Synthesis of [7-15N]-GTPs for RNA structure and dynamics by NMR spectroscopy

Several isotope-labeling strategies have been developed for the study of RNA by nuclear magnetic resonance (NMR) spectroscopy. Here, we report a combined chemical and enzymatic synthesis of [7-15N]-guanosine-5′-triphosphates for incorporation into RNA via T7 RNA polymerase-based in vitro transcription. We showcase the utility of these labels to probe both structure and dynamics in two biologically important RNAs. Graphical abstract Supplementary Information The online version contains supplementary material available at 10.1007/s00706-022-02892-1.


Introduction
RNAs, once thought of as an intermediate in the flow of genetic information from DNA to proteins, are now credited with playing a central role in many cellular functions [1][2][3][4]. As a result, RNAs have increasingly become the target of structural and therapeutic efforts [5][6][7]. Among the different techniques available to study RNA, nuclear magnetic resonance (NMR) spectroscopy is particularly useful [8,9]. NMR, unlike X-ray crystallography, permits the study of RNA structure and dynamics in solution on a wide-range of timescales spanning picoseconds-to-seconds [10]. Nevertheless, poor chemical shift dispersion and broad linewidths limit the broad application of NMR to understand the structure and dynamics of RNA [11].
Taken together, our synthetic route provides 4a and 4b in three chemical steps and one enzymatic step with a single chromatographic purification. Furthermore, while the present work used [ 1 H 6 ]-and [ 2 H 6 ]-d-ribose, our method enables the coupling of 3 to any ribose source to create a versatile assortment of atom-specifically labeled rNTPs for use in in vitro transcription and NMR.

NMR characterization of atom-specifically labeled RNA
Our reason for synthesizing 4a and 4b was to characterize the structure and dynamics of biologically important RNAs. Thus, we used 4a and 4b along with unlabeled ATP, CTP, and UTP to make two RNAs by in vitro transcription: a 4b-labeled 27 nt fragment from the human cytoplasmic A-site ribosomal RNA (A-site) and a 4a-labeled 35 nt fragment from domain 5 of the group II intron ribozyme from brown algae (domain 5) (Fig. 1).
As a first application, we employed a two-bond ( 2 J H8N7 ) 2D heteronuclear single quantum coherence (HSQC) experiment on 4a-labeled domain 5 RNA. We obtained a wellresolved 2D spectrum, showing all 13 guanosine H8-N7 resonances (Fig. 2a). A necessary NMR parameter for structure determination is proton-proton distances, which is provided by nuclear Overhauser effect spectroscopy (NOESY) experiments. While these data are informative, crowded proton-proton NOEs can be resolved into a third dimension with 13 C-or 15 N-editing. As a second application, we employed a 15 N-edited 3D NOESY HSQC experiment on 4a-labeled domain 5 to reveal all protons within ~ 5 Å of guanosine H8 [38]. In A-helical RNA, guanosine H8 protons show strong NOE cross-peaks to 5′-neighboring H2′ protons (Fig. 2b), which are traditionally difficult to assign due to severe overlap with other ribose protons (i.e., H3′, H4′, H5′, and H5″) [37][38][39]. We therefore used the chemical shift of guanosine N7 to resolve NOE cross-peaks of H2′ protons to H8. Examples of cross-peaks are shown for helical residues G8 and G22, as well as residues G15 and G26 from the apical loop and internal bulge, respectively (Fig. 2c).
As a final application, we probed the dynamics of 4b-labeled A-site RNA. Two common relaxation parameters in biomolecules are the longitudinal (R 1 ) and transverse (R 2 ) relaxation rates [40,41]. While R 1 measures the rate of recovery of the z-magnetization to equilibrium, R 2 reports on the rate of the decay of x-and y-magnetization [40]. An alternative method to obtain R 2 is a transverse rotating-frame (R 1ρ ) experiment wherein magnetization is aligned along an effective field whose direction is dependent upon the power of the radio frequency (RF) field and its offset [42]. We therefore employed pseudo-2D HSQC-based experiments to determine R 1 and R 1ρ relaxation rates of guanosine H8 protons (Fig. 3a). We obtained rates for 5 of the 9 guanosines in A-site, with R 1 and R 1ρ values ranging from 1.75 to 2.00/s and 15.26 to 21.25/s, respectively. Interestingly, G16 showed the highest R 1 and lowest R 1ρ , indicative of increased flexibility, suggesting the G:C base pair preceding the tetraloop is not stable. These data contribute to our understanding of the dynamic motions within the A-site RNA [43,44].

Conclusion
We report the synthesis of atom-specifically labeled [7][8][9][10][11][12][13][14][15] N]-GTPs for use in in vitro transcription to make RNA for NMR analysis. Our synthetic routes include a combined chemical and enzymatic approach, using inexpensive commercially available starting materials. To demonstrate the utility of our new labels, we introduced them into two RNAs via in vitro transcription to permit incorporated, respectively. Nucleotides labeled with 4a and 4b are numbered and shown in blue straightforward NMR structure and dynamics measurements. We anticipate these labels will aid efforts to probe in greater detail the structure, interactions, and dynamics of biologically and medically RNAs.

Experimental
Commercially available reagents were used without further purification unless explicitly stated. All reagents used for the synthesis of 3 were purchased from Sigma-Aldrich. All reactions were carried out under nitrogen or argon atmosphere. All non-commercially available enzymes were expressed and purified in-house using established methods [28]. DNA templates for in vitro transcription of RNAs were purchased from Integrated DNA Technologies (IDT, Coralville, IA) and used without further purification. Chromatographic purification was carried out using boronate affinity resin with eluent specified. 1 H NMR spectra were recorded on a Bruker Avance I 300 MHz or a Bruker Avance Neo 400 MHz spectrometer, 13 C NMR spectra were recorded on a Bruker Avance III 500 MHz spectrometer, 15 N NMR spectra were recorded on a Bruker Avance III 700 MHz spectrometer, and 31 P NMR spectra were recorded on an Avance III Bruker 800 MHz spectrometer with a triple resonance cryogenic probe. Samples were maintained at a temperature of 25 °C. All NMR experiments for RNA were performed in D 2 O and all chemical shifts were reported in ppm (parts per million). All RNA spectra were referenced to DSS (4,4-dimethyl-4-silapentane-1-sulfonic acid). Nitrogen-15 and Carbon-13 chemical shifts were indirectly referenced using the ratio of the gyromagnetic ratios of proton to 15 N (0.101329118) and 13 C (0.251449530), respectively [45,46]. All NMR experiments for compounds 1, 2, and 3 were performed in DMSO-d 6 or D 2 O. The chemical shifts of compounds 1, 2, and 3 were referenced to the residual protonated solvent signal of DMSO-d 6 (2.5 ppm) or D 2 O (HDO 4.7 ppm) as previously reported [47]. The 15 N dimension was referenced using the liquid ammonia referencing implemented in the Topspin software suite.

, 6 -D i a m i n o -5 -[ n i t ro s o -1 5 N ] py r i m i d i n -4 -o l ( 1 , C 4 H 5 N 4 15 NO )
To begin the chemical synthesis of [7-15 N]guanine, 3.65 g 2,6-diaminopyrimidin-4-ol (29.00 mmol) and 2.30 g 15 N-labeled sodium nitrite (33.30 mmol) were dissolved in 40 cm 3 of 3 M sodium hydroxide. The homogeneous solution obtained from the previous step was then added dropwise to 50 cm 3 glacial acetic acid, while stirring and cooling on ice, and gave rise to a pink precipitate. The precipitate was isolated by centrifugation, washed with cold water, ethanol, and diethyl ether, and dried in high vacuum (1 × 10 -2 mbar) on a vacuum line for 8 h to give pure compound 1. Yield: 3.55 g (79%); 1

, 6 -D i a m i n o -5 -[ a m i n o -1 5 N ] p y r i m i d i n -4 -o l ( , C 4 H 7 N 4 15 NO)
After vacuum drying, compound 1 was resuspended in 80 cm 3 boiling water and 9.96 g sodium dithionite (57.23 mmol) was added in several portions to give a pink suspension. The reaction was kept at 100 °C and the pink suspension became yellow in color. The mixture was then cooled in an ice bath for 30 min. The yellow solid was collected by filtration and then resuspended in 65 cm 3    The reaction was incubated at 37 °C for 12 h. After confirming successful triphosphate formation by 31 P NMR, crude compound 4a was purified by boronate affinity chromatography (eluent A: 1 M triethylamine pH 9; eluent B: acidified water pH 4), lyophilized to a powder, and resuspended in Ultrapure water. Yield: 6.8 mg (~ 90%).

RNA preparation
RNAs were synthesized via in vitro transcription. The reactions were carried out in a 10 cm 3

NMR experiments
All experiments on 4aand 4b-labeled samples were carried out at 25 °C on a Bruker 600 MHz magnet Avance III spectrometer with a TXI triple resonance probe in 5 mM Bruker optimized Shigemi NMR tubes. 2D- 15 15 N dimension, the spectral width was set to 5 ppm and the carrier to 240 ppm. The two-bond ( 2 J H8N7 ) 2D HSQC experiment was performed with 128 scans and 100 and 1024 complex points in the 15 N and 1 H dimensions, respectively. The 15 N and 1 H carrier frequencies were set to 235.5 ppm and 4.7 ppm, respectively; the spectral width for the 15 N and 1 H dimensions were set to 9 and 12 ppm, respectively. The 15 N-edited 3D NOESY HSQC experiment was carried out on a 600 MHz magnet Avance III spectrometer with a TXI triple resonance probe using previously described pulse sequences [38]. For each 3D data set, 200 × 128 complex points were used for the indirect 1 H and 15 N dimensions along with 64 transients and 10% non-uniform sine-weighted Poisson-gap sampling [49]. The spectral width of the 15 N and 1 H dimensions were set to 5.2 and 10.0 ppm, respectively; the carrier was set to 235.1 and 4.7 ppm, for the 15 N and 1 H dimensions, respectively. R 1 and R 1ρ relaxation rates on the guanosine-N7 nitrogen atoms were measured using a pseudo-2D HSQC detected experiments. These experiments were carried out with 16 scans and 1024 time domain points in the 1 H dimension, using a spectral width and carrier of 16 and 4.7 ppm, respectively. For the R 1 experiments, delay times of 0.02 (X2), 0.04, 0.08, 0.15, 0.30, 0.40, 0.60, 0.80, and 1.00 s were used. For the R 1ρ experiments, delay times of 8 (X2), 16, 20, 32, 48, 60, 80, and 100 ms were used. The strength of spin-lock field (ω 1 ) was 1.9 kHz. Calibration of the spin-lock field was carried out as previously described [50]. R 1 and R 1ρ relaxation rates were determined by fitting intensities to a monoexponential decay and errors were estimated from duplicated delay points. All NMR data were collected at 25 °C with a recycle delay of 1.5 s and analyzed using TopSpin 4.0 and NMRViewJ [51].
provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.