2′-O-Trifluoromethylated RNA – a powerful modification for RNA chemistry and NMR spectroscopy†‡

New RNA modifications are needed to advance our toolbox for targeted manipulation of RNA. In particular, the development of high-performance reporter groups facilitating spectroscopic analysis of RNA structure and dynamics, and of RNA–ligand interactions has attracted considerable interest. To this end, fluorine labeling in conjunction with 19F-NMR spectroscopy has emerged as a powerful strategy. Appropriate probes for RNA previously focused on single fluorine atoms attached to the 5-position of pyrimidine nucleobases or at the ribose 2′-position. To increase NMR sensitivity, trifluoromethyl labeling approaches have been developed, with the ribose 2′-SCF3 modification being the most prominent one. A major drawback of the 2′-SCF3 group, however, is its strong impact on RNA base pairing stability. Interestingly, RNA containing the structurally related 2′-OCF3 modification has not yet been reported. Therefore, we set out to overcome the synthetic challenges toward 2′-OCF3 labeled RNA and to investigate the impact of this modification. We present the syntheses of 2′-OCF3 adenosine and cytidine phosphoramidites and their incorporation into oligoribonucleotides by solid-phase synthesis. Importantly, it turns out that the 2′-OCF3 group has only a slight destabilizing effect when located in double helical regions which is consistent with the preferential C3′-endo conformation of the 2′-OCF3 ribose as reflected in the 3J (H1′–H2′) coupling constants. Furthermore, we demonstrate the exceptionally high sensitivity of the new label in 19F-NMR analysis of RNA structure equilibria and of RNA–small molecule interactions. The study is complemented by a crystal structure at 0.9 Å resolution of a 27 nt hairpin RNA containing a single 2′-OCF3 group that well integrates into the minor groove. The new label carries high potential to outcompete currently applied fluorine labels for nucleic acid NMR spectroscopy because of its significantly advanced performance.


Supporting Figures
Supporting Figure 1  Electronic Supplementary Material (ESI) for Chemical Science. This journal is © The Royal Society of Chemistry 2020 NMR measurements of natural and 2'-OCF3 modified RNA RNA samples were lyophilized as triethylammonium salts and dissolved either in 280 µL or 400 µL NMR buffer (15 mM Na[AsO2(CH3)2]•3H2O, 25 mM NaCl, 3 mM NaN3, in D2O or 9/1 H2O/D2O, pH 6.5) and transferred into restricted volume Shigemi tubes or standard 5 mm NMR tubes. Sample concentrations varied between 0.1 and 1 mM and experiments were run at 298 K unless otherwise stated. All NMR experiments were conducted on a Bruker 600 MHz Avance II+ NMR or a 700 MHz Avance Neo NMR both equipped with a Prodigy TCI probe. 1D 19 F-NMR spectra were typically acquired using the following parameters: spectral width 10 ppm, o1p -60 ppm, 32k complex data points. 128 scans were collected with a recycling delay of 1 s resulting in an experimental time of 4 minutes. For the 2D 19 F-13 C HMQC experiments at natural 13 C abundance the following parameters were used: spectral width in the indirect 13 C dimension was set to 10 ppm, and the spectral width in the direct 19 F dimension was set to 10 ppm. A total of 64 complex points was collected S04 in the indirect 13 C dimension (acquisition time = 21 ms) and 1024 complex points were collected in the direct 19 F dimension (acquisition time = 91 ms). 768 scans were collected with a recycling delay of 1 s resulting in an experimental time of 16 h. The carrier frequency was placed at -58 ppm in the 19 F dimension and in the 13 C dimension at 120 ppm. The 1 JCF coupling constant was set to 270 Hz.
The 19 F-19 F-EXSY experiment was conducted using a Bruker standard NOESY pulse sequence (noesygpph). The following parameters were used: spectral width in the indirect 19 F dimension was set to 10 ppm, and the spectral width in the direct 19 F dimension was set to 10 ppm. A total of 32 complex points was collected in the indirect 19 F dimension (acquisition time = 3 ms) and 1024 complex points were collected in the direct 19 F dimension (acquisition time = 91 ms). The carrier frequency was placed at -58 ppm in both 19 F dimensions. 64 scans were collected with a recycling delay of 1 s resulting in an experimental time of 1 h for each EXSY spectrum. The experiment was run at 11 mixing times ranging from 50 to 800 ms with repeat experiments at 150 and 500 ms. Spectral processing and peak integration were performed using Topspin 4.0.8. All subsequent steps were performed using in-house written software written in Matlab (The MathWorks, www.mathworks.com) according to an earlier published protocol. [1] Errors in the extracted rate constants were determined by Monte Carlo analysis where peak intensities were randomly modulated according to the signal to noise levels in the 2D correlation maps.
Crystallization of SRL-OCF3 RNA 27 nt RNA fragments corresponding to E. coli 23 S rRNA sarcin-ricin loop (SRL) modified at position A2670 or C2667 with 2'-O-trifluoromethyl were used for crystallization. RNA was dissolved at a 190 µM concentration in a buffer made of Na EDTA (1 mM, pH 8.0) and Tris-HCl (10 mM, pH 8.0). The dissolved RNA was heated at 55 °C and cooled down to 10 °C using a temperature-controlled device equipped with a Peltier element. Crystals of A2670-modified RNA were grown as very fragile needles (monoclinic form) for about one month at 4 °C using vapor diffusion method by mixing one volume of RNA sample and one volume of reservoir buffer composed of NaCl (80 mM), KCl (12 mM), MgCl2 (20 mM), Na[AsO2(CH3)2]•3H2O (40 mM, pH 7.0), MPD (35 %), and spermine tetrahydrochloride (12 mM). Crystals were flashfrozen in liquid ethane without further cryoprotection. Crystals of C2667-modified RNA were grown at 20 °C as platelets chunks (tetragonal form) and cryoprotected as described previously, [2] and finally flash-frozen in liquid ethane for data collection. The collection of X-ray diffraction data has been done on the X06DA beamline at the SLS synchrotron, Villigen, Switzerland. Processing of the data was done with the XDS Package [3] and the structure was solved by molecular replacement with MOLREP [4] using the related PDB ID 3DVZ unmodified SRL RNA model. Several crystals from the tetragonal form were discarded due to strong merohedral twinning. Both structures were refined with the PHENIX package. [5] Models were built using Coot. [6] A slight kink in the RNA backbone is observed close to the modification in all three molecules from the asymmetric unit of the monoclinic form (Supporting Figure 4).
Coordinates have been deposited with the PDB database (PDB ID 6ZYB for C2667-modified SRL and 6ZXZ for A2670-modified SRL).

N 4 -Acetyl-2'-O-(trifluoromethyl)cytidine
To a suspension of N-bromosuccinimide (NBS, 2.0 g, 11.3 mmol) in dichloromethane (14.5 mL), hydrogen fluoride in pyridine complex (70 % HF, 14.6 g, 13.3 mL) was added at -78 °C, and subsequently, a solution of compound C4a (1.39 g, 2.25 mmol) in dichloromethane (14.5 mL) was added to the mixture. The reaction mixture was allowed to warm to 0 °C and stirred for three hours. Upon completion, the reaction mixture was poured carefully into a flask containing saturated sodium bicarbonate solution (250 mL) and sodium thiosulfate solution (10 %, 180 mL) while vigorously stirring. Further bicarbonate solution was added until neutral pH was obtained. The aqueous layer was extracted several times with chloroform/isopropanol (3/1). Organic phases were combined, dried over sodium sulfate and evaporated.