Studying DNA G-Quadruplex Aptamer by 19F NMR.

In this study, we demonstrated that 19F NMR can be used to study the thrombin-binding aptamer (TBA) DNA G-quadruplex, widely used as a model structure for studying G-quadruplex aptamers. We systematically examined the structural feature of the TBA G-quadruplex aptamer with fluorine-19 (19F) labels at all of the thymidine positions. We successfully observed the structural change between the G-quadruplex and the unstructured single strand by 19F NMR spectroscopy. The thermodynamic parameters of these DNA G-quadruplex aptamers were also determined from the 19F NMR signals. We further showed that the 19F NMR method can be used to observe the complex formed by TBA G-quadruplex and thrombin. Our results suggest that 19F NMR spectroscopy is a useful approach to study the aptamer G-quadruplex structure.

Introduction

G-quadruplexes are four-stranded DNA or RNA structures consisting of G-tetrads formed by four guanines via Hoogsteen base pairing. Recent studies have shown that DNA and RNA G-quadruplexes are related to human disease and cancer.[1] For instance, we have demonstrated that the higher-order DNA G-quadruplex structure in telomeric-overhang DNA protects against degradation by double-strand breaks and nuclease hydrolysis.[2]

Recently, G-quadruplex structures are suggested as targets for drug development based on their biological activities including gene regulation, telomere maintenance, transcription, and DNA replication.[3−9] We found that human telomere RNA forms G-quadruplex structures.[10−13] Recently, we demonstrated that a RNA telomere G-quadruplex binds to hnRNPA1, a heterogeneous nuclear ribonucleoprotein (hnRNP) A1, and plays an important role in telomere end protection.[14] These reports suggested that the G-quadruplex structure has a significant biological function for various events in vivo. Thus, studies of physicochemical properties and structural characteristics of G-quadruplex structures will lead to a better understanding of the biological functions of G-quadruplexes.

G-quadruplex aptamers obtained using the in vitro selection technique have been reported against a wide variety of targets as aptamer-based therapeutics and diagnostics.[15] It is especially known that a thrombin-binding aptamer (TBA) is a well-characterized aptamer in the sense of structure and stability,[16−19] acting as an anticoagulant agent inhibiting the activation of fibrinogen as well as platelet aggregation induced by thrombin. The structure of a TBA is a chairlike and antiparallel G-quadruplex structure that consists of two G-quartets connected by two TT loops and a single TGT loop (Figure a).[20] Here, we used 19F NMR to examine the structural feature of a TBA G-quadruplex aptamer. 19F NMR was used to evaluate significant conformations of nucleic acids. 19F is an ideal conformational probe because of its highly sensitive and low background signals in biological samples.[21−35] We have recently reported the application of 19F NMR spectroscopy for characterizing RNA G-quadruplex conformations.[36] We have also successfully used 19F NMR spectroscopy to study the interactions between RNA G-quadruplexes and ligand molecules and proteins.[37]

Figure 1

(a) Schematic cartoon of the NMR structure of the G-quadruplex aptamer. (b) Sequences of the 19F-labeled oligonucleotides synthesized in this study and chemical structure of 19F-labeled 2′-deoxyuridine having 3,5-bis(trifluoromethyl)phenyl group.

Here, we systematically labeled the TBA G-quadruplex with 19F-labeled nucleotide at each one of the thymidine positions and successfully examined the structural feature of the aptamer by 19F NMR (Figure b). The thermodynamic parameters of the DNA G-quadruplex aptamers were determined. The 19F NMR method was also used to study the interactions of TBA G-quadruplex and thrombin complex.

Results and Discussion

To prepare the required 19F-labeling, 5-iodo-2′-deoxyuridine was reacted with 1-ethynyl-3,5-bis(trifluoromethyl)benzene by Sonogashira coupling (Scheme S1 and Figures S1–S9). We first investigated the conformation of the 19F-labeled oligonucleotides by circular dichroism (CD) spectroscopy. The CD spectra of with and without 19F-labeled TBA oligonucleotides in the presence of 100 mM potassium ions have positive and negative peaks at 295 and 265 nm, respectively, which are characteristic of an antiparallel chairlike G-quadruplex structure in agreement with previous reports.[38] This indicates that the incorporation of the 19F-labeled nucleoside in the loop region does not affect the folding of the G-quadruplex (Figure S10).

We next investigated the structural behavior of oligonucleotides using 19F NMR spectroscopy. TBA DNA G-quadruplex as a model structure has been well studied. In particular, Galeone et al. have suggested that the minor loops TT are involved in the interaction with the thrombin.[19] To test the influence of T on TBA aptamer by 19F NMR in detail, we label TBA at all T positions. Figure a shows the 19F NMR spectra of all 19F-labeled oligonucleotides (T3, T4, T7, T9, T12, and T13) in the presence or absence of potassium ions. It can be clearly observed that the new peaks appear with the addition of 100 mM KCl compared with 0 mM KCl for all of the oligonucleotides. The results suggested that the corresponding resonances of the DNA structures can be used to distinguish the DNA structures of G-quadruplexes. For instance, for 19F-labeled TBA oligonucleotide (T12) in the absence of potassium ions, the single-stranded T12 indicated a single peak at −62.88 ppm, whereas in the presence of 100 mM potassium ions, a new peak of G-quadruplex at −62.57 ppm appeared and the original peak of the single strand (−62.88 ppm) disappeared. When the temperature was increased to 70 °C, the G-quadruplex peaks disappeared, and only the single-strand peak (−62.88 ppm) was detected in the 19F NMR spectrum (Figure b). To further confirm that signal at −62.88 was from the G-quadruplex, we performed a 1H NMR experiment by the observation of imino protons from the G-quadruplex (Figure b). Eight peaks of the TBA G-quadruplex were found at 11.5–12.5 ppm corresponding to the single peak in the 19F NMR spectrum. At the denaturing temperature (70 °C), no peak was observed in the 1H NMR. In accordance with the 19F NMR spectrum, we only observed the peak of the unstructured single strand. We observed that the label of 19F showed different positions and degree of chemical shift. The structural environment of the 19F-label strongly influences the 19F NMR signals. Thus, the 19F at different positions of TBA aptamer may result in the corresponding resonances. Therefore, these results suggest that the 19F NMR method can strictly discriminate the G-quadruplex structure. NMR signals depend on whether the molecule of interest is freely available for tumbling in a solution.[39,40] In general, molecules display big tumbling rates due to their random coil and increase in temperature, leading to an overall increase in the tumbling rates and, consequently, a degree of chemical shift of 19F NMR.

Figure 2

(a) 19F NMR spectra of 19F-labeled oligonucleotides in the presence or absence of KCl. Condition: [DNA] = 100 μM, [KCl] = 100 mM, [Tris–HCl buffer (pH 7.0)] = 10 mM. (b) 1H imino proton NMR of T12 corresponding to 19F NMR at 25 or 70 °C.

We next performed a temperature-dependent experiment to confirm the assignment of the two signals in the 19F NMR (Figure a). When the temperature was increased from 30 to 65 °C, the intensity of the signals for the G-quadruplex decreased (red circle for T3, T4, T7, T9, T12, and T13), whereas that of the signals for the single-strand DNA increased (blue circle). The result indicated the conversion of the G-quadruplex to the single-strand DNA at higher temperatures. When heated to approximately 60 °C, only the peaks of the single strand showed a strong intensity, whereas the G-quadruplex peaks completely disappeared. This result indicated that all of the G-quadruplexes unfolded to a single strand at high temperatures.

Figure 3

(a) 19F NMR spectra of 19F-labeled oligonucleotides at different temperatures. Blue and red spots indicated single strand and G-quadruplex. Temperatures indicated on the right. (b) Melting profiles derived from the relative peak areas of the 19F signals on different temperatures.

Because the 19F NMR spectroscopy can monitor the conformational transition of the DNA aptamer, we further characterized the melting process using the 19F signals at various temperatures (Figure b). According to the melting curves estimated from the 19F NMR peak areas, T9 was more stable than other G-quadruplexes. The quantitative characterization of the thermodynamics of the process can be easily performed by the resonances of different conformations (Table and Figure S11). The enthalpy change (ΔH) indicates that the stability of G-quadruplex is an enthalpic definition (ΔH = −192.6 kJ/mol for T9), suggesting large enthalpic contributions to the stability of the G-quadruplex.

Table 1

Thermodynamic Parameters of G-Quadruplex Transitiona

sample	ΔH (kJ/mol)	ΔS (J/(mol K))	ΔG298K (kJ/mol)
T3	–166.7	–524.4	–10.4
T4	–168.9	–533.4	–9.9
T7	–163.7	–525.3	–7.1
T9	–192.6	–596.7	–14.7
T12	–174.4	–548.2	–11.0
T13	–175.1	–552.1	–10.5

van’t Hoff plots are determined the thermodynamic parameters. The experimental errors (ΔH and ΔS) were ±5 kJ/mol and ±10 J/(mol K).

Inspection of the molecular model of T9 suggests that the 3,5-bis(trifluoromethyl)benzyl group is located near the G8 residue of the TGT loop (Figure a), which provokes a substantial decrease in the distance between fluorine and H8 and H2′ of G8. The F–H8 and F–H2′ distances were approximately 2.4 and 1.9 Å, respectively (Figure b). The close F–H8 and F–H2′ contacts are indicative of a pseudo-hydrogen-bond of F–H8 and F–H2′, such as those found in the G-quadruplex containing 2′F–G residue.[41] Therefore, the enhanced stability of T9 may be due to the F–H8 and F–H2′ interactions via pseudo-hydrogen-bonds, both of which are lacking in the other G-quadruplexes.

Figure 4

(a) Molecular model of the G-quadruplex formed by T9. (b) Distances between F–H8 and F–H2′ in T9. dG and 19F-labeled 2′-deoxyuridine are indicated in magenta and yellow, respectively. 19F is light blue.

We further investigated the interaction of G-quadruplex aptamer with thrombin and inhibitory activity. As shown in Figure S12, following the addition of thrombin to the G-quadruplex aptamer sample, a new signal was detected in the 19F NMR spectrum. The signal was attributed to the complex between the G-quadruplex aptamer and thrombin, suggesting that 19F NMR spectroscopy is a useful tool for studying the binding events of G-quadruplex aptamer and protein targets. To test the inhibitory activity of 19F-labeled TBA oligonucleotides, we measured the prothrombin times of T4 and T9. To compare the influence of 19F-labeled T on inhibition, we selected the T9 with chemical modification at the central TGT loop and the T4 at the lateral TT loop. The T9 with a higher stability is expected to present an inhibitory activity, in which T4 is used as a control. The T9 oligonucleotide showed adequate inhibitory activity in plasma clotting as compared to T4 and the untreated condition (Figure S13).

Conclusions

We demonstrated that 19F NMR spectroscopy can be used to effectively analyze the G-quadruplex aptamer structure. On the basis of the 19F NMR signals, we quantitatively determined Tm and thermodynamic parameters of a G-quadruplex aptamer. Furthermore, observation of 19F NMR chemical shift changes was also shown to be a useful approach for monitoring the interactions of G-quadruplex aptamer and protein target.

Experimental Section

DNA Synthesis and Purification

All of the DNAs were synthesized on the 1.0 μmol scale with an NTS DNA/RNA synthesizer (Nihon Techno Service Co., Ltd.), using the solid-phase phosphoramidite chemistry. After automated synthesis, the oligonucleotides were detached from the support and deprotected according to the manufacturer’s protocol. All of the oligonucleotides were purified by RP-HPLC (JASCO) using an appropriate linear gradient of 50 mM ammonium formate in H2O and 50 mM ammonium formate in 1:1 acetonitrile/H2O. The fractions containing the product were collected and lyophilized to give purified oligonucleotides. The purified oligonucleotides were desalted by NAP5 columns (GE Healthcare) and identified by a matrix-assisted laser desorption/ionization-time-of-flight mass spectrometer (MALDI-TOF MS) (negative mode).

MALDI-TOF MS

The matrix for MALDI-TOF MS was 1:1 mixture of 3-hydroxypicolinic acid (3HPA) in 1:1 acetonitrile/H2O saturated solution and 0.5 M ammonium citrate aqueous solution. 1 μL of DNA sample was mixed with 1 μL of matrix solution. A spot of 1 μL of the sample–matrix mixture was placed on a MALDI target plate (MTP 384 ground steel, Bruker) and allowed to air dry at room temperature. The spectrum was measured using a matrix-assisted laser desorption/ionization-time-of-flight mass spectrometer (MALDI-TOF MS) on Bruker autoflex II mass spectrometer (negative mode) with dT8 ([M – H]−: 2370.603) and dT17 ([M – H]−: 5108.376) as an external calibration standard.

CD Measurements and Analysis of CD Melting Profile

CD spectra were measured using a JASCO J-820 CD spectrophotometer. The spectra were recorded using a 1 cm path length cell. Samples were prepared by heating the oligonucleotides at 95 °C for 5 min and gradually cooling them to room temperature. The melting curves were obtained by monitoring 295 nm. Solutions for CD spectra were prepared as 0.3 mL samples at 5 μM strand concentration in the presence of 100 mM KCl, 10 mM Tris–HCl (pH 7.0).

19F NMR Experiments

NMR experiments were performed on a Bruker AV-400M spectrometer. Trifluoroacetic acid as an external standard (−75.66 ppm) was used. The NMR experiment with DNA G-quadruplex (100 μM) was performed in 10 mM Tris–HCl (pH 7.0), 100 mM KCl containing 10% (v/v) D2O.

Prothrombin Time Measurement

Prothrombin times were measured by using Thrombotrack (Sanko Junyaku Co., Ltd.). Human plasma (100 μL) and 3 μL of oligonucleotides (T4 and T9) were incubated at 37 °C for 1 min at 0.25, 0.5, and 1 μM final concentrations. Then, 200 μL of thromboplastin reagent (tissue factors) was added. The prothrombin time measurement was performed in triplicate, and the average value was calculated.

Molecular Modeling

The initial model of 19F-labeled TBA (T9) was generated from the NMR solution structure of the d(GGTTGGTGTGGTTGG) (PDB code 148D) using the BIOVIA Discovery Studio 2016 (Accelrys). The molecular dynamics simulation was performed by the CHARMm program in BIOVIA Discovery Studio 2016.