Stable Isotopic Evidence for the Widespread ... - ACS Publications

Research Article pubs.acs.org/journal/ascecg

Stable Isotopic Evidence for the Widespread Presence of OxygenContaining Chemical Linkages between α‑Cellulose and Lignin in Poaceae (Gramineae) Grass Leaves Youping Zhou,*,†,§,○ Xijie Yin,‡,○ Hubiao Yang,∥ Jing Su,‡ Huimin Yu,† Yan Wang,§ Shuixiu Zhou,⊥ and Saša Zavadlav†,# †

School of Chemistry & Chemical Engineering, Shaanxi University of Science & Technology, Longshuo Rd, Xi’an, 710021, China Laboratory of Oceanic & Coastal Geology, The Third Institute of Oceanography, State Oceanic Administration, 178 Daxue Rd, Xiamen, 361005, China § CAS Key Laboratory of Coastal Environmental Processes and Ecological Remediation, Yantai Institute of Coastal Zone Research, Chinese Academy of Sciences, 17 Chunhui Rd, Yantai, 264003, China ∥ Tropical Pasture Research Centre, Tropical Crops Genetic Resources Institute, Chinese Academy of Tropical Agricultural Sciences, Baodao Xincun, Danzhou, 571737, China ⊥ Department of Internal Medicine, The University Hospital, Xiamen University, 172 Daxue Rd, Xiamen, 361005, China # Department of Forest Yield and Silviculture, Slovenian Forestry Institute, Večna pot 2, 1000 Ljubljana, Slovenia ‡

S Supporting Information *

ABSTRACT: The chemical linkage between α-cellulose and lignin in plant cell walls has long been a controversial topic and crucial to devising effective strategies for sustainable biomass and bioenergy utilization. In this new contribution, we surveyed 80 Poaceae (Gramineae) species grown in tropical Hainan Island (China) to test the hypothesis that the presence of oxygen-containing chemical linkage in Poaceae species is widespread. Our innovative natural abundance oxygen isotopic analysis allowed us to infer that more than 1/3 of the species investigated has chemical (ether) linkages between α-cellulose and lignin in their leaf cell walls, with a species-specific 2−89 oxygen-containing bonds for every 1000 glucose units. However, the presence of such linkage appears to be phylogeny-dependent. On average, species of C3 photosynthetic mode are found to have more extensive oxygen-containing linkages than those of C4 photosynthetic mode. Our finding challenges the conventional view that no chemical bonds between α-cellulose and lignin are present in higher plant cell walls and calls for new strategies for further understanding of the chemical linkage between the two major constituents of cell walls. This is especially important in the context of renewed and growing interests in biomass, bioenergy, and plant cell wall structure studies. KEYWORDS: Plant cell wall, Chemical linkage, α-Cellulose, Lignin, Stable oxygen isotopes

■

and plant-environment relationship6−9 can also benefit from an unambiguous understanding of these chemical linkages. Earlier work10,11 has convincingly established the existence of chemical bonds between hemicellulose and lignin and that the chemical linkages between lignin monomers and the hydroxyl group of hemicelluloses are in the form of benzyl ethers, (γ)esters, and phenyl glycosides. However, the lack of an effective method to directly target the possible chemical bonds between lignin and α-cellulose has experienced slow progress in this regard. Lawoko et al.12,13 demonstrated, by fractionating the lignin-carbohydrate complex (LCC) of native wood, the existence of covalent linkages between lignin and all

INTRODUCTION Quantitative and reliable understanding of the interactions within and between the three major biochemical components (lignin, α-cellulose, and hemicellulose) that constitute higher plant cell walls1 is important. For example, the choice of effective methods for conversion of lignocellulosic materials for biofuel production depends on the exact nature of the chemical bonds within and among them, while plant engineering2 for better digestibility (nutritive values) for human and animal consumption relies on the understanding of the biochemical compositions and the linkages between and among them. Furthermore, predicting the degradability of plant materials and their role in global scale carbon cycle3,4 will be assisted by an unhindered appreciation of the complexity of cell wall biochemical components. Finally, exploiting isotopes at natural abundance to understand the physiology of cell wall formation5 © 2017 American Chemical Society

Received: December 16, 2016 Revised: January 22, 2017 Published: February 22, 2017 3250

DOI: 10.1021/acssuschemeng.6b03075 ACS Sustainable Chem. Eng. 2017, 5, 3250−3260

Research Article

ACS Sustainable Chemistry & Engineering

Table 1. Sample Information and Isotopic Measurement Results for HCLC and α-Cellulose Products for the 80 Investigated Grass Species from the Poaceae Familya

3251


Research Article

ACS Sustainable Chemistry & Engineering Table 1. continued

The parameters for making corrections to HCLC and α-cellullose products according to eq 3 are fair_moisture = 0.05; δ18Oair_moisture = −12.9‰; α = 1.008. P: perennial. A: annual. *: C4 subtype uncertain. HCLC: holocellulose−lignin complex. meas: measured. corr: corrected. sd: standard deviation. min: minimum Δenrich‑unenrich,α‑cellulose or percentage of O-containing linkage between α-cellulose and lignin. Numbers in the brackets are the standard deviations. C4 subtyping was mainly based on the anatomic criteria described in the work of Hattersley and Watson55 and anatomical data from refs 56−61 was also referenced. Highlighted in red are the species identified to have O-containing chemical linkages between α-cellulose and lignin even if the maximum of contamination from hemicellulose is taken into account. C3 and C4 typing were based on δ13C measurements (data not shown here) of bulk leaf materials. NADP-ME: nicotinamide adenine dinucleotide phosphate-malic enzyme. NAD-ME: nicotinamide adenine dinucleotide-malic enzyme. PCK: phosphoenolpyruvate carboxylase. a

removal of extractives with solvents, lignin with bleaching and hemicellulose with strong alkali) from plant material so that the delignification step follows the hemicellulose removal step. As lignin is known to be extensively bonded to hemicelluloses, the hydrolytic removal of hemicelluloses ensures that not only the hemicellulose is entirely removed but also the part of the lignin that is linked to α-cellulose is preserved, leaving the α-celluloselignin complex (α-CLC) intact. The removal of hemicelluloses also opens up the lignin-carbohydrate “network”, thus facilitating the subsequent removal of lignin by bleaching. Oxygen isotopic analysis of α-cellulose products isolated from bleach-treated α-CLC in solutions of contrasting oxygen isotopic compositions allows for an inference to be made with regard to the presence/absence of oxygen-containing chemical bonds between α-cellulose and lignin. The application of this isotope-based method led to the finding that oxygencontaining chemical bonds between α-cellulose and lignin in Zea mays leaf cell walls exists, however, a similar conclusion could not be drawn for Araucaria cunninghamii wood. Since the publication of the innovative natural abundance isotope-based method, independent evidence in support of Zhou et al.26 findings has been provided. For example, Studer et al.27 showed that the amount of glucose released by hydrolysis from plant wood is negatively correlated with lignin content, a finding consistent with the inference described above that higher lignin content is most likely to be associated with a more extensive chemical linkage between α-cellulose and lignin and

carbohydrates in wood but fell short of specifically demonstrating the presence of oxygen-containing chemical bonds between lignin and α-cellulose. Jin et al.14 presented an indirect method to probe the presence of covalent linkage between lignin and cellulose based on comparative 1H NMR measurements of the degree of substitution of H in carboxymethylated LCC before and after being hydrolyzed in deuterated sulfuric acid (D2SO4/ D2O). Although the results from this study could be interpreted as the presence of a chemical linkage between lignin and αcellulose, they still fell short of elucidating the nature and the extent of such linkage. The failure of the past work to prove/disprove the presence of chemical linkage between lignin and α-cellulose was because past methods failed to isolate the α-cellulose-lignin complex (αCLC). Indeed, none of the past methods (i.e., acid degradation to target ether bonds,11 alkali degradation to target ester bonds,15−17 degradation to target glycosidic linkages,18,19 methylation to target ether bonds,20−22 and the more recent method23 of 13C-labeled precursor usage in differentiating xylem followed by NMR analysis) have made attempt to purify α-CLC. The problem is exacerbated by the fact that there are fewer opportunities for chemical linkages between the αcellulose and lignin as the cellulosic hydroxyl (−OH) groups are largely bound up in the formation of the crystalline structure of α-cellulose.24,25 To overcome such shortcomings, Zhou et al.26 altered the “standard” procedure of α-cellulose isolation (i.e., sequential 3252


Research Article


Figure 1. Flowchart for the isolation and purification of α-cellulose from field grown grass leaf samples collected from Hainan Island. island in the South China Sea. The relatively small size and small altitudinal range of this island ensure the cross-island climate varies little, making it an ideal location to sample field grown Poaceae species for phytochemical analysis. In addition, previous field surveys for germplasm resources and ecophysiological studies33,34 have established that there are 224 Poaceae species (all monocotyledonous with parallel venation pattern belonging to 189 genera) across the island, of which 189 species utilize C4 and 35 species utilize C3 photosynthetic carbon assimilation mode. This enabled us to look at whether these two different photosynthetic modes might influence the structure of the plant cell wall. A total of 80 species were randomly sampled across the island for healthy leaves over a narrow time window (24−28 June 2015) to minimize the effect of age and environment on the presence and extent of chemical linkages between lignin and α-cellulose. Over one-third of the species found on this island have been sampled to ensure the studied species are representative of the tropical members of the Poaceae family (species and photosynthetic modes are listed in Table 1, Table S1, and also in part in Zhou et al.35). Sample Processing for Isotopic Analysis. For isotopic and chemical analysis of the collected plant samples, we followed the method of Zhou et al.26 with minor modifications (Figure 1). Briefly, samples were first washed in deionized water (Merck Millipore, Shanghai, R ≥ 18.2 MΩ.cm), oven-dried at 70 °C, and then powdered and sieved (400 mesh). The uniformly sized powdered samples were then exhaustively and sequentially extracted with analytical grade methanol (CAS number 67-56-1) /dichloromethane (CAS number 75-09-2 (2/1, v/v), acetone (CAS number 67-64-1, all chemicals were obtained from the Shanghai Chemical Reagent Research Institute Co.

thus, lower releasable glucose. In studying the interaction behavior between native lignin and α-cellulose, Zhang et al.28 concluded that herbaceous biomass has more extensive chemical linkages between lignin and α-cellulose compared to woody biomass and that their finding was cited as “completely consistent with the results of Zhou et al.”26 Tropical grass species in the Poaceae (Gramineae) family have important economic, agricultural, and horticultural values.29−32 Since our previous work showed there are oxygen-containing chemical linkages between α-cellulose and lignin in the leaf of Zea mays, a typical Poaceae species of C4 photosynthetic mode, we suspect that such linkage is common in Poaceae family. In this paper, we report on the presence/ absence of these oxygen-containing linkages between αcellulose and lignin, using the stable-isotope-based method, in 80 species of tropical grasses in the Poaceae family in their native state. Our results show that such chemical linkage is widespread.

■

MATERIALS AND METHODS

Poaceae Species Chosen for the Test. If present, the extent of any chemical linkage between α-cellulose and lignin in plant cell walls could vary depending on a range of different interacting factors including the plant genetics, environmental conditions, plant growth stage, and tissue selection.12,13 To minimize environmental and age effects, we chose the Poaceae species from Hainan Island, a small 3253


Research Article

ACS Sustainable Chemistry & Engineering Ltd., Shanghai), water, and finally absolute ethanol (CAS number 6417-5, Sigma-Aldrich, Shanghai) to remove nonstructural components. After centrifuging at 3000g on a Beckman Allegra 64R benchtop centrifuge (Beckman Coulter, Shanghai), the pelleted holocellulose− lignin complex (HCLC) was oven-dried overnight at 70 °C. After taking an aliquot for δ18O analysis, the remainder was then hydrolyzed in a sealed glass centrifuge tube under dry N2 (to prevent the ambient CO2 entering into the alkaline solution and changing the pH and the oxygen isotopic composition of the hydrolytic solution) in concentrated NaOH (17%, w/v; CAS 1310-73-2, Aladdin, Shanghai) aqueous solution for 2 h at 70 °C to remove hemicellulose. Under these conditions, the glycosidic bonds joining the hemicellulose carbohydrate monomers, and the ester bonds linking both hemicellulose to lignin, and α-cellulose to lignin (if present) would have been completely cleaved, leaving the constituent ether bonds (stable to alkaline conditions) of both lignin and between lignin and α-cellulose, intact. The HCLC pellets were washed thoroughly with deionized water (δ18O = −5‰) at 70 °C for 2 h, settled by centrifuging at 3000g, and then dried with absolute ethanol. The dried samples were then kept sealed and stored in a desiccator filled with dry N2 before being analyzed for isotopic composition following the procedure described below (Stable Oxygen Isotope Analysis). In the delignification step, bleaching solution (15 g NaClO2 (CAS number 7758-19-2, Aladdin, Shanghai) and 10 mL CH3COOH (CAS number 64-19-7, Aladdin, Shanghai) were dissolved in 1050 mL of water (see also Table 1 of the work of Zhou et al.26) of contrasting oxygen isotopic compositions (δ18O = −5‰ and 298‰ prepared by diluting 18O-water (97%-18O, CAS Number 14314, Sigma-Aldrich, Shanghai) with −5‰ deionized water, for the unenriched and enriched solutions, respectively) were added to α-CLC and the bleaching was allowed to proceed at 70 °C. To avoid changes in the isotopic composition of the bleaching media by exposure to ambient CO2 and other oxygen-containing compounds, the headspace of the reaction tubes were filled with N2. We have shown in our previous work26 that after 1.5 h of bleaching the recovered α-cellulose products are both chemically and oxygen-isotopically stable. To ensure the completeness of bleaching, in this work, we extend the bleaching time to 2 h and replaced the bleaching solution after 1 h to maintain the oxidative power of the bleach. On completion of bleaching, the samples were immediately frozen in liquid N2 and then freeze-dried in a Christ Alpha 1−4 LD plus freeze drier (BMH., Beijing, China) before being washed with copious amounts of deionized water (−5‰) at 70 °C for 2 h to allow complete replacement of any adsorbed bleaching solution with water, pelleted by centrifuging at 3000g, then washed with absolute ethanol and dried. The dried samples were then sealed and kept in a desiccator filled with dry N2 before being analyzed for δ18O following the procedure described below (Stable Oxygen Isotope Analysis). In our previous work,26 we have shown that (i) our method has produced α-cellulose hemicellulose of high purity: contamination from lignin was beyond detection while that from hemicellulose was less than 4% as a result of conducting the hemicellulose removal step before delignification (bleaching), which opened up the “nanopores” formed by the microfibrils of cellulose and interconnecting hemicellulose where the lignin resides and therefore facilitating the oxidative removal of lignin, (ii) the method could keep the possible linkage between α-cellulose and lignin in the isolated α-CLC intact, and (iii) the method was effective and novel in probing the presence of oxygen-containing linkage between α-cellulose and lignin in the isolated α-CLC. However, we have not quantified (a) the effect of O from NaClO2 and CH3COOH in changing the δ18O of processing media, (b) the effect of adsorption of water from different bleaching solutions onto the final α-cellulose product, and (c) the effect of possible exposure to laboratory air while sitting the sample on an autosampler tray before being analyzed for 18O/16O ratio. As these effects may alter/skew the isotopic compositions of HCLC and αcellulose products, we added three more tests below to address these issues.

Effect of NaClO2 and CH3COOH on the Isotopic Compositions of Bleaching Solution. Although NaClO2 and CH3COOH do not change isotopically in the solution, the isotopic composition of bleaching solutions may change when bleaching starts, as NaClO2 is gradually spent and its oxygen ends up in the solution (see Figure 8 of Zhou et al.26). Therefore, it is warranted to know how the isotopic compositions of bleaching solutions change over the course of reaction as such information is needed for reliable and quantitative interpretation of the isotopic differences between the α-cellulose products isolated from the isotopically contrasting bleaching media. It is difficult to predict quantitatively how the bleaching of lignin will change the isotopic composition of bleaching solution because the mechanism of lignin destruction by the bleach is far from clear. A general consensus is that during bleaching, the aromatic rings of the lignin monomers are first cleaved to form dicarboxylic acids of various sizes and with various substitutions but ultimately are completely oxidized to CO2.36 Thus, we decided to monitor the isotopic compositions of bleaching solutions in the 2 h bleaching course by taking aliquots at 0.5, 1, 1.5, and 2 h. Time Needed to Completely Replace Bleaching Solution with Deionized Water after the Bleaching Step. To determine the time needed for complete replacement of the bleaching solution adsorbed onto the final α-cellulose product, we randomly chose five pairs (enriched versus unenriched) of α-cellulose samples (processed from I. cylindrica, P. maximum, B. bladii, M. floridulus, and P. orbuculare) on completion of bleaching and immediately froze them in liquid N2. Following freeze-drying, the samples were dispersed in deionized water (δ18O = −5‰) at 70 °C to allow replacement of adsorbed water. Aliquots taken at 0.5, 1, 1.5, 2, and 3 h were frozen in liquid N2 and then freeze-dried. The freeze-dried samples were kept sealed in a desiccator filled with dry N2 before being analyzed for δ18O following the procedure described below (Stable Oxygen Isotope Analysis). Effect of Exposure to Laboratory Air Moisture on the Isotopic Compositions of HCLC and α-Cellulose. As described below (Stable Oxygen Isotope Analysis), all HCLC and α-cellulose samples and standards were loaded into (but not sealed) tin capsules and dried in an oven for a minimum of 7 days to allow complete equilibration of adsorbed moisture with laboratory air.26 Immediately prior to being analyzed, the tin capsules were taken out from the drying oven and tightly sealed. To estimate the maximum effect of laboratory air moisture on the measured isotopic compositions of the HCLC and α-cellulose samples, we cryogenically collected laboratory air moisture with a dry ice-ethanol trap under the suction of a pump following the procedure described in the work of Gan et al.37 and Zhou et al.38 Laboratory air temperature and relative humidity were kept constant and recorded and were used to calculate the true isotopic compositions for the δ18O of the HCLC and α-cellulose using the following equation.

⎧ ⎡ 80 ⎤ 16 δ18O true ≈ ⎨δ18Omeasured ⎢ (1 − fair moisture ) + f ⎥ ⎣ 162 ⎩ 18 air moisture ⎦ 16 ⎫ − (δ18Oair moisture + 1000ln α)fair moisture ⎬ 18 ⎭ ⎧ 80 ⎫ /⎨ (1 − fair moisture )⎬ ⎩ 162 ⎭ (1) where δ18Otrue, δ18Omeasured, δ18Oair moisture are the true and measured isotopic compositions of samples, and the isotopic compositions of laboratory air, respectively, while fair moisture and α are the weight percentage of adsorbed air moisture in the samples and isotopic fractionation factor between air moisture and the adsorbed moisture under oven conditions (temperature 70 °C; relative humidity 50%), respectively. We use 1.008 for α, the equilibrium isotope fractionation (EIF) factor at 20 °C between liquid water and surrounding air moisture39 as an approximation for the EIF between water adsorbed onto cellulose and surrounding air moisture as there is no reported EIF factor for the latter, and 5% for fair moisture according to Saurer and Siegwolf40 and Stamm.41 The measured laboratory moisture had a 3254


Research Article

ACS Sustainable Chemistry & Engineering δ18O value of −12.9‰. 16/18 and 80/162 refer to the O percentage of water (H2O) and α-cellulose (C6(H2O)5), respectively. For hemicellulose, the percentage is slightly higher as the dominating basic unit in hemicelluloses has a chemical formula of C5(H2O)4. Stable Oxygen Isotope Analysis. The 18O/16O analysis was conducted using a minor modification of the method described in Farquhar et al.42 on a Flash HT (1112 series) EA (elemental analyzer) coupled to a MAT 253 IRMS (isotope ratio mass spectrometer) via a Conflo III (Thermo Fisher Scientific GmbH, Bremen). For routine analysis, 0.2−0.3 mg of solid samples (HCLC, α-cellulose and isotope standards) were pyrolyzed in a glassy-carbon lined ceramic tube in the Flash HT at a temperature of 1360 °C where oxygen in the samples was quantitatively converted to CO.43,44 After being dried in-line with a MgClO4 trap and separated from other gases on a GC column (stainless steel, 10 m long, 10 mm o.d.) filled with 5 Å molecular sieve (SANTIS, cat. No. SA990724), the CO was carried to the IRMS by a He (99.999%) flow via an open split, where the C18O/C16O ratio (representing the 18O/16O of the sample) was analyzed. The reference used in the analysis was high purity CO (99.999%) with a δ18OVSMOW of 8.8‰ calibrated by two international oxygen isotope reference materials: benzoic acid (IAEA-601) with a δ18OVSMOW of 23.2‰ and Ag3PO4 with a δ18OVSMOW of 21.7‰ (purchased from Elemental Microanalysis Ltd., Part No: B2207, Certificate No: BN180097). The deionized water used for hydrolysis and bleaching solution preparation, together with recovered bleaching solutions at various stages of bleaching and trapped laboratory air moisture (all 1 μL) were analyzed in the same way as the solids after being sealed in smoothwalled tin capsules with a pneumatically controlled crimper. To ensure batch to batch comparability, the temperature and humidity in the mass spectrometry facility was kept constant throughout the analytical course. All isotopic measurements are reported in the standard delta notation relative to V-SMOW: δ18O = (Rsample − Rstandard)/ Rstandard*1000‰, where Rsample and Rstandard are the 18O/16O ratios of sample and standard, respectively. External precisions of repeated analysis of isotopic standards gave a value better than 0.3‰ for benzoic acid (IAEA-601) and 0.25‰ for in-house water standard (−5.5‰, calibrated against IAEA VSMOW2), respectively, over a period of 7 days of the analysis. All isotopic data for the samples (average values and standard deviations) was analyzed in triplicate. Prior to isotopic analysis, all samples were loaded into tin capsules (unsealed) and dried at 70 °C for a minimum of 7 days to allow complete re-equilibration of any residual (adsorbed) water with laboratory air moisture (we have proved that a minimum of 5 days is needed for the residual water in α-cellulose to re-equilibrate with laboratory air moisture, see Table S1). The dried samples were then tightly wrapped (but not sealed), therefore it was still possible for further isotopic change due to the (possible) subsequent progressive adsorption of air moisture while the samples sat in the autosampler tray. To minimize the isotopic effect of such exposure, only six samples per batch were loaded into the autosampler at any one time thus no sample was exposed to the ambient laboratory atmosphere for more than 1 h. Statistical Analysis. Two-tailed F-test was used to test if the variances of the two groups (C3 and C4) are equal and t test was used to test if the means of the two groups are significantly different using the inbuilt statistic functions of Microsoft Excel (2010) .

Figure 2. Isotopic variations of the bleaching media during the course of bleaching experiments with a randomly chosen HCLC sample (processed from Centotheca lappacea (L.) Desv.). Both the enriched and unenriched media showed a slight increase in δ18O values as bleaching progresses, but the difference between the two media was almost unchanged.

unenriched solutions during the 2-h-long bleaching course, therefore the contribution of NaClO2 and CH3COOH as a whole to the isotopic difference between α-celluloses recovered from enriched and unenriched media was negligible. Minimum Time to Replace Residual Processing Medium. The isotopic changes for the five randomly chosen pairs (enriched versus unenriched) of α-cellulose products (processed from I. cylindrica, P. maximum, B. bladii, M. floridulus, and P. orbuculare) over a period of 3 h heating at 70 °C in deionized water to establish the minimal time for the complete replacement of residual (adsorbed) water are presented in Figure 3 (for data please see Table S3). No significant isotopic changes were observed after 0.5 h heating in deionized water, indicating that 0.5 h was sufficient to

■

RESULTS AND DISCUSSION Effects of NaClO2 and CH3COOH, Adsorption of Water from Bleaching Solutions onto α-Cellulose and Exposure to Laboratory Air Moisture. The isotopic variations of the bleaching media during the course of bleaching experiments with a randomly chosen HCLC sample (processed from Centotheca lappacea (L.) Desv.) are presented in Figure 2 (for data, please refer to Table S2). It can be seen that although both enriched and unenriched media showed a slight increase in δ18O values as bleaching progressed, there was no significant variation in the isotopic difference between the enriched and

Figure 3. Isotopic change for five randomly chosen pairs (enriched versus unenriched) of α-cellulose products (processed from I. cylindrica, P. maximum, B. bladii, M. floridulus, and P. orbuculare) over a period of 3 h heating at 70 °C in Milli-Q water to establish the minimal time for the replacement of residual (adsorbed water) to complete. 3255


Research Article


the xylosyl unit, only one is locked in the glycosidic bond which is affected by hydrolysis, while the other three are locked in the nonhydrolyzable nonglycosidic bonds. Although we have shown our altered method26 can produce α-cellulose with impurity less than 4%, in our calculations we used a more conservative value of 10% for F.46,47 With these values, the maximum Δ18Ohemicellulose,enrich−unenrich was estimated to be approximately 1.9‰. Moisture Adsorption. Hygroscopicity is a property of hydroxyl-group-abundant plant materials. Theoretically, there is no absolutely dry native plant material which always carries −OH groups; there is always some water adsorbed onto the surface of the amorphous domain and some trapped in the “nanopores” which are formed by hemicellulose and α-cellulose fibrils networks of the crystalline domain.48 When exposed to laboratory air moisture, exchange between adsorbed and trapped water and air moisture will occur. The exchangeability, and final equilibrium moisture content, EMC, depends on the amount of the −OH per unit mass and also on the crystallinity, which determines the accessibility of water trapped in the crystalline domain of the materials.48,49 Under the analytical conditions used in our work, placing the samples in the autosampler tray for up to 1 h resulted in no significant changes in the δ18O values (Figure 4). This can be attributed to (a) the exchangeable water, when compared to that of the inaccessible (inexchangeable under the conditions used) water in the samples, was quantitatively insignificant, and (b) 1 h exposure to laboratory air was too short a time period for the exchange between adsorbed moisture and air moisture to be in equilibrium. In other words, much more time would be required before any appreciable isotopic change could be observed. Although analytical delays did not result in isotopic changes, adsorbance of laboratory air moisture, while being dried in an oven at 70 °C for a minimum of 7 days, could result in changes to the isotopic composition. Rewriting eq 1, the correction made to the measured isotopic compositions of the samples can be expressed as

completely replace the adsorbed residual water, following the bleaching experiments. Exposure of α-Cellulose to Laboratory Air Moisture while in the Autosampler Prior to Isotopic Analysis. The isotopic changes to an E. atrovirens α-cellulose sample and a standard over a period of 1 h while placed in the autosampler tray are presented in Figure 4 (see also Table S4 for data). It

Figure 4. Isotopic change of an α-cellulose product (processed from E. atrovirens) and an α-cellulose standard (from Sigma-Aldrich) over a period of 1 h sitting on an autosampler tray.

can be seen that, under the analytical conditions used in this work, the maximum change observed for the measured isotopic composition was