Petrochemicals and petroleum-based products such as plastics are widely integrated into our lifestyles and make a major, irreplaceable contribution to virtually all product areas. Increasing petroleum costs have provided an opportunity for a number of renewable bio-based chemicals or plastics, in addition to the bio-based fuels, to become economically competitive. How­ever, full commercialization of renewable commodity chemicals to replace the currently exploited petrochemicals is critically tied to production cost. Therefore, development of low cost fermentation routes, increased microbial biocatalyst efficiency and productivity and increased final fermentation titer are desired.

5.1

Flowsheeting programs, e. g. Aspen Plus, HYSYS and ChemCad, may be used to perform rigorous material and energy balance calculations, with the use of detailed equipment models, to determine the flow rates, composition and energy flow for all streams in the process. Because of their flexibility, the programs have many advantages when comparing different process configu­rations or scenarios in terms of overall efficiency, minimum energy demand or lower production cost. Also, they serve as a powerful tool when performing sensitivity analyses, due to the ease of changing a certain parameter. All flow­sheeting programs are based on a modular approach where each module is a mathematical model of a unit operation. The fundamental equations needed to accurately describe standard process equipment, such as columns (distil­lation, absorption, etc.), heat exchangers, pumps, reactors and splitters, are normally available as part of the program. The actual simulation is performed by arranging different unit operation modules into a complete flowsheet that represents the process to be simulated.

Construction of a process model in a flowsheeting program can be sum­marized in the following three steps.

• Flowsheet definition: The flowsheet defines the process configuration. It shows all streams entering the system as well as all unit operations and their interconnecting streams. The flowsheet also indicates all product streams that will be determined by the simulation program.

• Chemical components: The user must specify all the chemical compo­nents to be used in the system. All necessary physical and thermodynamic properties must be defined for each component. Normally, a database con­taining these properties for a large number of chemical compounds is included in the flowsheeting software. In general, the size of this database, which varies greatly between different simulators, determines the cost of the flowsheeting program. If data for some compounds are missing the user has to define them.

• Operating conditions: For every unit operation the user has to specify the operating conditions, such as temperature, pressure, heat duties, etc. In addition, all input streams have to be completely defined. Enough infor­mation has to be provided to result in a single steady-state solution based on material and energy balances coupled with phase equilibrium equa­tions.

2.1

The transport of pentose sugars in S. cerevisiae occurs through hexose trans­porters [80,81], albeit with an affinity one to two orders of magnitude lower than for hexose sugars [47,82]. Therefore, pentose transport was early con­sidered a rate-controlling step for ethanolic pentose fermentation [47]. Nev­ertheless, a metabolic control analysis study demonstrated that transport controlled xylose conversion only in strains with high XR activity, and only at low xylose concentrations [83]. Few reports exist on expression of pen­tose transporters in pentose-utilizing S. cerevisiae strains [80,84]. The most effective approach has been the overexpression of the galactose permease Gal2 in recombinant arabinose-fermenting S. cerevisiae [71]. This is in part due to the difficulty in actively expressing heterologous membrane proteins. In a recent breakthrough, the first active heterologous expression of a glu — cose/xylose facilitated diffusion transporter and a glucose/xylose symporter from Candida intermedia [85] in S. cerevisiae [86] was reported. So far the fermentation performance of these strains has not been reported. Never­theless, the expression of heterologous xylose transporters opens up new metabolic engineering strategies to further increase the rate of xylose utiliza­tion in xylose-fermenting S. cerevisiae strains.

4.2

Despite the status of S. cerevisiae as a proven industrial microorganism, con­ferring the ability to rapidly convert pretreated cellulose to ethanol is a daunt­ing proposition. Apart from essential traits, such as high ethanol yield and productivity, industrial strains need to concurrently ferment both hexoses and pentoses under robust industrial conditions that require minimum nutri­ent requirements and high ethanol and inhibitor tolerance. In addition, these strains also have to hydrolyze cellulosics and thus need to produce and se­crete heterologous hydrolases at high enough levels to sustain hydrolysis and fermentation of cellulosics to ethanol (Table 1). Before contemplating these

challenges, it is worth considering the evolutionary development of S. cere­visiae as microorganism of choice for ethanol production.

Through the serendipitous duplication of its entire genome about 100 mil­lion years ago, followed by the further duplication of the alcohol dehydroge­nase (ADH) genes < 80 million years ago, the S. cerevisiae sensu stricto yeast (comprised of 14 Saccharomyces species) adapted the “make-accumulate — consumption strategy” for ethanol production [27,28]. This ability is largely attributed to its overriding glucose repression circuit that suppresses respi­ration of glucose and other C6 sugars above 20-40 mM threshold concentra­tions in the presence of oxygen, a characteristic feature of Crabtree-positive yeasts [29]. This strategy provided the ancestor of S. cerevisiae with an ad­vantage over its competitors because high ethanol levels (concentrations ex­ceeding 4% v/v) are toxic to most other microbes. Once S. cerevisiae has colonized a niche by producing ethanol levels often exceeding 10% v/v from readily available hexoses, the produced ethanol is reconsumed if oxygen is present. These yeasts therefore developed two distinct alcohol dehydrogenase enzymes through the duplication of the ADH genes for the production and

Table 1 Features required from S. cerevisiae as successful CBP microorganism (modified from [2,26])

Suitability of currently available strains of S. cerevisiae

Essential traits:

Only hexoses by native industrial strains. Partial pentose utilization has been engineered in some laboratory and industrial strains Most industrial strains Most industrial strains Most industrial strains

Primarily multicopy expression in laboratory strains

Laboratory and some industrial strains

Manipulated laboratory and some industrial

strains (maltose and glucose utilization)

Most laboratory and industrial strains

Most industrial strains

Some industrial strains, particularly wine

strains

Laboratory and some industrial strains

subsequent utilization of the ethanol: Adh1 that is constitutively produced and is required for ethanol production, and Adh2 that is only induced in the absence of C6 sugars and is necessary for ethanol consumption.

Regardless of the processes used for biomass hydrolysis, CBP-enabling microorganisms may encounter a variety of toxic compounds derived from biomass pretreatment and hydrolysis that could inhibit microbial growth, particularly in the presence of ethanol [30]. However, industrial strains of S. cerevisiae have been adapted to handle stress conditions, such as high ethanol and sugar concentrations (hence osmotolerance), in fermenting sim­ple hexoses (glucose, fructose, galactose, and mannose) or disaccharide (su­crose and maltose) streams. It also has a natural hardiness against inhibitors and has the ability to grow at low oxygen levels. These features confer to S. cerevisiae a general robustness in industrial process conditions [28]. S. cere­visiae has proven itself as a robust ethanol producer in traditional large — scale processes, and therefore presents itself as platform organism for plant biomass conversion to products such as ethanol [2].

3

Detailed kinetic studies have been reported in the literature for several re­combinant strains of Z. mobilis from NREL capable of utilizing both glucose and xylose. The initial evaluation by Zhang et al. [10] involved the batch culture growth of the strain CP4 (pZB5) on medium containing 25 g L-1 glu­cose and 25 gL-1 xylose. Batch and continuous culture studies on strain 39676 (pZB4L) were reported subsequently by Lawford et al. [31,33,34]. This strain was derived from the host ATCC 39676 transformed with a plasmid derived from pZB4. Final product values for 40 gL-1 glucose/40 gL-1 xylose medium included 4.04 g L-1 xylitol as well as 36.6 g L-1 ethanol [49] although it should be noted that xylitol levels with this particular recombinant strain were unusually high. Further studies reported by Lawford and Rousseau [35] focused on kinetic and energetic evaluations of strain CP4 (pZB5) in batch and fed-batch fermentations. Kinetic characterization of the chromosoma­lly integrated xylose/arabinose strain AX101 (derived from ATCC 39676) was also reported [37,38].

To determine which of the strains was likely to be most suitable for larger scale ethanol production, a comparative evaluation in batch and continuous

culture of strains CP4(pZB5) and ZM4(pZB5) was carried out by Joachimsthal et al. [28]. From the results it was found that ZM4(pZB5) was capable of con­verting a mixture of 65 g L-1 glucose and 65 g L-1 xylose to more than 60 g L-1 ethanol in 48 h in batch culture with an ethanol yield of 0.46 gg-1, with this latter strain demonstrating superior specific sugar uptake and ethanol pro­duction rates. The results for ZM4(pZB5) are shown in Fig. 3 together with the values of comparative kinetic parameters in Table 2. Higher sugar con­centrations (75 gL-1 each sugar) resulted in incomplete xylose utilization (80 h) presumably due to increasing ethanol inhibition of xylose assimila — tion/metabolism at ethanol concentrations of 65-70 gL-1.

The results for continuous culture with ZM4 (pZB5) and medium contain­ing 40gL-1 glucose and 40 gL-1 xylose are shown in Fig. 4 [28]. While the concentration of glucose was close to zero at dilution rates up to D = 0.15 h-1, increasing residual xylose at dilution rates higher than 0.08 h-1 indicated that the maximum volumetric rate of xylose uptake for the culture had been exceeded. The maintenance energy coefficient (m) under these conditions was estimated by extrapolation as 1.6±0.2gg-1 h-1 (within 95% confidence limits) based on linear regression analysis of the data from Fig. 4a for the maximum specific sugar uptake rate (glucose and xylose) vs. dilution rate (D) (Fig. 4b). A “true biomass yield” of 0.044 gg-1 was determined from the inverse of the gradient of this linear plot. For similar experimental con­ditions, closely related values were observed by Lawford and Rousseau for strain CP4 (pZB5) [34]. However, Lawford and Rousseau noted, when ob-

Mm: maximum specific growth rate (h-1)

(qs)m: maximum specific sugar uptake rate (gg-1 h-1)

(qp)m: maximum specific ethanol production rate (gg-1 h-1) (Yx/s): overall cell yield (based on total sugar utilized) (gg-1) (Yp/s): overall ethanol yield (based on total sugar utilized) (gg-1)

served over the lower dilution rate range of D = 0.04-0.08 h-1, that both strain CP4 (pZB5) and a biomass hydrolysate adapted variant of 39676(pZB4L) exhibited values of m and “true biomass yield” that were significantly lower [35].

Results with a potentially high productivity cell recycle system using a membrane bioreactor are shown in Fig. 5 [29]. From Fig. 5(a), at sugar con­centrations of 50 gL-1 glucose and 50 gL-1 xylose and D = 0.1 h-1, an ethanol productivity of 5 g L-1 h-1 was achieved with an ethanol yield based on total sugars utilized (Yp/s) = 0.50 gg-1. No decline in specific ethanol productivity was evident up to 70 h, however as shown in Fig. 5(b), a decrease in total vi­able cells was observed after an initial steady state (40-50 h). This indicates that for effective longer term operation, high cell concentrations should be

— 2.5

— 2 ^ Li 3

CO со ш

Є о

-1 s

— 0.5

0. 02 0.04 0.06 0.08 0.1

Dilution rate (h’1)

Fig.4 (a) Kinetics of ethanol production by Z. mobilis ZM4 (pZB5) in continuous culture on medium containing 40 gL-1 glucose and 40 gL-1 xylose (T = 30 °C, pH = 5.0). Sym­bols: biomass •; glucose ♦; xylose □; ethanol ▲ (b) Effect of dilution rate on specific rates of total sugar uptake (qs) and ethanol production (qp). Estimation of maintenance energy (m) value at D = 0 by extrapolation. Symbols: qs °; qp A achieved by less stressful methods than membrane-based cell recycling (e. g., by use of flocculent cells and cell settling).

2.5

As seen in the review of major biofuel producers, a common policy instru­ment used to support the industry is direct government program funding, in the form of contracts, loans, grants, or fiscal guarantees. It is difficult to evaluate the effectiveness of direct funding by comparing different countries, where synergistic policies (such as renewable fuel mandates, excise tax ex­emptions, etc.) or simply more favorable market conditions may play a role in determining capacity. However, within a single country it may be easier to see the impact of direct funding on the establishment of biofuel cap­acity. The bioethanol industry in the USA has been chosen for an analysis of the effectiveness of direct funding towards establishing biofuel produc­tion capacity. For the purpose of this study, direct funds are considered to be funds earmarked for all aspects of research, development and demon­stration, including all biofuel production as well as biomass production for general energy purposes. When different funding sources were considered, the only real criteria applied to warrant their inclusion in this study were (1) that the funds be applicable to research, development, and demonstra­tion (RD&D) projects for bioethanol, including construction or modification of production facilities, and (2) that bioethanol is accounted as an eligible product. Funding sources that recognized bioethanol as a co-product of mate­rial or bioenergy generation were also included. Estimates of the cumulative, total funding available to support the bioethanol industry are shown in Fig. 2. Canada is included in this graphic for comparison’s sake.

In Fig. 2, direct funds available in each state are indicated by the shading on the map, from blue (base levels of cumulative funding provided by the federal government as of 2005) to light or dark red (additional state fund­ing, depending upon the cumulative amount of funds available as of 2005). Existing bioethanol production capacity for 2005 is indicated by the yel­low circles, logarithmically sized according to the scale indicated. Additional bioethanol production capacity expected to be online as of 2007 is indicated by the dark orange circles, again plotted logarithmically. The graph indicates that bioethanol production is likely to be found where funding is available for infrastructure development, biomass procurement, and plant operation. Each of the major bioethanol-producing states has followed a different ap­proach in creating these incentives. Each approach represents a successful strategy for attracting the industry and expanding bioethanol production capacity.

In Illinois, the primary incentive offered to bioethanol producers is the Illinois Renewable Fuels Development Program, which offers up to US $ 5.5 million per facility in grants for the construction or retrofitting of renewable fuels plants, provided that they are a minimum of 114 million L in capacity and that the total grant award does not exceed 10% of total construc­tion costs, or US $ 0.026 L-1 of additional biofuels capacity created [73]. Both bioethanol and biodiesel production facilities are currently the primary re­cipient of these funds. In addition, the Renewable Energy Resources Program offers funding at various levels to promote the development and adoption of renewable energy within the state. With two new plants under construction in 2006, the total funding available to the bioethanol industry is estimated at approximately US$ 30.15 million [23]. Currently, Illinois has five operat­ing facilities with a capacity of 5.1 billion Lyear-1, while two new facilities are under construction [18].

In Iowa, a number of innovative programs are in place. The Iowa Re­newable Fuel Fund’s Financial Assistance Program offers a combination of forgivable and traditional low-interest loans for projects involving biomass and alternative fuel technologies, while the Alternative Fuel Loan Program of­fers zero-percent interest loans for up to half the cost of biomass or alternative fuels related fuel production projects, up to a maximum of US $ 250 000 per facility [74]. Approximately 20% of the money awarded under this program is in the form of forgivable loans, while the remaining 80% are low-interest loans. A number of other incentives, including the Ethanol Infrastructure Cost-Share Program, provide incentives for installation or conversion of E85 refueling stations [75].

In Minnesota, the chief incentive is the Ethanol Production Incentive. Originally, this incentive provided direct payments to producers at a rate of approximately US $ 0.052 US L-1 bioethanol, although the passage of bill SF 905 (2003) has reduced this amount to US $ 0.034 L-1 from 2004-2007. In 2007, the original incentive will be restored and producers may be reim­bursed for lost incentive if funds are available. The total fund available is US $ 37 million, although there is a cap of US $ 3 million per producer, which essentially means that producers of more than 15 million L year-1 are ineli­gible for extra incentive [76]. Perhaps due to this restriction in funding, the program has resulted in the establishment of 15 individual facilities by 2006 with a total production capacity of 1.9 billion Lyear-1 [18]. The Ethanol Pro­duction Incentive expires June 30, 2010 [77]. Ethanol infrastructure grants are also available to help upgrade service stations for dispensation of E85 fuels [23]. Minnesota has also enacted legislation for a bioethanol blend man­date, currently enforcing a 10% bioethanol blend for consumers (to increase to 20% bioethanol in 2013) [77].

In South Dakota, the Ethanol Production Incentive is designed as a dir­ect payment of US $ 0.052 L-1, with a maximum of US $ 1 million annually or US $ 10 million in total to any single facility. Unlike the incentives described for Minnesota, Illinois, or Nebraska, this particular program is targeted spe­cifically at bioethanol from cereal grains and expires this year [22]. While this level of support is lower than in many other states, South Dakota also has an excise tax exemption on bioethanol which provides additional financial in­centive for production. Currently, South Dakota has 11 operating facilities, with four additional plants under construction and a total production cap — acityof 2.2 billion L year-1 [18].

In Nebraska, the main program is the Ethanol Production Incentive, which offers a tax credit of US $ 0.048 L-1 bioethanol for up to 60 million L of annual production per facility, or 473 million L in total production over the course of a 96-month consecutive period [78]. This credit, which will expire in 2012, is limited to a total of US $ 22.5 million. As a tax credit, these funds can be con­sidered to be defrayed costs in direct support of the industry [23]. Nebraska currently has a production capacity of 1.8 billion L annually in ten facilities, with three new installations currently under construction [18].

As these examples demonstrate, a range of policy tools have been deployed in areas with significant bioethanol production capacity. The tools of pro­duction incentives, tax exemptions, direct loans, and cost-share schemes are shown to be effective in attracting capacity to individual jurisdictions, and the tools are shown to be flexible in achieving different results. The Min­nesota example, in particular, shows the potential impacts of small changes to policy. By limiting the capacity to which the incentive applied, the state gov­ernment was able to spur the creation of many individual facilities, which will in turn have a direct impact on jobs and the local economy. It is important to remember, however, that each of these strategies build upon the US fed­eral government’s strong commitment to research and development. Without that commitment, the rapidly improving technology that makes these facili­ties possible would not exist. However, it is interesting to note the differences that small amounts of local funding might have on productivity.

In Fig. 3, the relation between state funding for biofuels is compared to ac­tual bioethanol production capacity, using the funding data and bioethanol production capacities for 2003 and 2005. The two years of data are dif­ferentiated by the shaded and white circles. In 2003, a strong correlation was found between state-level funding and bioethanol production capacity (r2 = 0.85). This indicates that direct funding likely played a role in attract­ing new bioethanol capacity, and thus it could be concluded that this is an effective policy tool. By 2005, the changes in production capacity and dir­ect funding levels in many states has reduced this correlation significantly (r2 = 0.64). It may be postulated that a shift is taking place, in which the amount of funding available to capital projects has become less import­ant in relation to some other factor, such as feedstock availability or mar-

ket influences. Indeed, follow-up analyses using corn production data [79, 80] indicate that in the same period, the relation between bioethanol pro­duction capacity and corn harvest figures on a state level show the op­posite trend. In 2003, the correlation between the two was fairly weak (r2 = 0.58), while in 2005, this correlation had grown stronger (r2 = 0.83). In 2003, availability of corn seemed to be less important than direct fund­ing for bioethanol facilities. It may be postulated that the rapid growth in bioethanol capacity seen to 2005, coupled with strong prices for bioethanol, has made feedstock availability more important than funding for construc­tion purposes.

4

Metabolic engineering strategies for pentose fermentation are developed to fi­nally generate strains that ferment pentose sugars to ethanol under industrial conditions, which may include suboptimal pH and an array of compounds which inhibit cellular metabolism. Industrial strains of S. cerevisiae, including baker’s yeast, generally out-compete most other microorganisms with regard to the properties required in industrial ethanol production [9,10,133-135], including ethanol productivity, ethanol tolerance, lignocellulose hydrolysate tolerance, and tolerance to low pH [136].

5.1

To eliminate the dependence of KO11 and LY01 on costly nutritional supple­mentation, a new ethanologenic E. coli strain was constructed. The starting strain SZ110, a derivative of KO11 modified for production of lactic acid in mineral salts medium (see Sect. 5.1), was re-engineered for ethanol produc­tion.

2.2.1

Conversion of SZ110 to LY168

Strain SZ110, a derivative of KO11, was engineered and metabolically evolved to produce lactic acid, as described in detail below [39]. Evolved derivatives of SZ110 produced D-lactate at 92% yield from 100 gL-1 glucose in inexpen­sive mineral salts media. Since this cheap and efficient utilization of large amounts of sugar is the desired biocatalyst behavior, strain SZ110 was chosen as the starting point for re-engineering of ethanologenic E. coli(Yomano et al., submitted). Conversion of this strain from lactic acid production to ethanol production involved several steps, beginning with deletion of the lactic acid production gene IdhA. The Z. mobilis PET operon, inserted at the pfl locus in KO11, was removed during engineering of SZ110 for lactic acid production by deletion of the entire focA-pflB region [39]. Since elimination of ackA and adhE prevents undesirable carbon loss, deletion of pflB is unnecessary and possibly limits acetyl-CoA levels. Therefore, the native pfl gene was restored in the re-engineered ethanologenic E. coli. To select for optimal integration of the Z. mobilis homoethanol pathway, a promoterless operon containing pdc, adhA, and adhB was randomly inserted by transposon.

Specific growth requirements of both the donor and recipient strains en­abled direct functional selection in minimal medium without antibiotics. Candidate ethanologenic strains were enriched by serial transfers in mineral salts medium. One clone was selected and designated LY160. Further evolu­tion of strain LY160 by serially subculturing into fresh mineral salts medium every 24 h for 32 days led to strain LY160im, an intermediate strain with con­tinued improvement in performance. It was determined that the Z. mobilis ethanol pathway in LY160im was integrated within rrlE, a 23S ribosomal RNA subunit, concurrent with the direction of transcription. The complex regu­lation of ribosomal RNA transcription is reviewed in [40,41]; the presence of two promoters results in high expression at high growth rates and basal expression at low growth rates and during stationary phase, making rrlE an excellent site for PET integration. The Pseudomonas putida short chain es­terase estZ gene was also integrated into the microbial biocatalyst to lower ethyl acetate levels in the broth. The final strain was designated LY168.

2.2.2

Biological conversion of biomass to hydrogen either proceeds through photo­fermentation or dark fermentation. In dark fermentation the yield is only 10-20% of the potential hydrogen amount that theoretically can be de­rived from organic matter ([7] and Westermann P, J0rgensen B, Lange L, Ahring BK, Christensen CH (2007) Int J Hydrogen Energy (accepted for pub­lication)). Typical hydrogen yields are from 0.52 mol H2/mol hexose, when molasses was the substrate in a batch culture of Enterobacter aerogenes [8], to 2.3 mol when glucose was the substrate in continuous culture of Clostrid­ium butyricum [9]. Besides the low hydrogen yield, a major problem of fermentative hydrogen production is hydrogen-consuming microorganisms such as methanogens and acetogenic bacteria. In these processes, hydrogen is inevitably converted into methane or acetate, respectively, unless the re­sponsible microorganisms are excluded by sterilization of the biomass before fermentation and inoculation with specific hydrogen-producing microbes, or the process is carried out under conditions adverse to the hydrogen utilizers. A combination of biohydrogen production with fuel cell technology is, how­ever, rather straightforward since the fuel cell technology is available [10]. An upgrading of produced gases might be necessary before they are introduced into the fuel cells [11].

As a stand-alone process, fermentative hydrogen production from biomass is currently not feasible due to the low yield attained.

3

The performance of the thermostable enzymes at a lower temperature, the 35 °C commonly used in SSF, was compared. The T. reesei deletion strains produced only low amounts of background cellulase activities, mainly due to the presence of native EGIII (Cel12A) and EGV (Cel45A). However, the deletion strains used for the production of thermoenzymes produced some hemicellulases. For practical use, any mesophilic background activity en­hancing the hydrolysis can be considered useful, but in order to evaluate the performance of the thermophilic enzymes, the level of remaining background activities was evaluated. The FPU activity in the background was negligible and the endoglucanase activity was very low as compared to the commercial preparations. Most of the endoglucanase activity, 85-90%, was inactivated during the thermal treatment at 60 °C, pH 6.5 for 2 h. Obviously, the EGV ac­tivity was the most stable remaining activity. Thus, the background activities originating from the T. reesei deletion strains had only a minor contribution to the total hydrolysis above 65 °C.

The actual hydrolysis performance of the new thermostable enzyme mix­tures on various pretreated lignocellulose substrates (spruce and corn stover) at 35 °C showed some variations as compared with the T. reesei enzymes: on

A

100 і

Fig. 7 Hydrolysis of steam pretreated washed spruce (a) and unwashed corn stover (b) by Celluclast (■) and (□) the thermostable enzymes (TM 1 for spruce and TM 2 for corn stover) at 35 °C. Enzyme dosages: Celluclast 5 FPU g-1 substrate, supplemented with 100 nkatNovozym 188 g-1 substrate; thermostable enzymes 5 FPU g-1 substrate, substrate concentration 10 gL-1, hydrolysis time 72 h at pH 5, triplicates with mixing
spruce, the sugar yield obtained by the thermophilic enzymes was generally lower and on corn stover higher than with the commercial T. reesei enzymes (Fig. 7). The result was the same, irrespective of the presence of the thermoxy — lanase in the preparation (TM 1 in Fig. 7 or TM 3 in Fig. 5). Thus, with this substrate the relatively lower cellulase activity at 35 °C is obviously the rea­son for the poorer hydrolysis at the lower temperature. In contrast, on the xylan-containing substrate, corn stover, the additional xylanase activity in the thermostable enzyme mixture had a more profound effect. The total xylanase activity was somewhat higher in the thermostable preparation, emphasising the importance of hemicellulases in the hydrolysis of substrates containing residual xylans. Further research would be needed to study in detail the struc­tural differences of both cellulose and hemicellulose in the two substrates and their impact on the performance of the enzyme patterns used.

10