Since the ABE fermentation is already a mature process (see Section 10.3), the biggest potential for optimization is offered by metabolic engineering. Prerequisite is the knowledge of genome sequence and development of genetic tools, which are both available for C. acetobutylicum and C. beijerinckii. Electroporation has been established as a method of choice for gene transfer to C. acetobutylicum (Mermelstein and Papoutsakis, 1993; Nakotte et al., 1998; Tyurin et al., 2000) and C. beijerinckii (Birrer et al., 1994). For C. acetobutylicum, transformation efficiencies of up to 105-107 transformants/pg plasmid DNA are reported (Mermelstein and Papoutsakis, 1993; Tyurin et al., 2000), but methylation of the DNA proved to be essential prior to transformation (Mermelstein and Papoutsakis,

1993) . Recently, a modular system for Clostridium shuttle vectors was described (Heap et al., 2009), which comprises the most common origins of replication and selective markers for clostridia. During the last few years, major improvements in gene inactivation were achieved as well. Previously, it was only possible to silence genes by antisense RNA techniques (Tummala et al., 2005; Wagner and Simons,

1994) or inactivate (and respectively replace) genes by homologous recombination (Tomas et al., 2005; Tummala et al., 2005). However, the latter is very time­consuming, since the recombination frequency of clostridia is generally not very high. Another problem is the lack of temperature-sensitive plasmids or counterselectable markers for this genus that necessitates the use of non-replicative plasmids, which are rapidly degraded inside the cell by DNases and endonucleases.

To overcome this issue, Soucaille et al. (2008) designed a mutant strain of C. acetobutylicum with an inactivated restriction endonuclease system and a deleted upp gene. This gene encodes an uracil phosphoribosyl-transferase, which catalyzes transformation of 5-fluorouracil into a toxic product and can now be used as a counterselective marker on a respective plasmid. An even faster method is provided by the so-called ClosTron system (Heap et al., 2007; Heap et al., 2010; Shao et al, 2007). This system allows the rapid creation of integration mutants based on a sequence-specific group II intron from Lactococcus lactis.

Several genes involved in solventogenesis were already overexpressed or inactivated in C. acetobutylicum (Table 10.5). Highest butanol titers (238 mM) were reported for a strain with an overexpressed adhE and an inactivated orf5 gene (Harris et al., 2001). orf5 is located directly upstream of the sol operon (Fig. 10.3) and was proposed to encode the repressor of that operon (SolR) (Nair et al, 1994b). However, a more detailed study revealed that its gene product was actually localized extracellularly (which is in contrast to a transcriptional regulator) and is involved in glycosylation-deglycosylation reactions (Thormann and Durre, 2002; Thormann et al., 2002). The repressing effect observed stemmed from an intergenic region between orf5 and the sol operon (Thormann et al, 2002).

In addition to increasing butanol yields, some studies also focused on elimination of by-products and the improvement of substrate utilization and tolerance to a variety of stresses. Acetone formation was reduced by inactivation of acetone — producing genes ctfA/B (Sillers et al., 2009; Soucaille, 2008; Tummala, Junne, and Papoutsakis, 2003) and adc (Jiang et al, 2009). In this context, efforts are also ongoing to engineer the C. acetobutylicum mutant strain M5, which lost the megaplasmid pSOL1 and thus does not produce acetone at all (Lee et al., 2009b; Sillers et al., 2008). The production of the acids acetate, butyrate and lactate was decreased by inactivating the phosphotransacetylase gene pta (Green et al., 1996; Soucaille, 2008) and/or the acetate kinase gene ack (Sillers et al, 2008; Soucaille, 2008), the butyrate kinase gene buk (Green et al., 1996; Harris et al., 2000; Soucaille, 2008), and the lactate dehydrogenase gene ldh (Soucaille, 2008), respectively. However, the elimination of more than one by-product at the same time (in order to design a homo-butanol producer) still remains a challenging task. Especially, the elimination of ethanol as a by-product might be critical, since most butyraldehyde and butanol dehydrogenases also show activity with acetyl-CoA and acetaldehyde, respectively. To create a more robust strain, aerotolerance was prolonged by inactivation of perR (Hillmann et al., 2008) and tolerance to butanol was improved by overexpression of the groESL operon (Tomas et al., 2003b). The latter resulted in a strain that showed 85% less butanol inhibition and a prolonged metabolism that yielded 40% higher butanol titers (Table 10.5). Efforts to improve the substrate utilization of C. acetobutylicum only showed minor success so far. Xylose utilization was improved slightly by introduction of a transaldolase gene talA from E. coli (Gu et al, 2009), and overexpression of cellulosome components resulted in formation of a minicellulosome (see Section 10.2; Sabathe and Soucaille, 2003).

a C. acetobutylicum mutant strain M5, which lost the megaplasmid pSOLI (containing genes adhE, ctfA/B, adc, orf5 and adhE2) and does not produce solvents.

Meanwhile, metabolic engineering also allows butanol production in other organisms, which are easier to handle such as E. coli, Bacillus subtilis or the yeast Saccharomyces cerevisiae (Atsumi, Cann, et al, 2008; Atsumi, Hanai, et al, 2008; Dijk and Raamsdonk, 2009; Donaldson et al, 2007; Inui et al., 2008; Liao et al, 2008; Nielsen et al., 2009; Steen et al, 2008), or have a significantly higher tolerance against butanol such as Pseudomonas putida (Nielsen et al., 2009; Ruhl et al, 2009), or have the ability to grow on abundant substrates like synthesis gas such as Clostridium ljungdahlii (Kopke, 2009). However, the respective butanol yields are (still) insignificant compared to those of solventogenic clostridia (Table 10.6).

(Continued)

Note: acdh67 = acetylating aldehyde dehydrogenase, adh1/2 = alcohol dehydrogenase, adhE/adhE2 = alcohol/aldehyde dehydrogenase, ald = butyraldehyde dehydrogenase, bcd = butyryl-CoA dehydrogenase, bdhB = butanol dehydrogenase, ccr = butyryl CoA dehydrogenase, crt = crotonase, erg10 = thiolase, etfA/B = electron transferring flavoproteins, fnr = oxygen transcriptional regulator, frdB/C = fumarate reductase, hbd = 3-hydroxybutyryl-CoA dehydrogenase, ilvA = threonine deaminase, ldhA = lactate dehydrogenase, leuA = 2-isopropylmalate synthase, leuB = 3-isopropylmalate dehydrogenase, leuC/D = isopropylmalate isomerase, n. d.a. = no data available, pta = phosphotransacetylase, ter = trans-2-enoyl-CoA reductase, thlA = thiolase, yqhD = alcohol dehydrogenase.