Although the heterocyst/nitrogenase-based system has been the most studied, some other known cyanobac — terial hydrogen-producing reactions could poten­tially be used for biological hydrogen production. These include the unicellular and nonheterocystous filamentous cyanobacteria, which possess nitrogenase and are able to fix nitrogen in nature. Two strategies are employed to avoid oxygen inhibition. In some uni­cellular species, oxygen evolution and nitrogen fixation (or hydrogen production) are separated in time since photosynthesis and nitrogen fixation are under circa­dian control with photosynthesis taking place during the day and nitrogen fixation being maximal during the night period. The filamentous cyanobacterium Tri — chodesmium uses a strategy of spatial segregation where nitrogen fixation occurs in cells located in the middle of the bundle carrying out the oxygen-sensitive nitroge- nase reactions and the others carrying out the normal photosynthetic reactions (Berman-Frank, 2001).

The unicellular cyanobacterium Cyanothece has been the subject of a number of recent studies demonstrating prolonged hydrogen production in the light mediated by nitrogenase. In one study, considerable hydrogen production (up to 465 mmol per milligram of chlorophyll per hour) was shown, the growth conditions were very stringent and hydrogen production was only observed when the culture was submitted to nitrogen starvation, sparged with argon to remove any oxygen formed through photosynthesis, supplemented with glycerol and cultivated under low light (Bandyopadhyay et al., 2010; Min and Sherman, 2010). Glycerol, in addition to serving as a possible additional energy source to sup­port nitrogenase activity, appears to release nitrogenase from diurnal control (Aryal et al., 2013). Another recent study found appreciable hydrogen and oxygen produc­tion with nitrogen-depleted cultures that were incu­bated under continuous illumination (Melnicki et al.,

2012) . Light saturation curves and photosynthesis inhi­bition studies indicate that the hydrogen is evolved indi­rectly from the fixed carbon produced through photosynthesis. Here again, the requirements for contin­uous illumination (it can hardly be energetically positive to produce hydrogen using artificial illumination) and for argon sparging raise serious hurdles to practicality. Thus, although a nice proof of principle, such a system would hardly be economically viable.

Many cyanobacteria also possess Hox, a soluble reduced nicotinamide adenine dinucleotide (NADH)- linked [NiFe] hydrogenase. This reversible hydrogenase is capable of hydrogen evolution, in particular when dark-adapted cells are reilluminated (Schwarz et al., 2010). As discussed above, this forms an electron valve, readjusting the poise of the photosynthetic apparatus, but activity is quickly inhibited with renewed oxygen evolution. A recent survey showed that a diversity of cya­nobacteria contains this enzyme and that there is great variability in both the amounts of hydrogen made by this enzyme and the pattern of hydrogen evolution (Kothari et al., 2012). This enzyme is also responsible for evolution during dark fermentation of endogenous reserves, principally glycogen, and hydrogen production by this pathway can be enhanced through lowering of the hydrogen partial pressure (Ananyev et al., 2012). At least in Synechocystis, hydrogen production by Hox can be increased by eliminating the master regulator AbrB2, which normally represses synthesis of Hox (Dutheil et al., 2012; Leplat et al., 2013). In a recent attempt to increase hydrogen production, heterologous expression of the [FeFe] hydrogenase from Clostridium acetobutylicum was carried out in the non-nitrogen-fixing cyanobacte­rium Synechococcus (Ducat et al., 2011). Active hydroge — nase was formed under proper conditions, but in vivo light-driven hydrogen production from this system was significant only when the cultures were incubated under an inert atmosphere and oxygenic photosynthesis was completely inhibited.

Ethanol

While hydrogen production, or at least direct bio­photolysis, can be driven directly by photosynthesis, all other biofuels must use the capacity of cyanobacteria to drive carbon dioxide fixation with photosynthetically derived energy, ATP and reductant. However, once fixed by the Calvin-Benson-Bassham cycle, the newly recycled carbon can be converted to useful biofuels through the introduction of novel (to cyanobacteria) metabolic pathways (Angermayr et al., 2009). The first such cyanobacterial-derived biofuel that was demonstrated was ethanol (Deng and Coleman, 1999; Dexter and Fu, 2009), and its production is the only cyanobacterial — produced biofuel under active investigation and com­mercial development (Algenol Biofuels: http://www. algenolbiofuels. com/). Algenol Biofuels is presently claiming production at "around $1.00 per gallon using sunlight, carbon dioxide and saltwater at production levels above 9000 gallons of ethanol per acre per year". At an average solar insolation for Florida of 19.8 MJ/day and since ethanol has a higher heating value of 29.7 MJ/ kg, this translates to a claim of a very impressive 2.8% conversion efficiency. Now another company, Joule Unlimited (http://www. jouleunlimited. com/), has step­ped into the picture, offering to sell SunFlow-E through its fuel company, Joule Fuels. Their process uses geneti­cally modified thermophilic cyanobacterium containing Moorella alcohol dehydrogenase, and their Web site claims are even more spectacular with targets of up to

25,0 gallons per acre (7.8% conversion efficiency) and $0.60 per gallon at full-scale commercial production.

Cyanobacteria can naturally produce relatively min­ute amounts of ethanol so at the simplest level, creating a cyanobacterium that produces higher levels of ethanol involves boosting flux through the ethanol pathway through the introduction of the key enzymes for conver­sion of pyruvate, generated by glycolysis of the fixed carbon, to ethanol, pyruvate decarboxylase (pdc) and alcohol dehydrogenase (adh). This alone is sufficient to produce low millimolar levels of ethanol in the medium upon prolonged (5—10 days) incubation and growth. Further increases, obviously necessary for practical pro­duction, have been achieved through a variety of means, including better transcriptional control and further metabolic engineering. Most of this development work is being done at private enterprise laboratories, but a recent published report (Gao et al., 2012) shows that impressive increases in yields can be achieved by inte­grating a foreign pdc and a native adh into the genome of Synechocystis and abolishing carbon flux into polyhy — droxybutyrate synthesis.