150 mm NaClO4, 2.5 mm Mg(ClO4)2, 5 mm MOPS buffer (pH 6.8).

To confirm this assumption, duplexes VI and VII bearing a T:P or a TPP:P pair, respectively, were investigated. Duplex VI is stabilized by 16.5±0.5 °C upon formation of the T‐HgII‐P pair (Figure S8a, Supporting Information). This stabilization is identical to the one observed for duplex V upon metal binding and photodeprotection (16±1 °C). The TPP nucleobase in duplex VII bears a substituent of similar size as TNPP. However, this substituent cannot be removed by irradiation. For duplex VII, T m increases by 5.3±0.9 °C upon the addition of HgII (Figure S8b, Supporting Information). Even though this increase is a bit smaller than that observed for duplex V (7.9±0.8 °C), the experiment clearly confirms the applicability of a caged nucleobase in HgII‐mediated base pairing. As anticipated, subsequent irradiation does not lead to a change in T m. Again, the formation of the metal‐mediated base pair can be confirmed CD‐spectroscopically. The binding of HgII to form a TNPP‐HgII‐P pair in duplex V evokes an increase in [θ] at ∼275 nm (Figure S9a, Supporting Information). The same observation is made for reference duplex VII upon the formation of the TPP‐HgII‐P pair (Figure S9c, Supporting Information). Generation of the final T‐HgII‐P pair in duplex V upon photodeprotection is accompanied by a blueshift of the positive Cotton effect and a decrease in [θ] at ∼245 nm (Figure S9a, Supporting Information). Again, the same effects are observed upon the formation of the T‐HgII‐P pair in reference duplex VI (Figure S9b, Supporting Information). Taken together, these data prove that caged nucleobases can be involved in metal‐mediated base pairing, provided that the complementary nucleobase is a bidentate ligand.

It is interesting to note that the O4‐protected thymine residue does not require deprotonation at its N3‐position to engage in metal‐mediated base pairing, due to its enol tautomeric form (Figure 1 b). In this respect, it appears to resemble cytosine, a nucleobase that is known not to form HgII‐mediated base pairs. The formation of a stable T(N)PP‐HgII‐P pair thus indicates that a simple protonation/deprotonation event cannot explain the preferential binding of HgII to thymine rather than cytosine and that additional (e.g. electronic) factors must exist, too.

To conclude, we have shown for the first time the light‐triggered formation of a metal‐mediated base pair, achieved by applying a caged thymidine residue. When using a bidentate ligand as the complementary nucleobase, an unprecedented stepwise duplex stabilization was accomplished. Here, the addition of HgII leads to the formation of a stabilizing metal‐mediated base pair involving the caged nucleobase. Subsequent photodeprotection results in an additional increase in stability. The possibility of using light as an external trigger for metal‐mediated base‐pair formation and the ability to use two orthogonal triggers for the stepwise formation of metal‐mediated base pairs of different stability significantly expands the scope of metal‐modified nucleic acids. In combination with DNA that switches its topology upon metal‐mediated base‐pair formation, interesting applications are anticipated.

Experimental Section

General

The phosphoramidites of TNPP and P were prepared according to published procedures.10b, 17 The TPP nucleoside was prepared in analogy to TNPP.10b Details are given in the Supporting Information. All other phosphoramidites were purchased (Glen Research). The oligonucleotides were synthesized and purified as described previously.17 The desalted oligonucleotides were characterized by MALDI‐TOF mass spectrometry (ODN1: calcd for [M+H]+: 3966 Da; found: 3967 Da; ODN2: calcd for [M+H]+: 4097 Da; found: 4096 Da; ODN3: calcd for [M+H]+: 3803 Da; found: 3803 Da; ODN4: calcd for [M+H]+: 4260 Da; found: 4262 Da; ODN5: calcd for [M+H]+: 4149 Da; found: 4150 Da; ODN6: calcd for [M+H]+: 3921 Da; found: 3920 Da). During oligonucleotide quantification, the following molar extinction coefficients were used: TNPP, ϵ 260=7.5 cm2 μmol−1;10b TPP, ϵ 260=4.2 cm2 μmol−1; P, ϵ 260=10.0 cm2 μmol−1.5b

The UV melting experiments were carried out on a UV spectrometer CARY 100 Bio (Agilent) in a 1 cm quartz cuvette. The UV melting profiles were measured in buffer (1 μm DNA duplex, 150 mm NaClO4, 2.5 mm Mg(ClO4)2, 5 mm buffer (pH 6.8: MOPS, pH 9.0: borate) either with or without Hg(ClO4)2 at a scan rate of 1 °C min−1 with detection at 260 nm. CD spectra were measured using a J‐815 spectropolarimeter (JASCO) at 10 °C in the same solution. Each irradiation experiment was performed for 1 min (at ca. 50 °C for duplexes I–III at pH 6.8 or at room temperature in all other cases) using a 500 W Hg/Xe arc lamp (Newport) equipped with a 1.5 inch water filter and a 335 nm longpass filter (Schott). NMR spectra were recorded on Bruker Avance(I) 400 and Avance(III) 400 instruments. NMR spectra were referenced to residual solvent peaks (CD3OD, CD2Cl2) or to tetramethylsilane (CDCl3).

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.