BioEnergy Research

, Volume 8, Issue 3, pp 1391–1400

Incorporation of Flavonoid Derivatives or Pentagalloyl Glucose into Lignin Enhances Cell Wall Saccharification Following Mild Alkaline or Acidic Pretreatments

  • John H. Grabber
  • Nicholas Santoro
  • Cliff E. Foster
  • Sasikumar Elumalai
  • John Ralph
  • Xuejun Pan
Article

DOI: 10.1007/s12155-015-9605-2

Cite this article as:
Grabber, J.H., Santoro, N., Foster, C.E. et al. Bioenerg. Res. (2015) 8: 1391. doi:10.1007/s12155-015-9605-2

Abstract

Partial substitution of normal monolignols with phenolic precursors from other metabolic pathways may improve the susceptibility of lignified biomass to chemical pretreatment and enzymatic saccharification for biofuel production. Flavonoids and gallate esters readily undergo oxidative coupling reactions, suggesting they could serve as alternate monomers for forming lignin in plants. To test this premise, primary cell walls of Zea mays (L.) were artificially lignified with normal monolignols plus various flavan-3-ol/phenolic ester derivatives, flavonol glycoside/gallate ester derivatives, or pentagalloyl glucose added as 0 or 45 % of the precursor mixture. Most alternate monomers readily copolymerized with normal monolignols, but wall-bound lignin was most efficiently formed with epicatechin, epicatechin gallate, epigallocatechin gallate, or hyperoside. Yields of glucose from a high-throughput digestibility platform were used to examine how lignin modifications affected the susceptibility of cell walls to enzymatic hydrolysis following alkaline or acidic pretreatments of different severities. With the exception of hyperoside, incorporation of alternate monomers into lignin improved yields of enzymatically released glucose by 18–60 % after mild alkaline pretreatment and by 6–34 % after mild acid pretreatment. Responses due to lignin modification diminished as pretreatment severity increased. Overall, our results suggest that apoplastic deposition of pentagalloyl glucose or gallated flavan-3-ols such as epicatechin gallate or epigallocatechin gallate for incorporation into lignin could be promising plant genetic engineering targets for improving sugar yields from grass biomass crops that are subjected to low-temperature alkaline pretreatments.

Keywords

Monolignols Genetic engineering Pretreatment Enzymatic hydrolysis Cellulosic biofuel 

Introduction

Biofuels produced via the fermentation of sugars released enzymatically from cellulose and hemicelluloses in plant cell walls could reduce greenhouse gas emissions, improve energy security, and strengthen rural economies in the USA and other countries [1]. Unfortunately, yields of fermentable sugars from biofuel crops are inherently low because lignin within the plant cell wall forms a hydrophobic barrier that hinders the penetration and action of saccharifying enzymes. As a result, cellulose and hemicelluloses must be liberated from lignin by costly chemical pretreatments, often with alkali or acid at high temperature, prior to enzymatic saccharification [2, 3, 4].

To improve the economic feasibility of biofuel production, numerous genetic engineering, mutagenesis, and breeding efforts with biomass crops are underway to modify lignin concentration or composition so that cell walls will be more amenable toward pretreatment and saccharification [5, 6, 7]. Similar efforts are underway to improve the pulping of woody crops for paper production and to enhance the inherent digestibility of forage crops for livestock production [8, 9, 10]. Work in this area suggest that lignification in plants is highly malleable, permitting dramatic shifts in the quantity and proportions of both normal monolignols and atypical precursors incorporated into the lignin polymer [11, 12]. These findings support the notion that plants could be genetically engineered to make use of precursors from alternate phenolic pathways to form lignins that are more amenable to processing [13, 14, 15].

To help focus lignin bioengineering efforts, we are employing a biomimetic model based on primary cell walls of maize to test a variety of plant-derived phenolics for their potential as alternate monomers for lignification. Previous studies suggest that this model successfully mimics the natural lignification process and the effect of lignin on the enzymatic degradation of both primary and secondary walled tissues in grasses [16, 17]. This biomimetic model is therefore a valuable tool for testing the compatibility of alternate monomers with lignification and their effects on cell wall utilization prior to molecular biologists expending considerable resources to engineer plants to produce the desired type of modified lignin. It is of course recognized that studies with bioengineered plants will ultimately determine the commercial utility of modified lignins and their suitability for plant growth, development, and/or resistance to various biotic and abiotic stresses.

Previous model studies demonstrated that the copolymerization of normal monolignols with a structurally related phenylpropanoid conjugate, coniferyl ferulate, dramatically improved the alkaline extractability of lignin and the subsequent enzymatic hydrolysis of cell walls [18]. Subsequent high profile in planta studies confirmed the value of genetically engineering coniferyl ferulate incorporation into lignin as a means of enhancing the enzymatic saccharification of chemically pretreated cell walls [19]. In more recent model work, incorporation of more diverse polyphenolic monomers such as epigallocatechin gallate or rosmarinic acid into lignin enhanced the inherent fermentability of cell walls by rumen microflora and their susceptibility to alkaline pretreatment and enzymatic saccharification [20, 21, 22]. The current study builds on previous work with epigallocatechin gallate to examine how incorporation of pentagalloyl glucose or various flavonoids with differing structural features (e.g., flavan-3–ol vs. flavonol backbones, varying degrees of phenolic hydroxylation, and phenolic ester substitution) into lignin affects the enzymatic saccharification of cell walls following alkaline as well as acid pretreatments. Most of the alternate monomers examined in the study are found in high natural abundance in one or more plant species, which should facilitate cloning of their biosynthetic genes for expression in biomass crops.

Materials and Methods

Lignification and Chemical Characterization of Cell Walls

As previously described [20], freshly isolated primary cell walls (~2.2–3.4 g dry weight) from maize containing wall-bound peroxidases were stirred in water and artificially lignified over a 14-h period by adding separate solutions of monomers and dilute H2O2. An equimolar mixture of coniferyl alcohol (1 or 1.3 mmol) and sinapyl alcohol (1 or 1.3 mmol), which are the primary monolignols in grasses and other angiosperms, were added to form low and normal lignin control cell walls. Lignification was also carried out by adding three-component mixtures of coniferyl and sinapyl alcohols with the following alternate monomers (Fig. 1): epicatechin, epigallocatechin, epicatechin gallate, epigallocatechin gallate, epicatechin vanillate, hyperoside, 2″-O-galloylhyperin, or pentagalloyl glucose. The glycosylated forms of quercetin flavonol/gallate esters (hyperoside and galloylhyperin) were used because of their greatly improved solubility in aqueous media. Ethyl gallate and another gallate ester of glucose, corilagin, were previously tested, but not included in the current study because they were poorly incorporated into wall-bound lignin [20]. When added, alternate monomers comprised about 45 % by weight of the total monomers added to cell walls, potentially yielding a shift in lignin composition within the range reported for mutant or transgenic plants with altered lignin biosynthesis [23, 12, 24]. The studies were also designed so that the potential amount of lignin formed from conventional plus alternate monomers (excluding non-phenolic glycoside units) would equal that of the normal lignin control. Nonlignified control cell walls were stirred in a solvent mixture similar to the final makeup of the reaction media used for lignifying cell walls. The aforementioned treatments were distributed between two experiments according to a randomized complete block design and each experiment was independently replicated twice in time using cell walls prepared from different batches of maize cell suspensions. The various controls and the treatment with epigallocatechin gallate were included in both experiments to facilitate treatment comparisons between experiments.
Fig. 1

Coniferyl alcohol and sinapyl alcohol are the primary monolignols used by angiosperms to form lignin. In our studies, we examined how partial substitution of normal monolignols with flavan-3-ol, flavan-3-ol/phenolic ester derivatives, glycosylated quercetin flavonol/gallate esters, or pentagalloylglucose influenced the susceptibility of maize cell walls to chemical pretreatment and enzymatic saccharification

Following lignification, cell walls were collected on glass-fiber filters, washed sequentially with water, acetone-water (9:1, v/v), and acetone to remove non-incorporated dehydrogenation products and water. Monomer incorporation was assessed by drying subsamples of the acetone-water filtrates in vacuo, redissolving dehydrogenation products in dioxane-water (1:1, v/v), and reading the absorbance at 280 nm with a spectrophotometer. In a similar manner, the absorbance of a reference dehydrogenation polymer formed in vitro from an equimolar mixture of coniferyl and sinapyl alcohols was read to provide a quantitative estimate of monomer incorporation into lignified control cell walls. Cell walls were transferred to sample jars, set overnight in a hood to allow the acetone to evaporate, and dried in an oven at 55 °C. Oven-dried lignified and nonlignified cell walls were weighed to estimate, by difference, the quantity of lignin formed in cell walls. This mass balance approach provides a semiquantitative estimate of lignin concentration because primary maize cell walls have a very low density (~50 mg cm−3) and are difficult to fully recover throughout the multistep workup process.

Duplicate subsamples (75 mg) of cell walls were dissolved in 12 M H2SO4 at 25 °C for 2 h and then hydrolyzed by 1.6 M H2SO4 at 100 °C for 3 h for colorimetric analysis of uronosyls [25] and gravimetric analysis of acid-insoluble Klason lignin [26]. Utilizing high-throughput procedures [27, 28], triplicate subsamples (1 mg) were heated at 50 °C for 3 h with 25 % acetyl bromide in glacial acetic acid for spectrophotometric analysis of lignin; additional samples were also heated at 121 °C for 1.5 h in 2 M trifluoroacetic acid for analysis of noncellulosic neutral sugars as alditol acetates by GC/MS and residual cellulosic glucose was determined by a colorimetric anthrone assay following dissolution in 72 % H2SO4.

Alkaline and Acid Pretreatment and Saccharification of Cell Walls

Pretreatment and enzymatic saccharification of cell walls was carried out using a high-throughput digestibility platform [29]. Briefly, cell walls were weighed (1.5 mg) into triplicate 2 mL centrifuge tubes and heated for 1 h at 60, 90, or 120 °C in 150 μL aqueous solutions containing 75, 150, or 250 mg g−1 NaOH or 150, 250, or 500 mg g−1 H2SO4 on a cell wall basis. Following cooling and pH adjustment with 650 μL of 50 mM acetate buffer (pH 4.8 with 100 mg L−1 sodium azide), enzymatic hydrolysis was carried out in sealed tubes for 20 h at 50 °C in a total volume of 0.8 mL with the following enzymes from Novozymes (activities on a g−1 cell wall basis): NS50013 cellulase (15 filter paper units), NS50010 β-glucosidase (30 cellobiase units), NS50012 enzyme complex (15 fungal β-glucosidase units), and NS50030 xylanase (15 Farvet xylan units). Tubes were then centrifuged at 1500×g for 3 min and 4 μL of the supernatant and 64 μL of glucose oxidase/peroxidase (GOPOD) reagent were pipetted in quadruplicate into 384-well plates to determine d-glucose according to the GOPOD assay (Megazyme, Ireland). Glucose yield from sequential pretreatment and saccharification was expressed as a proportion of glucose originally contained in cell walls. The same enzyme hydrolysis conditions and GOPOD assay were also used to determine glucose yields from cell walls without the use of chemical pretreatments. Previous studies have shown that fungal enzymes can extensively degrade cellulose, hemicelluloses, and pectin from nonlignified primary maize cell walls with monomers comprising −90 % of the glucose released [30, 31]. Hence yields of monomeric glucose were used as a gauge to assess the susceptibility of cell walls to enzymatic saccharification, before and after chemical pretreatment. The digestibility platform also measures monomeric xylose production, but this data was not utilized because monomers comprised only ~10 % of the xylose released; low yields of monomeric xylose can be attributed to fungal enzyme preparations lacking the full array of activities needed to completely degrade highly branched xylans in primary maize walls [30, 31].

Statistical Analyses

Data were subjected to mixed model analyses by running PROC MIXED [32]. In all analyses, lignification treatment, pretreatment temperature, and pretreatment chemical loading and their two-way and three-way interactions were considered fixed effects. Experimental replication was also considered a fixed effect, but all two-, three-, and four-way interactions of experimental replication with lignification treatment, pretreatment temperature, and pretreatment chemical loading were considered random effects. If F tests were significant (P ≤ 0.05), then least square means of fixed effects were compared at P = 0.05 using t tests performed by a SAS pdmix800 macro [33]. Least square means of nonlignified controls were also compared to all lignified treatments at P = 0.05 using approaches described by Piepho et al. [34]. Unless noted otherwise, treatment differences described in the text were significant at P = 0.05.

Results and Discussion

Compositional Analysis and Cell Wall Lignification

Averaged across both experiments, nonlignified cell walls prepared from maize cell suspensions contained 218 mg g−1 of cellulosic glucose, 58 mg g−1 of noncellulosic glucose, 232 mg g−1 of arabinose, 180 mg g−1 of xylose, 108 mg g−1 of uronosyls, 86 mg g−1 of galactose, 8 mg g−1 of rhamnose, 4 mg g−1 of mannose, and 2 mg g−1 of fucose. As reported previously [20], the nonlignified cell walls used for these studies also contained an average of 19 mg g−1 of ferulate monomers and dimers [20] and 21 mg g−1 of Klason lignin (Tables 1 and 2). The actual quantity of lignin was probably <5 mg g−1 because covalent attachment and incomplete hydrolysis of ferulates, structural polysaccharides, and protein tends to elevate Klason values [30, 16]. Overall, the composition of nonlignified cell walls was characteristic of primary cell walls of maize [35, 36], with the polysaccharide component composed of about one-quarter cellulose and three-quarters of other noncellulosic polymers, likely highly branched arabinoxylans with smaller amounts of mixed-linked glucans, xyloglucans, and pectins.
Table 1

Lignin and glucose content of primary maize cell walls before and after artificial lignification with binary mixtures of coniferyl alcohol (CA) and sinapyl alcohol (SA) or with trinary mixtures of CA, SA, and various flavan-3-ol and flavan-3-ol/phenolic ester derivatives (experiment 1)

Lignification treatment

Lignin (mg g−1)a

Glucose (mg g−1)

Mass balance

Klason

Acetyl bromide

None (nonlignified control)

26b

58b

280b

CA:SA (low lignin control)

149 b

164 bc

264 bc

CA:SA (normal lignin control)

151 ac

175 a

184 ab

260 c

CA:SA: epicatechin

139 ab

182 a

193 a

263 bc

CA:SA: epigallocatechin

112 b

147 b

142 d

279 a

CA:SA: epicatechin gallate

143 ab

179 a

176 ab

267 b

CA:SA: epigallocatechin gallate

164 a

173 a

166 bc

268 b

CA:SA: epicatechin vanillate

136 ab

153 b

145 cd

270 b

aKlason lignin data previously published [20]

bWithin columns, the least square mean for the nonlignified control differs at P = 0.05 from the overall least square mean of all lignified treatments

cWithin columns, least square means for lignified treatments with unlike letters differ (P = 0.05)

Table 2

Lignin and glucose content of primary maize cell walls before and after artificial lignification with binary mixtures of coniferyl alcohol (CA) and sinapyl alcohol (SA) or with trinary mixtures of CA, SA, and epigallocatechin gallate, hyperoside, galloylhyperin, or pentagalloyl glucose (experiment 2)

Lignification treatment

Lignin (mg g−1)a

Glucose (mg g−1)

Mass balance

Klason

Acetyl bromide

None (nonlignified control)

16b

70b

272b

CA:SA (low lignin control)

162 b

165 c

242 a

CA:SA (normal lignin control)

176 ac

194 a

200 b

236 a

CA:SA: epigallocatechin gallate

173 a

186 a

176 c

245 a

CA:SA: hyperoside

194 a

179 a

222 a

247 a

CA:SA: galloylhyperin

174 a

159 b

211 ab

241 a

CA:SA: pentagalloyl glucose

179 a

160 b

174 c

238 a

aKlason and mass balance lignin data previously published [20]

bWithin columns, the least square mean for the nonlignified control differs at P = 0.05 from the overall least square mean of all lignified treatments

cWithin columns, least square means for lignified treatments with unlike letters differ (P = 0.05)

Normal lignin control cell walls were efficiently lignified with coniferyl and sinapyl alcohols; absorbance readings at 280 nm of combined reaction medium and washes collected after lignification vs. a reference dehydrogenation polymer indicated that ≥96 % of added monomers became incorporated into wall-bound lignin. Semiquantitative mass balance calculations and the commonly used Klason and acetyl bromide lignin assays all indicated that mixtures of normal monolignols with epicatechin, epicatechin gallate, and epigallocatechin gallate also readily formed wall-bound lignin at levels comparable to that in the normal lignin control (Tables 1 and 2). In general, the three lignin methods also indicated that inclusion of epigallocatechin and epicatechin vanillate modestly reduced lignification relative to the normal lignin control (Table 1), but the methods differed in their estimation of how the remaining alternate monomers affected lignification. Relative to the normal lignin control, addition of hyperoside, galloylhyperin, or pentagalloyl glucose had no effect on mass balance estimates of lignin (Table 2). By contrast, acetyl bromide lignin concentrations were increased by hyperoside, not affected by galloylhyperin, and decreased by pentagalloyl glucose, whereas galloylhyperin and pentagalloyl glucose decreased Klason lignin concentrations.

The discrepancy in lignin estimates for cell walls lignified with hyperoside, galloylhyperin, or pentagalloyl glucose (Table 2) was probably related to mass contributed by glycoside units in these monomers and the expected fate of these units during lignin assays. Unlike reactive flavonoid and gallate units, glycoside units do not directly participate in oxidative polymerization reactions; we therefore did not consider the mass of glycoside units when formulating monomer mixtures. Incorporation of glycosylated flavonoids or gallates into lignin would, therefore, lead to relatively high estimates of lignin by the mass balance approach. Relatively high estimates of lignin were also observed with the spectrophotometric acetyl bromide procedure, likely in part because degradation of glycoside units would increase absorbance in the 270–280-nm region used to quantify lignin [37]. Degradation of carbohydrates probably also contributed to unreasonably high estimates of acetyl bromide lignin in nonlignified controls, where very small quantities of lignin have been detected by the Klason procedure or other methods [16]. In addition, reaction of acetyl bromide with lignin containing quercetin units from hyperoside or galloylhyperin should yield ketone products that have a greater absorption coefficient than conventional lignin or modified lignin containing units from flavan-3-ol/phenolic ester derivatives (Fachuang Lu, personal communication, 2014). Finally, acetyl bromide lignin concentrations were based on a standard absorption coefficient (17.75 L g−1 cm−1) for lignin in grasses, which may not hold for lignin formed in part with atypical monomers. By contrast, the Klason procedure gave lower and perhaps more accurate estimates of lignin because glycoside units should be removed during acid hydrolysis. Klason values may, however, be slightly elevated by incorporation of ferulates, structural polysaccharides, and protein into lignin [16]. Conversely, the Klason method can somewhat underestimate concentrations if lignin is partially soluble in acid [38]. Overall, the Klason and acetyl bromide methods gave comparable estimates of lignin in cell walls artificially lignified with normal monolignols or normal monolignols plus flavan-3-ol/phenolic ester derivatives, but the Klason method may give more accurate estimates for cell walls containing only small amounts of lignin or lignin formed in part with glycosylated or ketone-containing monomers.

Enzymatic Hydrolysis of Cell Walls Before and After Pretreatment

We used a high-throughput single tube process for sequentially pretreating and enzymatically digesting samples to assess the availability of fermentable sugars based on the yield of monomeric glucose produced [29]. In the platform, NaOH or H2SO4 at three levels of chemical loading and three heating temperatures were used to help delineate whether alterations in lignin composition affected the susceptibility of cell walls to alkaline or acidic pretreatments and enzymatic saccharification by cellulases amended with a wide range of carbohydrases. Based on compositional assays, the total concentration of cellulosic plus noncellulosic glucose ranged from 236 to 280 mg g−1 of cell wall, with lignified treatments having lower concentrations than nonlignified controls (Tables 1 and 2).

On a mg g−1 glucose basis, enzymatic hydrolysis without pretreatment released an average of 502 mg g−1 of glucose from nonlignified controls and 230 to 318 mg g−1 of glucose from lignified treatments in experiment 1 (Table 3). Partially substituting normal monolignols with epigallocatechin gallate or other flavan-3-ol and flavan-3-ol/phenolic ester derivatives increased glucose yields from lignified cell walls by 16 to 38 %. In contrast, incorporation of epigallocatechin gallate, quercetin flavonol/gallate esters, or pentagalloylglucose in experiment 2 had no effect on glucose yields from lignified cell walls (Table 4). Glucose yields also appeared to be lower in experiment 2, averaging 319 mg g−1 for nonlignified controls and 158 mg g−1 for lignified treatments. Cell walls used in this experiment had relatively high levels of ferulate cross-linking [20], which may account for their lower inherent degradability and the lack of differences among lignification treatments [30, 31]. Previous studies with cell walls from experiments 1 and 2 found that incorporation of epigallocatechin gallate or other flavan-3-ols with gallate ester or pyrogalloyl units substantially increased the inherent degradability lignified cell walls incubated with rumen microflora [20].
Table 3

Klason lignin content and proportion of glucose released by batch pretreatment and enzymatic saccharification of maize cell walls artificially lignified with binary mixtures of coniferyl alcohol (CA) and sinapyl alcohol (SA) or with trinary mixtures of CA, SA, and various flavan-3-ol and flavan-3-ol/phenolic ester derivatives (experiment 1)

Lignification treatment

Klason lignin (mg g−1)

Proportion of enzymatically released glucose (mg g−1 glucose)a

No Pretreatment

NaOH (°C)

H2SO4 (°C)

60

90

120

60

90

120

None (nonlignified control)

26b

502b

855b

808b

740b

662b

835b

904b

CA:SA (low lignin control)

149 bc

254 bc

643 d

706 bc

694 c

317 c

678 a

799 ab

CA:SA (normal lignin control)

175 a

230 c

553 e

667 c

671 d

306 c

618 b

777 abc

CA:SA: epicatechin

182 a

267 b

719 c

740 ab

718 b

379 b

640 b

717 c

CA:SA: epigallocatechin

147 b

318 a

768 ab

723 b

691 c

426 a

702 a

800 ab

CA:SA: epicatechin gallate

179 a

308 a

780 ab

748 ab

719 b

384 ab

678 a

749 bc

CA:SA: epigallocatechin gallate

173 a

304 a

796 a

730 b

715 b

387 ab

676 a

771 abc

CA:SA: epicatechin vanillate

153 b

311 a

759 b

783 a

750 a

396 ab

691 a

826 a

aGlucose released from pretreated cell walls by enzymatic saccharification following aqueous pretreatment at various temperatures with a 150 mg g−1 loading of NaOH or a 250 mg g−1 loading of H2SO4

bWithin columns, the least square mean for the nonlignified control differs at P = 0.05 from the overall least square mean of all lignified treatments

cWithin columns, least square means for lignified treatments with unlike letters differ (P = 0.05)

Table 4

Klason lignin content and proportion of glucose released by batch pretreatment and enzymatic saccharification of maize cell walls artificially lignified with binary mixtures of coniferyl alcohol (CA) and sinapyl alcohol (SA) or with trinary mixtures of CA, SA, and epigallocatechin gallate, hyperoside, galloylhyperin, or pentagalloyl glucose (experiment 2)

Lignification treatment

Klason lignin (mg g−1)

Proportion of glucose released enzymatically (mg g−1 glucose)a

No Pretreatment

NaOH (°C)

H2SO4 (°C)

60

90

120

60

90

120

None (nonlignified control)

16b

319b

815b

774b

718

525b

839b

878b

CA:SA (low lignin control)

162 bc

184 a

503 d

654 b

700 ab

235 a

625 b

776 a

CA:SA (normal lignin control)

194 a

152 a

408 e

596 c

677 b

228 a

553 d

735 abc

CA:SA: epigallocatechin gallate

186 a

149 a

652 b

724 a

710 ab

206 a

601 bc

710 bc

CA:SA: hyperoside

179 a

156 a

483 d

663 b

706 ab

224 a

586 cd

680 c

CA:SA: galloylhyperin

159 b

157 a

623 c

719 a

727 a

234 a

669 a

746 ab

CA:SA: pentagalloyl glucose

160 b

148 a

735 a

745 a

722 a

235 a

665 a

741 abc

aGlucose released from pretreated cell walls by enzymatic saccharification following aqueous pretreatment at various temperatures with a 150 mg g−1 loading of NaOH or a 250 mg g−1 loading of H2SO4

bWithin columns, the least square mean for the nonlignified control differs at P = 0.05 from the overall least square mean of all lignified treatments

cWithin columns, least square means for lignified treatments with unlike letters differ (P = 0.05)

Yields of glucose produced by enzymatic hydrolysis were substantially increased by alkaline or acidic pretreatment of cell walls. Glucose yields were influenced by complex interactions involving lignification treatments, pretreatment types (alkaline vs. acid), pretreatment temperature, and pretreatment chemical loading. Results are illustrated in Fig. 2 for nonlignified and lignified control cell walls and for cell walls lignified with epigallocatechin gallate, which was among the most desirable alternate monomers tested for enhancing glucose yields following pretreatment. With alkaline pretreatments, NaOH loading and temperature both substantially influenced glucose yields from cell walls. For nonlignified controls, all alkaline pretreatments produced high yields of glucose, but the optimal loading of NaOH for maximizing yields declined with increasing temperature. In contrast, normal lignified controls had lower glucose yields under all alkaline pretreatment conditions, but yields were maximized with moderate to high NaOH loadings of 150 to 250 mg g−1 at temperatures of 90 to 120 °C. Cell walls lignified with epigallocatechin gallate had an intermediate response to alkaline pretreatments compared to nonlignified and lignified controls, but glucose yields were maximized at all temperatures with a NaOH loading of 150 mg g−1. These findings are in agreement with a previous study [21], which demonstrated that incorporation of epigallocatechin gallate into lignin substantially enhanced the delignification and enzymatic saccharification of conventionally prepared alkaline-insoluble residues recovered after alkaline pretreatment at 70 to 130 °C with a NaOH loading of 150 mg g−1. These results also corroborate the use of the high-throughput digestibility platform to rapidly assess how alterations in lignin composition affect the susceptibility of cell walls to various chemical pretreatments and enzymatic saccharification. In the case of acid pretreatments, the high-throughput digestibility platform indicated temperature usually had a greater effect than H2SO4 loading or lignification treatment on glucose yields from cell walls. Acid pretreatment at 120 °C maximized glucose yields from nonlignified and lignified controls, but the effects of H2SO4 loading and lignification treatment were more pronounced at 60 and 90 °C (Fig. 2).
Fig. 2

Glucose yields from nonlignified cell walls (top) and cell walls artificially lignified with normal monolignols (center) or normal monolignols plus epigallocatechin gallate (bottom) following pretreatments with NaOH (left) or H2SO4 (right) and enzymatic saccharification. Glucose yields were averaged from experiments 1 and 2

Based on the aforementioned observations, glucose yields from a subset of alkaline and acid pretreatments were used to assess how modified lignin prepared with various epicatechin, quercetin galactoside, or gallate derivatives influenced cell wall saccharification. For alkaline pretreatments, a NaOH loading of 150 mg g−1 at temperatures of 60, 90, and 120 °C were examined because these conditions maximized glucose yields at 120 °C and best delineated yield differences among lignification treatments, particularly at lower temperatures (Fig. 2.). Hence following alkaline pretreatment at 60 °C, lignified cell walls containing epicatechin, epicatechin gallate, epigallocatechin gallate, epigallocatechin, epicatechin vanillate, galloylhyperin, or pentagalloyl glucose yielded 116 to 244 mg g−1 more glucose by enzymatic saccharification than normal or low lignin controls having the same Klason lignin content (Tables 3 and 4). Thus, most of the tested flavonoid and gallate ester derivatives improved glucose yields from lignified cell walls following alkaline pretreatment, but responses were greater if the alternate monomers contained flavan-3-ol or gallate ester units rather than flavonol glycoside or vanillate ester units. Among the alternate monomers tested, epicatechin gallate, epigallocatechin gallate, and pentagalloyl glucose proved the most beneficial for enabling high yields of glucose to be produced from lignified cell walls subjected to mild alkaline pretreatment. This suggests that gallated flavan-3-ols are the most desirable subgroup of flavonoids that could serve as alternate monomers for enhancing the digestibility of lignified cell walls. Results with pentagalloyl glucose suggest that polygalloyl esters may have utility as alternate lignin monomers for enhancing digestibility, but as reported here and in previous work [20], gallate esters appear to be less efficiently incorporated into wall-bound lignin than most flavan-3-ols. As previously described in detail [21, 22], o-diol and especially o-triol, and ester functional groups in gallated flavan-3-ols and pentagalloyl glucose might enhance the delignification and enzymatic saccharification of cell walls by limiting the cross-linking of lignin to structural polysaccharides, introducing alkali-labile linkages into the lignin backbone, and increasing the alkaline solubility of lignin.

As alkaline pretreatment temperature increased from 60 to 120 °C, yields of glucose from cell walls lignified with alternate monomers remained at high levels or modestly declined while yields increased from lignified controls. As a result, differences in glucose yield from modified lignins vs. lignified controls diminished or were eliminated as alkaline pretreatment temperature increased. The exception was hyperoside, whose incorporation into lignin conferred little or no boost in cell wall glucose production relative to lignified controls under alkaline pretreatment conditions.

As noted above, glucose yields from cell walls lignified with alternate monomers such as epigallocatechin gallate often declined as alkaline pretreatment temperatures increased from 60 to 120 °C. High NaOH loading exacerbated this response, particularly in nonlignified controls (Fig. 2.). Overall, this suggests that modified lignins or their degradation products did not accelerate the normal alkali-catalyzed degradation of polysaccharides during pretreatment or adversely effect glucose production by enzymatic hydrolysis or its detection by the GOPOD assay. Further work is, however, needed to determine if alkaline degradation products from modified lignins adversely affect the fermentation of glucose or other sugars into biofuels or other products.

For acid pretreatments, temperatures of 60, 90, and 120 °C with an intermediate H2SO4 loading of 250 mg g−1 were examined because these conditions appeared to best delineate glucose yield differences among lignification treatments (Fig. 2). Hence, after acid pretreatment at 60 or 90 °C, lignification treatments including epicatechin, epicatechin gallate, epigallocatechin gallate, epigallocatechin, epicatechin vanillate, galloylhyperin, and pentagalloyl glucose yielded 40 to 109 mg g−1 more glucose than lignified controls having the same Klason lignin content (Tables 3 and 4). Under the same acid pretreatment conditions, the hyperoside lignification treatment produced glucose yields that were similar to the normal lignin control. Acid pretreatment at 120 °C maximized glucose yields, but eliminated glucose yield differences between modified lignins and lignified controls having the same concentrations of Klason lignin.

Overall, the absence or presence of ester linkages and the degree of o-diol and o-triol substitution of alternate monomers incorporated into lignin did not appear to consistently affect glucose yield responses of cell walls subjected to acid pretreatment with 250 mg g−1 of H2SO4. Thus, the mechanism(s) responsible for improved yields of enzymatically released glucose from acid-pretreated cell walls containing these modified lignins are difficult to ascertain. Dilute acid pretreatment primarily improves enzymatic saccharification by hydrolyzing and removing noncellulosic polysaccharides from the cell wall matrix [4]. At suboptimal acid pretreatment temperatures of 60 or 90 °C, this process might be aided by reduced cross-linking of noncellulosic polysaccharides to modified lignins containing o-diol or o-triol functional groups [21, 22]. By contrast at 120 °C, the effects of cross-linking or other mechanisms were likely superseded by extensive acid-catalyzed depolymerization and removal of noncellulosic polysaccharides, which would expose cellulose to extensive enzymatic hydrolysis.

Increasing H2SO4 loading from 250 to 500 mg g−1 at 120 °C reduced glucose yields by an average of 95 mg g−1 from cell walls lignified with epicatechin or hyperoside, whereas yields from other lignification treatments remained constant or slightly increased (data not shown). This suggests that acid-derived products from epicatechin and hyperoside may have inhibited glucose production by hydrolytic enzymes, reduced detection of glucose by the GOPOD assay, or accelerated the degradation or conversion of glucose into other products. Further testing would be needed to determine the mechanism responsible for reduced glucose yields and to assess whether acid-catalyzed degradation products derived from epicatechin, hyperoside, or other potential alternate monomers adversely affect the fermentation of sugars into biofuels or other products.

Conclusions

Most flavonoid and gallate ester derivatives improved glucose yields from lignified cell walls following alkaline pretreatment, but gallated flavan-3-ols or polygalloyl esters such as epicatechin gallate, epigallocatechin gallate, or pentagalloyl glucose appeared to be the most promising lignin bioengineering targets for improving sugar yields from cell walls that will be subjected to mild alkaline pretreatment prior to enzymatic saccharification. The primary benefit of incorporating these alternate monomers into lignin was to substantially decrease the severity (e.g., temperature) of alkaline pretreatment required to produce high yields of cell wall glucose by enzymatic saccharification. Thus, if bioengineered into biomass crops, these modified lignins might permit rapid and efficient alkaline pretreatment at sub-boiling temperatures (below 100 °C), which should eliminate the need for costly pressure vessels and reduce energy inputs that currently limit the feasibility of cellulosic biofuel production. Our studies, however, suggest that these modified lignins would be less useful for acid-pretreated biomass because modest gains in glucose yields observed at low temperatures were lost at the high temperatures (e.g., 120 °C) required to maximize yields. Finally, before initiating plant bioengineering efforts, additional studies should be carried out to ensure that alkaline or acidic degradation products from lignin containing pentagalloyl glucose, epicatechin gallate, or epigallocatechin gallate do not adversely affect the fermentation of sugars into biofuels or other products of commercial interest.

Acknowledgments

This work was funded by Stanford University’s Global Climate and Energy Project (GCEP) and by USDA-ARS in-house funds. CF, NS, and JR were funded by the DOE Great Lakes Bioenergy Research Center (DOE BER Office of Science DE-FC02-07ER64494). The authors thank Novozymes (Franklinton, NC) for generously providing enzymes for this research. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture.

Copyright information

© Springer Science+Business Media New York (outside the USA) 2015

Authors and Affiliations

  • John H. Grabber
    • 1
  • Nicholas Santoro
    • 2
  • Cliff E. Foster
    • 2
  • Sasikumar Elumalai
    • 3
  • John Ralph
    • 4
  • Xuejun Pan
    • 3
  1. 1.USDA-ARS, U.S. Dairy Forage Research CenterMadisonUSA
  2. 2.D.O.E. Great Lakes Bioenergy Research CenterMichigan State UniversityEast LansingUSA
  3. 3.Department of Biological Systems EngineeringUniversity of Wisconsin–MadisonMadisonUSA
  4. 4.Department of Biochemistry and D.O.E. Great Lakes Bioenergy Research Center, Wisconsin Energy InstituteUniversity of Wisconsin–MadisonMadisonUSA