Chemo-enzymatic synthesis of site-specific isotopically labeled nucleotides for use in NMR resonance assignment, dynamics and structural characterizations

Abstract

Stable isotope labeling is central to NMR studies of nucleic acids. Development of methods that incorporate labels at specific atomic positions within each nucleotide promises to expand the size range of RNAs that can be studied by NMR. Using recombinantly expressed enzymes and chemically synthesized ribose and nucleobase, we have developed an inexpensive, rapid chemo-enzymatic method to label ATP and GTP site specifically and in high yields of up to 90%. We incorporated these nucleotides into RNAs with sizes ranging from 27 to 59 nucleotides using in vitro transcription: A-Site (27 nt), the iron responsive elements (29 nt), a fluoride riboswitch from Bacillus anthracis(48 nt), and a frame-shifting element from a human corona virus (59 nt). Finally, we showcase the improvement in spectral quality arising from reduced crowding and narrowed linewidths, and accurate analysis of NMR relaxation dispersion (CPMG) and TROSY-based CEST experiments to measure μs-ms time scale motions, and an improved NOESY strategy for resonance assignment. Applications of this selective labeling technology promises to reduce difficulties associated with chemical shift overlap and rapid signal decay that have made it challenging to study the structure and dynamics of large RNAs beyond the 50 nt median size found in the PDB.

Figure 1.

TROSY improvements for IRE and HCV site-specifically labeled samples. (A) The base region of 1′,5′,8-¹³C3 GTP labeled IRE. Comparisons of the HSQC and TROSY pulse sequences show an average intensity improvement of 2.9 ± 0.5. The inset displays the overlaid slices of the highlighted residue. (B) The base region of 1′,5′,8-¹³C3 ATP labeled HCV. Again comparisons of the HSQC and TROSY pulse sequences show an average intensity improvement of 3.3 ± 1.0. The inset displays the overlaid slices of the highlighted residue.

Figure 2.

The improvement of HSQC versus TROSY pulse sequences for IRE (A) and HCV (B).

Figure 3.

The ¹³C-TROSY CEST profile for the fluoride riboswitch recorded at 35°C. Measurements were made on the C1′ (gray) and C6 (black) site of a 1′,6-¹³C-1,3-¹⁵N-5-²H UTP. Data were globally fit to a 2-site full Bloch–McConnell equation. The residue shown is proposed to be involved in an unstructured to base-paired transition involved in the formation of a pseudoknot. The B. anthracis construct had exchange parameters of k_ex = 617 ± 54 s^-1 and p_b = 3.0 ± 0.1%.

Figure 4.

CPMG fits for HCV SARS (A) and A-site (B). (A) The CPMG data revealed k_ex for select residues in the Human corona virus RNA element. Data were measured on the C1’ position of a 1′,8-¹³C labeled sample. (B) CPMG was run at 14.1 (▪) and 18.8 (•) T static field strengths for the A-site RNA. A global fit revealed k_ex = 3800 ± 200 s^-1 and p_b = 1.8 ± 0.1%. These results were in agreement with previously reported values (8). Measurements were made on the C2′ position of 2′,8-¹³C labeled sample. The A1493 position that these curves belong to has been implicated in the discrimination of cognate and near-cognate tRNAs in the ribosome.

Figure 5.

(A) NOESY walk of the bacterial A-site RNA using alternatively labeled nucleotides. Starting at the H1′ of C14, the connectivity from the sugar H1′/H2′ to C8/C6 of the N+1 base enables sequential assignment all the way to the H1′ of G19. This allowed the consecutive assignment of 6 residues present in the helical environment. Connectivities follow the placed arrows and move sequentially from cyan to magenta to yellow to black. (B) Pymol representation of the NOESY walk using the same coloring scheme as used in the NOESY walk (PDB: 1A3M (98)). (C) Predicted versus actual chemical shift values for the assigned residues as determined by NMRViewJ (One Moon Scientific). The offset from the central line represents how far the assigned resonance is from its predicted value. Blue circles represent resonances that are less than 0.1 ppm from predicted values, red circles have predicted ppm more than 0.1 ppm away from observed chemical shift.