There is a wave of debate whether biofuels production and use effectively reduces carbon emissions. Undoubtedly, the universal answer does not exist yet. To assess environmental effects of GHG reductions one should consider the combined net effects of the energy technology associated with biofuels, carbon emissions, land conversion and agricultural production. These lead in fact to two types of effects: GHG reduction from land conversion for biofuels feedstock production (direct impact) and GHG reduction from off-site land conversion for biofuels feedstock production (indirect impact).

Accounting for these effects creates the opportunity to measure direct and indirect emission reductions. It is important for policy makers to obtain, as precisely as possible, a picture of the regulation’s potential on biofuels production. It is crucial for example, given that the majority of policy support is in the form of a subsidy, to understand all net effects from biofuels feedstock production (and consequent biofuels use) on GHGs to efficiently assess the subsidy rate. Current debate mainly focuses on the quantification of indirect effects. These results are difficult to quantify because an increased dependence on biofuels would mean increased demand for land to meet the requirements of off-site land conversion. As a consequence, significant zero (or negative) net impacts on climate change (i. e. in terms of increasing GHG emissions) would result. The risk of considerable carbon emission coupled with land use has been, until present, mostly ignored. Few studies (Hill et al., 2006; Zah et al., 2007; Searchinger et al., 2008) assessed the magnitude of increasing emissions from land-use changes, and there is still concern on the quantification issue for indirect effects. Substantial efforts are therefore needed to address the correct measurement of indirect effects on GHGs from land-use changes for biofuels feedstock production.

The conversion of land for agricultural activities (i. e. from forests to agricultural lands) causes considerable carbon emissions through time because this is released at consecutive stages during the conversion process. Positive net carbon costs would be obtained with the benefits arising from displacement effects of fossil fuels emissions gained over new land use for biofuels production. However, since time plays an important part when computing net benefits, it becomes essential for policy makers to consider a ‘justified’ period of time consisting of the lifetime of indirect effects of land-use changes. Some studies (Righelato and Spracklen, 2007) consider a 30-year time a justified period for indirect effects to occur. This is though based on the average time frame of ethanol plants and, as a consequence, the land change occurs as long as 30 years when ethanol feedstock production most probably takes place. Other studies (Renewable Fuels Agency, 2008) consider the payback period (the time that land conversion needs to give positive GHG impacts) of biofuels production arguing that most of carbon effects are intensified during the first ten years of land conversion because the release of

carbon is more sensitive. Marshall (2009) argues on two time periods for the lifetime of biofuels feedstock production: The first is a ‘project horizon’, the effective time period needed for biofuels feedstock to grow on a specific (converted) land. In essence, the time for which the converted land is planned to be used for feedstock production. This period could also be shortened or amplified according to changes occurring in biofuels technologies or at policy level (i. e. changes in the subsidy rate). The second is ‘impact horizon’ which considers the environmental aspect (carbon emissions) over the converted land for biofuels feedstock production. This would undoubtedly be not necessarily as long as the project horizon time span because its effects are generally prolonged over time. While, in fact, GHG reductions linked to biofuels production terminate as soon as the biofuels production (on that land) ceases, the consequent emission reductions would still remain in place (Marshall, 2009). Therefore, the distinction between these two time effects is important to assess effective policies for adequate land use. Knowing about the time periods for project and impact horizons would also mean recognising economically viable biofuels land-use changes and, consequently, efficient carbon emission strategies.

A similar issue to consider for measuring net indirect effects of land conversion is an ‘efficient’ discount rate for comparing the outcomes of various projects for land-use changes into biofuels activities. Some (Howarth, 2005) argue against high discount rates which reflect time uncertainty for future outcomes in investments for biofuels activities. Others (Marshall, 2009) assert that discounting functions should also be seen under a physical carbon content perspective. The aim is that comparisons across investments activities for setting up biofuels production should also be performed so that environmental considerations for payback mechanisms are consistent with sustainable practices.

Other environmental impacts of biofuels production can be found in numerous life-cycle assessments, mainly for biodiesel, in the transport sector (Booth et al, 2005; Bozbas, 2008). These studies normally conclude with recognising the positive effects in terms of GHG emission reductions. As concerning other pollutants, biodiesel and ethanol production also produce zero emissions in terms of sulphur dioxide (which, in general, is emitted during the burning of fossil fuels). Relevant reductions can also be seen in carbon monoxide and hydrocarbons (Nwafor, 2004; Schmidt, 2004). The literature seems controversial about the nitrogen oxide and dioxide emissions. Nitrogen oxide emissions in vehicles using a biodiesel engine are found at slightly higher levels than those in a conventional diesel engine. However, a modification of the engine would reduce these levels, and therefore this negative effect could be considered of no relevance (Booth et al., 2005). Nitrogen dioxide emissions would instead occur from biofuels feedstock processes which have potential effects on the ozone layer (Franke and Reinhardt, 1998).

Feedstock processes either for biodiesel or for ethanol production also present three further environmental effects such as fertility of soils, biodiversity and hydrological impacts (Kartha, 2006). Furthermore, large-scale use of monoculture

for biofuels production also has an impact on the environment through the excess use of fertilisers and pesticides. Biofuels feedstock production significantly affects the ecosystem either boosting biodiversity or threatening existing species and the natural habitat. On one hand, the use of set-aside lands for biofuels feedstock production causes, for example, water pollution (because of the use of fertilisers and pesticides) and affects local biodiversity. On the other hand, biofuels production offers a good example of biodiversity protection compared to other conventional agricultural practices. In several countries (e. g. Brazil) existing regulation requires leaving a proportion of land to natural flora and fauna to preserve biodiversity losses (Turley et al., 2002). Biofuels production poses a number of challenges to the management of soil fertility. First, recycling of small organic and plant nutrients is possible. Second, current agricultural practices (in particular in developing countries) for soil management depend on the wasted crop (though this is more relevant for biomass feedstock than biofuels). In addition, feedstock nutrients can be retrieved during land conversion processes and applied to the crop field for biofuels production rather than putting these in landfills. Finally, hydrological effects are also important. Some bioenergy crops require the same amount of water irrigation as food crops (i. e. sugar cane). However, as for food crops, it is essential for bioenergy crops to be guaranteed water infiltration from rainfall to avoid inefficiencies from water wastes.

Sorghum bicolor crop, also known as sweet sorghum (Fig. 4.9), is heat tolerant and is one of the most drought-resistant crops as it has the capability to remain dormant

during the driest periods, so it can grow in marginal land (Yuan et al., 2008). Sweet sorghum is one of the most promising candidates for bioethanol production in developing countries because it produces grains with high starch content, stalks with high sucrose content and leaves and bagasse with high lignocellulosic content (Smith and Buxton, 1993). It has been found that the production of ethanol from the hemicellulose and cellulose in bagasse is more favourable than burning it to make power in North China, although ethanol produced from the juice is very sensitive to the price of sugar (Gnansounou et al., 2005).

Methanol is the most commonly used alcohol in biodiesel production, as shown in Table 6.1, mainly because of its high reactivity and relatively low cost. However, sustainable methods of methanol production are currently not economically viable. It is typically produced from syngas that is in turn produced from a non­renewable source, namely natural gas. In addition, methanol is the most toxic and has the most deleterious effect on the biocatalyst activity compared to other alcohols. On the other hand, ethanol can be easily formed from renewable sources by fermentation. Using ethanol that is produced from renewable resources for biodiesel production makes the process entirely ‘green’.

The reason lipase-based biodiesel production has not reached commercial potential at present is the high cost of the enzyme and the loss of its activity. The main reason for the loss of activity is due to the inhibition effect of alcohol. As mentioned in Section 6.3, lipases are inactivated by contact with insoluble methanol that exists as drops in the oil, due to the strong polarity of the methanol that strips the active water from the enzyme’s active site (Lara and Park, 2004). Another potential problem that arises with the use of lipases is the by-product glycerol inhibition of the lipase due to its strong adsorption onto its surface. To overcome the problem of methanol inhibition of lipase, its amount should always be kept below its solubility limits in oil. To achieve this efficiently, a stepwise addition of methanol in a way to keep its amount below its solubility limit has been proposed (Shimada et al., 1999; Shimada et al., 2002). However, this solution does not take into account the problems with glycerol inhibition.

Other ways to overcome the problem include the use of an organic solvent, which are mainly used to dissolve the methanol and eliminate the stripping of the water molecules required for enzyme activation. The use of organic solvents also helps to reduce the effect of the by-product inhibition by dissolving the produced glycerol and to reduce the viscosity of the reaction media. Organic solvents such as и-hexane and ether have been studied (Tweddell et al., 1998; Oliveira and Oliveira, 2001; Al-Zuhair et al., 2008); however the solubility of methanol and glycerol in these solvents is low, and the above problems probably persist. Since the solubility of methanol is higher in 1,4-dioxane, using it as a solvent results in an increased yield of biodiesel production from triolein (Iso et al., 2001). However, large amounts of this solvent that make up 90% of reaction media were required to obtain reasonable conversion. The main disadvantages of using organic solvents are substrate dilution and the requirement of the addition of solvent recovery unit. On the other hand, it was found that when using t-butanol, a long-chain alcohol that does not inhibit the enzyme, as a solvent the enzymatic process is improved (Wang et al., 2006; Royon et al., 2007). t-Butanol dissolves both methanol and glycerol and at the same time it is not a preferred substrate for lipase that does not act as tertiary alcohols. The advantages of using t-butanol with immobilized lipase is further discussed in Section 6.8.

An alternative approach is to replace methanol with a different acyl acceptor such as methyl acetate (Wei et al., 2004). The reaction of triglyceride with methyl acetate is known as interesterification, which is similar to transesterification with the main difference being that the main by-product is triacetyl-glycerol rather than glycerol. Unlike methanol, methyl acetate has no negative effect on enzymatic activity and almost no loss in lipase activity detected even after being continuously used for 100 batches (Du et al., 2004; Wei et al, 2004). However, the main dis­advantage of this approach is that it proceeds in a much slower rate compared to when methanol is used in appropriate concentrations (Wei et al, 2004). In addition, the removal of the by-product tri-acetyl-glycerol is more difficult than glycerol.

Ionic liquids, which are salts that are liquid at room temperature, have also been proposed to replace conventional organic solvents often with improved process performance. Ha et al. (2007) assessed the effectiveness of using several types of ionic liquids in a lipase-catalyzed production of biodiesel from soybean oil and methanol. They reported that higher percent conversions were achieved using hydrophobic ionic liquids as compared to solvent-free systems. The cost of these ionic liquids is expected to hinder their commercial application.

Pure sugar is relatively easily converted to ethanol by a biochemical pathway (fermentation) in several steps. Production of a sugar solution from sugarcane and/or sugar beet is quite straightforward since fermentable carbohydrates can be obtained just by extracting the raw material with water. In the case of starch-based raw material, the process becomes a little bit more complicated. The fermenting organisms need mono — or disaccharides to produce ethanol, and since starch is a polymer, the fermentation rate is very slow. In order to increase the rate, the polymer has to be broken down into monomers. The polymeric starch in wheat or corn is stored in granules. When the granules are heated in water, the hydrogen bonds between the polymer chains are broken and a water solution of starch is formed. To this solution, enzymes called amylases and amyloglucosidases are added. These enzymes then hydrolyse the glycosidic bonds between the monomers which results in a solution of fermentable monomers. The amylases randomly cut the bond between two sugar units, reducing the chain length, while the amyloglucosidase peels off one sugar unit at a time from the ends of the chains. By themselves, these enzymes are very inefficient in cleaving the starch polymer into monomers, but working together (synergic effect) gives a very efficient hydrolysis.

The main parameters for consideration in the process integration are the feedstock, bioreactor configurations and methods for enhancing the process efficiency.

12.4.1 Feedstock

The range of organic matter types that can be subjected to anaerobic digestion is wide; from low organic load wastes such as municipal sewage to high organic load wastes (such as the organic fraction of the municipal solid wastes). Specifically, the industrial sector (e. g. breweries, sugar mills, distilleries, food­processing industries, tanneries, and paper, pulp industries) generate large amounts of organic wastes. Food products and agro-based industries contribute 65-70% of the total industrial wastewater in terms of organic load (Zafar, 2008). The term ‘residues’ is often preferred to ‘waste’ whenever possible, in order to assign a value of exploitable resource to the organic matter rather than being a problem to be solved. The term ‘feedstock’ is also often preferably used to describe the material fed to a digester.

Feedstocks with a high biomethane potential can be classified as follows:

• Agricultural (livestock manure, agricultural residues, animal mortalities, energy crops)

• Industrial (wastewater, sludges, by-products, slaughterhouse waste, spent beverages, biosolids)

• Municipal (sewage sludge, organic fraction of municipal solid waste)

A database including the biochemical methane potential of various feedstocks can be found at the website (http://www. emu-bioconversion. eu/). Data are collected (but not validated) from literature and, after standardisation of the units, the chemical composition and the methane potential of various feedstocks are included in the database.

The following sections provide a brief discussion of other environmental impacts that in addition to global warming are associated with biofuel systems:

Biodiversity

Agriculture and forestry have been the main drivers for biodiversity loss globally. For example, more land was converted to cropland over 30 years between 1950 and 1980 than over 150 years between 1700 and 1850 (MEA, 2005). Therefore, there is also a potential for biofuel crops to alter local habitats and resources in a way that will affect native species. These effects will depend on the crop, its density, duration and distribution on the landscape and any regular inputs, including water and chemicals (The Royal Society, 2008). Biodiversity loss can also occur due to direct effects of land-use change. For example, if set-aside land in Europe is used to grow biofuel crops, impacts on biodiversity will need to be evaluated because some of these areas are more biodiverse than farmlands (Critchley and Fowbert, 2000). Intensified cultivation of biofuel crops could also lead to new pests and diseases which could in turn lead to increased use of pesticides/herbicides, causing further environmental damage.

Introducing new, particularly more invasive, species into an area could lead to the displacement of local biodiversity. Eucalyptus, some Miscanthus species and switchgrass all exhibit some features of invasiveness (The Royal Society, 2008).

However, there is also some evidence that under certain circumstances biodiversity could increase. For example, large-scale short rotation coppice (SRC) such as willow can provide benefits for some bird species, butterflies and flowering plants (Anderson and Fergusson, 2006).

Therefore, it is important that the overall risks and benefits for biodiversity be evaluated appropriately for bioenergy feedstocks. The Royal Society (2008) recommends using a risk assessment framework that covers the following:

• the full life cycle of biofuel production;

• the invasiveness potential of the crop;

• potential interactive effects of the biofuel crop with other pressures in the area (e. g. drought stress);

• the impacts on ecosystems; and

• changes in these risks under a future climate.

However, the lack of data represent a significant barrier in addressing biodiversity on a life cycle basis as biofuel crops have not yet been assessed for their impacts on biodiversity. Furthermore, currently there is no agreed methodology on estimating the impacts on biodiversity in LCA.

Water use

Water is used throughout the life cycle of biofuels, from feedstock to biofuel production. However, water use is usually not included in LCA or other evaluations of environmental sustainability of biofuels. The main reasons are the lack of data and an agreed methodology for estimating the water footprint. Although there are some data available on water use for crops, water requirements through the rest of the supply chain are not available. This is not an issue specific to biofuels only but also to other systems — as water has started to become a global issue only relatively recently, the need for information on water consumption in different productive systems has only come to light recently. Consequently, no current LCA databases contain reliable data on water consumption so that it is not possible to provide reliable estimates of water usage on a life cycle basis.

Pyrolysis refers to material chemical change caused by the application of thermal energy in the presence of air or nitrogen (Fukuda et al, 2001). The pyrolysis of different triglycerides was used for fuel supply in different countries during the First and Second World Wars. For instance, a tung oil pyrolysis batch system was used in China as a hydrocarbon supply during the Second World War (Lima et al, 2004).

Different types of vegetable oils produce large differences in the composition of the thermally decomposed oil (Lima et al., 2004). Many kinds of vegetable oil species have been subjected to pyrolysis conditions. Some of these vegetable oils are soybean (da Rocha Filho et al., 1993; Lima et al., 2004), rapeseed (Billaud et al., 1995), palm tree (Alencar et al., 2002; Lima et al., 2004,), castor (Lima et al., 2004), safflower (Billaud et al., 1995), olive husk (Demirbas et al., 2000) and tung (Chand and Wan, 1947). Soybean oil pyrolysed distillate, which consists mainly of alkanes, alkenes and carboxylic acids, has a cetane number of 43, exceeding that of soybean oil (37.9) and the ASTM (American Society for Testing and Materials) minimum value of 40. The viscosity of the distillate was higher than the ASTM specification for diesel fuel but considerably below that of soybean oil (Lima et al., 2004). Short-term engine tests have been successfully carried out on this fuel (Hu et al., 2000). Used frying cottonseed oil pyrolysate has also been investigated (Knothe et al., 1997).

Biological production of hydrogen (bio-hydrogen) has received special attention during the last decade because it can be operated at an ambient temperature and pressure and is more environmentally friendly compared to other processes (Mohan et al., 2007). Due to the low cost and regeneration properties, biotechnology of hydrogen production might be the most important way for energy production in the near future (Balat and Balat, 2009). Furthermore, it offers sustainable supply of hydrogen with low pollution and high efficiency from a variety of renewable resources (Cheong and Hansen, 2006; Wu and Chang, 2007). Biological hydrogen production can be classified into the following groups:

1 Direct biophotolysis: The process uses the photosynthetic capability of green algae and cyanobacteria to split water by the directly absorbed light energy and concomitant transfer of electrons to a hydrogenase or a nitrogenase for H2 production (Kovacs et al., 2006).

2 Indirect biophotolysis: This process involves a photosynthetic biomass production step and an anaerobic dark fermentation of the biomass to produce H2. Several models to achieve indirect biophotolysis have been developed. These systems use algae in most cases and intend to exploit their capability to produce high biomass yield per surface. Main research includes production of algal biomass, which is rich in easily fermentable storage carbohydrates (Benemann, 1998).

3 Biological water-gas shift reaction.

4 Photo-fermentation.

5 Dark fermentation: This process is able to use biomass provided by a photosynthetic solar energy conversion system to H2 production (Keasling et al., 1998). Dark microbial H2 production is driven by the anaerobic metabolism of the key intermediate, pyruvate. The complete oxidation of glucose would yield a stoichiometry of 12 mol. H2 per mole of glucose, but in this case, no energy would be gained to support growth and metabolism of the producing organism (van Niel et al., 2002).

In a proposed integrated system, dark fermentation and photo-fermentation are combined in order to achieve maximal conversion of the substrate to biohydrogen (deVrije and Claassen, 2002).

Starch, cellulose or hemicellulose content of wastes, carbohydrate-rich food industry effluents or waste biological sludge can be further processed to convert the carbohydrates to organic acids and then to hydrogen gas by using proper bioprocessing technologies.

4.5 Acknowledgements

Authors want to give their sincere thanks to the following people and organisations for their generosity in letting us use their photos: Shu Suehiro, Hannes Grobe, Fabio Visentin, Huw Williams, Professor Krishna K. Shrestha (Central Department of Botany, Tribhuvan University), Hans Hillerwaer, Botanische tuin (TU Delft),

J. M. Garg, Paul Fenwick, Rich Weber (Native Trees of Indiana website), Iwata Kenichi (Okayama University), Mehmet Karatay, Moreno Clementi, Mehmet Karatay, Marco Schmidt (Senckenberg Research Institute), Teck Long Chen (Suriachem), Daniel Georg Dohne, Biofuels Center of North Carolina, Jose Maria Fernandez (University of Cordoba) and Brad Lashua. Authors gratefully acknowledge support for this research from the Spanish Ministry of Education and Science (ENE2007-65490/ALT and HI2008-0229) and from the Spanish International Development Cooperation Agency (AECID, PCI-C/019212/08).

Basic information on the kinetics of enzymatic reactions is available in the books of Bailey and Ollis (1986), Dutta (2008) and Shuler and Kargi (2001). These books present a biological background and provide a comprehensive introduction to biochemical engineering. Introduction to the genetic sequencing for producing proteins from recombinant DNA is also available in these books. However, an interested reader should refer to more specialized books (Martin and Christopher, 1990; Leskovac, 2003; Cook and Cleland, 2007) for more profound information, where much more is included about the structures of enzymes and the kinetics and mechanisms of enzymatic reactions. Enzymatic kinetics mechanism, relative rates of steps along the reaction pathway, and chemical mechanism, including acid-base chemistry and transition state structure for mono-, bis — and tri-substrate reactions are explained, and numerous general experimental protocols and kinetic data interpretation are described. In addition, a comprehensive catalog of enzymes in general, and lipase in particular, and their uses in modern manufacturing are available in the book of Polaina and Maccabe (2007) and the book of Uhlig (1998). These books survey general enzyme characteristics and discuss their microbiological origin, and stability of each enzyme. In addition, the most important industrial enzymes in use today are examined including immobilized enzymes.

As far as biodiesel is concerned, the book of Pahl (2005), ‘Biodiesel: Growing a New Energy Economy’, offers a comprehensive review from the history of the diesel engine to the development of the biodiesel industry, past, current and future. In addition, detailed information and news updates are available on the webpage of

the National Biodiesel Board (NBB) (http://www. biodiesel. org/). NBB is the

national trade association representing the biodiesel industry in the United States.

C. acetobutylicum was the first completely sequenced Clostridium. The genome sequence of the type-strain ATCC 824 (Weyer and Rettger, 1927) was already released in 2001 by Nolling et al. This provided valuable information and helped in further understanding of solventogenic clostridia. Meanwhile, the genome sequence of another major solventogenic Clostridium species, C. beijerinckii NCIMB 8052, is available too (JGI, 2005). While the genome of C. acetobutylicum consists of a 3.94-Mbp chromosome and a 192-kbp megaplasmid pSOLl, C. beijerinckii contains no megaplasmid but has a significantly larger chromosome with a size of 6.0 Mbp.

One surprising finding in the genome sequence of C. acetobutylicum was the presence of 11 genes, whose products were unambiguously identified as cellulosome components (Nolling et al, 2001; Sabathe, Belaich, et al., 2002). Although overexpression of single genes led to functional proteins (Lopez — Contreras et al., 2003; Perret et al., 2004) and also a minicellulosome could be produced in vivo (Sabathe and Soucaille, 2003), C. acetobutylicum is unable to ferment cellulose (Lee et al., 1985a; Lopez-Contreras et al., 2003).

However, xylan, another major component of lignocellulose besides cellulose, can be degraded by C. acetobutylicum as well as C. beijerinckii (Lemmel et al., 1986; Qureshi et al., 2006), and respective enzymes have been isolated and characterized (Ali et al., 2004; Ali et al., 2005; Lee and Forsberg, 1987; Lee et al., 1985b; Lee et al., 1987).

Several a-amylases for degradation of starch were found in C. acetobutylicum. Two of them were purified and analyzed in detail (Annous and Blaschek, 1994; Paquet et al., 1991), but only one of the respective genes (amyP = CAP0168) has been confidently identified from the genome sequence (Sabathe, Croux, et al., 2002).

Mono — and disaccharides are then taken up by phosphoenolpyruvate-dependent phosphotransferase systems, which are already well described in C. acetobutylicum and C. beijerinckii (see Table 10.3). Only galactose is transported by a non­phosphotransferase mechanism (Mitchell and Tangney, 2005).

Inside the cell, sugars are metabolized to pyruvate, hexoses via glycolysis and pentoses via the pentose phosphate pathway. Respective genes were found in the genome sequence of C. acetobutylicum (Nolling et al., 2001) and C. beijerinckii (JGI, 2005). Pyruvate is then converted to acetyl-CoA by a pyruvate:ferredoxin — oxidoreductase, one of the most oxygen-sensitive enzymes known (Meinecke et al., 1989). Lactate and acetoin are also produced from pyruvate (Fig. 10.2), catalyzed by lactate dehydrogenase (Freier and Gottschalk, 1987) and acetolactate synthase plus acetolactate decarboxylase, respectively, while the formation of acetate, butyrate, ethanol, acetone, isopropanol and butanol starts from acetyl-CoA (Fig. 10.2).

Acetate is produced via acetyl phosphate by successive action of phosphotransacetylase (Pta) and acetate kinase (Ack) (Fig. 10.2). The latter was purified from C. acetobutylicum and was shown to be a highly specific enzyme (Winzer et al, 1997). The respective genes pta and ack are located in a common operon (CAC1742-CAC1743; Fig. 10.3) on the genome of C. acetobutylicum (Boynton et al., 1996) and are also present in C. beijerinckii, arranged in exactly the same order. Butyrate is formed in analogous reactions from butyryl-CoA (Fig. 10.2). The respective enzymes phosphotransbutyrylase (Ptb) and butyrate kinase (Buk) are already characterized in detail (Hartmanis, 1987; Thompson and Chen, 1990; Wiesenborn et al., 1989b). The corresponding genes ptb and buk are clustered in a common operon on the genome of C. acetobutylicum (CAC3075-CAC3076; Fig. 10.3; Cary et al, 1988; Walter et al., 1993) and C. beijerinckii, respectively. Gene expression is relatively stable over the whole growth, similar to pta and ack (Fig. 10.3; Alsaker and Papoutsakis, 2005). However, in contrast to the acetate-producing enzymes, butyrate kinase activity can also be detected during solventogenesis (Andersch et al, 1983; Hartmanis and Gatenbeck, 1984). This might be attributed to a second butyrate kinase (BKII) found in C. acetobutylicum, whose physiological function is yet unknown (Huang et al., 2000).

Butyryl-CoA itself is produced from two molecules of acetyl-CoA by successive action of thiolase (ThlA), 3-hydroxybutyryl-CoA dehydrogenase (Hbd), crotonase (Crt) and butyryl-CoA dehydrogenase (Bcd) (Fig. 10.2). The respective genes are clustered on the genome of C. acetobutylicum and C. beijerinckii in the bcs operon (CAC2708-CAC2712; Fig. 10.3; Boynton et al., 1996), except the thiolase gene thlA (CAC2873) that is organized monocistronically (Stim-Herndon et al., 1995). Furthermore, a second thiolase operon was found on the megaplasmid of C. acetobutylicum (Winzer et al., 2000). While its physiological role is still unknown, ThlA has already been studied in detail (Wiesenborn et al., 1988). Hbd has been purified and characterized from C. beijerinckii (Colby and Chen, 1992) and Crt from a non-specified C. acetobutylicum strain (Waterson et al., 1972). Data on Bcd are scarce, but Inui et al. (2008) demonstrated that a pair of electron­transferring flavoproteins (EtfA/B), whose genes are also part of the bcs operon,

1 2000 4000 6000 8000 10000

10.3 Arrangement of the genes associated with solvent formation in C. acetobutylicum and their expression profile (according to Alsaker et al., 2004).

is essential for the activity of this enzyme. In Clostridium kluyveri, Bcd was shown to form a stable complex with EtfA/B (Herrmann et al, 2008; Seedorf et al., 2008), which is also involved in energy conservation via an Rnf complex (Herrmann et al., 2008; Seedorf et al., 2008). However, no Rnf complex is present in C. acetobutylicum, while respective genes were found in C. beijerinckii.

Formation of solvents starts with an acetoacetyl-CoA:acetate/butyrate-CoA transferase CtfA/B, which converts the previously produced acids acetate and butyrate into the respective acyl-CoA derivatives and acetoacetate (Fig. 10.2). While the recycled acetyl-CoA and butyryl-CoA are used for the production of alcohols such as ethanol and butanol via a number of aldehyde and alcohol dehydrogenases (see below), an acetoacetate decarboxylase Adc splits acetoacetate into CO2 and acetone, which is in some strains of C. beijerinckii further reduced to isopropanol by action of a primary/secondary alcohol dehydrogenase (Ismaiel et al., 1993).

CoA transferase and acetoacetate decarboxylase have already been studied in detail (Chen, 1993; Gerischer and Durre, 1990; Petersen and Bennett, 1990; Schaffer et al., 2002; Wiesenborn et al., 1989a). In C. acetobutylicum, the respective genes are located on the megaplasmid directly next to each other in a convergent direction (Fig. 10.3). While adc forms a monocistronic operon, ctfA/B are arranged in the sol operon together with the genes orfL (encoding a small peptide of still unknown function) and adhE (coding for a bifunctional butyraldehyde/butanol dehydrogenase) (Fischer et al., 1993; Zickner et al., 1993). This distinguishes C. acetobutylicum from all other solventogenic clostridia, where adc is part of the sol operon and adhE is replaced by an ald gene encoding an aldehyde dehydrogenase (Berezina et al., 2009; Toth et al., 1999). A typical oA-dependent promoter was found upstream of adc (Gerischer and Durre, 1992), whereas two promoter sequences were deduced for the sol operon by primer extension experiments (Fischer et al., 1993; Nair et al., 1994b). However, reporter gene studies revealed that only the distal sequence P1 represents a promoter and the proximal P2 is obviously an mRNA processing site (see below; Thormann et al., 2002). Both operons show a similar expression profile with a massive upregulation at transition from acidogenesis to solventogenesis (Fig. 10.3; Alsaker and Papoutsakis, 2005). The signal leading to this induction is still unknown, but it is proposed that various extra — and intracellular parameters such as temperature, low pH and high concentration of (undissociated) acetic and butyric acid (Ballongue et al., 1985; Gottwald and Gottschalk, 1985; Huesemann and Papoutsakis, 1986; Monot et al., 1983; Terracciano and Kashket, 1986), limiting phosphate or sulphate concentrations (Bahl, Andersch, and Gottschalk, 1982; Kanchanatawee and Maddox, 1990), levels of butyryl phosphate and butyryl-CoA (Boynton et al., 1994; Gottwald and Gottschalk, 1985; Zhao et al., 2005), ATP/ ADP ratio and NAD(P)H level (Grupe and Gottschalk, 1992) are involved. All these factors result in less negative supercoiling and thus relaxation of DNA. This change in topology could serve as a transcriptional sensor by allowing or restricting regulatory proteins to bind (Wang and Syvanen, 1992; Wong and Bennett, 1996; Ullmann and Durre, 1998; Ullmann et al., 1996). Several binding motifs have been identified in intergenic regions of sol and adc; three binding sites for the master regulator of sporulation Spo0A were found upstream of the adc promoter and another 0A box is located upstream of the sol promoter (Ravagnani et al., 2000). Gel retardation and targeted mutation experiments confirmed binding of phosphorylated Spo0A to these sites (Hollergschwandner, 2003; Ravagnani et al., 2000), thus proving that the regulatory networks of solventogenesis and sporulation are linked. Furthermore, a potential binding site for the catabolite control protein CcpA (cre sequence; Feustel, 2004; Nold, 2008) and three imperfect repeats (R1, R2 and R3; Thormann et al., 2002; Scotcher et al., 2003) were found upstream of the sol promoter. The region downstream of the sol promoter forms a very complex secondary structure with several predicted stem loops. This structure seems to be important for processing of the adhE mRNA (Thormann et al., 2002), but might also affect the adhE expression negatively (Scotcher et al, 2003).

The adhE transcript finally yields two different products, the mature bifunctional enzyme and the C-terminal alcohol dehydrogenase, probably due to a second translation start within the same mRNA (Thormann et al., 2002). While overproduction of AdhE in C. acetobutylicum leads to increased NAD+-dependent butyraldehyde/acetaldehyde dehydrogenase activity and NADH-dependent butanol/ethanol dehydrogenase activity (Nair et al, 1994a), only minor aldehyde dehydrogenase activity could be detected after purification (Thormann, 2001). Heterologous overexpression in Escherichia coli resulted in no enzyme activity at all (Lorenz, 1997; Nair et al., 1994a). In addition to AdhE, two butanol dehydrogenases BdhA and BdhB (sometimes also referred to as BdhI and BdhII) are present in C. acetobutylicum. Although these isoenzymes have a high identity, BdhB has a significantly better affinity to butyraldehyde than BdhA (Petersen et al., 1991; Walter et al., 1992). The respective genes are organized in monocistronic operons directly next to each other on the genome of C. acetobutylicum (Fig. 10.3). In C. beijerinckii, three different butanol dehydrogenases could be identified (Chen, 1995) in addition to the aldehyde dehydrogenase Ald (Toth et al., 1999; Yan and Chen, 1990).

Under special conditions of high NAD(P)H availability, a second bifunctional alcohol/aldehyde dehdrogenase AdhE2 is formed in C. acetobutylicum, representing the first case of an organism that possesses two such enzymes (Fontaine et al., 2002). Such conditions could be induced by addition of artificial electron carriers such as methyl viologen dyes (Rao and Mutharasan, 1986, 1987) or growth on reduced substrates such as glycerol (Fontaine et al., 2002) and lead to alcohologenic fermentations with butanol and increased levels of ethanol, but no acetone as products (Girbal et al., 1995). The gene adhE2 is also located on the megaplasmid of C. acetobutylicum and organized as a monocistronic operon approximately 47 kbps upstream of the sol operon. A homologous gene is present in the genome of C. beijerinckii as well.

Several types of bioreactors have been used for hydrogen production via water biophotolysis (direct and indirect). These reactors require adequate entry of light,
which could be sunlight or another artificial light, such as red light. To maximize the area of incident light, which allows high cell growth and hydrogen production, the reactor design should provide a high surface-to-volume ratio. In addition, it should allow sterilization and easy handling. Furthermore, the photobioreactor should be an enclosed system so that the produced hydrogen could be collected without any losses. Photobioreactors can be mainly divided into three types: vertical column reactor, tubular type and flat panel photobioreactor. A summary of bioreactor types and their properties is provided in Table 13.2 (Dutta et al., 2005).