C-Labeled Lignocellulose Using Multidimensional Solution NMR and

Sep 6, 2013 - Graduate School of Medical Life Science, Yokohama City University, 1-7-29 Suehirocho, Tsurumi-ku, Yokohama 230-0045, Japan. §. Biomass ...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Comprehensive Signal Assignment of 13C‑Labeled Lignocellulose Using Multidimensional Solution NMR and 13C Chemical Shift Comparison with Solid-State NMR Takanori Komatsu†,‡ and Jun Kikuchi*,†,‡,§,# †

RIKEN Center for Sustainable Resource Science, 1-7-22 Suehiro-cho, Tsurumi-ku, Yokohama 235-0045, Japan Graduate School of Medical Life Science, Yokohama City University, 1-7-29 Suehirocho, Tsurumi-ku, Yokohama 230-0045, Japan § Biomass Engineering Research Program, RIKEN Research Cluster for Innovation, 2-1, Hirosawa, Wako 351-0198, Japan # Graduate School of Bioagricultural Sciences and School of Agricultural Sciences, Nagoya University, 1 Furo-cho, Chikusa-ku, Nagoya-shi 464-8601, Japan ‡

S Supporting Information *

ABSTRACT: A multidimensional solution NMR method has been developed using various pulse programs including HCCH-COSY and 13C-HSQC-NOESY for the structural characterization of commercially available 13C labeled lignocellulose from potatoes (Solanum tuberosum L.), chicory (Cichorium intybus), and corn (Zea mays). This new method allowed for 119 of the signals in the 13C-HSQC spectrum of lignocelluloses to be assigned and was successfully used to characterize the structures of lignocellulose samples from three plants in terms of their xylan and xyloglucan structures, which are the major hemicelluloses in angiosperm. Furthermore, this new method provided greater insight into fine structures of lignin by providing a high resolution to the aromatic signals of the β-aryl ether and resinol moieties, as well as the diastereomeric signals of the β-aryl ether. Finally, the 13C chemical shifts assigned in this study were compared with those from solid-state NMR and indicated the presence of heterogeneous dynamics in the polysaccharides where rigid cellulose and mobile hemicelluloses moieties existed together.

T

variety of different kinds of sugar residues, and these materials exist predominantly in the plant cell wall.12,13 These compounds are synthesized in golgi and then transferred to the cell wall. Lignin is an aromatic polymer formed from the radical coupling of coniferyl alcohol, sinapyl alcohol, and pcoumaryl alcohol, with the reaction being regulated by laccase, lignin peroxidase, and manganese peroxidase.14 The aromatic rings of lignin, which are referred to as guaiacyl, sryngyl, or phydroxyphenyl, depending on their degrees of methylation, are connected by several linkage structures, including β-aryl ether (β-O-4), phenylcoumaran (β-5), resinol (β−β), dibenzodioxin (5-5/4-O-β), and spirodienone (β-1) moieties. NMR is one of the most powerful analytical techniques available for the evaluation of biological complexes. For example, NMR has been used to identify metabolites in biological fluids15,16 and to characterize complex macromolecules from biological and environmental sources.17−19 Although solid-state NMR is more desirable than solution NMR in terms of the plant cell wall remaining intact for the analysis,20,21 solution NMR still has advantages in terms of its

he cell walls of plants have been identified as a prospective source of carbon stock because they consist predominantly of lignocelluloses, which are complexes composed of cellulose, hemicelluloses, and lignin. The supramolecular structures of lignocelluloses provide plants with mechanical strength, resistance to environmental stress, and a high level of tolerance toward microbial degradation. Biomass engineering technologies1,2 and genetically engineered plants3−8 have been developed to allow for the effective use of lignocellulosic biomass. The key point of this study is to focus on noncellulosic materials. The development of an analytical method for the characterization of noncellulosic biomass is therefore important for determining how biomass engineering technologies decompose lignocelluloses, as well as investigating the effects of genetic engineering on the structure of lignocellulose. Cellulose is the homopolymer of β-(1,4)-linked D-glucopyranose (Glcp) with a high degree of polymerization. Natural cellulose microfibrils contain two different crystalline forms, including cellulose Iα (triclinic crystal) and Iβ (monoclinic crystal).9−11 In higher plants, cellulose is synthesized at the plasma membrane by rosette terminal complexes, with the Iβ form being produced as the dominant form. Pectin and hemicellulose are heteropolysaccharides, which consist of a © 2013 American Chemical Society

Received: July 17, 2013 Accepted: August 23, 2013 Published: September 6, 2013 8857

dx.doi.org/10.1021/ac402197h | Anal. Chem. 2013, 85, 8857−8865

Analytical Chemistry

Article

were mixed with 660 μL of a 4:1 (v/v) mixture of DMSO-d6 and pyridine-d5 and heated at 323 K for 30 min before being centrifuged at 14 000g for 5 min. The supernatant was transferred into 5 mm NMR tubes. NMR Spectroscopy. Solution NMR spectra were recorded on an AvanceII-700 spectrometer (Bruker, Billerica, MA, USA), equipped with an inverse triple resonance cryogenic probe with a Z-axis gradient for 5 mm sample diameters operating at 700.153 MHz for 1H and 176.061 MHz for 13C. All of the NMR samples were maintained at 313 K. The chemical shifts were referenced to the methyl groups of DMSO on the tetramethylsilane scale (40.03/2.58 ppm (δ13C/δ1H)). Several 2D and 3D NMR experiments were also employed, including 13C-HSQC, ct-HSQC,43 13C-HSQC-TOCSY, 13CHSQC-NOESY, and HCCH-COSY experiments.44 The proton 90 degree pulse value was checked and set for each sample. Relaxation delay values of 1.5 s were used in all of the 2D experiments, and 360 complex f1 (13C) and 1024 complex f2 (1H) points were recorded. The spectral widths of the f1 and f2 dimensions for the 2D experiments were 90 and 16 ppm, respectively. The mixing times for MLEV16 in the 13C HSQCTOCSY spectra and the NOE buildup times in the 13C HSQCNOESY spectra were set to 45 and 300 ms, respectively. Three dimensional HCCH-COSY spectra were acquired using the States-TPPI method. A relaxation delay value of 1.0 s was used, and 2048 complex f1 (1H), 144 complex f2 (13C), and 96 complex f3 (1H) points were recorded. The spectral widths of the f1, f2, and f3 dimensions were 14, 80, and 10 ppm, respectively. The spectra were processed with NMRPipe and analyzed using NMRDraw. The spectra from the 2D experiments were Fourier-transformed with the 60-degree shifted squared cosinebell apodization function. The data were zero-filled to 512 (f1) and 1024 (f2) data points, and an automatic polynomial baseline correction was applied in both dimensions. The 3D HCCH-COSY spectrum was Fourier-transformed with the squared cosine-bell apodization function, and the data were zero-filled to 1024 (f3), 512 (f2), and 256 (f1) points. An automatic polynomial baseline correction was subsequently applied in all dimensions, and the linear prediction was applied to the f1 and f2 dimensions, respectively. The 13C chemical shifts from the solution NMR experiments were compared with the 13C chemical shifts from the solid-state NMR experiments in our previous study19 with respect to each residue. To compare the chemical shifts, the mean unsigned (absolute) error (MUE), the root-mean squared error (RMSE), and the unsigned maximum error (ME) values for each comparison were calculated.

high-resolution and the abundance of available pulse programs. The gel-state solution NMR method was recently developed to analyze whole cell wall samples in a single 13C-HSQC spectrum using DMSO with or without a cosolvent, such as Nmethylimidazole or pyridine.22−25 Solvent systems using ionic liquids have also been developed to dissolve whole cell wall components and obtain high-resolution solution NMR spectra.26−29 Several solution NMR studies have been reported in the literature, including investigations toward understanding the changes in lignocellulose during plant growth,30 fungal degradation processes of lignocellulose,31 tissue-specific characterization of lignocellulose,32 and the relationship between lignocellulose structure and enzymatic saccharification efficiency.33 Unfortunately, however, it can be difficult to make signal assignments in the complicated spectra acquired using this technique because the signals are frequently broad as a consequence of the fast transverse relaxation from the high viscosity of the samples. Furthermore, the assignment of the polysaccharide signals in an NMR spectrum can be challenging because of the small variations in their functional groups.34 The most advanced discipline in NMR over the last couple of decades has been the structural biology of biopolymers, with particular emphasis on proteins.35−37 For investigations toward the structural biology of proteins using solution NMR, techniques involving the incorporation of the stable NMR active 13C and 15N isotopes into overexpressed proteins in combination with multidimensional NMR spectroscopy have resulted in dramatic advances in structural determination. We have also developed analytical methods for signal assignment in metabolomics using 13C labeling in conjunction with threedimensional NMR. Pulse programs using transfer magnetization through the 13C−13C skeletons of molecules such as HCCH-COSY and 13C−13C TOCSY have also be used in combination with the 13C labeling technique to provide significant advances in the signal assignment of metabolites in plants, animals, and bacteria.38−42 The same techniques could have also been used in the structural characterization of lignocellulosic biomass. In the current study, we have conducted a comprehensive assignment of the signals of lignocellulose using multidimensional NMR with commercially available 13 C labeled lignocellulose from potatoes (Solanum tuberosum L.), chicory (Cichorium intybus), and corn (Zea mays). We have also developed methods for the structural characterization of lignocellulosic biomass using different pulse programs. Furthermore, we have provided some discussion of the fine methods used to resolve the signals from lignin. Finally, we have compared the chemical shifts assigned in this study with those from the solid-state NMR assigned in our previous study.19

■

RESULTS AND DISCUSSION Comprehensive Signal Assignment of Cellulose. Four distinguishable β-D-Glcp residues in the cello-oligosaccharides derived from cellulose were assigned and identified as the inner (1,4)-β-D-Glcp, the nonreducing end of the (1,4)-β-D-Glcp, reducing end of the (1,4)-α-D-Glcp, and reducing end of the (1,4)-β-D-Glcp. The assignments of the H6 and C6 signals of the reducing end were ambiguous because they were overlapping with those of the inner and nonreducing ends. Hemicellulose. Four distinguishable D-Xylp residues and three distinguishable L-Araf residues were assigned in the xylans, which were the major hemicelluloses in the angiosperm secondary cell wall. These D-Xylp residues were identified as the inner (1,4)-β-D-Xylp, the nonreducing end of the (1,4)-β-D-

■

EXPERIMENTAL SECTION Materials and Methods. U-13C labeled lignocellulose samples from potatoes, chicory, and corn (>97 atm% 13C) were purchased from IsoLife (Wageningen, Netherland). Dimethyl sulfoxide-d6 (DMSO-d6, 99.9% D) and pyridine-d5 (99.5% D, +0.05% v/v TMS) were purchased from Cambridge Isotope Laboratories (Andover, MA, USA) Sample Preparation. Samples of the dried U-13C labeled lignocellulose (60 mg) were pulverized for 6 h with a Pulverisette 5 (Fritsch GmbH, Idar-Oberstein, Germany), and the resulting pulverized powders were extracted three times with distilled water at 323 K for 10 min. Ball-milled samples 8858

dx.doi.org/10.1021/ac402197h | Anal. Chem. 2013, 85, 8857−8865

Analytical Chemistry

Article

Figure 1. 13C-HSQC spectra of 13C lignocellulose from potatoes, chicory, and corn. (a−c) Represent the regions of the sugar and lignin linkages. (d−e) Represent the aromatic regions in the lignin. There were no aromatic signals in the spectrum of the lignocellulose from corn. (f) Partial structures of lignocellulose and the corresponding abbreviations in the spectra. A list of chemical shift and references of assignments are shown in Tables 1 and S1, Supporting Information, respectively.

which were the major hemicelluloses in the primary cell wall of a dicot plant. D-Xylp was identified as (1,6)-α-D-Xylp in the xyloglucans, and L-Araf was identified as (1,2)-α-L-Araf in the arabinoxyloglucan. Lignin. In the aromatic region of the spectra, there were several signals derived from the lignin. Aromatic signals corresponding to G and S units were detected, although no signals corresponding to the H unit, which exists only in compound middle lamella,48 were detected. Several linkage structures (i.e., β-O-4, β−β, and β-5) and cinnamyl alcohol end

Xylp, the reducing end of the (1,4)-α-D-Xylp, and the reducing end of the (1,4)-β-D-Xylp in xylans. Several different branched structures derived from the α-L-Araf in the glucuronoarabinoxylan (GAX) were reported. In GAX, β-D-Xylp is sometimes substituted at the O3 position with a-L-Araf. Furthermore, a part of α-L-Araf can be substituted at the O2 position with β-DXylp.45 Infrequently, β-D-Xylp is substituted at the O2 position with α-L-Araf.46,47 These L-Arafs were identified as (1,3)-α-LAraf, 2-O-Xylp-substituted (1,3)-α-L-Araf, and (1,2)-α-L-Araf in GAX. The D-Xylp and L-Araf were assigned in the xyloglucans, 8859

dx.doi.org/10.1021/ac402197h | Anal. Chem. 2013, 85, 8857−8865

Analytical Chemistry

Article

Table 1. List of the Chemical Shifts Assigned in this Studya residues cellooligosaccharide

(1,4)-β-D-Glcp (1,4)-β-D-Glcp (non reducing end) (1,4)-β-D-Glcp (reducing end) (1,4)-α-D-Glcp (reducing end)

xylans

(1,4)-β-D-Xylp (1,4)-β-D-Xylp (non reducing end) (1,4)-β-D-Xylp (reducing end) (1,4)-α-D-Xylp (reducing end) 3-O-Araf-substituted (1,4)-β-DXylp (1,3)-α-L-Araf 2-O-Xylp-substituted (1,3)-α-LAraf (1,2)-α-L-Araf

xyloglucans

(1,6)-α-L-Xylp (1,3)-α-L-Araf

starch

(1,4)-α-D-Glcp (1,4)-α-D-Glcp’

lignin

β-O-4(S)erythro β-O-4(S)threo β-O-4(G)erythro β-O-4(G)threo β-5 β−β cinnamyl alcohol

a

shift (ppm) δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H δ13C δ1H

1

2

3

4

103.2 4.55 103.6 4.48 97.2 4.56 92.5 5.15 102.2 4.47 102.4 4.42 98.0 4.47 92.7 5.09 102.2 4.47 107.6 5.60 105.9 5.71 108.1 5.45 99.6 4.84 109.8 5.17 100.6 5.32 101.1 5.25

73.4 3.31 73.7 3.25 74.8 3.23 72.0 3.47 73.0 3.30 73.7 3.25 75.0 3.21 72.6 3.39 73.4 3.50 80.6 4.07 87.6 4.26 81.5 4.13 72.7 3.37 82.0 4.15 72.3 3.53 70.4 3.32 104.6 6.87 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 103.6 6.79 N.A. N.A.

75.2 3.58 77.2 3.41 75.2 3.52 71.9 3.83 74.4 3.50 76.7 3.36 75.1 3.45 71.5 3.77 73.7 3.76 78.3 3.88 76.2 4.07 77.8 3.94 73.4 3.82 77.6 3.88 73.6 3.92 73.8 3.74

80.5 3.57 70.5 3.31 81.1 3.56 81.1 3.56 75.8 3.72 69.9 3.49 76.2 3.67 76.2 3.67 75.7 3.82 86.7 4.23 85.8 4.26 85.6 4.18 70.5 3.48 84.1 4.05 79.3 3.58 73.8 3.66

75.2 3.58 76.9 3.42 75.2 3.44 70.2 3.94 63.6 3.36, 66.2 3.25, 63.4 3.29, 59.2 3.71, 63.3 4.12 62.3 3.67 61.9 3.72 61.8 3.67 62.3 3.59 61.4 3.63 72.0 3.83 72.9 3.48

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A. N.A. N.A. N.A. N.A. N.A. N.A.

N.A.

N.A.

N.A.

N.A.

N.A. N.A. N.A. N.A. N.A. N.A. N.A.

5

N.A. N.A. N.A.

6

7 (α)

8 (β)

72.6 5.09 72.0 5.12 72.1 4.98 71.4 4.97 87.3 5.65 85.4 4.80 128.6 6.60

86.3 4.32 87.3 4.22 84.1 4.50 84.9 4.45 53.6 3.61 54.0 3.16 128.8 6.35

9 (γ)

60.6 3.80, 3.99 61.5 3.90, 3.66 N.A. N.A. N.A. N.A. 4.06 3.90 3.99 3.80

60.8 3.78, 3.87 61.3 3.85, 3.69 104.6 6.87 N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 103.6 6.79 N.A. N.A.

60.0 3.60, 60.0 3.60, 60.4 3.44, 60.4 3.44, 63.4 3.88 71.5 4.29 61.9 4.24

3.93 3.93 3.81 3.81

N.A. means not assigned in this study.

of the heteronuclear polarization transfer was improved using uniform 13C labeling. In addition, 13C labeling provided an application for HCCH three-dimensional experiments including HCCH-COSY, which is designed to transfer magnetization via 13C−13C spin−spin coupling and is effective for the chemical shift assignments of the inner-residues of the sugar and lignin because of the transfer of magnetization along the13C-skeltons. We previously reported using HCCH-COSY in combination with 13C stable isotope labeling as an effective tool for resonance assignment in plants, bacteria, and animal metabolites.42,40 The development of a method for the

groups were assigned in the aliphatic region of the spectra. Four different types of β-O-4 linkages (i.e., β-O-4 syringyl erythro, βO-4 syringyl threo, β-O-4 guaiacyl erythro, and β-O-4 guaiacyl threo) were observed in the 13C HSQC spectra (Figure 1 and Table 1). Analytical Methodologies for Signal Assignments of 13 C Labeled Lignocellulose Using Multidimensional NMR. Using multidimensional NMR to analyze 13C labeled lignocellulose allows for dramatic improvements in the sensitivity compared with naturally abundant lignocellulose. For example, in the 1H−13C inverse experiments, the efficiency 8860

dx.doi.org/10.1021/ac402197h | Anal. Chem. 2013, 85, 8857−8865

Analytical Chemistry

Article

between H1 of 2-O-Xylp substituted (1,3)-α-L-Araf and H2, 3, 4, and 5 of (1,4)-β-D-Xylp were obtained at 102.2/5.71 (δC1‑xyl/ δH1‑ara), 73.4/5.71 (δC2‑xyl/δH1‑ara), 73.7/5.71 (δC3‑xyl/δH1‑ara), 75.7/5.71 (δC4‑xyl/δH1‑ara), and 63.3/5.71 ppm (δC5‑xyl/δH1‑ara). These results demonstrate that the technique of ct-HSQC can be used to provide a greater insight into the fine structures in lignin. For example, ct-HSQC effectively resolved the signals in an aromatic region (Figure 4a). Furthermore, the NOE correlations in the 13C-HSQC-NOESY spectrum at 72.6, 86.3, 60.0/6.87, (δCα,β,γ‑ASe/δH‑syringyl2/6) 85.4, and 71.5/6.79 (δCα,γ_C/δH‑syringyl2/6) revealed that the signals at 104.6/6.87 and 103.6/6.79 ppm were the 2 and 6 nuclei of β-O-4 syringyl erythro and β−β, respectively (Figure 4b). In a similar way, ctHSQC allowed for the resolution of the diastereomeric signals at the α position of the β-O-4 linkage (Figure 4c). In the 13CHSQC-TOCSY spectrum, 1H−1H correlations at 86.3/5.09 (δCβ‑A‑Se /δ Hα‑A‑S ), 87.3/5.12 (δ Cβ‑A‑St /δHα‑A‑S ), 72.1/4.50 (δCα‑A‑G/δHβ‑A‑Ge), and 71.4/4.45 (δCα‑A‑G/δHβ‑A‑Gt) ppm indicated that the signals at 72.6/5.09, 72.0/5.12, 72.1/4.98, and 71.4/7.97 ppm were the α nuclei of the β-O-4 syringyl erythro, β-O-4 syringyl erythro, β-O-4 syringyl erythro, and β-O-4 syringyl erythro groups, respectively (Figure 4d). ct-HSQC is a version of HSQC that has been optimized for uniformly 13C labeled molecules where the typical variable 13C evolution period has been replaced by a constant-time evolution period and the JCC evolutions have been refocused leading to the removal of any line broadening in the 13C dimension.43 ctHSQC is a powerful method for the analysis of crowded polysaccharide regions in 2D NMR spectra. However, this technique requires optimization to a constant time length (=2T) in accordance with the 1JCC value of the aliphatic polysaccharide, anomeric polysaccharide, or aromatic lignin carbons that modulate the intensities of the signals according to Πcos(2TπJCC) (Figure S3, Supporting Information). Typical values of 1JCC were estimated in high-resolution 13C-HSQC spectra (Figure S4, Supporting Information), indicating optimal values of constant time length in ct-HSQC were 22.2 ms for aliphatic signals (1JCC = 45 Hz), 20 ms for anomeric signals (1JCC = 50 Hz), and 13.9 ms for aromatic signals (1JCC = 72 Hz). Characterization of Cell Wall Structure by Plant Family. The methods described in this study could be used to characterize the differences in the plant cell wall structures of different plant families (Table S2 and Figure S1, Supporting Information). In the current study, 13C labeled lignocellulose samples from corn (Zea mays), potatoes (Solanum tuberosum L.), and chicory (Cichorium intybus) were subjected to NMR analysis. The xylan produced by gramineous plants is referred to as GAX, where Araf is frequently substituted with a 1,4-β-DXylp chain. In contrast, the xylan produced by dicot plants is referred to as glucuronoxylan (GX), where the Araf is never substituted.47 Signals consistent with the β-D-Xylp and α-L-Araf residues were identified in the 13C-HSQC spectrum of the 13C labeled lignocellulose from corn which were glycosyl residues in glucuronoarabinoxylan. In contrast, the 13C-HSQC spectrum of the 13C labeled lignocellulose from potato and chicory contained signals that were consistent with β-D-Xylp, with no α-L-Araf signals being detected. These results indicated that GAX was present in corn, whereas GX was present in the potatoes and chicory. These results were therefore in agreement with the current botanical knowledge.12 XAT1 and XAT2, which are members of the glycosyl transferase family 61,

resonance assignment of a bacterial cell-wall using HCCHCOSY and HCCH-TOCSY on high-resolution magic-angle spinning (HR-MAS) NMR has also been reported.49,50 As shown in Figure 2, HCCH-COSY is also effective for the

Figure 2. HCCH-COSY spectra for the interior signal assignments of lignocellulose. The signal assignments of the (1,4)-β-D-Glcp from the cello-oligosaccharide, (1,4)-β-D-Xylp from the xylans, and (1,3)-α-LAraf from the glucuronoarabinoxylan (GAX) are shown as representative examples.

resonance assignment of lignocellulose because it separates crowded signals into the additional dimension. Advances in multidimensional NMR have therefore provided greater insights into hemicellulosic signals, although previous studies of whole cell wall analyses using solution NMR could not have focused on the aliphatic signals of polysaccharides. NOE-based analysis is an effective way to analyze interresidue correlations in polysaccharides because of the absence of a large J-coupling in glycosidic bonds. The results of the inter-residue correlation analysis performed using 13C-HSQCNOESY are shown in Figure 3. In the 13C-HSQC-NOESY spectrum, NOE cross peaks representing the inter-residues correlations between H1 of (1,3)-α-L-Araf and H1, 2, 3, 4, and 5 of (1,4)-β-D-Xylp were obtained at 102.2/5.60 (δC1‑xyl/δH1‑ara), 73.4/5.60 (δC2‑xyl/δH1‑ara), 73.7/5.60 (δC3‑xyl/δH1‑ara), 75.7/5.60 (δC4‑xyl/δH1‑ara), and 63.3/5.60 (δC5‑xyl/δH1‑ara). Furthermore, the NOE cross peaks representing the inter-residue correlations

Figure 3. 13C-HSQC-NOESY correlations for the inter residue assignments of lignocellulose. (a) 13C-HSQC-NOESY spectrum of the lignocellulose from corn. (b) Partial structures of GAX which were detected in the 13C-HSQC-NOESY analysis. 8861

dx.doi.org/10.1021/ac402197h | Anal. Chem. 2013, 85, 8857−8865

Analytical Chemistry

Article

Figure 4. Fine structure analysis of lignin using ct-HSQC. (a) Spectra of 13C-HSQC (upper) and ct-HSQC (lower) in the aromatic regions. ctHSQC resolved aromatic signals. (b) 13C-HSQC-NOESY spectrum focused on correlations between the aromatic and aliphatic protons of lignins. (c) Spectra of 13C-HSQC (upper) and ct-HSQC (lower) in the α position of the β-O-4. Ct-HSQC resolved diastereomeric signals. (d) 13C-HSQCTOCSY spectrum focused on the correlations between the α and β protons of β-O-4.

encode xylan arabinosyltransferases in rice and wheat51 and play a key role in GAX biosynthesis. Xyloglucans are the major hemicelluloses in the primary cell walls of dicot plants, and the methods disclosed in this study provided a good means of characterization that provided results in agreement with the current botanical knowledge. A number of different structures have been reported for xyloglucans.52 These materials generally consist of a 1,4-linked β-D-Glcp chain with approximately 75% of the β-D-Glcp residues substituted at the O6 position with α-D-Xylp, and the α-D-Xylp residues themselves sometimes substituted at the O2 position with β-DGalp and α-L-Fucp. The xyloglucan produced by solanaceous plants is referred to as arabinoxyloglucan (AXG). This material possesses an unusual structure, in that only 40% of the β-D-Glcp residues are substituted at the O6 position with α-D-Xylp and about half of α-D-Xylp residues are substituted at the O2 position with α-L-Araf.53,54 In the 13C-HSQC spectrum of the 13 C lignocellulose from potatoes, the signals representing the αD-Xylp and α-L-Araf in the glycosyl residues of AXG were identified. In contrast, the 13C-HSQC spectrum of 13C labeled lignocellulose from chicory contained signals that were consistent with α- D-Xylp and no signals for α- L-Araf. Furthermore, the 13C-HSQC spectrum of the 13C lignocellulose from corn did not contain any α-D-Xylp or α-L-Araf signals. These results therefore indicated that the potatoes contained AXG, the chicory contained a common xyloglucan consisting of β-D-Galp and α-L-Fucp, and that the corn contained no xyloglucans. These results were therefore in agreement with the current botanical knowledge.12 The side chain structures in xyloglucans depend on their glycosyl transferases.55 XXT1 and XXT2 have been reported as xylosyltransferase.56 In contrast, MUR3 was reported as a galactosyltransferase,57 whereas MUR2/FUT1 was reported as a fucosyltransferase.58 The role of arabinosyltransferases, however, remains unclear.55 It has been suggested that MUR3 homologues in solanaceous plants act as arabinosyltransferases rather than galactosyltransferases.57 Comparison of the Chemical Shifts from the Solution and Solid-State NMR. A comparison of chemical shifts from the solution NMR assigned in the current study with those assigned in our previous study using solid-state NMR16,59 has been provided in Figure 5. Chemical shift differences between the solutions and solid-state NMR are shown in Table 2 as

Figure 5. 13C chemical shift comparison of the solution and solid-state NMR. The signals corresponding to the C4 and C6 positions of cellulose were significantly different.

MUE, RMSE, and ME. The RMSE values in cellulose were 4.3, 4.2 (Iβ crystal), and 2.1 ppm (amorphous) and were higher than those of the hemicelluloses. A similar trend was also observed in the MUE and ME values. In addition, the chemical shift differences of cellulose were large at the C4, C5, and C6 positions. The differences in the 13C chemical shifts between the solution and solid-state NMR analyses reflect differences in the conformations of the materials in the different states. In the solid-state NMR of cellulose, the chemical shifts are given from the crystalline conformation, whereas the chemical shifts in solution NMR are given as the average of a number of different conformations. Chemical shifts effectively supply information that can be used to evaluate the conformations of different macromolecules such as lignocellulosic biomass and large proteins.60−62 The relationships between the chemical shifts and their conformations were investigated by quantum chemistry calculations using the D-glucose, D-cellobiose, and cellobiose units of crystalline cellulose.63 The results of that study revealed that the conformation of the hydroxymethyl group strongly affected the chemical shifts of the C4, C5, and C6 positions. In the trans-gauche (tg) conformation, the chemical shifts of the C4 and C6 positions were shifted downfield compared to gauche−trans (gt) and gauche−gauche (gg) conformations. In the tg conformation, however, the chemical shifts of the C5 position were shifted upfield 8862

dx.doi.org/10.1021/ac402197h | Anal. Chem. 2013, 85, 8857−8865

Analytical Chemistry

Article

Table 2. Comparison of Isotropic 13C Chemical Shifts from the Solution and Solid-State NMR Experiments cellulose Iβ

C1 C2 C3 C4 C5 C6 MUE RMSE ME a

cellulose Iβ′

amorphous cellulose

β-D-Xylp (xylans)

α-L-Araf (GAX)

α-D-Xylp (xyloglucans)

α-L-Araf (AXG)

δ13C solution [ppm]

δ13C solida [ppm]

δ13C solution [ppm]

δ13C solida [ppm]

δ13C solution [ppm]

δ13C solida [ppm]

δ13C solution [ppm]

δ13C solida [ppm]

δ13C solution [ppm]

δ13C solida [ppm]

δ13C solution [ppm]

δ13C solida [ppm]

δ13C solution [ppm]

δ13C solida [ppm]

103.2 73.4 75.2 80.5 75.2 60.6 3.4 4.3 7.7

105.8 72.8 75.6 88.2 71.3 66.0

103.2 73.4 75.2 80.5 75.2 60.6 3.2 4.2 8.6

104.1 71.5 74.8 89.2 72.7 65.3

103.2 73.4 75.2 80.5 75.2 60.6 1.6 2.1 4.0

105.0 72.8 75.8 84.6 75.5 62.8

102.2 73.0 74.4 75.8 63.6

102.5 73.8 74.7 77.2 63.8

107.6 80.6 78.3 86.7 62.3

108.4 81.7 78.2 85.8 62.3

99.6 72.7 73.4 70.5 62.3

99.7 72.4 73.8 70.4 62.2

109.8 82.0 77.6 84.1 61.4

110.0 82.0 77.4 84.8 62.2

0.6 0.8 1.4

0.6 0.7 1.1

0.2 0.2 0.4

0.4 0.5 0.9

Isotropic 13C chemical shifts in solid-state NMR were used in our previous study.19

compared to the gt and gg conformations. These calculations were in agreement with our experimental results because the Iβ cellulose was in the tg conformation. The conformation around the glycosyl-linkage can also affect the chemical shifts of the C1 and C4 positions. Furthermore, having the OH hydrogen atom in a gauche conformation led to a downfield shift in the chemical shift of the carbon at the γ position. In amorphous cellulose, the 13C chemical shifts of the C5 position in solid-state NMR became very similar to those of the solution NMR. In contrast, the 13C chemical shifts of the C4 and C6 positions in the solid-state NMR were higher than those of the solution NMR, with a similar trend being observed in the crystalline cellulose. The hydroxymethyl groups of amorphous cellulose possess tg and gt conformations together, with both conformations contributing to the 13C chemical shifts. The results of the theoretical study indicated that the 13C chemical shifts of the C5 position in the gt conformation were lower than those in the tg conformation. In contrast, the shift of the C4 and C6 positions in the gt conformation were lower than those in the tg conformation. Although these differences made the chemical shifts differences between the solid-state and solution NMR analyses small, chemical shift differences still existed. These results indicated that the conformational distribution in the amorphous cellulose was different from the solution state. Furthermore, Mori et al.64 reported the local structure of amorphous cellulose and its distribution using experimental chemical shifts, molecular dynamics, and quantum chemistry calculations. The results of that particular study indicated that amorphous cellulose possesses a quasi-stable structure. The differences in the chemical shifts between the solution and solid-state NMR analyses could also be derived from the decomposition of cellulose into lower molecules during the pulverization procedure used in the preparation of samples for solution NMR analysis. Signals from reducing and nonreducing end sugar residues of 1,4-β-Glcp in 13C-HSQC spectra strongly indicate that chains of 1,4-β-Glcp became shorter than natural cellulose. The chemical shift differences in the solution and solid-state NMR of cellobiose, which is a dimer of 1,4-β-Dglucose, were reported by Tang and Belton.65 In this report, the 13 C chemical shift differences between the solution and solidstate NMR of the C4 position were significantly larger than those of the other carbon nuclei, although the molecular weight observed in the solution and solid states was obviously the same. These results strongly suggested that the chemical shift

differences observed between the solution and solid-state NMR analyses of cellulose were derived from conformational differences. The 13C chemical shifts of hemicelluloses in solution NMR (i.e., β-D-Xylp and α-L-Araf in xylans, and α-D-Xylp and α-L-Araf in xyloglucans) are similar to those in the solid-state NMR, suggesting that the hemicelluloses have no significant structure in bulk. These results lead to two different considerations, including (1) these materials possess random conformational structure, and (2) additionally, rotations of the molecules are fast, like in solution state. Our previous study indicated that these similarities in the chemical shifts between the solution and solid-state NMR occur as a consequence of the latter consideration because the hemicelluloses decayed to a much lesser extent than cellulose in the dipolar dephasing experiments under hydrated conditions.19 The supramolecular structure of lignocellulose must therefore be inhomogeneous, with the rigid and static structures of cellulose and the mobile and dynamic nonstructural hemicelluloses and lignin existing together.

■

CONCLUSION There are two key points to this Article. The first of these points is the application of multidimensional NMR analyses including three-dimensional NMR analysis to 13C labeled lignocellulose dissolved in organic solvent. These multidimensional NMR analyses have provided us with an unambiguous and comprehensive series of signal assignments for lignocellulosic biomass, with a significant number of nuclei from the lignocellulose being identified. Furthermore, we believe that our studies represent a significant advance in the NMR analysis of the aliphatic region of polysaccharides (Figure S5, Supporting Information), which has traditionally been a particularly challenging area for NMR analysis because of the structural similarities in this region. The methods developed in this Article and the resulting chemical shift information will become powerful tools for plant cell wall studies. The second key point concerns the comparison of individual 13 C isotropic chemical shifts from solution and solid-state NMR analyses. Differences in the chemical shifts in solution and solid-state NMR can come from conformational differences between the two states. Conversely, this result indicated that hemicelluloses possess no significant structure. The hemicelluloses, however, may possess some partial particular structure, for example, cellulose binding structure, although 8863

dx.doi.org/10.1021/ac402197h | Anal. Chem. 2013, 85, 8857−8865

Analytical Chemistry

Article

(15) Barding, G. A.; Fukao, T.; Beni, S.; Bailey-Serres, J.; Larive, C. K. J. Proteome Res. 2012, 11, 320. (16) Lewis, I. A.; Schommer, S. C.; Hodis, B.; Robb, K. A.; Tonelli, M.; Westler, W. M.; Suissman, M. R.; Markley, J. L. Anal. Chem. 2007, 79, 9385. (17) Langeslay, D. J.; Beecher, C. N.; Naggi, A.; Guerrini, M.; Torri, G.; Larive, C. K. Anal. Chem. 2013, 85, 1247. (18) Simpson, A. J.; Simpson, M. J.; Soong, R. Environ. Sci. Technol. 2012, 46, 11488. (19) Komatsu, T.; Kikuchi, J. J. Phys. Chem. Lett. 2013, 4, 2279. (20) Dick-Perez, M.; Wang, T.; Salazar, A.; Zabotina, O. A.; Hong, M. Magn. Reson. Chem. 2012, 50, 539. (21) Dick-Perez, M.; Zhang, Y. A.; Hayes, J.; Salazar, A.; Zabotina, O. A.; Hong, M. Biochemistry 2011, 50, 989. (22) Mansfield, S. D.; Kim, H.; Lu, F. C.; Ralph, J. Nat. Protoc. 2012, 7, 1579. (23) Kim, H.; Ralph, J. Org. Biomol. Chem. 2010, 8, 576. (24) Yelle, D. J.; Ralph, J.; Frihart, C. R. Magn. Reson. Chem. 2008, 46, 508. (25) Kim, H.; Ralph, J.; Akiyama, T. Bioenergy Res. 2008, 1, 56. (26) Qu, C.; Kishimoto, T.; Hamada, M.; Nakajima, N. Holzforschung 2013, 67, 25. (27) Cheng, K.; Sorek, H.; Zimmermann, H.; Wemmer, D. E.; Pauly, M. Anal. Chem. 2013, 85, 3213. (28) Samuel, R.; Foston, M.; Jaing, N.; Cao, S. L.; Allison, L.; Studer, M.; Wyman, C.; Ragauskas, A. J. Fuel 2011, 90, 2836. (29) Qu, C.; Kishimoto, T.; Kishino, M.; Hamada, M.; Nakajima, N. J. Agric. Food Chem. 2011, 59, 5382. (30) Rencoret, J.; Gutierrez, A.; Nieto, L.; Jimenez-Barbero, J.; Faulds, C. B.; Kim, H.; Ralph, J.; Martinez, A. T.; del Rio, J. C. Plant Physiol. 2011, 155, 667. (31) Martinez, A. T.; Rencoret, J.; Nieto, L.; Jimenez-Barbero, J.; Gutierrez, A.; del Rio, J. C. Environ. Microbiol. 2011, 13, 96. (32) Watanabe, T.; Shino, A.; Akashi, K.; Kikuchi, J. Plant Biotechnol.Nar 2012, 29, 163. (33) Yamamura, M.; Noda, S.; Hattori, T.; Shino, A.; Kikuchi, J.; Takabe, K.; Tagane, S.; Gau, M.; Uwatoko, N.; Mii, M.; Suzuki, S.; Shibata, D.; Umezawa, T. Plant Biotechnol.-Nar 2013, 30, 25. (34) Kato, K.; Sasakawa, H.; Kamiya, Y.; Utsumi, M.; Nakano, M.; Takahashi, N.; Yamaguchi, Y. BBA-Gen. Subj. 2008, 1780, 619. (35) Wuthrich, K. Science 1989, 243, 45. (36) Williamson, M. P.; Havel, T. F.; Wuthrich, K. J. Mol. Biol. 1985, 182, 295. (37) Sakata, E.; Yamaguchi, Y.; Kurimoto, E.; Kikuchi, J.; Yokoyama, S.; Yamada, S.; Kawahara, H.; Yokosawa, H.; Hattori, N.; Mizuno, Y.; Tanaka, K.; Kato, K. EMBO Rep. 2003, 4, 301. (38) Bingol, K.; Zhang, F. L.; Bruschweiler-Li, L.; Bruschweiler, R. J. Am. Chem. Soc. 2012, 134, 9006. (39) Bingol, K.; Zhang, F. L.; Bruschweiler-Li, L.; Bruschweiler, R. Anal. Chem. 2012, 84, 9395. (40) Chikayama, E.; Suto, M.; Nishihara, T.; Shinozaki, K.; Hirayama, T.; Kikuchi, J. PloS One 2008, 3. (41) Kikuchi, J.; Hirayama, T. Methods Mol. Biol. 2007, 358, 273. (42) Sekiyama, Y.; Chikayama, E.; Kikuchi, J. Anal. Chem. 2011, 83, 719. (43) Vuister, G. W.; Bax, A. J. Magn. Reson. 1992, 98, 428. (44) Kay, L. E.; Ikura, M.; Bax, A. J. Am. Chem. Soc. 1990, 112, 888. (45) Hoije, A.; Sandstrom, C.; Roubroeks, J. P.; Andersson, R.; Gohil, S.; Gatenholm, P. Carbohydr. Res. 2006, 341, 2959. (46) Pitkanen, L.; Tuomainen, P.; Virkki, L.; Aseyev, V.; Tenkanen, M. J. Agric. Food Chem. 2008, 56, 5069. (47) Ebringerova, A.; Heinze, T. Macromol. Rapid Commun. 2000, 21, 542. (48) Terashima, N.; Fukushima, K. Wood Sci. Technol. 1988, 22, 259. (49) Li, W.; Lee, R. E. B.; Lee, R. E.; Li, J. H. Anal. Chem. 2005, 77, 5785. (50) Lee, R. E. B.; Li, W.; Chatterjee, D.; Lee, R. E. Glycobiology 2005, 15, 139.

no methodology currently exists that would allow them to be detected. A further improvement in the current methodology would be necessary to investigate the site-specific structures and develop our understanding of the structures of hemicellulose and lignocellulose.

■

ASSOCIATED CONTENT

* Supporting Information S

Assignment notes, partial structure of GAX, constant time series of 1H projections of ct-HSQC spectra, and comparisons of the signal assignments between this study and our previous report. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Author Contributions

The manuscript was written with contributions from of all the authors. All of the authors have given their approval to the final version of the manuscript. Notes

The authors declare no competing financial interests.

■

ACKNOWLEDGMENTS This research was supported in part by Grants-in-Aid for Scientific Research (Grant No. 25513012, to J.K.) and the Advanced Low Carbon Technology Research and Developmental Program (Grant No. 200210023, ALCA to J.K.) from the Ministry of Education, Culture and Sports. The authors wish to thank Keiko Okushita (Yokohama City University) for her help with the initial stage of this study.

■

REFERENCES

(1) Hasunuma, T.; Rondo, A. Biotechnol. Adv. 2012, 30, 1207. (2) Ragauskas, A. J.; Williams, C. K.; Davison, B. H.; Britovsek, G.; Cairney, J.; Eckert, C. A.; Frederick, W. J.; Hallett, J. P.; Leak, D. J.; Liotta, C. L.; Mielenz, J. R.; Murphy, R.; Templer, R.; Tschaplinski, T. Science 2006, 311, 484. (3) Chen, F.; Dixon, R. A. Nat. Biotechnol. 2007, 25, 759. (4) Demura, T.; Ye, Z. H. Curr. Opin. Plant Biol. 2010, 13, 299. (5) Kubo, M.; Udagawa, M.; Nishikubo, N.; Horiguchi, G.; Yamaguchi, M.; Ito, J.; Mimura, T.; Fukuda, H.; Demura, T. Gene Dev. 2005, 19, 1855. (6) Baba, K.; Park, Y. W.; Kaku, T.; Kaida, R.; Takeuchi, M.; Yoshida, M.; Hosoo, Y.; Ojio, Y.; Okuyama, T.; Taniguchi, T.; Ohmiya, Y.; Kondo, T.; Shani, Z.; Shoseyov, O.; Awano, T.; Serada, S.; Norioka, N.; Norioka, S.; Hayashi, T. Mol. Plant 2009, 2, 893. (7) Hu, W. J.; Harding, S. A.; Lung, J.; Popko, J. L.; Ralph, J.; Stokke, D. D.; Tsai, C. J.; Chiang, V. L. Nat. Biotechnol. 1999, 17, 808. (8) Kaida, R.; Kaku, T.; Baba, K.; Oyadomari, M.; Watanabe, T.; Nishida, K.; Kanaya, T.; Shani, Z.; Shoseyov, O.; Hayashi, T. Mol. Plant 2009, 2, 904. (9) Atalla, R. H.; Vanderhart, D. L. Science 1984, 223, 283. (10) Nishiyama, Y.; Langan, P.; Chanzy, H. J. Am. Chem. Soc. 2002, 124, 9074. (11) Nishiyama, Y.; Sugiyama, J.; Chanzy, H.; Langan, P. J. Am. Chem. Soc. 2003, 125, 14300. (12) Scheller, H. V.; Ulvskov, P. Annu. Rev. Plant Biol. 2010, 61, 263. (13) Somerville, C.; Bauer, S.; Brininstool, G.; Facette, M.; Hamann, T.; Milne, J.; Osborne, E.; Paredez, A.; Persson, S.; Raab, T.; Vorwerk, S.; Youngs, H. Science 2004, 306, 2206. (14) Vanholme, R.; Demedts, B.; Morreel, K.; Ralph, J.; Boerjan, W. Plant Physiol. 2010, 153, 895. 8864

dx.doi.org/10.1021/ac402197h | Anal. Chem. 2013, 85, 8857−8865

Analytical Chemistry

Article

(51) Anders, N.; Wilkinson, M. D.; Lovegrove, A.; Freeman, J.; Tryfona, T.; Pellny, T. K.; Weimar, T.; Mortimer, J. C.; Stott, K.; Baker, J. M.; Defoin-Platel, M.; Shewry, P. R.; Dupree, P.; Mitchell, R. A. C. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 989. (52) Vincken, J. P.; York, W. S.; Beldman, G.; Voragen, A. G. J. Plant Physiol. 1997, 114, 9. (53) Eda, S.; Kato, K. Agric. Biol. Chem. Tokyo 1978, 42, 351. (54) York, W. S.; Kolli, V. S. K.; Orlando, R.; Albersheim, P.; Darvill, A. G. Carbohydr. Res. 1996, 285, 99. (55) Lerouxel, O.; Cavalier, D. M.; Liepman, A. H.; Keegstra, K. Curr. Opin. Plant Biol. 2006, 9, 621. (56) Faik, A.; Price, N. J.; Raikhel, N. V.; Keegstra, K. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 7797. (57) Madson, M.; Dunand, C.; Li, X. M.; Verma, R.; Vanzin, G. F.; Calplan, J.; Shoue, D. A.; Carpita, N. C.; Reiter, W. D. Plant Cell 2003, 15, 1662. (58) Perrin, R. M.; DeRocher, A. E.; Bar-Peled, M.; Zeng, W. Q.; Norambuena, L.; Orellana, A.; Raikhel, N. V.; Keegstra, K. Science 1999, 284, 1976. (59) Okushita, K.; Komatsu, T.; Chikayama, E.; Kikuchi, J. Polym. J. 2012, 44, 895. (60) Kikuchi, J.; Asakura, T. J. Biochem. Biophys. Methods 1999, 38, 203. (61) Kikuchi, J.; Asakura, T.; Hasuda, K.; Ito, T.; Ohwaku, K.; Araki, H.; Williamson, M. P. J. Biochem. Biophys. Methods 2000, 42, 35. (62) Williamson, M. P.; Kikuchi, J.; Asakura, T. J. Mol. Biol. 1995, 247, 541. (63) Suzuki, S.; Horii, F.; Kurosu, H. J. Mol. Struct. 2009, 921, 219. (64) Mori, T.; Chikayama, E.; Tsuboi, Y.; Ishida, N.; Shisa, N.; Noritake, Y.; Moriya, S.; Kikuchi, J. Carbohydr. Polym. 2012, 90, 1197. (65) Tang, H. R.; Belton, P. S. Solid State Nucl. Magn. Reson. 2002, 21, 117.

8865

dx.doi.org/10.1021/ac402197h | Anal. Chem. 2013, 85, 8857−8865