Historically, the need to carry out large numbers of complex calculations in a sequential manner, in order to undertake sufficient sampling to produce meaningful results, has limited the size of a system that could be reasonably modeled to hundreds of atoms in the 1980s, to millions of atoms now. That limit will continue to be pushed larger but there are new problems associated with systems of even hundreds of thousand atoms. The state-of-the-art computational methods are only now at the point that the smallest cellulosic systems found in nature, such as the plant cell wall cellulose microfibril, can be modeled with confidence. Together with the current force field development, computer hardware technology, and numerical methodologies for high performance computing, the stage is set to probe cellulose and its structures and reactions and answer questions that have been as recalcitrant as the cellulose itself. Reported modeling studies of the cellulose preparations (37, 50-57) are among the few examples of computational structural studies of cellulose. However, Nimlos and coworkers (58) have recently shown that MD simulations of protein-cellulose interactions can shed light on the as yet unknown nature of those interactions. Beyond simple MD simulations, we will discuss the kinds of numerical simulations and the properties that can be studied, quantified, and predicted.

The pretreatment conditions affect the solubilization, recovery, and composition of the solubilized hemicelluloses. Depending on the raw material and pretreatment conditions, high molecular weight fractions, oligosaccharides, monomers, or sugar degradation prod­ucts are formed. Longer pretreatment times (at around 190°C) and additives, such as SO2, lead generally to better recovery of monomers. Less severe pretreatment conditions lead to solubilization of xylan and/or formation of oligosaccharides which can be hydrolyzed into monomers by enzymes (101). The total enzymatic hydrolysis of substituted oligomers needs the synergistic action of endoenzymes and accessory enzymes. The raw material, pretreat­ment conditions, and thus the structure of the solubilized hemicellulose oligomers should be known for the identification of enzymes needed.

The solubilized fraction from steam-pretreated birch wood contained about 10% acetyl groups, which were liberated with a culture filtrate from T. reesei (77). Synergy between xylanases, p-xylosidase, and acetyl esterase of T. reesei was shown to be essential for the production of xylose from steamed birch xylan. Hydrolysis of the high molecular weight fraction of steamed birch wood hemicellulose by xylanase alone produced only about 10% of the amount of xylose produced by the whole set of enzymes (see Table 10.2). The work of characterizing oligomers and solubilized polymers from other substrates and pretreatment techniques is ongoing, with the details being scarce and widely varied. In essence, the best enzymes for a specific hydrolysate must be worked out empirically on a case-by-case basis. Knowledge of the structures involved, either in the native feedstock or hydrolysate, will provide significant clues; however, the lack of details regarding enzyme substrate specificities will still necessitate extensive screening and synergy studies.

Acknowledgment

This work was supported by the US DOE Office of the Biomass Program.

In many respects, the AFEX process is the alkaline equivalent of sulfur dioxide-catalyzed steam explosion pretreatment (13). In the AFEX process, biomass is treated with liq­uid anhydrous ammonia at temperatures between 60 and 100°C and pressures of 250­300 psig with residence times of about 5 minutes (20). The pressure is then released, result­ing in a rapid expansion of the ammonia gas that causes swelling and physical disruption of biomass fibers and partial decrystallization of cellulose, along with some lignin solubiliza­tion and re-arrangement and some solubilization of hemicellulose primarily to oligomeric sugars (10, 14). AFEX is typically conducted at high solids loadings (about 40% solids) and high ammonia loadings (about 1.0 g NH3/g dry feedstock), although the rapid expansion and high volatility of ammonia may permit near-complete recovery and recycle of ammo­nia (20, 38). The associated complexity and costs of ammonia recovery processes may be significant and must be better understood in order to assess the commercial potential ofthe AFEX process (21).

AFEX has been shown to deacetylate and increase the digestibility of biomass (39­41), although it does require that both cellulose and hemicellulose be enzymatically hy­drolyzed due to limited hemicellulose hydrolysis during AFEX pretreatment. The AFEX pretreatment is more effective on agricultural residues and herbaceous crops, with limited effectiveness demonstrated on woody biomass and other high-lignin feedstocks (14). AFEX has largely been practiced as a bench scale technique, although a larger, continuous ver­sion of AFEX based on extrusion technology, known as FIBEX, has been developed and tested (42).

HG-methyltransferase (HG-MT) catalyzes the methylesterification of HG at the C-6 car­boxyl group of some GalA residues by transfer of a methyl group from S-adenosylmethionine to HG. The name HG-MT is preferred, rather than the former name pectin methyltransferase, to distinguish HG-MT from the enzymes that methylate RG-I or RG — II. HG-MT activity has been identified in microsomal membranes from mung bean (280, 281, 283), flax (227, 282), tobacco (284), and soybean (285). Membrane-bound HG-MTs from flax (333, 334), and tobacco (335) could be solubilized using detergent. HG-MT has been localized to the Golgi (227-229, 336) and its catalytic site has been shown to face the Golgi lumen (229). In vitro biochemical studies showing that UDP-GalA stimulates HG-MT activity in intact membrane vesicles (284, 308) and that polygalacturonic and pectin can function in vitro as acceptors for HG-MT in detergent-permeabilized membranes support a model in which at least a small stretch of HG is synthesized prior to its methylation by HG-MT in the Golgi. The observation that some of the HG-MTs in detergent-permeabilized membranes from flax and soybean show a preference for partially esterified pectin (228,285, 337) over polygalacturonic acid further suggest that multiple HG-MTs may exist that differ in their specificity for HG of differing degrees of methylation. The question of whether such HG-MTs are preferentially involved in the initial methylation of HG or in the methylation of more highly esterified HG remains to be resolved. The observation that overexpression of an Arabidopsis S-adenosylmethionine (SAM) synthetase gene in flax leads to a concomitant increase in SAM synthetase activity and of pectin methylesterification in the wall, with no increase in HG-MT activity, suggests that the degree of methylesterification of HG maybe regulated, at least in part, at the level of substrate (i. e., SAM) concentration (338).

The gene encoding HG-MT has not yet been unambiguously identified. Two apparent HG-MT isozymes, PMT5 and PMT7, were reported from flax (337) and efforts to purify these apparent isozymes resulted in the identification of an additional small polypeptide with HG-MT activity designated PMT18. The definitive identification of one or more of these polypeptides as HG-MT has not yet been reported. Thus, the proposition that the 18-kDa protein is a subunit of the 40- and 110-kDa proteins (337) has not been substantiated.

Recently, Mouille and colleagues (287) have identified a putative methyltransferase as the gene mutated in the Arabidopsis mutant quasimodo2. Qua2-1 plants are dwarfed and have a 50% reduction in HG. Although confirmation of enzyme activity of QUA2 is required to establish if it indeed encodes an HG-MT, the reduced HG phenotype of qua2 plants along with the Golgi localization of QUA2-GFP fusions and the putative methyltransferase domain in QUA2 are consistent with a role as an HG-MT. Further work on QUA2 and the 29 QUA2-related proteins in Arabidopsis may shed light on the identity of multiple methyltransferases required for pectin synthesis.

UDP-Xyl is primarily synthesized by UDP-GlcA decarboxylase (UGlcA-DC, also named UDP-Xylose Synthase, Uxs) from UDP-GlcA. The enzyme has a tightly bound NAD+, which participates first in the oxidation of UDP-GlcA to the UDP-4ketohexose intermediate resulting in decarboxylation (removal of COOH as gas, CO2) and formation of UDP — 4ketopentose. In the second stage, the NADH-bound enzyme reduces the UDP-4ketopentose to UDP-Xyl resulting in the release of NAD-bound enzyme (403). Multiple distinct UXS isoforms encoding this enzyme activity were reported in Arabidopsis (431), rice (474), barley (475), and tobacco (463). Uxs isoforms are very specific enzymes and act only on UDP-GlcA. The 4-epimer of UDP-GlcA, UDP-GalA is not a substrate for Uxs (431). Uxs is active as a dimer and inhibited by UDP-Xyl. In plants, UDP-Xyl is made in the cytosol and in the endomembrane system (473). Phylogeny analysis classified the six Uxs isoforms from Arabidopsis into three distinct clades: Type A (1 isoform, At3g53520); Type B (2 isoforms, At3g62830, At2g47650), and Type C (3 isoforms, at5g59290, At3g46440, At2g28760). Type A and B isoforms have an N-terminal extension (~ 120 aa long) which encodes longer proteins compared to Type C Uxs isoforms. Type A and B UXS isoforms are predicted Type II membrane proteins with the catalytic domain facing the endomembrane lumen (431). Expression of Type B Uxs isoforms in plants confirmed that Uxs2 is an integral membrane protein. Expressing of a Uxs2-GFP construct in tobacco leaves was shown to localize the chimeric fusion protein in the Golgi apparatus (473). Uxs3 belongs to the Type C clade and the isoform was found in the cytosol as predicted.

As shown in Figure 7.1, the pathway to the monolignols 3-5 includes two methylation steps. These have been shown to be catalyzed by caffeic acid O-methyltransferases (COMTs) and caffeoyl CoA O-methyltransferases (CCOMTs). COMT activity was first detected in cambial tissue isolated from an apple tree species by Finkle and Nelson (134). It was thought for much time that COMTs could catalyze the formation of both ferulic (11) and sinapic (13) acids from caffeic (10) and 5-hydroxyferulic (12) acids, respectively. However, in 1989, Pakusch et al. (135) described an enzyme, CCOMT, capable of converting caffeoyl CoA (15) into feruloyl CoA (16). Additionally, downregulation of the COMT gene later unambiguously

NADPH

Figure 7.9 (A) Structure of the substrate-binding pocket of NADP+ binary form of AtCAD5 showing the catalytic Zn2+ ion (red sphere) tetrahedrally coordinated by Cys47, His69, Cys163 and Glu70 (blue). The NADP+ molecule (orange) is held by Val192, Ser211, Ser212, Ser213, Lys216 and Gly275 (green) (133). [Possible hydrogen bonds are shown as black dotted lines.] (B) Proposed proton shuttling mechanism during the reduction process in the active site of the AtCAD5. Solid arrows indicate the movement of two electrons among the functional groups during substrate reduction (133). The possible hydrogen bonds involved are shown with dotted lines. (Reprinted from Organic and Biomolecular Chemistry, vol. 4, Youn, B., Camacho, R., Moinuddin, S. G.A., Lee, C., Davin, L. B., Lewis, N. G. & Kang, C., Crystal structures and catalytic mechanism of the Arabidopsis cinnamyl alcohol dehydrogenases AtCAD5 and AtCAD4, pp. 1687-1697, Copyright 2006, with permission from The Royal Society of Chemistry.) (Reproduced in color as Plate 20.)
established its role in syringyl (S) unit formation and not the initial step involving G-lignin deposition (136). Genes encoding COMT and CCOMT were first reported in 1991 by Bugos etal. (137)/Gowri etal. (138) and by Schmitt etal. (139), respectively, with the actual biochemical/biophysical function for COMT being determined later by Atanassova et al. (136).

In the TAIR database, there are 17 genes putatively annotated as COMTs and five as CCOMTs. Investigations in our laboratory have indicated that out of the 17 COMTs only one, AtCOMT1, was capable of methylating caffeic (10)/5-OH ferulic (12) acids, caffeyl (20)/5-hydroxyconiferyl (22) aldehydes and caffeyl (2)/5-hydroxyconiferyl (4) alcohols in vitro. The efficiency of the reaction is higher when the aldehydes 20 and 22 are used as substrates (Zhang etal., manuscript in preparation). Only two CCOMTs, AtCCOMT1 and AtCCOMT2, have been investigated: caffeoyl (15) and 5-hydroxyferuloyl (17) CoAs are the preferred substrates for AtCCOMT1, with 10 and 12 not being converted into ferulic (11) and sinapic (13) acids, respectively. AtCCOMT2 more efficiently methylated quercetin (Takahashi etal., manuscript in preparation).

of lignin structure

With the incremental advances made in the study of lignins, such as by application of NMR and mass spectroscopic analyses, together with thioacidolytic degradation, further refinement of possible lignin representations were attempted for these polymers in spruce (15) and wheat straw (314). The Brunow et al. (15) model speculated the existence of random-linked structures, such as the 28-unit structure shown (Figure 7.3C), and also that dibenzodioxocin (substructure V, Figure 7.2D) and 5-5′ linkages (substructure VII) served as important branching points. The envisaged structure (Figure 7.3C), while acknowledging herein its hypothetical basis as emphasized by the researchers themselves, would nevertheless contain on a per monomer basis 10 potentially cleavable thioacidolysis monomers (35%), with the remaining eighteen units (65%) being releasable as 8-5′ (3), 5-5′ (2), 8-1′ (2), 4-0-5′ (1), and 8-8′ (1) linkedmoieties — where one of the latter has a presumed C-5 linkage to an adjacent lignin subunit, respectively. As gleaned from inspection of this proposed “hypothetical” structure, much of it would presumably be readily susceptible to thioacidolysis degradation. However, release of such fragments/substructures in the rel­ative amounts experimentally determined has not been observed. Furthermore, nor has it been established that either the dibenzodioxocin or 5,5′ substructures serve as branching points. Thus, this proposed structure again does not meet experimental scrutiny.

A linear (8-5′) linked lignin macromolecule has also been proposed (314), based on mass spectrometric analysis (APCI-MS, MS/MS, and MALDI-TOFMS) of the extracted lignin from wheat straw using the AVIDEL (315) procedure. Such proposed structures, however, need to be verified, quantified and placed in context with the existing chemistry of lignins, in order to assess what, if any, their relative merits and contributions are to lignin macromolec­ular configuration. Whatever the limitations of this approach, these researchers sought to obtain needed primary sequence data, unlike many of the previous highly speculative lignin structural models which did not.

In this chapter, quantum mechanical calculations investigating the energetics and kinetics of these reactions are discussed. The hydrolysis of xylobiose and the degradation of xylose were investigated using quantum mechanical molecular dynamics simulations and with static electron structure theory. The molecular dynamics simulations sample a large portion of the potential energy surface of a reacting system and can help identify probable reac­tion pathways. The sugar degradation reactions were carried out both in the absence and presence of the surrounding water molecules. Static quantum mechanical approaches pro­vide more accurate energies, allowing the determination of reaction energy barriers, which are used to compare the likelihood of competing reaction pathways. These approaches when applied to degradation reactions of xylose provide considerable insight into the likely mechanisms and kinetics. Further application of the static techniques to xylobiose has shed light on the relative reaction rates of hydration and dehydration for xylo-oligomers and xylan.

Recent analysis of two major cellulases from the proteome of R. albus 8 showed that both carried cellulose-binding modules, but neither possessed an obvious dockerin sequence (25). These enzymes, therefore, appear to be non-cellulosomal, although they might still be retained on the cell surface by other mechanisms. Meanwhile, a number of enzymes from R. albus 8 were also found to carry partially homologous C-terminal sequences that show remarkably broad polysaccharide-binding specificities (26). Although these sequences have potential roles in binding to dietary substrates, it is not ruled out that they might mediate binding of the polylpeptides to carbohydrates on the cell surface, thus providing an anchoring mechanism. Other enzymes from R. albus 8 and from several other R. albus strains have, however, been reported to possess dockerins (27) and a large family of dockerin-containing proteins has been detected in the genome of R. albus 8 (28). The role and importance of cellulosome organization in this species, therefore, seems less clear than in R. flavefaciens, and this attribute may even vary between strains of R. albus.

Based on studies in two R. albus strains (8 and 20) type IV pili have been proposed to be involved in binding cells to cellulose (29-31). The pilAl protein is thought to be a major component of the pius of R. albus 20, while pilA2 is more likely to play a role in pilus synthesis (31).

Fungal species play a critical role in plant biomass decay and can be divided into the fol­lowing three categories: (a) saprophytic fungi, which prefer dead and decaying material; (b) parasitic fungi, which prefer colonizing the biomass of the living host; and (c) mycorrhizal fungi, which form partnerships with specific plant species, mostly with living trees. Most fungi are saprophytic and are found in nature growing on dead organic matter. They are effective at secreting enzymes that degrade large polymers such as cellulose, hemicellulose, lignin, pectin, starch, and protein found in the organic matter to release nutrients that can be taken up and used by the fungi. The filamentous fungi are widespread in nature and for most part are obligate aerobes. Typically, fungi colonize substrates with tubular, branching hyphae that collectively form the mycelium. The hyphae cell walls composed of p-glucans and chitin embedded in a matrix that includes p-glucans and glycoproteins. As shown in Figure 15.1, growth is by extension of the hyphal tips that enables the mycelium to spread and penetrate throughout the substrate allowing new areas to be colonized as the nutrients are used from the earlier colonized substrate. Extensive communication and nutrient trans­fer is possible through the hyphae and between different regions of the mycelium. Older parts of the mycelium may die after nutrients in the substrate are exhausted and material from these areas can be transported to younger regions of the mycelium. Many ascomycetes and basidiomycetes produce a group of small (70-120 amino acids) amphiphilic proteins that promote attachment of hyphae to hydrophobic surfaces (9, 10). Hydrophobins are

self-assembling proteins that have been demonstrated to function at hydrophilic — hydrophobic interfaces to form films and to function as surfactants (10). Hydrophobins have been proven to lower the surface energy of water allowing the fungal hyphae to pene­trate the air-water interface and grow up into the air (11,12). Because of their adhesive and surface activity, hydrophobins are an interesting feature of biomass colonization.

Producing plant cell wall-degrading enzymes is widespread in fungi and has been de­scribed in the anaerobic fungi found in ruminants (13) and all other subdivisions of aerobic fungi including members of the Zygomycetes, Ascomycetes, Basidiomycetes, and Deuteromycetes. A major ecological activity of many fungi, especially members of the Ba- sidiomycetes, is the decomposition oflignocellulosic materials such as wood and other plant material. Fungal breakdown of biomass occurs by succession where the species of one fun­gal community alters the substrate enough to allow other species to become established and colonize (14). Two types of wood rot are known: brown rot, in which the cellulose is preferentially used and the lignin is not metabolized and white rot where both cellulose and lignin is used. White rot, brown rot, and soft rot fungi are recognized among those that colonize dead wood. Soft rot fungi degrade cellulose and hemicellulose in the wood under conditions of high moisture content leaving the wood soft or spongy without being fully degraded. Soft rot has also been reported to occur in dry environments and under other extreme conditions (15). Soft rot decay of wood reportedly occurs at historic sites in Antarctica where extreme environmental conditions inhibit wood decay from other fungi

(15) . Soft rot is caused primarily by fungi classified in the phylum Ascomycota with the decay characterized by cavities that form within the biomass structure. Soft rot fungi can also cause a progressive degradation of secondary cell walls that can be completely degraded except for the middle lamella between the cells (15). The white rot fungi exhibit a large amount of diversity, but are generally members of the Basidiomycetes or other higher fungi, and produce enzymes that are degraded lignin in addition to cellulases and hemicellulases

(16) . The hyphae of white rot fungi rapidly colonize wood by growing within the lumen of the cells and degrading the cell walls. Although they can degrade the substrate, white areas of the wood remain where the lignin and hemicellulose have been removed ahead of the cellulose. Under conditions where the biomass is moist members of the Basidiomycota will rapidly degrade cellulose and lignin by generating oxidants such as hydroxyl radicals.

Parasitic fungi are the second largest group, of whose members do considerable dam­age to growing plants. Indeed, root diseases and mycorrhizal systems have a similarity regarding parasitism. Root pathogens, such as Rhizoctonia, Fusarium, Verticillium, Sclero — tinia, and Pythium are stimulated by roots. These fungi progressively invade and colonize the meristematic root tissues and ultimately cause necrosis.

The rhizosphere is the region immediately outside the root that generally has more mi­crobial activity than the surrounding soil. Mycorrhiza refers to the symbiotic association between plant roots and fungi classified as ectomycorrhizae where the fungi form an extensive sheath around the root, and arbuscular mycorrhizae where the fungal mycelium becomes embedded with root tissue. Mycorrhizal associations are widespread and are found on most plants in diverse environments. Although the production of hydrolytic enzymes has been described in many mycorrhizal fungi, most do not use cellulose for metabolism but obtain carbon from root secretions (17). Mycorrhizal fungi have been demonstrated to have weak cellulase and endopolygalacturonase activities (17). Hohnjec etal. described the presence of three different endoglucanases with different preferences for sugar bonds and four different pectinolytic or polygalacturonate-degrading enzymes in G. mosseae and G. intraradices — colonized Medicago roots (18). Cellulase, pectinase, and xyloglucanase activities have been found in the external mycelium of arbuscular mycorrhizas (19). The production of these hydrolytic enzymes is thought to allow one for the modification of the extracellular matrix of the root and allow fungal colonization (18).