NMR on ¹⁵N-labeled RNA

In this chapter we discuss and use ¹⁵N-labeled RNA for NMR experiments. The labeling of the nucleobase nitrogens allows for detailed monitoring of the base-pairing interactions. For example, a G nucleotide could be involved in hydrogen-bonding via its N1 and N2 (Watson-Crick) as well as via its N7 (Hoogsteen edge) and N3 (Sugar edge) positions. The nitrogen chemical shifts in RNA span a wide range (70-240 ppm) and are ideal probes as they are very sensitive to changes in their surrounding environment. The typical chemical shift ranges (Watson-Crick) are depicted in Table 1 however they may substantially vary with the structural context, the protonation state (such as amino, imino and quaternary) and the strength and type of hydrogen-bonding. The additional ¹⁵N dimension in ¹H¹⁵N correlation spectra is extremely helpful to distinguish between the different imino and amino resonances of canonical (Watson-Crick) and non-canonical base-pairs. Moreover, the long-range scalar coupling along the N-H···N hydrogen-bond allows for direct elucidation of base-pairing patterns by means of 2D and 3D ¹H¹⁵N¹⁵N COSY type experiments. In addition the ¹⁵N heteronucleus can conveniently be exploited for the measurement of relaxation parameters and residual dipolar couplings to obtain structural and dynamics information.

Experiment Setups for ¹⁵N-labeled RNA:

The homonuclear and heteronuclear scalar couplings form a complex network in the RNA. They are essential for the design of specific NMR experiments (based on through-bond coherences). Moreover, by determining their values, they can provide important structural information (such as torsion angles, stereochemical assignment and presence/strength of hydrogen-bonds). The incorporation of ¹⁵N-isotope labeling thus allows for the use and detection of their associated J-couplings in the RNA bases (as depicted in Figure 1).

_{References (Figure 1):}

_{¹ https://doi.org/10.1016/S0079-6565(97)00023-X}

_{² https://doi.org/10.1023/B:JNMR.0000034358.12599.d1}

_{³ https://doi.org/10.1023/a:1008367405025}

_{⁴ https://doi.org/10.1021/ja981513x}

The ¹⁵N chemical shifts of the imino regions for G:N1 at 145 ppm and U:N3 at 160 ppm are well separated. This facilitates unambiguous identification of U and G in 2D ¹H¹⁵N correlation spectra, such as the 2D ¹H¹⁵N SOFAST-HMQC and the 2D ¹H¹⁵N BEST-TROSY (Figure 2). The resolution and sensitivity are better in the TROSY version, especially for large RNAs (> 50 nucleotides) and high fields. Both experiments should be tested and optimized for signal intensity and resolution by varying the parameters temperature, buffer composition, and excitation bandwidth. In many cases, a complete resonance assignment requires a combination of low-temperature spectra (favorable solvent exchange, better detection of labile base-pairs, usually better for G:H21/H22) and higher-temperature spectra (favorable for lineshapes of stable base-pairs and often for A:H61/H62).

As described in the previous chapter, each signal in the imino resonance region can be assigned to one stable N-H···N or N-H···O hydrogen-bond. The canonical AU and GC base-pairs have one imino-mediated hydrogen-bond each, and GU or UU wobble base-pairs have two imino-mediated hydrogen-bonds each. The spectra in Figure 2 show six canonical base-pairs that are observable at 303 K. Non-canonical base-pairs would be be observed as signals that are shifted to the upfield region in both dimensions with respect to G and U iminos in canonical base-pairs (Figure 3, right panel).

The exocyclic amino groups of G, C, and A also act as hydrogen-bond donors in base-pairs. Amino resonances have typical ¹⁵N chemical shift ranges, corresponding to the nucleobase type. Amino groups are observed in the 2D ¹H¹⁵N FAST-HSQC (Figure 3), which can be recorded either exclusively for the amino region (~60-110 ppm ¹⁵N; ~ 5.5-9 ppm ¹H) or for both imino and amino resonances as illustrated in Figure 3. Typically, the C:N4 groups resonate most downfield in the ¹⁵N dimension and have the most favorable exchange properties, so they are well observable over a wide temperature range. The right side spectrum of Figure 3 also shows some special signals such as the protonated C:N3, which is due to an elevated pK_a for this imino nitrogen.

In the long-range 2D ¹H¹⁵N LR-FAST-HSQC, the non-exchangeable aromatic signals H2 and H8 are correlated with the nitrogen atoms via the ²J(HN) coupling (10–15 Hz). The experiment enables assignment of the aromatic purine resonances in the absence of ¹³C labeling. Especially the A:H2- N1 / N3 region is well-separated at 220 ppm / 210 ppm (Figure 4).

The HNN-COSY experiment provides direct evidence for N-H···N hydrogen-bonds in the RNA structure. Two distinct magnetization transfers are employed within the HNN-COSY experiment, First, the H-N INEPT transfers the magnetization from the imino-proton to the hydrogen-bond donor nitrogen via the ¹J_NH coupling. Subsequently, the N-N COSY transfers part of the magnetization from the hydrogen-bond donor nitrogen across the hydrogen-bond onto the acceptor nitrogen via ²J_NN coupling. Part of the magnetization remains on the donor nitrogen, so that chemical shift in the indirect ¹⁵N dimension evolves on the hydrogen-bond donor and also acceptor nitrogen. After back transfer of magnetization to protons, magnetization evolves on the imino-protons in the direct ¹H dimension.

For all N-H···N hydrogen-bonds this results in a set of two peaks with identical ¹H and their distinct (donor and acceptor nitrogen) ¹⁵N chemical shifts. The signal intensity is proportional to -sin²(2πJ_NNΔ) for the cross and cos²⁡(2πJ_NN Δ) (Δ = transfer delay) for the diagonal peaks (Figure 1).

Therefore, the intensity ratio of cross and diagonal peaks depends on the ²J_NN coupling constant. The coupling constant typically ranges from 5 to 8 Hz and depends on the geometry of the hydrogen-bond more than on the type of base-pair.

The HNN-COSY experiment can unveil different types of N-H···N base-pairs based on the characteristic ¹⁵N chemical shifts of the donor and acceptor nitrogen. This includes Watson-Crick base-pairs (A-U & G-C), but also Watson-Crick-Hoogsteen, Hoogsteen-Hoogsteen, or Watson-Crick-Sugar Edge base-pairs. Usually, setup as a 2D experiment is favored due to the decreased measurement time, but in case of overlap in the ¹H dimension, an additional ¹⁵N dimension could possibly resolve the signals.

The detection of imino-proton resonances in the direct ¹H dimension limits the HNN-COSY to stable base-pairs for which the imino protons are sufficiently protected from solvent exchange. In Figure 5, the HNN-COSY spectrum of the LR-HP RNA at 303 K is shown. At this temperature, the imino proton resonances of G1, G11, U15 and U16 are broadened due to increased solvent exchange rates. While still observable in 1D ¹H spectra, these resonances are not observed in the HNN-COSY. The remaining G-C base-pairs G2-C23 & C5-G20 are observed and show the diagonal peak G:H1-G:N1 as well as the cross-peak G:H1-C:N3, which can be identified by the C:N3 ¹⁵N chemical shift of 190-200 ppm. Similarly, for the A-U base-pairs A3-U22, A4-U21, U6-A19, U7-A18 and U8-A17 the diagonal peak U:H3-U:N3 and cross-peak U:H3-A:N1 with a characteristic A:N1 ¹⁵N chemical shift of ~220 ppm, are observed. Non-canonical base-pairs are not present in the LR-HP RNA. They can be identified and assigned based on the ¹⁵N chemical shifts provided in Table 1. For exchange broadened imino proton resonances, the LR-HNN-COSY is recommended, because non-exchangeable protons are detected instead of the imino protons.

The LR-HNN-COSY detects non-exchangeable protons, thereby also enabling the observation of base-pairs with rapidly exchanging and therefore non-observable imino protons. In contrast to the HNN-COSY, there are different approaches to detect on non-exchangeable protons while still transferring magnetization via ²J_NN coupling across the hydrogen-bond. The most common approach relies on a H-N INEPT transfer starting on the A:H2 proton and transferring to the A:N1 via a ²J_NH coupling compared to ¹J_NH in the HNN-COSY. Subsequently, the NN-COSY element is applied with the only difference to the HNN-COSY being that the transfer starts on the hydrogen-bond acceptor nitrogen.

In the example spectra measured on the 24mer RNA, A:H2-U:N3 & A:H2-A:N1 correlations are observed for all five A-U base-pairs (Figure 6) which are also observed in the HNN-COSY. Additionally, three A:H2-A:N1 resonances are observed, which do not show a A:H2-U:N3 cross-peak across the hydrogen-bond, indicating that these adenosines are not involved in base-pairs at 303 K. Due to the similar size of coupling constants for ²J_N1H2 = 15 Hz and ²J_N7H8 = 11 Hz, also purine H8-N9 resonances are observed, which can be useful for identification of Hoogsteen base-pairing. A major drawback is that the LR HNN-COSY only allows observation of A-U base-pairs. There are additional pulse sequences which correlate C:H5 and A:H2 with G:N1 and U:N3 for A-U & G-C base-pairs, respectively, but the magnetization transfer requires ¹³C labeling (Dallmann et al.)

The 2D ¹H¹⁵N CPMG-NOESY is a 2D ¹H¹⁵N HSQC-alike correlation experiment employing an additional NOESY mixing time to derive distance information for exchange-broadened amino protons (Mueller et al. and Mulder et al.). Briefly, the magnetization transfer from proton to nitrogen in exchange-broadened amino groups is realized via a CPMG pulse-train. The CPMG-INEPT significantly enhances the sensitivity for all resonances that are subject to chemical or conformational exchange rates lower than the CPMG frequency. The F1 dimension resolves imino and amino resonances according to their respective ¹⁵N chemical shift. Thus, the diagonal peaks in the CPMG-NOESY correspond to the ¹H¹⁵N HSQC and the cross-peaks are NOE signals between these resonances (Figure 7). In the direct ¹H dimension, additional NOE cross-peaks are often observed, most prominently U:N3-A:H2 and C:N4-C:H5. The 2D ¹H¹⁵N CPMG-NOESY thus complements information gained from 2D ¹H¹H NOESY, 2D ¹H¹⁵N imino-HSQC and 2D ¹H¹⁵N amino-HSQC spectra. Since conformational exchange due to bond rotation is different for the exocyclic amino groups of A, C, and G, respectively, the experiment should be recorded at different temperatures.

When using ¹⁵N-labeled RNA, the signals of the amino protons can be suppressed in a 2D ¹H¹H NOESY using a ¹⁵N-filter. This is particularly useful for the spectral region that shows the cross-peaks between the aromatic and ribose protons (F2: 9-7 ppm / F1: 6.5-4 ppm). A more resolved and cleaner spectrum can be obtained when the overlapping amino proton signals are suppressed (Figure 8). For the ¹⁵N-filter to work effectively, it must be used before and after the t1 time, respectively. This leads to a reduction in sensitivity.

A very suitable application of isotope-filter (X-filter) experiments is selectively labeled RNA. As described before in the introductory chapter, RNA can be easily labeled in a nucleotide-specific manner, (e.g. selective ¹⁵N-labeling of AU or GC or GU). For selectively labeled RNA, the 1D ¹H and 2D ¹H¹H correlation spectra can be simplified by specifically selecting either the ¹⁵N-bound (¹⁵N-edited) or ¹⁴N-bound (¹⁵N-filtered) proton signals via the X-filter modules. The X-filters can be applied in both the direct and/or the indirect dimension of 2D ¹H¹H correlation experiments. For example, the combination of X-filter 1D ¹H spectra enables the separate detection of U (¹H¹⁵N, ¹⁵N-edited) and G (¹H¹⁴N, ¹⁵N-filtered) signals for an 48mer RNA (Figure 9), thereby facilitating assignment of the imino proton resonances from crowded regions in the non-filtered 1D ¹H spectra.

A double-half-X-filter NOESY (G. Otting and K. Wüthrich 1989) combines all possible variations of the X-filter modules in F1 and F2, leading to 4 distinct 2D ¹H¹H NOESY spectra (¹⁴N-¹⁴N; ¹⁴N-¹⁵N; ¹⁵N-¹⁴N; ¹⁵N-¹⁵N). Importantly, in the ¹⁴N-¹⁵N and ¹⁵N-¹⁴N-combinations, the NOESY spectra are diagonal-free, which offers additional resolution for NOEs close to the diagonal (Figure 10). This experiment is especially useful for selectively labeled RNA-samples, in order to reduce spectral overlap and resolve assignment ambiguities.

The 3D ¹H¹⁵N¹H NOESY-SOFAST-HMQC shows correlations to all protons within NOE vicinity (maximum of about 5-6 Å) of G:H1 and U:H3 while providing additional resolution in the imino region via the ¹⁵N dimension (in F2) (Figure 11). This possibly resolves ambiguities in crowded regions of the 2D ¹H¹H imino-NOESY. To enhance resolution in the ¹⁵N dimension, the HMQC plane should be optimized by “folding” the signals via reduction of the spectral width in F2. In favorable cases, the ¹⁵N spectral width can be reduced below 9 ppm.

Insight into the dynamic landscape of RNA structures is essential to understand its function, especially in cases where multiple, interconverting conformations (short-lived and long-lived) are present. For ¹⁵N-labeled samples, the imino groups of the bases (in Watson-Crick base-pairs: G:N1-H1 and U:N3-H3) can serve as reporters to study the structural dynamics of the RNA. Since these are present in the base-paired regions of RNA, they primarily probe the dynamics associated with base-pair opening and refolding events (occurring in the µs-s time regime). Apart from that, fast motions (on a timescale <ns) can be observed which can either be associated with changes in the hydrogen-bond geometry or with tumbling of motionally decoupled folded domains. All of these dynamics can be investigated by several NMR experiments which provide manifold information at atomic resolution (reviewed in Bothe et al). The T1, T2, T1rho, heteronuclear NOE (hetNOE) and ZZ-exchange are notable experiments that can be used to derive populations and rate constants of the dynamics and motions associated with conformational changes. Additional ¹³C-labeling would allow to probe the dynamics using the aromatic and ribose reporter signals (Chapter on ¹³C-labeled RNA), thus allowing to investigate both the structured and unstructured motifs in the RNA. Together, the T1, T2 (or T1rho) relaxation rates and the hetNOE values provide a comprehensive picture of the dynamic processes for the rigid and flexible regions covering a broad timescale. The combined data from these experiments is typically further analyzed using a model-free approach (Lipari and Szabo); yielding the order parameter (S², ranging between 0 for disordered and 1 for ordered) and time constants describing the internal motion and exchange contribution.

The ¹⁵N{¹H}-hetNOE experiment provides information on the fast dynamics (ps-ns timescale) by measuring the ¹H¹⁵N dipole-dipole interactions and is a key experiment to probe RNA flexibility. The experiment consists of two 2D ¹H¹⁵N HSQC-type experiments: one in presence of proton saturation and one reference spectrum in absence of saturation. Both spectra are acquired in an interleaved fashion (as pseudo 3D). The hetNOE value is defined as the ratio of the signal intensities (or volumes): hetNOE = I_saf/I_ref. For the ¹⁵N{¹H}-hetNOE, the values typically range from 0.8 to 1.0 for rigid regions, while lower or negative values indicate significant internal mobility.

The ¹⁵N-T1 experiment measures the spin-lattice (longitudinal) relaxation time. The experiment involves a series of 2D ¹H¹⁵N HSQC-type experiments with varying relaxation delay times. Typically, 8 or more spectra (with delays ranging from 100 ms to 1-2 s) are recorded in a pseudo-3D fashion where the intensity (or volume) of the signals decrease exponentially over time (I_t = I₀ × e^-t/T1). The ¹⁵N-T1 relaxation time is obtained by fitting of the data to the exponential decay function. Generally, a mono-exponential behavior is observed, but in case of strong anisotropic motions non-exponential signal decay behavior can occur. The obtained relaxation times reflect on the fast internal motions which occur on the ps-ns timescale. Hereby, a shorter T1 time (faster recovery delay) corresponds to a a higher relaxation rate (R1 = 1/T1) and indicates higher mobility.

The ¹⁵N-T2 experiment measures the spin-spin (transverse) relaxation time The experiment usually includes a Carr-Purcell-Meiboon-Gill (CPMG) pulse train. Analogous to ¹⁵N-T1, the ¹⁵N-T2 experiment also involves a series of 8 or more 2D ¹H¹⁵N HSQC-alike experiments with varying relaxation delay times (usually ranging from 0 to 100-200 ms), in this case generated by variable repetition of the CPMG element. Likewise, the ¹⁵N-T2 times are obtained by fitting of the data to a mono-exponential decay, but contributions from exchange processes or anisotropic rotational diffusion may lead to deviations. Unlike T1, the T2 relaxation time is sensitive to both fast (<ns) and slow (us-ms) motions (exchange) and is reduced by a larger molecular size (slower global tumbling). A longer T2 time (slower coherence delay) corresponds to a lower R2 rate and indicates higher mobility. Moreover, the T1/T2 ratios depend on the reorientational motion of the NH bond and thereby provide an estimate on the (local) rotational correlation times (τ_c). Increased T1/T2 ratios indicate longer τ_c values, which correspond to a slower tumbling and thereby a more rigid RNA region.

The ¹⁵N-T1rho experiment is closely related to the T2 experiment. Here, a spin-lock is applied in a rotating frame either performed on- or off-resonance. This spin-lock suppresses relaxation contributions from time-scales that correspond to the spin-lock frequency, thus effectively suppressing (conformational) exchange effects. Deriving T2 from T1rho requires corrections for the effective magnetic field angle using the off-resonance position of the spin-lock, spin-lock field strength, and T1.

The ¹H¹⁵N ZZ-exchange experiment resolves the interconversion in the slow exchange regime (dynamics on the ~ms-s timescale), when distinct signals are observable for each conformational state. In the ZZ-exchange experiment, additional ZZ cross-peak signals are detected between the signals of the interconverting conformations. This is a result from the exchange of Z magnetization during the mixing time. Careful optimization of this ZZ mixing time is crucial to pick up the interconversions on their distinct timescales. For quantitative analysis of the associated populations and rate constants, recording a series of 2D spectra with different mixing times is required.

Analysis to obtain the rates, values and other dynamic parameters for the hetNOE, T1, T2, T1rho and ZZ-exchange can be conveniently performed using the Bruker Dynamics Center. It provides an integrated workflow that covers all steps, including peak picking, signal integration, automated or user‑guided data fitting, and plot generation and also allows for the import of existing peak assignments in various formats.

Residual dipolar couplings (RDCs) offer a unique method for the validation of an proposed RNA structure. The RDCs of a a freely tumbling RNA are averaged to zero, because RDCs are dependent on the orientation of the bond vector in relation to the magnetic field of the spectrometer (Tjandra et al.). Therefore an alignment in the sample medium is needed to measure RDCs. (Hansen et al.). When the RDCs are observable in an aligned medium, they are however part of the ¹J-coupling. It is therefore necessary to obtain two samples one without aligned medium and one with aligned medium to determine the difference in coupling, originating from the RDC contribution.

To give a short summary of the most common alignment media for RNA samples:

Pf1 phages are rod-shaped viruses are the most widely used alignment medium for RNA. They are easy to handle and are tolerant to high salt concentrations and pH shifts (Hansen et al., Bailor et al.). As it is the most common alignment medium for RDC, this tutorial focuses on their usage.

Bicelles are mixtures of macromolecules (DMPC (dimyristoyl-phosphatidylcholine) and DHPC (dihexanoyl-phosphatidylcholine)) that at certain ratios form a liquid crystalline medium of disc shaped particles (Ottiger et al., Tjandra et al.). They tend to be sensitive to ionic conditions, which can make additional optimization steps necessary. Low temperatures are also not favorable for measurements with this medium (Prestegard et al.)

A further method is the use of polyacrylamid gels (Sass et al., Tycko et al.) as a mechanic alignment medium. The advantages are chemical inertness and longterm stability, however the process of inserting the viscous solution in the NMR tube can be difficult.

Another liquid crystalline solution can be formed by a mixture of n-Alkyl-poly(ethylene glycol)/n-alkyl alcohol and glucopone/n-hexanol or glucopone/n-hexanol (PEG) (Rückert et al., Alvarez-Salgado et al.). Due to the uncharged nature of these molecules they are insensitive to pH and only affected by extreme salt concentrations. They are chemically stable and easy to handle, due to low viscosity.

RDCs are used to verify an existing structure (.pdb, .cif) or to filter and refine an existing collection of structural models. These structures might be derived directly from a database like the PDB, an MD simulation or a structure prediction program. Because RDCs depend on the relative orientation of N-H bonds with respect to the principal axis of the RNA aligned in the external B₀ field, the degree of alignment is crucial for reliable analysis.

Preparation of phage samples

To prepare reference and aligned samples, the final concentration of RNA in the two samples should be the same in both tubes. As HSQC and HMQC are very sensitive experiments, a concentration of 100-200 µM is sufficient. Both samples prepared in the usual buffer, while the aligned sample additionally contains the aligning component. In case of PF1 phages 8 mg/mL is a good starting point but optimal concentrations vary with the RNA size. Larger RNAs need a lower concentration of the alignment medium. The amount of aligning component is proportional to the RDC value, but too high concentrations might interfere with folding of the RNA or cause unwanted interactions and line broadening.

Unfortunately the alignment medium can be, a source of RNase contamination, and samples should therefore be measured shortly after preparation. It is important to measure both samples at the same spectrometer to reduce field depended influence on the RDC.

Before measuring actual RDCs, a 1D ²H is recorded to check the proper alignment of the RNA sample. Due to the anisotropic property of the aligned sample, the solvent signal of D₂O (²H₂O) in a 1D ²H is a doublet. A reasonable value for small RNAs is around 20-25 Hz, and smaller for larger RNAs. The size of the coupling constant depends on the degree of alignment and therefore on the concentration of Pf1 phages, as well as the temperature and the magnetic field. For comparison between data sets it therefore is relevant to always compare the respective ²H splitting (Hansen et al.).

Measurement and Analysis of RDCs

By comparing the coupling of two samples, one with aligned medium and one without alignment, also called a reference sample, the observable coupling of a peak changes. The reference sample with freely tumbling RNA only displays ¹J_N,H-coupling (J_ref), while in the aligned sample the dipolar coupling also contributes to the observed coupling: ¹J_N,H + ¹D_N,H (J_align). To determine the dipolar coupling both samples are measured, and the couplings are subtracted to yield only ¹D_N,H.

The experiments recorded for coupling determination are recorded as in-phase-anti-phase (IPAP) ¹H¹³C-IPAP-HSQC (¹H-¹³C coupling) or ¹H¹⁵N-IPAP-SFHMQC (¹H-¹⁵N-coupling) (Ottiger et al.). IPAP spectra have the advantage that they can be split into in- and anti-phase terms resulting in less overlap than a common HMQC without decoupling. By determining the different peak positions in all four spectra (two for reference and two in the aligned sample) the coupling constants can be determined (Figure 12).

Processing of IPAP spectra is done differently compared to other experiments. Initially the in‑phase and anti-phase components must be separated via “split ipap” in the command line. Afterwards the resulting spectra should be processed with 16 k to 32 k points in the F2 dimension and with 4 k in the F1 dimension, to allow a precise determination of the peak center and the coupling constant.

To determine the RDC (D), the measured J-coupling (J_align and J_ref) need to be subtracted as follows:

D = J_align - J_ref

For ¹⁵N‑labeled samples only the imino N-H bonds (G:H1, U:H3) are usable as amino protons show an averaged RDC due to their free rotation. As the imino protons are typically only measurable for rigid stem structures they give little insights into more dynamic regions. Here ¹³C labeling is advantageous as the A:C2H2, C5H5, C6H6, C8H8 and C1’H1’, can be investigated for a wealth of RDC parameters.

Once the RDC values are determined, they can be used in structure validation and orientation predictions. While a short stem, loop construct might be of little interest, a RNA with one or more junctions and or bulges can be an interesting target for RDC investigation, as the orientation of the different stems can be investigated.

One possible calculation program to evaluate the agreement of a structure with the RDC values is PALES. The RDC values together with a structure coordinate file are used as input and the RDCs are fitted to that structure. The overall agreement of the back calculated RDCs of the supplied structure with the experimental RDCs is given by PALES with the Corr R (Pearson’s linear correlation coefficient) and Q (Q-factor) values, but individual residue agreement is also given. PALES might not work, if the coupling values are too big or small. In this case the values can be scaled up or down by a uniform factor. Note that down scaling causes less error in the values, than up scaling. PALES is freely accessible (here) and a complete description of its functions has been published by Zweckstetter.

In the ¹⁵N-detected BEST-TROSY the direct detection occurs via the G:N1 and U:N3 (Schnieders et al.) (Figure 13). The main advantage compared to ¹H‑detected spectrum lies in the higher tolerance to imino proton exchange, yet the sensitivity is lower compared to ¹H-detected spectrum. The decreased signal broadening compared to the ¹H‑detected version with increasing RNA size, is especially interesting as signal overlap, which is hereby decreased, is one of the main limits of investigating larger RNA. However the sensitivity of this experiment increases with the external magnetic field, making it especially interesting for facilities with access to > 800 MHz spectrometers As the spectrum records the imino peaks (G:H1, U:H3) it is noteworthy that imino peaks of non canonical base-pairs tend to display broader signals (Schnieders et al.).

< Back to Overview

Next chapter: Assignment of ¹³C/¹⁵N-labeled RNA >