The systematic analyses ofvarious transgenic/mutant lines has identified important trends in lignin macromolecular assembly. These analyses (71, 72,131,132,215) have also eliminated non-monolignol entities, such as feruloyl tyramine (60), acetosyringone (61), etc. as being involved in core lignification in contrast to previous assertions (173, 174). Additionally, as summarized in Figure 7.14, modulation of the monolignol-forming pathways in the same organisms gave H, G, 5OH-G, and S-enriched lignins whose subset of identifiable, cleavable, interunit linkages were apparently invariant of hydroxylation/methoxylation patterns of aromatic ring substitution. In the case of H-monolignol deposition, however, this was prematurely terminated at a “metabolic checkpoint” (72, 215), the reasons for which need to be established. A similar situation also held for the po/y-p-hydroxycinnamaldehydes (71), with severe adverse effects being noted on plant structure overall and thus on vascular integrity. For the H, G, 5OH-G, and S enriched lignins, this subunit invariance would not be expected a priori as the H, G, 5OH-G, and S monolignols differ in having from 5 (H) to 3 (S) potential sites available for radical-radical coupling. Additionally, the catechol nature of 5-hydroxyconiferyl alcohol (4) represents yet another potentially confounding feature, as discussed earlier.

Yet the data obtained in lignin subunit characterization and frequency suggest a limited substrate degeneracy during the dehydropolymerization step, i. e., whereby the amounts of (cleavable) 8- O-4′ and 8-1′ interunit linkages present are kept directly proportional to lignin content, but apparently invariant of hydroxyl/methoxyl group aromatic ring substitution pattern. Moreover, for the H — and G-enriched lignins, the quantifiable amounts of releasable 8-5′, 5-5′, and 5-O-4′-derived dimers followed similar trends, whereas neither the 8-8′ linked ligballinol (58) and pinoresinol (69) were released in any appreciable amount. The latter is likely consistent with their presumed covalent modification during native lignin formation, perhaps at C-5. Analyses of the S-enriched Arabidopsis line also provided useful insights: While only ~32% of the linkages could be accounted thus far via thioacidolysis, the 8-8′ linked syringaresinol (70) subunits were only present in very small amount as their thioacidolysis derivatives. These data suggest that the S-enriched lignin was not composed primarily of S-units linked through 8-O-4′ and 8-8′ interunits as previously envisaged (i. e., substructures Ic and IIIc, Figure 7.2D), the reasons for which need to be determined. Additionally, the COMT knockout line (in Arabidopsis) resulted in an apparently equivalent reduction of G/S monomeric moieties leading to benzodioxane formation. Such data, for the COMT knockout mutant line in Arabidopsis, provisionally suggest that in the cell wall types normally forming S-units, each of the 5-hydroxyconiferyl alcohol (4) moieties is linked to a coniferyl alcohol (3) moiety; this is again envisaged to place another severe limitation on how native lignin macromolecular configuration is achieved.

Taken together, the data have nevertheless begun to provide useful insight into the trends involved in lignin macromolecular assembly, and give an impetus to explore and determine lignin primary structure. We do not consider, however, that these data are consistent with a “random assembly/combinatorial chemistry” model with 1066 or more isomers, but instead reflect limited substrate degeneracy during native lignin formation, as is observed in many areas of metabolism. It is presumably significant that several of the preceding enzymes in the monolignol-forming pathway (4CL, CCR, and CAD) are also substrate versatile. This, in turn, may have an important bearing on the limited substrate degeneracy observed for lignin macromolecular assembly.

Quantum mechanical calculations were used to obtain energies for reactants, transition states, and products so that reaction energetics and barriers could be obtained. The Gaussian 03 (29) suite of programs were used in our calculations, which were designed to obtain minima in potential energy surfaces corresponding to stable molecular species and saddle points that correspond to transition states. For this study, the hybrid density functional, B3LYP (25, 26), was used to obtain molecular geometries and energies. More accurate energies were obtained for the monosaccharides using the complete basis set (CBS) approach termed CBS-QB3 (30), in which the geometry is optimized using B3LYP/6-311G(d, p) and the energy is extrapolated to the complete basis set limit with MP2. Comparison of results from these techniques to the experimental values for the G2 set of molecules (31) shows that the standard deviation for the B3LYP technique with a split-level basis set (6-31G[d, p]) is about (32) 3 kcal mol-1 while the standard deviation for CBS-QB3 is (30) 1.2 kcal mol-1. The B3LYP technique can underestimate transition states (33-37) by up to 5 kcal mol-1, but the CBS-QB3 approach provides much better estimates of transition state energies (38).

The starting geometries for stable species were selected from the low energy conformers from CPMD calculations and literature results. The optimized geometries of reactants and products had no imaginary vibrational frequencies, whereas transition states had exactly one. Transition states were also confirmed by visual inspection of the motion of the imaginary frequency and by intrinsic reaction coordinate (IRC) calculations (39, 40). Reaction energy barriers, Ea, were determined as the difference in total energies of the transition state and the reactant, including the zero-point energy.

Fibrobacter succinogenes and F. intestinalis are specialized cellulolytic bacteria that belong to a divergent phylum of Gram-negative anaerobes (42). Significant Fibrobacter populations can be detected by molecular probing (8), although they are poorly represented in ampli­fied 16S clone libraries (43). Genome sequencing of F. succinogenes S85 has identified 113 genes that have likely roles in plant cell wall degradation, including 40 cellulases and 29 xy — lanases (44). As in Gram-positive cellulolytic anaerobes such as the ruminococci, complex organization involving multiple catalytic domain and substrate-binding modules is a feature of F. succinogenes enzymes (1, 45). Dockerin-like domains have not been detected, however, and the organization of cellulolytic enzymes on the cell surface remains unclear. Some en­zymes apparently share a basic C-terminal region, but it has not been established whether this is involved in cell surface attachment. Gene complement and cellulolytic activity are apparently well conserved among strains related to F. succinogenes S85, but with evidence of significant sequence divergence in F. intestinalis and some F. succinogenes-related strains (44, 46, 47).

F. succinogenes apparently lack xylose isomerase activity (48) and fail to utilize the break­down products of xylan, despite possessing an array of xylanases. R. flavefaciens strains vary in their ability to utilize xylo-oligosaccharides, and appear to vary in respect of their xylose isomerase genes (49). For these bacteria, xylanase activity, therefore, appears to be required primarily for degrading the matrix polysaccharides thus facilitating access to the glucan components of the plant cell wall.

Microorganisms associated with the rhizosphere have been demonstrated to increase root exudation through production of plant hormones and by physically damaging the roots (45). These actions create a nutrient-rich rhizosphere zone that is naturally colonized by many beneficial or sometimes pathogenic microorganisms. Bacteria and fungi have a considerable impact on plant growth, development, and productivity. The numerous inter­actions between bacteria, fungi, and roots may have beneficial, harmful, or neutral effects on the plant, the outcome being dependent on the type of symbiont interaction and the soil conditions (46). “Helper bacteria” may promote the formation of fungal communities in decaying biomass litter (47). Little is known about the ecology of these helper bacteria; however, an analogy to the role played by bacteria present in mycorrhizal fungi can be made (47). In the rhizosphere of plants, and perhaps in the decaying biomass pile, fungi are always accompanied by another important group of microorganisms, the bacteria which also prosper in the organic-rich environment (mostly sugars, amino acids, and organic acids) released from the roots and mycorrhizal fungi. Among them, there is a large cate­gory of growth-promoting rhizobacteria that influence plant growth directly or indirectly by releasing a variety of compounds, from mineral nutrients, to phytohormones, and an­timicrobial compounds. Rhizobacteria have been demonstrated to promote plant growth directly through production of plant hormones such as auxins (48), gibberellins (49), and ethylene (46). Production of indole-3-ethanol or indole-3-acetic acid (IAA), compounds belonging to the auxins, have been reported for several bacterial genera, such as Frankia (50-52), Klebsiella and Enterobacter (53), and Bacillus (49). Most importantly, for rhizobac — teria to act beneficially, they must be able to efficiently colonize and multiply in the plant rhizosphere.

Electron microscopic studies and molecular methods revealed large bacterial populations associated with mycorrhizal roots and extraradical hyphae (54-56). In some cases, bacterial endosymbionts were discovered in fungal hyphae (55,57). The establishment of mycorrhizas on roots is affected by the microbial populations of the rhizosphere, and especially by some bacteria, which can have either a positive or a negative effect on mycorrhiza formation. Garbaye (47, 58) defined a new bacterial category, the mycorrhization helper bacteria that strongly promoted ectomycorrhiza formation. FreyKlett and coworkers (59) suggested that these helper bacteria stimulate the growth of fungal mycelia, thus increasing the probability of a root-mycelium encounter. Some mycorrhiza helper bacteria were noncultivable, and were identified only by molecular ecological methods.

The primary walls of the grasses and most other commelinid monocotyledons except the palms (Arecaceae) are rich in glucuronoarabinoxylans (GAXs) (3) (Table 4.1). Examination
of the primary cell walls of the grass maize by transmission electron microscopy after prepar­ing them by fast freeze, deep etch, and rotary shadowing showed similar bridges between the cellulose microfibrils to those described for pectin-rich primary walls [Section 4.4.1.1, (151)]. Again, two co-extensive, but independent, polymer networks have been proposed as models of the architecture of these walls (152). The bridging molecules are considered to be GAXs carrying few side chains. GAX with low degrees of substitution may interact reversibly through hydrogen bonding with surfaces of cellulosic microfibrils and may do so in the twofold helical conformation (153). With increasing degrees of substitution the affinity for microfibril surfaces decreases, presumably due to the increase in the threefold conformation which would be unfavorable for association with cellulose and additionallyto increased steric hindrance to interaction imposed by the Araf substituents themselves. GAXs with low degrees of substitution with arabinosyl units, as well as (1^3,1^4)-p-D-glucans, have been assumed to coat the cellulose microfibrils (152). However, it is possible that, as in the primary walls ofmungbean (Vigna radiata) (150), non-cellulosic polysaccharides are ad­sorbed onto only a small proportion of the microfibril surfaces. Highly substituted GAXs and small proportions of pectic polysaccharides are considered to comprise the second network. The primary cell walls of all other commelinid monocotyledons (excluding the Arecaceae), including the other families that contain (1^3,1^4)-p-D-glucans in their cell walls (17), presumably have similar wall architectures. The high molecular weight (1^3,1^4)-p-D — glucans show high viscosities in solution and at high concentrations can be induced to form gels. Associations between (1^3,1^4)-p-D-glucan chains are proposed to be due to junction zone formation between pairs of consecutive cellotriosyl units (154) and puta­tively there are also non-covalent associations between the (1^3, 1^4)-p-D-glucan and heteroxylan components in the wall matrix (155).

Except for cellulose and callose, all of the other plant polysaccharides appear to be syn­thesized in the Golgi or to pass through the Golgi en route to the cell wall. The available experimental results do not currently lead to an unambiguous picture of how the diverse wall polysaccharides are synthesized in, and travel through the Golgi to the wall. Immunoelec — tron microscopy using antibodies to XG, HG, and RGI showed that these polysaccharides are found in Golgi vesicles but not in the ER (106,234). Different types of Golgi cisternae contain different sets of glycosyltransferases. Thus, the functional organization of the biosynthetic pathways of complex polysaccharides is consistent with these molecules being processed in a cis-to-trans direction like the N-linked glycans. RG-I and HG polysaccharides appear to be synthesized in cis — and medial-cisternae (235). Methylesterification of the carboxyl groups of the galacturonic acid residues in the polygalacturonic acid domains occurs mostly in medial cisternae, and arabinose-containing side chains of the polygalacturonic acid domains are added to the nascent polygalacturonic acid/rhamnogalacturonan-I molecules in the trans — cisternae. In root tip cortical parenchyma cells, anti-RG-I and the anti-XG antibodies are shown to bind to complementary subsets of Golgi cisternae, and several lines of indirect evidence suggest that these complex polysaccharides may also exit from different cisternae (224). On the other hand, xyloglucan and polygalacturonic acid/rhamnogalacturonan-I can be synthesized concomitantly within the same Golgi stack.

O-linked arabinosylation of the hydroxyproline residues of extensin occurs in cis — cisternae, and the glycosylated proteins pass through all cisternae before they are packed into secretory vesicles in the monensin-sensitive, trans-Golgi network (224). The (3-1,4-linked D-glucosyl backbone of xyloglucan is synthesized in trans-cisternae, and the terminal fucosyl residues on the trisaccharide side chains of xyloglucan are partly added in the trans-cisternae, and partly in the trans-Golgi network (235). It has been shown by immuno-electron mi­croscopy using anti-a-L-Fuc-(1^2)-D-Gal antibodies, that fucosylated XyG first appears in the lumen of the trans-Golgi and trans-Golgi network before vesicle mediated secretion to the cell wall (224, 400). This activity appeared to be spatially distinct from galactosyl — and xylosyltransferase activity (401, 402).

Rajai H. Atalla, John W. Brady, James F. Matthews, Shi-You Ding, and Michael E. Himmel

6.1 Introduction

In this chapter, we focus on the accessibility of native celluloses in plant cell walls to hy­drolytic agents and the manner in which accessibility is modified by dehydration and thermal chemical pretreatments. The primary barrier to enzymatic hydrolysis of celluloses in living plants is their encrustation with lignin. To overcome this barrier, it is necessary to remove the lignin chemically or to fragment the tissue to expose the cellulose; that is, the cell wall must be deconstructed chemically or mechanically. Chemical deconstruction is usually carried out at high temperatures. Mechanical deconstruction is also usually carried out at elevated temperatures to facilitate removal of hemicelluloses, if the primary objective is hydrolysis of the cellulose to glucose. It is therefore important to assess the effects of temperature on the state of aggregation of cellulose. Past studies have shown that temperature elevation almost always results in tighter aggregation of the cellulose chains in the microfibrils in a manner that reduces their accessibility to hydrolytic agents. It is important, therefore, to understand the state of aggregation in the living plant prior to dehydration and to better understand the transformations that arise as a consequence of dehydration in the course of pretreatment of plant biomass.

Microscopic evidence suggests that the microfibrils and nanofibrils of cellulose in higher plants possess a long-period helical twist in their native state. Though the microscopic ev­idence has revealed the twist in bacterial and algal celluloses, recent theoretical analyses indicate that higher plant celluloses also possess a helical twist that is more pronounced because of its shorter period, and this indeed has been confirmed through atomic force microscope (AFM) imaging. It is important, therefore, to review the evidence regarding the helical twist characteristic of the native state and consider how the highly ordered biolog­ical structures, which are species — and tissue-specific, are transformed as a consequence of dehydration into states that are less species — and tissue-specific. We will also consider the driving forces responsible for the dehydration. Finally, we will consider how pretreatment processes might be modified to preserve as much as possible the inherent accessibility of celluloses in their native state.

Biomass Recalcitrance: Deconstructing the Plant Cell Wall for Bioenergy. Edited by Michael. E. Himmel © 2008 Blackwell Publishing Ltd. ISBN: 978-1-405-16360-6

pC3H is also considered to have an important regulatory role in carbon allocation (77) to the monolignol/lignin-forming pathway, at least as far as the G and S segments are involved. Two different mutant/transgenic lines (Arabidopsis and alfalfa) altered in pC3H activity have thus far been obtained using standard biotechnological manipulations (72, 213, 214). These manipulations resulted in one of the most dramatic examples of disparity seen in morphology between two different species, i. e., as a result of the same genetic target (pC3H) (72,215).

In this regard, the chemically generated (ethane methyl sulfonate, EMS) Arabidopsis ref8 mutant, isolated following screening of ~ 100 000 plants, contains a random mutation in the pC3H gene (213). The line was severely dwarfed (Figure 7.13A), reaching only ~3% of the height of wild type at maturity (215). It was also sterile and its proto-/metaxylem and xylary fiber elements had collapsed in the stem cross-sections examined (Figure 7.12B). This latter observation presumably reflects the very low levels of lignin, reduced by ~64% relative to wild type, and which is primarily of H character (H:G:S ratios of ~85:8:7) (215). This observation contrasts with previous assertions that the ref8 lignin was exclusively H-derived (175,213).

Although pC3H has a rate-limiting role in lignin biosynthesis and can thus be expected to produce vascular defects, the multiple and extreme morphological characteristics of ref8 suggested that additional pleiotropic effects were manifested in this line (215). Because of tissue availability limitations, the ref8 line was not readily amenable to more detailed analy­ses and thus these questions were not explored further. On the other hand, the alfalfa pC3H-I line was generated using a standard antisense transgenic method (214) with presumably less “unintended” effects resulting than that of chemical mutagenesis. In stark contrast to the Arabidopsis ref8, the alfalfa pC3H-I has a phenotype nearly like that of wild type (Figure 7.13B), with vessel anatomy also being very similar to wild type (data not shown) (72).

The remarkably different outcomes of modulation of pC3H between the Arabidopsis ref8 mutant and alfalfa pC3H-I lines (215) reflect further the weaknesses of current methods in the genetic manipulation of plants, and of our very limited understanding of said ma­nipulations. Indeed, it is for reasons such as the unknown (pleiotropic) effects observed herein that some segments of humanity currently distrust the use of genetically modified organisms. From a more scientific perspective, these limitations also highlight the need for comprehensive, careful, systematic data collection in order to more accurately assess,

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

Figure 7.13 (continued) of Botany, vol. 94, Patten, A. M., Jourdes, M., Brown, E. E., Laborie, M.-P., Davin, L. B. & Lewis, N. G., Reaction tissue formation and stem tensile modulus properties in wild type and p — coumarate-3-hydroxylase downregulated lines of alfalfa, Medicago sativa (Fabaceae), pp. 912-925, Copy­right 2007, with permission from the Botanical Society of America. (D) Plant Journal, vol. 13, Piquemal, J., Lapierre, C., Myton, K., O’Connell, A., Schuch, W., Grima-Pettenati, J. & Boudet, A.-M., Downregulation of cinnamoyl-CoA reductase induces significant changes of lignin profiles in transgenic tobacco plants, pp. 71-83, Copyright 1998, with permission from Blackwell;and (F) The Plant Cell, vol. 19, Mitsuda, N., Iwase, A., Yamamoto, H., Yoshida, M., Seki, M., Shinozaki, K. & Ohme-Takagi, M., NAC transcription factors, NST1 and NST3, are key regulators of the formation of secondary walls in woody tissues of Ara­bidopsis, pp. 270-280, Copyright 2007, with permission from the American Society of Plant Biologists.] (Reproduced in color as Plate 22.) analyze, and understand the effects of such manipulations. Clearly, we need to get such manipulations “right,” for example, before contemplating dedicating large swaths of lands to lignocellulosic-modified organisms.

The alfalfa pC3H line also had lignin contents significantly reduced (>68% of wild type) (72, 215), which were again mainly of H-character (H:G:S ratios of ~77:12:11) as to be predicted (31,34,35,77). Significantly, biomechanical testing of stem vascular tissue integrity showed essentially no differences in tensile dynamic modulus properties, at least under the test conditions employed (72), i. e., where both the pC3H-I and wild-type lines apparently displayed more or less equivalent structural integrity of the vascular apparatus. More detailed anatomical analyses ofthe stem tissue inboththe pC3H-I and wild-type lines also established that the vasculature contained lignin-deficient, cellulose-enriched, gelatinous layers that are characteristic of reaction (tension) wood (see Figure 7.12C). Moreover, this occurred both earlier in growth/development and to a greater extent in the pC3H-I line (72), suggesting that vascular/structural properties are (at least in part) maintained by partially “switching” metabolic outcome to that of reaction tissue formation. Although the latter tissue is (both constitutively and inducibly) employed to control branching and realignment of leaning stem(s) back to the vertical, there are many applications, such as in forestry and in pulp/paper manufacture, where its formation is not desirable; gelatinous fibers, for example, result in poor paper strength quality due to decreased bonding properties (216).

Analyses of the lignins in both pC3H-I and wild-type lines were most instructive: the 13C NMR spectroscopic analysis of solubilized lignin derivatives established the presence of a H-enriched lignin (215) in pC3H-I as predicted (31, 34, 35, 77), with substructures I-VIII (Figure 7.2D) being in evidence. More importantly, the thioacidolysis degradation products (monomers and dimers) released, which corresponded to cleavage products from subsets of 8-0-4′, 8-5′, 5-0-4′, 3-3′ (5-5′), and 8-1′ (substructures I/II, IV, VI, VII, and VIII in Figure 7.2D) lignin interunit linkages, were determined quantitatively at vari­ous stages of stem growth/development/lignification until maturation (215) (Figures 7.14A, F-I). The frequency of the interunit linkages indicated that they were directly proportional to increasing lignin content, at least for the (monomeric/dimeric) cleavable 8-0-4′, 8-1′, 8-5′, 5-0-4′, and 3-3′ (5-5′) moieties being released. Significantly, the 8-8′-linked dimer (lig — ballinol 58) was not released as such (Figure 7.14J), even though its “substructure IIIa” was readily detectable by NMR spectroscopic analyses. This observation is also consistent with the previous “conundrum” of pinoresinol-like substructures IIIb not being released either from native lignins, whether by acidolysis (193) or thioacidolysis (217). These substructure­like moieties are, nevertheless, also readily detectable by NMR spectroscopic analyses. In direct contrast, synthetic lignin DHPs readily release such entities (218) upon cleavage. This, in turn, is presumably indicative of a very different mode of lignin assembly/organization in vivo than that occurring randomly in vitro.

Moreover, when compared to the lignin levels and interunit linkage frequencies in the wild-type line, the same correlations and trends exist for all levels of lignin deposition, with the amounts of individual products released and quantified apparently being invariant of methoxyl group composition. Significantly, the H-monolignol (1) did not seamlessly replace the G/S components (3/5), and thus presumably was unable to extend deposition into the domains normally designated for G/S monolignols in the lignifying cell wall(s). This, in turn, indicated that the H-monolignols (1) presumably could not substitute fully for either coniferyl (3) or sinapyl (5) alcohols, this representing yet another constraint on

Figure 7.14 Correlations of current best estimates of amounts of (A-E) total monomeric derivatives re­leasable by thioacidolysis as a function of estimated AcBr lignin contents);(F) 8-1′ dimeric derivatives after thioacidolysis follow by desulphurization as a function of estimated AcBr lignin contents;8-5′ (G), 5-5′ (H) and 5-0-4′ (I) dimeric derivatives after thioacidolysis followed by desulphurization as a function of total releasable monomeric derivatives by thioacidolysis;(J) 8-8′ dimeric derivatives after thioacidolysis followed by desulphurization as a function of total S unit releasable monomeric derivatives by thioacidol­ysis.

the notion of a seamless random coupling/combinatorial chemistry being in effect (175). On the other hand, the overall amounts of H-moieties appeared to increase slightly over that of wild-type levels, but where designated patterns of a subset of interunit linkage frequencies (for the products released) were being maintained (relative to wild-type lignin) until the overall deposition process was prematurely terminated — for reasons still to be determined. Such reasons for terminating this process could include the plant perceiving that the biophysical properties of the H-lignin being deposited in the cell wall were, for example, limiting and/or defective in some way, e. g., through feedback inhibition. Nevertheless, the partial replacement of G/S moieties by H-monomers can presumably be explained via limited substrate degeneracy during lignin template polymerization.

Reduction in lignin contents have also been reported with HCT silencing in Nicotiana benthamiana and Arabidopsis (219, 220). For example, RNA silencing was achieved in Ara — bidopsis (ecotype Columbia) by transformation with a hairpin repeat of a portion of the HCT gene whereas in N. benthamiana, a tobacco rattle virus (TRV)-based virus-induced gene silencing (VIGS) system was used (219). In Arabidopsis, this resulted in stunted plants with apparently reduced lignin levels based on collapsed xylem cells (220). The yields of thioacidolysis monomers 54-56 of stems from 2-month-old plants were, however, very low, i. e., ~75.7 ^mol g-1 dry sample for the wild type and only ~1.0 ^mol g-1 dry sample for the silenced line (HCT-). Such values for the wild type (control) are uncharacteristically low and presumably indicate that the lines were of an unknown maturation state, and/or were defective in some manner, and/or the analyses had yielded unreliable data; in our studies and others, the thioacidolysis yields are reproducibly ~350 ^mol g-1 dry sample (cell wall residue, CWR) for wild-type lines at plant maturity (see Figures 7.14A-E). Thus, these data again underscore the need to determine at the minimum lignin contents and compositions at several stages until maturation has definitively been reached, rather than the single point analyses that have become typical of most studies in this field. On the other hand, the H:G:S ratios of 1:83:16 in wild type were now 83:12:2 in the HCT- line in agreement with HCT involvement in G/S lignin biosynthesis. In N. benthamiana, plants ranging from no visual growth phenotype to severely stunted phenotypes were also obtained, yet with estimated lignin levels decreased only by ~ 15% of wild type in the HCT-silenced plants as measured by the Klason method, i. e., the results for both species were apparently quite different in terms of effects on lignification. Although the reasons are not yet known for these differences, Arabidopsis and tobacco differ in having only one HCT homologue in the former, whereas in the latter both HCT/HQT enzymes are present. Yet, HQT silencing was also studied in tomato (Lycopersicum esculentum) where it was shown to only affect chlorogenic acid (26) levels but not lignin content/composition (111).

As mentioned above, one of the biggest limitations of classical MD simulations is the inability to make and break covalent bonds during a simulation. A potential solution to this is to describe the system quantum mechanically instead of classically. This would seem to be an ideal situation since then electron densities are explicitly included in the calculation. However, such simulations are far too computationally expensive to be used for molecular dynamics simulations of proteins. An alternative approach, first proposed in 1976 by Warshel and Levitt (59), was to combine a quantum mechanical (QM) potential with a molecular mechanical (MM) potential to form a hybrid QM/MM potential. In this approach, the parts of the protein and substrate that are directly involved in the enzyme reaction are calculated using QM potential functions, and the remaining atoms are treated using a classical MM potential. The coupling of a QM and MM potential allows just the reaction center to be studied quantum mechanically while keeping the calculation complexity low by using a more approximate MM potential elsewhere. This partitioning of the system allows calculations on systems significantly larger than would be possible with pure QM approaches and at the same time enables calculations such as reactions to be studied for which classical MM potentials are not appropriate. A number of commonly used MD codes provide support for QM/MM simulations including AMBER and CHARMM. QM/MM simulations still remain relatively expensive, however, although recent advances in AMBER (60) are bringing the cost of QM/MM simulations, for systems with up to 100 QM atoms, to a point where the cost is approximately double what the standard classical simulation would cost. Thus, QM/MM approaches will form an important tool for looking at the actual mechanism of hydrolysis within a cellulase enzyme as it degrades cellulose fibers.

Endomannanases (EC 3.2.1.78) catalyze the random hydrolysis of p-D-1,4 mannopyranosyl linkages within the main chain of mannans and various polysaccharides consisting mainly of mannose, such as glucomannans and galactoglucomannans. Far fewer mannanases are characterized as compared with the numerous xylanases. The mannanase of Trichoderma reesei has a similar multi-domain structure to several cellulolytic enzymes; i. e., the protein contains a catalytic core domain which is connected by a linker to a cellulose-binding domain (20,21). The CBD increases the action of T. reesei mannanase onfiber-bound glucomannan, even though the catalytic domain can efficiently degrade crystalline mannan (22).

The main hydrolysis products from galactoglucomannans and glucomannans are manno — biose, mannotriose, and various mixed oligosaccharides. The hydrolysis yield is dependent on the degree of substitution as well as on the distribution of the substituents (23). The hydrolysis of glucomannans is also affected by the glucose to mannose ratio. Some man — nanases are able to hydrolyze not only the (3 -1,4-bond between two mannose units, but also the bond between the adjacent glucose and mannose units (24,25). Interestingly, some mannanases are able to degrade mannan crystals quite efficiently (26, 27).

Enzymes needed for further hydrolysis of the released manno-oligomers produced by endoenzymes are p-mannosidase (1,4-p-D-mannoside mannohydrolase, EC 3.1.1.25) and p-glucosidase (EC 3.2.1.21). p-mannosidase and p-glucosidase catalyze the hydrolysis of manno-oligosaccharides by removing successive mannose or glucose residues from the non-reducing termini. The p — xylosidase of T. reesei and p — mannosidases of A. niger are also able to attack polymeric xylan and mannan, respectively, liberating xylose and mannose by successive exo-action (28, 29).