5.4.1 Upstream Factors: Biomass Collection and Delivery

The “billion ton vision” is a program to access a billion tons of dry biomass per year to produce bioethanol (and other biofuels) to replace 30% of U. S. gasoline consump­tion by 2030.69 One of the key parameters in a large-scale restructuring of the U. S. national fuel industry is that of supplying biomass raw material at such a high rate and at an economically acceptable cost.

The logistics and transportation costs of such large amounts of low-value, high- volume raw materials have only recently attracted serious consideration. Canadian studies comprise the most detailed considerations of these highly practical ques­tions now that cellulosic ethanol facilities are nearing industrial reality. For an agri­cultural economy (and climate) like Canada’s, wood is highly likely to be a large fraction of the biomass supply, initially from forest harvest residues and “energy plantations” on marginal farmland.70 For wood chips, larger production plants (up to 38 million dry tons of biomass/year) are more economic than smaller units (2 million tons), and truck delivery is limited to such small units by issues of traf­fic congestion and community acceptance.71 Combined road and rail shipping, that is, initial collection by truck followed by trans-shipping to rail, is only economic when the cost per unit distance of the rail sector is less than the trucking-only mode because of the incremental fixed costs: for woody material, the minimum economic rail shipping distance is 125 km (78 miles), whereas for cereal straw, the minimum distance extends to 175 km (109 miles).72 Existing rail networks impose, however, a serious restriction (that of their location), supplying only sites close to already positioned track; road transport is more versatile. Factoring in additional consid­erations, including air emissions during transport, definitely favors rail transport.73 Policy changes and new infrastructure investment appear therefore to be inevitable if the development of bioethanol production is not to be inhibited by objections of cost and pollution.

Focusing on corn stover and wheat straw as raw material inputs, a study of North Carolina concluded that more than 80% of these resources were located in the coastal area; four ethanol plants with feedstocks demands of between 146,000 and 234,000 dry tons/year required collection radiuses between 42 and 67 km (26-42 miles).74

The siting of production facilities to minimize transportation costs implies a contradiction with the economies of scale possible with larger production units. This is made more likely if relatively marginal biomass inputs such as municipal solid waste are to be considered.75 Urban fringes might be close to existing landfill sites and also within short distances of field crop residues, wastes from horticultural industries, and seasonal supplies of tree and plant residues from urban parkland. As discussed previously (section 5.3.2), the ability to design and build smaller-scale bioethanol production units, especially if they can utilize a variable and adventitious supply of feedstocks, would be highly beneficial to match the fragmented nature of the cheapest likely raw materials.

This C4 dicarboxylic acid is one of the key intermediates of glucose catabolism in aerobic organisms (including Homo sapiens) but can also be formed anaerobically in fermentative microbes (figure 8.4). In either case, CO2 is required to be “fixed” into organic chemicals; in classical microbial texts, this is described as “anaplerosis,” acting to replenish the pool of dicarboxylic and tricarboxylic acids when individual com­pounds (including the major industrial products citric, itaconic, and glutamic acids) are abstracted from the intracellular cycle of reactions and accumulated in the extracellular medium. Under anaerobic conditions, and given the correct balance of fermentation

GLUCOSE

products, a net “dark” fixation of CO2 can occur, and it is this biological option that has been most exploited in the development of modern biosynthetic routes:

• The rumen bacterium Actinobacillus succinogenes was discovered at Michigan State University and commercialized by MBI International, Lansing, Michi — gan.1617 Succinate yields as high as 110 g/l have been achieved from glucose.

• At the Argonne National Laboratory, Argonne, Illinois, a mutant of E. coli unable to ferment glucose because of inactivation of the genes encod­ing lactate dehydrogenase and pyruvate formate lyase spontaneously gave rise to a chromosomal mutation that reestablished glucose fermentative capacity but with an unusual spectrum of products: 1 mol succinate and 0.5 mol each of acetate and ethanol per mole of glucose consumed.1819 The second mutation was later mapped to a glucose uptake protein that, when inactivated or impaired, led to slow glucose transport into the cells and avoided any repression of genes involved in this novel fermentation.20 The result is a curious fermentation in which redox equivalents are balanced by a partition of carbon between the routes to succinate and that to acetate and ethanol (in equal measures), pyruvate being “oxidatively” decarboxylated rather than being split by pyruvate formate lyase activity or reduced to lac­tic acid, both routes lost from wild-type E. coli biochemistry in the paren­tal strain (figure 8.5). The maximum conversion of glucose to succinate by this route is 1 mol/mol, a carbon conversion of 67%; succinate titers have

with PTS

glucose

uptake

Oxaloacetic acid

Succinic acid oxidation value = 0 0.5 mole

oxidation value = +1
1 mole

FIGURE 8.5 Redox balance in the fermentation of glucose to succinic acid by Escherichia coli.

reached 75 g/l. Because E. coli is only facultatively anaerobic, biomass in the fermentation can be generated rapidly and to a high level under aerobic conditions, O2 entry then being restricted to transform the process to one of anaerobic metabolism.21 The same organism can successfully utilize both glucose and xylose in acid hydrolysates of corn straw and generate succi­nate as a fermentation product.22

• In wild-type E. coli, glucose fermentations produce complex mixtures of acid and nonacidic products, in which succinate may be only a minor component (chapter 2, section 2.2). Nevertheless, the succinate titer can be greatly increased by process optimization, and Indian researchers achieved more than 24 g/l within 30 hours with laboratory media and

17 g/l in 30 hours in a fermentor with an economical medium based on corn steep liquor and cane sugar molasses.23 24 The same group at the University of Delhi have enhanced succinate productivity with Bacteroides fragilis, another inhabitant of the human gut and intestine but an obli­gately anaerobic species.25 26

• In complete contrast, an aerobic system for succinate production was designed with a highly genetically modified E. coli, using the same glu­cose transport inactivation described above but also inactivating possible competing pathways and expressing a heterologous (Sorghum vulgare) gene encoding PEP carboxylase, another route for anaplerosis.2728 A succinate yield of 1 mol/mol glucose consumed was demonstrated, with a high pro­ductivity (58 g/l in 59 hours) under fed-batch aerobic reactor conditions. The biochemistry involved in this production route entails directing carbon flow via anaplerotic reactions to run “backward” through the tricarboxylic acid cycle, that is, in the sequence:

PEP ^ oxaloacetate ^ malate ^ fumarate ^ succinate

• At the same time, succinate is produced in the “forward” direction by block­ing the normal workings of the cyclic pathway (with a necessary loss of carbon as CO2) and the activation of a pathway (the “glyoxylate shunt”) nor­mally only functioning when E. coli grows on acetate as a carbon source:

citrate ^ isocitrate ^ succinate + glyoxylate

• The enzymes catalyzing the final two steps in the pathway from PEP, malate dehydratase and fumarate reductase, can be overexpressed in bacterial spe­cies and are the subjects of two recent patent applications from Japan and

Korea.29,30

• To return to anaerobic rumen bacteria, Anaerobiospirillum succiniciprodu — cens had a short but intense history as a candidate succinate producer.3133 Fermenting glucose in a medium containing corn steep liquor as a cost effective source of nitrogen and inorganic nutrients, succinate titers reached

18 g/l from 20.2 g/l of glucose, equivalent to a conversion efficiency of 1.35 mol/mol.34

• The same research group at the Korean Advanced Institute of Science and Technology, Daejeon, Republic of Korea, then isolated a novel rumen bacterial species, Mannheimia succiniciproducens, and has determined its complete genomic sequence as well as constructing a detailed metabolic network for the organism.35-37 Mutants of this microbe can produce suc­cinate with much reduced amounts of other acids and can anaerobically ferment xylose and wood hydrolysate to succinate.38 39

There are grounds to predict that overexpressing genes for anaplerotic pathway enzymes would enhance succinate production (and, in other genotypes or fermenta­tion conditions, the accumulation of other acids of the tricarboxylic cycle); experi­mental evidence amply confirms this prediction.40-45 With the capabilities to perform metabolic computer-aided pathway analysis with known gene arrays, comparison of succinate producers and nonproducers and between different species would be expected to greatly accelerate progress toward constructing the “ideal” microbial cell factory. Comparison of E. coli and M. succiniciproducens suggested five target genes for inactivation but combinatorial inactivation did not result in succinate over­production in E. coli; two of the identified genes — ptsG (the glucose transport system) and pykF (encoding pyruvate kinase, the enzyme interconverting PEP and pyruvic acid), together with the second pyruvate kinase gene (pykA) — increased succinate accumulation by more than sevenfold, although succinate was still greatly outweighed by the other fermentation products (formate, acetate, etc.).46 Eliminating the glycolytic pathway below PEP will clearly aid succinate production, but the pro­ducing cells will by then be highly dependent on organic nutrients (including many amino acids) for growth and maintenance.

The combination of several related technologies is particularly appealing for commercial production of succinic acid:

• Optimization of the Argonne National Laboratory strains for succinate pro­duction by the Oak Ridge National Laboratory, Oak Ridge, Tennessee

• Innovations for succinate recovery from the fermentation broth at Argonne National Laboratory

• An improved succinic acid purification process47

• The development of catalytic methods for converting succinic acid to 1,4- butanediol and other key derivatives at Pacific Northwest National Labora­tory, Richmond, Washington48

These advances have moved biologically derived succinic acid close to commercial­ization as a component of the first genuine biorefinery.

The solids remaining at the end of the fermentation (distillers dried grain with solids, or DDGS: see figure 1.21) are a high-protein animal feed — a saleable by-product that has been suggested to be so commercially desirable that reduced ethanol yields could be tolerated to support its increased production, although, in practice, high — sugar residues pose severe practical difficulties to DDGS drying and processing.122 The rapid rise of ethanol production from cornstarch has, however, demanded some remarketing of this coproduct:269

• The product is less dark because sugars are more efficiently fermented and less available to react chemically and caramelize in the dried product.

• The essential amino acid contents are higher (figure 4.13).

• Although ruminant animals can certainly benefit from feeding with DDGS, pigs are geographically much closer to ethanol plants in the midwestern United States.

Reducing phosphate content would widen the use of DDGS by addressing animal waste disposal issues, and the development of more efficient methods for removing water in the preparation of the DDGS could greatly reduce processing costs.270 Adding on a second fermentation (or enzymic biotransformation), a dry-grind processing to generate plant oils and a higher-value animal feed from the DDGS, and separating more useful and saleable fine chemicals from the primary fermentation would increase the total mass of recovered bioproducts to the maximum achievable (figure 4.14).271 Pric­ing is crucial because the increased supply of DDGS is likely to significantly reduce its market price, and its alternate use as the feedstock for further ethanol production itself has been worthy of investigation: steam and acid pretreatments can convert the residual starch and fiber into a substrate for yeast-based ethanol production with a yield 73% of the theoretical maximum from the glucans in the initial solids.272

A much simpler option is to realize the potential in the fermented solids to provide nutrients and substrates for a new round of yeast (or other ethanologen) growth and eth­anol production: such spent media (“spent wash,” stillage, or vinasse) can be recycled in the process known as “backsetting,” found to be beneficial for yeast growth and a practical means of reducing water usage in a fuel alcohol facility.273 Backsetting is not without its accepted potential drawbacks, including the accumulation of toxic nonvola­tiles in the fermentor, increased mash viscosity, and dead cells causing problems with

High Value Animal
Feed

Carbon Dioxide Plant Oils Inositol Succinic Acid Glycerol Ethanol

FIGURE 4.14 Projected recovery of product and coproducts from the ethanol fermentation of corn starch. (Data from Dawson.271)

viability measurements, but as a crude means of adapting the fermentation to a semi­continuous basis, it has its advocates on both environmental and economic grounds. Furthermore, a study by Novozymes demonstrated that mixtures of fungal enzymes could decrease vinasse viscosity and liberate pentose sugars from soluble and insoluble arabinoxylans that would be suitable for fermentation by a suitable pentose-utilizing ethanologen.274-277 Portuguese work has also shown that such a pentose-rich product stream can be the starting point for the fermentative production of xylitol (a widely used noncalorific sweetener) and arabinitol by the yeast Debaryomyces hansenii.278,279 As a support for the immobilization of yeast cells, brewer’s spent grains were a very effective means of supplying “solid-phase” biocatalysts for ethanol production from molasses, and the solids from bioethanol plants could serve a parallel function.280

Brazil has, by far, the longest continuous history of devising methods for eco­nomically viable disposal for vinasse and the solid waste product (bagasse), espe­cially because neither had at times been considered to be saleable and both could even represent negative value as incurred disposal costs:281 [43]

• Important for minimizing fertilizer use, material dissolved in the digestion wastewater represents 70% of the nutrient demand of sugarcane fields.

6.2.1 The Renascence of an Old Chemistry for Biomass-Based Fuels?

The generation of a combustible gas, synthesis gas (“syngas”), from biomass was discussed briefly in chapter 2 (section 2.1). Technologies for the conversion of coal and natural gas to liquid fuels were also included in chapter 5 (section 5.6) as part of a survey of different strategies for adapting to potentially dwindling crude oil reserves. The chemistry of gas-to-liquid fuel transformations was developed in the first quarter of the twentieth century and utilized extensively in Germany during World War II; further evolution led to commercial production processes being initi­ated for peacetime purposes in the 1990s.84,85

The essential step, known as the Fischer-Tropsch (FT) reaction, can be written as

nCO + 2nH2 ^ [CH2]n + nH2O,

where [CH2]n represents a range of hydrocarbons, ranging from low-molecular-weight gases (n = 1, methane), by way of gasoline (n = 5-12), diesel fuel (n = 13-17), and as far as solid waxes (n > 17). The reaction requires catalysts for realistic rates to be achieved, usually iron or cobalt (although transition metals will function effectively) at high temperatures (180-350°C) and high pressures; the higher the temperature, the higher the proportion of gas and liquid hydrocarbon products.

To date, no process has been commercialized from plant biomass feedstocks, and the FT technology could be described as “radical” or “nth” generation for biofuels were it not that the key elements of the chemistry and production options are reason­ably well established in industrial processes with fossil inputs; in a climate of high crude oil prices, the environmental desirability of low-sulfur diesel, and the drive to commercialize otherwise unmarketable natural gas in remote locations are impor­tant synergies (table 6.6).86 FT biomass-to-liquid fuel (FT-BtL) from lignocellulosic sources is particularly attractive because of the high CO2 emission reduction poten­tial (up to 90% when substituting conventional gasoline and diesel) and the ability to use woody materials from low-grade land, thus avoiding the pressures on land use in OECD countries contemplating agriculture-based bioethanol or biodiesel production on a large scale.79 The principal barrier to large-scale biomass FT-BtL appears to be the suboptimal mixture of gases in syngas as prepared from plant materials: the lower the molar ratio of H2:CO, the more the proportion of high-molecular-weight products formed in the FT reaction, but biomass gasification results in a wide range of H2:CO ratios, often with an excess of CO, together with appreciable amounts of CO2, methane, and higher hydrocarbons as well as smaller amounts of condensable tars and ammonia.87

The methane can be transformed to CO and H2 by a number of different reac­tions, including the uncatalyzed (but again high-temperature and high-pressure)

processes:88

CH4 + O2 ^ CO2 + 2H2O and CH4 + H2O ^ CO + 3H2

Partial removal of CO (and formation of additional H^ is possible by the water — shift reaction:

CO + H2O ^ CO2 + H2

Finally, the physical removal (adsorption) of CO2 (an inert gas for FT reactions) is relatively straightforward, but a higher-yielding process can be devised (at least, in principle) by including a catalytic reduction of the CO2 to using multiple FT reactors in series with an intermediate water removal step:89

CO2 + 3H2 ^ [-CH2-] + 2H2O

Complete wood-based FT-BtL production involves, therefore, a multistage pro­cess, incorporating biomass pretreatment, syngas purification, and optional syngas recycling, plus gas turbine power generation for unused syngas and, for FT die­sel, a hydrocracking step to generate a mixture of diesel, naphtha, and kerosene (figure 6.8).87,88

The assessment of corn-derived ethanol was the most extensive of the reports (60% of the total printed pages in the final collection of papers) and formed a notional blueprint for a facility sited in Illinois with a projected working life of 20 years and operating costs of approximately 950/gallon of hydrous ethanol (table 5.1). The final factory gate selling price was computed to be $1.05/gallon (1978 prices) in the base case of the 50-million-gallon/year capacity including the results of a 15% dis­counted cash flow/interest rate of return analysis; the selling price was a little lower (980/gallon) with twice the annual capacity but considerably higher ($1.55/gallon) at only 10 million gallons/year. The quoted comparative price for refinery gasoline was 400/gallon; after allowing for the lower energy content of ethanol (70% of that of gasoline — chapter 1, section 1.3), the “real” cost of corn-derived ethanol would have been $1.50/gallon for the 50-million-gallon facility, that is, 3.75-fold higher than gasoline at that time.

Various options were explored in the study to define the sensitivity of the required selling price for ethanol:

1. The DOE required the analyses to define a selling price that would cover not only the annual operating expenses but also yield a return on equity, the base case being a 15% discounted cash flow/interest rate of return; increas­ing this factor to 20% resulted in a higher selling price ($1.16/gallon for the base-case scenario).

2. Lengthening the depreciation schedule from 10 to 20 years increased the selling price by 20/gallon.

3. Increasing the working capital to 20% of the total production cost increased the selling price by 30/gallon.

4. A higher investment tax credit (50%) would reduce the selling price by 20/gallon.

5. Financing only 80% of the plant investment could reduce the selling price by 100/gallon.

6. For every 10% rise in the price of corn, the selling price would increase by 80/gallon (after allowing for a triggered rise in the selling price of the solid animal feed coproduct).

7. For every 10% rise in the price of the animal feed coproduct, the selling price would decrease by 40/gallon.

8. Replacing local coal by corn stover as the fuel for steam generation would increase the selling price by 40/gallon — although a lower total investment (by approximately $1 million) would have been an advantage resulting from the removal of the need for flue gas desulfurization.

All of these changes are comparatively minor, and other quantified changes to the over­all process were likely to have been equally small: ammonium sulfate (a coproduct aris­ing from flue gas desulfurization) was only generated in small amounts, approximately 3 tons/day, and no allowance was made for capturing and selling the CO2 generated in the fermentation step. No denaturant was included in the final cost breakdown.

Alternative feedstocks were also explored. Milo (grain sorghum) offered a slight reduction in the selling price of ethanol (down to $1.02/gallon) but was considered a

small-acreage crop at that time. Both wheat and sweet sorghum were likely to increase the final factory gate selling price to $1.31/gallon and $1.40/gallon, respectively. Although wheat and milo grain could be processed in essentially the same equipment used for corn, sweet sorghum required a higher investment in plant facilities.

The best taxonomically and physiologically characterized examples of H2 pro­ducers are clostridia, but other genera (including bacilli), as well as a microbial flora from anoxic marsh sediments and other environments are known that are capable of the H2 production and either the ABE “solvent” fermentation (chap­ter 6, section 6.3.3), the accumulation of one or more of the ABE trio, carbox­ylic acids (acetic, butyric, etc.) and/or other products (acetoin, 2,3-butanediol, etc.).26

A wide spectrum of carbon sources supports H2 production at rates up to 1,000 ml/hr/g cells at a maximum yield of 4 mol of H2/mol glucose with the stoichiometry:26

C6H12O6 + 2H2O ^ 2CH3COOH + 2CO2 + 4H2

This reaction is sufficiently exothermic (to support microbial growth). The yield of H2 is, however, subject to feedback inhibition by H2, requiring that the partial pres­sure of the gas be kept low to avoid problems with growth rate or a shift to acid production.

Another H2-forming fermentation has butyric acid as its major acidic product: C6H12O6 ^ C3H7COOH + 2CO2 + 2H2

although the molar production of H2 is only half that of acetate-accumulating strains.

The key enzyme in heterotrophic H2 producers is hydrogenase, an enzyme that catalyzes the reoxidation of reduced ferredoxin (Fd), an iron-containing protein reduced by ferredoxin-NAD and pyruvate-ferredoxin oxidoreductases, with the lib­eration of molecular hydrogen (figure 7.5):27

2Fd2+ + 2H+ ^ 2Fd3+ + H2

A summary of hydrogenase-containing bacteria is given in table 7.2.

Hydrogenases are a diverse group of enzymes and are often cataloged on the basis of the metal ion they contain as an essential component of the active site.28 The fastest H2-evolving species under laboratory conditions, Clostridium acetobutylicum, produces two different hydrogenases:29-31

• An iron-containing enzyme, whose gene is located on the chromosome

• A dual-metal (nickel, iron) enzyme whose gene is located on a large plasmid

Glucose

I Butyric Acid

Fructose-1, 6-bisphosphate

1

up to 0.59 ?

up to 0.9

?

up to 1.5

The iron-dependent hydrogenase from C. acetobutylicum has a specific activ­ity eightfold higher than similar enzymes from green algae even when all three enzymes are expressed in and purified from the clostridial host.32 The active sites of iron-dependent hydrogenases may be the simplest such structures yet studied at the molecular level. Of enormous potential importance for the industrial development of hydrogenases is the finding that even simple complexes of iron sulfide and CO mimic hydrogenase action.33 The crucial structure involves two iron atoms with different valency states at different stages of the reaction mechanism (figure 7.6). These findings raise the possibility of rational design of improved hydrogenases by the binding of novel metal complexes with existing protein scaffolds from known enzymes.

In contrast, nickel-iron bimetallic hydrogenases possess complex organometal — lic structures with CO and cyanide (CN-) as additional components, the metal ions bound to the protein via multiple thiol groups of cysteine resides, and an impor­tant coupling between the active site and iron-sulfur clusters.34,35 Multigene arrays are required for the biosynthesis of mature enzyme.36 Nevertheless, progress has been impressive in synthesizing chemical mimics of the organometallic centers that contain elements of the stereochemistry and atomic properties of the active site.37 The enzyme kinetics of nickel-iron hydrogenases remain challenging, and it is pos­sible that more than one type of catalytic activation step is necessary for efficient functioning in vivo.38 39

Such advances in basic understanding will, however, open the door to replacing expensive metal catalysts (e. g., platinum) in hydrogen fuel cells by iron — or iron/ nickel-based biocatalysts — the sensitivity of many hydrogenases to inhibition by O2

OC CO

Fe1 — Fe1———— Fe0 — Fe1

H’H H- ■

I I

Fe11—Fe1 ———— Fe11—Fe1

H+

FIGURE 7.6 Simple organometallic complexes as biochemical mimics of hydrogenase enzymes. (Modified from Darensbourg et al.33)

is a serious drawback but, of the many organism known to produce hydrogenases, some contain forms with no apparent sensitivity to O2 and can function under ambient levels of the gas.40

Constructing a large-scale bioethanol industry also implies a major change in the industrial landscape: whereas oil refineries are predominately coastal, biorefiner­ies would be situated in agricultural areas or (with the development of a mature industry) close to forests and other biomass reserves. Gasoline distribution to retail outlets is without doubt a mature industry — in the United States, shipments of 6.4 billion liters of petroleum and petroleum products are made each day, 66% by pipe­line (in 320,000 miles of pipeline), but only 4% by truck and 2% by rail; from the Gulf Coast to New York, shipping costs for gasoline amount to only 0.8 0/l.76 It was the early development of a national distribution system for gasoline that decided the use of this fuel rather than ethanol for the emerging automobile industry before 1920.77

To support nationwide consumption of E10, cellulosic ethanol would be 61% of the total, the remainder being corn-derived; assuming that switchgrass will be a major contributor to the feedstock mix, ethanol production would be centered in a wide swathe of states from North Dakota to Georgia, whereas demand would have geographic maxima from west to east (figure 5.7). Ethanol shipping would be predominantly by truck or rail until the industry evolved to take over existing petroleum pipelines or to justify the construction of new ones; linear optimization showed that shipping by truck would entail a cost of $0.13/l ($0.49/gallon), whereas rail transport would entail lower costs, $0.05/l ($0.19/gallon). In contrast, gasoline transportation to retail outlets only incurs costs of $0.003/l ($0.01/gallon).76 The same study concluded that national solutions, although they would spur innovation and eventually lead to economies of scale, would increase shipping distances and add to total truck movements; an investment of $25 billion would be required for a dedicated ethanol pipeline system, “just to make petroleum pipelines obsolete in the long-term.”

Tax incentives and subsidies are, therefore, highly likely to be features of policy making relevant to the adoption of biofuels in OECD economies generally. Fund­ing the ethanol supply chain will be crucial; minimizing shipping costs implies the construction of as many production sites as possible, based on the use of raw materi­als from multiple geographical areas (forest, agricultural wastes, dedicated energy crops, municipal solid waste, etc.), ideally to match the likely distribution of major urban demand centers for ethanol blends (figure 5.7).

Xylitol was a significant biochemical feature of the metabolic routes for the xylose presented to ethanologenic cultures in hydrolysates of the hemicelluloses from ligno — cellulosic biomass (chapter 2, section 2.3, and chapter 3, section 3.2). With an organ­oleptic sweetness to human taste approximately equivalent to sucrose, however, it is a fine chemical product in its own right as a low-calorie sweetener — xylose sugars
are not metabolized by the human consumers of xylitol-containing chewing gums (figure 3.2). Other envisaged uses include15

• Production of anhydrosugars (as chemical intermediates) and unsaturated polyester resins

• Manufacture of propylene and ethylene glycols as antifreeze agents and unsaturated polyester resins

• Oxidation to xylonic and xylaric acids to produce novel polymers (polyes­ters and nylon-type structures)

The production of xylitol for use as a building block for derivatives essentially requires no technical development, and if the xylose feedstock is inexpensive (as a product of biomass processing), then the production of xylitol could be done for very low cost.

The accumulation of xylitol during ethanologenesis from lignocellulosic sub­strates is, of course, unwanted and quite undesirable — for process as well as for eco­nomic reason (chapter 3, section 3.2). Viewed as an economically valuable product, xylitol formation and production acquire a different biotechnological perspective, and patenting activity has recently been intense (table 8.2). Biochemical efforts also continue to locate and exploit enzymes for bioprocessing hemicelluloses and hemi — cellulosic waste streams, for example:

•

An L-xylulose reductase identified from the genome sequence of the fil­amentous mold Neurospora crassa has been heterologously produced in E. coli for the production of xylitol.49

• A P-xylosidase from Taloromyces emersonii has been shown to be superior to the enzyme from the industrial fungus Hypocrea jecorina in releasing xylose from vinasse, the solid-waste material from ethanol fermentations.50

• Xylan residues in hemicelluloses can be variably esterified with acetyl, fer — uloyl, and p-coumaryl residues; hemicellulose deacetylating esterases have been characterized from fungal species and shown to only be highly effec­tive when mixed in multiplexes of enzymes capable of using all the possible structures as substrates for enzyme action.51

• Feruloyl esterases have recently been designed as novel chimeric forms with cellulose and hemicellulose binding proteins to improve their efficien­cies with plant polymeric substrates.52,53

• Ferulic acid has a number of potential commercial applications as an anti­oxidant, food preservative, anti-inflammatory agent, photoprotectant, and food flavor precursor; major sources — brewer’s spent grain, wheat bran, sugarbeet pulp, and corn cobs — make up 1-2% of the daily output of the global food industry, and ferulic acid can be released from brewer’s grain and wheat bran by feruloyl esterases from the thermophilic fungus Humi — cola insolens.54

Once acquired, a hemicellulose hydrolysate contains a variety of hexoses and pentoses; yeast highly suitable for the production of xylitol from the xylose present in the mix may, however, preferentially utilize glucose. One solution to this problem is to remove the rapidly utilized hexoses before removing the cells and replacing them with a purposefully xylose-grown cell batch, as was demonstrated with dilute acid hydrolysates of corn fiber.55 With Saccharomyces cerevisiae expressing the Pichia stipitis gene for xylose reductase, the presence of glucose inhibited xylose uptake and biased the culture toward ethanol production; controlling the glucose concentra­tion by feeding the fermentation to maintain a high xylose:glucose ratio resulted in a near-quantitative conversion of xylose to xylitol, reaching a titer of 105 g/l xylitol.56 With a double-recombinant strain of S. cerevisiae carrying the xylose reductase genes from both P. stipitis and Candida shehatae, quite different microbial biochem­istry occurred in a more process-friendly formation (at close to theoretical levels) of xylitol from a mixture of xylose (the major carbon source) and glucose, galactose, or mannose as the cosubstrate — indeed, the presence of the cosubstrate was manda­tory for continued metabolism of the pentose sugar.57 In a gene-disrupted mutant of C. tropicalis, with no measurable xylitol dehydrogenase activity, glycerol proved to be the best cosubstrate, allowing cofactor regeneration and redox balancing, with a xylitol yield that was 98% of the maximum possible.58

Bioprocess engineering for xylitol production appears straightforward; aerated cultures with pH regulation and operated at 30°C appear common. With Brazilian sugarcane bagasse as the source of the hemicellulosic sugars, xylitol from the har­vested culture broth was recovered at up to 94% purity crystallization from a clari­fied and concentrated broth.59

Xylitol dehydrogenase enzymes catalyze reversible reactions. The catabolism of xylitol proceeds via the formation of D-xylulose (figure 3.2). Fungal pathways of L-arabinose utilization can include L-xylulose as an intermediate, a much less common pentulose sugar.60-63 Such “rare” sugars are of increasing interest to metabolic bio­chemists because it has long been appreciated that many naturally occurring antibi­otics and other bioactives contain highly unusual sugar residues, sometimes highly modified hexoses derived by lengthy biosynthetic pathways (e. g., erythromycins); many secondary metabolites elaborated by microbes may only exhibit weak anti­biotic activity but are or can easily be converted into chemicals with a wide range of biologically important effects: antitumor, antiviral, immunosuppressive, anti- cholesterolemic, cytotoxic, insecticidal, or herbicidal.64 To the synthetic chemist, therefore, the availability of such novel carbohydrates in large quantities offers new horizons in development novel therapeutic agents: for example, L-xylulose offers a promising route to inhibiting the glycosylation of proteins, including those of viruses.65

Equally, such unconventional chemicals may have very easily exploited proper­ties as novel fine chemicals. The hexose D-tagatose is an isomeric form of the com­monly occurring sugar D-galactose (a hexose present in hemicelluloses, figure 1.23) and has attracted interest for commercial development.66 Because of its very rare occurrence in the natural world, a structure such as d-tagatose presents a “tooth — friendly” metabolically intractable sugar to human biochemistry. Fortuitously, the enzyme L-arabinose isomerase includes among its spectrum of possible substrates d-galactose; the enzyme can be found in common bacteria with advanced molecular genetics and biotechnologies, including E. coli, Bacillus subtilis, and Salmonella typhimurium and, when expressed in suitable hosts, can convert the hexose into d — tagatose with a 95% yield.67 Even more promising for industrial use is the efficient bioconversion of d-galactose to d-tagatose using the immobilized enzyme, more active than free L-arabinose isomerase and stable for at least 7 days.68 Both the enzyme and recombinant L-arabinose isomerase-expressing cells can be used in packed-bed bioreactors, the cells being particular adaptable to long production cycles.69,70

Synthetic chemistry offers only expensive and low-yielding routes to the rare sugars, but uncommon tetroses, pentoses, and hexoses can all be manufactured with whole cells or extracted enzymes acting on cheap and plentiful carbon sources, including hemicellulose sugars; bioproduction strategies use an expanding toolkit of enzymes, including D-tagatose 4-epimerase, aldose isomerase, and aldose reduc­tase.71-73 Together, the rare sugars offer new markets for sugars and sugar derivatives of at least the same magnitude as that for high-fructose syrups manufactured with xylose (glucose) isomerase.

Iogen’s choice of a wheat straw feedstock was made on practical and commercial grounds from a limited choice of agricultural and other biomass resources in Canada available on a sufficiently large scale to support a bioethanol industry (section 4.1). Wheat, as a monoculture, is inevitably subject to crop losses arising from pathogen infestation. Modern biotechnology and genetic manipulation offer novel solutions to the development of resistance mechanisms as well as yield improvements through increased efficiency of nutrient usage and tolerance to drought, and other seasonal and unpredictable stresses. But the deliberate release of any such genetically modi­fied (GM) species is contentious, highly so in Europe where environmental cam­paigners are still skeptical that GM technologies offer any advantage over traditional plant breeding and are without the associated risks of monopoly positions adopted by international seed companies, the acquisition of desirable traits by “weed” species, and the horizontal transfer of antibiotic resistance genes to microbes.282 The positive aspects of plant biotechnology have, in contrast, been succinctly expressed:

Genetic transformation has offered new opportunities compared with traditional breed­ing practices since it allows the integration into a host genome of specific sequences leading to a strong reduction of the casualness of gene transfer.283

Because large numbers of insertional mutants have been collected in a highly manip — ulable “model” plant species (Arabidopsis thaliana), it has been possible for some years to inactivate any plant gene with a high degree of accuracy and certainty.

In comparison with natural gas-based FT syntheses, biomass requires more inten­sive engineering, and gas-cleaning technology has been slow to evolve for indus­trial purposes — although for the successful use of biomass, it is essential because of the sensitivity of FT catalysts to contaminants. In year 2000 U. S. doller terms, investment costs of $200-340 million would be required for an industrial facility, offering conversion efficiencies of 33-40% for atmospheric gasification systems and 42-50% for pressured systems, but the estimated production costs for FT die­sel were high, more than 10 times those of conventional diesel.89 Two years later,

Biomass Air, O2

Syngas

the same research group from the Netherlands was predicting the same produc­tion costs and concluded that, unless the environmental benefits of FT diesel were valued in economic terms, the technology would only become viable if crude oil prices rose substantially.90 This did, in the event, occur (figure 5.1), and the cost differential has undoubtedly narrowed — although with no signs of a surge in investor confidence.

If it could be produced economically, using an energy crop such as switchgrass as the substrate, FT diesel rates better than E85 (from corn-derived ethanol) as a biofuel in assessments performed by the Argonne National Laboratory (figure 6.9).9192 FT diesel greatly outperformed E85 for total fossil fuels savings and also exhibited much reduced emissions of total particulates, sulfur oxides, and nitric oxides — although it fared worse than E85 using the criterion of total CO. Compared with conventional diesel fuel, FT diesel had higher total emissions of volatile organic carbon, CO, and nitric oxides (figure 6.9).

An experimental biofuel conversion technology, presently explored only in the Netherlands, is that for HydroThermalUpgrading diesel.93 At high temperature (300-350°C) and pressure, wet biomass feedstocks such as beet pulp, sludge, and bagasse can be converted to a hydrocarbon-containing liquid that, after suitable refining, can be blended with conventional diesel in any proportion without engine adjustments. A pilot plant remains the focus for further process optimization and development.