Multidimensional High-Resolution Magic Angle Spinning and Solution-State NMR Characterization of (13)C-labeled Plant Metabolites and Lignocellulose

Abstract

Lignocellulose, which includes mainly cellulose, hemicellulose, and lignin, is a potential resource for the production of chemicals and for other applications. For effective production of materials derived from biomass, it is important to characterize the metabolites and polymeric components of the biomass. Nuclear magnetic resonance (NMR) spectroscopy has been used to identify biomass components; however, the NMR spectra of metabolites and lignocellulose components are ambiguously assigned in many cases due to overlapping chemical shift peaks. Using our (13)C-labeling technique in higher plants such as poplar samples, we demonstrated that overlapping peaks could be resolved by three-dimensional NMR experiments to more accurately assign chemical shifts compared with two-dimensional NMR measurements. Metabolites of the (13)C-poplar were measured by high-resolution magic angle spinning NMR spectroscopy, which allows sample analysis without solvent extraction, while lignocellulose components of the (13)C-poplar dissolved in dimethylsulfoxide/pyridine solvent were analyzed by solution-state NMR techniques. Using these methods, we were able to unambiguously assign chemical shifts of small and macromolecular components in (13)C-poplar samples. Furthermore, using samples of less than 5 mg, we could differentiate between two kinds of genes that were overexpressed in poplar samples, which produced clearly modified plant cell wall components.

Figure 1. Diagrammatic illustration of our method for characterization of ¹³C-labeled plant metabolites and lignocellulose.

(a) Intact sample and DMSO/pyridine extraction of ¹³C-labeled poplar were used for HR-MAS and solution-state NMR, respectively. (b) Signal assignments of metabolites and lignocellulose components could be achieved by 3D HCCH-COSY and HCCH-TOCSY. (c) Based on signal assignments of lignocellulose components, application to two transgenic ¹³C-poplar samples were compared with wild-type poplar.

Figure 2. HR-MAS ¹H-¹³C HSQC spectrum of ¹³C-poplar, measured without sample extraction.

Peaks in the spectrum corresponding to metabolites were assigned by 3D HCCH-COSY experiments and matched by standard metabolites and SpinAssign. Ala, Alanine; Glu, Glutamic acid; Phe, Phenylalanine; Gly, Glycine; Ile, Isoleucine; Lys, Lysine; Leu, Leucine; Met, Methionine; Asn, Asparagine; Gln, Glutamine; Arg, Arginine; Ser, Serine; Thr, Threonine; Val, Valine; GABA, γ-amino butyric acid. The arrows show Glucose signals.

Figure 3. HR-MAS 3D HCCH-COSY spectrum of ¹³C-poplar, measured without sample extraction.

2D ¹H-¹H planes at (a) 29.94, (b) 54.01 (folded spectrum; 94.01), (c) 54.01 (94.01), and (d) 69.94 ppm of ¹³C axis of the 3D ¹H-¹H-¹³C spectrum are shown. Red lines connect the ¹H-¹³C-¹³C-¹H cross peaks of the assigned metabolites. Asn, Asparagine; Arg, Arginine; Lys, Lysine; Gln, Glutamine.

Figure 4. Solution-state ¹H-¹³C HSQC spectrum of ¹³C-poplar extracted in DMSO/pyridine.

Peaks were assigned by 3D HCCH-TOCSY experiments and matched on the basis of the results reported by Kim et al. (2010). Peak numbers in the figure correspond with those listed in Table 1. (a) Anomeric region. (b) Aliphatic region.

Figure 5. Analysis of (1, 4)-β-d-Glcp using solution-state 2D HSQC and 3D HCCH-TOCSY spectra of ¹³C-poplar extracted in DMSO/pyridine.

(a) 2D ¹H-¹³C HSQC spectrum. Crossed marks show C1-6 signals of (1, 4)-β-d-Glcp assigned by 3D HCCH-TOCSY. (b) 2D ¹H-¹H planes at 60.2, 74.6, 79.9, 74.6, 72.7, and 62.7 (folded spectrum; 102.7) ppm of ¹³C, which correspond to the C6, C5, C4, C3, C2, and C1 of (1, 4)-β-d-Glcp, slicing the 3D ¹H-¹H-¹³C spectrum. Red transverse lines connect ¹H-¹³C-¹³C-¹H cross peaks and vertical dashed lines connect corresponding signals between 3D and 2D spectra.

Figure 6. Solution-state ¹H-¹³C HSQC spectra of two transgenic ¹³C-poplar samples of less than 5 mg each dissolved in DMSO/pyridine, following subtraction of wild-type spectrum.

Peak numbers correspond with those in Table 1. (a) VND6. (b) VND7. Blue signals are positive, and red signals are negative.