Hydrogen Photoproduction by Immobilized N₂-Fixing Cyanobacteria: Understanding the Role of the Uptake Hydrogenase in the Long-Term Process

Abstract

We have investigated two approaches to enhance and extend H₂ photoproduction yields in heterocystous, N₂-fixing cyanobacteria entrapped in thin alginate films. In the first approach, periodic CO₂ supplementation was provided to alginate-entrapped, N-deprived cells. N deprivation led to the inhibition of photosynthetic activity in vegetative cells and the attenuation of H₂ production over time. Our results demonstrated that alginate-entrapped ΔhupL cells were considerably more sensitive to high light intensity, N deficiency, and imbalances in C/N ratios than wild-type cells. In the second approach, Anabaena strain PCC 7120, its ΔhupL mutant, and Calothrix strain 336/3 films were supplemented with N₂ by periodic treatments of air, or air plus CO₂. These treatments restored the photosynthetic activity of the cells and led to a high level of H₂ production in Calothrix 336/3 and ΔhupL cells (except for the treatment air plus CO₂) but not in the Anabaena PCC 7120 strain (for which H₂ yields did not change after air treatments). The highest H₂ yield was obtained by the air treatment of ΔhupL cells. Notably, the supplementation of CO₂ under an air atmosphere led to prominent symptoms of N deficiency in the ΔhupL strain but not in the wild-type strain. We propose that uptake hydrogenase activity in heterocystous cyanobacteria not only supports nitrogenase activity by removing excess O₂ from heterocysts but also indirectly protects the photosynthetic apparatus of vegetative cells from photoinhibition, especially under stressful conditions that cause an imbalance in the C/N ratio in cells.

INTRODUCTION

Many species of cyanobacteria and eukaryotic microalgae are capable of water biophotolysis, a light-dependent photosynthetic reaction that results in water oxidation with concomitant generation of molecular H₂ and O₂. Among cyanobacteria, N₂-fixing heterocystous species are the most promising candidates as potential H₂ producers (1). In these species, H₂ is generated mainly by the nitrogenase enzyme as an obligatory by-product of N₂ fixation (2). Although nitrogenase is sensitive to inactivation by O₂, N₂ fixation in heterocyst-forming species is protected by localization in specialized heterocyst cells. The maintenance of microoxic conditions within heterocyst cells enables efficient N₂ fixation, even under atmospheric levels of O₂ (2, 3).

Hydrogen formed in heterocysts is usually recycled by uptake hydrogenase (Hup) (4, 5). The uptake of H₂ may contribute to the decrease in partial pressure of oxygen inside the heterocyst cells via a Knallgas (oxyhydrogen) reaction, which is beneficial for sustained N₂ fixation (6), and may also serve as an additional source of electrons for the nitrogenase enzyme and other processes (3). As a result, the vast majority of heterocystous strains isolated from the natural environment show very little net H₂ photoproduction under ambient air conditions (1, 2). Under nitrogen deficiency, nitrogenase catalyzes solely the reduction of protons to H₂ (2), thus improving H₂ photoproduction yields (7). In addition to uptake hydrogenase, many heterocystous cyanobacteria contain bidirectional (Hox) hydrogenase. This enzyme may also participate in H₂ recycling, though at substantially lower rates (4, 5, 8, 9). The primary function of the bidirectional hydrogenase in cyanobacteria is for regeneration of NAD(P)⁺ from NAD(P)H during dark fermentation with concomitant evolution of H₂ (8). This enzyme, however, is also responsible for short-term H₂ photoproduction and is believed to prevent overreduction of the photosynthetic electron transport chain at the onset of illumination, especially in anaerobic environments (9–11).

The real potential of cyanobacterial H₂ photoproduction has yet to be fully explored. This is at least partly due to criticism of the high-energy requirement of nitrogenase-dependent H₂ photoproduction, which requires at least 4 mol of ATP per mole of H₂ produced (12). Heterocystous species are the most promising cyanobacteria for H₂ photoproduction as they are the only organisms capable of driving the H₂ photoproduction process under strictly autotrophic conditions, in the presence of atmospheric levels of O₂, and for extended periods of up to a few months under solar light intensities and periodic light conditions (13, 14). While the efficiency of H₂ photoproduction by any known species is currently far below requirements for commercial application, there are several promising approaches for significant improvement.

The H₂ photoproduction yields in cultures of heterocystous cyanobacteria can be significantly enhanced via elimination of uptake hydrogenase activity in cells. A good example of this is the PK84 strain, a chemically induced mutant of Anabaena variabilis with impaired hydrogenases that, in contrast to the parental strain, is able to produce H₂ in air containing 2% CO₂ (15, 16). This mutant has also demonstrated H₂ generation under outdoor conditions, where it has performed successfully for up to 40 days (14). There have been other mutant strains lacking HupL or HupS subunits of uptake hydrogenase which have also demonstrated more efficient H₂ photoproduction (17–19). Among these are the ΔhupL and ΔhupL ΔhoxH mutants of Anabaena strain PCC 7120 (20) that were used in the current study. The inactivation of uptake hydrogenase in cyanobacteria not only affects H₂ production but also results in a prominent loss of nonrecycled metabolic energy that leads to a significant change in the overall metabolic equilibrium. This change induces a cascade of compensatory mechanisms affecting both oxygen reduction mechanisms in heterocysts and processes providing reducing equivalents for different metabolic activities, including N₂ fixation (21).

Another promising approach for improving H₂ production yields in cyanobacteria is to screen for strains with naturally high H₂ photoproduction activity. The H₂ production capacity of promising strains can be further enhanced using genetic engineering techniques, including elimination of the hup genes (as described above), modification of the active center of nitrogenase (22, 23), and changes in the light-absorbing properties of both vegetative cells and heterocysts in a similar manner to approaches employed in green algae (24) and purple bacteria (25). Our recent screening of the University of Helsinki Cyanobacteria Collection (UHCC) revealed several cyanobacterial strains with promising H₂ photoproduction rates (26). One of these strains, Calothrix strain 336/3, a filamentous N₂-fixing cyanobacterium, showed efficient H₂ photoproduction both in suspension cultures and immobilized films (26, 27). The yields and the rates of H₂ photoproduction of Calothrix 336/3 were higher than in most other N₂-fixing heterocystous species and were comparable to the rates in the ΔhupL mutant of Anabaena PCC 7120 (26, 27).

Future efforts into improving H₂ yield should also target the development of better light utilization capacities in cyanobacterial cultures and the direction of absorbed light energy solely into H₂ production rather than biomass accumulation. For better light utilization per surface area unit, future technology requires the entrapment of algae and cyanobacteria in thin films consisting of a few layers of photosynthetic cells. The concept of an artificial leaf has been recognized (28), and approaches for making such a leaf are under development. There are several different techniques that can be considered for further improvements, including natural biofilm formation on translucent matrices and latex coatings (29). Latex coatings (28, 30), including the wet coalescence approach (31), are promising methods for the entrapment of phototrophic organisms, but these methods are not usually as efficient as hydrogels for the immobilization of H₂-producing microalgae and cyanobacteria. Recently, a technique for entrapment of eukaryotic microalgae within thin-layer alginate matrices has been developed, which has demonstrated 1.5% conversion efficiency of light energy to H₂ (32). This same technique has also worked efficiently for immobilization of N₂-fixing heterocystous cyanobacteria (27). The entrapment of cyanobacterial cells within thin alginate matrices under nitrogen-deficient conditions reduced the accumulation of cell biomass, improved H₂ photoproduction yields, and prolonged the duration of the H₂ production process.

In the present work, we evaluate possible routes to long-term and efficient H₂ photoproduction by immobilized cells of N₂-fixing heterocystous cyanobacteria by (i) supplementing CO₂ periodically to the films producing H₂ under an Ar atmosphere, (ii) investigating the effect of different gas atmospheres (Ar, air, and N₂) on H₂ production and cell fitness, and (iii) applying periodic short-term air treatments to the films producing H₂ under an Ar atmosphere. We demonstrate that H₂ photoproduction yield can be enhanced through the restoration of cell metabolism, particularly photosynthetic activity in N-starved immobilized cells, by periodic CO₂ additions and air treatments. Our results also demonstrate the important role of uptake hydrogenase in photoprotection of the photosynthetic apparatus in the long-term process.

MATERIALS AND METHODS

Strains and growth conditions.

The wild-type (WT) Anabaena PCC 7120 strain was obtained from the Pasteur Culture Collection (Paris, France). The ΔhupL and ΔhupL ΔhoxH mutants of Anabaena PCC 7120 lacking uptake hydrogenase and both uptake and bidirectional hydrogenases, respectively, were kindly provided by H. Sakurai. The Calothrix 336/3 strain was selected from the UHCC as has been described previously (26). In this study, we always used cells grown under diazotropic conditions. Stock cultures of all strains were pregrown at room temperature in 150-ml Erlenmeyer flasks containing 50 ml of Z8 medium (33) without combined nitrogen (Z8x). The medium for growing stock cultures of the ΔhupL and ΔhupL ΔhoxH strains was supplemented with 25 μg ml⁻¹ spectinomycin or 25 μg ml⁻¹ spectinomycin plus 10 μg ml⁻¹ neomycin, respectively. The flasks were illuminated from the top with cool-daylight fluorescence lamps (Lumilux T8 15W/865; Osram) with light intensity of about 30 μmol photons m⁻² s⁻¹ photosynthetic active radiation (PAR). Before immobilization, cultures were transferred into 500-ml Erlenmeyer flasks containing 300 ml of Z8x medium and grown under ∼50 μmol photons m⁻² s⁻¹ in a growth chamber at 22°C. During this stage, antibiotics were not added to the ΔhupL and ΔhupL ΔhoxH mutant cultures. All flasks were sparged continuously with sterile air.

Cell immobilization and H₂ photoproduction experiments.

The details of the immobilization technique, which was originally developed for the entrapment of green algae in thin Ca²⁺-alginate films (32) and later adapted for cyanobacterial cells, are described in Leino et al. (27). After immobilization, the 3-cm² Ca²⁺-alginate strips were transferred into 23-ml vials containing 5 ml of Z8x medium, purged with Ar for 20 min, supplemented with 6% CO₂, and placed in a growth chamber at 26°C under continuous overhead illumination with cool-white fluorescent lamps (∼150 μmol photons m⁻² s⁻¹; Philips Master TL-D T8 15W/840). During long-term experiments, the medium in the vials was periodically changed to avoid general (nonnitrogen) nutrient deprivation.

At the beginning of each experiment, the headspace of all vials contained Ar supplemented with 6% CO₂. This CO₂ concentration was selected based on our previous data (27) that showed the optimal H₂ photoproduction activities in Anabaena and Calothrix 336/3 cultures. Depending on the approach used, the gas phase in the headspace of the vials was replaced periodically (at the beginning of each incubation cycle) with the following: a new portion of Ar supplemented with 6% CO₂ (Ar plus CO₂), pure air (air), air supplemented with 6% CO₂ (air plus CO₂), or N₂ supplemented with 6% CO₂ (N₂ plus CO₂). In the case of air treatments, the headspace in the vial was regenerated using Ar with 6% CO₂ after 16 to 20 h. This was performed to restore efficient H₂ photoproduction by the films. The H₂ and O₂ contents in the headspace of vials were monitored once a day using a gas chromatograph (GC) (Clarus 500; PerkinElmer, Inc.) equipped with a thermal conductivity detector and a molecular sieve 5A column (60/80 mesh). All experiments were repeated several times with at least three vials for each individual set.

Cell fitness.

After 12 days of incubation under different atmospheres (as described above), cells were recovered from alginate by solubilization of the films in 50 mM Na-EDTA solution (pH 7.0). The cells were washed three times using Z8x medium and centrifuged (16,100 relative centrifugal force [RCF] for 2 min). The cells were diluted with Z8x to adjust the optical density at 750 nm (OD₇₅₀) to values of 0.25, and 2 ml was transferred to the wells of a 24-well cell culture plate. The cells were then regrown under standard growth conditions at 50 μmol photons m⁻² s⁻¹ for 6 days, and the OD₇₅₀ was measured daily to identify differences in cell fitness.

Nitrogenase activity assay.

Nitrogenase activity was determined by an acetylene reduction assay (34). The alginate films with entrapped cells were transferred into 23-ml vials containing 5 ml of Z8x medium, purged with Ar, supplemented with 10% acetylene, and placed in a growth chamber at 26°C under continuous overhead illumination with cool-white fluorescent lamps (∼150 μmol photons m⁻² s⁻¹; Philips Master TL-D T8 15W/840) for 18 h. For ethylene determination, 10-μl samples from the headspace were injected into a GC (PerkinElmer Autosystem) equipped with a flame ionization detector (FID) and a CP-CarboBond column (Varian). Helium was used as a carrier gas, and calibration was performed with 1% ethylene (AGA, Finland). Enzyme activity was calculated on the basis of the chlorophyll (Chl) a content of the cells and per film area.

Chl determination.

The Chl a content in the alginate films was assayed in randomly chosen strips after solubilization of the alginate matrices in 50 mM Na-EDTA solution (pH 7.0). The cells were washed once with Z8x medium by centrifugation. Then, Chl a was extracted from the pelleted cells with 90% methanol and determined either spectrophotometrically at 665 nm (35) or using high-performance liquid chromatography (HPLC). The data obtained spectrophotometrically were used for determination of the specific rates of H₂ photoproduction (see Fig. 2A), while the data obtained using HPLC are presented in Fig. 5.

FIG 2.

Chl a content (A) and effective PSII yield (B) in alginate films with entrapped Anabaena PCC 7120 and ΔhupL cells. The measurements were performed before (0 h) and after (12 days) the treatment of the samples under Ar, air, and N₂ atmospheres supplemented with 6% CO₂. The Chl a contents in films were measured spectrophotometrically (34). The average values from three films are presented ± standard deviations. Significant differences between the pre- and posttreated samples are indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 5.

The visual changes in the pigment composition (A) and Chl a content (B) in alginate films with entrapped Calothrix 336/3, Anabaena PCC 7120, and ΔhupL cells during the long-term experiment. The experimental conditions are the same as described in the legend of Fig. 4. The Chl a contents in films were measured at the end of the experiment (∼600 h) by HPLC. The average values from three films are presented ± the standard deviations. Significant differences between the 0 h and posttreatment samples are indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001.

For HPLC analysis, the harvested cell pellets were frozen in liquid nitrogen and stored at −80°C before processing. The pigments were separated from 100% methanol extracts by HPLC (model 1100; Agilent Technologies, Germany) equipped with a diode array detector and a reverse-phase C₁₈ end-capped column (LiChroCART 125-4; Merck KGaA, Darmstadt, Germany). An acetonitrile-methanol-Tris buffer mixture (720:80:30) was used as a mobile phase. The Chl a standard was purchased from DHI Lab Products (Hørsholm, Denmark).

Photochemical activity.

The photochemical performance of cyanobacteria entrapped in alginate films was evaluated using a Dual-PAM 100 system (Walz, Effelrich, Germany). The films were removed from vials and placed in the center of a leaf holder (Dual-BA; Walz). The steady-state Chl a fluorescence level (F_t) was determined during the application of actinic red light intensity of ∼50 or 150 μmol photons m⁻² s⁻¹ for 5 min. Saturating light pulses of 300-ms duration and light intensity of 3,000 μmol photons m⁻² s⁻¹ were applied to determine a maximum fluorescence level at light (F_m′). The effective photosystem II (PSII) yield, Y(II), was calculated as (F_m′ − F_t/F_m′).

Statistical testing.

Student t tests were applied for significance at 95% using GraphPad Prism, version 5, for Windows (GraphPad Software, San Diego, CA, USA).

RESULTS

The effect of CO₂ supplementation cycles on long-term H₂ photoproduction activity.

Efficient H₂ photoproduction in N₂-fixing heterocystous cyanobacteria is only possible in the absence of N₂, where nitrogenase catalyzes solely the reduction of protons to H₂ (2). This activity also requires the presence of CO₂ (27). Therefore, in the first approach to achieving long-term H₂ photoproduction, the alginate films with entrapped wild-type and ΔhupL cells of Anabaena PCC 7120 were subjected to a number of CO₂ supplementation cycles, where each cycle was initiated by the change of gas phase in the headspace of vials to Ar containing 6% CO₂ (Fig. 1). A lack of CO₂ has previously resulted in a significant attenuation of H₂ photoproduction starting from the first cycle and loss of H₂ evolution on the conclusion of the next cycle (27). As shown in Fig. 1, CO₂ supplementation cycles allowed extension of the overall H₂ photoproduction period of immobilized ΔhupL cells to at least 600 h. Nevertheless, both strains demonstrated a gradual loss of H₂ photoproduction activity over the time of the experiment, whereby supplementation of 6% CO₂ could not completely recover earlier production rates (Fig. 1 and Table 1).

FIG 1.

Kinetics of long-term H₂ photoproduction (A) and changes in specific (B) and per area (C) rates of acetylene reduction by Anabaena PCC 7120 and its ΔhupL mutant entrapped in thin Ca²⁺-alginate films. For H₂ photoproduction, arrows indicate the points where the atmosphere of the vials was replaced with Ar supplemented with 6% CO₂. For acetylene reduction, white bars indicate Anabaena PCC 7120, and black bars indicate the ΔhupL mutant. Error bars show the standard deviations of triplicate samples. In the case of acetylene reduction, the difference between two strains was significant (P < 0.001) only for the 290-h point, when activities were calculated on a per area basis (C).

TABLE 1.

Summary of kinetic parameters of O₂ evolution and H₂ photoproduction in the wild-type Anabaena PCC 7120 and the ΔhupL mutant entrapped in thin Ca²⁺-alginate films

Cycle no.	Strain/genotype	Kinetic profilea
		Maximum specific rate of H2 production (μmol/mg Chl a/h)	Net H2 photoproduction yield (mol/m2)	Maximum specific rate of O2 evolution (μmol/mg Chl a/h)	Net O2 evolution yield (mol/m2)
1	Anabaena PCC 7120	7.1 ± 3.0***	0.07 ± 0.02***	60.0 ± 8.8***	0.19 ± 0.01**
	ΔhupL	31.9 ± 9.1	0.31 ± 0.08	78.1 ± 10.0	0.33 ± 0.06
2	Anabaena PCC 7120	2.9 ± 2.1***	0.03 ± 0.02**	18.2 ± 8.3	0.15 ± 0.01
	ΔhupL	16.6 ± 4.5	0.23 ± 0.09	21.4 ± 8.2	0.24 ± 0.09
3	Anabaena PCC 7120	1.8 ± 0.8**	0.01 ± 0.01*	19.4 ± 9.4	0.16 ± 0.04
	ΔhupL	10.9 ± 5.2	0.08 ± 0.05	22.3 ± 5.1	0.19 ± 0.03
4	Anabaena PCC 7120	1.5 ± 1.1*	0.002 ±.0.001	15.6 ± 9.3	0.13 ± 0.06*
	ΔhupL	9.2 ± 7.0	0.05 ± 0.04	12.8 ± 7.2	0.07 ± 0.01
5	Anabaena PCC 7120	1.1 ± 0.8	0.001 ± 0.001*	13.5 ± 9.1	0.12 ± 0.07
	ΔhupL	4.9 ± 3.8	0.02 ± 0.01	14.4 ± 4.4	0.11 ± 0.06
Total	Anabaena PCC 7120		0.12 ± 0.02***		0.74 ± 0.11*
	ΔhupL		0.70 ± 0.18		0.93 ± 0.14

^aThe long-term H₂ photoproduction activities in immobilized cells were maintained by periodic purging of the headspaces of the vials with Ar supplemented with 6% CO₂. Values represent an average of three independent experiments (with three vials in each) ± the standard deviations. Significant differences between the two strains are indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001.

As expected, H₂ photoproduction activity was more pronounced in the ΔhupL mutant due to the lack of uptake hydrogenase in this strain (20). Indeed, in the beginning of the first cycle, the nitrogenase activity measured as C₂H₂ reduction and calculated on the basis of Chl (specific nitrogenase activity) or per area unit was only slightly higher in the ΔhupL mutant than in the wild-type strain (Fig. 1B and C). After 290 h, nitrogenase activity calculated per area unit declined about 2-fold in the ΔhupL mutant but increased slightly in the wild-type strain (Fig. 1C). However, no significant changes were observed in specific rates of C₂H₂ reduction between the two strains (Fig. 1B, 290-h samples).

The maximum specific H₂ production rates for both strains (Table 1) were highest in the first cycle, with the ΔhupL strain demonstrating a significantly higher maximum rate (32 μmol H₂ mg Chl a⁻¹ h⁻¹) than Anabaena PCC 7120 (∼7 μmol H₂ mg Chl a⁻¹ h⁻¹). In five cycles (three independent experiments) (Table 1), ΔhupL cells entrapped in alginate film produced a total 0.7 mol m⁻² of H₂ in 780 h (∼33 days), while the films with Anabaena PCC 7120 produced only around 0.12 mol m⁻² of H₂ in 600 h (25 days).

As was observed for H₂ photoproduction, immobilized cells also demonstrated a decline in O₂ evolution over the course of experimental cycles (Table 1). The most pronounced change in the O₂ production rate occurred after the first cycle, where a decrease of about 70% was observed for both strains. In later cycles, the maximum O₂ production activity stabilized, and the gap between rates and yields of the two strains decreased (Table 1). The sum of final O₂ yields was higher in vials with ΔhupL mutant films than in vials containing Anabaena PCC 7120 films (0.93 and 0.74 mol m⁻² after 5 cycles, respectively). The reason for this is not clear since both O₂ evolution by photosystem II (PSII) and its consumption in respiratory pathways are responsible for the final O₂ yields.

The effects of different gas compositions (Ar, air, and N₂) in CO₂ supplementation cycles on cell fitness.

To study the possible negative effects of nitrogen deficiency on H₂ photoproduction activity in immobilized cyanobacteria and to clarify the role of uptake hydrogenase in adaptation processes, films with entrapped cells of Anabaena PCC 7120 and its ΔhupL mutant were placed under different atmospheres (Ar, air, or N₂). All atmospheres were supplemented with 6% CO₂, and the gas phase in the headspace of vials was regenerated to its initial composition every fourth day.

Under our experimental conditions, the ΔhupL mutant significantly outperformed the wild-type strain in H₂ production (see Fig. S1 in the supplemental material). In contrast to the wild-type strain, which produced H₂ only under the Ar atmosphere, the ΔhupL mutant also produced H₂ under atmospheres both with N₂ plus 6% CO₂ and air plus 6% CO₂, but at substantially reduced rates compared to the samples with Ar plus 6% CO₂. The loss of H₂ photoproduction yields in the presence of N₂ is a well-known phenomenon for N₂-fixing cyanobacteria and is caused by the change in the stoichiometry of H⁺ reduction by the nitrogenase enzyme (2). Nevertheless, these experiments allowed us to compare cell fitness of immobilized Anabaena PCC 7120 and ΔhupL cells under different experimental conditions.

Despite a very high volumetric cell density in the films (24), the alginate matrix did not completely prevent cell growth in the presence of high levels of N₂ and CO₂, where significant increases in Chl a content (Fig. 2A, air and N₂ samples) and OD₇₅₀ (data not shown) were determined. However, under the Ar atmosphere, where immobilized cells were subject to N deficiency, both Anabaena PCC 7120 and ΔhupL strains demonstrated a decline in Chl a over the course of the experiment. This was more pronounced in the films containing ΔhupL mutant cells, where Chl a decreased by ∼35% (Fig. 2A). Interestingly, under the atmosphere of air plus 6% CO₂, the Chl a content of the ΔhupL mutant did not change, but a significant increase was observed for the wild-type strain. Both conditions led to a significant inhibition of the effective PSII yield in the entrapped ΔhupL cells, where Y(II) declined from ∼0.34 to 0.15 (Ar plus 6% CO₂) and 0.17 (air plus 6% CO₂) after 284 h (∼12 days) of continuous treatments with respective gas compositions (Fig. 2B). The replacement of air with N₂ in the headspace of vials improved both Chl a content (Fig. 2A) and Y(II) (Fig. 2B) in the ΔhupL mutant.

After 12 days, the alginate-entrapped cells were recovered from the films to suspensions and equilibrated to the same cell density. Cell fitness was then monitored as regrowth under standard conditions for another 6 days (Fig. 3). Cells that were recovered from the films immediately after immobilization (i.e., not undergoing pretreatment) were used as a control. The results of the regrowth experiment confirmed that the cells recovered from the ΔhupL mutant films were indeed adversely affected under Ar and air atmospheres supplemented with 6% CO₂, with final OD₇₅₀ values that were significantly less than those of the respective control. The wild-type strain grew slowly only after pretreatment with Ar plus 6% CO₂ but still recovered considerably faster than the ΔhupL mutant. Indeed, following a 3-day lag period under standard growth conditions, the wild type pretreated with Ar plus 6% CO₂ finally reached an OD₇₅₀ that was not significantly different from that of the control. The pretreatment of the ΔhupL films under N₂ plus 6% CO₂ did not significantly affect cell fitness compared to the control ΔhupL cells which were grown under standard conditions without any pretreatment. It is worth mentioning that the control ΔhupL cells without any pretreatment regrew significantly more slowly (one-sided t test, P < 0.001) than the control wild-type Anabaena PCC 7120 (Fig. 3).

FIG 3.

Regrowth (OD₇₅₀) of the cells recovered from alginate films. Control films did not undergo any treatment (0 h), whereas treated films were subject to 12 days under an Ar, air, or N₂ atmosphere supplemented with 6% CO₂. The suspension cultures were regrown diazotrophically in 24-well plates under 50 μmol m⁻² s⁻¹ and an air atmosphere. Error bars indicate the means ± standard deviations from three biological repetitions. Significant differences between the treated and nontreated samples are indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001. d, days.

To further determine the possible impact of deletion of HupL on growth, we inoculated the control cells at two different values of OD₇₅₀ (0.1 and 0.5) and cultivated them under standard growth conditions (air with light intensity of 50 μmol photons m⁻² s⁻¹). When the cells were inoculated at an OD₇₅₀ of 0.5, only a very small difference was observed between the growth of wild-type Anabaena PCC 7120 and the ΔhupL mutant cells after the second day (see Fig. S2 in the supplemental material). The highest difference in the growth of the WT and ΔhupL cells was observed when the cells were inoculated at the lowest OD₇₅₀ of 0.1, indicating a possible high light sensitivity of the ΔhupL mutant cells.

The effects of periodic short-term treatments with air and air plus 6% CO₂ on long-term H₂ photoproduction activities in immobilized cultures.

The results obtained from the experiments described above indicated that long-term and efficient H₂ photoproduction might be achieved by protecting the fitness of immobilized cells through periodic and short-term exposure of films to a N₂-containing atmosphere. It was decided that these treatments should be applied between CO₂ supplementation cycles (Ar plus 6% CO₂) and for a time that is short enough to exclude significant cell growth inside the films but long enough to ensure the recovery of photosynthetic and nitrogenase activities in immobilized cyanobacteria. This condition was achieved by changing the gas phase to air for a short treatment of 16 to 20 h.

Since the continuous treatment with air plus 6% CO₂ applied in the cell fitness experiments adversely affected the ΔhupL mutant (Fig. 2 and 3), our next approach involved two conditions: periodically (every 3 to 4 days) changing the gas phase of the vials to contain (i) air with ambient levels of CO₂ (air sample) or (ii) air supplemented with 6% CO₂ (air plus CO₂). After about 16 to 20 h of either treatment to restore some level of cell fitness (Fig. 4, insets), the headspace was returned to Ar supplemented with 6% CO₂ in order to achieve maximal H₂ photoproduction capacity. In this experiment, the effect of treatments with air or air plus CO₂ on H₂ photoproduction was evaluated for four different strains: Anabaena PCC 7120, Calothrix 336/3, and ΔhupL and ΔhupL ΔhoxH mutants of Anabaena PCC 7120. Since the ΔhupL ΔhoxH strain demonstrated very similar results as the ΔhupL mutant, the data are not shown.

FIG 4.

Long-term H₂ (A, C, and E) and O₂ (B, D, and F) photoproduction yields from Calothrix 336/3 (A and B), Anabaena PCC 7120 (C and D), and the ΔhupL mutant of Anabaena PCC 7120 (E and F) entrapped in alginate films. In contrast to the experiment shown in Fig. 1, the cumulative yields are presented. The gas phase in the headspace of the vials was renewed periodically with Ar supplemented with 6% CO₂ (Ar + CO₂), air (air), or air supplemented with 6% CO₂ (air + CO₂). In the case of Calothrix 336/3, the kinetic curves from the films treated with air plus CO₂ with lower Chl a content per area unit are also presented. The insets represent the typical H₂ (inset E) and O₂ (inset F) photoproduction curves during a single cycle of air treatment.

Periodic treatments of immobilized cells with air, or air containing 6% CO₂, significantly increased H₂ production yield in Calothrix 336/3 compared to the N-deprived control cells (Fig. 4A, Ar + CO₂). With periodic air treatments, entrapped Calothrix 336/3 cells produced 0.17 mol H₂ m⁻² (600 h), and the addition of CO₂ to air treatments increased the yield to 0.26 mol H₂ m⁻². The effect of air treatments was even more pronounced in thin films of Calothrix 336/3 with a lower Chl a content per surface area unit (113 versus 215 mg Chl m⁻² in the original films). When treated with air supplemented with 6% CO₂, the low-Chl a thin films with entrapped Calothrix 336/3 cells reached more than 0.5 mol H₂ m⁻² by 650 h (Fig. 4A).

In contrast to the Calothrix strain, no significant changes were observed in Anabaena PCC 7120 under any treatments (Fig. 4C). For this strain, the rate of H₂ production declined dramatically after ∼70 h in all samples, but cells continued producing H₂ at very low rates (similar to the Ar-treated curves presented in Fig. 1 and in Fig. S1 in the supplemental material) to a final value of approximately 0.1 mol H₂ m⁻². As expected, H₂ photoproduction yields were significantly higher in the vials with the ΔhupL mutant (Fig. 4E). The ΔhupL cells entrapped in alginate produced the highest level of H₂ (1.06 mol m⁻²) when treated with air. In line with the effects observed during extended H₂ production under CO₂ supplementation treatments (Fig. 2 and 3), the supplementation of 6% CO₂ during the periodic 16- to 20-h air treatments actually decreased the H₂ photoproduction yield of this ΔhupL strain (0.61 mol m⁻²) (Fig. 4E). However, this yield was still higher than that of both wild-type strains.

Cyanobacterial cells entrapped in films did not produce any (both wild-type strains) or only negligible amounts (ΔhupL mutant) of H₂ during periods of air treatment (Fig. 4E, inset). These results were independent of the presence or absence of introduced CO₂. Indeed, during exposure to air, cyanobacterial cells entrapped in the films actively fixed N₂ (as detected by GC) (data not shown) and therefore demonstrated significantly diminished stoichiometry of H₂ production (2, 4, 5). Interestingly, during the periods of air treatment (no CO₂ supplementation), all strains also demonstrated decreased O₂ evolution (Fig. 4F, inset). The inhibitory effect often remained noticeable during the first ∼24 h after the return of the gas to Ar. After this lag period, the O₂-evolving ability of the cells recovered and led to accumulation of O₂ in the headspace of vials.

Although air treatments, through restoring the fitness of immobilized cells, improved overall H₂ photoproduction yields of Calothrix 336/3 and ΔhupL cells, this approach could not drive continuous H₂ photoproduction. All strains finally stopped producing H₂ although this endpoint was reached at different times for the different strains (Fig. 4A, C, and E). Importantly, the periodic treatments of cells with air supplemented with 6% CO₂ significantly improved the O₂ production yields in all strains (Fig. 4B, D, and F), but the effect was less pronounced in the ΔhupL mutant films (Fig. 4F). The air-only treatments slightly improved O₂ yields in both wild-type strains compared to untreated cells (Fig. 4B and D), with the ΔhupL mutant (Fig. 4F) demonstrating a more pronounced increase. These results demonstrate the potential benefit of periodic air treatments for the recovery of photosynthetic apparatus and cell fitness in N-starved cells. However, we cannot directly link the O₂ yield determined from the headspace with the photosynthetic activity of the cells due to difficulties in distinguishing between photosynthetic O₂ evolution and O₂ consumption by terminal oxidases.

Changes in chlorophyll content and photochemical activity of immobilized cells during the long-term H₂ photoproduction experiment.

For a more sensitive study of photosynthetic machinery, we measured Chl a content and effective PSII yield, Y(II), of the samples during long-term H₂ photoproduction experiments where treatments with air or air plus CO₂ were applied. Changes in the pigment composition of immobilized cyanobacteria during experiments were easily seen in the color of the films (Fig. 5A). The first noticeable change in the pigment composition occurred soon after the second cycle and progressed thereafter. To quantify Chl a more accurately, we used HPLC (Fig. 5B).

Under N deprivation, both the Anabaena PCC 7120 strain and its ΔhupL mutant demonstrated not only progressive decreases in H₂ and O₂ production activities but also significant degradation of Chl a in the films (Fig. 5B, Ar + CO₂). The most noticeable degradation of Chl a was detected in the ΔhupL mutant, in which Chl a degraded by 70% compared to the same culture in the beginning of the experiment (0 h). Interestingly, this mutant also showed significant Chl bleaching when the cells were regularly treated by air supplemented with 6% CO₂ (Chl a degraded by almost 80% on average), replicating the effect observed for the continuous treatment with air plus 6% CO₂ (Fig. 2). The treatments of the films with air only supported the Chl a contents in both Anabaena PCC 7120 and its ΔhupL mutant, with that of the ΔhupL mutant remaining approximately constant and that of the wild type slightly increasing (Fig. 5B, air). In the control (Ar plus CO₂) films with entrapped Calothrix 336/3, the Chl a level actually increased from the beginning of the experiment by approximately 20%. Air treatments further increased the Chl content in the immobilized Calothrix 336/3 films by around 50 and 60% in samples treated with air plus 6% CO₂ and air, respectively. Therefore, it is possible that some minor growth of Calothrix 336/3 occurred in the alginate films, although at a considerably lower rate than would occur in suspension cultures.

Measurements of photochemical activity from the surface of the films using the Dual-PAM 100 system were performed at the beginning (1 h) and in the middle (∼455 h) of long-term experiments. These measurements demonstrated that photosynthetic activity was significantly impaired in all studied strains under long-term N deprivation conditions (Fig. 6, Ar + CO₂ films). In these films, the effective PSII yield, Y(II), declined from 0.33 to 0.16 for Anabaena PCC 7120, from 0.29 to 0.18 for Calothrix 336/3, and from 0.34 to 0.06 the ΔhupL mutant after ∼455 h of the experiment. The periodic treatments of all three entrapped strains with air, which brought nitrogen back to the system, considerably restored PSII yield Y(II) to levels of 0.32 in Anabaena PCC 7120, 0.38 in Calothrix 336/3, and 0.39 in the ΔhupL mutant. Periodic changes of the medium during long-term experiments did not impart a noticeable effect on the restoration of photosynthetic apparatus (data not shown), thus eliminating the possible effects of a shortage of other nutrients. Interestingly, the entrapped cells periodically treated by air supplemented with 6% CO₂ adjusted their photochemical activity differently: Anabaena PCC 7120 demonstrated a similar PSII yield (0.30), and Calothrix 336/3 showed an even higher activity (0.44), whereas the ΔhupL mutant demonstrated significantly decreased PSII yield (0.12) compared to the cells treated with only air (Fig. 6).

FIG 6.

The effect of different treatments on the PSII yield, Y(II), in alginate films with entrapped Calothrix 336/3, Anabaena PCC 7120, and ΔhupL mutant cells. Measurements of photochemical activity from the surface of the films were performed after 1 h and in the middle (∼455 h) of the long-term experiment. The experimental conditions are the same as described in the legend of Fig. 4.

DISCUSSION

Supplementation of CO₂ prolongs H₂ production in immobilized N₂-fixing heterocystous cyanobacteria under an Ar atmosphere.

H₂ photoproduction in heterocystous, filamentous N₂-fixing cyanobacteria is mediated mainly by the nitrogenase enzyme located in the heterocyst cells, where H₂ is coproduced in the reaction of N₂ fixation. It is a highly energy-demanding process requiring 16 mol of ATP for every mole of N₂ fixed and H₂ produced. Although the efficiency of the reaction toward H₂ evolution improves significantly in the absence of N₂ (see Fig. S1 in the supplemental material), where nitrogenase exclusively catalyzes ATP-dependent reduction of H⁺ to H₂, 4 mol of ATP is still required for every mole of H₂ evolved. Consequently, under photoautotrophic conditions, where CO₂/HCO₃⁻ is the only carbon source, H₂ photoproduction in heterocystous cyanobacteria depends solely on the photosynthetic activity of vegetative cells that fix CO₂ and provide heterocysts with energy in the form of sucrose (36). Thus, long-term H₂ photoproduction by cyanobacteria requires CO₂ supply into the cultures (37, 38). Indeed, periodic supplementations of 6% CO₂ into vials containing Ca²⁺-alginate films with entrapped Anabaena PCC 7120 and ΔhupL mutant cells under an Ar atmosphere allowed us to sustain H₂ photoproduction for 600 h and 780 h, respectively (Fig. 1 and Table 1). The effect of 6% CO₂ supplementations was similar to that observed previously by our group for Calothrix 336/3 (27), where supplementations prolonged H₂ photoproduction for even longer periods (936 h) than the ΔhupL mutant in the present work, but at lower maximum H₂ yields per cycle.

Although supplementation of the cells with CO₂ does prolong H₂ production in heterocystous cyanobacteria, this is clearly not the only factor influencing the long-term performance of this process. Indeed, under an Ar atmosphere, we observed H₂ photoproduction yields to steadily decline in each subsequent cycle of CO₂ supplementation for both Anabaena strains (Fig. 1). Since the ΔhupL mutant does not have uptake hydrogenase, the decline of H₂ photoproduction yields for this strain could not be linked to an increase of H₂ uptake activity. Additionally, the ΔhupL ΔhoxH mutant of Anabaena PCC 7120, deficient in both hydrogenases, showed approximately equivalent kinetics of H₂ photoproduction as the ΔhupL strain (data not shown), indicating that the bidirectional hydrogenase is not involved either. The specific nitrogenase activity of the wild-type and ΔhupL strains did not show significant changes during treatments (Fig. 1B). Moreover, there was no significant difference in the specific nitrogenase activity between the wild-type and ΔhupL strains, which agrees well with previous data that Masukawa and coauthors obtained for suspension cultures (20). However, when the activity was calculated per area of film (Fig. 1C), a decrease in nitrogenase activity was observed in the ΔhupL mutant. This decrease could be due to the lysis of ΔhupL cells under N-deprived conditions. Thus, the decline of H₂ photoproduction yields in each subsequent cycle is likely caused by nitrogen, rather than carbon, deficiency.

Long-term nitrogen deprivation, but not anaerobiosis, affects cell fitness of immobilized cells under conditions suitable for efficient H₂ photoproduction.

Comparison of cell fitness of Anabaena PCC 7120 and its ΔhupL mutant after 12 days under Ar and N₂ atmospheres supplemented with 6% CO₂ clearly showed that Anabaena cultures entrapped in alginate films do not suffer from microoxic conditions since the immobilized cells continue growing under an atmosphere with N₂ plus 6% CO₂ (Fig. 2). These cells also showed the fastest recovery of growth in suspension culture after treatment (Fig. 3). The situation changed dramatically under the atmosphere of Ar plus 6% CO₂, where the cells suffer from nitrogen deficiency leading to imbalance in the C/N ratio. The incubation of both wild-type and ΔhupL cells under an Ar atmosphere in the presence of high CO₂ was harmful to cell fitness, as determined by regrowth experiments (Fig. 3). The situation was more dramatic in the ΔhupL mutant films, where the cell fitness was affected not only under Ar but also under air atmospheres supplemented with 6% CO₂ (Fig. 3). It should be noted that the decrease of O₂ partial pressure in vials, achieved by changing the headspace atmosphere to N₂ plus 6% CO₂, recovered the Chl a content (Fig. 2A), Y(II) (Fig. 2B), and cell fitness (Fig. 3) of this strain. These results indicate that the ΔhupL mutant might be more sensitive to excess O₂ than the wild-type strain.

Long-term nitrogen deprivation decreases photosynthetic yield and H₂ production activity in immobilized cyanobacteria, particularly in the ΔhupL strain.

Prolonged cultivation of alginate-entrapped cyanobacterial cells in N-deprived conditions caused a significant loss of photochemical activity in Calothrix 336/3 and in both Anabaena cell types, with a more pronounced effect observed in the ΔhupL mutant (Fig. 2B and 6, Ar plus CO₂ films). Under these conditions, presumably due to impaired protein biosynthesis, cyanobacteria could not efficiently repair the photosynthetic apparatus. A distinct loss of Chl a was also observed in both Anabaena cell lines under the atmosphere with Ar plus 6% CO₂ (Fig. 2 and 5). Contrasting this, the Chl a content in the Calothrix 336/3 strain was not affected by N deficiency. Moreover, this strain demonstrated a very small amount of growth, even under an Ar atmosphere (Fig. 5, Calothrix films with Ar plus CO₂).

The prolonged (Fig. 2B) or periodic (Fig. 6) treatments of two wild-type and ΔhupL cells with air, which brought nitrogen back to the system, allowed the cells to restore the photosynthetic apparatus. Interestingly, treatments with air supplemented with 6% CO₂ had different effects on the different species, whereby the treatment further improved (Calothrix 336/3), did not change significantly (Anabaena PCC 7120), or negatively affected (ΔhupL mutant), PSII yield compared to that of air-treated cells (Fig. 6). In the case of the ΔhupL mutant, this negative effect was accompanied by a pronounced bleaching of the films (Fig. 2 and 5). Under prolonged incubation with air plus 6% CO₂, the ΔhupL cells were undergoing lysis, which was confirmed by the extremely slow recovery of the cells during the regrowth experiment (Fig. 3).

The response of H₂ production to periodic 16- to 20-h air treatments was strain specific. For example, the ΔhupL mutant produced more H₂ when treated with air only (Fig. 4E), Calothrix 336/3 required the continuous presence of CO₂ in the headspace of the vials for improved production (Fig. 4C), and the wild-type Anabaena strain did not improve H₂ production under any condition tested (Fig. 4A). It has previously been shown that continuous supplementation of cells with N₂ suppressed the H₂-producing activity of Anabaena PCC 7120 and its hydrogenase(s)-deficient mutants (38). In that experiment, cultures were supplemented with 1% N₂ every day, which could have led to inhibition of H₂ production due to allocation of electrons to N₂ fixation and/or caused a metabolic shift in the C/N balance of microoxic cultures. A similar suppression was observed in our experiments under N₂ and air atmospheres supplemented with 6% CO₂ (see Fig. S1 in the supplemental material), whereas periodic 16- to 20-h air treatments, which were employed to provide nitrogen to the cultures for restoration of photosynthetic activity and cell metabolism, enabled some high H₂ photoproduction yields in ΔhupL and Calothrix 336/3 strains (Fig. 4).

While declines in photochemical activity were observed under nitrogen deficiency, the increase of H₂ uptake capacity over the time period of experiments could also influence H₂ production levels for the wild-type strains. It is known that the alginate matrix significantly restricts diffusion of gas both to and from immobilized cells (32, 39), thus increasing intracellular levels of H₂ and O₂. Such elevations in H₂ may enhance the activity (40, 41) and expression level of the uptake hydrogenase enzyme (42), while O₂ accumulated inside the films may favor the oxyhydrogen or Knallgas reaction (5, 6). However, not all heterocystous cyanobacteria show positive regulation of hupSL genes by H₂ (43). Since the nitrogenase activity (both specific and per area basis) in the films with the entrapped wild-type strain of Anabaena PCC 7120 did not change significantly by the time of the sharp decline in net H₂ photoproduction yield at around 100 h (Fig. 1 and 4C; see also Fig. S1 in the supplemental material), the decrease could only be caused by an increase in the rate of H₂ consumption. Clearly, more investigations are required for resolving the role of hydrogenase(s) in this process.

Uptake hydrogenase protects filaments during long-term N deprivation.

Comparison of net H₂ and O₂ yields in the wild-type strain of Anabaena PCC 7120 and its uptake hydrogenase mutant (Table 1) demonstrates a direct link between H₂ uptake and respiration in these strains. Over five cycles, the ΔhupL mutant produced a total of 0.58 mol m⁻² more H₂ than the wild-type strain. Assuming 100% consumption of H₂ in the oxyhydrogen reaction with a molar ratio of 2 to 1 H₂/O₂, the absence of this process would bring an additional 0.29 mol m⁻² O₂ into the final O₂ yields for the wild-type strain, yielding almost equivalent O₂ production in both strains (0.93 versus 1.03, respectively). However, this link between H₂ uptake and respiration was not as pronounced in air-treated samples (Fig. 4) due to a high sensitivity of the H₂ production and H₂ uptake enzymes to O₂. Similar results have been observed for Anabaena variabilis ATCC 29413 and its hydrogenase-impaired PK84 mutant (44). Thus, the decline in the rate of H₂ photoproduction in the wild-type strain over time is partly caused by the recycling of H₂ through the oxyhydrogen reaction. The other wild-type strain evaluated in our study, Calothrix 336/3, also demonstrated H₂ uptake activity in vials treated with Ar plus CO₂ by the end of each CO₂ supplementation cycle (Fig. 4A). This cyanobacterium possesses the genes encoding the uptake hydrogenase enzyme and demonstrates in vitro H₂ uptake activity (45). Therefore, it is likely that H₂ uptake in Calothrix 336/3 is also connected to the respiratory chain although the oxyhydrogen reaction in this strain might be less pronounced, leading to higher H₂ photoproduction yields (Fig. 4A) than in the Anabaena PCC 7120 wild-type strain (Fig. 4C).

When linked to the respiratory chain, uptake hydrogenase removes excess O₂ in heterocysts and provides the cells with ATP via the oxyhydrogen reaction. Both conditions are beneficial for the nitrogenase enzyme (3, 5). The importance of this reaction becomes clear in our long-term experiments, especially during periods of air treatment, when cyanobacterial cells experience an excess of O₂. Indeed, the ΔhupL mutant lacking uptake hydrogenase showed significant degradation of Chl a and photochemical activity not only in the films treated with Ar plus CO₂ but also, most importantly, in the films that were treated periodically with air supplemented with 6% CO₂ (Fig. 5 and 6). After prolonged incubation under an atmosphere with air plus 6% CO₂, the immobilized ΔhupL mutant cultures showed reduced cellular fitness (Fig. 3). In contrast, the cell fitness of the Anabaena PCC 7120 strain was affected only under the atmosphere with Ar plus 6% CO₂, while still recovering significantly faster than the ΔhupL strain. The reduction of O₂ partial pressure by placing the films under the atmosphere with N₂ plus 6% CO₂ was the only treatment able to fully recover the cell fitness of the mutant (Fig. 3).

We suggest that an increased level of O₂ inside the vegetative cells of the ΔhupL mutant, occurring due to efficient photosynthesis (in air plus CO₂) and to limited gas exchange through the alginate matrix, led to the penetration of a significant amount of O₂ into the heterocysts via narrow terminal pores that connect heterocysts with vegetative cells (46). In this case, the ΔhupL mutant was unable to fix N₂ efficiently during the treatment with air plus 6% CO₂ and thus failed to restore the photosynthetic apparatus. Subsequently, a dramatic shift in the intracellular C/N ratio would have caused a strong metabolic imbalance for ΔhupL cells. Under this condition, the ΔhupL mutant demonstrates significant chlorosis (Fig. 5) and pronounced inactivation of the photosynthetic apparatus (Fig. 6) in contrast to both wild-type strains, which possess active uptake hydrogenase. A similar protective role of uptake hydrogenase has been demonstrated in the unicellular cyanobacterium Cyanothece strain PCC 7822, where its involvement in the protection of nitrogenase is clear due to a single compartmentalization whereby, under diazotrophic conditions, the ΔhupL strain failed to grow even under atmospheric air (47).

In the absence of 6% CO₂ during air treatments, films with entrapped ΔhupL cells do not produce O₂ efficiently (Fig. 4F, inset) and can maintain microoxic conditions, probably through the upregulation of alternative pathways (21). This may include the expression of heterocyst-specific flavodiiron proteins, which are known to be involved in the protection of nitrogenase by redirecting excess electrons to O₂ (48–50). It is also possible that under low CO₂ levels, highly inducible Flv1A and Flv3A flavodiiron proteins, localized to the vegetative cells of Anabaena PCC 7120 (48) and an active photorespiratory pathway (51, 52), could contribute to lowering intracellular O₂ levels in vegetative cells, thereby also protecting heterocyst nitrogenase when uptake hydrogenase is absent. As a result, short-term, air-only treatments of ΔhupL mutant films resulted in the highest H₂ photoproduction yields in immobilized cultures (Fig. 4E).

We propose that the uptake hydrogenase in heterocystous cyanobacteria is not only important for supporting the high activity of the nitrogenase system by removing excess O₂ in heterocysts and providing cells with ATP via the oxyhydrogen reaction but also indirectly involved in protecting the photosynthetic apparatus of vegetative cells, especially under N-deprived conditions. Indeed, the diazotrophic filament heterocysts and vegetative cells are completely interdependent, and changes within one cell type affect dramatically another cell type. Although the absence of uptake hydrogenase in heterocysts can be partly compensated for by the upregulation of alternative mechanisms, these may fail to satisfy cyanobacteria in long-term culture, especially immobilized cultures, where the matrix polymer significantly restricts diffusion of gases and nutrients in both directions. In this context, the elimination of uptake hydrogenase in cells, while increasing H₂ photoproduction yield and rate, may significantly decrease the duration of the process. Clearly, more detailed studies are required for understanding the role of the H₂ uptake mechanism in immobilized strains.