Cross-Correlated Relaxation between H1' Chemical Shift Anisotropy

Koichi Kato , Hiroaki Sasakawa , Yukiko Kamiya , Maho Utsumi , Michiko Nakano , Noriko Takahashi , Yoshiki Yamaguchi. Biochimica et Biophysica Acta (B...
0 downloads 0 Views 69KB Size
5

2006, 110, 5-8 Published on Web 12/09/2005

Cross-Correlated Relaxation between H1′ Chemical Shift Anisotropy and H1′-H2′ Dipolar Relaxation Mechanisms in Ribonucleosides: Application to the Characterization of Their Anomeric Configuration Kumar Pichumani, Tilak Chandra, Xiang Zou, and Kenneth L. Brown* Department of Chemistry and Biochemistry, Ohio UniVersity, Athens, Ohio 45701 ReceiVed: October 10, 2005; In Final Form: NoVember 15, 2005

Cross-correlated nuclear spin relaxation between 1H chemical shift anisotropy (CSA) and 1H-1H dipolar relaxation mechanisms in ribonucleosides in solution phase are observed and used to identify their anomeric configuration. Only R-ribonucleosides showed the presence of cross-correlated spin relaxation through differential spin-lattice relaxation (T1) of the H1′ doublet. Dependence of the magnitude and the orientation of the H1′ CSA tensor values on the glycosidic torsion angle and the fast time-scale internal motions present in the ribose moiety play a significant role in the characterization of the anomeric configuration of the nucleosides via cross-correlated relaxation.

Introduction Naturally occurring ribonucleosides are in β anomeric configuration, with the exception of the vitamin B12 family of compounds in which an R-ribonucleoside is included in the structure.1 There have been a limited number of NMR studies of the R-ribonucleosides aimed at understanding their role in overall nucleic acid structure and in the vitamin B12 molecule. It has been shown using Dreiding models and experimentally that R-deoxyoligonucleotide segments can also form regular Watson-Crick base pairs in the antiparallel or parallel duplexes and that such duplexes are resistant to cell nucleases and block the functions of reverse transcriptase.2-4 Such properties suggest that R-anomeric nucleotides have potential as antisense and antiviral chemotherapeutic agents. R-Nucleosides are also being used in the design of fluorescent probes.5 Consequently, structural and stereochemical characterization of the anomeric configuration of ribonucleosides and their dynamics are crucial to an understanding of their biological function and conformational and thermodynamic stability. Also, such knowledge could be helpful to understand why nature prefers only β-anomers to build nucleic acids. Cross-correlations between different relaxation mechanisms in NMR have widespread applications in structure determination and dynamics of biomolecules.6-18 This includes two studies using the 15N/13C CSA and C-H dipolar cross-correlations in labeled RNA molecules (β-anomeric configuration).19,20 Ravindranathan et al.19 discussed the role of internal motions of the ribose on the dipole-CSA cross-correlations between the 15N CSA in the base of the nucleotide and the C-H dipolar interaction located in the ribose unit. In this communication, we investigate H1′-H2′ Dipolar and H1′ chemical shift anisotropy (CSA) cross-correlated NMR relaxation measurements on * Corresponding author. E-mail: [email protected] (Kenneth L. Brown).

10.1021/jp055774m CCC: $33.50

various ribonucleosides as model compounds and their use as a direct and convenient method to determine the anomeric configuration of ribonucleosides. Also, we analyze how the fast time-scale internal motions in the ribonucleosides due to pseudorotation in the ribose moiety modulate the H1′-H2′ dipolar-CSA (H1′) interaction in a distinct way, so that nucleosides in the R configuration show significant detectable cross-correlation effects in the proton nuclear spin-lattice relaxation, whereas no such effects are observed for β-nucleosides. To our knowledge, this is the first comparative study of (H1′-H2′) dipolar-CSA (H1′) cross-correlated relaxation measurements in ribonucleosides in both the R and β anomeric configurations. Experimental Section The standard T1 inversion recovery experiment was used with a small flip angle mixing pulse (15°). During the mixing time 2′ of this experiment, longitudinal two spin-order H1′ z Hz is cre1′ ated by the cross-correlation between H CSA and H1′-H2′ dipolar relaxation and is only observed by the use of a small flip angle (∼15°) mixing pulse as the differential longitudinal relaxation of each component of the H1′ doublet.17 It is important to note that, with the 90° mixing pulse used in the conventional experiment, two spin-order is converted into multiple quantum coherences and cannot be observed. The difference between the intensities of the doublet components is directly related to the cross-correlation rate (∆1′,1′2′) by which the two spin-order is created by the H1′ CSA and H1′-H2′ dipolar relaxation mechanisms and is given by the following expression for the isotropic molecular reorientation:

∆1′,1′2′ ) -

(401 )ω D H

g 1′2′(∆σ)1′

© 2006 American Chemical Society

τc (1 + ω2τc2)

6 J. Phys. Chem. B, Vol. 110, No. 1, 2006

Letters

Figure 1. Schematic representation of R and β ribonucleosides studied in this article.

where

( )

µo 2γH2 D1′2′ ) p 4π r3 1′2′ and

1 g ) (3 cos2 θ - 1)∆σ1′ (∆σ)1′ 2 r1′2′ is the internuclear distance between the H1′ and H2′ protons, θ is the angle between the principal axis of the H1′ CSA tensor and the H1′-H2′ internuclear vector, γH is the proton gyromagnetic ratio, ωH is the Larmor frequency of the 1H nuclear spin, g and τc is the total correlation time of the molecule. (∆σ)1′ is the geometry-dependent chemical shift anisotropy of the H1′ proton (an axially symmetric CSA tensor is assumed). The 1H T1 inversion recovery experiment with the mixing pulse of 15° and the T1 and T2 experiments were performed on three ribonucleosides of R and β anomeric configurations whose structures are shown in Figure 1. The nucleosides were synthesized as described in the literature,21-24 and 1H and 13C assignments were made using COSY, NOESY, HMQC, and previous NMR data available on similar molecules.21-24 As 2′ described above, the amount of the H1′ z Hz created by dipoleCSA cross-correlations between H1′-H2′ and H1′ CSA is calculated from the small mixing pulse T1 inversion recovery experiment data for the three R and β ribonucleosides. Results and Discussion The spectra showing the relaxation behavior at different mixing times in the inversion recovery experiment for the H1′ proton doublet for the R- and β-dimethylbenzimidazole (2a and 2b in Figure 1) are shown in Figure 2. The amount of the 2′ H1′ z Hz created by dipole-CSA cross-correlations between 1′ H -H2′ and H1′ CSA, R1, and R2 values of the H1′ doublets are given in Table 1. Since the R1 and R2 values of both components of the doublet are almost the same (within experimental error), 1′ we have given the average values. In the table, the Hz0 , the 1′ equilibrium magnetization of the H spin, is calculated as the sum of the intensities of the H1′ doublet, and the two spin-order is normalized with respect to it. The values of the normalized two spin-order clearly show that there is no effect of dipoleCSA cross-correlations in the ribonucleosides in the β-anomeric configuration. The dipole-CSA cross-correlation rate depends on (1) the internuclear distance between H1′ and H2′ (dipolar interaction), (2) the magnitude of the H1′ CSA tensor and its orientation with respect to the internuclear dipolar vector, and (3) the dependence of the H1′ CSA tensor on internal motions due to the pseudorotation of the ribose moiety. The internculear distance for the R-

Figure 2. 1H inversion recovery spectra of a mixture of 2a and 2b (80% and 20%) for different mixing times. The H1′ differential recovery for the R-anomer of 2a indicates the presence of dipole-CSA crosscorrelations, whereas the β anomeric proton doublet shows no such effect. The H1′ chemical shift and JH1′-H2′ values of 2a and 2b are 6.48 ppm, 4.95 Hz and 6.13 ppm, 4.24 Hz, respectively. Experiments were done on an INOVA 500 MHz Varian NMR spectrometer at 298 K. This sample was dissolved in deuterated CD3OD.

TABLE 1: R1 and R2 Values of H1′ and Normalized 2′ Proton-Proton Two Spin-Order H1′ z Hz Values for the Three Different Ribonucleosides in Both r and β Configurations Shown (Figure 1) anomeric ribonucleoside configuration 1a 1b 2a 2b 3a 3b

R β R β R β

R1 (s )

R2 (s )

2′ (H1′ z Hz )/ 1′ (Hz0)‚100

1.05 ( 0.08 0.93 ( 0.06 1.09 ( 0.09 0.96 ( 0.06 0.93 ( 0.05 0.83 ( 0.04

6.28 ( 0.24 7.95 ( 0.30 5.14 ( 0.21 7.75 ( 0.26 6.20 ( 0.31 8.12 ( 0.44

9.3 none 7.06 none 5.89 none

-1

-1

and β-anomers of similar ribonucleosides are known from their X-ray structures,24 and the estimated ratio of (D1′2′)R/(D1′2′)β is 2.83, providing one of the reasons why cross-correlation effects are seen to be stronger for the R-ribonucleosides. However, the dipole-CSA cross-correlation rate also strongly depends on the magnitude and the orientation of the H1′ CSA tensor, (3 cos2 θ -1). Different H1′ chemical shift values of R- and β-anomers hint at the differences in the magnitude of the CSA tensor. The orientation of the tensor with respect to the proton-proton dipolar vector largely depends on the geometry of the molecule, and the crystal structure for similar ribonucleosides suggests that it may be close to the magic angle for the β anomer, which makes the dipole-CSA cross-correlation rate very small. Exact proton CSA tensor values require solid-state NMR methods, which are in progress. While this manuscript was in preparation, an excellent article appeared in the latest web edition of J. Am. Chem. Soc. by Sychrovsky et al., which shows how the C1′ and N1/9 CSA tensors and their corresponding cross-correlated relaxation rates depend on the glycosidic torsion angle and sugar pucker motions in 2′-deoxynucleosides using quantum chemical calculations. Similarly in our case, R- and β-anomers have significantly different glycosidic torsion angles (the difference is roughly 5060°), and we therefore expect a similar and significant dependence in the H1′ CSA tensor values resulting in a significant amount of dipole-CSA cross-correlations only for R anomers. It is known that the sugar moieties in nucleosides are involved in a dynamically interconverting two-state conformational

Letters

Figure 3. 1H inversion recovery spectra of the molecule 1a in a 70% CD3OD/30% glycerol mixture. Experiments were done on an INOVA 500 MHz Varian NMR spectrometer at 298 K. The H1′ chemical shift and JH1′-H2′ values are 6.57 ppm and 5.44 Hz, respectively.

equilibrium, C(2′)-exo and C(3′)-endo (N-type) and C(2′)-endo and C(3′)-exo (S-type), which occurs on the nanosecond time scale at room temperature.19,26,27 The presence of such fast timescale motions may be evident from the significantly lower H1′ R2 values (Table 1) of the R anomeric proton, which corresponds to the increased presence of fast time-scale motions. We performed the small mixing angle inversion recovery experiment on the vitamin B12 molecule (it is a ribonucleotide complex with R-dimethylbenzimidazole (Figure 2a) as a nucleoside) to see if there are any dipole-CSA cross-correlation effects on the H1′ of the ribonucleoside due to its inclusion in the structure of the vitamin B12, which significantly reduces the fast time-scale motions present in the free form. The amount of 2′ normalized proton-proton two spin-order H1′ z Hz in this case was determined to be much smaller (1.08%) compared to 7.06% (Table 1) when in its free form as a ribonucleoside. This decreased dipole-CSA cross-correlation effect must be partly due to the fact that fast time-scale motions in the ribose moiety become highly restricted when it is included in the structure of the vitamin B12 molecule.1 Further, we performed the inversion recovery experiments on the R-benzimidazole ribonucleoside in 30% glycerol to see if there was any effect of the slower rotational correlation time due to the solvent viscosity, and the results are shown in Figure 3. We observed the existence of dipole-CSA cross-correlations via a similar differential longitudinal relaxation pattern for the H1′ doublet, which remained resolved despite a small increase in the 1H line widths. This can be interpreted to mean that fast time-scale internal motions are not affected by the change in viscosity, but the effective overall motion has been attenuated, although not enough to overlap the doublet signal. The existence of internal motions at higher viscosity has been observed on the 13C line shapes of ATP molecules and also for the rotation of the phenyl group of ethidium bromide using picosecond timeresolved fluorescence spectroscopy.28,29 This suggests that the internal motion is not controlled by the static viscosity of the

J. Phys. Chem. B, Vol. 110, No. 1, 2006 7 solvent but by the high-frequency components of the frictional forces of the viscous solvent. In addition to the internal motions, nucleic acids and nucleosides are known to have a large amount of motional anisotropy. This has led us to conclude that the 1H-1H dipolar and 1H CSA cross-correlated relaxation observed in this study is strongly influenced by fast time-scale internal motions in the ribose sugar rather than the overall molecular motion and therefore contributes significantly to longitudinal cross-correlation rates for R-ribonucleosides but not for βribonucleosides, where such motions may be lesser in magnitude. It is important to note that longitudinal cross-correlations are more sensitive to faster molecular motions because they affect only single quantum transition probabilities (in the case of weakly coupled spin systems). Therefore, it may be argued that R-ribonucleoside molecules are relatively less rigid than their β anomeric counterparts where we do not see any crosscorrelation effects. It is known that contributions from polypeptide backbone fluctuations to the conformational entropy and heat capacity of globular proteins result in higher thermal stability of these proteins.30 In a similar fashion, the internal motions present in the R-ribose moiety could contribute to the conformational entropy/heat capacity of the rarely found Rribonucleosides, both in their free form as well as in the structure of vitamin B12 and nicotinamide adenine dinucleotide, thus enhancing their structural and thermodynamic conformational stability. Ab initio and quantum chemical studies to investigate the H1′ CSA dependence on the sugar pucker and glycosidic torsion angle and solid-state NMR studies to estimate the 1H CSA tensor components, and more relaxation studies using anisotropic motional models would be needed to substantiate the exact reasons for our experimental observations. Conclusion It has been shown that cross-correlated relaxation between dipolar and 1H CSA in three ribonucleosides is observed and can be related to their anomeric configuration. In addition to the differences in the strength of the proton dipolar relaxation, dependence of H1′CSA tensor magnitude and the orientation on the glycosidic torsion angle as well as the fast time-scale internal motions present in the ribose might be selectively reinforcing the observable presence of cross-correlations in the R-ribonucleosides and also contribute to thermodynamic conformational stability in the structure of the vitamin B12 molecule.

1H-1H

Acknowledgment. We thank Prof. Anil Kumar, Indian Institute of Science, Bangalore, India, and Prof. B. D. N. Rao, Indiana University-Purdue University at Indianapolis (IUPUI) for their useful discussions. This research was supported by the National Institute of General Medical Sciences grant GM48858 (to K.L.B.). References and Notes (1) Brown, K. L. Chem. ReV. 2005, 105, 2075. (2) Morvan, F.; Rayner, B.; Imbach, J.-L.; Lee, M.; Hartley, J. A.; Chang, D.-K.; Lown, J. W. Nucleic Acids Res. 1987, 15, 3421. (3) Aramini, J. A.; Van De Sande, J. H.; Germann, M. W. Biochemistry 1996, 35, 9355. (4) Aramini, J. A.; Kalisch, B. W.; Pon, R. T.; Van De Sande, J. H.; Germann, M. W. Biochemistry 1997, 36, 9715. (5) Bielecki, L.; Skalski, B.; Zagorowska, I.; Verrall, R. E.; Adamiak, R. W. Nucleosides Nucleotides Nucleic Acids 2000, 19, 1735. (6) Kumar, A.; Grace, R. C. R.; Madhu, P. K. Prog. Nucl. Magn. Reson. Spectrosc. 2000, 37, 191. (7) Fushman, D.; Cowburn, D. Methods Enzymol. 2001, 339, 109. (8) Levitt, M. H.; DiBari, L. Phys. ReV. Lett. 1992, 69, 3124. (9) Wang, T.; Cai, S.; Zuiderweg, E. R. P. J. Am. Chem. Soc. 2003, 125, 8639.

8 J. Phys. Chem. B, Vol. 110, No. 1, 2006 (10) Dittmer, J.; Kim, C.-H.; Bodenhausen, G. J. Biomol. NMR 2003, 26, 259. (11) Pervushin, K.; Riek, R.; Wider, G.; Wu¨thrich, K. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 12366. (12) Fischer, M. W. F.; Zeng, L.; Majumdar, A.; Zuiderweg, E. R. P. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 8016. (13) Kumar, P.; Kumar, A. J. Magn. Reson., Ser. A 1995, 115, 155. (14) Kumar, P.; Kumar, A. J. Magn. Reson., Ser. A 1996, 119, 29. (15) Werbelow, L. J. Phys. Chem. 1990, 94, 6663. (16) Tjandra, N.; Bax, A. J. Am. Chem. Soc. 1997, 119, 8076. (17) Dalvit, C.; Bodenhausen, G. Chem. Phys. Lett. 1989, 161, 554. (18) Pichumani, K. Cross Correlation studies in relaxation of coupled spins in NMR. Ph.D. Thesis. Indian Institute of Science, Bangalore, India, 1996. (19) Ravindranathan, S.; Kim, C. H.; Bodenhausen, G. J. Biomol. NMR 2003, 27, 365. (20) Duchardt, E.; Richter, C.; Ohlenschla¨ger, O.; Go¨rlach, M.; Wo¨hnert, J.; Schwalbe, H. J. Am. Chem. Soc. 2004, 126, 1962.

Letters (21) Brown, K. L.; Chandra T.; Zou, S.; Valente, E. J. Nucleosides Nucleotides Nucleic Acids 2005, 24, 1147. (22) Chandra, T.; Brown, K. L. Tetrahedron Lett. 2005, 46, 2071. (23) Chandra, T.; Zou, X.; Brown, K. L. Tetrahedron Lett. 2004, 45, 7783. (24) Mukaiyama, T.; Hashimoto, Y.; Hayashi, Y.; Shoda, S. Chem. Lett. 1984, 557. (25) Ghose, A. K.; Sanghvi, Y. S.; Larson, S. B.; Revankar, G. R.; Robins, R. K. J. Am. Chem. Soc. 1990, 112, 3622. (26) Plavec, J.; Roselt, P.; Fo¨ldesi, A.; Chattopadhyaya, J. Magn. Reson. Chem. 1998, 36, 732. (27) Altona, C.; Sundaralingam, M. J. Am. Chem. Soc. 1973, 95, 2333. (28) Rao, B. D. N.; Ray, B. D. J. Am. Chem. Soc. 1992, 114, 1566. (29) Sommer, J. H.; Nordlund, T. M.; McGuire, M.; McLendon, G. J. Phys. Chem. 1986, 90, 5173. (30) Seewald, M. J.; Pichumani, K.; Stowell, C.; Tibbals, B. V.; Regan, L.; Stone, M. J. Protein Sci. 2000, 9, 1177.