2H-13C correlation solid-state NMR for investigating dynamics and water accessibilities of proteins and carbohydrates

Abstract

Site-specific determination of molecular motion and water accessibility by indirect detection of ²H NMR spectra has advantages over dipolar-coupling based techniques due to the large quadrupolar couplings and the ensuing high angular resolution. Recently, a Rotor Echo Short Pulse IRrAdiaTION mediated cross polarization (^RESPIRATIONCP) technique was developed, which allowed efficient transfer of ²H magnetization to ¹³C at moderate ²H radiofrequency field strengths available on most commercial MAS probes. In this work, we investigate the ²H-¹³C magnetization transfer characteristics of one-bond perdeuterated CD _n spin systems and two-bond H/D exchanged C-(O)-D and C-(N)-D spin systems in carbohydrates and proteins. Our results show that multi-bond, broadband ²H-¹³C polarization transfer can be achieved using ²H radiofrequency fields of ~50 kHz, relatively short contact times of 1.3-1.7 ms, and with sufficiently high sensitivity to enable 2D ²H-¹³C correlation experiments with undistorted ²H spectra in the indirect dimension. To demonstrate the utility of this ²H-¹³C technique for studying molecular motion, we show ²H-¹³C correlation spectra of perdeuterated bacterial cellulose, whose surface glucan chains exhibit a motionally averaged C6 ²H quadrupolar coupling that indicates fast trans-gauche isomerization about the C5-C6 bond. In comparison, the interior chains in the microfibril core are fully immobilized. Application of the ²H-¹³C correlation experiment to H/D exchanged Arabidopsis primary cell walls show that the O-D quadrupolar spectra of the highest polysaccharide peaks can be fit to a two-component model, in which 74% of the spectral intensity, assigned to cellulose, has a near-rigid-limit coupling, while 26% of the intensity, assigned to matrix polysaccharides, has a weakened coupling of 50 kHz. The latter O-D quadrupolar order parameter of 0.22 is significantly smaller than previously reported C-D dipolar order parameters of 0.46-0.55 for pectins, suggesting that additional motions exist at the C-O bonds in the wall polysaccharides. ²H-¹³C polarization transfer profiles are also compared between statistically deuterated and H/D exchanged GB1.

Keywords: Cellulose; Molecular motion; Plant primary cell walls; RESPIRATIONCP; Trans-gauche isomerization.

Figure 1.

2D ²H-¹³C correlation pulse sequence, involving ²H RESPIRATION excitation, ²H t₁ evolution, ^RESPIRATIONCP from ²H to ¹³C, and ¹³C detection.

Figure 2.

Simulated ²H quadrupolar spectra for varying quadrupolar coupling constants, MAS frequencies, and asymmetry parameters. (a) Simulated spectra for 15 and 20 kHz MAS for C_Q values from 50 kHz to 250 kHz. (b) Simulated spectra for C-D order parameters from 0.20 to 0.50. (c) Simulated spectra as a function of η for C_Q = 170 kHz.

Figure 3.

²H-¹³C polarization transfer efficiencies η_D→C as a function of ^RESPIRATIONCP contact time for one-bond ¹³C-²H and two-bond ¹³C-O-²H spin systems. The transfer efficiencies, indicated on the left y-axis, are related to the enhancement factors I_D→C/I_C or I_D→C/I_H→C, shown on the right y-axis, according to Eq. 4 and 5. (a) Polarization transfer of ²H, ¹³C-labeled bacterial cellulose. (b) Polarization transfer of CDN-labeled Val. (c) Polarization transfer of ²Hβ-labeled Ala. (d) Polarization transfer of ¹³C-labeled and H/D exchanged D-glucose and Arabidopsis cell walls. The data were obtained under 15 kHz or 10 kHz MAS with the indicated short-pulse ^RESPIRATIONCP field strengths.

Figure 4.

(a) Representative 2D ²H-¹³C correlation spectrum of mixed CDN-labeled Val and ²Hβ-labeled Ala, measured under 15 kHz MAS with a ^RESPIRATIONCP field strength of 62.5 kHz. (b) ²H cross sections of Val Cα and Cγ as a function of the short-pulse ^RESPIRATIONCP field strength. The 62.5 kHz cross sections (red) are overlaid with the 50 and 35 kHz cross sections to illustrate differences in sideband intensities. The 35 kHz spectrum shows intensity distortions compared to the 62.5 and 50 kHz spectra.

Figure 5.

2D ²H-¹³C correlation spectra of ²H, ¹³C-labeled bacterial cellulose. (a) Chemical structure of cellulose. (b) The ²H-¹³C ^RESPIRATIONCP spectrum has the same intensity distribution as the ¹³C DP spectrum. (b) 2D ²H-¹³C correlation spectrum, measured under 15 kHz MAS. (c) ²H cross sections of iC4, sC4, iC6, and sC6. Best-fit simulations (red) for $\overline{\eta }=1$ indicate that the surface cellulose C6 is motionally averaged at 293 K. Alternative fit assuming $\overline{\eta }=0$ gives a similar ${\overline{C}}_Q$ value, but does not match the experimental spectrum, indicating that the spectrum is sensitive to the asymmetry parameter of motion.

Figure 6.

²H-¹³C correlation spectra of ¹³C-labeled and H/D exchanged D-glucose. (a) Comparison of ¹H-¹³C CP and ²H-¹³C ^RESPIRATIONCP spectra. The C5 signal is suppressed in the ²H-¹³C CP spectrum due to its lack of directly bonded OD. (b) 2D ²H-¹³C correlation spectrum, measured under 15 kHz MAS. (c) ²H cross sections of C1 (93 ppm) and C6 (64 ppm). Best-fit simulations give rigid-limit O-D quadrupolar couplings.

Figure 7.

²H-¹³C correlation spectra of ¹³C-labeled and H/D exchanged Arabidopsis cell wall. (a) Comparison of ¹H-¹³C CP and ²H-¹³C ^RESPIRATIONCP spectra. (b) 2D ²H-¹³C correlation spectrum, measured under 15 kHz MAS at 273 K. Cellulose structure is shown. (c) 72-ppm ²H cross sections of C2 and C5. Best-fit simulation was obtained with two $\overline{C_Q}$ values of 187 kHz and 50 kHz with weighting factors of 74% and 26%, respectively. For comparison, the 1D ²H MAS spectrum with RESPIRATION-4 excitation is shown. (d) 2D RMSD contour plot for determining the best-fit quadrupolar couplings (marked by a white cross) for the C2 and C5 cross section.

Figure 8.

(a) ¹H-¹³C CP and ²H-¹³C ^RESPIRATIONCP spectra of uniformly ¹³C,¹⁵N labeled and 70% ²H-labeled GB1. (b ¹H-¹³C CP and ²H-¹³C ^RESPIRATIONCP spectra of ¹³C-labeled and H/D exchanged GB1. The different intensity patterns of the ²H-¹³C spectra are consistent with the distinct deuteron distributions in the two samples.