Nanoscale Structure of the Cell Wall Protecting Cellulose from

Environ. Sci. Technol. 2011, 45, 1107–1113

Nanoscale Structure of the Cell Wall Protecting Cellulose from Enzyme Attack F A B R I Z I O A D A N I , * ,† G A B R I E L L A P A P A , † ANDREA SCHIEVANO,† GIOVANNI CARDINALE,‡ GIULIANA D’IMPORZANO,† AND FULVIA TAMBONE† Gruppo RICICLA, Dipartimento di Produzione Vegetale, Universita` degli Studi di Milano, Via Celoria 2, 20133 Milano, Italy, and Centro Ricerche ENEA, Casaccia, Via Anguillarese 301, 00123 S. Maria di Galeria, Roma, Italy

Received June 15, 2010. Revised manuscript received November 29, 2010. Accepted December 9, 2010.

The cell wall structure protects cellulose from enzymatic attack and its successive fermentation. The nature of this protection consists in the very complex macroscopic and microscopic structure of cell wall that limits transport. Explaining this kind of protection is critical in future research to improve cell polymer availability for enzymatic attack. This research shows that the complete description of the cell wall topography at a nanoscale level allows a mechanistic understanding of cellulose protection. For this purpose, we used gas adsorption methods (CO2 at 273 K and N2 at 77 K) to detect mesoporosity (pore size of 1.5-30 nm diameter; MeS) and microporosity (pore size of 0.3-1.5 nm diameter; MiS) of the cell wall of five energy crops, i.e., giant cane, rivet wheat straw, miscanthus, proso millet, and sorghum. The presence of both hemicelluloses in the spaces between cellulose fibrils and the unhydrolyzable and highly cross-linked lignocarbohydrate complex (LCC) determines a microporous (80% pores having diameters below 0.8 nm) structure of the cell wall that prevents the cellulase enzymes from coming into direct contact with the cellulose, as their sizes exceed the cell wall pore size. On the other hand, the removal of the hemicelluloses and of the LCC complex determines a reduction of the MiS and an increase of the available surface for enzymatic attack, i.e., pores >5 nm diameter. This was confirmed by the good negative (r ) -0.87, P < 0.001, n ) 11) and positive (r ) 0.78, P < 0.005, n ) 11) correlations found for microporosity and mesoporosity (pores of diameters >5 nm), respectively, vs the glucose production, by cellulase enzyme attack in specific enzymatic hydrolysis tests performed on biomass samples.

Introduction The bioethanol industry is expanding at an extraordinary rate (1-3) and lignocellulosic material (mainly plant cell wall) could represent an abundant source of ethanol (4), i.e., second-generation bioethanol. However, the development of processes related to the production of biofuels utilizing lignocellulosic biomass (such as agricultural residues, * Corresponding author phone: +39 0250316546; fax: +39 0250316521; e-mail: [email protected]. † Universita` degli Studi di Milano. ‡ Centro Ricerche ENEA. 10.1021/es1020263

 2011 American Chemical Society

Published on Web 12/21/2010

forestry wastes and thinning, waste paper, and energy crops) is still in the early stage of research and development (5) due to lack of supporting studies to improve biodegradability of these feedstocks. Efficient deconstruction of biomass is essential to the widespread use of cellulose polymer. The use of plant biomass to produce bioethanol involves the enzymatic degradation of cell wall polysaccharides to monosaccharide and its successive fermentation (6); on the other hand, cell wall structure has a natural recalcitrance that limits enzyme activity (7, 8). The term recalcitrance is used to describe the phenomenon by which plant tissues exhibit natural resistance against microbial and enzymatic deconstruction (7, 9). Despite recent progress, the nature of plant biopolymer recalcitrance remains unclear and new methodological approaches may be promising tools to identify the ultrastructure and the chemical topography of plant cell walls (7). There is currently increasing recognition of the importance of elucidating and studying in depth the complexity of the cell wall of higher plants, including nanostructure property and its relationship to recalcitrance, thereby employing this knowledge to develop the efficacy of enzymatic hydrolysis of plant polymers. In fact, it was reported that biochemical catalyzes are limited in their actions by the very complex macroscopic and, above all, microscopic structure of cell wall that limits transport (7). A mature cell wall is formed out of biomacromolecules whose main constituent is the cellulose, a homopolymer of β-1,4-linked glucose units forming the core of the microfibril structures under crystalline type (10, 11), that represents the more dominant substrate of the cell wall making up 40.6-51.2% of the wall material (6). The complexity of the cell wall gives to it physical properties, i.e., the presence of porosity and porosity surface area, that play an important role in the cell wall biological digestion (12). The study of the cell wall porosity was a subject of research in the past using different approaches, since the use of a molecular probe of known dimension was proposed (13). Later, the freezefracture electron microscopic technique (14), solute exclusion technique (15), mercury porosimeter, electron micrograph NMR, gas adsorption and others (16) were used. The methods used to detect porosity can affect measurements, thereby showing drawbacks and limitation (12). For example, some methods (14, 15) provide a single mean value or range of pore size, and they are not able to provide quantitative information about pore size distribution and data about surface area (12). The gas adsorption method is an alternative physical-chemical technique that is commonly used for cell wall porosity and surface area measurements (12, 17, 18). This method allows one to describe pore size distribution and their quantification, resulting also in the measure of the total surface area of pores. N2 adsorption at 77 K is the most widely used method, although this approach is limited in the detection of the micropores (pores below 1.5 nm diameter). Nevertheless, the use of CO2 adsorption technique at 273 K overcame this problem, allowing detection of pores of 0.3-1.5 nm diameter (19, 20). This is a key point because cell wall architecture is characterized by a complex network of polymers, determining the presence of pores of diameters in the range of the both mesoporosity (1.5-50 nm diameter) and microporosity (0-1.5 nm diameter) (10, 21, 22). This porosity determines limitations in the enzyme accessibility into the cell wall structure and, thereby, a limited cell wall surface available for the enzymatic attack (12). In fact, the access of cellulase enzymes to crystalline cellulose was found greatly reduced by the presence above all of hemicelluloses but also of pectin, lignin, and structural VOL. 45, NO. 3, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1107

protein that coat the crystalline cellulose, forming a microporous structure not accessible to enzymes (7). Cellulases were reported, in fact, to have finite dimensions (in the range of 4-13 nm) (23, 24) greater than those of cell wall pores, so that they are unable to come into direct contact with crystalline cellulose. This fact leads to the presence of a natural recalcitrance of the cell wall to the enzymatic attack that is related to the cell wall structure at nanoscale level (7). In this paper, we defined and measured the recalcitrant nature of cellulose in the cell wall, giving a mechanistic understanding of the phenomenon, using gas adsorption methods (CO2 at 273 K and N2 at 77 K) to describe the topography at a nanometer-scale level of the cell wall of five biomasses, i.e., giant cane (Arundo donax L.), rivet wheat (Triticum turgidum L.), miscanthus (Miscanthus giganteus L.), proso millet (Panicum miliaceum L.), and sorghum (Sorghum bicolor L.).

Materials and Methods Biomass Samples. The biomasses came from croplands cultivated during the year 2007-2008 in the Po Valley (North Italy). Agronomical details are reported in the Supporting Information (Table S1). The biomass samples, composed of the whole plant, were brought to the laboratory and chopped to 1 cm length. Part of the samples were dried in an oven at 35 °C overnight and at 45 °C until it had a constant weight, and then, it was ground to a 1 mm size. Dried samples were stored for successive analyses. Cell wall of onions (Allium cepa var. Jumbo) were prepared such as those previously reported (14). Physical and Chemical Biomass Treatments. Hemicelluloses were removed by steam explosion (SE) technology, performed by a batch steam explosion plant (Research Centre Trisaia; ENEA, Italy) with a capacity of 0.5 kg biomass. SE treatments were performed by steam at 220 °C and 15 bar, for 7 min. In particular, about 500 g of wet weight plant mass were used by adopting a mass to water ratio of 1:1. Treated samples were successively dried in an oven at 35 °C overnight and then at 45 °C until they had constant weight, and then, they were ground to a 1 mm size. Lignocarbohydrate complexes (LCC) were isolated by treating dried samples (ground to 1 mm) with H2SO4, 13.50 mol L-1, for 24 h at 4 °C (25). After drying, the acid-insoluble LCCs were recovered and stored for successive analyses. Mass losses after both physical and chemical treatments were determined by weighing samples before and after treatment. CP MAS 13C NMR Analysis. The solid-state CP MAS 13C NMR (13C NMR) spectra of the biomass samples were acquired on dried samples at 10 kHz on a Bruker AMX 600 spectrometer (Bruker BioSpin GmbH, Rheinstetten) using a 4 mm CP MAS probe. The pulse repetition rate was set at 0.5 s, the contact time at 1 ms, the line bordering at 80 Hz, and the number of scans was 1600. The chemical shift scale of 13C NMR spectra was referred to tetramethylsilane (δ ) 0 ppm). The spectra were elaborated using TOPSPIN 1.3 software (Bruker BioSpin GmbH, Rheinstetten, Germany). Cellulose crystallinity, measured by Crystallinity Index (CrICPMAS), was determined from the areas of crystalline (86-92 ppm) and amorphous (79-86 ppm) C-4 signals: CrI ) [A86-92 ppm/(A79-86 ppm + A86-92 ppm) × 100%]. Gas Adsorption Measurements. Micro- and mesoporosity surfaces (m2 g-1; MiS and MeS, respectively) and corresponding volumes (cm3 g-1; MiV and MeV) were detected on dried samples by gas adsorption measurements conducted using a Porosimeter NOVA 2200e Quantachrome Surface Area and Pore Size Analyzer (Quantachrome, Boynton Beach, FL, USA). Analyses were carried out using CO2 at 273 K for MiS and MiV and N2 at 77 K for MeS and MeV determination, respectively. Analyses were preceded by a vacuum-degassing procedure performed at 80 °C for 16 h. MiS and MiV refer 1108

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 3, 2011

to the surface and the volume of pores with diameter in the range of 0.3-1.5 nm, subdivided into two classes: ultramicropores (pores with diameter about 0.8 nm) and supermicropores (pores with diameter in the range of 0.8-1.5 nm). MeS and MeV refer to the surfaces of pores with diameters in the range of 1.5-30 nm. Surface and volume were calculated by applying the density functional theory (DFT) (26-28) method to both CO2 (MiS and MiV measurement) and N2 adsorption data. The DFT software version 2.2 available with Quantachrome (Quantachrome, Boynton Beach, FL, USA) was used for calculations. All measurements were performed at least in triplicate. Fractal dimension D (29) was obtained from the isotherm adsorption performed by N2 at 77 K, where D assumes values between 2 (smooth and regular surface) and 3 (very rough surface) and was calculated using the Neimark-Kiselev (NK) method (30). Enzymatic Hydrolysis Test. Enzymatic hydrolysis tests were performed according to the method used for characterizing cellulose enzyme activity (31). The tests were performed on dried and ground samples. Two hundred milligrams of pure cellulose or cellulose contained in plants were suspended in 10 mL of sodium acetate buffer solution, 50 mmol L-1, and adjusted to pH 5.00 ( 0.05, with HCl, 2 mol L-1. Cellulase enzyme (endo-1,4 β-glucanase, EC 3.2.1.4 from Trichoderma reesei ATCC26921, Sigma Aldrich) was added (0.5 units of cellulase mL-1 of liquid suspension) to the suspension (32). The mix was incubated at 37 °C, and glucose formation was determined at points where it was linear with time (approximately 2 h), maximizing differences between different theses (33, 34). Then, the suspensions were filtered to determine glucose concentrations in the solutions. The enzymatic generation of nicotinamide-adenine dinucleotide, that is formed stoichiometrically from the amount of glucose using an enzymatic kit test (Boehringer Mannheim enzymatic tests of the R-Biopharm AG, Darmstadt, Germany), was measured by a spectrometer at 340 nm (Varian CARY 50 BIO Spectrophotometer). Manufacturer recommendations were followed for each enzyme test. Statistical Analyses. All statistical analyses were performed using the SPSS statistical software (version 17; SPSS, Chicago, IL).

Results and Discussion Meso- and Microporosity Detections. The studied biomass samples showed mesopores (MeS) in the range of 1.5-30 nm diameter with the most frequent (mode value) pore size value of 3.44 ( 0.49 nm diameter and a surface area (MeS) in the range of 0.78-1.17 m2 g-1 dry matter (dm; average of 0.91 ( 0.15 m2 g-1 dm, n ) 5); Figure 1a) (Supporting Information, Table S2). These results agree with data reported by other authors who, using the same analytical technique, indicated 2-8 nm as typical pore diameters of the cell wall of crop plants (35) or reported for wheat straw pores of 1.5-2.5 nm as predominant pore class size diameters (12). The CO2 gas adsorption method revealed the presence of microporosity in all biomass samples (pore of 0.3-1.5 nm diameter), with high micropore surface (MiS, range of 38.2-53.8 m2 g-1 dm and average of 48.1 ( 6.7 m2 g-1 dm, n ) 5; Figure 1b; Supporting Information, Table S2) compared to MeS (Figure 1a; Supporting Information, Table S2). The MiS corresponded to proportional micropore volume ranging from 0.013 to 0.019 cm3 g-1 dm (average of 0.016 ( 0.002 m2 g-1dm, n ) 5; Supporting Information, Tables S2 and S3). The division of MiS into two subclasses (ultramicropores: 0.3-0.8 nm diameter; supermicropores: 0.8-1.5 nm diameter; Supporting Information,Table S3) revealed that cell wall microporosity was mainly formed by ultamicropores, i.e., 82.7 ( 8.1% of the total MiS surface measured (n ) 5; data calculated from Supporting Information, Table S3).

FIGURE 1. (a) Mesoporosity surface (MeS) determined for untreated biomass samples and samples treated by steam explosion and acid hydrolysis. (b) Microporosity surface (MiS) determined for untreated biomass samples and samples treated by steam explosion and acid hydrolysis. Data for biomass samples after acid hydrolysis were reported, also referred to as the dry matter of untreated sample and calculated by the equation: Pc ) [Pd × (1000 - LH2SO4)/1000] in which, Pc and Pd represent MiS (m2 g-1) and MeS (m2 g-1) calculated and detected, respectively, and LH2SO4 represents the dm losses after acid hydrolysis (g kg-1dm). Errors bars ( SD. Histogram bars marked by the same letter are not statistically different (P < 0.05) within (small letters) and between (capital letter) biomasses studied, according to Tukey test. A criticism to our approach in measuring mesoporosity and microporosity consists of the fact that porosity detection on oven-dried biomass sample could have decreased surface area and pore volume, because of pore collapse during cell wall drying (17). In order to support our findings, and such as requested by one reviewer, we measured by the gas adsorption method the mesopority in the cell wall of an onion (Allium cepa var. Jumbo) to provide a comparison with the result previously reported using another method, i.e., the freeze-fracture electron microscopic technique (14). Our results showed that the 80% of the mesoporosity in the range of 2-25 nm diameter (Supporting Information, Table S4) was below 10 nm, in agreement with previous findings (14). In addition, the size of mesopores measured on our biomass samples (mode value of 3.44 ( 0.49 nm) agrees with pores size measured by the solute exclusion technique on a vegetative cell wall (pores size in the range of 3.5-5.2 nm) (15). These findings drove us to exclude that water affected mesoporosity measurements.We declined, also, the hypothesis that microporosity measured by gas adsorption was affected by water because cellulose fibrils are located in the cell wall, orienting their hydrophobic faces externally (7),

and because lignification increases hydrophobicity of cell wall excluding water from micropores (16). Cell Wall Structure. In describing the architecture of the cell wall, various models (10) indicated cellulose microfibril with uniform diameters within the range of 2.4-3.2 nm (21, 36) with a distance of about 3.6 nm (center to center) between two microfibrils. The space between cellulose fibrils is filled by hemicelluloses attached to cellulose by hydrogen bonds, reducing its size to 0.4 nm (36), i.e., in the ultramicroporosity range. This cell wall characteristic could explain why the biomasses studied in this work showed relatively low mesoporosity surface and, on the contrary, high microporosity surface (above all ultramicroporosity surface). Moreover, the measured ultramicropores showed, on average, a pore size diameter of 0.55 nm, i.e., very similar to the size of cellulose-hemicellulose structure in the cell wall (0.4 nm) (36, 37). Therefore, it can be stated that the ultramicroporosity that we measured effectively represented the space between cellulose fibrils filled by hemicelluloses. To confirm our hypothesis, the hemicelluloses were removed from the biomasses studied to get new MiS measurements, expecting their removal to cause a MiS VOL. 45, NO. 3, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1109

TABLE 1. Crystalline Indexes and Fractal Surface Dimension (D) Measured for the Biomass Samples before and after Hemicelluloses Removing samples

crystallinity index (CrICPMAS)a % fractal dimension (D)b

raw material after steamc

Giant Cane 36.3 49.8

3.00 ( 0.02 2.91 ( 0.07


Miscanthus 37.8 50.4

3.00 ( 0.06 2.81 ( 0.06


Proso Millet 34.7 51.8

2.99 ( 0.11 2.81 ( 0.07


Sorghum 34.6 47.7

3.02 ( 0.09 2.76 ( 0.04


Rivet Wheat (Straw) 36.3 48.3

2.88 ( 0.02 2.77 ( 0.05

Purchased Cellulosed 53 (52)e

2.53 ( 0.03

a

CrI was determined from the CP MAS 13C NMR areas of the crystalline (86-92 ppm) and amorphous (79-86 ppm) C4 signal: CrICPMAS ) A86-92 ppm/(A79-86 ppm + A86-92 ppm) × 100. b D was determined by surface adsorption isotherm using N2 at 77 K, applying the Neimark-Kiselev method; D assumes a value between 2 (smooth surface) and 3 (for rough surface). c Steam explosion treatment performed at 220 °C for 7 min. d Sigma 20, S3504, Sigma Co., St Louis, MO, USA. e Data indicated by Sigma.

reduction. For this purpose, we treated our biomass samples by steam explosion (performed at 220 °C for 7 min), a physical method commonly used to remove hemicelluloses (38). Hemicelluloses are easily removed, as steam allows the hydrogen bonding, which held the hemicelluloses attached

to fibril cellulose, to break (6, 39). Even though lignin was reported to be affected by SE (39, 40), we considered this as a minor effect with respect to hemicelluloses removal, as we observed in all samples studied a relative increase of aromatic-C (lignin) as consequence of the removal of the O-alky-C (cellulose + hemicellulose) contents (see 13C NMR data; Table 2), such as, also, confirmed by the literature (40). The preferential removal of hemicelluloses vs cellulose was suggested by the increase of the CrICPMAS (Table 1) calculated for treated vs untreated biomasses from 13C NMR area (Table 2), being CrICPMAS widely used in the literature (e.g. (39, 41),) to describe hemicelluloses removal by chemical-physical treatments. In addition, we found that CrICPMAS of commercial pure crystalline cellulose (Sigma 20, S3504, Sigma Co., St Louis, MO, USA) was very similar to that of the SE-treated samples (Table 1). With respect to the raw materials, the total MiS of the SE-treated biomasses decreased by more than half after the SE-treatment (MiS average reduction of 59.1 ( 5.2%, n ) 5; Figure 1b). Microporosity reduction interested above all ultramicropore class (pores class of 0.3-0.8 nm in size; Supporting Information, Table S3) and, as the average of the five biomasses studied, about 86.9 ( 19.3% of the total MiS reduction was due to the smallest pores. We assumed, therefore, that MiS reduction was due to the opening of the space between cellulose fibrils, as a consequence of the removal of hemicelluloses. This fact could lead to an increase of the mesoporosity surface, such as, effectively, we registered for all biomass samples studied (Figure 1a; Supporting Information, Table S4). Cellulose fibrils, with low hemicelluloses contents, were reported to be compact and smooth contrarily to fibers with high hemicellulose content that show a porous surface (rough surface) (42). The decrease of fractal dimension (D), that we measured for treated vs untreated biomass samples (Table 1), confirmed these findings, having D values of 2 reported

TABLE 2. Assignments and Relative Area of CP MAS 13C NMR Bands for Untreated Biomass Samples and Biomass Samples after Steam Explosion and Sulfuric Acid Hydrolysis Treatments band range (ppm)

c

0-50

50-115

115-160

160-210

biomass samples

aliphatic C bonded to other aliphatic chains or to H

O-CH3 or N-alkyl-C O-alkyl-C di-O-alkyl-C

aromatic-C, phenol-C, or phenyl ether-C

carboxyl-C + keto-C

raw materiala after steamb after hydrolysisc

10.77 13.63 27.74

Giant Cane 76.16 67.49 38.07

8.50 13.40 24.77

4.58 5.48 9.42


8.22 14.16 24.87

Miscanthus 75.73 68.35 40.06

10.73 12.90 25.98

5.32 4.59 9.09


8.94 15.09 28.41

Proso Millet 78.97 63.88 36.77

8.16 14.91 25.64

3.94 6.12 9.18


11.31 18.06 27.63

Sorghum 75.71 63.53 38.91

8.00 12.76 23.35

4.98 5.64 10.11


10.26 16.70 29.17

Rivet Wheat (Straw) 73.10 65.07 39.46

10.09 12.78 23.32

6.55 5.45 8.05

a Untreated biomass samples. b Biomass samples treated by steam explosion treatment performed at 220 °C for 7 min. Biomass samples after acid hydrolysis with H2SO4 13.50 mol L-1 at 4 °C for 24 h.

1110

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 3, 2011

FIGURE 2. (a) Glucose production after incubation at 37 °C for 2 h of biomass samples before and after hemicelluloses removal, and cellulose by dosing cellulase enzyme (cellulase from Trichoderma reesei ATCC26921, Sigma Aldrich; 0.5 unit of cellulase 20 mg-1 pure cellulose or cellulose contained in plants). Glucose in the liquid medium was detected spectrometrically by measuring the enzymatic generation of nicotinamide-adenine dinucleotide that is formed stoichiometrically from the amount of glucose, using an enzymatic kit test (Boehringer Mannheim enzymatic tests of the R-Biopharm AG, Darmstadt, Germany). Errors bars ( SD. Histogram bars marked by the same letter are not statistically different (P < 0.05) within (small letters) and between (capital letter) biomasses studied, according to Tukey test. (b) Correlation found for microporosity vs glucose production, by cellulase enzymatic hydrolysis tests performed on biomass samples. (c) Correlation found for mesoporosity (pores of diameters >5 nm) vs the glucose production by cellulase enzymatic hydrolysis tests performed on biomass samples. for smooth and regular surfaces and a value of 3 for a very rough surface (29, 43). Availability of the Cell Wall Surface to Enzymatic Attack. A surface area of mesopores larger than 4 to 5 nm was reported to be associated with the total surface area available for enzymatic attack, because enzymes cannot penetrate into smaller pores (35). In agreement with this interpretation, some authors calculated the potential available surface for enzymatic attack of 40% of the total cell wall area when pores with a radius higher than 3 nm were considered accessible to the enzyme (12). Therefore, we assumed that the available surface for enzymatic attack is given by pore diameters above 5 nm. After hemicelluloses removal, microporosity severely decreased (Figure 1b) causing an increase of the mesoporosity and, in particular, total mesoporosity increased, on average, by 33.4 ( 15.8% (n ) 5) and by 70.1 ( 21.1% (n ) 5) when pores >5 nm diameter were considered (data calculated from Supporting Information, Table S4). Therefore, the presence of more available surface and the exposure of the crystalline cellulose should increase the accessibility of cellulase enzymes to the crystalline cellulose. This fact was confirmed by the statistically significant higher glucose yields (yield ) mg glucose g-1 cellulose) obtained from the cellulose of SE-treated biomass samples as compared to the untreated samples, when the purchased cellulase enzymes (cellulase from Trichoderma reesei ATCC26921, Sigma Aldrich) were dosed on biomass samples (Figure 2a). Therefore, microporosity and mesoporosity results should explain cellulose accessibility by enzymes and, hence, glucose production, such as confirmed by the good correlation (r ) -0.92, P < 0.001, n ) 10) found for glucose production vs MiS. Unexpectedly, MeS did not correlate with glucose yield (r ) 0.43, P > 0.22, n ) 10), unless only pores potentially accessible to the enzymes, i.e., pores >5 nm diameter, were considered (glucose production vs MeS; r ) 0.76, P < 0.05, n ) 10).

Lignocarbohydrate Complex Contributes to Cell Wall Protection from Enzymatic Attack. Until now, we assumed that hemicelluloses represent the only cellulose protection in the cell wall, but MiS measured on biomass samples treated by the SE indicated that, although a strong reduction of the MiS occurred (Figure 1b), cell wall structures were characterized by high ultramicroporosity (ultramicropores were 85.3% of the total MiS measured). It is well-known that an additional structural protection of the cell wall (37) is represented by the lignin-carbohydrate complex (LCC) (44). The treatment of our biomass samples with concentrated sulfuric acid (H2SO4 13.50 mol L-1) for 24 h at 4 °C allowed one to isolate their LCCs, although the use of strong acid could alter the final products (45). 13C NMR performed on the LCC fractions (Table 2) suggests that they were formed by a mix of lignin, structural proteins, and carbohydrates covalently linked (unhydrolyzable bonds). In addition, the high MiS measured (Figure 1; average of the five biomass samples of 79.13 ( 5.39 m2 g-1 dm, n ) 5; Supporting Information, Table S2) indicated the LCCs to be highly crosslinked. These results agree with those reported in literature (46), which indicated the presence of pores below 0.6 nm were higher in lignin than in the other main constituents of wood and that delignification of wood determines a decrease of the total microporosity of 0.3-0.6 nm. On the basis of these results, we assumed that LCC can operate the protection of the cellulose fibrils by a mechanism similar to that described for the hemicelluloses, i.e., a microporous structure not accessible to enzymes. Measuring the MiS of LCC surfaces in terms of weight of the bulk biomass samples, by taking the weight loss after acid hydrolysis into consideration (Supporting Information, Table S5), we calculated that the MiS accounted for 40.9 ( 7.5% (average of the five samples, calculated from Supporting Information, Table S2) of the total MiS measured on bulk biomass samples Moreover, MiSs were, also, very similar to those detected on biomass samples, after hemicellulose VOL. 45, NO. 3, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1111

removal by SE (data referred to the bulk samples; Figure 1b; Supporting Information, Table S2), indicating that LCCs were not greatly affected by acid hydrolysis. Hence, the role of hemicelluloses in protecting cellulose was attested and the fact that LCC may, also, play an important role in cellulose protection was confirmed. In any case, cellulose protection by LCC should be demonstrated by ascertaining whether cellulose is more hydrolyzed by cellulase after LCC removal, by direct measurement. Since LCC is difficult to eliminate without destroying or deeply altering cellulose fibrils, we measured the contribution of LCC to cellulose protection by testing cellulase activity on commercial microcrystalline cellulose (Sigma 20, S3504, Sigma Co., St Louis, MO, USA). To better simulate plant cellulose without any protection, the purchased cellulose was chosen so that it had a CrICPMAS similar to that of the SE-treated biomass samples (CrICPMAS-cellulose of 53%, to be compared with CrICPMAS-biomass of 49.6 ( 1.64%, the latter calculated as an average of the five biomass samples after SE; Table 1). When cellulase was dosed on purchased cellulose, the glucose yield (yield ) mg glucose g-1 cellulose) was much higher than those obtained from bulk samples but, more interestingly, it was higher, also, than those obtained on SE-treated samples (Figure 2a). These findings indicate that the LCC network contributed effectively to protect cellulose and that the mechanism was the same as that indicated for hemicelluloses, i.e., high MiS that limits enzyme penetration. This was confirmed by the fact that the purchased cellulose exhibited lower MiS (9.35 ( 4.2 m2 g-1 dm; Supporting Information, Table S2) and fractal value D (Table 1), than both bulk and SE-treated samples. Interestingly, when we added data obtained for cellulose (glucose yield, MiS and MeS) to our sample data set, the correlation coefficients found for MiS and MeS (pores

Nanoscale Structure of the Cell Wall Protecting Cellulose from

Recommend Documents