Experiments were conducted with xylobiose in the microwave reactor and results are shown in Figure 9.8 for the decomposition of xylobiose and the formation of xylose at nominal temperatures of 125 and 145°C. As can be seen from these plots, the xylobiose is quickly converted to xylose at these temperatures, i. e., within 600 seconds for 125°C and within 120 seconds for 145°C. Furfural formation was less than that observed for decomposition of pure xylose, suggesting that the furfural arose from the decomposition of xylose and not xylobiose. The mass balances in these experiments were poor (80-85%), due probably to the poor calibration of xylobiose. Because of the expense of this compound, we were only able to use single-point calibrations. However, as Figure 9.9 shows, the chromatographs were clean and did not indicate the formation of other products. These results seem to confirm the

Time/min

results from our calculations that xylobiose will preferentially undergo hydrolysis relative to dehydration. We conclude that low DP xylo-oligomers will preferentially undergo hydrolysis to form smaller oligomers and xylose as opposed to dehydration to form furans in solution containing only catalytic acids.

Interest in lignocellulose breakdown in the gut has centered largely on the rumen, mainly be­cause ofthe nutritional and economic importance ofruminal degradation, but also because of its relative accessibility as a gut microbial ecosystem. The large intestine of herbivores such as horses, however, harbor anaerobic communities of similar diversity and complexity (82) and include anaerobic bacteria, fungi, and protozoa. The major cellulolytic bacteria again appear to species of Ruminococcus and Fibrobacter (83). There is no indication that omnivores, including man, carry cellulolytic eukaryotes, but cellulolytic bacteria related to Ruminococcus spp. have been isolated (84). In addition, there is evidence that unknown Clostridium-related bacteria from human faecal samples can attach closely to wheat bran (85).

12.3 Conclusions

It is increasingly clear that the pool of DNA sequences coding for polysaccharide-degrading activities within the rumen microbial community has been subject to horizontal genetic exchange, including at some stage exchange between prokaryotic and eukaryotic microor­ganisms. The protein structures to which these catalytic and binding domains contribute, however, show an extraordinary diversity among rumen microorganisms. It seems safe to assume that the macromolecular organization of enzymes, adhesion mechanisms, and trans­port systems plays a crucial role in determining the ecological niche occupied by a given species. The details of this relationship, however, remain to be clarified.

Complex multidomain organization in polysaccharidases seems to be a feature of the primary cellulolytic species of anaerobic bacteria and fungi found in the rumen. At least in R. flavefaciens and cellulolytic rumen anaerobic fungi, a significant proportion of these polypeptides area organized into multienzyme cellulosome-type complexes that are an­chored to the cell surface. While it is assumed that these complexes play a key role in cellu­lose and plant cell wall decomposition, as in cellulolytic Clostridia (86), functional evidence is still lacking. Furthermore, the mechanisms by which R. albus and F. succinogenes, two cellulolytic rumen bacteria that produce complex multidomain enzymes but have not been reported to display cellulosome organization, anchor their enzymes to the cell surface and achieve efficient plant cell wall breakdown remain unclear. Cellulolytic protozoa apparently achieve plant cell wall breakdown despite producing soluble enzymes of relatively simple structure, from the limited information currently available. These organisms may present a special case, however, since digestion is assumed to occur within food vacuoles that also contain ingested cellulolytic bacteria.

Analysis of completed genomes for representative cellulolytic microorganisms should soon provide crucial insights into their degradative enzyme systems. This information will need to be illuminated by functional studies, however, with a continuing requirement for gene transfer methodologies that can be applied to rumen microorganisms.

Acknowledgment

The Rowett Research Institute receives funding from the Scottish Executive Environment and Rural Affairs Department.

Collectively, the genomes of the total microbiota found in nature, referred to as the metagenome (37), contains vastly more genetic information than is contained in the cultivable subset. However, the genetic complexity of a microbial community at a spe­cific site is influenced by many environmental factors. Re-association of total community DNA extracted from different environmental distinct sites has revealed that the community genome size can equal that of 6000-10 000 Escherichia coli genomes in unperturbed organic soil, but only 350-1500 genomes in arable or heavy metal-polluted soils (83, 94). These estimates are conservative, since genomes representing rare and unrecovered microorgan­isms were probably not included in the analysis. As expected, Torsvik and Ovreas (2) could recover less than 40 genomes by culturing methods which emphasizes the need for develop­ment of novel methods and approaches to provide new insight into the relationship between phylogenetic and functional diversity of these communities as ecosystems.

DNA sequencing continues to be one of the most important platforms for the study of biological systems. With the development of improved sequencing technologies that en­hance the speed, sensitively and throughput, it has become feasible to sequence the entire metagenome of an environmental sample (95). Culture-independent genomic analysis of microbial communities using metagenomics is revealing that soil and ocean environments are more genetically and potentially more biochemically diverse than previously thought (96). This involves the cloning and analysis of large genomic DNA fragments isolated from a mixed community. The metagenomic library can then be screened for functional or tax­onomic genes of interest or sequenced by shotgun sequencing. Most environments contain communities far too complex for it to be possible to sequence a complete metagenome, and even the simple communities contain micro-heterogeneity that makes most genome reconstructions simplified versions of reality. Reconstruction of community metagenomes was initially pursued for viral communities in the ocean and human feces (97-99) and has since been attempted in an acid mine drainage (AMD) biofilm (100) and the Sargasso Sea (101). The AMD biofilm community was ideal for complete metagenome sequencing be­cause 16S rRNA gene sequencing indicated that there were three bacterial and three archaeal species in the biofilm. Marine communities contain far greater species richness, on the order of 100-200 species per milliliter of water (102, 103), making the sequencing and assembly effort considerably more difficult. Further out on the continuum of biological complexity is soil, with an estimated species richness on the order of 4000 species per gram of soil (35, 102, 103). Sequencing the soil metagenome requires faster and less expensive sequencing technology than currently available.

Recently, we initiated in collaboration with the JGI the sequencing of the metagenome of a microbial community actively decaying poplar biomass under anaerobic conditions. The predominance of microbial enumeration in the biomass pile is represented by this large anaerobic core zone. In addition to some cellulolytic fungi, bacteria of the order Clostridiales, many of which have strong cellulolytic activities, were found to dominate this specific microbial community. The estimated composition and the distribution of bacte­rial members of this community were determined based on 16S rRNA gene sequencing (S. Taghavi and D. van der Lelie, unpublished). It should be noted that we were able to cultivate several members of this community and characterize their cellulolytic activities. Interestingly, none of these cultivable species represented the dominant community mem­bers, stressing the importance to use a cultivation-independent approach to characterize the composition and metabolic potential of this complex microbial community.

4.3.2.1 Polysaccharide-polysaccharide cross-linking.

4.3.2.1.1 DIMERIZATION OF ESTERIFIED HCA

In grasses, hydroxycinnamate esters on GAX (see Section 4.2.1.2.2) in primary and lignified secondary walls are covalently cross-linked by oxidative dimerization of HCA units on neighboring AX chains by radical coupling reactions (32, 79) (Figure 4.3a). The homo — and hetero-dehydrodimers formed involve mostly FA and sometimes SA (80); pCA does not appear in these dehydrodimers. Dehydrotrimers (32, 81) and a dehydrotetramer (82) of FA have also been described and could participate in cross-linking. A number of isomeric cross-linking homo — and hetero-dimers have been encountered involving both the aromatic ring and the propenoic acid side chain of the HCA (Figure 4.4) (83, 84).

4.3.2.1.2 ESTERIFIED HCA CYCLODIMERIZATION

In addition to dimerization of ester-linked HCA by oxidative coupling, cross-linking dimers are formed from FA and/or pCA monomers by photodimerization (83). These ester-linked cyclodimers are cyclobutane derivatives (truxillic acids and truxinic acids) (Figure 4.3b). Both the esterified HCA dimers and cyclodimers are readily released by treatment with dilute alkali (0.5 M NaOH, 20°C, 18 h) (85).

4.3.2.1.3 DIRECT ASSOCIATIONS BETWEEN POLYSACCHARIDES

Direct covalent linkages between pectic polysaccharides and xyloglucans in angiosperm walls have been reported (86-88), but the detailed chemistry of the linkage(s) is not known.

Figure 4.3 Three modes of covalent cross-linking involving feruloylated GAXs: (a) a dehydrodiferulate cross-link(5-5), (b)aferulatecyclodimer (truxillicacid) cross-link, and (c)tyrosyl-ferulatecrosslinkbetween a protein and a feruloylated GAX. Ph = 4-Hydroxy-3-methoxy-benzene.

Studies of the Nicotiana plumbaginifolia T-DNA nolac-H18 callus mutant lead to the iden­tification of the mutated gene, NpGUT1. NpGUT1 has 60% sequence homology to animal glucuronosyltransferases that synthesize heparin sulfate. Complementation of the nolac — H18 mutant with the NpGUT1 gene corrected the non-organogenesis and weak intercellular attachment phenotypes of the mutant. Cell walls of the nolac-H18 mutant contained 86% reduced levels of glucuronic acid, a reduction that was associated with the pectin-enriched fraction of the walls. RG-II from mutant walls was devoid of glucuronic acid, leading the authors to propose that NpGUT1 encodes RG-II-p-1,2GlcAT that transfers GlcAonto the l — fucose in RG-II side chain A (181). The mutant RG-II in the nolac-H18 showed 82% reduced

RG-II dimer formation, providing further support that RG-II is modified in the mutant. Taken together these results show that NpGUTl encodes a putative RG-II: p-1,2GlcAT. En­zymatic confirmation of the activity of the encoded protein has not yet been reported.

5.4.9.2 RG-II:methyltransferase (RG-II:MT)

RG-II contains 2-O-methylfucose and 2-O-methylxylose. It is not known if the methyl group is added at the stage of the nucleotide-sugar or after the sugar is added to RG-II. The genes encoding the methyltransferases have not been identified.

A pectin methyltransferase activity detergent-solubilized from suspension-cultured flax cells was able to transfer methyl groups from S-adenosylmethionine onto RG-II isolated from wine (334). Enzyme reactions containing RG-II had sevenfold great methyltransferase activity than reactions without exogenous acceptor and the radiolabeled product synthesized had a size similar to RG-II monomers and RG-II dimers (158, 334). It was not established where in RG-II the methyl group was added and thus, the methylation may have represented methylesterification of the HG backbone of RG-II, or alternatively, could have been due to methyletherification of RG-II since RG-II contains methyl groups on non-galacturonic glycosyl residues (e. g., 2- O-methyl xylose and 2- O-methyl fucose (352, 360) of side chain residues. The location of the methylation in RG-II and the identity of the potentially novel enzyme activity that catalyzes its incorporation into RG-II have not been established.

The eight-carbon acid sugar KDO, 3-deoxy-D-manno-2-octulosonic acid, is a primary sugar constituent in various types of cell surface extracellular polysaccharides and liposaccharides of Gram-negative bacteria. In plants, KDO is found only in RG-II. Synthesis of CMP-KDO in plants requires the activities of three enzymes.

1 KdsA, KDO-8-P synthase, catalyzes a condensation of PEP, phosphoenolpyruvate, and phosphorylated monosaccharide, D-arabinose 5-phosphate (A5P) in the presence of metal. Functional genes encoding KDO-8-P synthase activity were isolated from various plant species (498-500) and the encoded proteins share ~50% amino acid sequence iden­tity with the bacterial proteins. In Arabidopsis, two gene isoforms (AtkdsAl, At1g79500; and AtkdsA2, At5g09730) were identified. The encoded isoforms share 93% amino acid sequence identity to each other. Interestingly, AtkdsAl is predominantly expressed in shoots, while AtkdsA2 transcript accumulates to a higher level in roots. The activity of the recombinant plant kdsA toward other phosphorylated-sugars, such as D-erythrose-4- phosphate (E4P) was not tested. Based on bioinformatics, the plant KdsA are predicted to reside in the cytosol.

2 KDO-8-P phosphatase activity removes the phosphate to form KDO. The nature and specificity of this phosphatase is unknown.

3 KdsB, CMP-KDO synthase (CMP-KDOs), catalyzes the transfer of the cytidylyl group (CMP) from CTP to KDO in the production of the unusual nucleotide-sugar, CMP-p-KDO. The resulting activated sugar has a half-life of about 30 minutes in so­lution. The maize gene homolog was functionally identified (501), and the homologous Arabidopsis protein sequence is encoded by At1g53000. In vitro, the recombinant maize CMP-KDO is capable of using both CTP andUTP as nucleotide (Bar-Peled, unpublished). Interestingly, the plant proteins that share 40-50% amino acid sequence identity to the bacterial KdSA proteins have a 50 amino acid N-terminal extension. Bioinformatic anal­ysis suggests that plant CMP-KDO is a Type Ib transmembrane protein with the catalytic domain predicted to face the cytosol. However, the subcellular location of the plant protein is uncertain. Since the function of RGII:2,3KDOT activity was not reported, it remains unclear if CMP or UMP-Kdo are the sugar donors.

7.5.3.1 Klason, acetyl bromide, and thioglycolic acid estimations

Of these three procedures, the two most commonly used for estimating gross lignin con­tents are the Klason and AcBr methods, respectively. The first method, albeit very routinely applied, is most uninformative, since it only measures insoluble material remaining follow­ing “digestion” with 72% H2SO4 (179). While generally reliable for mature woody plant stem material, it has substantial limitations when generically applied to both herbaceous and immature woody tissues [discussed in Anterola and Lewis (77) and references therein], as well as for other tissues, such as bark. Indeed, as long ago as 1986, Leary et al. (180) carried out Klason lignin analyses on 15 samples of hardwoods, softwoods, and grasses; it was determined from these studies that various non-lignin components, such as tannins, were present in the “lignin” isolated thereby making the analyses ofsolely insoluble material suspect for various sample types.

<———————————————————————————————————————

Figure 7.10 Sulphonation, alkaline nitrobenzene oxidation, and thioacidolysis products. (A) Sulphonated lignin-derived monomers (38-41) (147). (B) Products formed by alkaline nitrobenzene oxidation of 8-0- 4′-linked model compounds 42-44. (C) Proposed mechanism for NBO lignin cleavage. [Adapted from Schultz and co-workers (189, 190).] (D) Thioacidolysis monomeric products 54-56 from 8-0-4′ model compounds 51-53. (E) Maximum amounts (and ranges) of monomeric/dimeric products from lignins that are typically released by alkaline nitrobenzene oxidation (NBO), (71, 132) thioacidolysis (71, 132) and permanganate oxidation;the bulk of the lignin biopolymer is largely unaccounted for.

Two recently published examples ofKlason lignin contents in juvenile poplar and tobacco stems serve to further illustrate the current state of disarray in acquiring reliable basic analytical data: A report of “lignin” contents in immature (approximately 1-year old) poplar (Populus tremula x Populus alba) stems gave values as high as ~32% (181), whereas others analyzing 3-month and 2-year-old poplar had indicated that levels were ~20% (182-184). Values of ~32% are well outside the ranges expected, since it is known that, for example, mature poplar wood tissues only have lignin levels ~ 18—21% (5). If lignin contents can be overestimated by up to nearly 60%, such approaches are unlikely to identify meaningful trends in lignin deposition/composition and assembly proper, as well as on the kinetics of lignin removal. Another quite similar example also occurred in tobacco stem analyses (69), whose sections were reported to contain ~40—50% lignin rather than the 20—25% expected. Both examples reflect simply a degree of unreliability in the data obtained through Klason lignin analyses, and thus a departure from the analytical rigor expected.

Similar concerns about the unreliability of thioglycolic acid lignin determinations have also been noted and critically evaluated (77). Two examples will again suffice: thioglycolic acid levels of presumed lignin contents in stem sections of 4CL downregulated Arabidopsis stems were considered to be ~50% lower than that of wild-type levels (185). On the other hand, reanalysis of this study by Anterola and Lewis (77) established that their alkaline nitrobenzene oxidation protocols had given > 115% recoveries of lignin-derived fragments, rather than the ~25% or so expected, and suggest unreliability in one or both procedures. These results again reflect significant departure from the data expected. A similar level of unreliability was also noted for COMT downregulated alfalfa analyses: lignin contents were considered to be reduced by ~50% (186), although many other studies (77) have demonstrated that there were no significant levels of reduction in lignin amounts following COMT downregulation.

The AcBr method “solubilizes” lignins, as well as other non-lignin phenolics, as bromi — nated derivatives; it has also been applied to numerous lignin determinations, using a generic extinction coefficient of e280 = 20.09Lg-1cm-1 (179,187). Again, while very routinely used, this method does not take into account the differences in extinction coefficients due to vari­ations in lignin monomeric H, G, and S compositions. In our more recent investigations using H-, G-, and S-enriched lignin isolates, the best current estimates of the actual extinction coefficients (, 280 nm) were established to be considerably different, i. e., (H) 15.3, (G) 18.6, and (S) 14.6 L g-1cm-1 for p-coumaryl (1), coniferyl (3), and sinapyl (5) alcohol-enriched lignins (71,188). These differences thus again underscore the need for both circumspection and scientific rigor in the study of lignins, i. e., as routinely expected for all other areas of natural product chemistry.

In the early molecular models, water was not included, but was later determined to be essential to carbohydrate structure and behavior. The solvent model types are classified as explicit solvent, in which solvent molecules are explicitly modeled using the same model type as the solute, or implicit, in which the effect of solvent is modeled by a functional form and is a function of solute configuration. Water is the most common solvent, though other solvents and mixed solvents are also used, especially to reproduce experimental solvent environments. There are several explicit water models, the most important ones being TIP3P, SPC, TIP4P, and TIP5P (41).

The implicit solvent models are based on the assumption that the major effect of solva­tion is encapsulated in its dielectric properties. In a simple protein of 2400 atoms, solvating with explicit water molecules for a cube with at least three water layers on each side in­creases the system size to 23 500 atoms. The simplest, and crudest, model simply uses a distance-dependent dielectric constant in the electrostatic term of Equation (8.1), so that the effect of the solvent is to mask the charge interaction between distant charges, assumed to have a dielectric medium between them, and to not mask at all when two charges are close to each other. The advantage of this method is that no explicit water molecules are included in the simulation and the cost of a distance-dependent dielectric constant is mini­mal, cutting the computational demand by a factor of ten or more. The drawback is that one does not model the solvation free energy correctly nor the dielectric environment inside a macromolecule.

The more sophisticated methods of modeling solvent implicitly are based on solving the Poisson equation. The most rigorous methods involve solving the partial differential equa­tions for the electrostatic potential on a grid, and are quite computationally intensive. These methods, commonly called Poisson-Boltzmann (PB) solvers (42, 43), are useful in accurate examination of electrostatic potentials around static macromolecules and are not often used for dynamics. Even though there are no explicit water molecules in a PB calculation, the computation of the electrostatic potential at each dynamics step is too costly to offset the savings. The Generalized Born approximation is used to provide a much more efficient method for parameterizing the Poisson equation which is very close to the rigorous solu­tion and provides reasonable solvation energies and other thermodynamic solvation effects (44-46). The detailed interactions with individual water molecules are missing as are the hydrodynamic effects, but for many modeling problems, a solvated simulation that repro­duces the ensemble averages of an explicitly solvated system can be performed at one fourth the computational cost. A second benefit of implicit solvent calculations is that the solvent response to solute changes is instantaneous at each step, rather than requiring many picosec­onds (thousands of steps) of equilibration of the thousands of individual water molecules in an explicit-water simulation. This solvation model is also very useful in preparing a system for fully solvated modeling and for finding probable mechanisms and structures for more detailed studies.

Ferulate side chains are found in cereal and hardwood xylans, as well as in many types of pectin. In xylan, they are ester-linked to the C-2 position of the arabinose side chain of the xylan backbone, where they function as crosslinkers through ether linkages to either ferulic acid on another xylan chain or to lignin components (79). This provides some three­dimensional stability to the polymer network (80). Ferulic acid esterases (FAEs) are active in cleaving ferulic acid from these polymers, though the specificity is again not clear. Some have been reported to act on coumaric as well as ferulic acid (81). Some are preferentially active on polymers while others are more active on substituted xylo-oligomers. There are also reports of FAEs having activity on both ferulated xylan and ferulated pectin (82). Atomic force microscopy studies have suggested that hydrolysis of ferulic acid bridges results in shorter, less-branched xylan chains (83). Other studies have reinforced the synergy between xylanase and FAE, including enhanced synergy in an FAE/xylanase fusion protein (11, 67, 84-86).

While no one particular pretreatment process can presently be viewed as the “ideal” ap­proach for all feedstocks or for all process circumstances, a well-accepted list of the desired properties of an ideal pretreatment process has been generated (3). Such an ideal pretreat­ment process:

• Produces a highly digestible pretreated solid

• Does not significantly degrade pentoses

• Does not significantly inhibit subsequent fermentation steps

• Requires little or no size reduction of biomass feedstock

• Can work in reactors of reasonable size and moderate cost

• Produces no solid-waste residues

• Has a high degree of simplicity

• Is effective at low moisture content

14.2 Physicochemical properties of pretreated biomass believed to affect cellulose digestibility

A significant effort has been expended, aimed at increasing our understanding of the factors that are the most critical in controlling the susceptibility of cellulosic substrates to enzymatic hydrolysis. A wide variety of physical and chemical properties of pretreated biomass and the cellulose remaining after pretreatment have been studied and reviewed (4-7). Changes in the physicochemical properties of pretreated lignocellulosics have been correlated with the enzymatic hydrolysis of the cellulose. There are general conclusions that have come out of these reviews that should be considered when trying to identify the critical properties in pretreated biomass. As stated by Coughlan (4), “There is considerable disagreement in the literature regarding the relative importance of the various factors that render cellulose so recalcitrant to hydrolysis.” Mansfield et al. (5) cautions that “when contemplating these characteristics (structural and physicochemical features of the substrate) and identifying potential contributing factors or limitations (to enzymatic hydrolysis) care must be taken to consider some undisputable principles: (i) all samples of insoluble cellulose (both native and pretreated) are structurally non-uniform; (ii) the pretreatment method and conditions can effectively alter the structure of the original cellulose; (iii) native cellulose contains inherent regions of highly ordered and disordered molecular polymers (i. e., crystalline and amorphous regions); (iv) considerable attention must be paid to the anatomical and structural ‘levels’ of organization (i. e., microfibril, fibril, or fiber) which is being modified or characterized during hydrolysis…”

A key factor for successful enzymatic conversion of biomass to fermentable sugars is the accessibility of the [3(1—>4) glycosidic bonds in cellulose to cellulase enzymes. Pretreatment regimes must be designed to remove substrate-specific barriers to cellulases to improve cel­lulose digestion. The precise nature of the obstacles encountered by cellulases in the complex biomass ultra-structure remains ambiguous. The effect of pretreatment is typically evalu­ated on the basis of improved enzyme digestibility and downstream ethanol production. The link between changes in cell wall chemistry/structure and cellulase digestibility is ultimately dependent on improved access to the cellulose microfibril. Accurate and direct assessment

of changes in enzyme accessibility is challenging primarily due to the complexities of both the cellulase system and the biomass. (8).

Some of the factors that could influence the rate of enzymatic hydrolysis of cellulose in pretreated lignocellulosic feedstocks are cellulose crystallinity, degree of cellulose poly­merization, feedstock particle size, the lignin barrier (content and distribution), substrate available surface area (pore volume), and cell wall thickness (coarseness). In addition, irre­versible binding of enzymes onto lignin is also influenced by the nature of the substrate (6, 7). Typical physicochemical properties of biomass obtained from the pretreatments used by the CAFI group are shown in Table 14.1. Zhang and Lynd (9) have attempted to take this a step further by building functional models of cellulose hydrolysis that incorporate substrate features in addition to concentration and the activities of multiple cellulase components.