CHEMSUSCHEM FULL PAPERS DOI: 10.1002/cssc.201400042

Effect of Lignin Chemistry on the Enzymatic Hydrolysis of Woody Biomass Zhiying Yu, Ki-Seob Gwak, Trevor Treasure, Hasan Jameel, Hou-min Chang, and Sunkyu Park*[a] The impact of lignin-derived inhibition on enzymatic hydrolysis is investigated by using lignins isolated from untreated woods and pretreated wood pulps. A new method, biomass reconstruction, for which isolated lignins are precipitated onto bleached pulps to mimic lignocellulosic biomass, is introduced, for the first time, to decouple the lignin distribution issue from lignin chemistry. Isolated lignins are physically mixed and reconstructed with bleached pulps. Lignins obtained from pretreated woods adsorb two to six times more cellulase than lignins obtained from untreated woods. The higher adsorption of enzymes on lignin correlates with decreased carbohydrate con-

version in enzymatic hydrolysis. In addition, the reconstructed softwood substrate has a lower carbohydrate conversion than the reconstructed hardwood substrate. The degree of condensation of lignin increases significantly after pretreatment, especially with softwood lignins. In this study, the degree of condensation of lignin (0.02 to 0.64) and total OH groups in lignin (1.7 to 1.1) have a critical impact on cellulase adsorption (9 to 70 %) and enzymatic hydrolysis (83.2 to 58.2 %); this may provide insights into the more recalcitrant nature of softwood substrates.

Introduction Enzymatic hydrolysis of lignocellulosic biomass is one of the key steps for sugar production and the subsequent conversion of sugars to biofuels and chemicals. The abundant woody biomasses in North America are great resources for sugar production through enzymatic hydrolysis. Softwood (SW), such as loblolly pine, accounts for over 80 % of planted forests in the southeast of America.[1] However, it has been widely observed that SW is more recalcitrant in terms of its enzymatic hydrolysis compared with hardwood (HW).[2] One might consider that the recalcitrant nature of SW is due to its larger physical dimension because SW fiber is larger in length (3 vs. 1 mm) and width (30–50 vs. 20–40 mm) than that of HW.[3] However, the digestion of fully bleached HW and SW were almost identical,[4] which indicated that the fiber dimension was not a critical factor. Many researchers have found that, even though the lignin contents of SW and HW substrates were reduced to similar levels by pretreatment, SW substrates were still more resistant to enzymatic hydrolysis than HW.[2b, 5] The different degree of hydrolysis between HW and SW at similar lignin levels depends on the lignin content and the methods of pretreatment.[2b] For example, it was found that, at similar lignin content of green-liquor (GL) pretreatment followed by delignifica[a] Dr. Z. Yu, Dr. K.-S. Gwak, T. Treasure, Dr. H. Jameel, Dr. H.-m. Chang, Dr. S. Park Department of Forest Biomaterials North Carolina State University 2820 Faucette Boulevard, Campus Box 8005 Raleigh, NC 27695-8005 (USA) E-mail: [email protected] Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cssc.201400042.

 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

tion-treated HW and SW, if the lignin content was beyond 15 %, SW substrate showed up to 40 % lower carbohydrate conversion than that of HW substrate. Ozone-delignified HW and SW substrates had more distinctive digestion than that of sodium chlorite delignified substrates.[2b] Therefore, the difference in lignin chemistry and structure in SW and HW substrates is considered to be an important reason for their digestibility. Lignin-derived inhibition is a major obstacle that restricts enzymatic hydrolysis of lignocellulosic substrate, especially for SW.[6] It has been suggested that the inhibition effect of lignin on enzymatic hydrolysis can be classified into three categories: 1) Enzymes can be adsorbed on lignin through hydrophobic interactions, electrostatic interactions, and/or hydrogen-bonding interactions. 2) Lignin in lignocellulosic materials acts as a surface barrier to block the accessible surface of carbohydrates through physical blockage on the surface and chemical blockage through lignin carbohydrate complex. 3) Soluble lignin derivatives can deactivate enzymes in liquid phase.[7] However, studies on the mechanisms of lignin-derived inhibition on enzymatic hydrolysis are very limited. One of the greatest challenges is to delineate the coupling factors of lignin chemistry and physical barrier of lignin. The pattern of lignin incorporated with polysaccharides might be another important factor other than lignin content that limits enzymatic hydrolysis. In addition, the distribution of lignin that specifies the pattern of a physical barrier might be equally as important as the mechanisms mentioned above. Herein, we developed a new method to decouple the lignin distribution issue from lignin chemistry. A unique technique, the reconstruction of lignocellulose, which uses isolated lignin ChemSusChem 0000, 00, 1 – 10

&1&

These are not the final page numbers! ÞÞ

CHEMSUSCHEM FULL PAPERS

www.chemsuschem.org

and bleached pulps to precipitate the isolated lignin back to Some of the key lignin structures that have potential impact pulp, was established, for the first time, to investigate ligninon enzymatic hydrolysis are listed in Table 1. After pretreatderived inhibition. In this study, milled wood lignin (MWL) was ment, the total OH groups in both HW and SW decreased. The isolated from untreated and pretreated woody biomass. By comTable 1. Characterization of MWL by 13C NMR spectroscopy. paring the enzymatic hydrolysis between reconstructed pulp and Structure HW MWLs SW MWLs physically mixed MWL with GL-HW Pine Au-SW[a] GL-SW eucalyptus maple Au-HW[a] bleached pulp at the same lignin total OH 1.70[b] 1.58 1.19 1.18 1.41 1.02 1.11 content, the factors of lignin aliphatic OH, primary 0.81 0.72 0.44 0.30 0.81 0.49 0.45 aliphatic OH, secondary 0.67 0.61 0.25 0.32 0.34 0.19 0.20 that limit enzymatic hydrolysis phenolic OH 0.22 0.25 0.51 0.55 0.27 0.34 0.46 were decoupled into chemistry total COOR 0.08 0.07 0.07 0.08 0.13 0.18 0.11 and a surface barrier. This reOMe 1.62 1.47 1.26 1.34 0.89 0.65 0.77 search is of significant novelty S/G[c] 2.89 1.34 0.47 0.56 – – – degree of condensation 0.02 0.23 0.52 0.42 0.48 0.64 0.69 and importance to elucidate the mechanisms of lignin chemistry [a] Au = autohydrolysis. [b] Amount: per Ar. [c] The syringyl-to-guaiacyl units of lignin (S/G) ratios in the original wood of eucalyptus, red maple, and loblolly pine were previously characterized as 2.73, 1.27, and 0, respecon enzymatic hydrolysis. tively.

Results and Discussion Characterization of MWLs The yield and lignin content of all MWLs are shown in Table S1 in the Supporting Information. MWL has been traditionally used as a representative source of original lignin in the substrates due to mild preparation conditions that maximally preserve the structures of lignin.[8] Although the yield of MWL is lower than that of cellulolytic enzyme lignin (CEL), MWL is considered an appropriate lignin for this study because CEL contains a significant amount of residual carbohydrates that may interact with added enzymes for the enzyme adsorption test.[6] The yield of purified MWL was a limiting factor, which depended on the extent of ball milling. Although the increase in the milling time increases the yield of MWL, it could result in a higher degree of lignin degradation.[8] Quantitative 13C NMR spectroscopy is recognized as the most-used NMR spectroscopy method for lignin characterization. This technique is informative, reliable, and, at the same time, relatively feasible.[9] The advantage of NMR spectroscopy over other spectroscopic techniques, such as IR, UV/Vis, and Raman spectroscopy, is that NMR spectroscopy has much higher resolution, which enables a larger amount of information to be obtained.[9] A quantitative 13C NMR spectroscopy example of nonacetylated and acetylated pine MWL is shown in Figure S1 in the Supporting Information. For MWLs obtained from SW, the integral of the d = 162–102 ppm region was set as the reference, assuming that it includes six aromatic carbon atoms and 0.12 vinylic carbon atoms.[10] It follows that the integral value divided by 6.12 is equivalent to one aromatic ring. For MWLs obtained from HW, the reference of integration at d = 160–102 ppm includes six aromatic carbon atoms because the amount of vinyl carbon atoms in cinnamyl alcohol and cinnamaldehyde in HW MWL is very limited; thus the contribution of vinyl carbon atoms can be neglected.[11] A very small resonance of C-1, representing carbohydrates, was observed at d = 102–90 ppm; this indicates that carbohydrates do not interfere with the analysis of lignin moieties in MWL.  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

total OH groups include primary aliphatic OH, secondary aliphatic OH, and phenolic OH. The significant decrease of primary and secondary aliphatic OH groups was the major reason for the overall decrease in total OH groups. A carboxylic group increased hydrogen bonding between lignin and the enzyme.[12] MWLs obtained from SW substrates generally possessed higher carboxylic groups than that of HW. However, due to the limited amount of carboxylic groups in MWLs, their impact on lignin–enzyme interactions was not clear. The content of methoxyl groups also decreased after pretreatment, and the S/G ratio of pretreated HW significantly decreased; this might be due to the complete loss of both OMe groups of the S unit or the conversion of S to G units by losing one of the OMe groups. The degree of condensation for MWLs of SW and HW were calculated in different ways.[9, 11] For MWL of SW, the degree of condensation is denoted as (3.00 h units) [(I125 103)na + M + 2  I] (h units = p-hydroxyphenyl moieties; M = vanillin moiety). I125 103 is the region of d = 125–103 ppm attributed to aromatic methine carbon atoms (CAr H). The theoretical value for CAr H in noncondensed guaiacyl units is 3.00, and the difference between it and the integral at I125 103 ppm is usually considered as the degree of condensation. However, some corrections should be made. For example, only two carbon atoms from p-hydroxyphenyl units resonate in this region. The signal of C-6 in vanillin moieties (M) is at d = 126 ppm. The chemical shifts of C-6 and C-5 in spirodienone (I) moieties are also higher than d = 125 ppm. In the case of HW, the degree of condensation is defined as (2 s + 3 g + 2 h) (I125 103) (s = syringyl moiety, g = guaiacyl moiety, h = p-hydroxyphenyl moiety). The theoretical amount of CAr H atoms can be calculated from the h/g/s ratio by considering the contribution of two carbon atoms of the s and h units and three carbon atoms of the g units in this region. The amount of tertiary aromatic carbon atoms (CAr H) in the lignin preparation was obtained from the integral at d = 125–103 ppm. Generally, MWLs of SW had a greater degree of condensation than that of MWLs of HW. Because G-type lignin moieties contain more non-condensed ChemSusChem 0000, 00, 1 – 10

&2&

These are not the final page numbers! ÞÞ

CHEMSUSCHEM FULL PAPERS carbon, it has greater chance of undergoing condensation. The major decrease of OH groups was due to a decrease in primary and secondary OH; thus carbon bonds of CAr C-a, CAr C-b, or CAr C-g were likely to be formed during pretreatments. The impact of lignin structure on enzyme adsorption and enzymatic hydrolysis is discussed in the following paragraphs.

Cellulase adsorption on MWLs and bleached pulps The adsorption of cellulase on lignin is an important consideration to understand lignin inhibition. The adsorption kinetics of cellulase was conducted on all MWLs, bleached hardwood (BHW) pulps, and bleached softwood (BSW) pulps, respectively (Figure 1). All materials reached their maximum adsorption in 30 min. BSW and BHW adsorbed almost half of the initially loaded cellulase. MWLs of pine and maple had similar adsorption performance, which showed about 18 % of the total cellulase adsorption, whereas MWL of eucalyptus adsorbed about 10 % of cellulase. The adsorption of cellulase on lignins from the untreated wood was much lower than that on bleached pulps. However, after pretreatment, the adsorption of cellulase on all of the MWLs increased dramatically. The cellulase ad-

Figure 1. Adsorption kinetics of cellulase on MWLs and bleached pulps. a) MWLs isolated from untreated woods and b) MWLs isolated from pretreated wood pulps.

 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemsuschem.org sorptions on MWLs of all the pretreated wood pulps were in the range of 40 to 68 %. MWL of GL-SW adsorbed the most cellulase (68 %), even more than that of bleached pulps. MWL of GL-HW adsorbed the least amount of cellulase (about 40 %), compared with the MWLs of the other pretreated wood pulps. The significant increase in the adsorption of cellulase on lignin isolated from pretreated pulps might be associated with changes of the lignin structures by pretreatment. As shown in Figure 2 a, cellulase adsorption increased with the increasing degree of condensation of lignin. A greater degree of condensation of lignin might lead to a more branched lignin structure, which probably results in more interactions with the enzyme. Also, it is likely that more enzymes might be trapped in the network structure of lignin due to the increased degree of condensation of lignin. The increase of cellulase adsorption was

Figure 2. The effect of characteristics of MWLs such as a) degree of condensation and b) hydroxyl content on cellulase adsorption.

ChemSusChem 0000, 00, 1 – 10

&3&

These are not the final page numbers! ÞÞ

CHEMSUSCHEM FULL PAPERS also related to the decrease in total OH. The major contributors to the changes of total OH were aliphatic OH groups; this indicated that the condensation reaction occurred during pretreatment. Alternatively, the increase of phenolic OH groups was accompanied by the increase of cellulase adsorption (Figure 2 b). Several studies mentioned that phenolic OH groups could be involved in the adsorption of cellulase to lignin.[13] Sewalt and co-workers compared the effect of hydroxypropylated organosolv lignin and steam-pretreated lignin on filterpaper digestion.[13a] They found that the hydroxypropylation of lignin with free phenolic sites resulted in increased cellulose hydrolysis. Pan reported that phenolic hydroxyl groups were critical to their inhibitory effects on enzymatic hydrolysis of cellulose.[13b] At a concentration of 10 mm, phenolic compounds showed 1–5 % more inhibition than nonphenolic ones. The inhibitory effect of lignin could be significantly removed by using hydroxypropylation to block free phenolic hydroxyl groups. Due to the correlation between OH groups and cellu-

www.chemsuschem.org lase adsorption, it was speculated that the changes in OH were associated with changes to the hydrophobicity of lignin. Impact of lignin chemistry on enzymatic hydrolysis The impact of lignin chemistry on enzymatic hydrolysis was studied by enzymatic hydrolysis of physically mixed substrate with 80 % bleached pulps and 20 % MWLs, as shown in Figure 3 a and b. It appeared that the presence of 20 % MWLs obtained from untreated woods and physically mixed bleached pulps had no impact on enzymatic hydrolysis because it was evidenced that the hydrolysis kinetics of bleached pulps mixed with MWLs of eucalyptus, maple, and pine were mostly the same as that of bleached pulps alone (Figure 3 a). MWLs isolated from untreated woods adsorbed much less cellulases than bleached pulps (Figure 1 a); therefore, the impact of lignin chemistry of MWLs obtained from untreated woods was limited for enzymatic hydrolysis.

Figure 3. Enzymatic hydrolysis of physically mixed and reconstructed lignocellulosic pulps. a) Enzymatic hydrolysis of physically mixed pulp and MWLs isolated from untreated woods. b) Enzymatic hydrolysis of physically mixed pulp and MWLs isolated from pretreated wood pulps. c) Enzymatic hydrolysis of reconstructed substrate using bleached pulps and MWLs obtained from untreated woods. The lignin contents in the reconstructed eucalyptus, maple, and pine were 21.1, 21.5, and 21.6 %, respectively. d) Enzymatic hydrolysis of reconstructed substrate using bleached pulps and MWLs obtained from pretreated woods. The lignin content in reconstructed Au-HW, Au-SW, GL-HW, and GL-SW were 20.6, 22.3, 21.5, and 21.9 %, respectively.

 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemSusChem 0000, 00, 1 – 10

&4&

These are not the final page numbers! ÞÞ

CHEMSUSCHEM FULL PAPERS In contrast, the presence of 20 % MWLs obtained from pretreated biomass showed a significant impact on enzymatic hydrolysis (Figure 3 b) and a 4–13 % decrease of 96 h carbohydrate conversion was observed. Particularly, the MWL of GLSW, which had a potential to adsorb up to 68 % of loaded cellulase, caused the 96 h carbohydrate conversion to decrease from 92.8 to 80.5 %. This is the most severe decrease for the hydrolysis of physically mixed substrate. Thus, enzymatic hydrolysis was largely affected by the enzyme adsorption properties of lignin. When the lignin alone was capable of adsorbing a similar amount of cellulase to that of bleached pulps, the presence of 20 % MWLs showed a noticeable impact on the enzymatic hydrolysis of physically mixed substrate (Figure 3 b). It had been suggested that the extent to which lignin adsorbed enzymes substantially depended on the nature of lignin itself, and lignin adsorption of enzyme decreased as the severity of pretreatment increased.[6] It was reported that the lignin isolated from lodgepole pine and steam-pretreated poplar decreased the hydrolysis yields of Avicel, whereas the other isolated lignins did not appear to decrease the hydrolysis yields significantly; this was likely to be because of the different adsorption properties of the isolated lignins.[6]

Impact of a physical barrier of lignin on enzymatic hydrolysis The effect of a physical barrier of lignin on enzymatic hydrolysis was revealed by the hydrolysis of reconstructed pulps (Figure 3 c and d). The reconstruction procedure had a minimal effect on the carbohydrate substrate, since the hydrolysis kinetics of BHW and BSW subjected to reconstruction without the addition of lignin had no difference on hydrolysis kinetics of the original BHW and BSW (data not shown). The buildup of a physical barrier of lignin in reconstructed lignocellulose was confirmed by the image taken by confocal microscopy (Figure 4). After reconstruction, lignins were observed to deposit on the surface of the fiber and primarily in the lumen.

www.chemsuschem.org This artificial biomass generated by the reconstruction process separated the issue of lignin distribution from lignin chemistry. Regardless of the lignin type used during reconstruction, lignin precipitated in the lumen and, to a lesser extent, on the outer surface. Thus, this is an ideal substrate to study the impact of lignin chemistry and physical barrier on enzymatic hydrolysis. Compared with the hydrolysis of physically mixed substrate, the reconstructed lignocellulosic substrates were more resistant to hydrolysis. For the reconstructed HW with MWLs obtained from untreated woods, the 96 h carbohydrate conversion dropped from 90.4 % to 83.2 and 78.4 % for the hydrolysis of reconstructed eucalyptus and reconstructed maple, respectively. The reconstructed pine showed a more pronounced decrease in the efficiency of enzymatic hydrolysis. At the end of 96 h of hydrolysis, the reconstructed pine pulps ended up with 71.3 % carbohydrate conversion, whereas the carbohydrate conversion of BSW was 88.9 % (Figure 3 c). The order of the decrease of carbohydrate conversion in reconstructed pulps followed the decrease in the S/G ratio of MWLs isolated from untreated woods (Table 1). With the decreasing of S/G ratio, the reconstructed pulps became more difficult for enzymatic digestion. It has been speculated that the guaiacyl unit of lignin formed a more cross-linked lignin structure than that of the syringyl unit of lignin (Table S2 in the Supporting Information), and thus, yielded a greater physical barrier against the access of enzymes.[14] In the case of reconstructed substrate with MWLs obtained from pretreated woods, over the course of 96 h of hydrolysis, the carbohydrate conversion decreased from over 90 % to 83.3 % when 20 % MWL of GL-HW was reconstructed with BHW pulps. For the other reconstructed substrates, the 96 h carbohydrate conversion decreased to a greater degree: 74.9, 64.3, and 47.1 % for Au-HW, Au-SW and GL-SW, respectively (Figure 3 d). For the same type of pretreatment, the reconstructed SW substrates always had lower carbohydrate conversion than those of reconstructed HW substrates. This trend was consistent with the observation in the case of hydrolysis of reconstructed lignocellulose with MWLs obtained from untreated wood. The order of decreased hydrolysis of the reconstructed substrates generally agrees with the increase trend of the adsorption capability of the corresponding MWLs. As shown in Figure 3 b and d, GL-SW had a stronger inhibition to enzymatic hydrolysis than that of Au-SW, whereas GLHW had a weaker inhibition than that of Au-HW. This result indicated that, although the same pretreatment method was applied, the extent of changes in lignin characteristics, which could be critical factors for enzymatic hydrolysis depended on the species, such as HW and SW. Notably, MWL does not fully represent the original whole lignin in wood or pretreated biomass. However, carbohydrate conversion is closely correlated with the measured lignin characteristics, as discussed in the next section. Lignin characteristics that inhibit enzymatic hydrolysis

Figure 4. A confocal image showing the lignin distribution in the reconstructed pine substrate.

 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

There are two factors of lignin that have a critical impact on enzymatic hydrolysis. As shown in Figure 5, for both physically ChemSusChem 0000, 00, 1 – 10

&5&

These are not the final page numbers! ÞÞ

CHEMSUSCHEM FULL PAPERS

www.chemsuschem.org structed substrate, lignins were precipitated on the carbohydrate; thus lignin formed a physical barrier. This physical barrier could block the access of enzyme to the carbohydrate surface. If enzymes were not accessible to the carbohydrate, they could adsorb on the lignin surface. Therefore, the lignin barrier enhanced the impact of lignin chemistry on limiting enzymatic hydrolysis. As a result, the impact of lignin chemistry and lignin structure was amplified in the case of the reconstructed substrate. In the real lignocellulosic biomass, the lignins distribute in each cell wall layer and middle lamella.[15] The more spreadable distribution of lignin is likely to cover more surface of the carbohydrate and lead to a greater obstacle to the accessibility of enzyme to carbohydrates. For example, as shown in Figure S2 in the Supporting Information, the pretreated biomasses (GL-HW and GL-SW, no reconstruction) with lignin contents of 21.5 and 26.5 % displayed 28.2 and 25.6 % lower carbohydrate conversion for 96 h enzymatic hydrolysis, respectively, than that of reconstructed pulps.

Conclusions

Figure 5. a) The effect of degree of lignin condensation on 96 h carbohydrate conversion and b) the effect of total hydroxyl group content on 96 h carbohydrate conversion.

mixed and reconstructed pulps, the yield of enzymatic hydrolysis decreased with increase of degree of condensation and the decrease of total hydroxyl groups for the samples tested. Once the degree of condensation reached 0.4, the further decrease in the degree of condensation of lignin had a minimal effect on enzymatic hydrolysis. The larger value of the degree of condensation implied that the CAr H bonds in lignin were converted into C C, CAr CAr, or CAr O bonds. This finding of a correlation between the same category of lignin structure in several different lignin samples and the carbohydrate conversion of enzymatic hydrolysis is reported for first time. Moreover, these two factors play a more significant role in limiting the enzymatic hydrolysis of reconstructed pulps. It was considered that, in the physically mixed substrate, lignin and bleached pulp were not directly linked together, so that enzymes could either adsorb on carbohydrate or on lignin. However, in the recon 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

The lignin-derived inhibition during enzymatic hydrolysis was investigated by using a new method, that is, biomass reconstruction, which could decouple the factors of a physical barrier of lignin from lignin chemistry. Lignins isolated from pretreated woody biomass adsorbed two- to sixfold more cellulase than lignins isolated from untreated woods; this indicated that inhibition during enzymatic hydrolysis was greater with lignins obtained from pretreated woody biomass. These lignins adsorbed a similar amount of cellulase to that of bleached pulps, and thus, the presence of 20 % lignins in the physically mixed substrate had a negative impact on enzymatic hydrolysis. The carbohydrate conversion of reconstructed lignocelluloses with about 20 % lignin decreased significantly relative to the hydrolysis of bleached pulps. The reconstructed softwood (SW) substrate had a lower carbohydrate conversion than that of the reconstructed hardwood (HW) substrate. The degree of condensation of lignin increased significantly after pretreatment, especially with SW lignins. The degree of condensation of lignin and total OH groups in lignin had an important impact on cellulase adsorption and enzymatic hydrolysis.

Experimental Section Materials Four types of wood chips (Eucalyptus globulus; red maple; loblolly pine; and mixed HW, which included oak, maple, poplar, and sweet gun) and two types of fully bleached pulps (BHW and BSW) used in this study were obtained from a mill in the southeastern USA. Commercial enzyme preparations of cellulase (NS50013), xylanase (NS50014), and b-glucosidase (NS50010) were kindly provided by Novozymes (Novozymes, Franklinton, NC, USA).

Pretreatments of woody biomasses GL and Au pretreatments, as representatives of mild alkaline and acidic pretreatments, respectively, were conducted on both HW ChemSusChem 0000, 00, 1 – 10

&6&

These are not the final page numbers! ÞÞ

CHEMSUSCHEM FULL PAPERS (mixed HW) and SW (loblolly pine). Both pretreatments were performed in an M/K bath digester (M/K system Inc., Danvers, MA, USA). GL was a mixture of sodium carbonate and sodium sulfide with a sulfidity of 40 %. The total titratable alkali (TTA) charge was 16 and 20 % for HW and SW, respectively. Wood chips (600 ovendried (od) g) were used in each batch and the ratio of GL to wood chips was 4:1 (v/w). The pretreatment conditions for SW were an H factor of 800 at 170 8C and an H factor of 400 at 160 8C for HW. After cooking, the solids were washed overnight with running tap water to remove residual chemicals and dissolved wood components completely. For the Au pretreatment, the same amount of wood chips (600 god od = oven dry) was loaded and the water to solid ratio was 4:1 (v/w). The pretreatment was conducted at 180 8C target temperature for 1 h and it took 45 min to reach the target temperature. After cooking, the remaining solid residues were washed under running tap water for at least 2 h and then soaked in water for 8 h. The resulting residual chips were then centrifuged to remove excessive water and obtain a uniform moisture content. After GL and Au pretreatment, the pretreated chips were fiberized by using a Bauer 148-2 disk refiner twice with a disk gap of 0.25 and 0.05 mm. After that, the wet pulp cake was applied to a pulp breaker to produce uniform and fluffy pulp.

Preparation of MWLs MWLs were isolated from seven different substrates: three un-pretreated wood (eucalyptus, maple, pine) and four pretreated wood (GL-HW, Au-HW, GL-SW, Au-SW), according to the procedure described in Figure 6. The wood chips and pulps were ground to pass a 20-mesh screen and then extracted with 1:2 (v/v) ethanol/ benzene for 24 h to remove the extractives. The ball milling of extractive free sawdust was conducted on a planetary micro mill pulverisette 7 classic line system (Fritsch, Idar-Oberstein, Germany). Each sample (2 od g) was milled at a speed of 600 rpm in a silicon nitride jar (45 mL) by 17 zirconium dioxide grinding balls. The milling time for eucalyptus and maple was 6 h, and for loblolly pine was 8 h. The pretreated pulps were milled for shorter times: 4 h for pretreated HW and 6 h for pretreated SW. Out of every 30 min of

www.chemsuschem.org milling, 15 min suspension was applied to prevent overheating. After ball milling, lignins in milled samples were extracted three times with 96 % (v/v) dioxane at 50 8C for 24 h each time. The solution was collected by centrifugation and concentrated for freeze drying. The crude MWL, isolated by freeze-drying, was dissolved in 90 % (v/v) acetic acid and then precipitated into 10 times of water. The purified MWLs were finally obtained by washed, centrifugation, and freeze-drying of the isolated MWLs. The total lignin content of all samples were determined by using a modified Klason lignin method derived from the TAPPI standard method T222pm-88, as previously described.[2b]

NMR spectroscopy analysis 13

C NMR spectroscopy analysis was conducted for both non-acetylated and acetylated (Ac) lignins to obtain overall structural information of MWLs (Figure S1 in the Supporting Information). The Ac MWL was prepared by dissolving MWL (40 mg) in pyridine (0.5 mL) and acetic anhydride (0.5 mL) and stored in a dark place at room temperature for 2 days to allow complete acetylation reaction. The Ac lignin was then recovered by rotary evaporation with ethanol three to five times to remove pyridine completely. MWL sample (40 mg), either MWL or MWL-Ac, was dissolved in [D6]DMSO (180 mL) and chromium(III) acetylacetonate (20 mL, 0.01 m) to obtain a 20 % (w/v) lignin solution. Chromium(III) acetylacetonate was added as a relaxant to provide complete relaxation of all nuclei; the presence of relaxant was assured to have no impact on the quantitative analysis of the spectra.[9] The Shigemi microtube filled with lignin solution was loaded into the NMR spectrometer. NMR spectra were recorded on a Bruker AVANCE 300 MHz spectrometer at 300 K by using [D6]DMSO as the solvent. Chemical shifts were referenced to tetramethylsilane (TMS; 0.0 ppm). A total of 25 600 scans lasting for about 21 h were collected under operating conditions of 908 pulse width, 1.2 s acquisition time, and 1.7 s relaxation delays.[16]

Measurement of cellulase adsorption on bleached pulps and lignins The kinetics of cellulase adsorption were determined by incubating 1 mg mL 1 of cellulase preparation with bleached pulps or MWLs at 1 % consistency for 5, 10, 20, 30, and 60 min at 50 8C with 180 rpm. The protein content of cellulase in the supernatant was measured by ninhydrin assay analysis. Aqueous bovine serum albumin (BSA) was used as a reference standard. Protein solutions (0.1 mL) with a maximum protein concentration of 1 mg mL 1 were mixed with 10 n NaOH (0.3 mL) and autoclaved at 121 8C for 20 min to degrade protein to amino acid completely. After cooling, acetic acid (0.5 mL) and 2 % ninhydrin reagent (0.5 mL) were added and mixed well, and then boiled for 10 min. The samples were then diluted with 99.5 % ethanol (4.2 mL), and centrifuged to remove byproducts and solids prior to UV/Vis analysis at a wavelength of 570 nm.[17]

Preparation of physically mixed lignocellulose and reconstructed lignocelluloses

Figure 6. Preparation of MWLs from raw woods and pretreated wood pulps.

 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

The MWLs obtained from SW were physically mixed or reconstructed with BSW, whereas the MWLs isolated from HW were processed with BHW. Reconstructed lignocellulosic pulps, which were prepared by using bleached pulps and MWLs, were used to mimic the structure of ChemSusChem 0000, 00, 1 – 10

&7&

These are not the final page numbers! ÞÞ

CHEMSUSCHEM FULL PAPERS lignocellulosic biomass. The total lignin content of reconstructed pulps was designed to be 20 %. Considering the empirical efficiency parameter (75–80 %) during reconstruction and the purity of MWLs (about 93 %, as shown in Table S1 in the Supporting Information), the MWL used in this study to prepare 20 % lignin content (0.25 od g) of the reconstructed lignocellulose (1.25 od g) was about 0.358 od g (0.25 g/93 %/75 % = 0.358 g). Based on this calculation, purified lignins (about 0.358 od g) were first dissolved in 0.1 n NaOH (7 mL) to obtain an alkali–lignin solution. Then, bleached pulp (1 od g) was mixed with the alkali–lignin solution and the consistency of the slurry was adjusted to 10 %. The mixture was allowed to stand overnight to allow the penetration of dissolved lignin into bleached pulps. Subsequently, 1 % sulfuric acid (6 mL) was added to the alkali–lignin pulp to precipitate lignin in the bleached pulps. The pH decreased to 2.5. After 1 h, excess acid was removed by filtration, and then the pulp was washed with deionized (DI) water (50 mL) to remove the salt. A control sample was prepared by treating only bleached pulp with this reconstruction method. The lignin content of the reconstructed lignocelluloses was determined by Klason lignin and found to be in the range of 20–22 %.

Determination of lignin distribution in the reconstructed lignocellulose Lignin distribution in reconstructed lignocellulose was characterized by using confocal laser scanning microscopy (CLSM). Lignocellulosic samples were stained with HPLC-grade acridine orange (AO; 3,6-bis(dimethylamino)acridine hydrochloride, Sigma–Aldrich, St. Louis, MO) at room temperature. In a glass vial (25 mL), sample (approximately 200 mg) was immersed in 12.5 mn AO solution (20 mL) prepared with DI water. The vial was sealed and placed in a dark box to avoid light and shaken frequently for 1 h. After staining, the samples were washed three times by carefully pouring the supernatant out of the vial and replacing the volume with DI water. The samples were then placed on a 3 in.  1 in. glass slide, covered with a No.1 coverslip, and sealed with wax. The imaging of dyed samples was performed by using a Carl Zeiss LSM 710 confocal workstation with a C-Apochromat 40  /1.1 water immersion objective lens. An argon laser at l = 488 nm was used as the excitation light source. Fluorescence emission between l = 515 and 540 nm was collected as the green channel and emissions above l = 590 nm were collected as the red channel. Image analysis was performed by using ZEN lite 2011 image analysis software. The images that appear green in color are rich in carbohydrate and areas that are red in color are rich in lignin. When AO interacts with carbohydrates, it remains in a monomeric form that fluoresces and emits light primarily in the green region of the visible-light spectrum. However, when AO interacts with the aromatic p electrons of lignin, the electron density of the molecule changes in such a way that causes other AO molecules to aggregate and this causes a fluorescence emission shift from green to red light.

Enzymatic hydrolysis Enzymatic hydrolysis was performed on physically mixed substrate, including 20 % (w/w) of MWL and 80 % (w/w) of bleached pulp, and reconstructed lignocellulose with about 20 % (w/w) lignin content. In a microcentrifuge tube (2 mL), substrate (0.1 od g) was soaked in sodium acetate buffer (pH 4.8, 100 mm, with 0.3 % sodium azide). The reaction was conducted at 5 % consistency in an air shaker at 50 8C with a shaking speed of 180 rpm. The loading of cellulase was 5 FPU g 1 (FPU = filter paper units) carbohydrate.  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemsuschem.org Xylanase and b-glucosidase were also used to enhance hydrolysis. The weight ratio of cellulase, xylanase, and b-glucosidase was 1:0.3:0.3. Hydrolysis kinetics were obtained by sampling the hydrolysate periodically. The sugar content in the hydrolysate was determined by HPLC (Agilent technology 1200 series, Palo Alto, CA, USA). The column system for HPLC consisted of a Shodex SP0810 column and a deashing cartridge filter precolumn. Liquid samples were filtered (filter = 0.22 mm) before analysis. Milli-Q water was used at a flow rate of 0.5 mL min 1. The hydrolysis yield of the substrate was evaluated by carbohydrate conversion, which was defined as the amount of hydrolyzed carbohydrate in the hydrolysates as a percentage of total carbohydrate in the starting substrate for enzymatic hydrolysis. Two parallel samples were used in all analyses and the data were presented as the mean of the duplicates.

Acknowledgements We are grateful for financial support from the Wood to Ethanol Research Consortium (WERC) and the Department of Energy (Award Number: DE-FG36-08GO88053). In addition, we express sincere appreciation to Novozymes North America for providing enzymes. The skillful assistances of Dr. Hanna Gracz, Dr. Douyong Min, Dr. Zhoujian Hu, and Dr. Ewellyn Capanema in the NMR experiments are greatly appreciated. Keywords: adsorption · biomass · enzymes · hydrolysis · lignin [1] T. Nordlie, UF-Led Consortium Garners $20 Million Grant to Improve Pine Forest Management, to be found under http://news.ifas.ufl.edu/2011/ 02/uf-led-consortium-garners-20-million-grant-to-improve-pine-forestmanagement/, 2011. [2] a) M. J. Taherzadeh, K. Karimi, Int. J. Mol. Sci. 2008, 9, 1621 – 1651; b) Z. Yu, H. Jameel, H.-m. Chang, S. Park, Bioresour. Technol. 2011, 102, 9083 – 9089; c) Y. Xue, H. Jameel, S. Park, BioResources 2012, 7, 602 – 615. [3] G. A. Smook in Handbook for Pulp and Paper Technologists (Ed.: G. A. Smook), Angus Wilde, Vancouver, 2002, pp. 10 – 20. [4] Z. Yu, K.-s. Gwak, H.-m. Chang, S. Park, H. Jameel, The 4th International Conference on Pulping, Papermaking and Biotechnology, Vol. 2 (Nanjing, P.R. China) 2012, pp. 1143 – 1147. [5] Z. Yu, H. M. Chang, S. Park, H. Jameel, 2010 TAPPI PEERS Conference and 9th Research Forum on Recycling, Vol. 2 (Norfolk, USA) 2010, pp. 1555 – 1580. [6] S. Nakagame, R. P. Chandra, J. N. Saddler, Biotechnol. Bioeng. 2010, 105, 871 – 879. [7] a) A. Berlin, M. Balakshin, N. Gilkes, J. Kadla, V. Maximenko, S. Kubo, J. Saddler, J. Biotechnol. 2006, 125, 198 – 209; b) T. W. Jeffries, Biodegradation 1990, 1, 163 – 176; c) S. Nakagame, R. P. Chandra, J. N. Saddler in Sustainable Production of Fuels, Chemicals, and Fibers from Forest Biomass, Vol. 1067 (Eds.: J. Zhu, X. Zhang, X. Pan), American Chemical Society, Washington DC, 2011, pp. 145 – 167; d) X. Pan, D. Xie, N. Gilkes, D. Gregg, J. Saddler, Appl. Biochem. Biotechnol. 2005, 124, 1069 – 1079; e) J. L. Asensio, A. Ard, F. J. CaÇada, J. Jimnez-Barbero, Acc. Chem. Res. 2013, 46, 946 – 954. [8] T. Ikeda, K. M. Holtman, J. F. Kadla, H. M. Chang, H. Jameel, J. Agric. Food Chem. 2002, 50, 129 – 135. [9] E. A. Capanema, M. Y. Balakshin, J. F. Kadla, J. Agric. Food Chem. 2004, 52, 1850 – 1860. [10] C.-L. Chen in Lignin and Lignan Biosynthesis, Vol. 697 (Eds.: N. G. Lewis, S. Sarkanen), ACS Publications, Washington DC, 1998, pp. 255 – 275. [11] E. A. Capanema, M. Balakshin, J. F. Kadla, J. Agric. Food Chem. 2005, 53, 9639 – 9649. [12] W. S. Brey, Physical Chemistry and Its Biological Applications, Academic Press, New York, 1978.

ChemSusChem 0000, 00, 1 – 10

&8&

These are not the final page numbers! ÞÞ

CHEMSUSCHEM FULL PAPERS [13] a) V. J. H. Sewalt, W. G. Glasser, K. A. Beauchemin, J. Agric. Food Chem. 1997, 45, 1823 – 1828; b) X. Pan, J. Biobased Mater. Bioenergy 2008, 2, 25 – 32. [14] J. H. Grabber, Crop Sci. 2005, 45, 820. [15] J. Fromm, B. Rockel, S. Lautner, E. Windeisen, G. Wanner, J. Struct. Biol. 2003, 143, 77 – 84. [16] T. Tammelin, M. sterberg, L.-S. Johansson, J. Laine, Nord. Pulp Pap. Res. J. 2006, 21, 444 – 450.

 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemsuschem.org [17] a) Z. Zhu, N. Sathitsuksanoh, Y. H. Percival Zhang, Analyst 2009, 134, 2267 – 2272; b) Z. Yu, H. Jameel, H.-m. Chang, R. Phillips, S. Park, Bioresour. Technol. 2013, 147, 369 – 377.

Received: January 11, 2014 Revised: March 18, 2014 Published online on && &&, 0000

ChemSusChem 0000, 00, 1 – 10

&9&

These are not the final page numbers! ÞÞ

FULL PAPERS Z. Yu, K.-S. Gwak, T. Treasure, H. Jameel, H.-m. Chang, S. Park* && – && Effect of Lignin Chemistry on the Enzymatic Hydrolysis of Woody Biomass

 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Breaking it down: A new method called biomass reconstruction is introduced to study the effect of lignin chemistry on enzymatic hydrolysis. Hydrolysis is inhibited by increasing degree of condensation and decreasing total hydroxyl content of lignin; this may be a reason for the high recalcitrant nature of softwood substrate.

ChemSusChem 0000, 00, 1 – 10

&10&

These are not the final page numbers! ÞÞ

Effect of lignin chemistry on the enzymatic hydrolysis of woody biomass.

The impact of lignin-derived inhibition on enzymatic hydrolysis is investigated by using lignins isolated from untreated woods and pretreated wood pul...
2MB Sizes 2 Downloads 3 Views