A fluorescent surrogate of thymidine in duplex DNA

is a new fluorescent thymidine mimic composed of 2'-deoxyuridine fused to dimethylaniline. exhibits the same pKa and base pairing characteristics as native thymidine residues, and its fluorescence properties are highly sensitive to nucleobase ionization, base pairing and metal binding.

Nucleobase analogs constitute an important family of fluorescent probes. 1 They can be positioned in nucleic acid structures with high precision, and their photophysical properties are highly sensitive to local polarity, 2 viscosity, 3 and pH. 4 These features facilitate specific monitoring of biochemical transformations, 5 conformational changes, 6 metal binding, 7 and base pairing interactions. 8 Nucleobase ionization can mediate proton-coupled folding, 9 metal binding, 10 and/or the catalytic activities of certain nucleic acids at neutral pH. 11 Thymidine (T) and uracil (U) are among the most inherently acidic residues, 12 but fluorescent analogs capable of reporting pyrimidine ionization in nucleic acids are scarce. Previously reported examples utilized biaryl or triaryl fluorophores, 4c,d having unreported or highly perturbed pK a values (pK a E 6.8) as compared to unmodified T and U residues (pK a E 9.5). 12 Here we report ''N,N-dimethylaniline-2 0 -deoxythymidine'' or '' DMA T'' that exhibits the same pK a and base pairing characteristics as thymidine, as well as fluorescence properties that can be used to monitor nucleobase ionization, base pairing and metal binding reactions in DNA. DMA T was designed to have the same Watson-Crick face and pK a as thymidine. To generate a push-pull fluorophore, an electron donating group was incorporated at the C6 position of a quinazoline core. 9,13 This position was selected because it is not in conjugation with N3-H, and therefore expected to have little or no impact on its acidity. Molecular orbital calculations predicted charge transfer from a dimethylaniline-centered HOMO to a pyrimidine-centered LUMO with a HOMO-LUMO energy gap (DE) = 2.44 eV (Fig. 1). DMA T was therefore predicted to be a ''push-pull'' fluorophore in its neutral form. In contrast, the DMA T anion has a pyrimidine-centered HOMO, and a larger DE = 2.81 eV. We therefore expected a blue-shift in DMA T fluorescence upon its deprotonation. The synthesis of DMA T (1) commenced from the previouslyreported nucleoside 2 (Scheme 1). 14 Buchwald-Hartwig coupling with Me 2 NH gave the known compound 3 in 82% yield. 9 Addition of fluoride ions to 3 afforded the new nucleoside DMA T (1) in a HOMOs, LUMOs, and their relative energies were calculated from DFToptimized geometries using LSDA/pBP86/DN**.
Under neutral conditions, DMA T (1) exhibits an exceptionally large Stoke's shift, with an absorbance maximum (l abs ) = 357 nm and emission maximum (l em ) = 522 nm (Table 1). To characterize its environmental sensitivity, the l abs and l em of DMA T (1) were measured in various water/dioxane mixtures (Table S2 and Fig. S1, ESI †). A linear correlation (R 2 = 0.980) with a large slope of 177 cm À1 kcal À1 mol À1 was obtained by plotting the Stoke's shift of DMA T (1) against Reichardt's solvent polarity parameter (E 30 T ). 15,16 Together these results confirm that DMA T (1) is a push-pull fluorophore. Interestingly, DMA T (1) exhibits a twofold higher quantum yield in D 2 O (f = 0.07) than H 2 O (f = 0.03, Table 1), suggesting that proton transfer with bulk solvent provides an effective nonradiative decay pathway. 7a To evaluate the fluorescence sensitivity of DMA T (1) towards nucleobase ionization, its absorbance and emission spectra were recorded at different pH values. Consistent with DFT calculations, the emission maximum of DMA T (1) shifted towards the blue with increasing pH (Fig. 3A). This was accompanied by a dramatic increase in fluorescence intensity. The absorbance and fluorescence changes were plotted against pH to determine a pK a = 9.5 AE 0.1, Fig. 3B. This value corresponds to the pK a of thymidine and uracil. 12 To facilitate the site-specific incorporation of DMA T into DNA, phosphoramidite 5 was prepared in two steps by standard DMT-protection and phosphitylation reactions (Scheme 2). Using automated DNA synthesis, DMA T was incorporated at one of four positions within the same, 21-residue DNA sequence (Table S3, ESI †). Three positions near the middle of the sequence (X13, X14 and X15) and a single position near the 5 0 terminus (X2) were selected in order to evaluate the impacts of variable flanking sequences and DNA end ''breathing'' motions, respectively. ‡ The identity and purity of the purified oligonucleotides were confirmed using analytical HPLC and HR-MS (Table S4 and  Circular dichroism (CD) and thermal denaturation experiments were used to assess the impact of DMA T on the global structure and stability of duplex DNA. DMA T-containing duplexes were prepared by heating and slow cooling with 1.1 equiv. of the complementary strand to give CD spectra consistent with the formation of B-form helices (Fig. S3, ESI †). 18 CD spectra were monitored as a function of temperature (Fig. S4, ESI †) to determine the melting temperature of each duplex (T m , Table 2).
DMA T-A-containing duplexes exhibited nearly identical T m values as the corresponding wild-type duplexes containing T-A base pairs (DT m = À0.2 to À1.7 1C). In contrast, duplexes containing a single C-A, A-A, DMA T-T or T-T mismatch at positions X13-X15 caused a large loss in thermal stability (DT m = À4.8 to À7.9 1C). End breathing motions of duplex DNA explain the relatively small changes when the mismatches were positioned at X2 ( Table 2).
The fluorescence properties of DMA T were highly sensitive to the global structure of the DNA containing it (Fig. S5, ESI, † and Table 3). At all three internal positions X13-X15, DMA T exhibited a two-fold higher quantum yield and blue-shifted l em in duplex   versus single-stranded DNA. Decreased probe hydration upon duplex formation is probably responsible for these differences, because the DMA T nucleoside (1) exhibited higher quantum yields and blue-shifted l em in organic versus aqueous solvents (Table S2 and Fig. S1, ESI †). At all four positions of incorporation, higher fluorescence anisotropy was observed in duplex DNA (r = 0.09-0.18) as compared to unfolded structures (r = 0.03-0.06), consistent with large losses in dynamic motions of the probe upon duplex formation ( Table 3).
The photophysical properties of DMA T were sensitive to matched versus mismatched base pairing in duplex DNA. DMA T exhibited a higher quantum yield and blue-shifted l abs and l em in DMA T-A base pairs as compared to DMA T-T, DMA T-G and DMA T-C mismatches (Table 4). Changes in probe hydration and base stacking are likely responsible for these differences, because similar trends were also observed when comparing duplex versus single-stranded DNA containing DMA T (Table 3). Taken together with the thermal denaturation results (Table 2), these data provide additional evidence that DMA T exhibits the same base pairing specificity as T.
Metal-mediated base pairing interactions serve as important recognition motifs in biological and material sciences. 10 For example, Hg II ions specifically bind to opposing thymine residues to form T-Hg-T base pairs 19 that can cause miscoding of DNA synthesis in vitro and possibly in vivo. 20 To evaluate the ability of DMA T-T to mimic T-T in duplex DNA, T m values were measured in the presence or absence of 1.0 equiv. of Hg II (Table 5 and Fig. S6-S8, ESI †). Only small increases in thermal stabilities (DT m = +0.8 to +1.5 1C) were observed when Hg II was added to duplexes containing a DMA T-T or T-T at position X2, whereas much larger increases were observed at positions X13-X15 (DT m = +2.9 to +6.0 1C). At all four positions, the same T m values were obtained for duplexes containing DMA T-Hg-T as T-Hg-T. In contrast, the addition of 1.0 equiv. of Hg II to duplexes containing a C-T mismatch resulted in no increase in thermal stability as compared to the mismatch alone (Table 5). Taken together, these results demonstrate the excellent mimicry of DMA T for thymidine residues in the demanding context of T-Hg II -T base pairs.
The fluorescence properties of DMA T can be utilized to monitor site-specific binding of Hg II ions to DMA T-T sites. Bi-phasic fluorescence quenching was observed upon addition of Hg II to duplex DNA containing a DMA T-T mismatch, giving a 95% decrease in fluorescence intensity upon addition of 3 equiv. of Hg II (Fig. 4). Similar results were obtained when DMA T was located at all four positions X2, X13, X14 and X15 (Table S5 and Fig. S9, ESI †). The biphasic quenching (Fig. 4B) mirrors the increases in T m values obtained when Hg II is added to duplexes containing a T-T mismatch (Fig. S6, ESI †). A comparison of these results reveals that the steep slopes observed between 0.0 to 1.0 equiv. of added Hg II are a result of T-T-specific binding, and the shallow slopes    Table 3 footnotes for experimental details and symbol definitions. between 2.0 to 3.0 equiv. are due to non-specific interactions. In contrast to DMA T-T, very little fluorescence quenching (À20%) was observed upon the addition of Hg II ions to duplex DNA containing a DMA T-A base pair, or a DMA T-G mismatch (Table S6 and Fig. S10, ESI †). Conversely, the addition of Zn II , Cu II , Mg II , Fe II , Ca II , Ag I , Cd II , Pd II and Ni II ions to duplexes containing a DMA T-T mismatch resulted in little or no change in DMA T fluorescence (Fig. S11, ESI †). These results are consistent with previous studies demonstrating a high degree of specificity between T-T mismatches and Hg II ions using thermal denaturation. 19 To the best of our knowledge, our results provide the first example of using a fluorescent nucleobase analog to monitor a specific binding reaction between DNA and Hg II ions. DMA T will therefore enable detailed kinetics analyses and large-scale screening efforts that are not feasible using other analytical techniques. 10d A variety of fluorescent nucleoside analogs are available for solid-phase synthesis of DNA and RNA. 1-8 However, the vast majority of these probes are quenched by their incorporation into duplex nucleic acids, where they can disrupt duplex stability by as much as a base pair mismatch. Here we report a new fluorescent thymidine mimic composed of 2 0 -deoxyuridine fused to dimethylaniline. According to thermal denaturation, fluorescence, and metal binding studies, DMA T exhibits the same base pairing characteristics as native thymidine residues. The quantum yield of DMA T (f = 0.03 in water) increases upon its incorporation into duplex DNA (f = 0.11-0.20), where its fluorescence properties are highly sensitive to nucleobase hydration, ionization, base pairing, and metal binding. Taken together, these results demonstrate that DMA T will enable a wide variety of studies aimed at characterizing biochemical transformations, conformational changes, site-specific metal binding, and base pairing interactions with single-nucleotide resolution.