ENZYMES FOR BIOENERGY
Our research on bioenergy is supported by the Department of Energy through the Great Lakes Bioenergy Research Center (http://www.glbrc.org).
Enzymes for Bioenergy Applications
Current scenarios for the conversion of lignocellulosic biomass to ethanol and other liquid transportation fuels utilize enzymes to depolymerize polysaccharides to their constituent fermentable sugars. However, the high cost of enzymes is a major hindrance to the development of a viable lignocellulosic ethanol industry.
A multitude of enzymatic activities are required for the conversion of lignocellulosic biomass into useful (fermentable) products. Known essential activities to degrade crystalline cellulose include cellobiohydrolase (CBH), endo-β1,4-glucanase (EG), and β-glucosidase (BG). A larger suite of enzymes is necessary to depolymerize hemicelluloses, including endo-β1,4-xylanase (EX), β-xylosidase (BX), α-arabinosidase, α-glucuronidase, and esterase. Currently available commercial enzyme preparations for the depolymerization of lignocellulosic materials are partially defined complex mixtures of the secreted proteins from filamentous fungi such as Trichoderma reesei that have been grown in the presence of inducers such as sophorose or lactose. Commercial enzyme “cellulase” mixtures contain between 80 and 200 proteins (Nagendran et al., 2009). Except for a few of the better-characterized cellulases and hemicellulases, the roles of most of these proteins in lignocellulose deconstruction are poorly understood (Banerjee et al., 2010a).
The way forward for the development of more efficient lignocellulose-degrading enzyme cocktails will require deeper and more precise knowledge about the specific enzymes that are involved in the degradation of lignocellulose. This knowledge is important for many reasons (Fig. 1). However, it is not possible to gain this knowledge working only with partially defined complex mixtures. The precise contribution of an individual enzyme can be established only by working with it in a purified state in a realistic enzyme cocktail (Banerjee et al., 2010b).
We have developed a platform, called GENPLAT, for construction and optimization of synthetic mixtures of enzymes. GENPLAT is described in Banerjee et al. (2010b). Key features of GENPLAT are the use of a bead mixing chamber for accurate and rapid pipeting of stover slurries; the use of 96-deep well plates; automated liquid handling for dispensing slurries, enzymes, and buffers; gentle mixing during digestion by end-over-end rotation; and automated colorimetric determination of Glc and Xyl. Fractional factorial experimental design uses Design-Expert™ software from Stat-Ease, Inc.
Using GENPLAT, we have developed cocktails with up to 16 components for release of Glc and Xyl from multiple pretreatment/biomass combinations combinations (Banerjee et al., 2010c; Banerjee et al., 2010d; Banerjee et al. 2010d Supplementary Data)(Fig. 2).
Alkaline Hydrogen Peroxide Pretreatment
Without pretreatment, yields of Glc and other sugars from most native biomass sources in response to enzyme treatments are very low. Most lignocellulosic scenarios include a thermochemical pretreatment prior to enzyme hydrolysis.
An ideal pretreatment would be effective (high sugar yields in a short time at low enzyme loading), simple (avoidance of multiple biomass handling steps), inexpensive (for capital equipment, energy, and/or chemical input), and compatible with high biomass loadings. It would also minimize water consumption and not release or generate inhibitors of downstream unit operations (i.e., enzyme digestion and fermentation). Current pretreatment methods are less than ideal in one or more of these properties.
Pretreatments using hydrogen peroxide at pH 11.5, herein referred to as alkaline peroxide (AHP), was shown by Gould and colleagues to be an effective process for ruminant and in vitro enzyme digestibility. AHP is relatively unstudied compared to other thermochemical pretreatments and has generally not been included in comparative studies and reviews.
We have found that AHP compares favorably against dilute base (0.25% NaOH) or ammonia fiber expansion (AFEX) as a pretreatment for corn stover, Miscanthus, and switchgrass. AHP pretreatment of corn stover yielded 18% more Glc than AFEX (69% of total available Glc vs. 52%) when hydrolyzed with a 16-component synthetic mixture at an enzyme loading of 15 mg protein/g glucan. Xyl yields were also higher from corn stover pretreated by AHP compared to AFEX (55% vs. 41% of available Xyl) (Banerjee et al., 2010d) (Fig. 3).
With AHP we can achieve close to maximum Glc release at reasonable enzyme loadings (Fig. 4).
A major advantage of AHP over most other pretreatments is that it requires no special reaction chambers. It operates at room temperature and atmospheric pressure and is easily scalable (Fig. 5). Ongoing research between our lab and those of David Hodge (MSU Department of Chemical Engineering) and Eric Hegg (MSU Department of Biochemistry) is aimed at understanding the mechanisms of AHP pretreatment and developing it into an economically practical pretreatment. See our recent paper on AHP optimization: Banerjee et al., 2011.