nitrogen fixation and hydrogen metabolism in cyanobacteria · nitrogen fixation and hydrogen...

MICROBIOLOGY AND MOLECULAR BIOLOGY REVIEWS, Dec. 2010, p. 529–551 Vol. 74, No. 41092-2172/10/$12.00 doi:10.1128/MMBR.00033-10Copyright © 2010, American Society for Microbiology. All Rights Reserved.

Nitrogen Fixation and Hydrogen Metabolism in CyanobacteriaHermann Bothe,1* Oliver Schmitz,2 M. Geoffrey Yates,3 and William E. Newton4

Botanical Institute, The University of Cologne, D-50923 Cologne, Germany1; Metanomics GmbH, Tegeler Weg 33, 10589 Berlin,Germany2; Fir Trees, Kingston Ridge, Kingston, Lewes, Sussex BN7 3JU, England3; and Department of Biochemistry,

Virginia Polytechnic Institute & State University, Blacksburg, Virginia 240614

INTRODUCTION .......................................................................................................................................................529MOLYBDENUM NITROGENASE...........................................................................................................................529ALTERNATIVE NITROGENASES ..........................................................................................................................532NITROGENASES IN CYANOBACTERIA ..............................................................................................................533

Occurrence of Nitrogenases in Heterocysts ........................................................................................................533Electron Transport to Nitrogenase in Cyanobacteria .......................................................................................534Alternative Nitrogenases in Cyanobacteria.........................................................................................................535Nitrogen Fixation in Nonheterocystous Cyanobacteria.....................................................................................536

HYDROGENASES IN GENERAL ...........................................................................................................................536HYDROGENASES IN CYANOBACTERIA.............................................................................................................537

Hydrogenase Types in Cyanobacteria ..................................................................................................................537Uptake hydrogenase ...........................................................................................................................................537Bidirectional hydrogenase .................................................................................................................................538

POTENTIAL FOR EXPLOITING CYANOBACTERIA IN SOLAR ENERGY CONVERSIONPROGRAMS FOR PRODUCTION OF COMBUSTIBLE ENERGY (HYDROGEN)..............................543

ACKNOWLEDGMENTS ...........................................................................................................................................546REFERENCES ............................................................................................................................................................546

INTRODUCTION

Biological (di)nitrogen fixation is catalyzed by the enzymecomplex nitrogenase, where the formation of molecular hydro-gen accompanies ammonia production according to equa-tion 1:

8H� � 8e� � N2 � 16MgATP 3 2NH3 � H2

� 16MgADP � 16Pi (1)

Whereas H2 formation by nitrogenases is unidirectional, H2

production by some hydrogenases is reversible, as shown inequation 2:

2H� � 2e�7 H2 (2)

N2 fixation and H2 formation are closely linked processes, ashas been known at least since a publication by Phelps andWilson in 1941 (39). Hydrogenase recycles the H2 produced inN2 fixation, thereby minimizing the loss of energy during nitro-genase catalysis. A rather simple scheme showing the relationshipbetween pyruvate degradation, N2 fixation, and production anduptake of H2, as occur in strict anaerobes such as Clostridiumpasteurianum or in the facultative anaerobe Klebsiella pneu-moniae, is shown in Fig. 1. However, H2 can also be producedindependently of N2 fixation, e.g., as an end product of fermen-tation, which can also take place in N2-fixing organisms.

As described in detail below, nitrogenases (Mo, V, and ho-mocitrate) and hydrogenases (Ni, CO, and CN�) contain un-

usual components in their prosthetic groups (Fig. 2 and 3) thatare not or only rarely employed elsewhere in nature. Theirroles and their biosyntheses pose fascinating questions that areas yet only partly resolved. Most cyanobacteria are aerobicorganisms producing O2 photosynthetically. They are generallynot exposed to environmental molecular H2. Despite this, andparadoxical at first glance, the capability to metabolize H2 isconstitutively expressed in many aerobic cyanobacteria. N2 fix-ation and H2 metabolism have been key research areas inmicrobiology over the years. Cyanobacteria are the best suitedorganisms for studies on the subject, because several of them,both unicellular and heterocystous forms, can be easily genet-ically modified by molecular techniques. Moreover, cyanobac-terial H2 production offers perspectives for potential applica-tions.

Both N2 fixation (153, 177) and H2 metabolism (226, 228)have been reviewed. Excellent accounts on cyanobacterial hy-drogenases (82, 212, 214) are available, and those articlesshould be consulted for primary references. The aim of thisreview is not to reiterate these subjects but to highlight factsand ideas, particularly on the physiology, that have not re-ceived much attention in the past. This review also emphasizesthe more recent developments and focuses on the fact thatnitrogenases and hydrogenases are common players in H2 me-tabolism. The restriction to cyanobacteria as the best candi-dates for applications appears to be timely.

MOLYBDENUM NITROGENASE

The longest-known and best-studied nitrogenase is the Monitrogenase, which occurs in all N2-fixing organisms with theexception of some CO-oxidizing bacteria (178). The Mo nitro-genase is encoded by the structural genes nifHDK. It consists of

* Corresponding author. Mailing address: Botanical Institute, TheUniversity of Cologne, Zulpicher Str. 47b, D-50923 Cologne, Ger-many. Phone: 49 221 470 2760. Fax: 49 221 470 5039. E-mail: [email protected].

529

on June 16, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

https://crossmark.crossref.org/dialog/?doi=10.1128/MMBR.00033-10&domain=pdf&date_stamp=2010-12-01

http://mmbr.asm.org/

two component proteins. Figure 2 shows the structure of a 2:1complex of the two components, which might approximate anelectron transfer transition state, with the larger component inthe center and one molecule of the smaller component at eachend (see the legend to Fig. 2 for more information). The nifHgene codes for the smaller, homodimeric (�2) protein, whichhas a molecular mass of about 64 kDa and is termed Feprotein, (di)nitrogenase reductase, or protein 2. Its prostheticgroup is a [4Fe-4S] cluster that bridges the subunit interfaceand is ligated by two cysteinyl residues from each subunit. Thiscluster accepts reducing equivalents from electron carriers

which are either ferredoxin or flavodoxin, depending on theorganism. Each subunit possesses a MgATP/MgADP bindingsite. When provided with MgATP and reductant, the Fe pro-tein undergoes a conformation change combined with a changeof its redox potential of ca. �200 mV. Docking to the largercomponent protein (Fig. 2) lowers the redox potential furtherto about �600 mV and is accompanied by an additional con-formation change. All these changes are prerequisites for thetransfer of one electron from the Fe protein to the largercomponent protein with concurrent MgATP hydrolysis. Mul-tiple electron transfers prepare the larger component for sub-

FIG. 1. A simple scheme showing the relationship between pyruvate degradation, ammonium and hydrogen formation by nitrogenase, andhydrogen uptake by hydrogenase. This pathway is typical in strict or facultative anaerobes but also proceeds in cyanobacteria.

FIG. 2. The structure of the 2:1 Fe protein-MoFe protein complex of the Azotobacter vinelandii nitrogenase stabilized by MgADP plus AlF4�.

Each Fe protein molecule (shown at the top left and bottom right of the complex in brown) docks directly over the interface between an �/� subunitpair of the MoFe protein (in black and gray), which occupies the center of the structure, to juxtapose its [4Fe-4S] cluster (in yellow) with a P cluster(in red) at this interface. One FeMo cofactor (in pale blue) is accommodated within each � subunit. The two � subunits (in gray) provide theinteractions among the two �/� subunit pairs (183) (Protein Data Bank [PDB] code 1N2C). (Adapted from reference 183 with permission fromMacmillan Publishers Ltd.)

530 BOTHE ET AL. MICROBIOL. MOL. BIOL. REV.


mbr.asm

.org/D

ownloaded from


strate binding and reduction. The Fe protein has the mostconserved amino acid sequence among all nitrogenase pro-teins. Therefore, the nifH gene is best suited for DNA probingwhen searches for the occurrence of nitrogenase in organismsor different environments are undertaken (181).

The larger component protein (MoFe protein, dinitroge-nase, or protein 1) is a tetrameric (�2�2) protein of about 240kDa. It contains two unique prosthetic groups, the P clusterand the MoFe cofactor (Fig. 3). Each �� dimer of the largernitrogenase protein binds one FeMo cofactor and one P clus-ter. The P cluster is composed of both a [4Fe-4S] subclusterand a [4Fe-3S] subcluster, which share one S2�. It sits at theinterface of the � and � subunits and is usually depicted as anintermediate in electron transfer from the Fe protein to theFeMo cofactor. However, there is no direct evidence to sup-port this supposition. The P cluster may have an N2 fixation-specific role through which it provides the impetus to commit

the reversibly bound N2 to the irreversible reduction pathway(70). The MoFe cofactor consists of 1 Mo atom, 7 Fe atoms, 9S atoms, and homocitrate, plus an as-yet-unidentified lightatom (or ion) at its center (Fig. 3). Although an educated firstguess might be that it is N based, this suggestion remainsunproven (see, for example, reference 234). The FeMo cofac-tor is the site of substrate binding and reduction. This clustercan again be subdivided into two subclusters, one [Mo-3Fe-3S]and one [4Fe-3S]. These are bridged by 3 S2� ligands and thelight atom. Homocitrate, which is bound to Mo by two Oligands, is required for full catalytic activity, but its specific roleremains unclear.

The substrate binding and reduction sites have not yet beenidentified definitively. The N2 molecule may be bound at acentral 4Fe-4S face, possibly with participation of the lightatom. The Mo-homocitrate entity would then not be directlyinvolved in catalysis but could determine the redox potential ofthe cofactor. Alternatively, N2 may be bound directly to and bereduced at the Mo-homocitrate part of the FeMo cofactor. Itis somewhat surprising that this issue has not yet been resolveddespite extensive research for many years. However, neitherthe FeMo cofactor nor any other part of the nitrogenase com-plex binds a substrate on its own. Substrate binding and reduc-tion commence only when both nitrogenase component pro-teins plus MgATP and reductant are available.

Nitrogenase catalyzes the reduction of many substratesother than N2, nearly all of which have a complete or partialtriple bond in common, e.g., HC'N (hydrocyanic acid),R,C'N (nitriles), RN'C (isonitriles), N2O (nitrous oxide),N'NON� (azide), and HC'CH (acetylene); the main ex-ceptions are H� and NO2

�. Of particular interest is the re-duction of C2H2 to C2H4. In contrast to carbon fixation re-search, where an easily manageable isotope (14C) is available,N2 fixation research suffers from the absence of a similar iso-tope of N. 13N is highly radioactive and very unstable, andbecause 15N is nonradioactive, its reduction can be determinedonly by the somewhat laborious technique of mass spectrom-etry. In contrast, the gases C2H2 and C2H4 can be easily andquickly separated and quantified with high accuracy by gaschromatography. Unless special questions (e.g., the determi-nation of the ratio between C2H2 and N2 reduction) are to beresolved, nitrogenase activity is routinely assayed by the C2H2

reduction method despite the fact that the ratio between N2

fixation and C2H2 reduction is not always 3:1. The reduction ofall nitrogenase substrates is inhibited by CO, with the excep-tion of H� conversion to H2 (see below).

The reduction of N2 but not that of all other nitrogenasesubstrates is accompanied by the evolution of one H2 moleculefor each N2 molecule that is reduced (203) (see equation 1).This formation of H2 could represent an activation step that isuniquely required for N2 binding (196). In the absence of anyother substrate, nitrogenase catalyzes an ATP-dependent re-duction of H�. The relationship(s) between the binding of N2,the other substrates, and inhibitors such as CO is apparentlyvery complex and at best only partly understood. The complex-ity of the situation is evidenced by the fact that N2 is a com-petitive inhibitor of C2H2 reduction but C2H2 is a noncompet-itive inhibitor of N2 reduction (179).

In addition to the three structural genes nifHDK, nitroge-nase expression requires altogether 20 genes in the enterobac-

FIG. 3. The structure of the FeMo cofactor of the Azotobactervinelandii nitrogenase MoFe protein with its � subunit-based ligatingamino acid residues (�Cys-275 and �His-442) and homocitrate. TheMo (red), Fe (gray), and S (pale green) atoms are individually colored.The identity of the central atom (blue) remains unassigned (PDB code1M1N). (Reprinted from reference 61 with permission from AAAS.)

VOL. 74, 2010 CYANOBACTERIAL NITROGENASES/HYDROGENASES 531


mbr.asm

.org/D

ownloaded from


terium Klebsiella pneumoniae, all of which are contiguouslylocated on the chromosome. In other bacteria, these genes areinterspersed throughout the genome, and other fix genes maybe necessary for nitrogenase synthesis and catalysis.

ALTERNATIVE NITROGENASES

Mo nitrogenase is now known to have two close relatives, theV nitrogenase and the Fe nitrogenase, but the distribution ofthese two enzymes appears to be haphazard (see below). Thediscovery of the alternative nitrogenases without molybdenumin their prosthetic groups can be regarded as a milestone innitrogenase research. Reviews on this subject are available (18,59, 167, 242). The aerobe Azotobacter vinelandii possesses genesets for all three different types of nitrogenases (Fig. 4). Underconditions of Mo sufficiency in the culture medium, A. vine-landii expresses nifHDK, encoding Mo nitrogenase. When Mois limiting but V is sufficiently available, A. vinelandii synthe-sizes a V nitrogenase with a VFe cofactor in the N2 bindingand reducing site through expression of the alternative struc-tural genes vnfHDGK. The occurrence of V in the prostheticgroup of an enzyme complex is remarkable because, other thanin V nitrogenase, the element V has only rarely been found tohave a biological function, e.g., in some uncommon peroxi-dases (95). When the concentrations of both Mo and V aregrowth limiting, A. vinelandii synthesizes a third nitrogenasewith an FeFe cofactor in the active site and encoded by thestructural genes anfHDGK.

All three nitrogenases are rather similar. They require botha larger and a smaller component protein for catalytic activityand possess the P cluster, with identical spectroscopic proper-ties, and a special cofactor for the substrate binding and re-ducing site. All three nitrogenases show extensive but not iden-tical amino acid sequence homologies. Most importantly, bothalternative nitrogenases possess the additional G gene locatedbetween the D and K genes and the resulting component pro-teins are, therefore, �2�2�2 heterohexamers. The � subunit hasno counterpart with similar sequence homologies elsewhere.

Its function has not been finally resolved, but it is apparentlyrequired for processing the apoprotein of the alternative ni-trogenases to the functional enzyme complex by assisting in theinsertion of the cofactor, as has been specifically shown for theV nitrogenase (45, 46). Remarkably, although the proteinsVnfG and AnfG are required for N2 fixation by A. vinelandii,they are not required for C2H2 reduction (45, 46, 228).

Both alternative nitrogenases can support growth of A. vine-landii, albeit with lower rates than Mo nitrogenase. Both N2

and C2H2 are poorer substrates for the alternative nitroge-nases than for the Mo enzyme. Whereas with Mo nitrogenasethe stoichiometry between ammonia production and H2 for-mation is about 2:1, as shown in equation 1, the reaction via theV nitrogenase proceeds optimally as shown in equation 3:

12H� � 12e� � N2 � 24MgATP 3 2NH3 � 3H2

� 24MgADP � 24Pi (3)

With Mo nitrogenase, virtually all electrons are allocated toC2H2 when it is the only substrate available. In contrast, C2H4

formation by V nitrogenase is accompanied by a significantproduction of H2. This H2 formation in the presence of eitherN2 or C2H2 seems to be even higher with the Fe nitrogenase,although these reactions have not been examined in compara-ble detail. These differences between the three nitrogenasesare not due to differences in the apparent Km values for N2 andC2H2 and are also not caused by restricted electron transferwithin or between the nitrogenase proteins (59). The differ-ences may lay in the rate-limiting step in the nitrogenase cat-alytic cycle (220), which is the final dissociation of the oxidizedFe protein-MgADP from the electron transfer complex.

The production of NH3 from N2 by the V nitrogenase isaccompanied by the release of the presumptive reduction in-termediate N2H4 (57). In addition, both the V and Fe nitro-genases reduce C2H2 beyond C2H4 to produce some C2H6.Although this ethane formation amounts to only about 3% ofthe total C2H2-reducing capacity, it can easily be assessed bygas chromatography and is therefore indicative for the expres-

FIG. 4. Genes coding for nitrogenases in two cyanobacteria and two other microorganisms. (Courtesy of Teresa Thiel, University of Missouri—St. Louis.)



mbr.asm

.org/D

ownloaded from


sion of an alternative nitrogenase in an organism (56). Monitrogenase does not catalyze the reduction of ethene, butsome H2-consuming methanogenic enrichment cultures havebeen reported to produce ethane from ethene apparently in-dependently of nitrogenases (118).

The apparently haphazard distribution of nitrogenases re-sults in some organisms having all three, some possessing onlyMo nitrogenase, and others having the Mo and V but not theFe nitrogenase or the Mo and Fe nitrogenases without the Vnitrogenase. Azotobacter vinelandii (19), Azotobacter paspali(129), Rhodopseudomonas palustris (155), and the archaeonMethanosarcina acetivorans (76) are the only organisms so faridentified that possess gene sets for all three nitrogenases. Thecombination of a Mo and a V nitrogenase is found in Azoto-bacter chroococcum, Azotobacter salinestris, and the archaeonMethanosarcina barkeri 27 (129) and in several cyanobacteria(see below). The Mo and Fe nitrogenases but not the V en-zyme occur in Clostridium pasteurianum, Azomonas macrocy-togeneses, and Azospirillum brasilense Cd (44) and in the pho-totrophs Rhodospirillum rubrum, Rhodobacter capsulatus, andHeliobacterium gestii (17).

Probes have been developed from vnfG and anfG to specif-ically amplify gene segments by PCR and to detect the alter-native nitrogenases in organisms. By this technique, Lovelesset al. (130) were able to isolate seven diazotrophs from aquaticenvironments that possess an alternative nitrogenase(s) andbelong to the fluorescent pseudomonads and azotobacteria ofthe gammaproteobacteria. Recently, 24 bacteria of the samegroup, one closely related to Enterobacter and another withsequences almost identical to those of Paenibacillus, were iso-lated from diverse habitats, all with an alternative nitroge-nase(s) (17). Besides in pseudomonads and azotobacteria, al-ternative nitrogenases occur only occasionally and inprokaryotes of totally unrelated taxonomic affinities. Therather close sequence similarities of the nitrogenase genessuggest that they may have arisen by gene duplication in theazotobacter-fluorescent pseudomonad group (17). In other or-ganisms, however, there is little correlation between vnfG andanfG sequences on the one hand and the phylogeny inferredfrom the 16S rRNA gene sequence data on the other. Thiscould mean that alternative nitrogenase genes may have beeninterspersed by lateral gene transfer among nonmembers ofthe azotobacter-pseudomonad group.

An indicator of this situation seems to occur in Methanosar-cina barkeri 227 (47). This archaeon possesses a D gene and aG gene with close sequence homologies to vnfDG from otherorganisms, particularly Anabaena variabilis. The vnfH gene isseparated from the vnfDGK cluster by two open readingframes (ORFs). Phylogenetic analysis indicates that this Hgene is a member of a separate cluster comprising anfH genesof several bacteria and is closely related to anfH fromRhodobacter capsulatus and Clostridium pasteurianum. Thiscluster might also include vnfH from A. vinelandii. In anothermethanogen, Methanococcus maripaludis, with only a singlenitrogenase, nifD and nifK cluster with the other genes for theMo nitrogenase, whereas the H gene is an amalgam of both Moand V nitrogenase H genes (113). Thus, vnfH and vnfDGK mayhave been acquired from other organisms by two independentgene transfers.

Such processes are difficult to understand because there is

no apparent selective pressure to acquire and maintain alter-native nitrogenases. Conditions in nature where Mo is growthlimiting in soils or aqueous habitats are unknown, and micro-organisms have high-affinity transport systems that effectivelymobilize Mo from habitats (168). These mobilizations mayresult in microzones of Mo depletion around microorganismswhere bacteria that can express an alternative nitrogenase(s)have a selective advantage (142). However, the isolation ofdiazotrophs with alternative nitrogenases from habitats withsufficient Mo concentrations (17) may indicate that these en-zymes could have other, but so far totally unresolved, functionsin nature. Otherwise, why would these genes, if redundant, beretained in organisms during evolution?

NITROGENASES IN CYANOBACTERIA

Occurrence of Nitrogenases in Heterocysts

Cell-free preparations of nitrogenases from all organismsare irreversibly damaged by O2, and different groups of micro-organisms have been versatile in developing various means toprotect their nitrogenases against the O2 of the air (153). Incyanobacteria, the O2 problem is enhanced by the photosyn-thetic production of this gas. Many filamentous cyanobacteriasolve the issue by cell differentiation. Under aerobic growthconditions, their vegetative cells perform photosynthetic O2

evolution and CO2 fixation, whereas nitrogenase resides inspecialized cells, the heterocysts (66). These differentiate fromvegetative cells by cell division and extensive metabolicchanges (133, 162). Photosystem II (PSII) is largely degradedin heterocysts so that they cannot perform the photosyntheticwater-splitting reaction. They are also unable to fix CO2 pho-tosynthetically. Vegetative cells provide photosyntheticallyfixed carbon, which may be exported as sucrose to the hetero-cysts (52). In turn, heterocysts provide nitrogen, likely as glu-tamine formed via ammonia generated by N2 fixation and bothglutamine synthetase and glutamate synthase (219). Alterna-tively, glutamine may be converted to arginine which is thenincorporated into the cyanophycin granule. This may be de-graded by cyanophycinase in a dynamic way depending on theN demand of heterocysts and vegetative cells (86).

Heterocysts possess a thick cell envelope composed of long-chain, densely packed glycolipids providing a barrier to gasexchange (9). The main diffusion pathway for O2 and N2 mightbe through the terminal pores (“microplasmodesmata”) (83)that connect heterocysts with vegetative cells. Walsby (230)suggested that transmembrane proteins make the narrow porespermeable enough and might provide a means of regulatinggas exchange. Residual O2 reaching the inside of the hetero-cysts might be immediately consumed by their high respiratoryactivity and also other reactions in these cells. In this way,heterocysts provide an anaerobic environment which allowsnitrogenase to function.

The occurrence of nonspecific intercellular channels be-tween heterocysts and vegetative cells has recently been con-firmed (149). Any analogy to the plasmodesmata of higherplants is misleading, however, because cyanobacteria do notpossess an endoplasmic reticulum. However, the export ofmetabolites might follow the source-sink gradient along theintercellular channels of both plants and cyanobacteria. Alter-



mbr.asm

.org/D

ownloaded from


natively, the periplasmic space between the peptidoglycanlayer and the outer membrane could constitute a communica-tion conduit for the transfer of compounds, since this space iscontinuous between heterocysts and vegetative cells (72).

Heterocyst formation from vegetative cells of Anabaena spe-cies takes about 24 h after the cells have suffered N depriva-tion. More than 500 proteins are differentially expressed inheterocysts during cellular transformation from vegetative cells(162), showing that this complex process is under the control ofmany genes. Master regulators are HetR, a serine-type pro-tease, and NtcA, a nitrogen control transcription factor incyanobacteria (115, 152, 160, 200). Expression of hetR is up-regulated by nitrogen deprivation, and this upregulation de-pends on NtcA (62). Heterocyst formation is also controlled bythe availability of 2-oxoglutarate, which provides the carbonskeleton for the incorporation of inorganic nitrogen and whichalso serves as a signal molecule of the organic carbon contentin the developing heterocysts (122, 161, 223). NtcA is the main2-oxoglutarate sensor for the initiation of heterocyst differen-tiation (239). The otherwise important signal protein PII,which is involved in regulation of nitrogen metabolism in bac-teria and plants, is apparently not required for heterocyst for-mation (240). Nitrogenase synthesis has a high demand for Fe.The uptake of this element is controlled by furA, whose ex-pression is also modulated by NtcA and HetR (128). Thereader is referred to review articles on this complex regulatorycascade (84, 94).

Before nitrogenase can be expressed in Anabaena sp. strain7120, a gene rearrangement has to occur within nifD. An 11-kbDNA element is excised by a specific enzyme (XisA), and thetwo fragments of nifD are ligated to allow nitrogenase tran-script formation to proceed. The excisase gene xisA is locatedon the excised DNA element. This gene rearrangement occursin heterocystous cyanobacteria, such as the best-studied spe-cies Anabaena variabilis (37) and Anabaena (Nostoc) PCC 7120(43), but not in nonheterocystous, N2-fixing forms (93). Similar

rearrangements happen during the late stages of heterocystdevelopment of some cyanobacteria. These include excisionwithin a special ferredoxin (fdxN) of a 55-kb element by XisFand excision of a 10.5-kb element within the large subunit ofuptake hydrogenase (hupL) (see below) mediated by XisC.These genetic elements may represent ancient viruses thathave come under the control of the host and are excised asrequired. Similar gene rearrangements were detected duringspore formation in bacteria. The subject has been reviewed(93), and newer publications on this subject are available (43,96, 198).

Electron Transport to Nitrogenase in Cyanobacteria

Electron transport to nitrogenase has been studied exten-sively in heterocystous cyanobacteria. Heterocysts have a veryactive ferredoxin- and photosystem I-dependent cyclic photo-phosphorylation (28) which generates the ATP for N2 fixation.These cells possess several ferredoxin-like Fe-S proteins. Ofthese, a special FdxH is expressed only in heterocysts and wasproposed to serve as the electron carrier to nitrogenase. How-ever, mutants with mutations in FdxH can still perform N2

fixation at a high rate (138), indicating that this protein can bereplaced by others. Another ferredoxin-like protein, FdxB(PatB), is specifically expressed in heterocysts (107). Neitherferredoxin was identified in a quantitative proteomic investi-gation of heterocysts (163).

Reducing equivalents for the reduction of ferredoxins can begenerated by several pathways (Fig. 5). In heterocysts, in thelight, ferredoxin can be reduced via photosystem I. Alterna-tively, either NAD(P)H and a dehydrogenase or H2 and up-take hydrogenase (see below) can feed in electrons at theplastoquinone site (or close to it). In darkness, ferredoxin canbe reduced by NAD(P)H and NAD(P)H:ferredoxin oxi-doreductase (FNR) present in heterocysts and vegetative cells.The reduction of ferredoxin can also be achieved by the pyru-

FIG. 5. Generation of reductant for N2 fixation in cyanobacteria. The details are explained in the text.



mbr.asm

.org/D

ownloaded from


vate phosphoroclastic reaction. Here, pyruvate and coenzymeA are cleaved to acetyl coenzyme A and CO2, and the remain-ing two electrons are transferred to ferredoxin. The enzymeinvolved, the pyruvate:ferredoxin oxidoreductase (PFO) is typ-ically distributed among anaerobes, either strict (Clostridium)or facultative (Escherichia coli).

A somewhat controversial issue arose regarding the occur-rence of PFO in cyanobacteria. The enzyme was originallyobserved in extracts from Anabaena variabilis (120) and wasthen characterized in much greater detail from Anabaena cy-lindrica (151). Extracts from the latter cyanobacterium cata-lyzed the pyruvate-dependent reduction of methyl viologen (asan artificial substitute of ferredoxin) with formation of CO2

and the synthesis of acetohydroxamate from the acetyl coen-zyme A produced. The reverse reaction, the synthesis of pyru-vate from acetyl coenzyme A, CO2, and reduced ferredoxin,was also demonstrated. This reaction is even more indicativefor the occurrence of the pyruvate:ferredoxin oxidoreductasebecause the pyruvate dehydrogenase complex is thermody-namically unable to catalyze this reaction.

Despite all this work, the occurrence of the phosphoroclasticreaction in cyanobacteria was not readily accepted in the lit-erature until 1993, when two groups independently publishedsequences of the nifJ gene, encoding PFO. The enzyme fromAnabaena sp. PCC 7120 was expressed only under Fe defi-ciency in the growth medium (12), whereas it was constitutiveand independent of the Fe content in A. variabilis (192). Thesequenced parts of the two nifJ genes showed only a low sim-ilarity of ca. 75%, in contrast to the sequences of other genesfrom the two organisms, which did not differ by more than 5%.The genome sequencing project for Anabaena 7120 then re-vealed that this cyanobacterium contained two nifJ genes andthat the two above-mentioned groups had each sequenced adifferent nifJ copy. All cyanobacterial PFO sequences clusterwith those from strict anaerobes, such as Clostridium or Desul-fovibrio (191). However, as shown by the lux reporter system,PFO is expressed both under aerobic growth conditions and inFe-replete medium in the unicellular, non-N2-fixing Synecho-coccus sp. PCC 7942. This cyanobacterium and other com-pletely sequenced unicellular cyanobacteria contain only onePFO. Their genomes also contain sequences for phosphotrans-acetylase and acetate kinase. Acetyl coenzyme A could, there-fore, be converted to acetyl-phosphate and then to ATP as afermentative generation of additional energy. Such ATP gen-eration has, however, never been verified experimentally incyanobacteria.

Under Fe deficiency conditions, some cyanobacteria synthe-size flavodoxin (formerly termed phytoflavin) instead of ferre-doxin (221). Despite statements to the contrary (12), fla-vodoxin effectively transfers electrons to nitrogenase whenproperly reduced (32). Flavodoxin exists in three redox states,the oxidized, semiquinone, and fully reduced (hydroquinone)forms. Only the hydroquinone/semiquinone couple, with anE0� of about �500 mV, can transfer electrons to nitrogenase incyanobacteria (32) and in Azotobacter (235). Reduction of fla-vodoxin to the fully reduced state does not occur effectivelyusing NAD(P)H [E0� of NA(P)H/NAD(P)� � �320 mV], butit can proceed via photosystem I or from pyruvate (E0� for thepyruvate cleavage �500 mV). Flavodoxin is constitutive inthe nonphotosynthetic aerobe Azotobacter vinelandii (225). It

remains to be elucidated under what conditions flavodoxin hasa physiological role in cyanobacteria. Fe deficiency is generallynot a constraint in nature that demands the expression offlavodoxin. The demonstration of flavodoxin, other flavopro-teins, and other ferredoxin-like electron transferring proteinsin heterocysts of Nostoc sp. PCC 7120 (162) in non-Fe-limitedcultures may indicate that other, still unresolved electrontransfer pathways operate in these specialized cells. Similarevidence may be derived from work with Nostoc punctiformeATCC 29133, where two ferredoxin-like electron transportproteins show a markedly increased abundance together withFNR in heterocysts (163). Flavodoxin was reported to enhancecyclic electron flow around photosystem I in salt-stressed cells(89), which may also occur in N2-fixing heterocysts.

Alternative Nitrogenases in Cyanobacteria

The occurrence of the V nitrogenase in cyanobacteria wasfirst inferred from physiological evidence with A. variabilis(111). Under Mo deficiency and with V in the culture medium,this cyanobacterium reduced significant amounts of C2H2 toC2H6 and also produced much more H2 than Mo-grown cells.Subsequently, Thiel and coworkers performed the molecularcharacterization in great detail (216). In A. variabilis, thevnfDGKEN genes occur as a cluster, whereas four other Hgenes, in addition to nifH, are interspersed on the chromosome(Fig. 4). A vnfH gene is located 23 bp from vnfDGK. EitherNifH or VnfH can act to complement either Mo or V nitro-genase. Two copies of the H gene exist in Nostoc punctiforme,which does not possess any other genes encoding an alternativenitrogenase (Fig. 4).

Among cyanobacteria, the V nitrogenase has been foundonly in A. variabilis, in an Anabaena isolate from the fern Azolla(154), in the southern Chinese rice field isolates AnabaenaCH1 and Anabaena azotica (26), and recently in one Nostocstrain and two Anabaena strains (141). Anabaena azoticathrives at high temperatures at which Azolla dies. A differentexpression pattern for the two cyanobacterial nitrogenases,possibly dependent on growth temperature, was suspected(26). In support of this idea, the V but not the Mo nitrogenaseof A. vinelandii has been found to be active at lower temper-atures (167). However, the specific activities of C2H2 reductionfor both Mo and V nitrogenase of A. azotica were found to bethe same over a range of temperature and light regimens (26).Thus, the V nitrogenase is unlikely to provide a selective ad-vantage for A. azotica at higher temperatures. Other condi-tions, such as Mo-deficient microzones around microbial col-onies, unusually high W concentrations (which block Monitrogenase synthesis), or high alkalinity (pH of 10), havebeen suggested, but not proven, to favor V nitrogenase geneexpression (222).

The close sequence similarity of the cyanobacterial vnfDGgenes to those of Methanosarcina spp. could indicate an ar-chaeal origin for the alternative nitrogenase similar to that forthe Mo enzyme (176). Alternatively, these two groups of or-ganisms with totally unrelated taxonomic affinities may haveretained these genes in evolution by chance.

Some physiological evidence has been presented for theexistence of the Fe nitrogenase in A. variabilis (112). However,the completely sequenced chromosome of this organism and of



mbr.asm

.org/D

ownloaded from


more than 30 other cyanobacteria did not reveal genes codingfor the Fe nitrogenase, and a nifH vnfH double mutant ofAnabaena variabilis did not grow diazotrophically (172). Thus,the evidence, particularly the positive results after hybridiza-tion with an anfH probe from Azotobacter vinelandii (112),must indicate the presence of some other sequence-relatedentity (possibly two other nifH copies [Fig. 4]). In the past,searches for nitrogenases were often based on probing with thenifH gene. However, sequences of anfH are significantly diver-gent from those of nifH and vnfH (59), and thus possibly acyanobacterial Fe nitrogenase, say, occurring on a plasmid,may have been missed by probing with the nifH gene.

In waters, cyanobacteria thrive under oxygenic conditionswhere Fe is generally limiting but Mo or V is abundantlyavailable (63). Those authors suggest that these conditions mayfavor the expression of Mo or V nitrogenase, whereas theconcentration of Fe is possibly too low to allow synthesis of theFe nitrogenase.

In 1995, two groups independently reported the existence ofa second Mo nitrogenase in A. variabilis (193, 217). The “clas-sical” Mo nitrogenase occurs only in heterocysts of this organ-ism. The second Mo nitrogenase is encoded by a separate setof nifHDK genes and is expressed in vegetative cells underanaerobic or, more precisely, low-O2-tension conditions be-cause these cells produce O2 photosynthetically. It resembles,by its expression under anaerobic conditions, the enzyme fromthe filamentous, nonheterocystous Plectonema (Leptolyngbya)boryanum (209). Its physiological and biochemical propertiesin A. variabilis have not been studied extensively. The distri-bution of this enzyme has recently been screened in severalcyanobacteria (141).

Nitrogen Fixation in Nonheterocystous Cyanobacteria

The literature on nitrogen fixation in nonheterocystous cya-nobacteria up to the mid-1990s was extensively reviewed (14).Therefore, this section concentrates on more recent results.

Many nonheterocystous cyanobacteria can fix N2, but almostall of them do so under anaerobic conditions, or, rather, underconditions of decreased O2 tension. Several of them wereshown to separate these two incompatible reactions, with pho-tosynthetic CO2 fixation being performed in the light and N2

fixation in darkness. Thus, at night, nitrogenase is not exposedto the photosynthetically produced O2 and respiration mightthen utilize most of the O2 of the air to provide anaerobicconditions, especially in dense cultures or in biofilms. How-ever, not all nonheterocystous cyanobacteria show this circa-dian rhythm. Gloeothece and Synechococcus (Cyanothece) spp.also fix N2 during the day and can grow slowly under contin-uous illumination. In the oceans, the filamentous Trichodes-mium may show a division of labor in which some cells performphotosynthesis whereas others fix N2 (14). However, a recentimmunological study (156) revealed that more than 77% of allcells were nitrogenase immunopositive, indicating that Tri-chodesmium does not develop heterocyst-equivalent cells. Im-munological studies indicated that nitrogenase in Plectonema,Gloeothece, and others is also uniformly distributed throughoutall cells, thus showing no preferential association with a cellstructure (14). Cyanobacteria did not develop O2 protectiondevices, such as changes in the enzyme’s conformation upon

exposure to excess O2 as in azotobacteria, production of leghe-moglobin as in the rhizobia, or reversible modification of theFe protein by ADP-ribosylation controlled by the DRAT/DRAG enzymes as in photosynthetic purple bacteria or azos-pirilla. Their respiratory activity does not seem to be extraor-dinarily high as in Azotobacter sp. or in heterocysts, to consumeall O2 reaching within the cells (66). Thus, N2 fixation in lightby these few aerobic cyanobacteria remains an enigma.

Cyanobacterial N2 fixation in the oceans contributes signif-icantly to the global N budget (15, 55, 202). In temperate areas,heterocystous species can form blooms in summer, but they aresomewhat unpredictable in time and location, as exemplifiedfor the fresh- and brackish-water species Aphanizomenon flos-aquae and the toxin-producing Nodularia spumigena (143). Themajor organisms in oceanic N2 fixation in areas of the warmertropical and subtropical regions of the Pacific Ocean are Tri-chodesmium sp. and the heterocystous Richelia intracellularis,which lives inside diatoms (74, 206). In other areas of thePacific Ocean, N2-fixing cyanobacteria, such as Crocosphaerawatsonii, and the non-N2-fixing Prochlorococcus marinus thrivein abundance (238). Other nanoplanktonic organisms may beeven more important there. Small uncultured cyanobacteriathat fix N2 but are unable to perform photosynthetic CO2

fixation and thus O2 evolution have now been recognized(237), and they are particularly active during winter in areas ofthe Pacific Ocean (117). They have not yet been characterizedproperly, but their nitrogenase DNA sequences resemble thoseof the “spheroid bodies” that occur in the fresh water diatomsRhopalodia gibba and Epithemia sp. (80). These diatoms growvery slowly on agar plates. During the time before the use ofmolecular biology techniques, physiological experiments dem-onstrated light-dependent C2H2 reduction by R. gibba evenwith the rather small amounts of cell material then available(71). More recently, DNA sequencing showed that the spher-oid bodies of R. gibba indeed possess the structural nitrogenasegenes (173). The spheroid bodies and uncultured marine cya-nobacteria either could perform cyclic phosphorylation or maybe completely dependent on a supply of both ATP and reduc-tant from organic carbon in the environment. These spheroidbodies, being N2-fixing entities within eukaryotic cells, mightattract special attention in the near future for potential appli-cations. They could serve as models in attempts to make plantsindependent from a supply with combined nitrogen by incor-porating an N2-fixing cyanobacterium into their cells.

The discovery of a new group of N2-fixing cyanobacteria mayappear to be totally unexpected. As mentioned above, nifH isvery much conserved during evolution, and probing with nifHsequences should allow one to detect all N2-fixing microorgan-isms in environmental samples. Recent studies showed thatmost of the bacterial DNA sequences from soil (for nifH aswell as for nosZ in denitrification and for the 16S rRNA genefor total bacteria) could be detected with the short DNAprobes available, but the gene sequences in total were entirelynew (60, 180).

HYDROGENASES IN GENERAL

The subject of hydrogenases has been extensively reviewed(226, 228). Therefore, just a few general facts will bementioned here. There are three classes of hydrogenases:



mbr.asm

.org/D

ownloaded from


(i) the [FeFe] hydrogenase, (ii) the [FeNi] hydrogenase, and(iii) the methylenetetrahydroxymethanopterin-containing en-zyme. The last enzyme is a homodimer, each subunit of whichcontains a low-spin, redox-inactive Fe atom which is involvedin H2 splitting or formation (201, 211, 229). This enzyme hasbeen found only in some methanogenic archaea. In all otherhydrogenases, iron occurs in Fe-S clusters.

The [FeFe] hydrogenases have a unique active center (the Hcluster) which produces about 100-fld higher activity than theother hydrogenases (229). The simplest [FeFe] hydrogenaseoccurs in green algae with only the H cluster as the prostheticgroup (91). The H cluster contains two Fe atoms and the twoligands CO and CN�, which are attached to both of the Featoms. In green algae, the H cluster is directly reduced byferredoxin. All other [FeFe] hydrogenases contain a relay ofadditional FeS centers (both 4Fe-4S and 2Fe-2S clusters) thatare involved in electron transfer from the external electronsource (reduced ferredoxin) to the H cluster deep inside thesemonomeric proteins. They possess hydrophobic channels fromthe surface to the active site (the H cluster) that provide accessfor protons and the egress of H2. [FeFe] hydrogenases functionmostly in the disposal of excess reductant generated duringfermentation under anaerobic conditions. However, theperiplasmic [FeFe] hydrogenase of Desulfovibrio vulgaris is in-volved in the utilization of H2 in sulfate reduction (171). Theenzyme occurs in anaerobes, such as the genera Clostridiumand Desulfovibrio, and in eukaryotes (in chloroplasts of greenalgae or in hydrogenosomes). It has not been detected incyanobacteria. This is true also for those cyanobacteria thatsynthesize starch (semiamylopectin) and could therefore beconsidered ancestors of chloroplasts (150). The evolutionaryorigin of the [FeFe] hydrogenase of green algae is a mystery yetto be resolved (132, 146).

The majority of hydrogenases in prokaryotes are Ni-contain-ing enzymes. The core enzyme is an �� heterodimer where thelarger subunit, of ca. 60 kDa, possesses the deeply buriedbinuclear NiFe active site (Fig. 6). The Fe in this center bindstwo CN� and one CO. The whole cluster is ligated to theprotein by the thiolate groups of four cysteines. The smallersubunit, of ca. 30 kDa, harbors FeS clusters (up to three) whichserve to transfer electrons from or to the NiFe active site. Asin the [FeFe] hydrogenases, there are hydrophobic channelsfrom the active site to the surface of this globular �� dimer.The Ni hydrogenases have a high affinity (low apparent Km) forH2, indicating that they act mostly in utilizing H2 in the differ-ent organisms. Indeed, they are often linked to nitrogenase,where they serve to utilize the H2 produced in N2 fixation.They are often membrane bound and feed electrons into therespiratory chain via either ubiquinone or a cytochrome atrespiratory complex III. Often they are synthesized with a longsignal peptide of 30 to 50 amino acid residues which is cleavedoff when the hydrogenase is folded and incorporated into themembrane. They may be subdivided into four groups by theirfunctions (227, 228).

In the oxidized form, [NiFe] hydrogenases are inactive dueto a bridging hydroxo ligand between the Ni and Fe atoms (Fig.6), and the different enzymes vary in their sensitivity to O2.When reduced, this ligand is removed by conversion to water,with the simultaneous reduction of Ni3� to Ni2�. The enzymecan then bind H2, probably at the Fe atom, and is then able to

catalyze the heterolytic cleavage to 2H� � 2e�. Details of thisenzymatic mechanism have been depicted previously (228).Remarkably, none of the [NiFe] hydrogenases transfers elec-trons to ferredoxin or to another low-potential electron carrier.The structure/function relationship of anaerobic gas-process-ing metalloenzymes has recently been summarized (73).

The biosynthesis of hydrogenase, including the synthesis ofthe metallocenter and the incorporation of the CO and CN�

ligands, has been studied extensively for hydrogenase 3 from E.coli by Bock and colleagues in Munich and has been reviewed(226, 228). The concentration of H2 in cells is sensed by hupUVgene products, which in other organisms are termed HoxBC.These proteins also catalyze the cleavage of H2 and can there-fore be considered an independent, regulatory hydrogenase,e.g., in Ralstonia eutropha (79).

HYDROGENASES IN CYANOBACTERIA

Hydrogenase Types in Cyanobacteria

The subject of hydrogenase types in cyanobacteria has beenrepeatedly reviewed (7, 81, 82, 91, 99, 134, 194, 199, 212, 214,222). The reader is particularly referred to the very detailedand elaborate review by Tamagnini et al. (214). Cyanobacteriacontain two different Ni hydrogenases defined by their physi-ological role as either an uptake or a bidirectional (reversible)enzyme. There is no evidence for an H2-sensing regulatoryhydrogenase encoded by hupUV. Cyanobacterial hydrogenasesdo not contain Se as do some hydrogenases in anaerobic bac-teria.

Uptake hydrogenase. The uptake hydrogenase is encoded bythe contiguous and cotranscribed genes hupSL and is associ-ated with nitrogenase functioning. Generally, intact N2-fixingcyanobacteria show very little net H2 production due to theefficient recycling of the gas by uptake hydrogenase. This H2

consumption proceeds by the respiration- and photosystem

FIG. 6. (A) Prosthetic group of [NiFe] hydrogenases in the oxi-dized, inactive form (Ni-A state [228]). (B) Upon reduction, it isconverted to the active form (Ni-S state). (Adapted from reference228, where further details can be found.)



mbr.asm

.org/D

ownloaded from


I-dependent pathways (33). In cyanobacteria, respiration andphotosynthesis share the cytochrome bc complex (respiratorycomplex III), from where the electrons are allocated either tothe donor side of photosystem I to generate reduced ferre-doxin or to respiratory complex IV accompanied by O2 con-sumption. Factors that control electron allocation to eitherphotosystem I or respiration in light-grown cyanobacteria havenot been elucidated. Likewise, the electron entry from H2 anduptake hydrogenase to either the plastoquinone pool or acytochrome b, as in Xanthobacter autotrophicus (184) and pre-sumably in Bradyrhizobium japonicum (65), is not known incyanobacteria. Transcription starts before hupS and terminatesimmediately after hupL; thus, the electron acceptor is notcotranscribed on this operon. The enzyme does not couplewith any other electron carrier with a redox potential morenegative than �300 mV, which explains its unidirectional phys-iological function and name. Uptake hydrogenase is mem-brane bound and has never been characterized in the homo-geneous form. Recent immunological studies confirmed itsassociation with the thylakoid membranes of three cyanobac-terial strains (195), which corroborates earlier studies withthylakoid preparations (reviewed, e.g., in reference 164). Thesequences indicate that the larger subunit (HupL) has a mo-lecular mass of about 60 kDa and that the smaller one (HupS)is about half that size.

In accordance with the postulates of Dixon (58), which weredeveloped for Rhizobium nodules, H2 utilization in cyanobac-teria likely functions (i) to remove O2 from the nitrogenase sitevia the respiratory oxyhydrogen (Knallgas) reaction, (ii) toregain ATP inevitably lost in H2 production during nitrogenasecatalysis, and (iii) to prevent a deleterious buildup of a highconcentration of H2 which affects nitrogenase activity. Such asituation might apply particularly to heterocysts. In addition,H2 uptake might provide additional reductant for N2 fixation,photosynthesis, and other reductive processes.

Rather simple physiological experiments, performed in stu-dent courses in the Cologne laboratory over the years, showthat N2 fixation (C2H2 reduction), e.g., by Anabaena variabilis,is much less sensitive to exposure to O2 when the assay mix-tures are supplemented with exogenous H2 (29, 34). Uptakehydrogenase-deficient mutants of several cyanobacteria pro-duce roughly three times more H2 than wild-type cells (forreferences, see reference 214). However, their growth ratesunder N2-fixing conditions are essentially the same (125).

In other bacteria, a twin-arginine signal peptide at the Nterminus and a hydrophobic motif, both presumably involvedin translocation and anchorage, are typical for many mem-brane-bound hydrogenases. Such motifs are missing from thecyanobacterial HupS and HupL, which also do not containsignatures indicative of membrane insertion. As in other or-ganisms, however, HupL contains the C-terminal extensionthat is cleaved off at the last step of maturation by a specificendopeptidase encoded by hupW.

In approximately half of the heterocystous strains (21, 213),hupL undergoes a rearrangement during the late state of het-erocyst differentiation before it can be transcribed. The exci-sion of the 9.5-kb element is catalyzed by the recombinaseXisC, with its gene located on this element. XisC is sufficient tocatalyze the site-specific recombination in hupL (43). Thephysiological advantage of such a site-specific recombination is

not obvious. Of the two best-studied heterocystous cyanobac-teria, Anabaena (Nostoc) sp. strain PCC 7120 shows this generearrangement but Anabaena variabilis ATCC 29413 does not.

Uptake hydrogenases occur in almost all N2-fixing microor-ganisms except for some Rhizobium strains (35, 36) andHerbaspirillum seropedicae (F. Pedrosa, personal communica-tion). In cyanobacteria, the enzyme is present in all N2-fixingspecies with the exception of an N2-fixing unicellular strain,Synechococcus sp. strain BG 043511 (132), and some Chroo-coccidiopsis isolates (see below). No uptake hydrogenase andnone of its genes have been unambiguously detected in non-N2-fixing cyanobacteria. It is not clear whether an uptake hy-drogenase is expressed in parallel with the second Mo nitro-genase which is active in vegetative cells of A. variabilis upontransition to anaerobiosis. Low transcript levels of hupSL havebeen reported for A. variabilis ATCC 29413 cells grown in thepresence of ammonia (231).

The formation of hupSL transcripts may be controlled byfactors such as Ni availability, anaerobiosis, the presence of H2,and the absence of combined nitrogen, among others, and mayproceed in parallel with heterocyst formation (92, 98, 231). Thetranscriptional regulator NtcA, which controls cyanobacterialgenes involved in nitrogen metabolism, has also been reportedto regulate hupSL expression (231). The NtcA binding site wasidentified 427 bp upstream of the transcriptional start site ofhupSL in A. variabilis, whereas most other NtcA binding sitesare located not more than 40 bp from the start site (231). TheNtcA binding site identified in Nostoc punctiforme ATCC29133 is TGTN9ACA, which differs from the optimal one,GTAN8TAC, and this might therefore indicate only weakbinding (98). A shorter promoter fragment, covering 57 bpupstream of and 258 bp downstream of the transcription startsite, was enough for high heterocyst-specific expression ofhupSL independent of NtcA (98). Surprisingly, hupSL expres-sion in A. variabilis ATCC 29143 was not regulated by H2

(231). This is in sharp contrast to the situation in Nostoc punc-tiforme and N. muscorum (10). In addition to transcriptionalregulation, uptake hydrogenase synthesis could also be con-trolled at the posttranslational level. This enzyme, but notbidirectional hydrogenase, is activated by thioredoxin (164). N2

fixation by cyanobacteria is largely light stimulated due to thedemand for reductant (reduced ferredoxin) and ATP. There-fore, activation of the uptake hydrogenase by photosyntheti-cally reduced thioredoxin makes sense physiologically becausemore H2 is produced by nitrogenase in the light than in dark-ness. The number of proteins activated by thioredoxin is highin cyanobacteria and chloroplasts, but the target enzymes differin the two entities (124).

Other transcription and translation cues will undoubtedly beresolved in the near future to further understanding of thesignal cascade involved in the synthesis of the uptake hydro-genase. The currently available data suggest that different cya-nobacteria differ markedly in their patterns of expression ofthis protein.

Bidirectional hydrogenase. After extensive controversy, thework of Houchins and Burris (100, 101) clearly showed thatN2-fixing cyanobacteria may contain another hydrogenase inaddition to the uptake enzyme. This reversible, bidirectionalhydrogenase, which catalyzes both H2 uptake and reducedmethyl viologen-dependent H2 evolution, was separated from



mbr.asm

.org/D

ownloaded from


the unidirectional, uptake enzyme in crude extracts ofAnabaena (Nostoc) sp. strain PCC 7120. Later, molecular bio-logical characterization showed that the bidirectional hydroge-nase in cyanobacteria is, surprisingly, a NAD(P)H-dependentenzyme (187). This finding had been marked as a milestone incyanobacterial hydrogenase research (212). The enzyme has apentameric structure encoded by the genes hoxEFUYH in A.variabilis. HoxYH constitutes the hydrogenase, which containsthe motifs for binding both Ni-Fe-S and Fe-S centers. HoxFUis the diaphorase part that transfers the electrons to NAD(P)�

and possesses binding sites for NAD(P)�, flavin mononucle-otide (FMN), and Fe-S centers. The enzyme complex containsa further HoxE subunit, which copurifies with the active bidi-rectional enzyme (188). The hoxE gene possesses a motif forbinding an Fe center and was therefore thought to couple theenzyme to the respiratory and photosynthetic electron trans-port chain on the thylakoids and also possibly at the cytoplas-mic membrane. However, the role of the hoxE gene producthas not been resolved yet, despite extensive research.

In organisms other than cyanobacteria, a pentamericNADH-dependent bidirectional hydrogenase is present inThiocapsa roseopersicina (174) and in Allochromatium vinosum(108, 127). The best-studied bidirectional hydrogenase, theNADH-dependent enzyme from Ralstonia eutropha, is en-coded only by hoxFUYH (75).

The locations of the five structural genes hoxEFUYH on thechromosome differ from one cyanobacterium to the next (25,199, 214). In some cyanobacteria, they are clustered on onepart of the chromosome, though interspersed with ORFs atdifferent positions. In others, they occur in two different partsof the genome separated by several kilobases of interveningDNA. Similar to the case for HupL of the uptake hydrogenase,HoxH of the bidirectional hydrogenase undergoes maturationat the C terminus catalyzed by a specific endopeptidase en-coded by hoxW. The expression of hydrogenase genes inSynechococcus sp. PCC 7942 is under the control of the circa-dian clock, as shown for two promoters of the gene cluster(186). When expressed, the native protein might function as adimeric assembly complex Hox(EFUYH)2 (188). In extracts, itcatalyzes both NAD(P)�-dependent H2 uptake and H2 evolu-tion with NADP(P)H as the electron donor (190).

Bidirectional hydrogenase is widespread in cyanobacteria. Itis present in unicellular, filamentous, and heterocystous spe-cies, where it occurs in both heterocysts and vegetative cells(213). The enzyme is apparently not present in marine cya-nobacteria isolated from the open ocean (132). It is expressedindependently of N2 fixation and thus is present in cells grownaerobically and with combined nitrogen. However, it is moreO2 sensitive than uptake hydrogenase, probably due to oxida-tion to its inactive state (51). When reduced, it can be purifiedas a pentameric complex (188).

The regulation of the expression of the bidirectional hydro-genase in cyanobacteria differs with the physical location of thehox genes on the chromosome in the species. In Synechococcussp. PCC 7942 (� Anacystis nidulans), the genes are organizedinto two clusters, hoxEF and hoxUYHWhypAB, and are regu-lated by three promoters, one before each of hoxE, hoxU, andhoxW (23, 186). In Synechocystis sp. PCC 6803, the hoxEFUYHgenes are cotranscribed, with the transcription start point lo-cated 168 bp upstream of the start codon (87, 158). Taking the

high diversity of the different cyanobacterial species into ac-count, expression of the bidirectional hydrogenase in cya-nobacteria seems to be species specific.

Over the last several years, significant progress has beenachieved in the identification of the transcription factors reg-ulating the expression of bidirectional hydrogenase, and detailsof the subject are found in a very recent review (159). NtcAdoes not seem to be the transcriptional activator, but a LexA-related protein (87, 158) and two members of the AbrB-likefamily (157) appear to be activators. In other organisms, LexAactivates the expression of a cascade of genes coding for en-zymes involved in either DNA repair or carbon starvation. ALexA-depleted mutant of Synechocystis sp. 6803 had lowerhydrogenase activity than the wild type, indicating that LexAoperates as a transcription activator of hox genes in this cya-nobacterium (87). The binding site of LexA upstream of hoxEof Synechocystis sp. PCC 6803 is, surprisingly, not clear (214).LexA may bind to a region from bp �198 to �338 from thetranslational start point (158), to the region from bp �592 to�690 bp from the hoxE start codon (87), or to both regions(159). The two distant LexA binding regions in the hox pro-moter could indicate the occurrence of a DNA loop involved ingene transcription (86, 159), which warrants experimentalproof. LexA may act as mediator of the redox-responsive reg-ulation of hox gene expression (5). In Synechocystis sp. strainPCC6803, LexA binds as a dimer to 12-bp direct repeats con-taining a CTAN9CTA sequence in target genes (170).

Abr proteins act as transcription factors of antibiotic resis-tance in organisms other than cyanobacteria. An AbrB-likeprotein (sII0359) was recently shown to interact specificallywith the promoter region of the hox genes and with its ownpromoter region (157). Whereas this AbrB-like protein worksas a transcription activator in Synechocystis sp. PCC 6803, an-other one of these regulator proteins (sII0822) acts as repres-sor of the hox gene expression, because they were significantlyupregulated in a completely segregated �sII0822 mutant (105).This transcription factor works in parallel to, but apparentlyindependently from, the long-known nitrogen transcriptionalcontrol element NtcA (97) in the regulation of the expressionof genes coding for nitrogen assimilation enzymes (105).

The cyanobacterial transcription factors, the LexA- andAbrB-like proteins, show significant divergences in their se-quences and functions from the counterpart proteins in otherorganisms, and their activity may be regulated by posttran-scription modifications (159). They are members of an appar-ently complex signal cascade that directs the expression of thebidirectional hydrogenase genes. Their expressions and inter-actions in responses to environmental cues might be a subjectof extensive research in the near future (159). The identifica-tion of other transcription factors of bidirectional hydrogenaseis to be expected (116).

Besides its inactivation by O2 and a non-light dependence(51), bidirectional hydrogenase seems to be activated by H2 onthe transcriptional or translational level or even on both. Theeffects of H2 on bidirectional hydrogenase synthesis are notunderstood and appear to vary with the organism and theculture conditions employed. In some cases, high hydrogenaseactivity could be the result of bacterial contamination of slime-forming cyanobacterial cultures.

The biosynthesis and maturation of the [NiFe] hydrogenase



mbr.asm

.org/D

ownloaded from


have been characterized for the enzyme from E. coli (20). Thehyp genes required for the synthesis of the hydrogenase aresimilar in E. coli and cyanobacteria and are scattered through-out the genomes of those cyanobacteria in which their occur-rence was examined (reviewed in reference 214). Both uptakeand reversible hydrogenases appear to utilize the same hupgene products for their biosynthesis. However, the last step,the maturation at the C terminus by endopeptidase, seems tobe specific for the two enzymes, with HupW catalyzing the finalcleavage of uptake hydrogenase and HoxW involved in pro-cessing the bidirectional enzyme (233). Both endopeptidasesare transcribed from their own promoters (67) and are undersimilar regulatory control as the hydrogenases they cleave (54).

In contrast to uptake hydrogenase, the bidirectional enzymeis soluble after breaking cyanobacterial cells. The exact loca-tion of the enzyme inside the cells is unknown (Fig. 7). Immu-nological (109) and membrane solubilization (110) studies in-dicated a location at/on the cytoplasmic membrane in Anacystisnidulans (Synechococcus PCC 6301). Other researchers withdifferent antibodies found a location in the cytoplasm, withsome preferential association to the thylakoids (213, 214).However, all these investigations with antibodies were per-formed before the true nature of the hydrogenase as a pen-tameric NAD(P)H-dependent complex was recognized.Clearly, this issue needs to be reexamined with newly raisedantibodies.

The physiological function of this constitutively expressedbidirectional hydrogenase in photosynthetic, aerobic cyanobac-teria has been hotly debated but remains controversial. Workwith mutants of Anabaena (Nostoc) sp. PCC 7120 (139) showedthat the bidirectional hydrogenase is unable to support N2

fixation. Its high affinity (low apparent Km value) for H2 sug-gests that the enzyme functions in H2 utilization under physi-ological conditions (99). Indeed, H2 uptake catalyzed by thebidirectional hydrogenase can support photosynthetic reac-tions such as CO2 fixation and also, to some extent, nitrite orsulfite reduction (215). The rates of these reduction reactionswith H2 as the only electron donor are low, however, comparedto these same photosynthetic activities with H2O as the elec-

tron source (31). Bacteria, such as Ralstonia eutropha orXanthobacter autotrophicus (185), are able to grow autotrophi-cally with H2 as the sole source of reductant and energy, andsome of them, such as Bradyrhizobium japonicum, can do soeven under N2-fixing conditions (204). H2-dependent growth indarkness has never been demonstrated for any cyanobacte-rium. Such anoxygenic growth is possible when energy is pro-vided by cyclic photophosphorylation and when the electronsare provided from Na2S or H2S in some cyanobacteria, such asOscillatoria limnetica (78). However, to our knowledge, H2-and photosystem I-supported growth has not yet been demon-strated in cells of Anabaena, Nostoc, or other autotrophic uni-cellular species when photosystem II is impaired by use ofdichlorophenyldimethylurea (DCMU). In the two facultativeanoxygenic cyanobacteria Oscillatoria limnetica and Aphano-thece halophytica, however, H2 was described to substitute forH2S in supporting CO2 fixation in a photosystem I-driven re-action (13).

In all organisms, the respiratory complex I consists of at least14 subunits, but only 11 in the cyanobacterial NADPH-dehy-drogenase complex I have as yet been identified. The diapho-rase genes hoxEFU show high sequence homologies to themissing three genes. Although it has been suggested that thehoxEFU gene products are used simultaneously by both the bi-directional hydrogenase and respiratory complex I (189), theexperimental evidence is against this suggestion. Mutants withmutation either in hoxF (102) or in hoxU (22) do not showbidirectional hydrogenase activity but have unimpaired respi-ratory activity. Furthermore, Nostoc PCC 73102 has no bidi-rectional hydrogenase activity at all but respires with ratescomparable to those of other cyanobacteria (22). This couldmean that cyanobacterial respiration partly circumvents respi-ratory complex I and utilizes the succinate dehydrogenasecomplex instead, as may be inferred from studies with mutants(49). Then the fate of the NAD(P)H generated in carboncatabolism has to be determined. The electron input pathwayinto respiratory complex I in cyanobacteria remains unknown(11).

Some authors consider the bidirectional hydrogenase to

FIG. 7. Possible coupling of the bidirectional hydrogenase to the cytoplasmic membrane in cyanobacteria. The HoxE subunit may serve as adevice for coupling to the membrane, but this has not been verified experimentally. Solubilization experiments indicate that the bidirectionalhydrogenase is loosely membrane bound (110).



mbr.asm

.org/D

ownloaded from


work in the transition from anaerobiosis in the dark to aerobicconditions in the light (6, 51, 88, 132). In order to avoid anoverload of reducing equivalents, the organisms react to dis-pose of the excess by generating a burst of H2 via photosyn-thetic electron transport, ferredoxin, FNR, NADPH, and hy-drogenase. Such sudden H2 production that lasts for onlyseconds up to few minutes, has been observed repeatedly.However, the physiological relevance of this observation isquestionable, because the sun does not rise so suddenly in themorning that it overreduces soil cyanobacteria. Furthermore,in aqueous habitats, turbulences are hardly so effective thatthey expose cyanobacteria to extremely high light intensitieswithin a very short time scale. Cyanobacteria may, however, beoverreduced when continuously exposed to too bright a lighton a very sunny day and then be forced to use hydrogenase asa valve for disposing of the excess of photosynthetically pro-duced reductants, as shown in laboratory cultures of Anabaenacylindrica (119).

As stated in an extensive review (207), the majority of cya-nobacteria are obligate photoautotrophs. Only few species areable to grow chemoautotrophically at the expense of a limitednumber of organic carbon compounds, and they do so with O2

as the terminal respiratory electron acceptor. Anaerobic che-moorganic growth is exceptional in cyanobacteria. Thus, mostspecies accumulate glycogen in the light, which they then haveto degrade in darkness. Glucose residues from glycogen areutilized via the oxidative pentose phosphate pathway, finallyresulting in pyruvate (208). Its further degradation is ham-pered by the fact that the tricarboxylic acid cycle is incompletein cyanobacteria because neither an oxoglutarate dehydroge-nase complex nor an oxoglutarate:ferredoxin oxidoreductase ispresent (208), which has been confirmed by recent large-scaleproteomic studies (162, 163). This prevents the complete deg-radation of the C2 moiety to CO2 and NAD(P)H. Cyanobac-teria apparently prefer to utilize NADP� rather than NAD� incatabolism (51), since several enzymes, such as isocitrate de-

hydrogenase (165) and glyceraldehyde-3-phosphate-dehydro-genase (166), are NADP� rather than NAD� dependent. Indarkness, most cyanobacteria have to generate their energy viathe oxidative pentose phosphate pathway: pyruvate, pyruvate:ferredoxin oxidoreductase, reduced ferredoxin, FNR, andNADPH (Fig. 8). By using the lux reporter system, it wasshown that the pyruvate:ferredoxin oxidoreductase is constitu-tively expressed, even in aerobically grown A. variabilis (191).In dense cultures, biofilms, mats, or cyanobacterial blooms, theamount of O2 may rapidly become insufficient to oxidize allNAD(P)H by respiration. Thus, the NADPH generated viapyruvate:ferredoxin oxidoreductase and FNR must then bereoxidized via the bidirectional hydrogenase in order to avoidoverreduction in the cells. The generation of H2 (E0� � �420mV for H2/2H�) from NAD(P)H [E0� � �320 mV forNAD(P)H/NAD(P)�] is thermodynamically unfavorable. It re-quires a 1,000-fold excess of reduced pyridine nucleotides, butthis can rapidly be generated in dark-kept cells under anaero-bic conditions. To prevent overreduction of the cells during thenight, reducing equivalents must be disposed of as H2 (Fig. 8).Similar to the case for pyruvate:ferredoxin oxidoreductase, bi-directional hydrogenase is also constitutively expressed underaerobic growth conditions. When cyanobacteria such asSynechocystis, Anabaena, or Nostoc sp. are transferred to dark-ness and anaerobiosis, H2 production begins immediately withouta distinct lag phase. High hydrogenase activity under anaerobicconditions was described long ago (99), and an increase in thehoxH (67, 68) or hoxEF (116) transcription levels during darkperiods was recently detected in different cyanobacteria.

Thus, cyanobacteria might have retained the genes codingfor these enzymes (hydrogenase and pyruvate:ferredoxin oxi-doreductase) of anaerobes because of their obligatory auto-trophy (of many species). The essential role of hydrogenaseduring fermentation of cyanobacteria has also been suggestedby others (224). As recently shown (218), cyanobacteria con-tain one petH gene that encodes two isoforms of FNR, one of

FIG. 8. Roles of bidirectional and uptake hydrogenases in cyanobacterial hydrogen metabolism. Bidirectional hydrogenase is active mainly inthe dark and under anaerobic conditions to dispose of reductants, whereas uptake hydrogenase functions in recycling the hydrogen lost duringnitrogen fixation.



mbr.asm

.org/D

ownloaded from


which accumulates under heterotrophic conditions. It needs tobe shown whether the latter is specifically involved in thefermentative degradation of pyruvate. The same question alsoapplies to the two isoforms of pyruvate:ferredoxin oxidoreduc-tase in heterocystous species. As mentioned above, acetyl co-enzyme A formed in pyruvate fermentation may be convertedto ATP by phosphotransacetylase and acetate kinase, but thisalso remains to be shown. ATP formation by this pathway mustbe accompanied by the formation of acetate, but the fate of anyacetate produced remains unknown.

In photosynthetic eukaryotic algae, hydrogenase is locatedin plastids (210). The ancestors of plastids are believed to beorganisms similar to the filamentous, heterocyst-forming, N2-fixing species of class IV of the cyanobacteria, related to thecurrent Nostoc or Anabaena spp. (53). If so, it is surprising that,during evolution, plastids have lost not only N2 fixation genesbut also both gene sets that encode the bidirectional and up-take hydrogenases. When hydrogenase occurs at all in plastids,it is an [FeFe] hydrogenase of a completely unknown origin.

Similarly, it is totally unclear how both hydrogenases havebeen acquired by cyanobacteria from bacteria over evolution-ary time. With respect to photosynthetic bacteria, the greennonsulfur bacterium Chloroflexus aurantiacus possesses bothuptake and bidirectional hydrogenases, which has led to theassumption that a Chloroflexus-like bacterium is the ancestorof C. aurantiacus and cyanobacteria (132). On the other hand,the first phototrophs may have been anoxygenic procyanobac-teria from which the Chlorobiaceae, Heliobacillaceae, Chloro-flexaceae, purple sulfur bacteria, and cyanobacteria descendedin parallel and independently of each other (148). The genesets of both cyanobacterial hydrogenases may have been ac-quired vertically or laterally. A lateral gene transfer is partic-ularly difficult to conceive for the bidirectional enzyme becauseits genes may be scattered throughout the genome of a species.Similarly, the loss of hydrogenase from one cyanobacterialisolate but not from another may be difficult to explain.

The unicellular cyanobacterium Chroococcidiopsis sp. (Fig.9A to C) is regarded as a fossil relict which may have proper-

ties related to those of the first O2-evolving cyanobacteriumdeveloped some 3 � 109 years ago (69). Chroococcidiopsis isbeing proposed as the organism best suited to go on explor-atory missions to Mars (48). Today, Chroococcidiopsis thrivesat sites with extremely hostile conditions (24). The strainsChroococcidiopsis thermalis ATCC 29380 (1) and CALU 758(197) were found to possess the bidirectional, but not theuptake hydrogenase and to fix N2 (reduce C2H2) under mi-croaerobic conditions. However, experiments performed in theCologne laboratory (106) showed that the hydrogenase activ-ities of Chroococcidiopsis sp. strain PCC 7203 exhibited someunusual features. Southern hybridizations and PCR experi-ments with probes from hupL and hoxH, hoxF, or hoxE devel-oped from A. variabilis sequences indicated the presence of thebidirectional hydrogenase but the absence of the uptake en-zyme in Chroococcidiopsis PCC 7203. In this cyanobacterialstrain, H2 and the bidirectional hydrogenase can support ni-trogenase activity (C2H2 reduction) but only at a rather lowconcentration of 0.3 to 0.5% O2 in the gas phase. Above thatconcentration, O2 is completely inhibitory, presumably by ox-idizing the NiFe center of the enzyme to its inactive oxidizedstate or (less likely) by affecting an extremely O2-sensitivenitrogenase in this organism. In more than 100 different ex-periments performed in air-free vessels, about 50% showed noH2-supported C2H2 reduction activity, whereas the outcomewas positive in the other half. However, the optimal O2 con-centration was 0.3% in one experiment and 0.5% in the next,depending on the concentration of cells in the assay vessels, thephotosynthetic O2 production activity of the cells, and thesuccess in getting the vessels air free. The activity in the pos-itive experiments must come from bidirectional hydrogenase,since any uptake hydrogenase is not so sensitive toward O2. NoC2H2 reduction activity was seen in the dark. The results in-dicate that the bidirectional hydrogenase of ChroococcidiopsisPCC 7203 can only poorly protect nitrogenase from damage byO2. Thus, the bidirectional hydrogenase may be a fossil relicttogether with the organism itself. In early geological times, itmay have served in fermentation and may have effectively

FIG. 9. Chroococcidiopsis sp., as isolated from the gypsum rock “Sachsenstein” near Bad Sachsa, Harz Mountains, Germany (24). Thiscyanobacterium is regarded as a fossil record ancestor of heterocystous cyanobacteria (69) (see the text). It now occupies ecological niches suchas the fissures in gypsum, where it might be exposed to light intensities that are low but still sufficient for photosynthesis. It forms packages of 16cells or multiples thereof (A). The gypsum shards can easily be peeled off by hand (B), and the greenish-blue layer consisting almost exclusivelyof Chroococcidiopsis below the shards then becomes visible (C).



mbr.asm

.org/D

ownloaded from


supplied reducing equivalents to nitrogenase. However, whenthe concentration of O2 in the atmosphere rose above 0.3 to0.5%, bidirectional hydrogenase may have been inactivated.Then, heterocysts that could better accommodate and protecttheir nitrogenase had to be developed. Indeed, Chroococcidi-opsis has been discussed as an ancestor of heterocyst-formingspecies (69).

POTENTIAL FOR EXPLOITING CYANOBACTERIA INSOLAR ENERGY CONVERSION PROGRAMS

FOR PRODUCTION OF COMBUSTIBLEENERGY (HYDROGEN)

Of all organisms, cyanobacteria have the simplest nutrientrequirement in nature. They thrive photoautotrophically onsimple inorganic media, and many of them do not need com-bined nitrogen in their medium. They can be grown with areasonably fast generation time of 2 to 3 h for unicellular forms(though not as fast as fermentative bacteria, such as E. coli,where the half-life [t1/2] can be close to 10 min). A laudablegoal is to generate clean energy, without generating green-house gases such as CO2 or NOx, by exploiting the photosyn-thetically produced reductant (ferredoxin) for H2 production.To do so demands the separation of the photosyntheticallyproduced O2 from H2 production. Research in this area startedaround 1973 during the first global energy crisis and has foundrenewed interest currently due to the concerns over globalwarming. Success in this area demands the continuous produc-tion of H2 over weeks or months, followed by effective utiliza-tion of the cyanobacterial cells produced. Cyanobacterial pro-teins are not optimal to feed to cattle but can be used as dietarysupplements with various positive effects for humans and ani-mals (77, 114). One obstacle is that neither cyanobacterialhydrogenase couples with the reduced ferredoxin generatedphotosynthetically. Presumably based on their own research

interests, different researchers favor the use of either hydro-genase or nitrogenase in solar energy conversion programs.

A comparison of the published rates of H2 formation suffersfrom the fact that different laboratories refer their data todifferent units. As a basis for comparing the various results, thefollowing gross estimates are made (Table 1). In all photosyn-thetic organisms, chlorophyll a constitutes 1 to 2% of the dryweight. Taking the average of 1.5%, the cyanobacterial dryweight can be estimated by multiplying the chlorophyll acontent by a factor of 67 (http://www.chebucto.ns.ca/ccn/info/Science/SWCS/DATA/PARAMETERS/CHA/cha). More-over, chlorophyll a has a molecular weight of slightly less than1,000, and 1 mg of chlorophyll corresponds to 20 to 25 mg ofcell protein (20 mg is used here). In photosynthesis, the unitcommonly used since the time of Willstatter and Stoll (232) ismg chlorophyll per h. The C/N ratio is around 6 in cells, andthe maximal photosynthetic CO2 fixation rates are roughly 100 mol/h � mg chlorophyll. Thus, the N2 fixation rate is unlikelyto exceed 20 mol NH4

� produced/h � mg chlorophyll. If allelectrons transferred to nitrogenase were reallocated to reduceH�, H2 production by cyanobacteria would be around 40 molH2 produced/h � mg chlorophyll, based on the fact that fourelectrons are needed for NH4

� production (with concomitantH2 evolution) but only two electrons are needed for H2 for-mation (equations 1 and 2). The data in Table 1 also use a gasmolar volume of 24 liters at 25°C.

On the basis of the considerations described above, the fewsignificantly higher activities reported in the literature (Table1) seem to require reassessment. If the experiments were notdone with great care, less than total chlorophyll could be re-leased from the cyanobacterial cells, which would lead to thehigh activities reported. With artificial photosystem I/Pt or Aunanoparticle biconjugates, maximal H2 production activitieswere 49 mol/mg chlorophyll � h (85), which are in the same

TABLE 1. Examples of published rates of cyanobacterial H2 formation

Cyanobacterium Enzyme ConditionH2 formation rate

( mol/mgchlorophyll � h)

Reference(s)

Synechocystis PCC 6803 Bidirectional hydrogenase Light � glucose 6 51Arthrospira (Spirulina)

maximaBidirectional hydrogenase Autofermentation, dark, 35°C 0.75–1.5 3

Anabaena azotica Mo nitrogenase C2H2 � H2 9 34V nitrogenase C2H2 � H2 40 34

Anabaena variabilis Mo nitrogenase C2H2 � H2 24 34Nostoc (Anabaena) sp.

PCC 7120Mo nitrogenase hupL/hoxL-deficient mutants 50 139

Anabaena variabilis Mo nitrogenase Immobilized cells, continuously for 5 months 0.06–0.6 134Anabaena variabilis Mo nitrogenase Immobilized cells, sustained production after

extensive CO2 fixation53 137

Anabaena variabilis Mo nitrogenase hupSL mutant 135 92Anabaena variabilis PK84 V nitrogenase Hydrogenase mutant 260 27 (Fig. 3)Anabaena cylindrica Bidirectional hydrogenase Anaerobic incubation, high light intensity 34 119Anabaena variabilis PK84 Mo nitrogenase Hydrogenase-defective isolate 19 126

Synechococcus PCC 7942 Clostridial hydrogenase I Heterologous expression, activity in cell extracts 45 8Wild type Activity in cell extracts 14

Thermosynechococcuselongatus

Uptake hydrogenase Modified PsaE protein of PSI coupled withRalstonia eutropha hydrogenase

0.3 103, 104



mbr.asm

.org/D

ownloaded from


range as the theoretically achievable formation with cyanobac-teria.

To overcome the problem that cyanobacterial hydrogenasesdo not couple with ferredoxin, the clostridial, ferredoxin-de-pendent hydrogenase I was heterologously expressed in theunicellular cyanobacterium Synechococcus PCC 7942 (8). Cellextracts of the genetically engineered isolate showed about3-fold-higher activity than the wild type. An alternative geneticapproach was to modify the photosystem I PsaE subunit fromThermosynechoccocus elongatus so that it linked to the O2-insensitive membrane-bound hydrogenase of Ralstonia eutro-pha and PSI from Synechocystis sp. PCC 6803 (104). This ar-tificial hydrogenase-PSI complex displayed light-driven H2

production, but only at low rates, and this activity was sup-pressed by ferredoxin and FNR (104). The latter problem wascircumvented by modifying the ferredoxin-binding site of PsaE(103). There have been other attempts with limited success toexpress a foreign hydrogenase in cyanobacteria (8) or a cya-nobacterial hydrogenase in a foreign organism (135). Ap-proaches with cyanobacteria are based on the assumption thatthe membrane-bound [NiFe] hydrogenases from Ralstoniaeutropha, R. metallidurans, Allochromatium vinosum, or othersare more O2 tolerant than the cyanobacterial enzymes (64).Since both the bidirectional and uptake hydrogenases of cya-nobacteria have never been biochemically characterized in thepure form, this assumption may not necessarily be true, par-ticularly for the bidirectional hydrogenase. This enzyme, withits complex of five HoxEFUYH subunits, may easily fall apartupon purification, and not necessarily due to any inferred O2

lability. The current state of attempts to develop heterologousand recombinant expression of hydrogenases for improving H2

formation by organisms has been summarized and reviewed(64, 131).

In intact cyanobacterial cells, H2 produced by nitrogenase ismore or less completely recycled by hydrogenase so that oftenalmost no net H2 production is detectable. Uptake hydroge-nase, but not the bidirectional enzyme, is effective in recyclingthe gas (139). Mutants defective in uptake hydrogenase show amuch higher H2 production than wild-type cells. This wasshown some years ago with mutants of Anabaena variabilisobtained by classical N-methyl-N�-nitro-N-nitrosoguanidine(NTG) mutagenesis (147) and more recently with strains thatwere defective in uptake hydrogenase due to site-directed mu-tagenesis (92, 140).

As recently published (34), Anabaena variabilis and A.azotica produce large amounts of H2 when incubated underhigh concentrations of H2 and C2H2 (Fig. 10A). This H2 pro-duction, on top of the H2 added, is higher in V- than inMo-grown cultures of A. azotica (34). The amount of H2

formed increases and C2H4 production decreases in parallelwith the concentration of H2 added to the vessels (Fig. 10B). Inline with these findings, a 2- to 4-fold increase of light-inducedH2 production was observed in Nostoc muscorum preincubatedunder argon and H2 (182). Although added C2H2 is known toinhibit the uptake hydrogenase (205), this observation does notexplain the effect of increasing amounts of H2. The effects ofH2 and C2H2 on nitrogenase itself and/or photosynthetic elec-tron flow to nitrogenase cannot mechanistically be explained asyet. However, the meaning of these findings is that all electronscoming to nitrogenase can be directed to produce H2, partic-

ularly in V-grown cells. The rate of 40 mol H2 producedreflects the maximal photosynthetic H2-forming potential ofcyanobacterial suspension cultures.

Such an interpretation of the data indicates that furthergenetic engineering of cyanobacteria, either by transferring analien hydrogenase or nitrogenase or by genetically manipulat-ing the acceptor side of photosystem I, is unlikely to enhancethe rate of cyanobacterial H2 production. The compilation ofthe data in Table 1 shows that maximal H2 production insuspension cultures is already achieved by coupling either ni-trogenase or hydrogenase to the cyanobacterial photosystem I.A temporal separation of the photosynthetic organic carbonformation (glycogen) in light followed by a fermentative deg-radation of these carbohydrates in the dark (3) is unlikely toenhance H2 production rates, although it would separate H2

and O2 production from each other. Apart from this, rates ofH2 production in strict fermentative bacteria (clostridia) are atleast 3 orders of magnitude higher than those in cyanobacterialfermentations. Therefore, clostridia or other fermentative bac-teria with a much more efficient [Fe-Fe] hydrogenase couldpossibly be coupled and exploited to degrade the cyanobacte-rial photosynthetically produced organic carbon for maximalH2 production.

The transfer of a hydrogenase which is insensitive to expo-sure to O2, either produced by genetic modification or takenfrom an alien organism, may facilitate but may not be obliga-tory for commercially acceptable rates of cyanobacterial H2

production. Genetic alterations of amino acids in the gas-substrate channel of hydrogenases changes their intramolecu-lar gas transport kinetics (121). Substitutions of two aminoacids at the end of the channel (valine and leucine, both withmethionine) make [NiFe] hydrogenase O2 tolerant, as shownfor the enzyme from Desulfovibrio fructosovorans (50). Similargenetic engineering of an [Fe-Fe] hydrogenase could be re-warding, since such an enzyme heterologously transferred tocyanobacteria could couple directly with ferredoxin and thephotosynthetically generated reducing power while being in-sensitive to the photosynthetically produced O2. However, aspointed out previously (64), heterologous expression of anysuch genetically modified hydrogenase in a cyanobacteriumalso requires transcription of host-specific response regulators,and, as outlined above, transcription factors likely show a de-gree of specificity for cyanobacteria, as is evidenced for LexA-and AbrB-like proteins of bidirectional hydrogenase (159) (seeabove).

A realistic chance of improving H2 production by using ei-ther nitrogenase or hydrogenase lies in optimizing the photo-synthetic electron flow for the generation of reductants, asoutlined by the late David Hall and coworkers (90) some yearsago. The light energy conversion efficiencies for H2 productionin suspension cultures are only ca. 1 to 2% and thus very low(136). However, these values refer to the radiant energy inci-dent on the cells rather than the energy absorbed, which isdifficult to determine. These efficiencies can hardly be im-proved in dense cyanobacterial suspension cultures with self-shadowing effects. However, immobilization of cyanobacteriaby adsorption on solid matrices or by entrapment in gels orpolymers may enhance the functional lifetime of cells and mayalso increase the number of heterocysts in filamentous cya-nobacteria. Indeed, immobilized cells were reported to show



mbr.asm

.org/D

ownloaded from


sustained high rates of H2 production (90, 134, 175) (Table 1).The light energy conversion efficiencies for H2 production mayalso be higher in immobilized cells than in suspension cultures.In addition, such an approach may enhance the lifetime of thecyanobacterial cells and thus may result in longer-lasting H2

production (90).Sulfur deprivation leads to inactivation of photosystem II

activity, resulting in anaerobiosis in the cultures and subse-quently enhanced H2 production, as shown first for the greenalga Chlamydomonas reinhardtii (145) and subsequently forcyanobacteria (4, 241). Cyanobacterial H2 production may alsobe augmented by altering the PSII/PSI ratio and by reducingthe content of phycobilisome antennae in the cells (16). Bothcyanobacterial hydrogenases are Ni enzymes. CyanobacterialH2 production could also be altered by the supply of Ni to thecells (2, 10, 164, 169). Limitations could prevent synthesis ofuptake hydrogenase, resulting in higher net H2 productionfrom nitrogenases in the cells. Excess Ni could favor bidirec-

tional hydrogenase synthesis and H2 production by this en-zyme. In addition, culture conditions can be optimized formaximal cyanobacterial H2 production (40, 41).

Activity may also be increased by artificially enhancing thenumber of heterocysts within filaments and thus nitrogenaseconcentrations, e.g., by use of chemicals such as 7-azatrypto-phan (30) or by site-directed mutagenesis (144, 123). A highnumber of 600 to 1,000 genes are estimated to be specificallyexpressed in recently differentiated heterocysts (42, 133). Themaster gene controlling the expression of heterocysts is hetR,and their suppression is regulated by the patS and hetN geneproducts (38, 42, 236). Overexpression of the hetR gene leadsto an enhancement of heterocyst frequency up to 29% inAnabaena (Nostoc) PCC 7120, but the remaining vegetativecells cannot perform CO2 fixation fast enough to meet thedemand of the filaments for organic carbon and reductants(38). Research over the next several years following up in such

FIG. 10. (A) H2 production by Anabaena azotica (V or Mo grown) and A. variabilis. The lower parts of the columns indicate the amount of H2 addedto the vessels by syringes and determined by gas chromatography at the start of the experiments. The gas phase was 85% argon and 15% C2H2 (vol/vol).Complete means gas-phase H2 (about 1 bar). (B) Inhibition of C2H2 reduction by increasing concentrations of H2 added to the assays, using Mo-grownA. azotica. The inhibition pattern was the same for V-grown A. azotica and for Mo-grown A. variabilis (not shown). The data are from reference 34.



mbr.asm

.org/D

ownloaded from


directions will reveal whether cyanobacteria can ever be ex-ploited for the realistic generation of new energies.

ACKNOWLEDGMENTS

We are indebted to Gudrun Boison (Mariefred, Sweden) for helpfuldiscussions and to Stefanie Junkermann (University of Cologne) forexpert technical assistance with some of the experiments.

REFERENCES

1. Almon, H., and P. Boger. 1988. Hydrogen metabolism of the unicellularcyanobacterium Chroococcidiopsis thermalis ATCC 29380. FEMS Micro-biol. Lett. 49:445–449.

2. Almon, H., and P. Boger. 1984. Nickel-dependent uptake-hydrogenase ac-tivity in the blue-green alga Anabaena variabilis. Z. Naturforsch. 39:90–94.

3. Ananyev, G., D. Carrieri, and C. G. Dismukes. 2008. Optimization ofmetabolic capacity and flux through environmental cues to maximize hy-drogen production by the cyanobacterium Arthospira (Spirulina) maxima.Appl. Environ. Microbiol. 74:6102–6113.

4. Antal, T. K., and P. Lindblad. 2005. Production of H2 by sulphur-deprivedcells of the unicellular cyanobacteria Gloeocapsa alpicola and Synechocystissp. PCC 6803 during dark incubation with methane or at various extracel-lular pH. J. Appl. Microbiol. 98:114–120.

5. Antal, T. K., P. Oliveira, and P. Lindblad. 2006. The bidirectional hydro-genase in the cyanobacterium Synechocystis sp. strain PCC 6803. Int. J.Hydrogen Energy 31:114–120.

6. Appel, J., S. Phunpruch, K. Steinmuller, and R. Schulz. 2000. The bidirec-tional hydrogenase of Synechocystis PCC 6803 works as an electron valveduring photosynthesis. Arch. Microbiol. 173:333–338.

7. Appel, J., and R. Schulz. 1998. Hydrogen metabolism in organisms withoxygenic photosynthesis: hydrogenase as important regulatory devices for aproper redox poising? J. Photochem. Photobiol. Biol. 47:1–11.

8. Asada, Y., Y. Koike, J. Schnackenberg, M. Miyake, I. Uemura, and J.Miyake. 2000. Heterologous expression of clostridial hydrogenase in thecyanobacterium Synechococcus PCC7942. Biochim. Biophys. Acta 1490:269–278.

9. Awai, K., and C. Wolk. 2007. Identification of the glycosyl transferase requiredfor synthesis of the the principal glycolipid characteristic of heterocysts ofAnabaena sp strain PCC 7120. FEMS Microbiol. Lett. 266:98–102.

10. Axelsson, R., and P. Lindblad. 2002. Transcriptional regulation of Nostochydrogenases:effects of oxygen, hydrogen and nickel. Appl. Environ. Mi-crobiol. 68:444–447.

11. Battchikova, N., and E. M. Aro. 2007. Cyanobacterial NDH-1 complexes:multiplicity in function and subunit composition. Physiol. Plant. 13:22–32.

12. Bauer, C. C., L. Scappino, and R. Haselkorn. 1993. Growth of the cya-nobacterium Anabaena on molecular nitrogen: NifJ is required when iron islimited. Proc. Natl. Acad. Sci. U. S. A. 90:8812–8816.

13. Belkin, S., and E. Padan. 1978. Hydrogen metabolism in the facultativeanoxygenic cyanobacteria (blue-green algae) Oscillatoria limnetica andAphanothece halophytica. Arch. Microbiol. 116:109–111.

14. Bergman, B., J. R. Gallon, A. N. Rai, and L. J. Stal. 1997. N2 fixation bynon-heterocystouis cyanobacteria. FEMS Microbiol. Rev. 19:139–185.

15. Berman-Frank, I., P. Lundgren, and P. Falkowski. 2003. Nitrogen fixationnad photosynthetic oxygen evolution in cyanobacteria. Res. Microbiol. 154:157–164.

16. Bernat, G., N. Waschewski, and M. Rogner. 2009. Towards efficient hydrogenproduction: the impact of antenna size and external factors on electron trans-port dynamics in Synechocystis PCC 6803. Photosynthesis Res. 99:205–216.

17. Betancourt, D. A., T. M. Loveless, J. Brown, and P. E. Bishop. 2008.Characterization of diazotrophs containing Mo-independent nitrogenases,isolated from diverse natural environments. Appl. Environ. Microbiol. 74:3471–3480.

18. Bishop, P., and R. D. Joerger. 1990. Genetics and molecular biology ofalternative nitrogen fixing systems. Annu. Rev. Plant Physiol. Plant Mol.Biol. 41:109–125.

19. Bishop, P., and R. Premakur. 1992. Alternative nitrogen fixing systems, p.736–762. In G. Stacey, R. H. Burris, and H. J. Evans (ed.), Biologicalnitrogen fixation. Chapman and Hill, New York, NY.

20. Bock, A., P. W. King, M. Blokesch, and M. C. Posewitz. 2006. Maturationof hydrogenases. Adv. Microbiol. Physiol. 51:1–71.

21. Bohme, H. 1998. Regulation of nitrogen fixation in heterocyst-formingcyanobacteria. Trends Plant Sci. 3:346–351.

22. Boison, G., H. Bothe, A. Hansel, and P. Lindblad. 1998. Evidence against acommon use of the diaphorase subunits by the bidirectional hydrogenaseand by respiratory complex I in cyanobacteria. FEMS Microbiol. Lett.37:281–288.

23. Boison, G., H. Bothe, and O. Schmitz. 2000. Transcriptional analysis ofhydrogenase genes in the cyanobacteria Anacystis nidulans and Anabaenavariabilis monitored by RT-PCR. Curr. Microbiol. 40:315–321.

24. Boison, G., A. Mergel, H. Jolkver, and H. Bothe. 2004. Bacterial life anddinitrogen fixation at a gypsum rock. Appl. Environ. Microbiol. 70:7070–7077.

25. Boison, G., O. Schmitz, B. Schmitz, and H. Bothe. 1998. Unusual genearrangement of the bidirectional hydrogenase and functional analysis of itsdiaphorase subunit HoxU in respiration of the unicellular cyanobacteriumAnacystis nidulans. Curr. Microbiol. 36:253–258.

26. Boison, G., C. Steingen, L. J. Stal, and H. Bothe. 2006. The rice fieldcyanobacteria Anabaena azotica and Anabaena sp.CH1 express vanadium-dependent nitrogenase. Archiv. Microbiol. 186:367–376.

27. Borodin, V. B., A. Tsygankov, K. K. Rao, and D. O. Hall. 2000. Hydrogenproduction by Anabaena variabilis PK84 under simulated outdoor condi-tions. Biotechnol. Bioeng. 69:479–485.

28. Bothe, H. 1969. Ferredoxin als Kofaktor der cylischen Photophos-phorylierung in einem zellfreien System aus der Blaualge Anacystis nidu-lans. Z. Naturforsch. 24b:1574–1582.

29. Bothe, H., E. Distler, and G. Eisbrenner. 1978. Hydrogen metabolism inblue-green algae. Biochimie 60:277–289.

30. Bothe, H., and G. Eisbrenner. 1977. Effect of 7-azatryptophan on nitrogenfixation and heterocyst formation in the blue-green alga Anabaena cylin-drica. Biochem. Physiol. Pflanz. 133:323–332.

31. Bothe, H., and G. Eisbrenner. 1981. The hydrogenase-nitrogenase relation-ship in nitrogen-fixing organisms, p. 141–150. In H. Bothe and A. Trebst(ed.), Biology of inorganic nitrogen and sulfur. Springer, Berlin, Germany.

32. Bothe, H., P. Hemmerich, and H. Sund. 1971. Some properties of phytofla-vin isolated from the blue-green alga Anacystis nidulans, p. 211–237. In H.Kamin (ed.), Flavins and flavoproteins. University Park Press-Butterworth,Baltimore, MD.

33. Bothe, H., and G. Neuer. 1988. Electron donation to nitrogenase in het-erocysts. Methods Enzymol. 167:496–501.

34. Bothe, H., S. Winkelmann, and G. Boison. 2008. Maximizing hydrogenproduction by cyanobacteria. Z. Naturforsch. 63c:226–232.

35. Brito, B., C. Baginsky, J. M. Palacios, E. Cabrera, T. Ruiz-Arguesco, and J.Imperial. 2005. Biodiversity of uptake hydrogenase systems from legumeendosymbiotic bacteria. Biochem. Soc. Trans. 33:33–35.

36. Brito, B., A. Toffanin, R. I. Prieto, J. Imperial, T. Ruiz-Arguesco, and R. M.Palacios. 2008. Host-dependent expression of Rhizobium leguminosarum bvviciae hydrogenase is controlled at the transcriptional and posttranslationallevels in legume nodules. Mol. Plant Microbe Interact. 21:597–604.

37. Brusca, J. S., C. D. Carrasco, and J. W. Golden. 1989. Excision of an11-kilobase-pair DNA element from within the nifD gene in Anabaenavariabilis heteterocysts. J. Bacteriol. 171:4138–4145.

38. Buikema, W. J., and R. Haselkorn. 2001. Expression of the Anabaena hetRgene from a copper-regulated promoter leads to heterocyst differentiationunder repressing conditions. Proc. Natl. Acad. Sci. U. S. A. 98:2729–2734.

39. Burns, R. C., and R. W. F. Hardy. 1975. Nitrogen fixation in bacteria andhigher plants. Mol. Biol. Biochem. Biophys. 21:1–189.

40. Burrows, E. H., F. W. R. Chaplen, and R. L. Ely. 2008. Optimization of medianutrient composition for increased photofermentative hydrogen production bySynechocystis sp PCC 6803. Int. J. Hydrogen Energy 33:6092–6099.

41. Burrows, E. H., W. K. Wong, X. Fern, F. W. R. Chaplen, and R. L. Ely. 2009.Optimization of pH and nitrogen for enhanced hydrogen production bySynechocystis sp PCC 6803 via statistical and machine learning methods.Biotechnol. Prog. 25:1009–1017.

42. Callahan, S. M., and W. J. Buikema. 2001. The role of HetN in mainte-nance of the heterocyst pattern in Anabaena sp PCC 7120. Mol. Microbiol.40:941–950.

43. Carrasco, C. D., S. D. Holliday, A. Hansel, P. Lindblad, and J. W. Golden.2005. Heterocyst-specific excision of the Anabaena strain PCC 7120 hupLelement requires xisC. J. Bacteriol. 187:6031–6038.

44. Chakraborty, B., and K. R. Samaddar. 1995. Evidence for the occurrenceof an alternative nitrogenase system in Azospirillum brasilense. FEMS Mi-crobiol. Lett. 127:127–131.

45. Chatterjee, R., R. M. Allen, P. W. Ludden, and V. K. Shah. 1996. Purifica-tion and characterization of the vnf-encoded aponitrogenase from Azoto-bacter vinelandii. J. Biol. Chem. 271:6819–6826.

46. Chatterjee, R., P. W. Ludden, and U. K. Shah. 1997. Characterization ofVNFG, the delta subunit of the vnf-encoded apodinitrogenase fromAzotobacter vinelandii: implications for its role in the formation of func-tional dinitrogenase 2. J. Biol. Chem. 272:3758–3765.

47. Chien, Y.-T., V. Auerbruch, A. D. Brabban, and S. H. Zinder. 2000. Analysisof genes encoding an alternative nitrogenase in the archaeon Methanosar-cina barkeri 227. J. Bacteriol. 182:3247–3253.

48. Cockell, C. S., A. C. Schuerger, D. Billi, E. I. Friedmann, and C. Panitz.2005. Effects of a simulated martian UV flux on the cyanobacterium,Chroococcidiopsis sp. 029. Astrobiology 5:127–140.

49. Cooley, C., C. A. Howitt, and W. F. J. Vermaas. 2000. Succinate:quinoleoxidoreductases in the cyanobacterium Synechocystis sp. PCC 6803: pres-ence and function in metabolism and electron transport. J. Bacteriol. 182:714722.

50. Cournac, L., A. L. De Lacey, A. Volbeda, C. Leger, B. Burlat, M. N., S.Champ, L. Martin, O. Sanganas, M. Haumann, V. M. Fernandez, B.Guigliarelli, J. Fontecilla-Camps, and M. Rousset. 2009. Introduction ofmethionines in the gas channel makes [NiFe] hydrogenase aero-tolerant.J. Am. Chem. Soc. 131:10156–10164.



mbr.asm

.org/D

ownloaded from


51. Cournac, L., G. Guedeney, G. Peltier, and P. M. Vignais. 2004. Sustainedphotoevolution of molecular hydrogen in a mutant of Synechocystis sp.strain PCC 6803 deficient in the type I NADH-dehydrogenase complex. J.Bacteriol. 186:1737–1746.

52. Currati, L., E. Flores, and G. Salerno. 2002. Sucrose is involved in thediazotrophic metabolism of the heterocyst-forming cyanobacteriumAnabaena sp. FEBS Lett. 513:175–178.

53. Deusch, O., G. Landan, M. Roettger, N. Gruenheit, K. V. Kowallik, J. F.Allen, W. Martin, and T. Dagan. 2008. Genes of cyanobacterial origin inplant nuclear genomes point to a heterocyst forming plastid ancestor. Mol.Biol. Evol. 25:748–761.

54. Devine, E., M. Holmquist, K. Stensjo, and P. Lindblad. 2009. Diversity andtranscription of proteases involved in the maturation of hydrogenases inNostoc punctiforme ATCC 29133 and Nostoc sp strain PCC 7120. BMCMicrobiol. 9:1–19.

55. Diez, B., B. Bergman, and R. El-Shehawy. 2009. Marine diazotrophic cya-nobacteria: out of the blue. Plant Biotechnol. 25:221–225.

56. Dilworth, M. J., R. R. Eady, R. L. Robson, and R. W. Miller. 1987. Ethaneformation from acetylene as a potential test for vanadium nitrogenase invivo. Nature 327:167–168.

57. Dilworth, M. J., and R. R. Eady. 1991. Hydrazine is a byproduct of dini-trogen reduction by the vanadium-nitrogenase from Azotobacter chroococ-cum. Biochem. J. 277:465–468.

58. Dixon, R. O. D. 1972. Hydrogenase in legume root nodule bacteroids,occurrence and properties. Arch. Microbiol. 85:193–201.

59. Eady, R. R. 1996. Structure-function relationship of alternative nitrogenase.Chem. Rev. 96:3013–3030.

60. Eilmus, S., C. Rosch, and H. Bothe. 2007. Prokaryotic life in a potash-polluted marsh with emphasis on N-metabolizing microorganisms. Environ.Pollut. 146:478–491.

61. Einsle, O., F. A. Tezcan, A. Andrade, B. Schmid, M. Yoshida, J. B. Howard,and D. C. Rees. 2002. Nitrogenase MoFe-protein at 1.16 Ångstrom reso-lution: a central ligand in the FeMo-cofactor. Science 297:1696–1700.

62. Elhira, S., and M. Ohmori. 2006. NrrA directly regulates expression ofhetR during heterocyst differentiation in the cyanobacterium Anabaena sp.strain PCC 7120. J. Bacteriol. 188:8520–8525.

63. Emerson, S. R., and S. S. Huestedt. 1991. Ocean anoxia and the concen-tration of molybdenum and vanadium in seawater. Mar. Chem. 34:177–196.

64. English, C. M., C. Eckert, K. Brown, M. Seibert, and P. W. King. 2009.Recombinant and in vitro expression systems for hydrogenases. new fron-tiers in basic and applied studies for biological and synthetic H2 production.Dalton Trans. 45:9970–9978.

65. Evans, H. J., A. R. Harker, H. Papen, S. A. Russell, F. J. Hanus, and M.Zuber. 1987. Physiology, biochemistry, and genetics of the uptake hydro-genase in rhizobia. Annu. Rev. Microbiol. 41:335–361.

66. Fay, P. 1992. Oxygen relations of nitrogen fixation in cyanobacteria. Mi-crobiol. Rev. 56:340–373.

67. Ferreira, D., F. A. L. Pinto, P. Morades-Ferreira, M. V. Mendez, and P.Tamagnini. 2009. Transcription profiles of hydrogenase related genes in thecyanobacterium Lyngbya majuscula CCAP 1446/4. BMC Microbiol. 9:67.

68. Ferreira, D., L. J. Stal, P. Moradas-Ferreira, M. V. Mendez, and P. Tama-gnini. 2009. The relation between N2-fixation and H2-metabolism in themarine filamentous nonheterocystous cyanobacterium Lyngbya aestuariiCCY 9616. J. Phycol. 45:896–905.

69. Fewer, D., T. Friedl, and B. Budel. 2002. Chroococcidiopsis and heterocysts-differentiating cyanobacteria are each other’s closest living relatives. Mol.Phylogenet Evol. 23:82–90.

70. Fisher, K., D. J. Lowe, P. Tavares, A. S. Pereira, B. H. Huynh, D. Edmond-son, and W. E. Newton. 2007. Conformations generated during turnover ofthe Azotobacter vinelandii MoFe protein and their relationship to physio-logical function. J. Inorg. Biochem. 101:1649–1656.

71. Floener, L., and H. Bothe. 1980. Nitrogen fixation in Rhopalodia gibba, adiatom containing blue-greenish inclusions symbiotically, p. 541–552. In W.Schwemmler and H. E. A. Schenk (ed.), Endocytobiology, endosymbiosisand cell biology. Walter de Gruyter & Co, Berlin, Germany.

72. Flores, E., A. Herrero, C. P. Wolk, and I. Maldener. 2006. Is the periplasmcontinuous in filamentous cyanobacteria? Trends Microbiol. 14:439–443.

73. Fontecilla-Camps, J., P. Amara, C. Cavazzo, Y. Nicolet, and A. Volbeda.2009. Structure-function relationship of anaerobic gas-processing met-alloenzymes. Nature 460:814–822.

74. Foster, R., and J. Zehr. 2006. Characterization of diatom-cyanobacteriasymbioses on the basis of nifH, hetR and 16S rRNA sequences. Environ.Mirobiol. 8:1913–1925.

75. Friedrich, B., and E. Schwartz. 1993. Molecular biology of hydrogen utili-zation in aerobic chemolithotrophs. Annu. Rev. Microbiol. 47:351–383.

76. Galagan, J. E., C. Nusbaum, A. Roy, et al. 2002. The genome of Methano-sarcina acetivorans reveals extensive metabolic and physiological diversity.Genome 12:532–542.

77. Gantar, M. S. 2008. Microalgae and cyanobacteria: food for thought. J.Phycol. 44:260–268.

78. Garlick, S., A. Oren, and E. Padan. 1977. Occurrence of facultative anoxy-

genic photosythesis among filamentous and unicellular cyanobacteria. J.Bacteriol. 129:623–629.

79. Gebler, A., T. Burgdorf, A. L. De Lacey, O. Rudiger, A. Martinez-Arias, O.Lenz, and B. Friedrich. 2007. Impact of alterations near the [NiFe] activesite on the function of the H2 sensor from Ralstonia eutropha. FEBS J.274:74–85.

80. Geitler, L. 1977. Zur Entwicklungsgeschichte der Epithemiaceae Epi-themia, Rhopalodia und Denticula (Diatomophyceae) und ihre vermutli-chen symbiontischen Sparoidkorper. Plant Syst. Evol. 128:265–275.

81. Ghirardi, M. L., A. Dubini, J. P. Yu, and P. Maness. 2009. Photobiologicalhydrogen-producing systems. Chem. Soc. Rev. 38:52–61.

82. Ghirardi, M. L., M. C. Posewitz, P. C. Maness, A. Dubini, J. Yu, and M.Seibert. 2007. Hydrogenases and hydrogen photoproduction in oxygenicphotosynthetic organisms. Annu. Rev. Plant Biol. 58:71–91.

83. Giddings, J. W., and L. A. Staehelin. 1978. Plasma membrane architectureof Anabaena cylindrica: occurrence of microplasmodesmata and changesassociated with heterocyst development and the cell cycle. Eur. J. Cell Biol.16:235–249.

84. Golden, J. W., and H. S. Yoon. 2003. Heterocyst development in Anabaena.Curr. Opin. Microbiol. 6:557–563.

85. Grimme, R. A., C. E. Lubner, D. A. Bryant, and J. H. Golbeck. 2008.Photosystem/molecular wire/metal nanoparticle biconjugates for the pho-tocatalytic production of H2. J. Am. Chem. Soc. 130:6308–6309.

86. Gupta, M., and N. G. Carr. 1981. Enzyme activities related to cyanophycinmetabolism in heterocysts and vegetative cells of Anabaena spp. J. Gen.Microbiol. 125:17–23.

87. Gutekunst, K., S. Phunpruch, C. Schwarz, S. Schuchardt, R. Schulz-Friedrich,and J. Appel. 2005. LexA regulates the bidirectional hydrogenase in the cya-nobacterium Synechocystis sp. PCC 6803. Mol. Microbiol. 58:810–823.

88. Gutthann, F., M. Egert, A. Marques, and J. Appel. 2007. Inhibition ofrespiration and nitrate assimlation enhances photohydrogen evolution un-der low oxygen concentrations in Synechocystis sp PCC 6803. Biochim.Biophys. Acta 1767:161–169.

89. Hagemann, M., R. Jeanjean, S. Fulda, M. Havaux, F. Joset, and N. Erd-mann. 1999. Flavodoxin accumulation contributes to enhanced cyclic elec-tron flow around photosystem I in salt stressed cells of Synechocystis spstrain PCC 6803. Physiol. Plant. 105:670–678.

90. Hall, D. O., S. A. Markov, Y. Watanabe, and K. K. Rao. 1995. The potentialapplications of cyanobacterial photosythesis for clean technologies. Photo-synth. Res. 46:159–167.

91. Happe, T., A. Hemschemeier, M. Winkler, and A. Kaminski. 2002. Hydro-genases in green algae. Do they save the algae’s life and solve our energyproblems? Trends Plant Sci. 7:246–250.

92. Happe, T., K. Schutz, and H. Bohme. 2000. Transcriptional and mutationalanalysis of the uptake hydrogenase of the filamentous cyanobacteriumAnabaena variabilis. J. Bacteriol. 182:1624–1631.

93. Haselkorn, R. 1992. Developmentally regulated gene rearrangements inprokaryotes. Annu. Rev. Genet. 26:113–130.

94. Haselkorn, R. 2005. Heterocyst differentiation and nitrogen fixation inAnabaena, p. 65–68. In Y.-P. Wang, M. Lin, Z.-X. Tian, C. Elmerich, andW. E. Newton (ed.), Biological nitrogen fixation, sustainable agricultureand the environment, Proceedings of the 14th International Nitrogen Fix-ation Congress. Springer, Dordrecht, Netherlands.

95. Hemrika, W., R. Renirie, H. L. Dekker, P. Barnett, and R. Wever. 1997.From phosphatases to vanadium peroxidases. A similar architecture of theactive site. Proc. Natl. Acad. Sci. U. S. A. 94:2145–2149.

96. Henson, B. J., L. E. Pennington, L. E. Watson, and S. R. Barnum. 2008.Excision of the nifD element in heterocystous cyanobacteria. Arch. Micro-biol. 189:357–366.

97. Herrero, A., A. M. Muro-Pastor, and E. Flores. 2001. Nitrogen control incyanobacteria. J. Bacteriol. 183:411–425.

98. Holmqvist, M., K. Stensjo, P. Oliveira, P. Lindberg, and P. Lindblad. 2009.Characterization of the hupLS promoter activity in Nostoc punctiformeATCC 29133. BMC Microbiol. 9:54.

99. Houchins, J. P. 1984. The physiology and biochemistry of hydrogen metab-olism in cyanobacteria. Biochim. Biophys. Acta 768:227–255.

100. Houchins, J. P., and R. H. Burris. 1981. Comparative characterization oftwo distinct hydrogenases from Anabaena sp. strain 7120. J. Bacteriol.146:215–221.

101. Houchins, J. P., and R. H. Burris. 1981. Occurrence and localization of twodistinct hydrogenases in the heterocystous cyanobacterium Anabaena sp.strain 7120. J. Bacteriol. 146:209–214.

102. Howitt, C. A., and W. F. J. Vermaas. 1997. Analysis of respiratory mutantsof Synechococcus 6803, p. 36. Abstr. IX Symp. Photosynth. Prokaryotes.The Vienna Academy of Sciences, Vienna, Austria.

103. Ihara, M., H. Nakamoto, T. Kamachi, I. Okura, and M. Maeda. 2006b.Light-driven hydrogen production by a hydrid complex of a [NiFe]-hydro-genase and the cyanobacterial photosystem I. Photochem. Photobiol. 82:1677–1685.

104. Ihara, M., H. Nishihara, K.-S. Yoon, O. Lenz, B. Friedrich, H. Nakamoto,K. Kojima, D. Honma, T. Kamachi, and I. Okura. 2006. Light-driven



mbr.asm

.org/D

ownloaded from


production by a hybrid complex of a [NiFe]-Hydrogenase and the cyanobac-terial photosystem I. Photochem. Photobiol. 82:676–682.

105. Ishii, A., and Y. Hihara. 2008. An AbrB-like transcriptional regulator,SII0822, is essential for the activation of nitrogen-regulated genes in Syn-echocystis sp. PCC6803. Plant Physiol. 148:660–670.

106. Jolkver, H. 2005. Untersuchung zur Funktion der bidirektionalen Hydro-genase als nitrogenaseschutzendes Enzym im einzelligen CyanobakteriumChroococcidiopsis. Thesis. The University of Cologne, Cologne, Germany.

107. Jones, K. M., W. J. Buikema, and R. Haselkorn. 2003. Heterocyst-specificexpression of patB, a gene required for nitrogen fixation in Anabaena strainPCC 7120. J. Bacteriol. 185:2306–2314.

108. Kellers, P., H. Ogata, and W. Lubitz. 2008. Purification. crystallization andpreliminary X-ray analysis of the membrane-bound [NiFe] hydrogenasefrom Allochromatium vinosum. Acta Crystallogr. F 64:719–722.

109. Kentemich, T., M. Bahnweg, F. Mayer, and H. Bothe. 1989. Localization ofthe reversible hydrogenase in cyanobacteria. Z. Naturforsch. 44c:384–391.

110. Kentemich, T., M. Casper, and H. Bothe. 1991. The reversible hydrogenasein Anacystis nidulans is a component of the cytoplasmic membrane. Natur-wisssenschaften 78:559–560.

111. Kentemich, T., G. Danneberg, B. Hundeshagen, and H. Bothe. 1988. Evi-dence for the occurrence of the alternative, vanadium-containing nitroge-nase in the cyanobacterium Anabaena variabilis. FEMS Microbiol. Lett.51:19–24.

112. Kentemich, T., G. Haverkamp, and H. Bothe. 1991. The expression of athird nitrogenase in the cyanobacterium Anabaena variabilis. Z. Naturfor-sch. 46c:217–111.

113. Kessler, P. S., J. McLarnan, and J. A. Leigh. 1997. Nitrogenase phylogenyand the molydenum dependence of nitrogen fixation in Methanococcusmaripaludis. J. Bacteriol. 179:541–543.

114. Khan, Z., P. Bhadouria, and P. Bisen. 2005. Nutritional and therapeuticpotential of Spirulina. Curr. Pharm. Biotechnol. 6:373–379.

115. Khudyakov, I. Y., and J. W. Golden. 2004. Different functions of HetR, amaster regulator of heterocyst differentiation in Anabaena sp. PCC 7120,can be separated by mutation. Proc. Natl. Acad. Sci. U. S. A. 101:16040–16045.

116. Kiss, E., P. B. Kos, and I. Vass. 2009. Transcriptional regulation of thebidirectional hydrogenase in the cyanobacterium Synechocystis 6803. J. Bio-technol. 142:31–37.

117. Kitajima, S., F. Hashihama, and S. Takeda. 2009. Latitudinal distributionof diazotrophs and their nitrogen fixation in the tropical and subtropicalwestern North Pacific. Limnol. Oceanogr. 54:537–547.

118. Koene-Cottaar, F. H. M., and G. Schraa. 1998. Anaerobic reduction ofethene to ethane in an enrichment culture. FEMS Microbiol. Ecol. 25:251–256.

119. Laczko, I. 1986. Appearance of a reversible hydrogenase activity inAnabaena cylindrica grown in high light. Physiol. Plant 67:634–637.

120. Leach, C. K., and N. G. Carr. 1971. Pyruvate:ferredoxin oxidoreductase andits activation by ATP in the blue-green alga Anabaena variabilis. Biochim.Biophys. Acta 245:165–174.

121. Leroux, F., S. Dementin, B. Burlat, L. Cournac, A. Volbeda, S. Champ, L.Martin, B. Guigliarelli, P. Bertrand, J. Fontecilla-Camps, M. Rousset, andC. Leger. 2008. Experimental approaches to kinetics of gas diffusion inhydrogenase. Proc. Natl. Acad. Sci. U. S. A. 105:11188–11193.

122. Li, J. H., S. Laurent, V. Konde, S. Bedu, and C. C. Zhang. 2003. An increasein the level of oxoglutarate promotes heterocyst development in the cya-nobacterium Anabaena sp. strain 7120. Microbiol. 149:3257–3263.

123. Liang, C.-M., M. Ekman, and B. Bergman. 2004. Expression of cyanobac-terial genes involved in heterocyst differentiation and dinitrogen fixationalong a plant symbiosis development profile. Mol. Plant Microbe Interact.17:436–443.

124. Lindal, M., and F. J. Florencio. 2003. Thioredoxin-linked processes incyanobacteria are as numerous as in chloroplasts, but targets are different.Proc. Natl. Acad. Sci. U. S. A. 100:16107–16112.

125. Lindblad, P., K. Christensson, P. Lindberg, A. Federov, G. Pinto, and A.Tsygankov. 2002. Photoproduction of H2 by wildtype Anabaena PCC 7120and a hydrogenase deficient mutant: from laboratory experiments to out-door culture. Int. J. Hydrogen Energy 27:1271–1281.

126. Liu, J. G., V. E. Bukatin, and A. A. Tsygankov. 2006. Light energy conver-sion into H2 by Anabaena variabilis mutant PK84 dense cultures exposed tonitrogen limitations. Int. J. Hydrogen Energy 31:1591–1596.

127. Long, M., J. Liu, Z. Chen, B. Bleijlevens, W. Roseboom, and S. P. Albracht.2007. Characterization of a HoxEFUYH type of [NiFe] hydrogenase fromAllochromatium vinosum and some EPR and IR properties of the hydro-genase module. J. Biol. Inorg. Chem. 12:62–78.

128. Lopez-Gomollon, S., J. A. Hernandez, S. Pellicer, V. E. Angarica, M. L.Peleato, and M. F. Fillat. 2007. Cross-talk between iron and nitrogenregulatory networks in Anabaena (Nostoc) PCC 7120: identification of theoverlapping genes in FurA and Ntc regulons. J. Mol. Biol. 374:267–281.

129. Loveless, T. M., and P. E. Bishop. 1999. Identification of genes unique toMo-independent nitrogenase systems in diverse diazotrophs. Can. J. Mi-crobiol. 45:312–317.

130. Loveless, T. M., J. R. Saah, and P. E. Bishop. 1999. Isolation of nitrogen-

fixing bacteria containing molybdenum-independent nitrogenases from nat-ural environments. Appl. Environ. Microbiol. 65:4223–4225.

131. Lubner, C. E., R. A. Grimme, D. A. Bryant, and J. H. Golbeck. 2010. Wiringphotosystem I for direct solar hydrogen production. Biochemistry 49:404–414.

132. Ludwig, M., R. Schulz-Friedrich, and J. Appel. 2006. Occurrence of hydro-genases in cyanobacteria and anoxygenic photosynthetic bacteria: implica-tions for the phylogenetic origin of cyanobacterial and algal hydrogenases.J. Mol. Evol. 63:758–768.

133. Lynn, M. E., J. A. Bantle, and J. D. Ownby. 1986. Estimation of geneexpression in heterocysts of Anabaena variabilis by using DNA-RNA hy-bridization. J. Bacteriol. 167:940–946.

134. Madamwar, D., N. Garg, and V. Shah. 2000. Cyanobacterial hydrogenproduction. World J. Microbiol. Biotechnol. 16:757–767.

135. Maeda, T., G. Vardar, W. T. Self, and T. K. Wood. 2007. Inhibition ofhydrogen uptake in Escherichia coli by expressing the hydrogenase from thecyanobacterium Synechocystis sp. PCC 6803. BMC Biotechnol. 7:25–37.

136. Markov, S. A., M. J. Bazin, and D. O. Hall. 1995b. The potential of usingcyanobacteria in photobioreactors for hydrogen production. Adv. Biochem.Eng. Biotechnol. 52:61–86.

137. Markov, S. A., R. Lichtl, M. J. Bazin, and D. O. Hall. 1995. Hydrogenproduction and carbon dioxide uptake by immobilised Anabaena variabilisin a hollow fibre photobioreactor. Enzyme Microb. Biotechnol. 17:306–310.

138. Masepohl, B., K. Scholisch, K. Gorlitz, C. Kiutski, and H. Bohme. 1997.The heterocyst-specific fdxH gene product of the cyanobacterium Anabaenasp. PCC 7120 is important but not essential for nitrogen fixation. Mol. Gen.Genet. 253:770–776.

139. Masukawa, H., M. Mochimaru, and H. Sakurai. 2002. Disruption of theuptake hydrogenase gene, but not of the bidirectional hydrogenase gene,leads to enhanced photobiological hydrogen production by the nitrogen-fixing cyanobacterium Anabaena sp. PCC 7120. Appl. Microbiol. Biotech-nol. 58:618–624.

140. Masukawa, H., M. Mochimaru, and H. Sakurai. 2002. The hydrogenasesand photobiological hydrogen production utilizing nitrogenase system incyanobacteria. Int. J. Hydrogen Energy 27:1471–1474.

141. Masukawa, H., X. Zhang, E. Yamazaki, S. Iwata, K. Nakamura, M. Mo-chimaru, K. Inoue, and H. Sakurai. 2009. Survey of the distribution ofdifferent types of nitrogenases and hydrogenases in heterocyst-forming cya-nobacteria. Mar. Biotechnol. 11:397–409.

142. Maynard, R. H., R. Premakur, and P. E. Bishop. 1994. Mo-independentnitrogenase 3 is advantageous for diazotrophic growth of Azotobacter vine-landii on solid medium containing molydenum. J. Bacteriol. 176:5583–5586.

143. Mazur-Marzec, H., and M. Plinski. 2009. Do toxic cyanobacteria pose athreat to the Baltic ecosystem? Oceanologia 51:293–313.

144. Meeks, J. C., and J. Elhai. 2002. Regulation of cellular differentiation infilamentous cyanobacteria in free-living and plant-associated symbioticgrowth states. Microbiol. Mol. Biol. Rev. 66:94–121.

145. Melis, A., and T. Happe. 2001. Hydrogen production. Green algae as asource of energy. Plant Physiol. 127:740–748.

146. Meyer, J. 2007. [FeFe] hydrogenases and their evolution: a genomic per-spective. Cell Mol. Life Sci. 64:1063–1084.

147. Mikheeva, L. E., O. Schmitz, S. V. Shestakov, and H. Bothe. 1995. Mutantsof the cyanobacterium Anabaena variabilis altered in hydrogenase activities.Z. Naturforsch 50c:505–510.

148. Mulkidjanian, A. Y., E. V. Koonin, K. S. Makarova, S. L. Mekhedov, A.Sorokin, Y. I. Wolf, A. Dufresne., F. Partensky, H. Burd, D. Kaznadzey, R.Haselkorn, and M. Galperin. 2006. The cyanobacterial genome core andthe origin of photosynthesis. Proc. Natl. Acad. Sci. U. S. A. 103:13126–13131.

149. Mullineaux, C. W., V. Mariscal, A. Nenninger, H. Khanum, A. Herrero, E.Flores, and D. G. Adams. 2008. Mechanisms of intercellular moleculeexchange in heterocyst-forming cyanobaceria. EMBO J. 27:1299–1308.

150. Nakamura, Y., J. Takahashi, A. Sakurai, Y. Inaba, E. Suzuki, S. Nihei, S.Fujiwara, M. Tsuzuki, H. Myashita, H. Ikemoto, M. Kawachi, H. Sekiguchi,and N. Kurano. 2005. Some cyanobacteria synthesize semi-amylopectintype a-polyglucans instead of glycogen. Plant Cell Physiol. 46:539–546.

151. Neuer, G., and H. Bothe. 1982. The pyruvate:ferredoxin oxidoreductase inheterocyts of the cyanobacterium Anabaena cylindrica. Biochim. Biophys.Acta 716:358–365.

152. Neunuebel, M. R., and J. W. Golden. 2008. The Anabaena sp. strain 7120gene all2874 encodes a diguanylate cyclase and is required for normalheterocyst development under high-light growth conditions. J. Bacteriol.190:6829–6838.

153. Newton, W. E. 2007. Physiology, biochemistry and molecular biology ofnitrogen fixation, p. 109–129. In H. Bothe, S. J. Ferguson, and W. E.Newton (ed.), Biology of the nitrogen cycle. Elsevier, Amsterdam, Nether-lands.

154. Ni, C. V., A. F. Yakuninin, and I. N. Gogotov. 1990. Influence of molybde-num, vanadium, and tungsten on growth and nitrogenase synthesis of thefree-living cyanobacterium Anabaena azollae. Microbiology 59:395–398.

155. Oda, Y., S. K. Samanta, F. E. Rey, L. Wu, X. Liu, T. Yan, J. Zhou, and C. S.Harwood. 2005. Functional genomic analysis of three nitrogenase isoen-



mbr.asm

.org/D

ownloaded from


zymes in the photosynthetic bacterium Rhodopseudomonas palustris. J. Bac-teriol. 187:7784–7794.

156. Ohki, K., and Y. Taniuchi. 2009. Detection of nitrogenase in individual cellsof a natural population of Trichodesmium using immunocytochemicalmethods for fluorescent cells. J. Oceanogr. 65:427–432.

157. Oliveira, P., and P. Lindblad. 2008. An AbrB-like protein regulates theexpression of the bidirectional hydrogenase in Synechocystis sp. strain 6803.J. Bacteriol. 190:1011–1019.

158. Oliveira, P., and P. Lindblad. 2005. LexA. A transcriptional regulatorbinding in the promoter region of the bidirectional hydrogenase in thecyanobacterium Synechocystis sp. strain PCC6803. FEMS Microbiol. Lett.251:59–66.

159. Oliveira, P., and P. Lindblad. 2009. Transcriptional regulation of the cya-nobacterial Hox-hydrogenase. Dalton Trans. 45:9990–9996.

160. Olmedo-Verd, E., E. Flores, A. Herrero, and A. M. Muro-Pastor. 2005.HetR-dependent and -independent expression of heterocyst-related genesin an Anabaena strain overproducing the NtcA transcription factor. J.Bacteriol. 187:1985–1991.

161. Olmedo-Verd, E., A. Valladares, E. Flores, A. Herrero, and A. M. Muro-Pastor. 2008. Role of two Ntc-binding sites in the complex ntcA genepromoter of the heterocyst-forming cyanobacterium Anabaena sp. strain7120. J. Bacteriol. 190:7584–7590.

162. Ow, S. Y., T. Cardona, A. Taton, A. Magnuson, P. Lindblad, K. Stensjo, andP. C. Wright. 2008. Quantitative shotgun proteomics of enriched hetero-cysts from Nostoc sp. PCC 7120 using 8-plex isobaric peptide tags. J. Pro-teome Res. 7:1615–1626.

163. Ow, S. Y., J. Noirel, T. Cardona, A. Taton, P. Lindblad, K. Stensjo, andP. C. Wright. 2009. Quantitative overview of N2 fixation in Nostoc puncti-forme ATCC 29133 through cellular enrichments and iTRAQ shotgunproteomics. J. Proteome Res. 8:187–198.

164. Papen, H., T. Kentemich, T. Schmulling, and H. Bothe. 1986. Hydrogenaseactivities in cyanobacteria. Biochimie 68:121–132.

165. Papen, H., G. Neuer, M. Refaian, and H. Bothe. 1983. The isocitratedehydrogenase from cyanobacteria. Arch. Microbiol. 134:73–79.

166. Papen, H., G. Neuer, A. Sauer, and H. Bothe. 1986. Properties of theglyceraldehyde-3-P dehydrogenase in heterocysts and vegetative cells ofcyanobacteria. FEMS Microb. Lett. 36:201–206.

167. Pau, R. N. 1991. The alternative nitrogenases, p. 37–57. In M. J. Dilworthand A. R. Glenn (ed.), Biology and biochemistry of nitrogen fixation.Elsevier, Amsterdam, Netherlands.

168. Pau, R. N., W. Klipp, and S. Steinkuhler. 1997. Molydenum transport,processing and gene regulation, p. 217–234. In G. Winkelmann and C. J.Carrano (ed.), Transition metals in microbial metabolism. Harwood, New-ark, NJ.

169. Pederson, D. M., A. Daday, and G. D. Smith. 1986. The use of Ni to probethe role of hydrogen metabolism in cyanobacterial nitrogen-fixation. Bio-chimie 68:113–120.

170. Petterson-Fortin, L. M., and G. W. Owttrim. 2008. A Synechocystis LexA-orthologue binds direct repeats in target genes. FEBS Lett. 582:2424–2430.

171. Pohorelic, B. K. J., J. K. Voordouw, E. Lojou, A. Dolla, J. Harder, and G.Voordouw. 2002. Effects of deletion of genes encoding Fe-only hydrogenaseof Desulfovibrio vulgaris Hildenborough on hydrogen and lactate metabo-lism. J. Bacteriol. 184:679–686.

172. Pratte, B., K. Eplin, and T. Thiel. 2006. Cross functionality of nitrogenasecomponents NifH1 and VnfH in Anabaena variabilis. J. Bacteriol. 188:5806–5811.

173. Prechtl, J., C. Kneip, P. Lockhart, K. Wenderoth, and U.-G. Maier. 2004.Intracellular spheroid bodies of Rhopalodia gibba have nitrogen-fixing ap-paratus of cyanobacterial origin. Mol. Biol. Evol. 21:1477–1481.

174. Rakhely, G., A. Kovacs, G. Maroti, B. D. Fodor, G. Csanadi, D. Latinovics,and K. L. Kovacs. 2004. Cyanobacterial-type, pentameric, NAD�-reducingNiFe hydrogenase in the purple sulfur photosynthetic bacterium Thiocapsaroseopersicina. Appl. Environ. Microbiol. 70:722–728.

175. Rashid, N., W. Song, J. Park, H. F. Jin, and K. Lee. 2009. Characteristics ofhydrogen production by immobilized cyanobacterium Microcystis aerugi-nosa through cycles of photosynthesis and anaerobic incubation. J. Ind.Eng. Chem. 15:498–503.

176. Raymond, J., J. L. Siefert, C. R. Staples, and R. H. Blankenship. 2004. Thenatural history of nitrogen fixation. Mol. Biol. Evol. 21:541–554.

177. Rees, D. C., F. A. Tezcan, C. A. Haynes, M. A. Walton, A. Andrade, O.Einsle, and J. N. Howard. 2005. Structural basis of biological nitrogenfixation. Philos. Trans. R. Soc. Lond. A 363:971–984.

178. Ribbe, M., D. Gadkari, and O. Meyer. 1997. N2 fixation by Streptomycesthermoautotrophicus involves a molybdenum-dinitrogenase and a manga-nese-superoxide oxidoreductase that couple N2 reduction in the oxidationof superoxide produced from O2 by a molybdenum CO-dehydrogenase.J. Biol. Chem. 272:26627–26633.

179. Rivera-Ortiz, J. M., and R. H. Burris. 1975. Interactions among substratesand inhibitors of nitrogenase. J. Bacteriol. 123:537–545.

180. Rosch, C., and H. Bothe. 2009. Diversity of total, nitrogen-fixing and deni-trifying bacteria in an acid forest soil. Eur. J. Soil Sci. 60:883–894.

181. Ruvkun, G. B., and F. M. Ausubel. 1980. Interspecies homology of nitro-genase genes. Proc. Natl. Acad. Sci. U. S. A. 77:191–195.

182. Scherer, S., W. Kerfin, and P. Boger. 1980. Increase of nitrogenase activityin the blue-green alga Nostoc muscorum (cyanobacterium). J. Bacteriol.144:1017–1023.

183. Schindelin, H., C. Kisker, J. Schlessman, J. B. Howard, and D. C. Rees.1997. Structure of ADP. AlF4

� stabilized nitrogenase complex and itsimplications for signal transduction. Nature 387:370–376.

184. Schink, B. 1982. Isolation of a hydrogenase-cytochrome b complex fromcytoplasmic membarnes of Xanthobacter autotrophicus GZ 29. FEMS Mi-crobiol. Lett. 13:289–293.

185. Schlegel, H. G., and U. Eberhardt. 1972. Regulatory phenomena in themetabolism of Knallgas bacteria. Adv. Microbiol. Physiol. 7:205–242.

186. Schmitz, O., G. Boison, and H. Bothe. 2001. Quantitative analysis of twocircadian clock-controlled gene clusters coding for the birectional hydro-genase in the cyanobacterium Synechoccus sp. PCC7942. Mol. Microbiol.41:1409–1417.

187. Schmitz, O., G. Boison, R. Hilscher, B. Hundeshangen, W. Zimmer, F.Lottspeich, and H. Bothe. 1995. Molecular biological analysis of a direc-tional hydrogenase from cyanobacteria. Eur. J. Biochem. 233:266–276.

188. Schmitz, O., G. Boison, H. Salzmann, H. Bothe, K. Schutz., S. Wang, andT. Happe. 2002. HoxE—a subunit specific for the pentameric bidirectionalhydrogenase complex (HoxEFUYH) of cyanobacteria. Biochim. Biophys.Acta 1554:66–74.

189. Schmitz, O., and H. Bothe. 1996. The diaphorase subunit HoxU of thebidirectional hydrogenase as electron transferring protein in cyanobacterialrespiration? Naturwissenschaften 83:525–527.

190. Schmitz, O., and H. Bothe. 1996. NAD(P)�-dependent hydrogenase activ-ity in extracts from the cyanobacterium Anacystis nidulans. FEMS Micro-biol. Lett. 135:97–101.

191. Schmitz, O., J. Gurke, and H. Bothe. 2001. Molecular evidence for theaerobic expression of nifJ, encoding pyruvate:ferredoxin oxidoreductase, incyanobacteria. FEMS Microb. Lett. 195:97–102.

192. Schmitz, O., T. Kentemich, W. Zimmer, B. Hundeshagen, and H. Bothe.1993. Identification of the nifJ gene coding for pyruvate:ferredoxin oxi-doreductase in dinitrogen-fixing cyanobacteria. Arch. Microbiol. 160:62–67.

193. Schrautemeier, B., U. Neveling, and S. Schmitz. 1995. Distinct and differ-entially regulated Mo-dependent nitrogen-fixing systems evolved for het-erocysts and vegetative cells of Anabaena variabilis ATCC 29413: charac-terization of the fdX1/2 gene regions as part of the nif1/2 gene clusters. Mol.Microbiol. 18:357–359.

194. Schutz, K., T. Happe, O. Troshina, P. Lindblad, E. Leitao, P. Oliveira, andP. Tamagnini. 2004. Cyanobacterial H2-production—a comparative analy-sis. Planta 218:350–359.

195. Seabra, R., A. Santos, S. Pereira, P. Monades-Ferreira, and P. Tamagnini.2009. Immunolocalization of the uptake hydrogenase in the marine cya-nobacterium Lyngbya majuscula CCAP 1446/4 and two Nostoc strains.FEMS Microbiol. Lett. 292:57–62.

196. Seefeldt, L. C., L. G. Dance, and D. R. Dean. 2004. Substrate interactionswith nitrogenase: Fe versus Mo. Biochemistry 43:1401–1409.

197. Serebryakova, L. T., M. E. Sheremetiva, and P. Lindblad. 2000. H2-uptakeand evolution in the unicellular cyanobacterium Chroococcidiopsis thermalisCALU 758. Plant Physiol. Biochem. 38:525–530.

198. Shah, G. R., R. Karunakaran, and G. N. Kumar. 2007. In vivo restrictionendonuclease activity of the Anabaena PCC 7120 XisA protein in Esche-richia coli. Res. Microbiol. 158:679–684.

199. Shestakov, S. V., and L. E. Mikheeva. 2006. Genetic control of hydrogenmetabolism in cyanobacteria. Russian J. Genet. 42:1272–1284.

200. Shi, Y. M., W. X. Zhao, W. Zhang, Z. Ye, and J. D. Zhao. 2006. Regulationof intracellular free calcium concentration during heterocyst differentiationby HetR and NtcA in Anabaena sp PCC 7120. Proc. Natl. Acad. Sci.U. S. A. 103:11334–11339.

201. Shima, S., O. Pilak, S. Vogt, M. Schick, M. S. Stagni, W. Meyer-Klaucke, E.Warkentin, R. K. Thauer, and U. Ermler. 2008. The crystal structure of[Fe]-hydrogenase reveals the geometry of the active site. Science 321:572–575.

202. Short, S. M., and J. P. Zehr. 2005. Quantitative analysis of nifH genes andtranscripts from aquatic environments. Methods Enzymol. 397:380–394.

203. Simpson, F. B., and R. H. Burris. 1984. A nitrogen pressure of 50 atmo-spheres does not prevent evolution of hydrogen by nitrogenase. Science224:105–1097.

204. Simpson, F. B., R. J. Maier, and H. J. Evans. 1979. Hydrogen-stimulatedCO2-fixation and coordinate induction of hydrogenase and ribulosebiphos-phate carboxylase in a H2-uptake positive strain of Bradyrhizobium japoni-cum. Arch. Microbiol. 123:1–8.

205. Smith, L. A., S. Hill, and M. G. Yates. 1976. Inhibition by acetylene ofconventional hydrogenase in nitogen-fixing bacteria. Nature 262:209–210.

206. Stal, L. J. 2009. Is the distribution of nitrogen-fixing cyanobacteria in theoceans related to temperature? Environ. Mirobiol. 11:1632–1645.

207. Stal, L. J., and R. Mozelaar. 1997. Fermentation in cyanobacteria. FEMSMicrobiol. Rev. 21:179–211.



mbr.asm

.org/D

ownloaded from


208. Stanier, R. Y., and G. Cohen-Bazire. 1977. Phototrophic prokaryotes: thecyanobacteria. Annu. Rev. Microbiol. 31:225–274.

209. Stewart, W. D. P., and M. Lex. 1970. Nitrogenase activity in the blue-greenalga Plectonema boryanum. Archiv. Mikrob. 73:250–260.

210. Stripp, S. T., and T. Happe. 2009. How algae produce hydrogen—newsfrom the photosynthetic hydrognase. Dalton Trans. 45:9960–9969.

211. Takeshi, H., K. Ataka, O. Pilak, S. Vogt, M. S. Stagni, W. Meyer-Klaucke,E. Warkentin, R. K. Thauer, S. Shima, and U. Ermler. 2009. The crystalstructure of C176 mutated [Fe]-hydrogenase suggests an acyl-iron ligationin the active site iron complex. FEBS Lett. 583:585–590.

212. Tamagnini, P., R. Axelsson, P. Lindberg, F. Oxelfelt, R. Wunschiers, and P.Lindblad. 2002. Hydrogenases and hydrogen metabolism of cyanobacteria.Microbiol. Mol. Biol. Rev. 66:1–20.

213. Tamagnini, P., J.-L. Costa, L. Almeida, M.-J. Oliveira, R. Salema, and P.Lindblad. 2000. Diversity of cyanobacterial hydrogenases, a molecular ap-proach. Curr. Microbiol. 40:356–361.

214. Tamagnini, P., E. Leitao, P. Oliveira, D. Ferreira, F. A. L. Pinto, D. J.Harris, T. Heidorn, and P. Lindblad. 2007. Cyanobacterial hydrogenasesdiversity, regulation and application. FEMS Microbiol. Rev. 31:692–720.

215. Tel-Or, E., L. W. Luijk, and L. Packer. 1977. An inducible hydrogenase incyanobacteria enhances N2-fixation. FEBS Lett. 78:49–53.

216. Thiel, T. 1993. Characterization of genes for an alternative nitrogenase inthe cyanobacterium Anabaena variabilis. J. Bacteriol. 175:6276–6286.

217. Thiel, T., E. M. Lyons, J. Erker, and A. Ernst. 1995. A second nitrogenasein vegetative cells of a heterocyst-forming cyanobacterium. Proc. Natl.Acad. Sci. U. S. A. 92:9358–9362.

218. Thomas, J.-C., B. Ughy, B. Lagoutte, and G. Ajlani. 2006. A second isoformof ferredoxin:NADP oxidoreductase generated by an in-frame initiation oftranslation. Proc. Natl. Acad. Sci. U. S. A. 103:18368–18373.

219. Thomas, J., J. C. Meeks, C. P. Wolk, P. W. Shaffer, S. M. Austin, and W.-S.Chien. 1977. Formation of glutamine from [13N]ammonia and [13N]dini-trogen, and [14C]glutamate by heterocysts isolated from Anabaena cylin-drica. J. Bacteriol. 129:1545–1155.

220. Thorneley, R. N. F., and D. J. Lowe. 1985. Kinetics and mechanism of thenitrogenase enzyme system, p. 222–284. In T. Spiro (ed.), Molydenumenzymes. J. Wiley, New York, NY.

221. Trebst, A., and H. Bothe. 1966. Zur Rolle des Phytoflavins im photosyn-thetischen Elektronentransport. Ber. Dtsch. Bot. Ges. 79:44–47.

222. Tsygankov, A. 2007. Nitrogen fixing cyanobacteria: a review. Appl. Bio-chem. Microbiol. 43:250–259.

223. Valladares, A., E. Flores, and A. Herrero. 2008. Transcription activation byNtcA and 2-oxoglutarate of three genes involved in heterocyst differentia-tion in the cyanobacterium Anabaena sp. strain PCC 7120. J. Bacteriol.190:126–6133.

224. Van der Oost, J., B. A. Builthuis, S. Feitz, K. Krab, and R. Kraayenhof.1989. Fermentation metabolism of the unicellular cyanobacterium Cyano-thece PCC 7822. Arch. Microbiol. 151:415.419.

225. Van Lin, B., and H. Bothe. 1972. Flavodoxin from Azotobacter vinelandii.Arch. Microbiol. 82:155–172.

226. Vignais, P. M., and B. Billoud. 2007. Occurrence, classification and biolog-ical function of hydrogenases: an overview. Chem. Rev. 107:4206–4272.

227. Vignais, P. M., B. Billoud, and J. Meyer. 2001. Classification and phylogenyof hydrogenases. FEMS Microbiol. Rev. 25:455–501.

228. Vignais, P. M., and A. Colbeau. 2004. Molecular biology of microbialhydrogenases. Curr. Issues Mol. Biol. 6:159–188.

229. Vogt, S., E. J. Lyon, S. Shima, and R. K. Thauer. 2008. The exchangeactivities of [Fe] hydrogenase (iron-sulfur-cluster free hydrogenase) frommethanogenic archaea in comparison with the exchange activities of [FeFe]and [NiFe] hydrogenases. J. Biol. Inorg. Chem. 13:97–106.

230. Walsby, A. E. 2007. Cyanobacterial heterocysts. terminal pores proposed assites of gas exchange. Trends Microbiol. 15:340–349.

231. Weyman, P. D., B. Pratte, and T. Thiel. 2008. Transcription of hupSL inAnabaena variabilis ATTC 29143 is regulated by NtcA and not by hydrogen.Appl. Environ. Microbiol. 74:2103–2110.

232. Willstatter, R., and A. Stoll. 1918. Untersuchungen uber die Assimilationder Kohlensaure. Sieben Abhandlungen. Springer, Berlin, Germany.

233. Wunschiers, R., M. Batur, and P. Lindblad. 2003. Presence and expressionof hydrogenase specific C-terminal endopeptidases in cyanobacteria. BMCMicrobiol. 3:8–20.

234. Yang, T., N. Maeser, M. Laryukhin, H. I. Lee, D. R. Daen, L. C. Seefeldt,and B. M. Hoffman. 2005. The interstitial atom of the nitrogenase FeMo-cofactor: Endor and ESEEM evidence that it is not a nitrogen. J. Am.Chem. Soc. 127:12804–12805.

235. Yates, M. G. 1972. Electron-transport to nitrogenase in Azotobacter chroo-coccum—Azotobacter flavodoxin hydroquinone as an electron donor. FEBSLett. 27:63–67.

236. Yoon, H. S., and J. W. Golden. 1998. Heterocyst pattern formation con-trolled by a diffusible peptide. Science 282:935–938.

237. Zehr, J. P., S. R. Bench, B. J. Carter, I. Hewson, F. Niazi, T. Shi, H. J.Tripp, and J. P. Affourtit. 2008. Globally distributed uncultivated oceanicN2-fxing cyanobacteria lack oxygenic photosystem II. Science 322:1110–1112.

238. Zehr, J. P., S. R. Bench, E. A. Mondragon, J. McCarren, and E. F. DeLong.2007. Low genomic diversity in tropical oceanic N2-fixing cyanobacteria.Proc. Natl. Acad. Sci. U. S. A. 104:17807–17812.

239. Zhang, C. C., S. Laurent, S. Sakr, L. Peng, and S. Bedu. 2006. Heterocystdifferentiation and pattern formation in cyanobacteria. a chorus of signals.Mol. Microb. 59:367–375.

240. Zhang, C. C., H. Pu, Q. S. Wang, S. Cheng, W. X. Zhao, Y. Zhang, and J. D.Zhao. 2007. PII is important in regulation of nitrogen metabolism but notrequired for heterocyst formation in the cyanobacterium Anabaena sp PCC7120. J. Biol. Chem. 282:33641–33648.

241. Zhang, Z., N. D. Pendse, K. N. Phillips, J. B. Cotner, and A. Khodursky.2008. Gene expression patterns of sulfur starvation in Synechocystis sp. PCC6803. BMC Genomics 9:344–354.

242. Zhao, Y., S. M. Bian, H. N. Zhou, and J. F. Huang. 2006. Diversity ofnitrogenase systems in diazotrophs. J. Integrative Plant Biol. 48:745–755.

Hermann Bothe received his Ph.D. fromGottingen University and his habilitationfrom Bochum University. He was Professorof botany and microbiology at the Univer-sity of Cologne, Germany, from 1978 and isnow retired. As a student of A. Trebst, Got-tingen/Bochum, Germany, he started towork on photosynthetic electron transportbefore he switched to nitrogen fixation, bothin cyanobacteria. He also studied aspects ofnitrogen fixation by associative bacteria,denitrification, arbuscular mycorrhiza, and heavy metal and salt resis-tance in plants. He has almost 200 publications in refereed journals.

Oliver Schmitz studied biology in Cologne,Germany, with his main focus on botany, ge-netics, and biochemistry, and completed hisdiploma thesis on arbuscular mycorrhiza in1991. In the course of his dissertation in thelaboratory of Professor Bothe, he specializedin hydrogen metabolism in cyanobacteria andobtained his Ph.D. in 1995 by characterizingthe bidirectional hydrogenase in unicellularand in N2-fixing cyanobacteria by means ofprotein purification and applying molecularbiology, resulting in the first identification of cyanobacterial hydrogenasegenes at that time. He worked as postdoctoral fellow in Susan Golden’sgroup at Texas A&M University, performing research on photosynthesisand the circadian clock in cyanobacteria. In 2001, he joined MetanomicsGmbH, a BASF Plant Science company specialized in applying metabo-lomics in the fields of plant biotechnology, pharmacology, diagnostics, andtoxicology. Currently, he is member of the management team and head ofthe Data Interpretation Health group at Metanomics.



mbr.asm

.org/D

ownloaded from


M. Geoffrey Yates received his B.Sc. fromthe University College of North Wales,Bangor, United Kingdom, and his Ph.D.from the University of Nottingham, UnitedKingdom, and then was Research Associateat Unilever Research Colworth House, Bed-ford, United Kingdom, at the BiochemistryDepartment of John Hopkins University,Baltimore, MD, and then at the Depart-ment of Biochemistry of Oxford University.For almost 30 years, he was Principal Scien-tific Officer at the BBSRC Unit of Nitrogen Fixation, University ofSussex, United Kingdom. For the last 15 years, he was Visiting Re-search Fellow at the Department of Biochemistry and Molecular Bi-ology, Federal University of Parana, Curita, Brazil. In recent years heworked on nitrogen fixation and hydrogen uptake in Azotobacter chroo-coccum, Azospirillum brasilense, and Herbaspirillum seropedicae.

William E. Newton received his B.Sc. fromthe Nottingham University and his Ph.Dfrom London University (both in the UnitedKingdom), and he then spent a postdoctoralyear at Harvard before spending 15 years atthe Charles F. Kettering Research Labora-tory in Yellow Springs, OH, as a member ofits nitrogen fixation group. He then becameResearch Leader for Plant Productivity atthe Western Regional Research Center(USDA-ARS) in Berkeley, CA, where hewas awarded the USDA Certificate of Merit. He also served as Ad-junct Professor at UC-Davis. In 1990, he moved to Virginia Polytech-nic Institute and State University (Virginia Tech) as Director of theBiotechnology Center and Professor of Biochemistry. He later servedas head of both the Biochemistry Department and the Department ofAnaerobic Microbiology. He was elected Fellow of the Royal Societyof Chemistry in 1992 and Fellow of the American Association for theAdvancement of Science in 1996.



mbr.asm

.org/D

ownloaded from


nitrogen fixation and hydrogen metabolism in cyanobacteria · nitrogen fixation and hydrogen...

Documents