Comparison of alignment tensors generated for native tRNA(Val) using magnetic fields and liquid crystalline media

Abstract

Residual dipolar couplings (RDCs) complement standard NOE distance and J-coupling torsion angle data to improve the local and global structure of biomolecules in solution. One powerful application of RDCs is for domain orientation studies, which are especially valuable for structural studies of nucleic acids, where the local structure of a double helix is readily modeled and the orientations of the helical domains can then be determined from RDC data. However, RDCs obtained from only one alignment media generally result in degenerate solutions for the orientation of multiple domains. In protein systems, different alignment media are typically used to eliminate this orientational degeneracy, where the combination of RDCs from two (or more) independent alignment tensors can be used to overcome this degeneracy. It is demonstrated here for native E. coli tRNA(Val) that many of the commonly used liquid crystalline alignment media result in very similar alignment tensors, which do not eliminate the 4-fold degeneracy for orienting the two helical domains in tRNA. The intrinsic magnetic susceptibility anisotropy (MSA) of the nucleobases in tRNA(Val) was also used to obtain RDCs for magnetic alignment at 800 and 900 MHz. While these RDCs yield a different alignment tensor, the specific orientation of this tensor combined with the high rhombicity for the tensors in the liquid crystalline media only eliminates two of the four degenerate orientations for tRNA(Val). Simulations are used to show that, in optimal cases, the combination of RDCs obtained from liquid crystalline medium and MSA-induced alignment can be used to obtain a unique orientation for the two helical domains in tRNA(Val).

Figure 1

(A) Sequence and secondary structure and (B) tertiary structure model of native E. coli tRNA^Val (Vermeulen, et al. 2005). (A) Long dashes represent canonical Watson-Crick base pairs and the short dashed lines represent non-canonical or tertiary interactions. Imino groups of the residues in black circles are involved in tertiary interactions, and their ¹D_HN values were omitted from SVD analysis. The two helical domains are labeled as the acceptor arm and the anticodon arm.

Figure 2

Scatter plots of experimental one-bond ¹H-¹⁵N imino ¹D_HN values for native E. coli tRNA^Val in 10 mg/mL Pf1 bacteriophage plotted against ¹D_HN values in the presence of (A) fd or fd-9ah8 bacteriophage and (B) bicelle or bilayer alignment media. (A) Squares are the ¹D_HN values in wild-type fd and circles are the ¹D_HN values in the longer fd mutant, fd-9ah8. (B) Filled squares are the ¹D_HN values in DMPC/DHPC bicelles, filled circles are the ¹D_HN values in SDS-doped DMPC/DHPC bicelles and open squares are the ¹D_HN values in C₈E₅/1-octanol bilayers. The lower right inset of each graph shows error bars representing the ±1.5 Hz error in the ¹D_HN values, and the Pearson’s correlation coefficient (R_P) values are given in the upper left of each graph. The lines are the best linear fit for each plot.

Figure 3

The uracil imino regions from the 2D ¹H, ¹⁵N doublet selected, sensitivity enhanced HQSC spectra of uniformly ¹⁵N-labeled native E. coli tRNA^Val in (A) DMPC/DHPC bicelles and (B) SDS-doped DMPC/DHPC bicelles. Spectra were collected at 600 MHz and 38 °C. The lower right inset of each spectrum shows the downfield shifted s⁴U8 imino resonance. Imino assignments for each residue are indicated in B.

Figure 4

Sauson-Flaumsteed projection maps illustrating the orientations of the principal components of the molecular alignment tensor for native E. coli tRNA^Val derived from alignment by (A) various filamentous bacteriophage and (B) bilayer or bicelle alignment media. The positions of the principal components S_xx, S_yy and S_zz of the alignment are indicated on each map. The legend in the lower right of each map shows the symbols for the principal components for the various alignment media. Due to the high rhombicity for several of the alignment tensors, the orientations of S_zz and S_yy are exchanged. Sauson-Flaumsteed projection maps were generated from the SVD output using the program REDCAT (Valafar and Prestegard 2004). 50 solutions are plotted for each alignment tensor.

Figure 5

Scatter plot of the experimental ¹D_HN values of native E. coli tRNA^Val in 10 mg/mL Pf1 plotted against ¹D_HN values for magnetic alignment at 800 MHz (filled squares) and 900 MHz (filled circles). The solid and dashed lines correspond to the best-fit line of the Pf1 data to the 800 MHz and 900 MHz data, respectively. The lower right inset shows an error bar representing the ±1.5 and ±0.8 Hz errors in the ¹D_HN values obtained from Pf1 and MSA-induced alignment, respectively, and the R_P values are given in the upper left of each graph. The lines are the best linear fit for each plot.

Figure 6

Sauson-Flaumsteed projection maps illustrating the orientations of the principal components of the molecular alignment tensor derived from alignment by (A) 10 mg/mL Pf1 and (B) MSA-induced alignment using the normalized average RDC values measured at 800 and 900 MHz, ¹D_HN(800, 900). Areas of green, blue and red correspond to the orientations of the principal components of the order tensor S_xx, S_yy, and S_zz, respectively.

Figure 7

Relative orientations of the principal components of the molecular alignment tensors for native E. coli tRNA^Val resulting from Pf1 (red) and MSA-induced alignment (blue). The magnitude and orientation of S_xx, S_yy and S_zz were determined by SVD analysis of the ¹D_HN values using the program REDCAT. The lengths of S_xx, S_yy, and S_zz represent their magnitudes and are relative to the S_zz for Pf1 alignment. For visualization purposes, the lengths of S_xx, S_yy, and S_zz for the MSA-induced alignment tensor are amplified 5-fold.