Bioresource Technology 103 (2012) 201–208

Contents lists available at SciVerse ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /bior tech

The lignin present in steam pretreated softwood binds enzymes and limitscellulose accessibility

Linoj Kumar, Valdeir Arantes, Richard Chandra, Jack Saddler ⇑Forest Products Biotechnology/Bio-energy Research Group, Forest Sciences Centre, University of British Columbia, 2424-Main Mall, Vancouver, BC, Canada, V6T1Z4

a r t i c l e i n f o

Article history:Received 25 May 2011Accepted 22 September 2011Available online 8 October 2011

Keywords:LigninUnproductive bindingCellulose accessibilitySteam pretreatment of softwoodsEnzymatic hydrolysis

0960-8524/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.biortech.2011.09.091

Abbreviations: BSA, bovine Serum Albumin; CBMule; CBU, cellobiase Units; FPU, filter Paper Unit; SO2,Pretreated Douglas-fir; SPDF, steam Pretreated Dpretreated Douglas-fir that was completely delignifisteam pretreated Douglas-fir that was partially decontent; SPDF-S41, steam pretreated Douglas-fir thatRetention Value.⇑ Corresponding author.

E-mail address: jack.saddler@ubc.ca (J. Saddler).

a b s t r a c t

The influence of cellulose accessibility and protein loading on the efficiency of enzymatic hydrolysis ofsteam pretreated Douglas-fir was assessed. It was apparent that the lignin component significantly influ-ences the swelling/accessibility of cellulose as at low protein loadings (5 FPU/g cellulose), only 16% of thecellulose present in the steam pretreated softwood was hydrolyzed while almost complete hydrolysiswas achieved with the delignified substrate. When lignin (isolated from steam pretreated Douglas-fir)was added back in the same proportions it was originally found to the highly accessible and swollen, del-ignified steam pretreated softwood and to a cellulose control such as Avicel, the hydrolysis yieldsdecreased by 9 and 46%, respectively. However, when higher enzyme loadings were employed, thegreater availability of the enzyme could overcome the limitations imposed by both the lignin’s restric-tions on cellulose accessibility and direct binding of the enzymes, resulting in a near complete hydrolysisof the cellulose.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction uble fraction while increasing the enzyme accessibility of the

Progress continues to be made in the bioconversion of lignocel-lulosic substrates to ethanol with the hope that it can, at leastpartially, provide a sustainable alternative to fossil derived trans-portation fuels such as gasoline (Solomon and Johnson, 2009;World Watch Institute, 2007). Softwoods have been shown to beone possible feedstock for an enzyme-based biomass-to-ethanolprocess, particularly in the Nordic countries, the Pacific Northwestand areas such as Russia where the taiga/boreal forest predomi-nates (Mabee et al., 2006; FAO, 2010). Compared to hardwoodsand agricultural crops, softwoods have an advantage in terms oftheir overall high hexose content, with these sugars being readilyfermented by conventional yeasts such as Saccharomyces cerevisiae(Matsushika et al., 2009). In the overall enzyme-based softwood-to-ethanol process, some form of pretreatment is a pre-requisiteso that the complex structure of lignocellulosic matrix can be dis-rupted. The pretreatment process enhances the fractionation andrecovers most of the hemicellulose derived sugars in the water sol-

ll rights reserved.

, carbohydrate Binding Mod-sulphur dioxide, SPDF, Steamouglas-fir; SPDF-CD, Steamed; SPDF-D38/D30/D22/D11,lignified to different ligninwas sulfonated; WRV, water

water insoluble cellulosic component. Among the many pretreat-ment processes that are currently advocated by various groups,steam pretreatment has been shown to be an effective pretreat-ment option from both a technical and economic perspective(Hendriks and Zeeman, 2009; Holtzapple et al., 1989). Several ofthe emerging pilot scale and demonstration bioconversion plantsare using variations of steam pretreatment due to the various tech-no-economic benefits of this process (IEA Task 39, 2011).

Recent work has shown that, among a range of lignocellulosicsubstrates pretreated at near optimal conditions so that good,overall sugar recovery can be achieved (i.e. both hemicelluloseand cellulose derived sugars), steam pretreated softwood was themost recalcitrant to enzymatic hydrolysis. Typically, high proteinloadings (�60 mg) were required before even 70% cellulose hydro-lysis could be achieved (Arantes and Saddler, 2011). When low en-zyme loadings (�5 FPU/g cellulose) were used to hydrolyze a rangeof softwoods, steam pretreated at near optimal conditions, poorhydrolysis yields (<21%) were obtained (Kumar et al., 2010). Evenwhen a relatively high pretreatment severity, which reduced thehemicelluloses recovery, was used, hydrolysis yields were still dis-appointingly low (Kumar et al., 2011). Although adequate hydroly-sis yields could be obtained at high enzyme loadings, it is clear thatthe economic feasibility of any lignocellulosic based ethanol pro-cess will require the use of low protein loadings (Donghai andJunshe, 2007; Gregg and Saddler, 1996; Merino and Cherry,2007). To enhance enzymatic hydrolysis of steam pretreated soft-woods at low enzyme input, it is important to increase our under-

202 L. Kumar et al. / Bioresource Technology 103 (2012) 201–208

standing of the prominent factors behind the poor hydrolysisyields obtained when lower enzyme loadings are employed.

Among the various substrate characteristics which influenceenzymatic hydrolysis, limited cellulose accessibility was suggestedto be the major factor that restricts the effectiveness of cellulosehydrolysis (Jeoh et al., 2007; Mooney et al., 1997; Wong et al.,1988; Zhang et al., 2006; Arantes and Saddler, 2010). In additionto characteristics within the cellulose itself (i.e. crystallinity, de-gree of polymerization, etc.), hemicellulose and lignin limit theaccessibility of the cellulase enzymes to cellulose (Jeoh et al.,2007; Varnai et al., 2010). Previous work has shown that SO2-cat-alyzed steam pretreatment of softwoods at severities in the rangeof 3.3 6 Log R0 6 3.6 resulted in good recovery of the hemicellulo-sic sugars in the water soluble fraction while the cellulose and lig-nin rich water insoluble fraction could be readily hydrolyzed if ahigh enough cellulase loading was used (Kumar et al., 2010).

It is known that there are various ways by which lignin limitsthe accessibility of cellulose in a steam pretreated softwood sub-strate. Lignin restricts the swelling of cellulose and therefore limitsthe accessible surface area of the cellulose available to the enzymes(Jeoh et al., 2007; Mooney et al., 1997, 1998; Wong et al., 1988;Zhang et al., 2006; Eriksson et al., 1991; Nakatani et al., 2008).Swelling of cellulose, resulting in an increase in the accessible sur-face area, is essential to achieving efficient cellulose hydrolysis(Arantes and Saddler, 2010). A decrease in molecular weight of lig-nin during pretreatment, together with its hydrophobicity and thesurface tension effects of water, result in the deposition of smallspherical lignin droplets on the fiber surface, likely limiting accessof the enzymes to the cellulose (Donohoe et al., 2008; Donaldsonet al., 1988). In addition, lignin also reduces the effectiveness ofenzymatic hydrolysis by unproductively adsorbing cellulases,thereby reducing the availability of the enzymes (Nakagameet al., 2010; Berlin et al., 2005, 2006; Tu et al., 2007). The relativecontribution of these two main mechanisms of lignin inhibition(limiting cellulose accessibility and unproductive binding of theenzymes) on a typical steam pretreated softwood substrate is notyet studied.

The role that lignin plays is also demonstrated by recent workwhich showed that partial or complete substrate delignification re-duced the enzyme requirement to attain reasonable hydrolysisyields by more than an order of magnitude (i.e. five to ten-fold)(Kumar et al., 2011; Yang et al., 2002). Similar improvements werealso observed when the cellulose component of the steam pre-treated substrate was made more accessible by treatments suchas neutral sulfonation which, instead of removing lignin, increasedits hydrophilicity, and promoted fiber swelling by incorporatingacid groups onto the lignin (Mooney et al., 1997; Kumar et al.,2011). However, sulfonation also resulted in a lignin with lessunproductive binding sites (Nakagame et al., 2010). It is apparentthat the relative contribution of the different roles of lignin andthe mechanism by which lignin primarily restrict enzymatichydrolysis is not yet fully understood. Further clarification on therole that lignin plays in pre- and post-treated biomass substratesshould help in optimizing pretreatment and/or subsequent posttreatment strategies to enhance the effectiveness of enzymatichydrolysis so as to obtain high sugar yields at low protein loadings.

Although, it is recognized that lignin restricts enzymatic hydro-lysis by both limiting cellulose accessibility and unproductivelybinding cellulases, previous studies have not tried to quantify therelative roles that lignin might play in either binding the enzymeor limiting the swelling of the cellulose and thus restrictingenzyme access. To evaluate the relative impact that lignin has onlimiting cellulose accessibility vs. unproductive binding of theenzymes, steam pretreated Douglas-fir was selectively delignifiedand the minimum enzyme dosage required to achieve a near com-plete conversion of the resulting pure cellulose (SPDF-CD) was

determined. To assess the extent that unproductive binding mightplay, BSA was used as a surrogate for cellulases and used as a probeto help assess how much the unproductive binding effects of lignincontribute to the generally poor hydrolysis yields obtained withhigh lignin containing substrates (Kawamoto et al., 1992; Palonenet al., 2004; Pan et al., 2005; Yang and Wyman, 2006; Rollin et al.,2011). By being able to assess the influence of one possible mech-anism of lignin inhibition when compared to the other we hoped todetermine what extent enzyme binding of lignin has on the en-zyme loading required. We also hoped to determine if hydrolysisyields are dependent on the overall accessibly of the cellulose tothe enzymes as, at very low enzyme loadings, lignin binding is stilllikely to be problematic, even with highly accessible cellulosicsubstrates.

2. Methods

2.1. Preparation of the wood samples for the pretreatment

Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) logs (approx-imately 9 logs) were collected from 27 year old trees found in thecoastal regions of British Columbia. The logs were debarked, split,chipped and screened to approximately 2 � 2 � 0.5 cm3. The mois-ture content of the wood chips was determined by drying a knownamount of chips in the oven at 101 �C overnight and measuring theweight loss. The moisture content was 10.2 ± 0.9% (measured intriplicate). The chemical composition of the wood chips is givenin the Table 1.

2.2. Pretreatment

Prior to steam pretreatment, wood chips containing 300 g dryweight in sealable plastic bags were injected with �12 g of SO2

(4% w/w of the substrate). Once the desired amount of SO2 enteredthe bags, the bags were immediately sealed and left for approxi-mately 12 h before steam pretreatment was carried out (Buraet al., 2009). Steam pretreatment was conducted in a 2-L StakeTech(Stake Tech., Ontario, Canada) steam gun at 200 �C for 5 min. Afterthe pretreatment, the entire slurry was removed and the water sol-uble and insoluble fractions were separated by vacuum filtrationusing a Buchner funnel. The water insoluble fraction was washedwith a volume of water equal to 10 times the dry weight of thestarting substrates and vacuum filtered. Final moisture content ofthe substrate was 77%.

2.3. Complete and partial delignification treatments

The complete delignification of the steam pretreated sampleswas conducted at room temperature according to the procedureby PAPTAC useful methods G10, which was the standardized delig-nification technique for subsequent DP determination of isolatedcelluloses (Pulp and Paper Technical Association of Canada (PAP-TAC), 1998). Sodium chlorite (0.8 g) and 0.1 ml glacial acetic acidwas added per g of steam pretreated substrate with a liquor towood ratio of 15:1. After an initial gentle mixing, the reaction mix-ture was kept in the fume hood for 3 h at room temperature, theslurry was filtered using a Buchner funnel and the water insolublefraction was washed extensively with water. This reaction and thewashing were carried out multiple times until the substrate wascompletely delignified. Substrates with various lignin contentswere prepared by using the same conditions of complete delignifi-cation, but changing the dosages of the sodium chlorite with therespect to the weight of the pulp. Briefly, 100 g of the steam pre-treated substrate (on a dry weight basis) was treated in a 5-L Erlen-meyer’s flask. Sodium chlorite (0.15–0.8 g) and 0.1 ml glacial acetic

Table 1Chemical composition of untreated Douglas-wood chips (% dry weight).

Arabinan Galactan Glucan Xylan Mannan Acid Insoluble Lignin Acid Soluble Lignin Acetone Extractives

% dry weight 1.2 (0.0)* 2.2 (0.0) 47.3 (0.4) 4.4 (0.1) 11.7 (0.1) 29.8 (0.8) 0.5 (0.1) 1.1 (0.4)

* Numbers in the brackets represent standard deviations from the average.

L. Kumar et al. / Bioresource Technology 103 (2012) 201–208 203

acid was added per g of steam pretreated Douglas-fir on a dryweight basis to obtain different degrees of delignification. Theliquor to wood ratio was 15:l. After an initial gentle mixing, thereaction mixture was kept in the fume hood for 3 h at roomtemperature. After the reaction, the slurry was filtered using aBuchner funnel and the water insoluble fraction was washedextensively with water (�5 times higher volume of water com-pared to the original reaction volume).

2.4. Neutral sulfonation

Neutral sulfonation treatment was carried out with minor mod-ification of the conditions used by Mooney et al. (1998). Briefly, thetreatments were conducted in a Parr reactor at 10% (w/w) consis-tency with the addition of 0.16 g Na2SO3/g dry weight of the sub-strate at a temperature of 160 �C for one hour, with a prior20 min to allow the fibers to come to temperature. After the addi-tion of the chemicals and the substrate, the final pH was adjustedto 7 at room temperature using 4 M H2SO4. The 16% (w/w) total al-kali charge is the chemical loading normally employed during sul-fite pulping at the commercial scale (Biermann, 1996). The treatedslurry was separated into solid and liquid fractions by filtering itthrough a P8 filter paper using a Buchner funnel and the resultingwet solid fraction (�70% (w/w) moisture content) was extensivelywashed with water.

2.5. Enzymatic hydrolysis

The water insoluble fractions after pretreatment/post treatmentwere enzymatically hydrolyzed in acetate buffer (50 mM, pH 4.8)at 50 �C with stirring in an orbital shaker at 150 rpm. The enzy-matic hydrolysis was conducted at 2% substrate consistency in a60-g reaction mixture in a 125-mL capacity Erlenmeyer’s screwtop flask. Filter paper activity of cellulase (Celluclast 1.5L, Novo-zymes, Bagsvaerd, Denmark) was 60.4 FPU/ml while the cellobiaseactivity of b-glucosidase (Novozymes 188, Bagsvaerd, Denmark)was 360 CBU/ml. The enzyme loading was 5 FPU cellulase and10 CBU beta glucosidase/g cellulose of the substrates. Tetracycline(40 lg/ml) and cycloheximide (30 lg/ml) were added to inhibitmicrobial contamination during the hydrolysis process (Seliget al., 2008). Sampling of 500 ll of supernatant was done at 12,24, 48 and 72 h from the reaction mixture. Enzymes were inacti-vated by incubating the aliquots of the reaction mixture on a hotplate at 100 �C for 5 min and subsequently stored at �20 �C untilsugar analysis was performed. For BSA blocking treatment, appro-priate amounts of BSA were added to substrates and incubated at50 �C for 60 min prior to the addition of enzymes.

2.6. Analytical methods

All analyzes were done in triplicate. The substrates were ana-lyzed for acid insoluble lignin and carbohydrates according to Tap-pi T222 om-88 (TAPPI, 1994a). The hydrolysate from this analysiswas retained and analyzed for soluble lignin by reading the absor-bance at 205 nm and using an absorptivity constant of 110 l/g-cm(Dence, 1992). Sugars were measured on a Dionex (Sunnyvale, CA)HPLC (ICS- 3000) equipped with an AS 50 auto sampler, ED50 elec-trochemical detector, GP 50 gradient pump and anion exchange

column (Dionex CarboPac PA1). Acetone soluble extractives wereestimated using Tappi method T204 om-88 with the followingmodifications (TAPPI, 1994b). Briefly, 10 g of air-dried samplewas extracted for 8 h with acetone with six cycles/h. The acetonein the round-bottomed flask was evaporated in the fume hoodand dried in an oven at 100 �C overnight, to determine the weightof extractives present in the sample flasks.

2.7. Isolation of lignin

Enzymatic residual lignin was isolated according to methods re-ported previously (Berlin et al., 2006). Cellulase (Celluclast 1.5L,Novozymes, Bagsvaerd, Denmark) and beta-glucosidase (Novo-zymes 188, Bagsvaerd, Denmark) were added at a loading of200 FPU/g cellulose and 200 CBU/g cellulose to the pretreatedDouglas-fir substrates in order to achieve near complete enzymatichydrolysis of cellulose and obtain lignin as pure as possible. Theenzymatic hydrolysis was conducted at the conditions describedin section 2.5, at a 2% (w/w) consistency. After hydrolysis, the lig-nin residue was recovered by centrifugation at 8000g for 30 minand re-suspended in distilled water to a final consistency of 2%(w/w). The suspension was sonicated at 40 kHz for 60 min in aTP 680DH ultrasonic water bath (Elma Hans Schmidbauer GmbH& Co., Singen, Germany). The washing was done three times andthe lignin residue was incubated overnight at 37 �C in 50 mMphosphate buffer, pH 7, containing 1 U/mL Pronase (Sigma Chemi-cal Company, USA). After protease treatment, the residual ligninwas heated at 100 �C for 15 min to deactivate the protease fol-lowed by washing four times with acetate buffer, repeating thewashing steps with 1 M sodium chloride solution and thereafterwith distilled water. The lignin preparations were freeze dried,ground manually, screened through a 180 mesh screen and storedin sealed vials at room temperature in a desiccator. The lignin wasanalyzed for lignin and carbohydrate content as described in sec-tion 2.6. The nitrogen content of the residual lignin was deter-mined using a LECO CN-2000 elemental analyzer (Leco Inc., MI,USA). The Klason lignin, carbohydrate, and nitrogen contents ofthe isolated lignin sample were 93.3 ± 0.6, 2.7 ± 0.8, and 0.6 ± 0.1,respectively.

2.8. Determination of the water retention value (WRV)

The WRV was measured following the Scandinavian test meth-od SCAN-C 62:00 in five replicates. Pulp pad making unit consist ofa 25 mm inner diameter plastic tube with a height of 60 mm with anylon mesh (pore size �100 um) at the bottom of the tube.Approximately 0.5 g of the substrate was disintegrated in 50 mLof water and soaked overnight. The resulting suspension was care-fully filtered through the unit using the suction so as to result inapproximately 5–10% dry matter in the pulp. Subsequently, thewhole test pad holding unit was placed in a 50-mL falcon tube(30 mm internal diameter and 100 mm height) with an open endedplastic tube supporting the unit at the bottom, thus leaving spacefor water accumulation at the bottom of the unit. The whole unitwas subsequently centrifuged at 3000g for 15 min in a laboratorycentrifuge (Thermo Fisher Scientific, Sorvall ST 16R). The wetcentrifuged sample pad was weighed and oven dried at 105 �Covernight and weighed again. The WRV is the percentage of

204 L. Kumar et al. / Bioresource Technology 103 (2012) 201–208

retained water (weight change of the substrate before and afterdrying) of the dried substrate, i.e. WRV ð%Þ ¼ wwet�wdried

wdried� 100

where wwet and wdried are the wet and dry weights of the substrate,respectively.

2.9. Determination of cellulose accessibility of pure cellulosicsubstrates

Accessibility of the cellulose for each substrate was measuredwith the Family 2a cellulose-specific carbohydrate binding module(CBM2a) of xylanase 10A from Cellulomonas fimi. Adsorption ofCBM2a to the sample was used as a quantitative indicator of the to-tal amount of available cellulose surface area (McLean et al., 2000,2002; Tomme et al., 1998). Samples containing 450 lg CBM2a and10 mg substrate (on a dry weight basis) were made up to a finalvolume of 1.5 mL of 50 mM Tris–HCl at pH 8.0. Samples were incu-bated at room temperature for 30 min followed by centrifugationat 16,000g for 10 min using Sorvall micro-centrifuge (Sorvall, Ashe-ville, NC). The concentration of CBM2a in the supernatant wasdetermined by measuring the absorbance of the solution at280 nm (by using a Cary 50 UV–Vis spectrophotometer – VarianInc., Mississauga, ON) using an extinction coefficient of27625 M�1 cm�1 (McLean et al., 2002).

3. Results and discussion

3.1. Influence of enzyme loading on the overall inhibitory effect oflignin

Steam pretreated softwoods have a greater degree of recalci-trance to enzymatic hydrolysis than hardwood and agriculturalresidues (Mabee et al., 2006; Kumar et al., 2010). Even afterincreasing the severity of pretreatment, only about 21% of the cel-lulose present in the water insoluble component was hydrolyzedwhen an enzyme loading of 5 FPU/g cellulose was used (Kumaret al., 2011). However, when this substrate was subsequently del-ignified, the same protein loading resulted in near complete hydro-lysis of the cellulose, indicating that lignin played an importantrole in limiting the enzymatic hydrolysis of steam pretreated soft-woods when low enzyme loadings are used (Kumar et al., 2010,2011). As it is known that lignin restricts enzymatic hydrolysisby both binding cellulases and restricting access to the cellulose,very high protein loading and prolonged hydrolysis times wereemployed (realizing this would not be economically viable), tosee if it is possible to completely hydrolyze a recalcitrant substratesuch as steam pretreated Douglas fir. Understandably, previouswork has tended to look at the influence of a lower range of en-

Fig. 1. (a) Effect of increasing cellulase loading (——) and BSA pre-incubation (- - -) on theInfluence of delignification (SPDF-CD) on the rate and extent of enzymatic hydrolysis of S(—) pre-incubation of BSA.

zyme loadings (20–120 FPU/g cellulose) (Pan et al., 2004; Tengborget al., 2001) as the primary goal was to assess the efficiency ofenzymatic hydrolysis using a reasonable range of enzyme loadings.However, in the work described here, it was important to ensurethat complete hydrolysis of the cellulose could be achieved, withenough protein and time available. It was apparent that, at veryhigh enzyme loadings (200 FPU/g cellulose) and given enough time(72 h), steam pretreated Douglas-fir (SPDF) could be completelyhydrolyzed (Fig. 1a). However, there was a dramatic difference atlow (less than 40 FPU/g cellulose) enzyme loadings, with slowand incomplete hydrolysis obtained even after 3 days of incuba-tion. Only 16% and 33% of the cellulose present in the SPDF washydrolyzed after 72 h at enzyme loadings of 5 FPU and 10 FPU/gcellulose, respectively (Fig. 1a).

3.2. Relative contributions of lignin towards enzyme binding andlimited cellulose accessibility

To try to differentiate the role that lignin might play in eitherenzyme binding or restricting enzyme access to the cellulose, Bo-vine Serum Albumen (BSA) was added to the substrate prior toenzymatic hydrolysis, with the intent that cellulase binding to lig-nin would be surrogated by prior saturation of the lignin by theBSA. Initially, the amount of BSA required to saturate the ligninbinding sites was optimized by monitoring the improvement inenzymatic hydrolysis of steam pretreated Douglas-fir achieved atdifferent BSA loadings. The prior addition of 2.5–5.0 mg/mL(123–245 mg/g of the substrate) of BSA was found to be the opti-mum range of BSA concentration that can achieve highestimprovement in enzymatic hydrolysis (data not shown). It wasclear that at higher enzyme loadings, prior addition of the BSAhad only a marginal beneficial effect after 72 h hydrolysis(Fig. 1a). However, at low enzyme loadings (5 FPU/g cellulose),prior addition of the BSA to the non-delignified steam pretreatedDouglas fir (SPDF) increased both the rate and extent of hydrolysissubstantially, with about 66% of the original cellulose hydrolyzedafter 72 h as compared to only 16% without prior BSA addition(Fig. 1b). When the BSA was added to the steam pretreated Douglasfir (SPDF) which had been delignified (SPDF-CD), no direct benefi-cial effect was observed (Fig. 1b). This result indicated that pre-incubation of the higher lignin containing substrate with BSAessentially enhanced the availability of cellulases which wouldhave otherwise been unproductively adsorbed to lignin. Althoughlignin binding of the enzymes played an important role in restrict-ing enzyme hydrolysis, it was also clear that lignin played animportant role in limiting the accessibility of the enzymes to thecellulose, as indicated by incomplete hydrolysis of the BSA-treated

extent of enzymatic hydrolysis of steam pretreated Douglas fir (SPDF) after 72 h. (b)PDF when low (5 FPU/g cellulose) enzyme loadings are used with (- - -) and without

Table 2Chemical composition and cellulose accessibility (WRV) of partial delignified/ligninmodified steam pretreated Douglas-fir (SPDF).

Substrate Lignin (%) Glucan (%) Water retention value (%)

SPDF – Control 42.8 (0.3)** 56.8 (0.6) 121.3 (10.3)SPDF-D38* 38.2 (0.1) 61.4 (0.3) 150.9 (12.2)SPDF-D30* 29.7 (0.3) 68.3 (0.2) 177.2 (3.1)SPDF-D22* 22.4 (0.2) 75.8 (0.4) 195.4 (3.7)SPDF-D11* 10.9 (0.6) 87.5 (0.3) 219.7 (8.1)SPDF-CD* BDL 99.3 (0.8) 250.2 (7.3)SPDF-S41* 41.1 (0.1) 58.3 (0.4) 170.6 (13.0)

* substrates refer to the partially delignified substrates. SPDF-CD refers to thecompletely delignified steam pretreated substrate and S41 refers to the neutralsulfonated steam pretreated Douglas-fir.** Numbers in the brackets represent standard deviations from the average.

Fig. 2. Influence of substrate delignification on 72 h enzymatic hydrolysis yield ofsteam pretreated Douglas-fir (SPDF) when low enzyme loadings (5 FPU/g cellulose)are used with (- - -) and without pre-incubation of BSA (—).

Fig. 3. Effect of neutral sulfonation (S-41) on the enzymatic hydrolysis of steampretreated Douglas-fir (SPDF) when using low enzyme loading (5 FPU/g cellulose)with (- - -) and without (—) BSA pre-incubation.

L. Kumar et al. / Bioresource Technology 103 (2012) 201–208 205

SPDF substrate at 5 FPU/g compared to the near complete hydroly-sis of the completely delignified, SPDF-CD substrate. These resultsindicated that the two main influences of lignin, in both directlybinding with the enzymes and restricting the swelling of the sub-strate were at play.

3.3. Influence of lignin removal/modification on cellulose accessibilityand unproductive binding

To try to further elucidate how the contribution of unproductivebinding of the enzymes to the lignin might be affected by increas-ing cellulose accessibility, the steam pretreated Douglas-fir sub-strate was sequentially delignified, using sodium chloritetreatment at room temperature, to ensure minimal changes tothe cellulose. This procedure is the standardized delignificationtechnique typically used for the determination of the degree ofpolymerization (DP) of isolated celluloses (PAPTAC, 1998).Although other workers had successfully used hot chlorite treat-ment on substrates with a lignin content as high as 30% (Varnaiet al., 2010; Hubbell and Ragauskas, 2010), a hot temperaturetreatment was avoided to minimize any possible detrimental ef-fects on the cellulose.

The partially delignified substrates had a lignin content of 38,30, 22 and 11% (corresponding to 18, 42, 62 and 84% removal of lig-nin) and were referred as SPDF-D38, SPDF-D30, SPDF-D22 andSPDF-D11, respectively (Table 2). As was expected, with decreasinglignin content there was an almost equivalent increase in theamount of cellulose present in the substrate, indicating the selec-tiveness of sodium chlorite delignification. The steam pretreatedDouglas-fir sample was also sulfonated at neutral pH to producea substrate (SPDF-S41) which differed significantly in terms of cel-lulose accessibility but was still quite similar in its lignin content(Table 2).

As previous work had shown that the water retention value(WRV) could provide a relatively good indication of cellulose swell-ing (Ogiwara and Arai, 1968, 1972; Luo and Zhu, 2010) the differ-ence in the cellulose accessibility between SPDF and thecorresponding delignified/lignin modified substrates was assessedusing this procedure. It was apparent that the accessibility of thecellulose, as indicated by the increase in the WRV, increased withincreasing lignin removal (Table 2), confirming earlier work thatshowed that chlorite delignification solubilized the high molecularweight lignin and consequently increased fiber swelling (Ahlgrenand Goring, 1971; Ahlgren et al., 1971). The sulfonated steam pre-treated substrate (S-41) also showed an increased WRV value. Thiswas likely due to incorporation of sulfonic acid groups during neu-tral sulfonation at the a or c position of the lignin side chain mak-ing the lignin more hydrophilic (Mooney et al., 1998; Sarkanen andLudwig, 1971), suggesting that the sulfonated substrate may alsobe more accessible to the cellulase enzymes.

To see if this increase in accessibility translated into betterhydrolysis at low enzyme loading, 5 FPU/g cellulose of enzymewas next added to the partially delignified substrates and the ex-tent of hydrolysis after 72 h, with and without prior addition ofBSA (Fig. 2) was determined. For the substrates without priorBSA addition, an almost linear increase in cellulose hydrolysiswas observed with increasing lignin removal. However, with thehigh lignin containing substrates, the prior addition of BSA had asignificant effect, resulting in substantially increased levels ofhydrolysis. As expected, for the completely delignified SPDF-CDsubstrate, prior addition of BSA resulted in little improvement(Fig. 2). It was apparent that the lignin-enzyme interaction thatled to unproductive binding did not decrease proportionately witha decrease in the lignin content, indicating that unproductive bind-ing was also dependent on other factors. Removal of more than 40%of the substrate’s lignin increased cellulose accessibility (WRV)from 120 to 177 (Table 2) while the beneficial effects of priorBSA addition decreased substantially (Fig. 2). It is likely that, forthe less accessible substrates, the unproductive binding of cellu-

Fig. 4. A comparison of the extent of Cellulose Binding Module (CBM) binding, usedto assess the likely cellulose accessibility of a range of pure cellulosic substrates.SPDF-CD refers to the completely delignified steam pretreated Douglas-fir.

206 L. Kumar et al. / Bioresource Technology 103 (2012) 201–208

lases at low protein loading predominates. As cellulose accessibil-ity increases, lignin plays a less significant role as an unproductiveadsorbent, even at high lignin concentrations, as there was enoughcellulose available for good hydrolysis.

To try to further elucidate the possible mechanisms involved,the hydrolysis profiles of the steam pretreated Douglas-fir thathad been sulfonated (SPDF-S-41), were determined again usinglow enzyme loadings (5 FPU/g), with or without prior addition ofBSA. Although neutral sulfonation resulted in a significant increasein both the rate and extent of hydrolysis, prior incubation with BSAhad a much more dramatic effect on the unsulfonated substrate(Fig. 3). This again suggested that when cellulose accessibility is in-creased (as a result of sulfonation), even though lignin is present inrelatively high concentrations, it plays a significantly lesser role inhindering the effectiveness of enzymatic hydrolysis. The small, butsignificant beneficial effect noted by prior BSA incubation with thesulfonated substrate (SPDF-S41) was also likely due to the decreasein the lignin’s hydrophobicity decreasing the unproductive adsorp-tion of the cellulase enzymes.

3.4. Influence of lignin’s unproductive binding with changes incellulose accessibility

While the water retention values (WRV) had provided a goodindication of the overall accessibility of the whole substrate, a

Fig. 5. Influence of added lignin, isolated from steam pretreated Douglas-fir (SPDF) on thcellulose, of never-dried, completely delignified, steam pretreated Douglas-fir (SPDF-CD

more representative probe that would both assess the specificcellulose surface accessibility and simulate the types of lignin-enzyme interactions that were likely to occur with realisticbiomass substrates would be useful. Previous work (Mclean etal., 2000, 2002; Rollin et al., 2011) had shown that measuring theadsorption profile of the cellulose-specific carbohydrate bindingmodule (CBM2a) could provide a quantitative indicator of the totalamount of available cellulose surface area. To see if the CBM probewould give an indication of the anticipated cellulose accessibility, arange of substrates was prepared, including the never dried, com-pletely delignified steam pretreated Douglas fir, the same substratethat had been oven dried as well as two cellulose controls, Aviceland Sigmacell. As mentioned earlier in the section 3.1, the never-dried SPDF-CD was accessible enough such that complete cellulosehydrolysis was obtained at an enzyme loading of 5 FPU/g cellulose,while the other cellulosic substrates were expected to have muchlower cellulose accessibilities and, consequently, proved to bemore recalcitrant.

As expected, the substrate with the highest cellulose accessibil-ity was the never-dried SPDF-CD while the same substrate when itwas oven dried was much less accessible to the CBM probe (Fig. 4),likely due to hornification of the cellulose fibers upon drying andthe resulting collapse of the cellulose fibers (Welf et al., 2005).The cellulose accessibility of Avicel and Sigmacell was similarand slightly better than that of the oven dried SPDF-CD (Fig. 4).Avicel and Sigmacell are dried during their manufacture (moisturecontent �5%) which also probably resulted in the collapse of thecellulose fibers (Luukkonen et al., 2001). When these substrateswere hydrolyzed over 72 h at an enzyme loading of 5 FPU/g cellu-lose, the SPDF-CD substrate was completely hydrolyzed while 59%,57%, and 53% of the original cellulose present in the respective Avi-cel, Sigmacell and oven-dried SPDF-CD was hydrolyzed. The factthat the cellulose content of the never dried and oven dried sub-strates (SPDF-CD) is the same indicated that the difference in theenzymatic hydrolysis must be due to differences in the celluloseaccessibility.

To try to further elucidate the roles that lignin might play inrestricting effective enzymatic hydrolysis, isolated lignin was ob-tained from the steam pretreated Douglas fir using previous re-ported methods (Berlin et al., 2006) and this lignin was added toeach of the four substrates described above (Avicel, Sigmacell,never-dried SPDF-CD and Oven-dried SPDF-CD) prior to carryingout the same hydrolysis procedure as described previously(5 FPU/g cellulose, 72 h). The lignin was added in the same propor-tions as it was originally found in the steam pretreated Douglas firsubstrate prior to delignification. There was a minimum decrease(less than 10%) in the hydrolysis of the never dried SPDF-CD sub-

e rate and extent of enzymatic hydrolysis at, (a) 5 FPU/g cellulose and, (b) 10 FPU/g).

L. Kumar et al. / Bioresource Technology 103 (2012) 201–208 207

strate while the oven dried substrates hydrolysis yield was reducedby about 48% and the Avicel and Sigmacell were reduced by about45%.

It seems that the degree by which isolated lignin de-creased the hydrolysis yields was critically dependent onhow accessible the cellulose was. When cellulose was highlyaccessible, such as in never-dried SPDF-CD, lignin had a minorrole in decreasing hydrolysis yields. Alternatively, when thecellulose is not readily accessible, such as occurred with theoven dried SPDF-CD or Avicel and Sigmacell, the decreasewas significant.

These results confirmed that lignin’s major role in restrictingenzymatic hydrolysis is primarily through limiting celluloseaccessibility to the enzymes. At low enzyme loadings (5 FPU/gcellulose), only 16% of the cellulose present in steam pretreatedDouglas-fir was hydrolyzed compared to near complete hydroly-sis of the corresponding, completely delignified substrate. Toconfirm that the beneficial effect of lignin removal observedwas due to increase of the specific cellulose surface area we nextcompared the hydrolysis profile of the steam pretreated Douglasfir (SPDF) substrate, this same substrate which had been com-pletely delignified (SPDF-CD) and this later substrate to whichhad been added back the same amount of lignin found in theoriginal SPDF substrate (Fig. 5). At an enzyme loading of5 FPU/g cellulose (Fig. 5a), the prior addition of the lignin backto the delignified substrate (SPDF-CD) did slow both the rateand degree of completeness of hydrolysis, although prolongedincubation (for a further 48 h) brought the final yields close towhat was obtained with the delignified substrate. When an en-zyme loading of 10 FPU/g cellulose was used, the prior additionof lignin to the delignified substrate had little effect, implyingthat the enough protein had been added to ensure that inhibi-tion caused by the proteins binding to the lignin had beenovercome.

4. Conclusions

This work compared the relative contributions that ligninplays in restricting cellulose hydrolysis through either bindingof the cellulases or limiting swelling of the cellulose and thereforelimiting accessibility. It was apparent that the inhibitory effect oflignin is dependent on the enzyme loading employed as the useof excessive cellulase loadings resulted in near complete cellulosehydrolysis of steam pretreated Douglas-fir. However, at low cellu-lase loadings (<10 FPU/g cellulose) the presence of lignin resultedin poor hydrolysis yields mainly due to the unproductive bindingof the enzymes and limited cellulose accessibility. However, whenthe lignin does not restrict cellulose swelling and the cellulose isreadily accessible to the enzymes, the impact of unproductivebinding on the hydrolysis yield is lower. When access to the cel-lulose was enhanced, even at a high lignin concentration, goodhydrolysis yields were obtained, thus minimizing lignins’ role asboth an unproductive adsorbent and a limitation to celluloseaccessibility.

Acknowledgements

The authors thank Asia Pacific Partnership (APP) Program forthe financial support of the study. We acknowledge Prof. DouglasKilburn for the generous gift of the CBM for our study. Our thanksto Mr. Keith Gourlay for his technical help with laboratory exper-iments and valuable advice for this study. We also thank allmembers of the Forest Products Biotechnology/Bioenergy Re-search Group UBC for their support and cooperation throughoutthe study.

References

Ahlgren, P.A., Goring, D.A.I., 1971. Removal of wood components during chloritedelignification of black spruce. Can. J. Chem. 49, 1272–1275.

Ahlgren, P.A., Yean, W.Q., Goring, D.A.I., 1971. Chlorite delignification of sprucewood – comparison of molecular weight of lignin dissolved with size of pores incell wall. Tappi J. 54, 737–740.

Arantes, V., Saddler, J.N., 2010. Access to cellulose limits the efficiency of enzymatichydrolysis: the role of amorphogenesis. Biotechnology for Biofuels. Available at:<http://www.biotechnologyforbiofuels.com/content/3/1/4>

Arantes, V., Saddler, J.N., 2011. Cellulose accessibility limits the effectiveness ofminimal cellulase loading on the efficient hydrolysis of pretreatedlignocellulosic substrates. Biotechnology for Biofuels. Available at: <http://www.biotechnologyforbiofuels.com/content/4/1/3>

Berlin, A., Gilkes, N., Kurabi, A., Bura, R., Tu, M.B., Kilburn, D., Saddler, J., 2005. Weaklignin-rinding enzymes – a novel approach to improve activity of cellulases forhydrolysis of lignocellulosics. Appl. Biochem. Biotechnol. 121, 163–170.

Berlin, A., Balakshin, M., Gilkes, N., Kadla, J., Maximenko, V., Kubo, S., Saddler, J.,2006. Inhibition of cellulase, xylanase and beta-glucosidase activities bysoftwood lignin preparations. J. Biotechnol. 125, 198–209.

Biermann, C.J., 1996. Handbook of Pulping and Paper Making. San Diego, AcademicPress Inc, pp. 741.

Bura, R., Chandra, R., Saddler, J., 2009. Influence of xylan on the enzymatichydrolysis of steam-pretreated corn stover and hybrid poplar. Biotechnol. Prog.25, 315–322.

Dence, C.W., 1992. The determination of lignin. In: Lin, S.Y., Dence, C.W. (Eds.),Methods in Lignin Chemistry. Springer, Berlin, pp. 33–61.

Donaldson, L.A., Wong, K.K.Y., Mackie, K.L., 1988. Ultrastructure of steam-explodedwood. Wood Sci. Technol. 22, 103–114.

Donghai, S., Junshe, S., 2007. Enhancing hydrolysis efficiency of cellulase andreducing the cost of saccharification. Prog. Chem. 19, 1147–1152.

Donohoe, B.S., Decker, S.R., Tucker, M.P., Himmel, M.E., Vinzant, T.B., 2008.Visualizing lignin coalescence and migration through maize cell wallsfollowing thermochemical pretreatment. Biotechnol. Bioeng. 101, 913–925.

Eriksson, I., Haglind, I., Lidbrandt, O., Salmen, L., 1991. Fiber swelling favored bylignin softening. Wood Sci. Technol. 25, 135–144.

FAO., 2010. Global Forest Resources Assessment, Food and AgriculturalOrganization of the United Nations (FAO) Forestry Paper 163: 266-271. Rome,Italy. Available at: <http://www.fao.org/forestry/fra/fra2010/en/> (last accessedon 29 August 2011).

Gregg, D., Saddler, J.N., 1996. A techno-economic assessment of the pretreatmentand fractionation steps of a biomass-to-ethanol process. Appl. Biochem.Biotechnol. 57–8, 711–727.

Hendriks, A.T.W.M., Zeeman, G., 2009. Pretreatments to enhance the digestibility oflignocellulosic biomass. Bioresour. Technol. 100, 10–18.

Holtzapple, M.T., Humphrey, A.E., Taylor, J.D., 1989. Energy-requirements for thesize-reduction of poplar and aspen wood. Biotechnol. Bioeng. 33, 207–210.

Hubbell, C.A., Ragauskas, A.J., 2010. Effect of acid-chlorite delignification oncellulose degree of polymerization. Bioresour. Technol. 101, 7410–7415.

IEA Task 39, 2011. Status of second generation biofuel demonstration plants, 2011.<http://biofuels.abc-energy.at/demoplants/projects/mapindex> (last accessedon 2 May 2011).

Jeoh, T., Ishizawa, C.I., Davis, M.F., Himmel, M.E., Adney, W.S., Johnson, D.K., 2007.Cellulase digestibility of pretreated biomass is limited by cellulose accessibility.Biotechnol. Bioeng. 98, 112–122.

Kawamoto, H., Nakatsubo, F., Murakami, K., 1992. Protein-adsorbing capacities oflignin samples. Mokuzai Gakkaishi 38, 81–84.

Kumar, L., Chandra, R., Chung, P.A., Saddler, J., 2010. Can the same steampretreatment conditions be used for most softwoods to achieve good,enzymatic hydrolysis and sugar yields? Bioresour. Technol. 101, 7827–7833.

Kumar, L., Chandra, R., Saddler, J., 2011. Influence of steam pretreatment severity onpost-treatments used to enhance the enzymatic hydrolysis of pretreatedsoftwoods at low enzyme loadings. Biotechnol. Bioeng. 108–10, 2300–2311.

Luo, X., Zhu, J.Y., 2010. Effects of drying induced fiber hornification on enzymaticsaccharification of lignocelluloses. Enzyme Microb. Technol. 48, 92–99.

Luukkonen, P., Maloney, T., Rantanen, J., Paulapuro, H., Yliruusi, J., 2001.Microcrystalline cellulose–water interaction – a novel approach usingthermoporosimetry. Pharm. Res. 18, 1562–1569.

Mabee, W.E., Gregg, D.J., Arato, C., Berlin, A., Bura, R., Gilkes, N., Mirochnik, O., Pan,X.J., Pye, E.K., Saddler, J.N., 2006. Updates on softwood-to-ethanol processdevelopment. Appl. Biochem. Biotechnol. 129, 55–70.

Matsushika, A., Inoue, H., Kodaki, T., Sawayama, S., 2009. Ethanol production fromxylose in engineered Saccharomyces cerevisiae strains: current state andperspectives. Appl. Microbiol. Biotechnol. 84, 37–53.

McLean, B.W., Bray, M.R., Boraston, A.B., Gilkes, N.R., Haynes, C.A., Kilburn, D.G.,2000. Analysis of binding of the family 2a carbohydrate-binding module fromCellulomonas fimi xylanase 10A to cellulose: specificity and identification offunctionally important amino acid residues. Protein Eng. 13, 801–809.

McLean, B.W., Boraston, A.B., Brouwer, D., Sanaie, N., Fyfe, C.A., Warren, R.A.J.,Kilburn, D.G., Haynes, C.A., 2002. Carbohydrate-binding modules recognize finesubstructures of cellulose. J. Biol. Chem. 277, 50245–50254.

Merino, S.T., Cherry, J., 2007. Progress and challenges in enzyme development forbiomass utilization. Adv. Biochem. Eng. Biotechnol. 108, 95–120.

Mooney, C.A., Mansfield, S.D., Tuohy, M.G., Saddler, J.N., 1997. The effect of lignincontent on cellulose accessibility and enzymatic hydrolysis of softwood pulps.

208 L. Kumar et al. / Bioresource Technology 103 (2012) 201–208

In: Proceedings of 1997 Biological Sciences Symposium, October 19–23, 1997,San Francisco Marriott, San Francisco, CA. Tappi Press, pp. 259–265.

Mooney, C.A., Mansfield, S.D., Touhy, M.G., Saddler, J.N., 1998. The effect of initialpore volume and lignin content on the enzymatic hydrolysis of softwoods.Bioresour. Technol. 64, 113–119.

Nakagame, S., Chandra, R.P., Saddler, J.N., 2010. The effect of isolated lignins,obtained from a range of pretreated lignocellulosic substrates, on enzymatichydrolysis. Biotechnol. Bioeng. 105, 871–879.

Nakatani, T., Ishimaru, Y., Iida, I., Furuta, Y., 2008. Micropore structure of wood:change in micropore structure accompanied by delignification. J. Wood Sci. 54,252–255.

Ogiwara, Y., Arai, K., 1968. Swelling degree of cellulose materials and hydrolysis ratewith cellulase. Text. Res. J. 38, 885–891.

Ogiwara, Y., Arai, K., 1972. Kinetic study of swelling effects on cellulase-hydrolysisreaction of cellulose materials. Nippon Kagaku Kaishi 1, 189–193.

Palonen, H., Tjerneld, F., Zacchi, G., Tenkanen, M., 2004. Adsorption of Trichodermareesei Cbh I and Eg II and their catalytic domains on steam pretreated softwoodand isolated lignin. J. Biotechnol. 107, 65–72.

Pan, X.J., Zhang, X., Gregg, D.J., Saddler, J.N., 2004. Enhanced enzymatic hydrolysis ofsteam-exploded douglas fir wood by alkali-oxygen post-treatment. Appl.Biochem. Biotechnol. 113–16, 1103–1114.

Pan, X.J., Xie, D., Gilkes, N., Gregg, D.J., Saddler, J.N., 2005. Strategies to enhance theenzymatic hydrolysis of pretreated softwood with high residual lignin content.Appl. Biochem. Biotechnol. 121, 1069–1079.

Pulp and Paper Technical Association of Canada (PAPTAC), 1998. Useful methods G10U. PAPTAC, Montreal, Canada.

Rollin, J.A., Zhu, Z.G., Sathitsuksanoh, N., Zhang, Y.H.P., 2011. Increasing celluloseaccessibility is more important than removing lignin: a comparison of cellulosesolvent-based lignocellulose fractionation and soaking in aqueous ammonia.Biotechnol. Bioeng. 108, 22–30.

Sarkanen, K.V., Ludwig, C.H., 1971. Lignins: Occurrence, Formation, Structure andReactions, John Wiley & Sons Inc, New York, pp. 916.

Selig, M., Weiss, N., Ji, Y., 2008. Enzymatic saccharification of lignocellulosicbiomass, NREL technical report, Laboratory analytical procedure. NREL/TP-510-42629. Available at: <http://www.nrel.gov/biomass/pdfs/42629.pdf>.

Solomon, B.D., Johnson, N.H., 2009. Valuing climate protection through willingnessto pay for biomass ethanol. Ecol. Econ. 68, 2137–2144.

TAPPI, 1994. TAPPI Test Method T222 om-88, ‘‘Acid-Insoluble Lignin in Wood andPulp.’’ In Tappi Test Methods. Technical Association of the Pulp and PaperIndustry, Atlanta, GA

TAPPI, 1994. TAPPI Test Method T-204 OM-88. ‘‘Solvent extractives of wood andpulp’’ Tappi Test Methods. Technical Association of the Pulp and Paper Industry,Atlanta, GA

Tengborg, C., Galbe, M., Zacchi, G., 2001. Influence of enzyme loading and physicalparameters on the enzymatic hydrolysis of steam-pretreated softwood.Biotechnol. Prog. 17, 110–117.

Tomme, P., Boraston, A., McLean, B., Kormos, J., Creagh, A.L., Sturch, K., Gilkes, N.R.,Haynes, C.A., Warren, R.A.J., Kilburn, D.G., 1998. Characterization and affinityapplications of cellulose-binding domains. J. Chromatogr. B Biomed. Sci. Appl.715, 283–296.

Tu, M., Chandra, R.P., Saddler, J.N., 2007. Recycling cellulases during the hydrolysisof steam exploded and ethanol pretreated lodgepole pine. Biotechnol. Prog. 23,1130–1137.

Varnai, A., Siika-aho, M., Viikari, L., 2010. Restriction of the enzymatic hydrolysis ofsteam-pretreated spruce by lignin and hemicellulose. Enzyme Microb. Technol.46, 185–193.

Welf, E.S., Venditti, R.A., Hubbe, M.A., Pawlak, J.J., 2005. The effects of heatingwithout water removal and drying on the swelling as measured by waterretention value and degradation as measured by intrinsic viscosity of cellulosepapermaking fibers. Prog. Pap. Recycl. 14, 1–9.

Wong, K.K.Y., Deverell, K.F., Mackie, K.L., Clark, T.A., Donaldson, L.A., 1988. Therelationship between fiber porosity and cellulose digestibility in steam-exploded pinus-radiata. Biotechnol. Bioeng. 31, 447–456.

World Watch Institute, 2007. Biofuels for Transport: Global Potential Implicationsfor Sustainable Agriculture and Energy. Earthscan, London, UK.

Yang, B., Boussaid, A., Mansfield, S.D., Gregg, D.J., Saddler, J.N., 2002. Fast andefficient alkaline peroxide treatment to enhance the enzymatic digestibility ofsteam-exploded softwood substrates. Biotechnol. Bioeng. 77, 678–684.

Yang, B., Wyman, C.E., 2006. BSA treatment to enhance enzymatic hydrolysis ofcellulose in lignin containing substrates. Biotechnol. Bioeng. 94, 611–617.

Zhang, Y.H.P., Cui, J.B., Lynd, L.R., Kuang, L.R., 2006. A transition from celluloseswelling to cellulose dissolution by o-phosphoric acid: evidence fromenzymatic hydrolysis and supramolecular structure. Biomacromolecules 7,644–648.