2.1.1

Engineering Scheme

The development of ethanologenic E. coli has included a combination of dir­ected engineering and metabolic evolution; the overall scheme is summarized in Fig. 3. The Z. mobilis homoethanol pathway (PET operon) was introduced

into E. coli in plasmids and these derivatives produced ethanol as the main fermentation product [13-16]. The PET operon was stably integrated into the chromosome at the pfl locus along with an antibiotic resistance marker; spon­taneous mutants exhibiting high ADH activity and high antibiotic resistance were selected to ensure high PET activity. Side reactions that drain carbon away from ethanol were eliminated either by mutation (frd — succinate) or physiologically (differences in Km for pyruvate) (Fig. 2). The resulting strain KO11 produced ethanol at a yield of 95% in complex media [17]. While it was originally reported that KO11 was derived from E coli B, it has recently been discovered that E coli W is the parental strain (Jarboe and Ingram, unpub­lished).

While the rate of ethanol production by KO11 is as high as yeast, the ethanol tolerance is lower than the commercially employed yeast strains. In complex media, KO11 shows a complete lack of growth in the presence of 35 gL-1 ethanol and only 10% survival from 30 s of exposure to 100 gL-1 ethanol [18]. Using strain KO11 as a starting point, mutant strains with sig­nificantly increased ethanol tolerance were isolated. The 3-month metabolic evolution consisted of alternating periods of selection in liquid media for in­creased ethanol tolerance and selection on solid media for increased ethanol production. The final product of this evolution, strain LY01, was able to grow in the presence of 50 gL-1 ethanol and had greater than 80% survival from 30 s of exposure to 100 g L-1 ethanol. The method of metabolic evolution used to derive LY01 from KO11 has proved to be successful and has been applied to the design of other ethanologenic biocatalysts and to the production of other commodity products, as described in Sect. 5.

2.1.2

R-phenylacetylcarbinol (PAC), an intermediate in the production of ephedrine and pseudoephedrine, is currently produced by the controlled addition of benzaldehyde to an actively growing culture of yeast (usually Saccharomyces cerevisiae). A decarboxylation/condensation biotransformation is effected by pyruvate decarboxylase (PDC) between pyruvate produced by the yeast and added benzaldehyde (see Fig. 9). Using this traditional process, 12-15 gL-1 PAC is usually produced in 10-12 h with a yield of 70% theoretical based on benzaldehyde [100].

Confirmation of PAC production from benzaldehyde and pyruvate using purified PDC from various sources, including Z. mobilis, S. carlsbergenis, S. cerevisiae, S. fermentati and S. delbrueckii, was demonstrated by sev­eral groups during the late 1980s to mid 1990s [101-105]. Bringer-Meyer et al. [106] isolated and characterized PDC obtained from Z. mobilis. By comparison with yeast PDC (Saccharomyces sp., Candida sp.), bacterial PDC (Zymomonas sp.) had a lower benzaldehyde affinity and was inhibited more strongly by benzaldehyde, even though its affinity for pyruvate was similar or higher than that of yeast PDC.

However, interest in PDC from Z. mobilis continued due to its greater sta­bility than yeast PDC at room temperature with an enzyme half-life in the absence of benzaldehyde of over 100 h [107,108]. Unlike yeast PDC, it is also able to utilize the lower cost acetaldehyde as an alternative substrate to pyru­vate for production of PAC [109]. Advances in site-directed mutagenesis tech­niques have facilitated the production of mutant PDC from Z. mobilis with greater carboligase activity and higher stability towards acetaldehyde [110]. This mutant enzyme, designated PDCW392M, resulted from replacement of the bulky tryptophan residue 392 with methionine. A continuous process with PDCW392M was used in a biotransformation process for conversion of acetaldehyde and benzaldehyde to PAC in an enzyme membrane reactor. A volumetric productivity (space-time yield) of 81 gl-1 day-1 was reported with final PAC concentration of 22 mM and molar yields of 45% (initial sub­strates), based on 50 mM reaction mixture of both aldehydes [111,112].

In further studies by Rosche et al. [113], a biphasic enzymatic biotransfor­mation system for production of PAC from acetaldehyde and benzaldehyde with Z. mobilis PDCW392 was evaluated. Higher concentrations of benzalde — hyde and PAC in the organic phase (octanol) provided protection for the aqueous phase PDC. As a result, a specific PAC production of 11 mg PAC U PDC-1 was achieved compared with 1.2 mg PAC U PDC-1 in the absence of an organic phase. A similar two-phase system has been developed sub­sequently for conversion of pyruvate and benzaldehyde to PAC using PDC from yeast (C. utilis) with higher concentrations and productivities being at­tained [114,115].

A similar aqueous/organic two-phase system has been used also to screen a number of yeasts and bacteria for the enantio-specific reduction of the al­pha, beta-unsaturated carbon bond in citral to produce citronellal [116]. In comparison to the bacteria tested, the eukaryotes showed at least 5-fold lower citral reductase activities. Bacterial strains were found to produce the (S)- enantiomer of citronellal preferentially with ee values > 99% for Z. mobilis and 75% for Citronella freundii. The possible use of a Z. mobilis biofilm biore­actor for production of other fine chemicals has been proposed also [117] as it has been demonstrated that increased tolerance to aromatic substrates such as benzaldehyde can occur with such a bioreactor.

5

The tested thermostable fungal enzymes, classified as cellobiohydrolases (CBHs) or endoglucanases (EGs) based on the activity determinations, were chosen by preliminary screening and characterisation. Several thermostable CBHs from various thermophilic organisms were purified and characterised (Voutilainen et al, manuscript in preparation). The gene of the most poten­tial CBH isolated from Thermoascus aurantiacus was fused with the T. reesei CBHI cellulose binding domain (CBM). In addition, an EG from Acremo — nium thermophilum, a в-glucosidase and a xylanase from T. aurantiacus were used to compose the thermostable mixtures. Fermenter supernatants pro­duced in pilot scale were used to obtain the thermostable cellulase mixtures. The optimal ratio of EG to CBH amount (measured as protein of the enzyme mixtures) was determined on the basis of FPU activity of the preparations. The highest FPU activity was obtained by an EG to CBH protein ratio of 3 : 10, which corresponded well to the respective ratio of the native T. ree — sei enzymes. This ratio also gave the highest sugar yields in the hydrolysis of the steam pretreated corn stover substrate (results not shown) and was used as the standard basis for various mixtures. Three different mixtures were used in this work, differing with respect to the xylanase activity (Table 3). The xylanase-free preparation (TM 1) was first used for the spruce substrate

and the xylanase-containing preparations (TM 2 and TM 3) for the corn stover substrate. As it has frequently been observed that xylanases enhance the hydrolysis of lignocellulosic substrates containing even low amounts of re­sidual xylan [9], preparations with xylanase activity were later used for both substrates.

7

Although S. cerevisiae cannot grow on D-xylose as the sole carbon source, its genome does contain genes that code for a non-specific NADH-dependent al­dose reductase (GRE3) and for a xylitol dehydrogenase (XYL2). It has been shown that over-expression of these native S. cerevisiae genes using endoge­nous promoters enabled a specific growth rate of 0.01 h-1 on D-xylose in shake flasks [64]. However, in these shake-flask cultures this engineered yeast strain converted D-xylose into xylitol with a yield of 55%. Under anaero­bic conditions, precluding respiratory NAD+ regeneration, the strain over­expressing the endogenous enzymes was unable to utilise D-xylose [64].

In addition to this metabolic engineering approach, the presence of en­dogenous genes for D-xylose-converting enzymes has been used in recent experiments by Attfield and Bell (2006), describing a non-recombinant S. cere- visiae strain that grows on D-xylose as the sole carbon source in aerobic shake flask cultures. In their study a combination of population genetics and evolutionary engineering [5,60] resulted in an increase in growth rate from extremely low, barely measurable growth rates to a specific growth rate of around 0.12 h-1 (a doubling time of less than 6 h) over a period of 1400 days. Apparently, this S. cerevisiae strain had evolved in such a way that the very low “background” xylose reductase and xylitol dehydrogenase activities, which were previously described as insufficient for growth on D-xylose [8], increased to levels that did enable growth. Indeed, subsequent analysis of the evolved strain showed that xylose reductase activity had increased fourfold and the xylitol dehydrogenase activity 80-fold relative to the parental strain. The ac­tual genes that underwent mutation have not yet been characterised. Although this very interesting study underlines the tremendous potential of evolution­ary approaches, the selection procedure inevitably resulted in a yeast strain displaying the characteristics of redox imbalance, such as xylitol production.

1.4

While hemicellulose represents a large potential biomass source that is not presently utilized, pretreatment is required for depolymerization of its sol­uble components. Many depolymerization techniques are available, but re­search in this laboratory has focused on hydrolysis with dilute mineral acid at modest temperatures [85,86]. Unfortunately, dilute acid hydrolysis produces toxins that negatively affect biocatalyst growth and metabolism (reviewed in [87]); many of these toxins are listed in Fig. 1. Recent work has focused on an increased understanding of the underlying mechanisms of toxicity and methods for toxicity quantification and reduction.

Furfural, a pentose sugar derivative, is present in hemicellulose hydrolysate at a concentration of 1-4 gL-1 [88] but can inhibit E. coli growth at con­centrations as low as 2.4 gL-1 [89,90]. While other aldehydes, such as 4- hydroxybenzaldehyde and syringaldehyde, are more toxic than furfural on a weight basis, the presence of furfural enhances the effect of other tox­ins [90]. Despite the observed toxicity, ethanologenic E. coli KO11 and LY01 and K. oxytoca P2 have demonstrated a native ability to transform furfural to furfuryl alcohol [91]; the size and substrate specificity of the LY01 furfural re­ductase suggests that it is a new type of alcohol-aldehyde oxidoreductase [92]. Strain LY01, which has higher ethanol tolerance than KO11, also has higher furfural tolerance: KO11 growth was completely inhibited by 3 gL-1 furfural but LY01 was not, although growth was reduced by more than 50% [90]. Con­trastingly, there is no difference in the syringaldehyde tolerance of the two strains [90].

The toxicity of representative alcohol, aldehyde, and acid components of hemicellulose hydrolysis were investigated and found to affect ethanologenic E. coli LY01 in various ways [90,93,94]. In all cases, toxicity was related to hy — drophobicity. The organic acid data suggests that aliphatic and mononuclear acids both inhibit biocatalyst growth and ethanol production by collapsing ion gradients and increasing the internal anion concentration, and not by in­hibiting central metabolic or energy pathways [93]. At least some inhibitors are present at sufficient concentrations to account for the observed growth in­hibition: 9 g L-1 of acetic acid in rich media inhibits LY01 growth by 50%, and acetic acid concentration in hydrolysate can exceed 10 gL-1.

While all of the tested aldehydes did inhibit growth, only furfural had an impact on ethanol production [90]. Alcohols have a lower toxicity than alde­hydes and acids and appeared to inhibit ethanol production primarily by inhibiting growth [94].

Total furan content is representative of total toxicity and can be estimated from UV spectra [95]. The adjustment of hydrolysate pH to 9-10 by the add­ition of Ca(OH)2, a process known as overliming, is an effective method of hydrolysate toxicity reduction [96]. LY01 was able to produce less than 1 g L-1 ethanol from hydrolysate adjusted only to pH 6.5-6.7 but produced 33 gL 1 ethanol from baggase hydrolysate that was overlimed to pH 11 [97].

4.4

Despite the high interest and rapid development of biomass-based fuels, it is not anticipated that oil-based fuels will be completely replaced by renewable fuels in a foreseeable future of 50 years [29]. Conventional refineries convert­ing crude oil to fuels, starting chemicals, and other products, therefore, will operate decades ahead. In the biorefinery literature it has been a common practice to compare conventional petroleum-based refineries with biorefiner­
ies [30,31], but to our knowledge a combination of the two refinery types has not previously been suggested. Integrating conventional and bio/catalytic refineries in the transition period from petroleum-based to biomass-based refineries might lead to several potential synergies with respect to processes, chemicals, and logistics.

Several process streams of intermediates, wastes, and heat from a conven­tional refinery might be utilized in a biorefinery (Fig. 6). Cooling water and some effluent water streams can be used as process water in the biorefinery. A conventional refinery has big volumes of low temperature energy, which could be exchanged and used as process energy in the biorefinery.

Products from the biorefinery can be used as input for various conven­tional refinery processes. As discussed elsewhere in this book, ethanol is mainly used as a blending component in gasoline products. Integrating the two refineries will improve the logistics of this mixed fuel production.

Hydrogen produced from fermentation processes of the biorefinery can for instance be used in the traditional hydrogenation processes of a conventional refinery. Methane produced in the biorefinery can be used as fuel gas, but also as a raw material for further catalytic reforming, producing more hydro­gen. It could also be used for production of H2/CO (synthetic gas), which is a feed gas for gas-to-liquids or methanol production. Introducing catalytic steps between the two refineries might further enhance the beneficial coup­ling since the hydrocarbon output from catalytic conversion of methane and ethanol might serve as a substrate for further refining and modification in the conventional refinery process streams.

Heat waste streams

Ethano

Fig.6 Combination of bio/catalytic refinery and petroleum-based refinery. cat indicates chemical catalytic conversion

5

Conclusion

In this chapter we have shown the potential of producing more than bioethanol out of biomass raw material. While carbohydrates will be the precursor for ethanol production, the rest of the biomass can be used for production of other fuels. By this integration the net energy production will increase and the CO2 reduction will be higher than in biorefineries with­out the integration. Furthermore, reuse of water and nutrient will allow for a more sustainable process with much lower environmental impact on the ecosystem.

Advances in the fields of genomics and metagenomics have dramatically re­vised our view of microbial biodiversity and the potential for biotechnological applications. In the last decade the revolution in computer processor speeds, memory storage capability, and expanding networks has made possible the large scale sequencing of genomes and management of large integrated databases over the Internet. Since the first microorganism, Haemophilus in­fluenzae, was sequenced in 1995, genome sequencing initiatives have resulted in over 300 sequenced organisms, including 27 archaeal, 337 bacterial, and 41 eukaryotic genomes. As of July 2006, more than 1500 prokaryotic and eukaryotic genome sequencing projects are underway [70,71]. The genome sequences of Escherichia coli and Saccharomyces cerevisiae were not only among the first to be published, but were also the first of wide relevance for the production of industrial biochemicals, including bioethanol. Given that the genome of a particular microorganism, following annotation, provides the theoretical enzyme reaction set, it serves as a preferred starting point for engineering metabolic pathways that will lead to significantly improved titer, yield, productivity, and performance of a microorganism [62].

Annotated genomes certainly compliment experimental designs; however, the design space that can be considered by visual inspection or classical hy­pothesis driven experimentation is limited given the high degree of connec­tivity of the metabolic network. Modifying a given enzyme or metabolite pool is likely to elicit a multilayered regulatory response that not only mitigates the original perturbation, but will shift the equilibrium of other enzymes, metabolite pools, or signalling pathways. To a large extent, this is why ran­dom mutagenesis approaches have been favored over targeted approaches, until recently. The first genome-scale in silico metabolic network model for E. coli was made available in 2000 and was among the first to demonstrate consistency between modeling predictions and in vivo physiology [72,73]. Specifically, the model was used to explore the relationship between acetate, succinate, and oxygen uptake rates when attempting to maximize growth rate, to confirm the hypothesis that E. coli under acetate and succinate car­bon limitations regulates its metabolic network to maximize growth rate. For industrial biotechnology process development, it is often desirable to shift carbon flux from biomass to product formation, thereby maximizing the yield of product on substrate.

The first eukaryotic genome-scale metabolic model was reported in S. cere — visiae in 2003 based on its annotated genome sequence and a thorough examination of online pathway databases, biochemistry textbooks, and jour­nal publications [74]. This genome-scale in silico model, by using a relatively simple synthetic medium, could predict 88% of the growth phenotypes cor­rectly, indicating that this model in many cases can predict cellular behavior. In one step further, Duarte et al. (2004) [74] used the S. cerevisiae genome — scale metabolic network constructed by Forster et al. (2003) [75] to generate a phenotypic phase plane (PhPP) analysis that describes yeast’s metabolic states at various levels of glucose and oxygen availability. Examination of the S. cerevisiae PhPP has led to the identification of two lines of optimal­ity: LOgrowth, which represents optimal biomass production during aerobic, glucose-limited growth, and LOethanol, which corresponds to both maximal ethanol production and optimal growth during microaerobic conditions. The predictions of the S. cerevisiae PhPP and genome scale model were compared to independent experimental data, and the results showed strong agreement between the computed and measured specific growth rates, uptake rates, and secretion rates. Thus, genome-scale in silico models can be used to system­atically reconcile existing data available for S. cerevisiae, particularly now that yeast resources, databases, and tools for global analysis of genomic data have been expanded and made publicly available, such as the Saccharomyces Genome Database [70,71].

Another major challenge of current biotechnology, especially in the lignocellulose-to-ethanol process, is to identify novel biocatalysts and en­zymes for enzymatic hydrolysis from the genomes of organisms and metage­nomic libraries. A large number of protein sequences deduced from the genomes of industrial microorganisms have no assigned function, as they exhibit low (or no) homology to the enzymes and/or proteins functionally characterized thus far [76]. The demand for identification of novel biomass­degrading enzymes as well as for heterologous protein production at higher efficiencies and reduced costs has catalyzed an interest in elucidating the genomic sequence of Trichoderma reesei — the most prolific producer of biomass-degrading enzymes. Diener et al. (2004) [77] has described the creation of a T. reesei strain QM6A large-insert BAC (bacterial artificial chro­mosome) library and its subsequent analysis, which was successfully used to identify both biomass degradation and secretion related genes. These data re­vealed the utility of a BAC library for use in assembly of the T. reesei genome and isolation of genomic sequences of industrial interest.

Even though the above study represents a direct application of sequenc­ing technology for identification of novel biomass-degrading enzymes, it is also often the case to use such high-throughput experimental techniques to elucidate mechanistic understanding of enzymes derived from random, nat­ural selective pressures. The research of Foreman et al. (2003) [78] using

T. reesei RL-P37, a strain that has been selected for improved production of

cellulolytic enzymes [79], is such an example. The mutation(s) that improved cellulase production concurrently improved the inducible expression of ancil­lary genes that do not have a known function in cellulose degradation. These results suggest significant regulatory points of convergence across the spec­trum of cellular processes involved in carbon sensing, signal transduction, and transcriptional regulation. These findings will likely have significant im­plications for the design of industrial processes for commercial production of biomass-degrading enzymes.

In conclusion, the vastly improved computational capability integrated with large-scale miniaturization of biochemical techniques such as BAC, PCR, and microarray chips has delivered significant amounts of genomic data to researchers all over the world [80]. This availability of data has led to an ex­plosion of genome analysis leading to many new discoveries and tools that are not possible in exclusively wet-lab experiments.

It is evident from the above applications of genomics coupled to in silico modeling that industrial biotechnology, and especially bioethanol produc­tion, can benefit from this technology platform both in the identification of metabolic engineering target genes to improve yields, titer, and productivity, and in the discovery of novel enzymatic catalysts. This is further reinforced by the various case studies to be presented in subsequent chapters, including the role genomics has played in the identification of thermostable cellulases, metabolic engineering for pentose and xylose utilization in S. cerevisiae and Pichia stipitis, development of ethanologenic bacteria, and development of Z. mobilis for bioethanol production.

3.2

While possible variations in the process of converting lignocellulosic biomass to ethanol are virtually endless, it can most simply be described as the in­tegration of five unit operations: (1) desizing, (2) thermochemical pretreat­ment, (3) enzymatic hydrolysis, (4) fermentation, and (5) ethanol recovery (Fig. 1). In the first step of the process, the delivered biomass must be made uniform in size to facilitate handling and transport via conveyor or screw drive and to provide a more consistent surface-to-mass ratio for thermo­chemical pretreatment. The pretreatment step is typically a short — (minutes) to long-term (hours) exposure to extremes of temperature (150-200 C), pH (<2 or >10) and pressure (2-5 atm) and may additionally involve a rapid pressure release that facilitates chemical infiltration and fiber explosion. Ide­ally, pretreatment produces a disrupted, hydrated substrate that is accessible to enzymatic attack, but avoids both the production of sugar degradation products and fermentation inhibitors. As discussed below, some pretreat­ments solubilize hemicellulose to oligomeric and/or monomeric sugars com­prised largely of pentoses that can be fermented independently or together with the glucose released from the cellulose fraction. In the next step, the pH is adjusted and enzymes are added to initiate cellulose hydrolysis to fer­mentable sugars. With pretreatments that do not solubilize the hemicellulose fraction, additional enzymes may be required to hydrolyze the hemicellulose

polymer. Hydrolysis typically is performed at pH 5 and 50 °C for 24-120 h, followed by addition of a fermentation organism to begin production of ethanol. In many cases (as described below) fermentation is initiated long be­fore hydrolysis has completed, since both the extent and speed of ethanol pro­duction can often be increased by combining the hydrolysis and fermentation steps. In the final step, ethanol is recovered via distillation, and remaining organic waste is burned for production of heat and/or power.

2.1

The regular and cross-linked cellulose chains form a very efficient barrier against penetration of the enzymes into the fibres. Swelling of the pores can be achieved by alkaline pretreatment through soaking of the material in an alkaline solution, such as NaOH, and then heating it for a certain time. The swelling causes an increase in the internal surface area, and a decrease in the degree of polymerization and crystallinity. Usually a major fraction of the lignin is solubilized together with some of the hemicellulose. A rather large fraction of the hemicellulose sugars are usually recovered as oligomers. Al­kaline pretreatment breaks the bonds between lignin and carbohydrates and disrupts the lignin structure, which makes the carbohydrates more accessible to enzymatic attack. As it acts mainly by delignification, it is more effective on agricultural residues and herbaceous crops than on wood materials, as these materials in general contain less lignin. For softwood species, which contain a large amount of lignin, a small or no effect has been observed. Pretreatment using lime instead of sodium hydroxide is another alkaline method, especially suited for agricultural residues, e. g. corn stover, or hardwood materials, such as poplar [24,25].

Dilute acid pretreatment is performed by soaking the material in dilute acid solution and then heating to temperatures between 140 and 200 °C for a certain time (from several minutes up to an hour). Sulphuric acid, at con­centrations usually below 4 wt %, has been of most interest in such studies as it is inexpensive and effective. The hemicellulose is hydrolysed and the main part is usually obtained as monomer sugars. It has been shown that materials that have been subjected to acid hydrolysis may be harder to ferment because of the presence of toxic substances [26-28].

Another approach is to use an organic or aqueous-organic solvent mixture with addition of an inorganic acid catalyst (H2SO4 or HCl), which is used to break the internal lignin and hemicellulose bonds. These methods are usu­ally referred to as organosolv processes [29]. In these cases the hydrolysed lignin is dissolved and recovered in the organophilic phase. It is important to thoroughly wash the material prior to enzymatic hydrolysis and fermenta­tion, as the solvents may act as inhibitors. Solvents that are used are typically methanol, ethanol, acetone, ethylene glycol, triethylene glycol and phenol. Some of these substances are explosive and highly inflammable and thus dif­ficult to handle.

3.3

Development of improved enzymes for the hydrolysis of the other major carbohydrate polymer present in lignocellulosic biomass is also of commer­cial interest, particularly to those utilizing neutral or alkaline pretreatments that leave much of the hemicellulose intact. To develop these enzymes, an industrial residue of the wheat starch industry was used as a model sub­strate. In Europe, wheat is one of the major substrates for production of fuel ethanol. Processing of wheat starch for glucose results in a by-product stream (vinasse) consisting mainly of the wheat endosperm cell wall ma­terial and leftover yeast cells following the fermentation of the starch. The hemicellulose by-product is approximately 33 wt % carbohydrates of which approximately 66 wt % is arabinoxylan. Arabinoxylans consist of a linear backbone of P-1,4-linked D-xylopyranosyl units that are partially substituted with arabinofuranosyls. The major portion of the arabinoxylan in indus­trial wheat fermentation residues is water-soluble [39], the water-insoluble arabinoxylan is quantitatively more abundant in cell walls isolated directly from unprocessed wheat endosperm [40]. Arabinoxylans are hydrolyzed to monosaccharides by acid treatment or by enzymatic hydrolysis. The enzy­matic hydrolysis is usually preferred because it allows for a more specific and controlled modification and fewer undesirable by-products, making it more suitable for microbial fermentation using organisms that can metabolize both xylose and arabinose [41].

The enzymatic degradation of arabinoxylans requires both side-group cleaving and depolymerizing enzymes. Cleavage of the side chains re­quires the action of several accessory enzyme activities, including a-L — arabinofuranosidases, a-glucuronidases, ferulic acid esterase, and acetyl — esterases [41,42]. Depolymerization requires endo-1,4-|3-xylanases that result in unbranched xylooligosaccharides, including xylotriose and xylobiose, and в-xylosidases that cleave xylobiose and attack the non-reducing ends of short chain xylooligosaccharides to liberate xylose [41].

The hydrolysis of arabinoxylan is critical for improved utilization of wheat hemicellulose in the ethanol industry. Three Novozymes cellulolytic and hemicellulolytic enzyme preparations, Celluclast 1.5 L, Ultraflo L, and Vis — cozyme L were tested in various combinations for their ability to liberate arabinose and xylose from water-soluble wheat arabinoxylan. The substrate was medium viscosity water-soluble wheat arabinoxylan from Megazyme (Bray). The three different enzymes were evaluated individually and also in 50 : 50 combinations to look for possible synergistic effects. Reactions were carried out at pH 5 and 50 °C followed by analysis of arabinose, galactose, glu­cose, xylose, xylobiose, and xylotriose by high-performance anion exchange chromatography (HPAEC) [43]. The molecular weight and distribution of water-soluble wheat arabinoxylan and hydrolyzates were determined by high — performance size exclusion chromatography (HPSEC).

In those reactions containing the individual enzyme preparations, the lev­els of arabinose and xylose increased with increasing enzyme dosage and time. Ultraflo L was superior to Celluclast 1.5 L and Viscozyme L in releasing the arabinose from the water-soluble wheat arabinoxylan, meaning that Ul — traflo L must contain a significant amount of a-L-arabinofuranosidase. Cellu­clast 1.5 L was the best enzyme preparation for liberating xylose, resulting in 26 wt % of the available xylose. Ultraflo L released 16 wt % while Viscozyme L released less than 1.5 wt %. In a mixture of 50 : 50 Celluclast 1.5 L and Ultra — flo L there was no interaction among the arabinose-releasing side activities since the same amount of arabinose was obtained as when the two individual enzyme preparations were used and then the arabinose total was combined. The Viscozyme L preparation exhibited a weak antagonistic effect with Ul — traflo L and Celluclast 1.5 L since the amount of arabinose actually decreased compared to that observed with the individual enzyme preparations. The re­sults indicated that the arabinose-releasing side activities of Viscozyme L had the same activity as those demonstrated by Ultraflo L and Celluclast 1.5 L. Another possibile but less likely explanation is the Viscozyme L contained a — L-arabinofuranosidase inhibitors [43]. The 50 : 50 mixture of Celluclast 1.5 L and Ultraflo L produced an increase in the release of xylose compared with the sum of the individual enzyme preparations (Fig. 8). The mixture released 59 wt % of the available xylose, which was 32 wt % more than the theoret­ical addition of the individual enzyme preparations alone. Combination of Ultraflo L and Viscozyme L showed no such synergism, but incubation of Cel­luclast 1.5 L and Viscozyme L showed a weak synergistic effect in liberating some of the xylose from the wheat arabinoxylan.

To further examine the synergistic affect between Celluclast 1.5 L and Ul — traflo L the amounts of xylobiose and xylotriose released during enzymatic hydrolysis were quantified using HPAEC for both individual and combined enzyme preparations. During the initial stage of incubation, Celluclast 1.5 L

liberated small amounts of both xylobiose and xylotriose, indicating the pres­ence of endo-1,4-^-xylanase activities. As hydrolysis continued, the released xylobiose and xylotriose was hydrolyzed to xylose, indicating the Cellu — clast 1.5 L contained one or more в-xylosidase activities.

Ultraflo L treatment resulted in continual liberation of both xylobiose and xylotriose. Ultraflo L showed a low release of free xylose indicating one or more endo-1,4-^-xylanase activities, but little в-xylosidase activity. The synergistic effect between Celluclast 1.5 L and Ultraflo L in releasing xylose is therefore likely to be a result of the action of a-L-arabinofuranosidase and endo-1,4-^-xylanase activities present in Ultraflo L and the в-xylosidase present in Celluclast 1.5 L [43].

Since a strong synergistic effect was observed with a 50 : 50 combination of Celluclast 1.5 L and Ultraflo L for the breakdown of arabinoxylan, a sec­ond study was conducted to look for similar effects and viscosity reduction in the fermentation residue, vinasse. The effects of enzyme dosage, optimal temperature, and pH were examined in hydrolysis of whole vinasse, vinasse supernatant, and washed vinasse sediment that was provided by Tate & Lyle, Amylum UK (Greenwich, UK). On whole vinasse, the enzyme-catalyzed re­lease of arabinose and xylose by the 50 : 50 combination of Ultraflo L and Cel­luclast 1.5 L decreased as the substrate concentration of the vinasse increased. The monosaccharide release also decreased when the substrate concentration of the vinasse increased. Release of arabinose and xylose from the vinasse sediment was very low. The release of arabinose from the whole vinasse var­ied from 40- 50 g arabinose per kilogram vinasse DM while xylose release was between 75-100 g xylose per kilogram vinasse DM after a 24 h hydrolysis. The

Ultraflo L:Celluclast 1.5 L mixture released 53-75 g arabinose and 75-115 gof xylose per kilogram of vinasse DM after a 24 h hydrolysis [44].

Significant viscosity reduction was obtained by enzyme-catalyzed degra­dation of arabinoxylans present in the fermentation residue stream, vinasse. However, there was limited hydrolysis of the insoluble arabinoxylans in the vinasse sediment. The efficiency of enzymatic degradation of the arabinoxylan in vinasse was dependent on enzyme dosing and substrate dry matter [44].

In an effort to narrow down the specific activities involved in the previous studies, the в-xylosidase from Celluclast 1.5 L was purified and used as a sup­plement to Ultraflo L enzyme preparation. When dosed at 0.25 g в-xylosidase protein per kilogram of arabinoxylan along with Ultraflo L, this enzyme mix released the same or more xylose as the enzyme mix consisting of 50 : 50 Ultraflo L and Celluclast 1.5 L (Fig. 9).

In order to determine the optimal enzyme mix for the hydrolysis of vinasse arabinoxylan, several recombinant enzymes were made and tested in various combinations. Genes were cloned and expressed in the fungal host A. oryzae. Based on our studies the optimal enzyme mix for vinasse hydrolysis consists of a-L-arabinofuranosidase from Meripilus giganteus, a-L — arabinofuranosidase II from Humicola insolens, and T. reesei в-xylosidase. A mixture of 25 : 25 : 50 of a-L-arabinofuranosidase from M. giganteus, a-L- arabinofuranosidase from H. insolens and в-xylosidase from T. reesei was determined to be optimal for maximizing arabinoxylan hydrolysis. The success of this work in identifying and exploiting synergism between hemicellulase component activities is currently being applied to other relevant lignocellulosic substrates that differ significantly in their hemicellulose composition.

5