• Open Access

Composition of cell wall phenolics and polysaccharides of the potential bioenergy crop –Miscanthus


V.V. Lozovaya, tel. +1 217 333 9465, e-mail: lozovaya@uiuc.edu


The composition and concentrations of cell wall polysaccharides and phenolic compounds were analyzed in mature stems of several Miscanthus genotypes, in comparison with switchgrass and reed (Arundo donax), and biomass characteristics were correlated with cell wall saccharification efficiency. The highest cellulose content was found in cell walls of M. sinensis‘Grosse Fontaine’ (55%) and in A. donax (47%) and lowest (about 32%) in M. sinensis‘Adagio’. There was little variation in lignin contents across M. sinensis samples (all about 22–24% of cell wall), however, Miscanthus×giganteus (M × g) cell walls contained about 28% lignin, reed – 23% and switchgrass – 26%. The highest ratios of cellulose/lignin and cellulose/xylan were in M. sinensis‘Grosse Fontaine’ across all samples tested. About the same total content of ester-bound phenolics was found in different Miscanthus genotypes (23–27 μg/mg cell wall), while reed cell walls contained 17 μg/mg cell wall and switchgrass contained a lower amount of ester-bound phenolics, about 15 μg/mg cell wall. Coumaric acid was a major phenolic compound ester-bound to cell walls in plants analyzed and the ratio of coumaric acid/ferulic acid varied from 2.1 to 4.3, with the highest ratio being in M × g samples. Concentration of ether-bound hydroxycinnamic acids varied greatly (about two-three-fold) within Miscanthus genotypes and was also the highest in M × g cell walls, but at a concentration lower than ester-bound hydroxycinnamic acids. We identified four different forms of diferulic acid esters bound to Miscanthus cell walls and their concentration and proportion varied in genotypes analyzed with the 5-5-coupled dimer being the predominant type of diferulate in most samples tested. The contents of lignin and ether-bound phenolics in the cell wall were the major determinants of the biomass degradation caused by enzymatic hydrolysis.


New dedicated energy crops are essential to provide lignocellulosic biomass to meet targeted demands for biofuel production in addition to corn grain and stover, which are the current US feedstocks for ethanol production (Barriere et al., 2003). A number of annual and perennial herbaceous crops and trees are currently touted as potential energy crops (Lewandowski et al., 2003; Pauly & Keegstra, 2008; Angelini et al., 2009). Perennial grasses are excellent candidates as biofuel feedstocks and have some advantages over other potential energy dedicated plants due to generally lower costs for establishment with the exception of Miscanthus×giganteus, reduced soil erosion and improved wildlife habitats (McLaughlin et al., 2002; Roth et al., 2005; Karp & Shield, 2008). Miscanthus, a fast growing perennial crop, is currently grown mainly in Europe where it is burned to generate power and has recently been introduced to the US Midwest. The most important Miscanthus species for bioenergy are Miscanthus sacchariflorus, Miscanthus sinensis, and the sterile triploid hybrid generated from a cross between these species, Miscanthus×giganteus (M × g) (Karp & Shield, 2008). M × g is known to yield harvestable biomass of up to 61 metric t ha1 and averaged 30 metric t ha1 across three Midwestern US locations, much greater than the 10 metric t ha−1 observed for a regionally adapted switchgrass cultivar (Heaton et al., 2008). Miscanthus stems senesce in the fall in response to cold temperature, reallocating nitrogen and nutrients to belowground rhizomes (Heaton et al., 2009). As a triploid M × g is sterile and must be clonally propagated from rhizome cuttings to generate mature stands that provide yields which can be maintained for 20 or more years of production (Clifton-Brown et al., 2007). Once established, the stands are so dense that there is no need for weed control. Because of its capacity to reallocate minerals and other nutrients to the rhizomes in the fall, M × g has extremely high nutrient use efficiency and has been shown to produce high levels of biomass over 15 years without the addition of fertilizer (Christian et al., 2008). The dried stems which consist primarily of cellulose, hemicelluloses and lignin are low in nitrogen and mineral content when harvested in the late fall or winter (Lewandowski et al., 2003; Heaton et al., 2009, 2010) and can be used for combustion or biofuel production. The large perennial root mass also leads to large amounts of carbon sequestration (Hansen et al., 2004).

These and other characteristics could make Miscanthus a leading candidate as a bioenergy crop. However, Miscanthus species have not been sufficiently characterized for their renewable potential for fuel production (Jones & Walsh, 2001), (Heaton et al., 2008, and Jessup, 2009). Even though valuable information was published recently describing biomass characteristics of various Miscanthus species grown in different locations (Jones & Walsh, 2001; Heaton et al., 2008; Jessup, 2009 and Hodgson et al., 2010a, b), a better understanding of the composition, structure, synthesis and degradation of cell wall constituents of Miscanthus species for potential bio-energy feedstock utilization is required in order to optimize strategies for crop improvement for processing into liquid biofuels. It is important to determine the cell wall composition that will provide optimal biomass conversion properties for biofuel production and at the same time would not negatively affect plant fitness. Since Miscanthus biomass for processing into biofuel can be harvested in the late fall or winter after allowing it to achieve maximal field drying, stalks must not lodge or experience significant degradation in the field. We analyzed a genotype M.×giganteus (M × g), which is the most widely established and grown accession in Europe. We have also collected samples of genetically variable M. sinensis genotypes of different mature plant heights, switchgrass (cv. ‘Cave-in-Rock’), a tall perennial C4 grass, native to the prairies of North America and of the giant reed species, Arundo donax, a perennial C3 grass that produces more biomass per acre per year than many other known biomass plants (Lewandowski et al., 2003).

To enhance hydrolysis of all lignocellulosic biomass to sugars and fermentation to alcohol cost-effective pretreatment (mechanical, physical, chemical or enzymatic) is required to modify or remove unwanted components, such as lignin and hydroxycinnamic acids that are implicated in cell-wall cross-linking (Iiyama et al., 1990; Ralph et al., 1995) which greatly affects accessibility of structural polysaccharides and their degradation (Grabber et al., 2004). In addition to removal of lignin such pretreatments could reduce cellulose crystallinity and increase the porosity, which can significantly improve the hydrolysis (McMillan, 1994). However, the relationships between cell wall phenolics content, their variable structure, and a variety of cross-linkages between cell wall components and saccharification of lignocellulosic biomass for biofuel production are not yet well understood (Chen & Dixon, 2007). Cell-wall degradability was reported to be largely regulated by lignin concentration, cross-linking, and hydrophobicity but not directly by most variations in lignin composition or structure (Grabber et al., 2004). Analysis of the relationships between lignin content/composition and chemical/enzymatic saccharification was carried out using alfalfa transgenic lines expressing antisense constructs for down-regulating different enzymatic steps in lignin biosynthesis [regulated by cinnamate 4-hydroxylase (C4H); hydroxycinnamoyl CoA: shikimate hydroxycinnamoyl transferase (HCT); coumaroyl shikimate 3-hydroxylase (C3H); caffeoyl CoA 3-O-methyltransferase (CCoAOMT); ferulate 5-hydroxylase (F5H); or caffeic acid 3-O-methyltransferase (COMT), Chen & Dixon, 2007]. The authors concluded that composition indeed has some effect as measured by S/G ration but most likely this effect is smaller than the effect of lignin concentration per se.

Existing information indicates that generation or selection of energy crop genotypes with decreased lignin content and cross-linkage levels in cell walls would result both in higher efficiency of biomass conversion and environmental benefits.

We have examined the cell wall polysaccharide and phenolic composition and concentrations in mature stems of several accessions and species of Miscanthus in comparison with switchgrass and reed, since these characteristics are important for efficient biomass processing to biofuels. This study was designed to provide baseline information on Miscanthus cell wall composition to assist future genetic engineering and breeding programs aimed at production of Miscanthus as an energy crop.

Materials and methods

Plant materials

We analyzed the mature tiller biomass samples of several fast growing grasses including Miscanthus genotypes from the UIUC Miscanthus germplasm collection grown in the UIUC experimental fields on the South Farm during two different seasons 2008 and 2010. Miscanthus germplasm collection consists of about 50 clonal accessions originally purchased from several US landscape nurseries. The M × g clone used in this investigation was originally procured from Chicago Botanical Garden which in turn was procured from Kew Garden.

To provide the homogeneity of field conditions in our experiments, several clones of the same genotype were planted near each other (2 ft between clones and 2 m × 2 m between genotypes). There was no a slope in the field, and the soil type was 154 A – Flanagan silt loam. In plots on the South Farm at the University of Illinois in Urbana-Champaign the Miscanthus genotypes in this study initiated flowering at various times during the summer from mid-June through late July. By early September all genotypes had completed flowering and initiated senescence where amino acids and other nutrients are translocated to the rhizomes to be stored for remobilization the next spring (Mutoh et al., 1968; Beale & Long, 1997). The aboveground biomass of all of the accessions in this study was completely dried up before late October or early November when the first hard freeze typically occurs indicating that senescence was completed. Samples were harvested from fully senesced dried plants in early February 2008 and 2010 at the end of the growing season terminated by a hard freeze and height and mass of plants were measured (Table 1 and Supporting Information for 2010 experiments).

Table 1.   Height and weight of mature plant of perennial crops harvested in Illinois in February, 2008
GenotypeHeight (cm)Weight (g)
  • *

    Mean±standard deviation.

Switchgrass ‘Cave-in Rock’140±5*2.5±0.8
Miscanthus giganteus240±940.8±6.5
(#1) Reed Arundo-donax330±30149.0±16
(#5) M. sinensis‘Grosse Fontaine’150±105.8±0.6
(#8) M.sinensis‘Adagio’120±611.5±1.8
(#13) M. sinensis‘Gracillius Nana’85±63.1±0.5
(#55) M. sinensis‘Silberfeder’165±67.5±2.2
(#57) M.sinensis130±42.7±0.5

All tillers within each biological replicate were excised at soil level with a random sampling and samples were dried at 60 °C until constant weight in both year experiments. Not less than four independent biological replicates per species were analyzed in 2008 and not less than three replicates in 2010. Each independent biological sample/replication consisted of combined biomass of two to four individual plants (two plants – for A. donax) with two additional analytical replications for each biological sample.

Cell wall and hemicelluloses extraction

Dried and milled plant material was homogenized in 80% (v/v) ethanol using Brinkmann Polytron Homogenizer (Kinematica AG, Switzerland, Luzern) at high speed for 2–3 min. Homogenate was heated for 1 h at 80 °C, cooled down to room temperature and centrifuged at 12 000 × g for 30 min. Pellet was resuspended again in 80% ethanol and procedure was repeated. Pellet (AIS) was washed three times with 85% acetone and air dried. Dry AIS was suspended in 0.5% aqueous SDS overnight, the residue was recovered and washed with water by filtration, washed with a 1 : 1 mixture of chloroform and methanol, rinsed with acetone and air-dried. Samples were resuspended in 0.1 m sodium acetate buffer at pH 5.0 and heated 20 min at 80 °C to initiate the removal of starch. After cooling starch was removed by treatment with α-amylase (Type IIA, Sigma-Aldrich, St. Louis, MO, USA) at 37 °C overnight. The extracted cell wall materials (CWM) were washed with water and acetone, dried and used for analyses.

CWM were resuspended in 4 m KOH containing 0.1% NaBH4 and shaken overnight at 130 rpm. Extracted hemicelluloses were separated from non-soluble residues by centrifugation at 12 000 × g for 30 min. Extraction of hemicelluloses from the pellets was repeated an additional two times and all collected supernatants were combined together, neutralized with acetic acid to pH 7, dialyzed against water and freeze-dried.

Analyses of monosaccharide composition and cellulose content

The monosaccharide composition of the hemicelluloses (1 mg) hydrolyzed with 2 m trifluoroacetic acid (TFA) was analyzed by high-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC–PAD) (Dionex, Sunnyvale, CA, USA) on a CarboPac PA20 column with post-column addition of 300 mm NaOH, using following gradient conditions: 0–0.05 min, 12 mm NaOH; 0.05–26 min, 0.65 mm NaOH; 26.1–46 min, 1–300 mm NaOH; 46.1–55 min, 12 mm NaOH. Monosaccharide standards were purchased from Sigma-Aldrich, and included l-Fuc (l-fucose), l-Rha (l-rhamnose), l-Ara (l-arabinose), d-Gal (d-galactose), d-Glc (d-glucose), d-Xyl (d-xylose), d-Man (d-mannose), d-GalA (d-galacturonic acid) and d-GlcA (d-glucuronic acid).

Total carbohydrate content was estimated by phenol-sulfuric assay (Dubois et al., 1956) using 1 mg of dry CWM or hemicellulose fraction and xylose as a standard. Amount of cellulose was estimated by treatment of CWM (10 mg) with acetic-nitric reagent (80% acetic acid: concentrated nitric acid, 10 : 1) (Sloneker, 1971). Undigested pellet was washed several times with water followed by acetone and air-dried. The dried residue was incubated for 1 h with 72% sulfuric acid at 30 °C, then diluted with water to 1 N sulfuric acid and incubated another 1 h at 120 °C. Aliquot (5 μL) of hydrolysate was used for phenol–sulfuric assay with glucose as standard.

Lignin measurement

Acetyl bromide lignin in stem samples was determined according to Fukushima & Hatfield (2004). Five milligrams of cell walls were placed in glass vials and 2 mL 25% acetyl bromide was added. Samples were incubated at 50 °C for 2 h, with occasional mixing. 1 mL reaction mixture after cooling was transferred to 15 mL centrifuge tube containing 2 mL 1 N NaOH, and then 1 mL 1 N hydroxylamine hydrochloride and 4 mL of acetic acid were added and after shaking the volume was made up to 10 mL with acetic acid. Optical density at 280 nm was measured against blank, which contained all reagents except cell wall. The specific absorption coefficient (SAC), 17.78 g−1 cm−1, was determined for purified HCl-dioxane lignin from Miscanthus giganteus according to Fukushima & Hatfield (2001) and lignin concentration was determined using the following equation: % lignin content={absorbance × 100}/SAC × sample conc (g−1) (Sasaki et al., 1996).

Analysis of ester- and ether-bound cell wall phenolics

Ester- and ether-bound cell wall phenolics were extracted and analyzed by HPLC as described by Lozovaya et al. (1999). Ester-bound phenolics: dried cell walls (15–20 mg) were extracted with 1 mL of 1 N NaOH at room temperature under nitrogen for 12 h and centrifuged for 10 min at 5000 g. Supernatant was carefully collected; the pellet was washed with 1 mL water and recollected. The supernatants after each washing were combined and acidified with HCl (conc.) to pH 2.0. Phenolic compounds were extracted with diethyl ether, dried under a nitrogen gas stream, redissolved in 50% methanol, and analyzed by HPLC. Ether-bound phenolics: the pellets obtained after cell wall treatment with 1 N NaOH were carefully transferred to 4 mL vials and 2 mL 4 N NaOH was added to each vial and sealed with PTFE caps. Vials were heated at 170 °C for 3 h in the oven. The reaction mixture after cooling was acidified by concentrated HCl to pH 2.0 (Lozovaya et al., 1999). Phenolic acids and aldehydes were extracted and analyzed as described for ester-bound phenolics. Separation of phenolics was achieved with a Waters 2690 Separation Module (Waters Corp., Milford, MA, USA) using a 250 mm × 4.6 mm Prevail C18, 5 μm, Column (Alltech Assoc., Deerfield, IL, USA). A linear gradient composed of water (pH 2.8 adjusted with acetic acid) and acetonitrile was used. Following injection of 10 μL of sample, acetonitrile was increased from 5% to 15% v/v over 10 min and then increased to 20% over 30 min and than to 60% over 40 min. The solvent flow was 1 mL min−1. A Waters 996 photodiode array (PDA) detector was used.

Identification of phenolic acids and aldehydes was carried out by comparison of retention times and UV spectra (PDA detector, Waters Corp.) of the eluting peaks and authentic standard compounds (p-hydroxybenzoic, vanillic acid, syringic acid, ferulic acid, coumaric acid, and p-hydroxybenzaldehyde, vanillin and syringaldehyde) that were purchased from Sigma (Sigma-Aldrich Corp., St. Louis, MO, USA). Diferulic acids were identified on the basis of their spectroscopic properties and molecular masses determined with LC-MS as we described earlier (Lozovaya et al., 2006). Phenolics were quantified as individual peak area units (converted to μm) based on the appropriate integration response factors.

Nitrobenzene oxidation

Pellets obtained after cell wall treatment with 1 N NaOH were transferred to 4 mL glass vials with 2 mL 2 N NaOH and 200 μL of nitrobenzene (Kajita et al., 1996). Vials were sealed with PTFE caps and heated at 170 °C in oven for 3 h. After cooling the samples, the excess nitrobenzene was extracted three times with 1 mL ether and residue was acidified by concentrated HCl to pH 2.0 and phenolics were extracted as described above for ester-bound phenolics.

Enzymatic hydrolysis of the cell walls

A portion of the plant material was treated with 1.3% sulfuric acid at 130 °C for 40 min (Chen & Dixon, 2007). Enzymatic hydrolysis of treated and acid untreated materials was done using a mixture of equal parts of cellulase from Trichoderma reesei and cellobiase from Aspergillus niger (Novozyme 188) purchased from Sigma-Aldrich Corp. (St. Louis, MO, USA) according to the procedure of National Renewable Energy Laboratory (LAP-009) and Ghose (1987).

Three independent measurements of biomass enzymatic degradability with and without acid pretreatments were performed using three replicates in each case. The patterns of genotypic variations were similar in these experiments and the combined results of these experiments are presented.

Statistical analysis

The grass cell wall component and saccharification data were analyzed statistically by one way analysis of variance (anova) with P<0.05, using the xlstat-pro (v.7.5.2) program (Addinsft, NY, USA) with at least four independent biological replicates per species used in 2008 experiment and three replicates in 2010. Regression analysis was performed using the statistics program built into microsoft excel. Bars with common letters were not significantly different using least significant difference (P=0.05).

Results and discussion

Main biomass components

We have examined the cell wall phenolics and polysaccharides in mature tillers of switchgrass (Cave-in-Rock), reed (A. Donax) and several Miscanthus genotypes (Miscanthus×giganteus– M × g; and M. sinensis: Grosse Fontaine – Ms#5; Adagio – Ms#8; Gracillius Nana-Ms#13; Silberfeder-MS#55 and Ms#57), that differ in their growth performance and biomass, all grown in the University of Illinois experimental fields (Table 1). The reed stem dry biomass contained less cell walls (80%) than switchgrass and all Miscanthus samples tested (varied at the range of ∼85–89%) (Fig. 1a). The cellulose content in cell walls varied greatly among these plants, with the highest content found in M. sinensis‘Grosse Fontaine’ (MS#5) (55%) and the lowest (32%) in M. sinensis‘Adagio’ (MS#8, Fig. 1c). There was no marked variation in lignin contents across M. sinensis samples (all about 22–24% of cell wall), however, M × g cell walls contained the highest lignin concentration (about 28%) (Fig. 1b). Therefore, differences in the ratio of cellulose/lignin in M. sinensis samples reflected mainly the differences in cellulose content (Fig. 2), being the highest in ‘Grosse Fontaine’ (MS#5) (2.4) across all samples tested. The high ratio of cellulose/lignin (2.1) was also found in A. donax, containing 47% cellulose in cell walls.

Figure 1.

 Biomass cell wall (CW) characteristics: cell wall concentration in dried mature tillers – (a); acetyl bromide lignin concentrations in cell walls – (b); cellulose and xylan concentrations in cell walls of several perennial grasses – (c). Cellulose was estimated by treatment of CWM with acetic-nitric reagent followed by phenol-sulfuric assay with glucose as a standard. Xylan was calculated using data from monosaccharide composition of hemicelluloses shown in Fig. 3. Bars with common letters were not significantly different using least significant difference (P=0.05). DW – dry weight.

Figure 2.

 Cellulose/lignin and cellulose/xylan ratios in cell walls of several perennial grasses. Bars with common letters were not significantly different using least significant difference (P=0.05).

Interestingly, it was just recently reported that in winter harvest M × g had the highest lignin content across fifteen Miscanthus genotypes analyzed by NIRS calibrated for acid detergent lignin (ADL) (Hodgson et al., 2010a, b) which is in a good agreement with our data, although our lignin content is higher than found in the publication cited. In that study M × g biomass contained also the highest concentration of cellulose, however, in our work the highest cellulose content was found in M. sinensis‘Grosse Fontaine’.

Very limited published information is currently available describing Miscanthus lignocellulosic biomass. The literature suggests that lignin content accounts for 18.30–20.99 and holocellulose for 69.78–78.63% of M. sacchariflorus dry matter according to studies summarized by Visser & Pignatelli (2001), while it was reported that dried biomass of M×g consists of 38% cellulose, 24% hemicellulose and 25% Klason lignin (de Vrije et al., 2002). Analysis of a genotype M. ogiformis (analogous species to M × g) showed that it contained 41.9% cellulose, 26.6% hemicellulose and only 13.3% lignin determined as acid lignin fiber (Magid et al., 2004). When ADL, acid detergent fibre, and neutral detergent fibre were measured in fifteen Miscanthus genotypes grown in five European locations using chemometric techniques combining near infrared reflectance spectroscopy (NIRS) and conventional chemical analyses the values obtained ranged from 7.6–11.5%, 41.2–52.9%, to 23.5–33.8% of dry biomass in the winter harvest, respectively (Hodgson et al., 2010a, b).

Such discrepancy in reported lignin contents could be due to the fact that various analytical procedures currently used for quantifying lignin in cell walls do not yield consistent results, and methods including acid treatments can solubilize some of the lignin, especially in grasses, which leads to underestimation of lignin concentrations (Lowry et al., 1994; Fukushima & Hatfield, 2004). The method used here was suggested to be used for prediction of biomass digestibility based on lignin levels (Fukushima & Hatfield, 2004) and it was shown that the levels of lignin determined by this method are better correlated with in vitro dry matter digestibility and in vitro cell wall digestibility than lignin contents determined by other methods.

Composition of hemicellulosic polymers

Monosaccharide composition of cell walls from these plants was typical for the grasses, where arabinoxylans are the predominant hemicellulosic polymers at the mature stage, although some variation in the amounts of xylose and glucose were observed (Fig. 3) with the highest content of xylose being present in A. donax, and lowest – in M. sinensis‘Grosse Fontaine’ (Ms#5), ‘Adagio’ (Ms#8) and ‘Silberfeder’ (Ms#55). Considering, that xylose is a main constituent of arabinoxylan, which is more prevalent in grasses and all types of plant secondary walls, we can assume that amount of xylose reflects the amount of arabinoxylan and therefore can estimate its amount in cell walls from analyzed plants (Fig. 1c). Xylans are tightly bound to cellulose microfibrils via hydrogen bonds, thereby stabilizing hemicellulose/cellulose networks. Calculation of the ratios of cellulose/xylan demonstrated that in M. sinensis‘Grosse Fontaine’ (Ms#5) plants this ratio is the highest, while in other samples it is much lower (Fig. 2).

Figure 3.

 Monosaccharide composition of hemicelluloses in cell walls of several perennial grasses. analyzed by HPAEC-PAD. Results are expressed in mol% and represent two independent extractions of 4M KOH fraction from two separately grown plants. Monosaccharides were identified according to retention times for standard monosaccharides (Sigma). Fuc, fucose; Ara, arabinose; Gal, galactose; Glc, glucose; Xyl, xylose; GalA, galacturonic acid; GluA, glucuronic acid. (Vertical bars represent the SD).

We did not observe significant differences in the amount of arabinose and glucuronic acid (Fig. 3), the monosaccharides which are constituents of the side chains of arabinoxylan and contribute to its cross-linking within cell wall. However, the ratios of Ara/Xyl were different for different plants (Fig. 4a), indicating that there is a variation in the level of xylan substitution with arabinose. The M. sinensis samples # 8, 55 and 57 had the highest and A. donax and M × g had the lowest Ara/Xyl ratios (Fig. 4a).

Figure 4.

 Arabinose/xylose (a) and arabinose/ferulic acid (b) ratios in cell walls of several perennial grasses. Amounts of arabinose and xylose were estimated from data of monosaccharide composition of hemicelluloses presented in Fig. 3. Bars with common letters were not significantly different using least significant difference (P=0.05).

The arabinose side chains of xylans in grasses often bear a feruloyl ester, which leads to the oxidative cross-linking of xylan chains through formation of diferulate bridges between arabinoxylans. Ferulate also links arabinoxylan to lignin via ester or ether bonds in grass cell walls. Ara/FerA ratios indicates the degree of arabinose substitution by ferulates in xylan which also varied in our samples (Fig. 4b) and was the highest in M. sinensis#5 and the lowest in samples of switchgrass (Fig. 4b). Such structural differences could affect formation of lignin/arabinoxylan cross-links and therefore, cell wall degradability.

Cell wall bound hydroxycinnamic acids

Substantial differences were found in contents and composition of ether- and ester-bound hydroxycinnamic acids across genotypes tested (Figs 5–7). About the same total content of ester-bound phenolics was found in different Miscanthus genotypes (23–27 μg/mg cell wall), while reed contained 17 μg/mg cell wall and switchgrass about 15 μg/mg cell wall. Coumaric acid was a major ester-bound phenolic compound in the cell walls of the plants analyzed (Fig. 5), and the ratio of coumaric acid/ferulic acid varied from 2.1 to 4.3 in plant materials tested here, with the highest ratio being in M × g samples (Fig. 7). Concentration of ether-bound hydroxycinnamic acids varied greatly within Miscanthus genotypes and was also the highest, but lower than concentration of ester-bound hydroxycinnamic acids in M × g cell walls, although about two-three-fold that found in samples Ms#8 and Ms#55 (Fig. 6). Cell walls of switchgrass and reed also had high content of ether-bound hydroxycinnamic acids (Fig. 6). In addition to coumaric and ferulic acids the ether bound fraction of our samples contained also the derivatives of sinapic acid which accounted for about 30–40% of the fraction. We identified four different forms of diferulic acid ester-bound to Miscanthus cell walls and their concentration and proportion varied in genotypes analyzed (Fig. 8), which probably also could affect accessibility of polysaccharides to enzymes during biomass saccharification. The 5-5-coupled dimer was a predominant type of diferulate in most samples tested, except M × g and, Ms#8 where about equal proportions of the 5-5 and 8-O-4′ bonds were found (Fig. 8). This is different from results published for maize cell walls where about half of cell wall ferulates were found to be able to form mainly 8-coupled dehydrodimers with small amounts of 5-5-coupled dimers (Grabber et al., 1995, 2000). Interestingly, alkaline treatment of bran cell walls released 5-5/8-O-4 dehydrotriferulates (Bunzel et al., 2003; Rouau et al., 2003), which could probably cross-link three arabinoxylan chains. Ferulate deposition takes place during formation of both primary and thickened secondary cell walls in grasses (MacAdam & Grabber, 2002; Jung, 2003), resulting in a heavily cross-linked polymeric structure (Grabber et al., 2004). Ferulate molecules could also provide initiation/nucleation sites for the lignin polymerization and deposition (Ralph et al., 1995, 1998). The p-coumarate in grass cell walls is mainly esterified to the lignin S units (Lu & Ralph, 1999) and its accumulation serves as an indicator of lignification, although the role of p-coumarate in wall development and degradability is less studied and understood.

Figure 5.

 Content and composition of cell wall (CW) ester-bound phenolics (VA, vanilic acid; SirA, syringic acid; pBAl, p-hydroxy benzealdehyde; VN, vanillin; SAL, syringaldehyde; CoumA, coumaric acid; FerA, ferulic acid; SinA, sinapic acid). (Vertical bars represent the SD).

Figure 6.

 Content and composition of cell wall (CW) ether-bound phenolics expressed as a sum of etherified phenolic acids and their derivatives: p-coumaric acid (CoumA), ferulic acid (FerA) and sinapic acid (SinA). Vertical bars represent the SD.

Figure 7.

 Ratios of cell wall ester-bound coumaric acid (CoumA) to ferulic acid (FerA). Bars with common letters were not significantly different using least significant difference (P=0.05).

Figure 8.

 Content and composition of cell wall (CW) ester-bound diferulates (DiFA). (Vertical bars represent the SD).

G units (ferulate and derivatives) are found in the cell wall bound phenolic fraction of all samples (varying from 53% to 64%) when nitrobenzene oxidation was carried out using starch free alcohol insoluble residue after removal of ester-bound phenolics by 1 m NaOH treatment, with the smallest proportion found for H units (p-coumarate and derivatives; 8–13% of fraction) (Fig. 9). This fraction consists of lignin and ether-bound hydroxycinnamic acids. Our results are in a good agreement with studies of linkages between hydroxycinnamic acids and lignins in cell walls of several grasses (Sun et al., 2002). They demonstrated that more than half of ferulic acid was associated with lignin fractions through ether bonds (51.6–68.3%) and to lesser extent through ester bonds (31.7–48.4%), however, the bulk of p-coumaric acid was ester bound to lignin (67.0–83.5%).

Figure 9.

 Composition of cell wall fraction included lignin and ether-bound phenolics (H, hydroxyl-benzoyl; G, guaiacyl; and S, sinapyl units). (Vertical bars represent the SD).

We present here the results for biomass samples harvested in February 2008. The same trends were found when samples of the same accessions and species were harvested in 2010 and analyzed for cell wall constituents (results are presented as Supporting Information) even though some differences in concentrations of these compounds were shown, which is expected as a result of different climatic conditions.

Biomass saccharification

Since the variations in cell wall constituents between different perennial grasses reported here can have an influence on the wall mechanical properties, extensibility and biodegradability (Jung & Deetz, 1993), we carried out saccharification of selected plant biomass that had different lignin/structural carbohydrate contents and composition using the laboratory analytical procedure of the National Renewable Energy Laboratory (LAP-009) as described by Chen & Dixon (2007). Figure 10 shows the differences in the efficiency of cell wall enzymatic hydrolysis without (a) and after (b) acid pretreatment across different samples. The percentage of sugar released from the biomass varied between 31.3% (M × g) and 36.3% (Ms#57) in 2008 samples (Fig. 10a) when enzymatic hydrolysis was carried out without acid pretreatment. However, when samples were pretreated with sulfuric acid, the degradability efficiency was increased with between 37% (M × g) and 51% (Ms#5 and 55, Fig. 10b) of the biomass being converted to sugars and the efficiency did not differ very much within the M. sinensis genotypes tested. The lowest degradation efficiency was found for M × g samples of both 2008 and 2010 harvests (Fig. 10 and Supporting Information).

Figure 10.

 Sugars (as glucose equivalent) released from grass cell walls (CW) by enzymatic hydrolysis without (a) and with (b) acid pretreatment. Bars with common letters were not significantly different using least significant difference (P=0.05).

There was a negative correlation between the lignin and ether bound phenolic contents and sugar released by both enzymatic hydrolysis alone and that after acid pretreatment (Fig. 11a–d). When a similar analysis was done with cellulose and ester bound phenolics (data not shown), there was no clear correlation with the saccharification efficiency. M × g that had the lowest efficiency, also had an increased concentration of lignin and ether bound phenolics in cell wall compared with other biomass tested (Fig. 1c), so that lignin content and cross-linking via ether bonds may be the major factor determining biomass degradation both with and without acid pretreatment. Recalcitrance to enzymatic saccharification with and without acid pretreatment was also directly proportional to lignin concentrations in stem cell walls of several transgenic alfalfa lines with modified lignin contents and composition achieved by downregulation of six lignin biosynthetic enzymes (Chen & Dixon, 2007). These studies with transgenic alfalfa also showed that for untreated stem samples, not only lignin content, but also its composition could affect efficiency of enzymatic hydrolysis. In our experiments the untreated Ms#5 sample, showing the highest enzymatic sugar release, had the highest cellulose/lignin and cellulose/xylan ratios, while M × g with decreased cell-wall degradability had the highest ether-bound phenolic concentration in addition to high lignin content and high coumaric acid/ferulic acid ratio. There was no clear correlation between saccharification and the ester bound diferulate contents and composition and also lignin composition in our grass samples.

Figure 11.

 Relationships between the cell wall (CW) component contents and saccharification of grass biomass. Each point represents an individual accession/species. Stem material was treated with cellulase and cellobiase for 72 h. (a–d) Total sugar released is presented as a function of either lignin content of untreated stems (a), lignin content of acid pretreated stems (b), content of ether-bound phenolics of untreated stems (c), content of ether-bound phenolics of acid pretreated stems (d).


These studies have characterized the cell wall compositions of a number of potential biomass plants and showed substantial difference in contents of main wall constituents.

The cellulose content varied between 32% and 55% while the lignin content varied at the range of 22–28% and xylan at the range of 13–30% of cell walls in plants compared. The highest ratios of cellulose/lignin and cellulose/xylan were in M. sinensis‘Grosse Fontaine’ across all samples tested.

The total content and composition of ester- and ether bound phenolics also differed in cell walls of plant tested with coumaric acid being a major hydroxycinnamic acid bound to cell walls and with the 5-5-coupled dimer being the predominant type of diferulate in most samples tested.

The differences in amounts and composition of arabinoxylans and lignin in the cell walls determine their rigidity and should affect the degradability of biomass. The cell wall hydroxycinnamic acids could also determine recalcitrance of lignocellulosic feedstocks to chemical or enzymatic saccharification during processing to biofuel; for example, it was found that ferulate or diferulate linkages between lignin and arabinoxylan negatively affected enzymatic polysaccharide degradation (Grabber et al., 1995, 1998a, b). It was also reported that the forage cell wall enzymatic or microbial degradability negatively correlated with both lignins and p-hydroxycinnamic acids (Jung & Deetz, 1993; Besle et al., 1994). In addition, xylans are composed of pentasaccharides, which are not suitable for fermentation by the currently existing yeast strains. Therefore, a reduced proportion of xylan within polysaccharide fraction in the cell walls of grasses used for bioethanol production could be a desirable feature.

We found that lignin content and cellulose/lignin and cellulose/xylan ratios could be the major determinants of Miscanthus biomass degradability. Further studies are needed to clarify the effects of genetic variation of cell wall composition, ester and ether cross-linking on biomass saccharification during processing to biofuel to define the set of criteria for genetic engineers and breeders to use for the improvement of Miscanthus biomass characteristics. Importantly, the data presented shows that there is considerable genetic variability across Miscanthus genotypes which can be used for breeding or engineering of varieties in the future which can maximize both the biomass production and degradability to be able to produce the most biofuel per unit land area possible.


This study was supported in part by funds from the Plant Science Institute, Iowa State University, 2008–2010, the Consortium for Plant Biotechnology Research Inc., Illinois Council on Food and Agricultural Research (C-FAR) and the USDA Cooperative State Research, Education and Extension Service, Hatch project numbers 802-352 and #802-309.