2d Hp-cpmg-hsqc

Fig. 6.4 Pulse schemes for MQ [37] (a) and TROSY [45] (b) HCN experiments. The thin and thick bars represent nonselective 90° and 180° pulses, respectively. S = 1.60 ms forJcvhv or 1.25 ms forJC6/8H6/8; A = 15 ms. The band-selective pulses are set as follows (500 MHz): proton 4 ms band-selective 180o REBURP pulses centered at 5.7 ppm and 7.8 ppm for H1' and H6/8 correlations, respectively; 3.0 ms REBURP centered at 90 ppm for HsCNb (a) or 2.5 ms REBURP centered at 140 ppm for HbCNb correlations in b; 2.0 ms IBURP-2 on 15N positioned at 158 ppm.

GARP decoupling of 13C and 15N is used during detection. The pulses are applied along the x-axis unless otherwise specified. Phase cycling: <^ = x, -x; ^ = 2(x), 2(-x); ^ = 8(x), 8(-x); ^, = 4(x), 4(-x); <p5 = abba, where a = x, -x, -x, x, and b=-x, x, x, -x. In addition, (p2 is incremented in the States-TPPI manner to achieve quadrature detection in the F1 dimension. The gradient pulses are for purging purposes only and need not to be in any specific ratios; gradients applied for 1 ms with the strengths of 2-10 G/cm are sufficient.

magnetization, they can be dephased by passive scalar interactions with neighboring nuclei, resulting in substantial decrease of attainable sensitivity. Decoupling of these interactions using band-selective pulses and constant time evolution periods is mandatory if the highest possible sensitivity is to be obtained. Because of a rather low proton density in oligonucleotides, the use of MQ coherence proved to produce a significant sensitivity im-

provement in the HCN experiments for the intra-residual correlation of H1', C1', N9/N1, C6/C8 and H8/H6 nuclei [36-38]. Sensitivity increase by a factor of 3.5-6.5 when compared to the original SQ versions were observed in the HCN correlations with incorporated band-selective XH pulses (Fig. 6.4a) on a 40-nucleotide RNA aptamer [37]. The MQ-HCN experiment was also combined with a CCH-TOCSY sequence in order to improve resolution for the ribose resonance assignment [39]. Sensitivity enhancement by incorporating the gradient scheme that preserves both coherence transfer pathways [40, 41] was further applied to the MQ-HCN-CCH-TOCSY and MQ-HCN-CCH-COSY experiments [42]. To correlate sugar-to-base resonance in the cases of severe overlap in both C1'-H1' and C6/8-H6/8 regions, 2D and 3D bi-directional HCNCH experiments have been proposed [43] that involve as many as five different frequencies (H1', C1', N1/9, C6/8 and H6/8). The correlation is established by a rectangular arrangement of crossover and out-and-back peaks in the proton-carbon correlated spectrum and can be further resolved in the nitrogen dimension in a 3D experiment. The somewhat lower sensitivity of the experiment as compared to other HCN correlation schemes is compensated by the fact that information obtained is equivalent to up to three out-and-back HCN experiments.

The HMQC approach has also been employed in the measurements of 3J(H3'P) and 3J(C4'P) coupling constants using a 2D {31P} spin-echo difference constant-time experiment [44] to increase sensitivity by a factor of 1.5-2.4 in a 17 kDa DNA-protein complex.

TROSY-type enhancements (transverse relaxation-optimized spectroscopy, see Chapt. 10) have also been incorporated into the experiments for nucleic acids in order to achieve substantial reduction of the relaxation rates. For the TROSY to be efficient, the contributions of the CSA (chemical shift anisotropy) and DD (dipole-dipole) mechanisms to transverse relaxation must be comparable in size, and their tensors must have similar orientations. In nucleic acids, the highest reduction occurs for aromatic carbons at magnetic fields between 12 and 19 T [45], which corresponds to proton resonance frequencies of approximately 500-800 MHz. The efficiency of the TROSY approach based on the interaction of 13C CSA with XH-13C DD was demonstrated for the 1H-13C correlation experiment [46] and HCN experiments for sugar-to-base through-bond correlation [45] (Fig. 6.4b). Further sensitivity increase of about 20% (with 17 kDa Antennapedia homeodomain-DNA complex) in the HCN-type experiment was reported by incorporating [13C-13C]-TROSY of base carbon nuclei [47]. Since the CSA values of sugar carbons are low, TROSY does not provide significant sensitivity advantage for the sugar moiety, and selective multiple-quantum (MQ) experiments [37] are more sensitive for the sugar nuclei [45] even at very high magnetic fields (Fig. 6.5). The recently proposed MQ-TROSY-HCN pulse sequence [48] combines the advantages of MQ-HCN for sugar-to-base and TROSY-HCN for intra-base correlations in a single experiment.

The TROSY effect was also used in a relayed HCCH-COSY experiment to correlate adenine H2/H8 resonances in uniformly 13C-labeled RNA molecules [49], and significant sensitivity over the existing HCCH-TOCSY version was reported. Magnetization is transferred simultaneously in an out-and-back manner from H2 and H8 to the three aromatic carbon spins, C4, C5 and C6, establishing thus the connectivity within the adenine base spin system.

In principle, TROSY is not limited to interference between dipole-dipole coupling and CSA. Pervushin [50] proposed the use of the differential conformational exchange-in-

Fig. 6.5 The dependence of transverse relaxation rates on static magnetic field B0 calculated for a C6 of Cytosine, b C8 of Guanine and c C1' (average value for all nucleosides). Solid, dashed and

dotted lines represent the slowly relaxing (TROSY) component of the CH doublet, multiple quantum and single quantum coherence, respectively.

duced transverse relaxation (CSX) in the ZQ (zero quantum) and DQ (double quantum) coherences to optimize transverse relaxation properties in the experiments that correlate spins and measure scalar couplings across the hydrogen bonds in nucleic acids.

Most of the early methods for assigning isotopically labeled nucleic acids were developed using RNA samples, since 13C and 15N isotopes can be incorporated in RNA more easily than in DNA. While most of the magnetization transfer pathways in RNA and DNA are the same, the latter contains thymine rather than uracil. To remove the ambiguity of intra- and inter-residue H6-CH3 peak assignment in thymine, the HCCCH through-bond method was proposed [51] as a more sensitive alternative to NOE or ROE.

Similarly, the assignment of exchangeable hydrogens based on imino-imino and imino-amino sequential and intrabase NOEs often fails in nonhelical regions such as bulges and loops. This problem has been addressed by through-bond TOCSY-based experiments linking the imino/amino hydrogen with the nonexchangeable protons on the same base [52-54]. In the case of pyrimidine bases, better sensitivity was reported using an experiment to correlate their imino/amino protons with H5 using consecutive INEPT steps [55].

An improved method for the 1H-31P correlated experiment (HP-CPMG-HSQC) was proposed [56] achieving ~ 2.5-fold sensitivity increase as compared to phosphorus-excited methods. The effective correlation is achieved by the application of a CPMG pulse train in the XY-16 expansion scheme during the periods of polarization transfer. The CPMG sequence suppresses the effect of conformational exchange and cross-correlated relaxation while achieving in-phase transfer in the 1H-1H coupling network. The pulse sequence can be combined with an additional magnetization transfer step such as NOESY to extend the correlation to better resolved sugar or base proton resonances, e.g. H1- The method does not require an isotopically enriched sample. This approach shows several advantages over the previously published heteronuclear TOCSY experiment proposed by Kellog to obtain the identical information [57, 58].

The pulse sequences used for the resonance assignments are not discussed in this chapter. For further details the interested reader is referred to the paper of Ref. [2] and references therein.

Was this article helpful?

0 0

Post a comment