Biosolar Hydrogen Production from Cyanobacteria

BiosolarH2: Synechococcus sp. PCC 7002

Arthrospira maxima CS-328 and strain JSC-1

Photosystem I-Hydrogenase Chimeras

Cyanobacteria and other photosynthetic bacteria are well-tuned biological devices that can harvest solar energy, the one limitless source of energy on Earth, and convert that energy into fuels, including hydrogen, fermentable biomass, and a variety of reduced carbon compounds. Cyanobacteria, because of their very limited requirements for growth, are particularly attractive organisms for such conversion processes, since they derive their reducing power from water. Furthermore, Synechococcus sp. PCC 7002 is a euryhaline organism capable of rapid growth over a very wide range of NaCl concentrations, potentially eliminating the problem of providing freshwater for large-scale growth. The Bryant lab is presently pursuing two different approaches to biosolar H2 production: one based on biological production (as part of a Multi-University Research Initiative sponsored by the Air Force Office of Scientific Research) and one based on direct coupling of an [Fe-Fe]-hydrogenase to Photosystem I (in collaboration with Dr. John H. Golbeck and supported by DOE). The MURI project seeks to identify robust cyanobacterial H2 producing cyanobacteria and to use systems biology approaches to modify and if possible improve these organisms for enhanced hydrogen production (see: http://www.princeton.edu/~catalase/). Current research has focused on the genetically tractable and rapidly growing cyanobacterium Synechococcus sp. PCC 7002 to study the regulation and manipulation of H2 production. We are additionally attempting to develop Arthrospira maxima strain CS-328 as a system for hydrogen production and are currently sequencing the genome of this organism as well as an interesting iron-tolerant cyanobacterium, strain JSC-1, as part of the MURI project.

The marine, unicellular cyanobacterium Synechococcus sp. PCC 7002 synthesizes a bidirectional [NiFe] hydrogenase (H2ase) encoded by five hox genes (hoxE, hoxF, hoxH, hoxU, hoxY). Maturation of this H2ase requires seven hyp/hox gene products (HypA, HypB, HypC, HypD, HypE, HypF and HoxW). We have studied the expression of these genes by reverse-transcription PCR (RT-PCR) and quantitative-RT-Real-time PCR (q-RT-PCR), through which we have begun to understand the regulatory mechanisms that will lead to optimized H2ase activity and H2 production in Synechococcus sp. PCC 7002 and other cyanobacteria (e.g., Arthrospira maxima).

In Synechococcus sp. PCC 7002, all hox and hyp genes cluster in a 13-kb region. RT-PCR was used to amplify specifically all intergenic regions. Although we cannot exclude that secondary transcription initiation sites occur, the data show that the 13 hox and hyp genes probably form an operon that is transcribed as hypE- hoxEFUY-hyp3-hoxHW-hypABFCD, (Fig. 1).

Fig. 1 (A) Gene organization of the hox-hyp operon in the genome of Synechoco-ccus sp. PCC 7002. (B) RT-PCR showing cotranscription of genes in the hox-hyp operon. For each intergenic region, primers were designed inside the coding sequence of the two neighboring genes.

Fig. 2a (left). Comparison of the relative changes of gene products with and without DBMIB treatment. The transcription levels of hox/hyp without DBMIB treatment are arbitrarily set to one and are thus not shown. Internal control data are also not shown. Fig. 2b (right). Comparison of the relative abundance of hox and hyp genes in WT without and after DBMIB treatment. The transcription level of hypB in WT is arbitrarily set to one. Blue bars: WT without DBMIB; red bars: WT after DBMIB treatment.

The transcription of hox and hyp genes is significantly enhanced when a WT culture is treated by 2 µM DBMIB. q-RT-PCR data confirm the constitutive expression of hox/hyp genes, and the elevation of transcript abundance of hox and hyp genes by DBMIB. hypA, hypB, hypF and hyp3 change more significantly than others, whereas changes of the hox genes are roughly in the same scale (Fig. 2a). The hox gene transcripts are more abundant than those for the hyp genes (Fig. 2b).

Fig. 3. The restriction map of ExKmPAQ1Sp, in which aadA (streptomycin/spectinomycin resistance) and expression cassette are in opposite transcriptional directions. Another version, ExKmPAQGm, has the aacC1 (gentamicin resistance) downstream of the expression cassette and in the same transcriptional direction.

We have engineered a novel, high-copy-number vector that can replicate in E. coli and be used to introduce gene(s) into a specific region of the high-copy number plasmid, pAQ1, of Synechococcus sp. PCC 7002 (Fig. 3). Several trials to successfully express/overexpress genes have revealed a wide range of applications for this vector (Fig. 4, 5.). To date, the following proteins have been expressed: NifU from Synechocystis sp. PCC 6803; HydA from Clostridium acetobutylicum; HydA from Bacteroides thetaiotaomicron; CrtG from Synechococcus elongatus sp. PCC 7942; and GlbN from Synechococcus sp. PCC 7002).

Fig. 4. (a) Immunoblot showing overexpression of NifU from ExKmPAQ1; (b) Immunoblot showing the heterologous expression of C. acetobutylicum hydA from ExKmPAQ1 (using the aphII promoter, P-km) and ExCpcPAQ1 (using the cpcBA promoter of Synechocystis sp. PCC 6803, P-cpcBA); c) Overexpression of B. thetaiotaomicron hydA by ExKmPAQ1.

Fig. 5. Left: Expression of holo-GlbN from ExKmPAQ1 in E. coli and Synechococcus sp. PCC 7002. (a) GlbN from E. coli, Coomassie-blue stained; (b) GlbN from E. coli, heme-stained, showing that heme is covalently bond with GlbN; (c) GlbN from Synechococcus sp. PCC 7002, Coomassie-blue stained; (d) GlbN from Synechococcus sp. PCC7002, heme-stained. GlbN may have isoforms indicated by arrows. Right: Absorption spectrum of purified GlbN from E. coli, showing the presence of heme in holo-GlbN.

H2 production is also found in cyanobacterial fermentation (Fig. 6 left). Two key enzymes (lactate dehydrogenase and pyrovate: ferredoxin oxidoreductase, LDH and nifJ, respectively) have been knocked out by interposon mutagenesis. △nifJ grows faster than WT in circadian dark/light growth, which is more obvious when glycerol is added to both △nifJ and WT. However △LDH seems not to be affected at all (Fig. 6). The effects and alternative metabolic pathways caused by these mutations are still under investigation.

Back to top

Cyanobacteria, because of their ability to photo-oxidize water, their nutritional simplicity, and their amenability for genetic manipulation, offer great potential as cellular factories for biosolar hydrogen production or alternatively for carbon-based, biomass and biofuels energy applications. The genus Arthrospira (Figure 1) comprises filamentous, non-heterocystous cyanobacteria that are generally found in tropical and subtropical regions in warm bodies of water with high carbonate/bicarbonate content, elevated pH, and salinity. Their large, gas-vacuolate filaments (3 to 12 µm in diameter) are easily collected by filtration and other physical means of separation, and blooms of these cyanobacteria have long been used as a food source by aboriginal peoples around the world. For example, the Aztecs were exploiting this organism from Lake Texcoco in Mexico before the arrival of Europeans. More recently, the phycologist Dangeard reported in 1940 that the Kanembu people in Chad consumed dihé, cakes of sun-dried cyanobacteria collected from the shores of small ponds around Lake Chad. While participating in a Belgian Trans-Saharan expedition 25 years later, the botanist Jean Leonard discovered a bloom of cyanobacteria covering the waters around the shores of Lake Chad. He realized the connection between the algal blooms and the dried dihé cakes sold in the markets. Arthrospira sp. additionally serve as the principal food source for millions of filter-feeding flamingoes in the soda lakes of the Rift Valley in central and eastern Africa. Carotenoids (principally echinenone, canthaxanthin, and astaxanthin) derived from these cyanobacteria are responsible for the characteristic pink coloration of the feathers of these birds, which are a principal source of food for migratory and endemic raptors from Europe, Asia, and Africa. Arthrospira spp. are rich in amino acids, vitamins, and g-linolenic acid, and these bacteria are cultivated on an industrial scale (see www.relfe.com/spirulina_health_benefits%20_4.htm) as a source of various specialty chemicals (e.g., b-carotene) and as a health food supplement under the name “Spirulina.” 16S rRNA analyses indicate that the genera Arthrospira and Spirulina are not closely related and that that the genus Arthrospira is a monophyletic group.

Arthrospira maxima strain CS-328 possesses several attributes that are extremely valuable in choosing a microbial system for biosolar H2 production. Arthrospira spp. are arguably among the most robust oxygenic phototrophs known. These filamentous cyanobacteria thrive under extreme environmental conditions in alkaline soda lakes (pH 9.5 to 11) at concentrations of carbonate that approach saturation (0.4 to 1.2 M), thereby facilitating selective growth in mass cultures (Figure 2). The native growth conditions for Arthrospira sp. are intolerable for most other microbes, although Halomonas and some Bacillus spp. commonly occur in similar environments and are often contaminants of non-axenic Arthrospira sp. cultures. As noted above, Arthrospira spp. are already grown on very large scale in commercial culturing facilities in open (non-sterile) aqua-cultures exposed to environmental conditions without major problems from contamination for this reason. Global annual production is currently estimated at over 3000 tons. Growth rates are fast and cell densities are high when compared to other cyanobacteria and microalgae. A. maxima strain CS-328 grows optimally at near saturating carbonate concentrations and at pH 9.5 to 11. It grows at twice the rate of A. platensis (UTEX strain #1926) and to a cell density that is about 2- to 2.5-fold higher. These properties also make A. maxima strain CS-328 potentially useful for the production of biomass for other energy-related applications.

Fluorescence measurements show that A. maxima CS-328 possesses the fastest and most efficient O2-evolving enzyme of all oxygenic phototrophs they have examined to date, including many cyanobacteria and several green algae. This attribute is absolutely essential for achieving high efficiency in extracting electrons and protons from water for use in H2 production. Under steady-state conditions Photosystem II (PSII) turnover rates in vivo approach 103 s-1 and continue at that rate, limited only by the available size and reoxidation rate of the PQ pool. The Kok parameters of the oxygen-evolving complex indicate the fewest photochemical misses by a factor of 5 at the highest turnover rates when compared to all other species of green algae and higher plants examined thus far. A. maxima CS-328 also contains a 3-5-fold larger plastoquinone pool than green algae (including C. reinhardtii), and thus has an intrinsically higher capacity for storage of electrons. This confers greater efficiency in electron transport under variable solar intensity. Hydrogen production occurs during periods of fermentative metabolism under anoxic conditions in the dark. Physiological studies, which have been conducted in the laboratory of collaborator Chuck Dismukes, indicate that appropriate management of the growth conditions can lead to significant enhancements of hydrogen production.

As a first step towards the long-term goal of manipulating Arthrospira maxima genetically, we are determining the genome sequence of the Arthrospira maxima CS-328 in collaboration with JGI-DOE. The information obtained will be used to facilitate “-omics” studies (i.e., metabolomics, proteomics, and transcriptomics) which will enable a systems biology approach to modifying and engineering an organism that shows excellent potential for fermentative hydrogen production and that could become an important means for producing carbon-based biomass and biofuels.

The genome sequence of a second cyanobacterium, which is a presently unnamed filamentous, non-heterocystous member of the family Oscillatoriaceae, is also being sequenced. Strain JSC-1 (Figure 2), was isolated by Dr. Igor Brown from the iron-depositing Chocolate Pots Hot Spring in Yellowstone National Park. Although the freshwater cyanobacterium Synechocystis sp. PCC 6803 grows optimally in the presence of 40 µM Fe+3, iron-tolerant cyanobacteria from these hot springs require Fe+3 concentrations up to 400 µM for maximal growth. One reason for the elevated requirement for Fe could be the high ratios of PSI to PSII exhibited by these strains. Some studies have reported that Fe+2 concentrations up to 1 mM significantly stimulated light- and dark-dependent bicarbonate uptake by cyanobacteria isolated from Chocolate Pots Hot Springs, and the authors of one such report suggested a specific role for elevated iron in photosynthesis by iron-tolerant cyanobacteria. Strain JSC-1 exhibits an interesting colonial morphology with an aerial, phototactic growth mode (Figure 3). When the organism is grown on plates, it resembles the fur of an animal or velvet as a result of the filaments that project upwards. This aerial growth would facilitate gas exchange and light penetration into the colonies of this organism. The colonial, filamentous growth mode of this cyanobacterium strongly suggests that, like other Section III organisms, it will perform a specialized form of apoptosis: the formation of necridia. Necridia are sacrificial cells that are selectively killed to allow fragmentation of filaments for dispersal of a cyanobacterium in its environment. Very little is known about this unusual process and its regulation. It is anticipated that the genome sequence of this unusual organism will significantly expand our knowledge of cyanobacterial diversity through comparative bioinformatics analyses of the genomic sequence information.

Fig. 2 Strain JSC-1 (courtesy of Dr. Igor Brown)

Figure 3. Colony morphology and phototactic behavior of JSC-1(image courtesy of Dr. Igor Brown)

Back to top

An economy based on dihydrogen (H2) depends on utilizing 143 kJ g-1 of energy to split the water molecule into dihydrogen and oxygen. The source of energy can be the combustion of fuels such as natural gas, oil, or coal or it can be supplied by nuclear, tidal, wind or solar energy. The latter has advantages over the others, as it is widely distributed, plentiful, and inexhaustible. Over the course of 3.5 ´ 109 years, living cells have evolved two types of light-dependent enzymes that carry out the conversion of solar energy into chemical bond energy. In cells, these enzymes photosystem II (PS II) and photosystem I (PS I), operate in series producing nicotinamide adenine nucleotide diphosphate (NADPH). Although it is not useful as a source of stored energy for human activity, NADPH is an indispensable molecule. Thus, were it possible to re-engineer PS I to generate H2 instead of NADPH, an energy source suitable for use in fuel cells would immediately become available. The objective of our work is to couple PSI directly with a hydrogenase, which can efficiently catalyze H2 evolution at high rates.

Goal:

To engineer both photosystem I and C. acetobutylicum [Fe-Fe]-H2ase  (HydA) in order to link them and as a result to have hydrogen evolution.

Approach:

1. To replace a Cys ligand of PsaC with Gly
2. To link an organic tether to PS I
3. To engineer the H2ase Cys97 ligand of distal [4Fe4S] cluster to Gly, Ser, Arg, Ala and Asp
4. To tether HydA directly to PS I

The pioneering work of Toney and Kirsch (1989) paved the way for chemical rescue of site-modified amino acids, restoring enzymatic activity using externally supplied organic molecules. Recently, research in this laboratory has shown that a missing cysteinyl ligand to a [4Fe-4S] cluster can be successful rescued by an alkyl thiolate (Antonkine et al., 2007). Thus, these observations led us to replace a similar cysteine ligand (residue 97) in HydA by different amino acids. If it is possible to rescue the distal [4Fe-4S] cluster with an external ligand, a mechanism to insure a rapid transfer electrons from PS I to HydA can likely be found. The change of the ligand to the distal  [4Fe-4S] cluster can be monitored using EPR spectroscopy. Thus, we can envisage three scenarios based in the amino acid characteristics. S and D have a hydroxyl side group that could potentially provide and oxygen  ligand to an Fe atom, whereas G, A and R do not:  i) an empty site would  result from the inability of a [4Fe-4S]^1+,2+ cluster to assemble in the modified site; ii) the missing Cys ligand at the modified site would lead to a presence of [3Fe-4S]^0,1+ cluster; or iii) a [4Fe-4S] cluster with altered redox and/or spectroscopic properties would assemble in the modified site.

Fig. 1 Amino acid sequence of C. acetobutylicum (CAI) HydA and C. pasteurianum (CPI) HydA. These sequences show a 70% identity.

Fig. 2 Model depicting the chimera between the HydA from Clostridium pasteurianum and PS I from cyanobacteria, respectively, PDB files 1feh and 1jb0. The  [Fe-S] clusters are shown in green, and PsaC, which harbors the FA and FB. [4Fe-4S] clusters is shown in orange. The white bar linking the two proteins represents a viologen dithiol tether used to rescue one [Fe-S] in each protein.

Fig. 3 The calculations where made accordingly to King et al (2006). * This mutant has two tags: a his-tag at amino-terminus and a strep-tag at the carboxyl terminus.

Fig. 4 12% SDS-PAGE left panel, Immunoblot analysis right panel; elution represents strep-tag HydA.

Work in progress:

• double selection of HydA with two different affinity resins
• correction factor for protein content
• reconstitution of all constructs and rescue of the missing cysteinyl ligand
• EPR spectroscopy

References:

Antonkine, M. L.; Maes, E. M.; Czernuszewicz, R. S.; Breitenstein, C.; Bill, E.; Falzone, C. J.; Balasubramanian, R.; Lubner, C.; Bryant, D. A.; Golbeck, J. H. 2007. Chemical rescue of a site-modified ligand to a [4Fe-4S] cluster in PsaC, a bacterial-like dicluster ferredoxin bound to Photosystem I. Biochim. Biophys. Acta 1767: 712-724.

King, P. W.; Posewitz, M. C.; Ghirardi, M. L.; Seibert, M. 2006. Functional studies of [FeFe] hydrogenase maturation in an Escherichia coli biosynthetic system. J. Bacteriol. 188: 2163-2172.

Back to top

Support from