HNN-COSY

Fig. 9.1 Basic HNN COSY pulse sequence. Narrow and wide pulses correspond to flip angles of 90° and 180°, respectively, whereas low-power (water flip-back) 90° l-l pulses are illustrated as smaller narrow pulses. Delays 8 = 2.25 ms 7=15 ms (can be shorter or longer) (a = 2.5 ms fb = 0.25 ms fc= 2.25 ms fd = 0.5 ms. Unless indicated, the phase of all pulses are applied along...

Fig. 9.2 Part of a quantitative- /NN HNN-COSY spectrum of a 1.5 mM uniformly 13C/15N-labeled intramolecular DNA triplex. This triplex consists of five Hoogsteen-Watson-Crick T A-T and three Hoogsteen C+ G-C base triplets. The spectral region corresponds to the 10 imino resonances of the Hoogsteen-Watson-Crick T A-T triplets. The data matrix consisted of 250 (q) X1024 (i2) data points (where n refers to complex points) with acquisition times of 45 ms (tn) and 85 ms...

Several modifications have been proposed for the basic HNN-COSY experiment. For example, frequency separations between amino and aromatic 15N resonances are typically in the range 100-130 ppm and therefore much larger than between imino 15N donor and aromatic 15N acceptor resonances. As has been pointed out by Majumdar and coworkers [33], such 15N frequency separations are too large to be covered effectively by the non-selective 15N pulses of the homonuclear HNN-COSY. They therefore designed a pseudo-heteronuclear H(N)N-COSY experiment, where selective 15N pulses excite the amino and aromatic 15N resonances separately to yield excellent sensitivity [33]. An inconvenience of this experiment is that the resonances corresponding to the amino 15N nuclei are not detected, and a separate spin-echo difference experiment was used to quantify the h2/NN values. A slightly improved version of this pseudo-heteronuclear H(N)N-COSY [35] remedies this problem by the use of phase-coherent 15N pulses such that both amino and aromatic 15N resonances can be detected in a single experiment. 

Another ingenious modification of the HNN-COSY scheme involves the replacement of the homonuclear 15N-COSY transfer by a 15N-TOCSY transfer [37]. As the homonuclear TOCSY transfer is twice as fast as the COSY transfer, a significant sensitivity increase is achieved. The application is, however, limited to cases where the 15N donor and acceptor resonances are at similar frequencies, so that the power of the 15N-TOCSY radio frequency pulses needed is not too strong. A very interesting application of this scheme was presented for the sensitive detection of very small (0.14 Hz, Tab. 9.1) h4/NiNi couplings in N1-H1 06=C6-N1 H-bonds of guanosine tetrads. 

For a pair of H-bonded histidine side chains, Ii3/Nn couplings in the range 8-11 Hz (Tab. 9.2) have been observed between a protonated 15Ne2-nucleus and an unprotonated 15Ne2 using an HNN-COSY-based technique [48]. These values are similar to the h2/NN coupling values observed for the imino and amino N-l I Nlorii ,, lc H-bonds of nucleic acid base pairs (Tab. 9.1). 

Figure 11 2D HNN-COSY spectrum of the 38-nt RNA construct from the consensus stem D of the cloverleaf domain of enteroviruses sequence is shown in the inset. Labeled peaks in the upper part of the spectrum arise from the one-bond correlations of the NH imino groups. Cross-peaks in the lower part of the spectrum (not labeled) arise due to the scalar coupling (2Jnn of 5-7 Hz) between 15N nuclei across the NH-N hydrogen bonds in Watson-Crick AU and GC pairs. They have the opposite phase compared with the diagonal NH peaks. Reproduced with permission from Z. Du J. Yu N. B. Ulyanov R. Andino T. L. James, Biochemistry 2004, 43, 11959-11972, with permission from the American Chemical Society. Copyright (2004) American Chemical Society.

Fig. 14. a) Correlation spectrum correlating the aromatic protons of histidines with the resonances, b) HNN-COSY Spectrum correlat-... 

All the calibrations and tests required for 2D quantitative /nn HNN-COSY spectrum have been described in the first protocol. The second protocoP contains the description of the details of 2D long-range H(N)CO TROSY pulse sequence along with the reference experiment sequence which accounts for relaxation losses. 

---