As discussed above, nonbiological production of hydrogen is energy intensive and often associated with the production of greenhouse gas. Biologically, hydrogen can be produced by a variety of microorgan­isms possessing one of several different hydrogenases. In the cyanobacteria, enzymes involved in hydrogen metabolism belong to one of the three families discussed above: Hox, Hup or nitrogenase. The uptake hydroge — nase (Hup) is not useful for hydrogen evolution since it is poised to work unidirectionally, toward the recy­cling of H2 into H+. When hydrogen is produced by a heterotrophic organism, an organic carbon source (ulti­mately derived from photosynthesis) is used to provide protons and chemical energy to fuel hydrogen evolu­tion. Ironically, this is also true for cyanobacteria carrying out direct or indirect biophotolysis, at least on the molecular level. As discussed above, a complete photosynthetic apparatus uses water as proton donor, releasing molecular oxygen (Figure 22.1). Thus, the high sensitivity of hydrogenase to this gas dictates that both reactions cannot occur in the same place at the same time. The solution found by Nature was the most obvious one: changing the timing (indirect biophotolysis) or the space (direct biophotolysis).

In indirect biophotolysis the cell uses the chemical energy stored through the capture of sunlight, as

NADPH and ATP (Figure 22.1), to fix CO2 into organic compounds. These energy reserve molecules are then consumed in the dark to drive cellular metabolism, including nitrogen fixation by nitrogenase. The separa­tion of these reactions occurs naturally in several cyano — bacterial species by circadian control and in these strains dark hydrogen production by either nitrogenase or the bidirectional hydrogenase is frequently reported (Praba — haran et al., 2010; Troshina et al., 2002). An interesting characteristic found in many of these strains is a burst of hydrogen production when cells are reilluminated. This phenotype was characterized as a function of the bidirectional hydrogenase and hydrogen production ceases quickly as the O2 produced by photosystem II (Figure 22.1) accumulates in the cell, inactivating hydrogenase The production of H2 is thought to serve as an electron sink, helping the cell return to the proper redox state for carrying out the light reactions. In prac­tice, indirect biophotolysis could possibly be done as a large-scale production using a two-stage cultivation system. In a first stage, the cells are cultivated in the light and biomass is formed through photosynthesis. When the desired cell concentration is achieved and the cells have stored enough fixed carbon, a dark anaerobic culti­vation could follow, favoring proton reduction to hydrogen by hydrogenase. Thus, the water-splitting reaction is separated from H2 production in time and space. This system has being already demonstrated, where nitrogen limitation was also used to induce glycogen accumulation and increase hydrogen produc­tion yield in the second stage through the nitrogenase enzyme (Huesemann et al., 2009). In a similar approach with Synechococcus sp., the carbon accumulated in the first stage was converted into hydrogen in a second stage by a [NiFe] hydrogenase (McNeely et al., 2010).

H2 PRODUCTION BY HETEROCYSTOUS CYANOBACTERIA

Solar energy capture and hydrogen evolution by some filamentous cyanobacterial strains proceeds natu­rally in the presence of oxygen by confining the oxygen-sensitive processes to the heterocyst, a cell type that emerged shortly after the oxygenation of the earth’s atmosphere in what has been called the Oxygen Catas­trophe or Great Oxidation Event 2.6 billion years ago (Kumar et al., 2010; Mariscal and Flores, 2010). In this case the evolved hydrogen is produced by nitrogenase whose expression is restricted to the heterocyst under normal aerobic conditions (Murry et al., 1984). A num­ber of mechanisms are employed to protect nitrogenase from oxygen damage; heterocysts lack photosystem II so do not produce oxygen, gas diffusion into the heterocyst is restricted by a unique cell wall structure, and hetero­cysts possess a very active membrane-bound respiratory system that consumes trace amounts of entering oxygen.

Even so, some continual synthesis of nitrogenase is necessary to replace oxygen-damaged nitrogenase (Murry et al., 1983).

As discussed above, since heterocysts lack a complete photosynthetic apparatus, the necessary reductant is derived from fixed carbon imported from the neigh­boring vegetative cells through specialized interconnect­ing pore structures (Mariscal and Flores, 2010). The imported sugar is sucrose (Lopez-Igual et al., 2010) and it is metabolized though the oxidative pentose pathway (Summers et al., 1995) Thus, hydrogen produc­tion by heterocysts is essentially indirect biophotolysis on a microscopic scale, and since the energy captured by photosynthesis is first stored as fixed carbon, the maximal possible theoretical conversion efficiencies are reduced.

However, this system has been attractive due to its inherent robustness and has been studied for almost four decades (Benemann and Weare, 1974). Very reason­able conversion efficiencies, sustained for days to weeks, were achieved in early studies using nitrogen-limited cultures. Under laboratory conditions where higher effi­ciencies can be expected, conversion efficiencies (total incident light energy to free energy of hydrogen pro­duced) were shown to be 0.4% (Weissman and Bene- mann, 1977). Cultures incubated under natural sunlight (Figure 22.3) were able to attain an average con­version efficiency of 0.1% (Miyamoto et al., 1979a). Remarkably, even though there have been a large num­ber of studies since, very little improvement in yields

has been obtained. Thus, recent reports of conversion efficiencies found =0.7% under laboratory conditions (Berberoglu, 2008; Sakurai and Masukawa, 2007; Yoon et al., 2006) and 0.03—0.1% with natural sunlight (Sakurai and Masukawa, 2007; Tsygankov et al., 2002). Similar low efficiencies have been found with thermo­philic strains, which at least have the possible advantage of requiring less cooling (Miyamoto et al., 1979b, c). There should be room for improvement as theoretical efficiencies with this nitrogenase-based system have been calculated to be around 4.6% (Hallenbeck, 2011).

Since observed conversion efficiencies are lower than predicted, different strategies might be employed in order to improve overall performance, which is critically important since light conversion efficiencies directly impact on the photobioreactor footprint (doubling effi­ciency should halve the required surface area for the same amount of fuel production). For one thing, genetic engineering could be applied to optimizing the size of the photosynthetic antenna, since part of the reduction in efficiency is thought to be due to inefficient use of light energy at high intensities where more photons are captured than can be used and the excess energy is wasted. Another point that could be addressed is the hydrogen producing catalyst. Since half of the photon requirement is needed to provide ATP to nitrogenase ac­tion, replacing it with a hydrogenase, which does not require ATP for proton reduction, should in principle have an energy sparing effect. In a recent attempt to verify this, the [FeFe] hydrogenase from Shewanella onei — densis was expressed in Anabaena sp. under the control of a heterocyst-specific promoter with the required matu­ration genes (Gartner et al., 2012). Although it could be shown that active hydrogenase was made under the proper conditions, the increase in hydrogen production above the levels due to the coexisting nitrogenase was disappointingly small. Of course, under these condi­tions the two enzymes compete for the reductant; the true test would be to do this in a strain lacking nitroge — nase activity. Finally, it might in principle be a possible way to increase hydrogen production by increasing het­erocyst frequency. However, heterocyst frequency might already be close to optimal since even in long-term studies the H2/O2 ratio is close to the desired stoichiom­etry of two, what one would expect for optimal coupling between oxygen-generating photosynthesis in the vege­tative cells and hydrogen production by heterocysts.