Excited States of Nucleic Acids Probed by Proton Relaxation Dispersion NMR Spectroscopy

Abstract In this work an improved stable isotope labeling protocol for nucleic acids is introduced. The novel building blocks eliminate/minimize homonuclear 13C and 1H scalar couplings thus allowing proton relaxation dispersion (RD) experiments to report accurately on the chemical exchange of nucleic acids. Using site‐specific 2H and 13C labeling, spin topologies are introduced into DNA and RNA that make 1H relaxation dispersion experiments applicable in a straightforward manner. The novel RNA/DNA building blocks were successfully incorporated into two nucleic acids. The A‐site RNA was previously shown to undergo a two site exchange process in the micro‐ to millisecond time regime. Using proton relaxation dispersion experiments the exchange parameters determined earlier could be recapitulated, thus validating the proposed approach. We further investigated the dynamics of the cTAR DNA, a DNA transcript that is involved in the viral replication cycle of HIV‐1. Again, an exchange process could be characterized and quantified. This shows the general applicablility of the novel labeling scheme for 1H RD experiments of nucleic acids.


N 6 -Benzoyl-8-13 C-adenosine (S4)
Compound S3 (1.76 g, 2.6 mmol) was dissolved in a mixture of pyridine and ethanol (30 ml, 1:1) and stirred in an ice bath at 0°C. A 2 M NaOH solution (25 ml in H 2 O/EtOH 1/1) was added and the reaction was stirred for 20 min. The strongly akaline solution was neutralized using DOWEX 50WX8-100 H + -form. The DOWEX resin was filtered off and washed with pyridine/ethanol 4:1 (200 ml). The solvents were evaporated and the crude yellow residue was co-evaporated twice with toluene. The slightly yellow residue was triturated with diethylether/CH 2 Cl 2 1/1 (100 ml). the resulting precipitate was isolated by filtration and dried in high vacuum. The crude product was directly used in the next step.

N 6 -Benzoyl-5'-O-(4,4'-dimethoxytrityl)-(8-13 C)adenosine (S5)
The crude product S3 (1.0 g, 2.69 mmol) from the previous step was co-evaporated three times with pyridine and dried in high vacuum for 30min. The oily residue was dissolved in anhydrous pyridine (10 ml) and stirred under argon at room temperature. Then, 4,4'-dimethoxytrityl chloride (DMT-Cl, 910 mg, 2.69 mmol) was added in three portions over a period of 1 h. After another 3 hours thin layer chromatography (TLC) showed complete conversion and the solvent was evaporated. The residue was co-evaporated three times with toluene. The resiude was dissolved in methylene chloride and the organic phase was washed once with saturated sodium bicarbonate solution and dried over anhydrous sodium sulfate. Compound S5 was purified by CC using a gradient of methanol in CH 2 Cl 2 from 1% to 7%. Yield: 1.34 g (1.99 mmol, 74%) TLC: 9/1 CH 2 Cl 2 /MeOH, R f =0.

8-13 C-Guanine (S8)
Morpholine (1.8 g, 20.7 mmol) and 13 C-formic acid (798 μl, 20.7 mmol) were combined at 0°C. The clear solution turned yellow and crystallized after a few minutes. 2,4,5-Triamino-6hydroxypyrimidine sulfate (1.65 g, 6.9 mmol) was added the mixture heated to 100°C. After 40 min. the temperature was raised to 200°C. The reaction was continued for 2 h. After cooling to room temperature ethanol was added and the mixture heated to 100°C for 30 min. The precipitate was isolated by filtration and transferred into a round bottom flask. The crude guanine was suspended in water and 2 ml concentrated aqueous ammonia (28%) was added. The suspension was concentrated on a rotary evaporator until a pH 7 was reached and then cooled to 4°C overnight. 8-13 C-Guanine S8 was isolated by filtration and then dried in high vacuum. To the suspension isobutyric anhydride (iBu 2 O, 6.2 ml, 37.2 mmol) was added and the mixture was refluxed at 170°C under argon. After three hours, the solution was cooled to room temperature and the solvent was removed by distillation. The residual oil was co-evaporated twice with methanol and recrystallized in ethanol/H 2 O (1/1). After filtration S9 was dried under high vacuum.

N 2 -Isobutyryl-2',3',5'-tri-O-benzoyl-(8-13 C)-guanosine (S10)
Compound S9 (1.0 g, 4.5 mmol) and ATBR (2.27 g, 4.5 mmol) were suspended in anhydrous toluene (20 ml). BSA (6.62 ml, 27 mmol) was added and the reaction was heated to 105°C under argon. After 1h the solution was allowed to cool to room temperature and (TMS)OTf (2.46 ml, 13.5 mmol) was added. The solution was heated again to 105°C and stirred for 1 h. The solution was diluted with CH 2 Cl 2 and washed with saturated NaHCO 3 solution. Dried over anhydrous Na 2 SO 4 , the solvents were evaporated and the crude product was purified by column chromatography using a gradient of MeOH in CH 2 Cl 2 from 0% to 3%. N 2 -Isobutyryl-(8-13 C)-guanosine (S11) Sodium ethoxide (NaOEt, 2.4 g, 35.3 mmol) was dissolved in 34 ml pyridine/EtOH 2/3 and S10 (2.8 g, 4.2 mmol), dissolved in 20 ml pyridine/EtOH 2/3, was added. The reaction was stirred for 25 minutes at room temperature. The strongly alkaline solution was neutralized by DOWEX 50WX8-100 H + -form. The DOWEX was removed by filtration and washed with pyridine/ethanol 3/2. The solvents were evaporated and co-evaporated twice with toluene. The residue was dissolved in water and the aqueous phase was washed with CH 2 Cl 2 . The aqeous phase was evaporated and the oily residue was co-evaporated with methanol. The resulting orange foam was dried under high vacuum. The crude product was directly used in the next step.
Yield: 1.47 g (4.15 mmol, 99%) N 2 -Isobutyryl-5'-O-(4,4'-dimethoxytrityl)-(8-13 C)-guanosine (S12) S11 (870 mg, 2.46 mmol) was co-evaporated three times with pyridine before being suspended in anhydrous pyridine (15 ml). The reaction was stirred under argon at room temperature for ten minutes. Then, DMT-Cl (1.25 g, 3.68 mmol) was added in two portions while the progress of the reaction was monitored with TLC. After stirred 4 h the solvent was evaporated and the residue was co-evaporated three times with toluene and then dried in high vacuum. The oily residue was then dissolved in methylene chloride and the organic phase was washed with saturated sodium bicarbonate, dried over anhydrous sodium sulfate. After evaporation crude S12 was purified by column chromatography using a gradient from 0% to 10% of methanol in CH 2 Cl 2 .

Analysis of proton relaxation dispersion data
Spectral processing was performed using nmrPipe and the nmrDraw software package. Peak integration was performed using either the nmrDraw or FuDA software packages. [2][3][4]

Analysis of CPMG Relaxation dispersion data
For the A-site RNA the peak intentities as a function of CPMG frequencies, I(ν CPMG ), were obtained using the FuDA software, whereas peak intensities for cTAR DNA were obtained by summing over 3x3 (t 1 x t 2 ) data points using the serT-script implemented in the nmrPipe package. Peak intensities were converted into effective relaxation rates via R 2,eff (ν CPMG ) = −1/T relax ·ln(I(ν CPMG )/I 0 ) where I 0 is the peak intensity in a reference spectrum recorded without the relaxation delay T relax . The uncertainty of the R 2,eff (ν CPMG ) rates was propagated from the corresponding uncertainty of I(ν CPMG ) and I 0 or set to 2% of R 2,eff (ν CPMG ), whichever was the largest. Values of the exchange rate, k ex , the population of the excited state, p b , chemical shift differences between states, Δϖ, and intrinsic relaxation rates ",$%% & that were assumed to be identical in the two exchanging sites, were extracted using the in-house written software CATIA by minimization of the following χ 2 target function: [5,6] [7] In all fits of experimental dispersion profiles we have assumed that the intrinsic relaxation rates were the same for both exchanging states. The exchange parameters k ex and p b were determined by only including those residues where (1) the two-site exchange model generated statistically significant improvements in fits over a model of no exchange at the 98% confidence level and (2) the maximum exchange contribution, R 2,eff (100 Hz) -R 2,eff (2000 Hz) > 3 s -1 . Thus, the relaxation dispersions of C7, C9, A10, A93, G94, U95 were included for the A-site RNA 1 and residues G4, A5, C19, G20, A21, C22 and C23 included for the cTAR DNA 2 for the determination of k ex and p b . In a subsequent fit, the obtained k ex and p b were kept fixed and the Δϖ for all residues determined. The uncertainties in the derived model parameters were obtained from the covariance matrix method. Supporting Figure 3. b) Selected proton relaxation dispersion curves of A-site RNA 1 at two/three magnetic field strengths (500 (red), 600 (green) and 800 (blue) MHz). Open circles represent data points, vertical bars the associated uncertainty. The fit is shown as a full-drawn line. b) Proposed exchange process for A-site RNA 1 with rates and populations derived from the relaxation dispersion curves in a. c) The relaxation dispersion profile of U90 exemplifies a "flat dispersion profile" when Δϖ ≈ 0 ppm, thereby showing the absent of homenuclear couplings and supporting the efficiency of the labelling scheme.
Supporting Figure 4. a) Proton relaxation dispersion curves of cTAR DNA 2 at two magnetic field strengths (500 (red) and 600 (green) MHz). Open circles represent data points, vertical bars the associated uncertainty. The fit is shown as a full-drawn line. b) Proposed exchange process for cTAR DNA 2 with rates and populations derived from the relaxation dispersion curves in a.  All pulse phases are assumed to be x, unless indicated otherwise. The 1 H carrier is placed on the water signal at a and moved to the middle of the H6/H8 region (7.6 ppm) during the CPMG element between b and c, whereas the 13 C carriers is in the middle of the C6/C8 region (140 ppm). 13 C decoupling during acquisition is achieved with a WALTZ-16 scheme applied at a field of 2.3 kHz. [12] The phase cycling used is: φ1 = {x -x}, φ2={x