Butanol has many desirable properties as a fuel and thus is a suitable target for modification of

cyanobacteria. In fact, as a fuel it is superior to ethanol, being less corrosive and less volatile. Thus, it can easily be mixed with hydrocarbon-based fuels and used in the same infrastructure. A number of recent studies have shown that engineered cyanobacteria can in fact make surprisingly high levels of this compound, at rates that in fact surpass published rates for ethanol production by engineered cyanobacteria. Since cyanobacteria pro­duce this fuel directly through photosynthetically driven CO2 fixation, it is appropriate to compare the pro­ductivity per area of this process, as presently described, with that required for other biofuels, be it growing corn to produce the necessary sugars, or growing algae to produce biodiesel. Such a comparison shows that butanol production by cyanobacteria could be much bet­ter than making fuels from corn and very comparable to making biodiesel from microalgae (Sheehan, 2009).

Many cyanobacteria are capable of producing volatile compounds, including higher alcohols, but the natural production levels are miniscule (Hasegawa et al., 2012). In order to make fuel molecules at significant quantities, new pathways must be introduced as well as changes made to the native metabolic pathways. Studies on creating heterotrophic bacterial strains capable of producing butanol demonstrated that two possible routes were useful: the 2-ketoacid pathway, normally involved in amino acid biosynthesis, and the acetyl-CoA pathway, found in organisms such as Clos­tridium that naturally produce butanol during fermenta­tion (Figure 22.4).

The first successful attempt in this direction was to engineer S. elongates to produce isobutyraldehyde through the 2-ketoacid pathway (Atsumi et al., 2009). Isobutyraldehyde is a precursor for isobutanol and other chemicals of interest and has the advantage of being highly volatile, easing its recovery from the culture broth thus removing product inhibition. The strategy applied consisted of boosting carbon flux through the pathway from pyruvate to 2-ketoisovalerate by integra­tion into the genome of three foreign genes, alsS, ilvC and ilvD, catalyzing these steps, as well as kivd from Lactococcus lactis, the gene encoding the ketoacid decar­boxylase enzyme that converts 2-ketoisovalerate to iso — butyraldehyde. Overall carbon flux was then increased by integrating an additional copy of Rubisco (rbcLS) and the resulting strain produced 6230 mg isobutyralde — hyde per liter per hour, a production rate that is higher than any other fuel molecule made by cyanobac­teria to date. Additionally, it was demonstrated that isobutanol could be formed if a foreign alcohol dehydro­genase (YqhD from E. coli) was introduced, but titers were lower, presumably due to product inhibition.

However, the isomer that is made by the 2-ketoisovalerate pathway is isobutanol, a fuel additive, but not nearly as desirable in itself as a fuel as n-butanol, the product of the acetyl-CoA pathway or the 2-ketobutyrate pathway (Figure 22.4). Metabolic engineering was used to create an n-butanol-producing strain of S. elongatus by introducing the hbd, crt, and adhE2 genes from C. acetobutylicum, the ter gene from Treponema denticola, and the atoB gene (instead of thl) from E. coli (Lan and Liao, 2011). However, n-butanol was only produced by this strain under anaerobic con­ditions, either in the light when photosystem II was inhibited by DCMU, or in the dark, which gave the highest production, a meager 20.8 mg/L/h. It was sug­gested that anaerobic conditions were necessary since some of the enzymes introduced are oxygen sensitive, severely limiting its usefulness. On the other hand, metabolic fluxes are obviously different during dark metabolism than during photosynthesis and the differ­ence could be in the supply of a key metabolite. In line with this, in a more recent attempt to create an n-butanol-producing strain, flux through acetyl-CoA was increased by substituting an irreversible ATP hydrolysis step leading to the formation of acetoacetyl CoA (Lan and Liao, 2011). Other improvements con­sisted of substituting NADPH-requiring enzymes for NADH enzymes. With these changes, it was possible to demonstrate light-dependent n-butanol production, but at 62.5 mg/L/h this is well below (by a factor of 100) the initial promising results with butyraldehyde. This system would need very significant improvement before it could be considered for practical biofuel production.