Improving Marine Algae for Bioenergy 289
Kaiser and Nelson 2016). Such strains showed a photosynthetic performance comparable to wild-type
C. reinhardtii when grown at 25ºC, but a seriously reduced PSII activity when shifted to 37ºC, which
allowed on-going hydrogen production for three days. Moreover, after this period, mutants returned to
the wild-type phenotype upon incubation at 25ºC. This behaviour may elicit a cyclic hydrogen production
system with a biomass-generating phase at 25ºC and a hydrogen-production phase at 37ºC.
According to what has been discussed before, an important goal to optimize hydrogen production
is to enhance the electron flow from ferredoxin into hydrogenase. The identification of the potential
surfaces of interaction between ferredoxin and hydrogenase (Long et al. 2008) raises the possibility of
designing hydrogenase mutants to form strengthened complexes with ferredoxin, thus improving electron
transfer. In this regard, the HydA2 double mutant D102K/T99K has shown increased association rate
with ferredoxin (Long et al. 2008, 2009). Alternatively, electrons can be redirected to ferredoxin by
decreasing their flow to the competitive acceptor ferredoxin:NADP reductase. This was achieved with
a ferredoxin mutant with lower affinity for the competitor but a similar interaction with hydrogenase,
resulting in higher hydrogen synthesis in vitro (Rumpel et al. 2014). A more straightforward approach has
been the replacement of hydrogenase by a ferredoxin-hydrogenase fusion (Fd-HydA), in C. reinhardtii.
The chimeric protein accumulated to lower amounts than the wild-type hydrogenase but showed a
4.5-fold increased hydrogen photoproduction rate per mol of enzyme, indicative of a more efficient
electron transfer from PSI to hydrogenase (Eilenberg et al. 2016). Importantly, the ferredoxin module in
the fusion protein is not active as an electron intermediate between PSI and hydrogenase, as evidenced
by in vitro experiments carried out with thylakoid preparations, where addition of wild-type ferredoxin
was required for hydrogen release (Yacoby et al. 2011). The higher activity of the chimeric hydrogenase
can be explained taking in account that ferredoxin:NADP reductase is physically attached to PSI. Thus,
the ferredoxin module in the hybrid protein may direct the hydrogenase to a closer proximity to PSI,
facilitating electron transfer. A direct electron transfer from PSI to hydrogenase, avoiding the ferredoxin
intermediate, was achieved by fusing PSI subunit PsaE from Thermosynechococcus elongatus with
the oxygen-tolerant [NiFe]-hydrogenase from Ralstonia eutropha. The hydrogenase-PsaE fusion was
assembled together with a PsaE-depleted PSI isolated from a Synechocystis sp. PCC6803 mutant. The
resulting hydrogenase-PSI complex was able to sustain light-driven hydrogen production in vitro (Ihara
et al. 2006). Alternatively, the electron current from ferredoxin may be diverted to hydrogen synthesis by
hampering the activity of competing pathways. The use of specific inhibitors redirecting the electron flow
towards nitrogenase and bidirectional [NiFe]-hydrogenase has indeed enhanced hydrogen production
in the cyanobacterium Anabaena siamensis TISTR 8012 (Khetkorn et al. 2012a). Synechocystis mutant
strains disrupted in the nitrate assimilatory pathway (potentially competing with hydrogenase for reducing
potential) also showed increased hydrogen production (Baebprasert et al. 2011). A C. reinhardtii strain
(CC-2803) lacking Rubisco, the enzyme catalyzing the first step in CO 2 fixation through the Calvin cycle,
is able to produce hydrogen without submitting the cells to sulfur-starvation (Hemschemeier et al. 2008).
Due to CC-2803 being light sensitive and grows only in the dark, a similar strategy was attempted with
C. reinhardtii Rubisco mutants, carrying a single substitution at the small subunit which decreased the
structural stability of the enzyme. These strains were characterized by Rubisco low levels and enhanced
proteolysis (Esquivel et al. 2006; Pinto et al. 2013). Mutant Y67A showed particularly high rates of
Rubisco degradation, even in sulfur-repleted cultures, and this was accompanied by PSII inactivation,
low photosynthetic rate, high expression of Fe-hydrogenases, and high levels of hydrogen production
(around 9-fold compared to the wild-type strain under the same conditions). Rubisco manipulation may
be a target not only to down-regulate the Calvin-cycle but also to increase oxygen consumption through
photorespiration, since this enzyme is located at a metabolic crossroad between these two pathways
(Marin-Navarro et al. 2010). Cyclic-electron transport from ferredoxin back into plastoquinone is
another electron waste, in terms of hydrogen productivity. Photosynthetic organisms promote transitions
between linear (state 1) and cyclic (state 2) transport to balance the energy absorbed by PSII and PSI. The
C. reinhardtii mutant stm6, which is permanently blocked at state 1 (Kruse et al. 2005; Volgusheva et al.
2013) showed a 9-fold enhanced hydrogen production compared to wild-type. This is another example
of the efficacy of multi-targeting strategies, since this mutant presents other relevant phenotypes, such as
increased respiration rate (accelerating the transition into anoxia) and increased starch reserves.
Kaiser and Nelson 2016). Such strains showed a photosynthetic performance comparable to wild-type
C. reinhardtii when grown at 25ºC, but a seriously reduced PSII activity when shifted to 37ºC, which
allowed on-going hydrogen production for three days. Moreover, after this period, mutants returned to
the wild-type phenotype upon incubation at 25ºC. This behaviour may elicit a cyclic hydrogen production
system with a biomass-generating phase at 25ºC and a hydrogen-production phase at 37ºC.
According to what has been discussed before, an important goal to optimize hydrogen production
is to enhance the electron flow from ferredoxin into hydrogenase. The identification of the potential
surfaces of interaction between ferredoxin and hydrogenase (Long et al. 2008) raises the possibility of
designing hydrogenase mutants to form strengthened complexes with ferredoxin, thus improving electron
transfer. In this regard, the HydA2 double mutant D102K/T99K has shown increased association rate
with ferredoxin (Long et al. 2008, 2009). Alternatively, electrons can be redirected to ferredoxin by
decreasing their flow to the competitive acceptor ferredoxin:NADP reductase. This was achieved with
a ferredoxin mutant with lower affinity for the competitor but a similar interaction with hydrogenase,
resulting in higher hydrogen synthesis in vitro (Rumpel et al. 2014). A more straightforward approach has
been the replacement of hydrogenase by a ferredoxin-hydrogenase fusion (Fd-HydA), in C. reinhardtii.
The chimeric protein accumulated to lower amounts than the wild-type hydrogenase but showed a
4.5-fold increased hydrogen photoproduction rate per mol of enzyme, indicative of a more efficient
electron transfer from PSI to hydrogenase (Eilenberg et al. 2016). Importantly, the ferredoxin module in
the fusion protein is not active as an electron intermediate between PSI and hydrogenase, as evidenced
by in vitro experiments carried out with thylakoid preparations, where addition of wild-type ferredoxin
was required for hydrogen release (Yacoby et al. 2011). The higher activity of the chimeric hydrogenase
can be explained taking in account that ferredoxin:NADP reductase is physically attached to PSI. Thus,
the ferredoxin module in the hybrid protein may direct the hydrogenase to a closer proximity to PSI,
facilitating electron transfer. A direct electron transfer from PSI to hydrogenase, avoiding the ferredoxin
intermediate, was achieved by fusing PSI subunit PsaE from Thermosynechococcus elongatus with
the oxygen-tolerant [NiFe]-hydrogenase from Ralstonia eutropha. The hydrogenase-PsaE fusion was
assembled together with a PsaE-depleted PSI isolated from a Synechocystis sp. PCC6803 mutant. The
resulting hydrogenase-PSI complex was able to sustain light-driven hydrogen production in vitro (Ihara
et al. 2006). Alternatively, the electron current from ferredoxin may be diverted to hydrogen synthesis by
hampering the activity of competing pathways. The use of specific inhibitors redirecting the electron flow
towards nitrogenase and bidirectional [NiFe]-hydrogenase has indeed enhanced hydrogen production
in the cyanobacterium Anabaena siamensis TISTR 8012 (Khetkorn et al. 2012a). Synechocystis mutant
strains disrupted in the nitrate assimilatory pathway (potentially competing with hydrogenase for reducing
potential) also showed increased hydrogen production (Baebprasert et al. 2011). A C. reinhardtii strain
(CC-2803) lacking Rubisco, the enzyme catalyzing the first step in CO 2 fixation through the Calvin cycle,
is able to produce hydrogen without submitting the cells to sulfur-starvation (Hemschemeier et al. 2008).
Due to CC-2803 being light sensitive and grows only in the dark, a similar strategy was attempted with
C. reinhardtii Rubisco mutants, carrying a single substitution at the small subunit which decreased the
structural stability of the enzyme. These strains were characterized by Rubisco low levels and enhanced
proteolysis (Esquivel et al. 2006; Pinto et al. 2013). Mutant Y67A showed particularly high rates of
Rubisco degradation, even in sulfur-repleted cultures, and this was accompanied by PSII inactivation,
low photosynthetic rate, high expression of Fe-hydrogenases, and high levels of hydrogen production
(around 9-fold compared to the wild-type strain under the same conditions). Rubisco manipulation may
be a target not only to down-regulate the Calvin-cycle but also to increase oxygen consumption through
photorespiration, since this enzyme is located at a metabolic crossroad between these two pathways
(Marin-Navarro et al. 2010). Cyclic-electron transport from ferredoxin back into plastoquinone is
another electron waste, in terms of hydrogen productivity. Photosynthetic organisms promote transitions
between linear (state 1) and cyclic (state 2) transport to balance the energy absorbed by PSII and PSI. The
C. reinhardtii mutant stm6, which is permanently blocked at state 1 (Kruse et al. 2005; Volgusheva et al.
2013) showed a 9-fold enhanced hydrogen production compared to wild-type. This is another example
of the efficacy of multi-targeting strategies, since this mutant presents other relevant phenotypes, such as
increased respiration rate (accelerating the transition into anoxia) and increased starch reserves.