Research in Microbiology 154 (2003) 157–164
www.elsevier.com/locate/resmic
Mini-review
Nitrogen fixation and photosynthetic oxygen evolution in cyanobacteria
Ilana Berman-Frank a,∗,1 , Pernilla Lundgren b , Paul Falkowski a,c
a Environmental Biophysics and Molecular Ecology Program, Institute of Marine and Coastal Sciences, Rutgers University, 71 Dudley Rd.,
New Brunswick, NJ 08901, USA
b Department of Botany, Stockholm University, SE-106 91 Stockholm, Sweden
c Department of Geological Sciences, Rutgers University, 71 Dudley Rd., New Brunswick, NJ 08901, USA
Received 29 October 2002; accepted 23 December 2002
First published online 7 January 2003
Abstract
The biological reduction of N2 is catalyzed by nitrogenase, which is irreversibly inhibited by molecular oxygen. Cyanobacteria are
the only diazotrophs (nitrogen-fixing organisms) that produce oxygen as a by-product of the photosynthetic process, and which must
negotiate the inevitable presence of molecular oxygen with an essentially anaerobic enzyme. In this review, we present an analysis of the
geochemical conditions under which nitrogenase evolved and examine how the evolutionary history of the enzyme complex corresponds
to the physiological, morphological, and developmental strategies for reducing damage by molecular oxygen. Our review highlights
biogeochemical constraints on diazotrophic cyanobacteria in the contemporary world.
 2003 Éditions scientifiques et médicales Elsevier SAS. All rights reserved.
Keywords: Cyanobacteria; Nitrogen fixation; Photosynthesis; Nitrogenase; Trichodesmium
1. Introduction
Although N2 is the most abundant gas in Earth’s atmosphere, it is extremely unreactive. Before it can be incorporated into biological molecules, N2 must be chemically
reduced to the equivalent of ammonia. The biological reduction of N2 is catalyzed by a multimeric enzyme complex,
nitrogenase. This enzyme is irreversibly inhibited by molecular oxygen and reactive oxygen species (see summary
in [55]). Oxygen stress on diazotrophic (nitrogen-fixing) organisms triggers a wide range of protective responses aimed
at deterring the inhibitory effects of oxygen on nitrogenase.
The level of resistance to oxygen stress and the mechanisms
involved vary among diazotrophs and influence niche selection. The evolutionary trajectory of adaptive mechanisms
that protect nitrogenase from molecular oxygen and reactive oxygen species can be discerned in physioecological
patterns in microbial morphology, biochemistry, physiology,
and community structure along a gradient from anaerobic to
* Correspondence and reprints.
E-mail address: [email protected] (I. Berman-Frank).
1 Present address: Faculty of Life Sciences, Bar Ilan University, Ramat
Gan, 52900, Israel.
fully aerobic environments. The cyanobacteria are the only
diazotrophs that actually produce oxygen as a by-product of
the photosynthetic process, and which must negotiate the inevitable presence of molecular oxygen with an essentially
anaerobic enzyme. Several recent reviews have focused on
oxygen protection mechanisms and have examined molecular, phylogenetic, physiological, morphological, and regulatory adaptations [1,4,8,26,30,31,73]. Here we briefly examine how, on geological time scales, changing environmental
conditions selected specific adaptations for diazotrophy and
oxygenic photosynthesis in cyanobacteria, and how the evolution of cyanobacteria influenced Earth’s geochemistry.
2. Setting the background
2.1. The structure of nitrogenase
Nitrogenase appears to be a highly conserved enzyme
complex. It is found in a diverse group of prokaryotes from
the Bacteria and Archea [81,82], but is not encoded in any
eukaryotic genome. Phylogenetic analyses suggest all extant
nitrogenases sequenced so far are derived from a single common ancestor and that the catalytic subunits of the enzyme
0923-2508/03/$ – see front matter  2003 Éditions scientifiques et médicales Elsevier SAS. All rights reserved.
doi:10.1016/S0923-2508(03)00029-9
158
I. Berman-Frank et al. / Research in Microbiology 154 (2003) 157–164
complex indicate the enzyme existed prior to oxygenation
of Earth’s atmosphere [10]. In many diazotrophs, nitrogenase comprises about 10% of total cellular proteins and consists of two components, an iron protein (Fe-protein) and an
iron-molybdenum protein (MoFe-protein). The Fe-protein is
a homodimer, composed of a single Fe4 S4 cluster bound between identical ∼32–40 kDa subunits. The Fe4 S4 cluster is
redox-active, and is similar to those found in small molecular weight electron carrier proteins such as ferredoxins [33].
It is the only known redox-active agent capable of obtaining
more than two oxidative states [58] and transfers electrons
to the MoFe-protein [11]. The MoFe-protein is an α2 β2 heterotetramer; the ensemble is approximately 250 kDa. Each
unit contains two types of clusters, a P cluster and a FeMoco
center (sometimes called the M (magnetic) center). The P
cluster is an Fe8 S7 center that functions as a conduit for
electron transfer, accepting electrons from the Fe4 S4 cluster of the iron protein (in conjunction with ATP hydrolysis) and donating them to the FeMoco center (an inorganic
structure of Fe7 Mo1 S9 and an organic component, homocitrate), the site of substrate reduction [33]. Whereas both
the Fe4 S4 and P clusters are inactivated by O2 , the Fe4 S4
cluster is much more susceptible and irreversibly damaged
in vitro [55]. This is reflected by the difference in typical
half-lives of the Fe-protein and MoFe-protein about 45 s and
10 min, respectively [31]. Complicating the situation is the
fact that O2 not only affects the protein structure but also inhibits the synthesis of nitrogenase in many diazotrophs. The
repression is both transient (lasting only a few hours) and
permanent [31]. Moreover, the toxicity of oxygen involves
other forms of reactive oxygen species (ROS) that have both
direct and indirect effects on nitrogenase and other components of the N fixation and assimilation pathways.
Alternative nitrogenases have been found that are homologous to the described enzyme, yet have vanadium or
iron substituting for molybdenum [20]. The catalytic efficiency of these alternative nitrogenases is lower than that
of the MoFe-nitrogenase; the specific activity of the VFenitrogenase is about 1.5 times lower than the MoFe-nitrogenase at 30 ◦ C [45]. In addition to variations in metal cofactors, the nitrogenase complex is non-specific and reduces
triple and double bond molecules other than N2 . These include hydrogen azide, nitrous oxide, acetylene, and hydrogen cyanide [11]. The non-specificity of this enzyme and the
alternative nitrogenases containing other metal co-factors
implicate the role of varying environmental pressures on the
evolutionary history of nitrogenase that could have selected
for different functions of the ancestral enzyme.
2.2. The historical setting
The function(s) of the primitive nitrogenases and the selective pressures on its evolution are debated and depend
on the scenarios defining the early atmosphere and oceans,
specifically the oxidation state and the form and concentration of greenhouse gases [24,70,71]. Physical models of so-
lar radiation [54] suggest that the early atmosphere would
have been mildly reducing with CH4 and CO2 serving as major greenhouse gases [15,39,40,53]. Under such conditions,
several scenarios are proposed. In the first, ammonia could
have been abundant initially and the primitive forms of nitrogenase may have evolved as a respiratory enzyme (N2 being
an accessible electron sink for anaerobic heterotrophs under
the reducing conditions) or a detoxyase utilized in detoxifying cyanides and other prevalent molecules in the ancient
oceans [24,56]. Any increase in ultraviolet radiation would
have caused rapid dissociation of NH3 in the atmosphere
with little fallout of ammonia to the oceans [22,39,40]. With
the loss of free ammonia and cyanides, nitrogenase would
have evolved to become the prevalent biological mechanism
for nitrogen acquisition prior to the oxygenation of the atmosphere and the advent of nitrification [22,24]. Another
possible scenario is that rapid weathering of fragments from
bolide impacts, combined with efficient sequestration of carbon in the mantle, resulted in very low atmospheric CO2
concentrations. These would have limited the amount of
NOx formation from N2 and CO2 , causing a crisis in fixed N
that would select for nitrogen-fixation in the early Archaean
eon ∼3.5 Giga annum before present (Ga) [40].
The change from an anoxic to an oxygenic atmosphere
posed major challenges to cyanobacteria (and in particular
to diazotrophic cyanobacteria). Present evidence based on
paleosols and the isotopic fractionation of sulfur places the
oxygenation of the atmosphere, resulting from cyanobacterial photosynthesis, between 2.4 and 2.2 Ga [25,32,38,66]
(Fig. 1). Oxygenation of the atmosphere resulted in a change
in the partial pressure of oxygen (pO2), from <4 × 10−6
prior to >0.03 atm [54,60] (Fig. 1). With the increase in
atmospheric oxygen, large-scale glaciation is documented,
possibly resulting from the electrophyllic attack on CH4 by
an increased abundance of hydroxyl radicals [39]. These dramatic shifts would have influenced the global sinks of reduced nitrogen, phosphate and metal co-factors required for
organismal growth. Phosphate availability may have limited
primary production due to high adsorption of phosphate by
iron-rich minerals. Consequently, sea-water phosphate concentrations are thought to have been more than tenfold lower
than the present values (0.15–0.6 µM versus the modern
value of 2.3 µM) [7]. Ammonia would be oxidized to nitrite and nitrate, and nitrification pathways would gain importance [23] (Fig. 1). The oxidation of ammonium to nitrate
in the upper ocean and the concomitant hypoxic deep ocean
would have facilitated a denitrification of the ocean, comparable to what is observed at the oxic/anoxic interface in the
contemporary Black Sea.
In addition to the above constraints, the availability of required trace elements such as Fe, Mo, and V would also influence the evolution of nitrogenase. Pyrite-derived prostetic
groups in enzymes are especially vulnerable to damage by
oxygen and reactive oxygen species (ROS) [77]; these lead
to irreversible oxidation of the Fe–S clusters of the Fe4 S4
protein. As described above, the nitrogenase enzyme is an
I. Berman-Frank et al. / Research in Microbiology 154 (2003) 157–164
159
Fig. 1. Evolution of the biogeochemical cycles of oxygen, CO2 , nitrogen (ammonia and nitrate), and the corresponding evolution of the metabolic pathways
of oxygenic photosynthesis and N2 fixation (presence of nitrogenase) in cyanobacteria. Dashed segment for oxygenic photosynthesis indicates the debated
origins of this process in cyanobacteria. Figure is modified from Falkowski [23].
Table 1
Changes in availability of Fe forms (Fe2+ , Fe3+ ) in seawater from the different epochs
Time frame (Myrs)
P O2 (atm)
pH
pE
SO2−
4 (mM)
HS− (µM)
Fe3+ (M)
Fe2+ (M)
Log Fe2+ /Fe3+
Mo (nM)
Archean
Proterozoic
Modern
4000–2500
2500–550
550–0
0.001
7
−3.69
2
100
10−15.7
10−8
7.6
1
0.01
7–8
13.1 (pH 7)–12.1 (pH 8)
2.1
0
10−8
5.8 × 10−18
−9.2
<10
0.2
8.2
13.2
28
0
10−8
4.5 × 10−20
−11.4
100
Fe speciation was calculated using MINEQL+ calculations using above conditions. Assumptions based on Holland (pers. comm.). Mo concentrations were
estimated according to Anbar and Knoll [3].
Fe-rich protein and the availability of iron influences N2
fixation in cyanobacteria from its direct effect on Fe-rich
protein synthesis, to effects on photosynthesis, growth and
global productivity [5,22,51]. In the anaerobic environments
of the Archean oceans, Fe would have been found predominantly in its reduced form (FeII) rather than FeIII (Table 1).
In the Archean and Paleoprotozoic eons, the availability of
Mo as a cofactor for nitrogenase would have been up to
90% lower (Table 1) due to its sequestration in sediments
as insoluble sulfide mineral complexes [3,21,77]. Limited
availability of Mo may have been exacerbated in the midProterozoic, where weathering, under a moderately oxidizing atmosphere, would have enhanced the delivery of sulfate (SO2−
4 ) to the deep ocean. Combined with primary production in the surface waters, this would have resulted in
extremely high H2 S concentrations and removal of Mo via
increased precipitation and formation of active thiomolybdate (MoS2−
4 ) [3]. Alternative nitrogenases (with Fe or V
replacing Mo) most certainly are older forms of the enzyme
complex. Although, presently, Mo-independent nitrogenases
have been found only in heterocystous diazotrophs (which
are phylogenetically the most recently diverging group, see
discussion below) but not in non-heterocystous species [4].
3. Nitrogenase in a changed environment: Multiple
adaptations for a single purpose
Recent phylogenetic analysis of the two pairs of nitrogen
fixation genes nifDK (encoding for alpha and beta subunits
160
I. Berman-Frank et al. / Research in Microbiology 154 (2003) 157–164
of nitrogenase) and nifEN (encoding for the two subunits
of the NifNE protein complex involved in biosynthesis of
the FeMo-cofactor) indicate that the enzyme’s evolutionary
history predates the divergence of the Archaea and bacteria
[24,78]. Fani et al. [24] propose a pathway of evolution for
the ancestors of the nifDK and nifEN operons via two intandem paralogous gene duplication events. The divergence
of a bicistronic operon controlling two genes would be
consistent with either a neutral or mildly reducing Archean
atmosphere.
During the course of planetary evolution, cyanobacteria
have co-evolved with the changing oxidation state of the
ocean and atmosphere to accommodate the machinery of
oxygenic photosynthesis and oxygen-sensitive N2 fixation
within the same cell and/or colony of cells. These strategies
have been generally simplified to distinguish between two
types of adaptations: spatial or temporal separation of photosynthesis and N2 fixation. However, recent research reveals
a more complex pattern that can be traced phylogenetically
and ecologically.
Although the abundance of anaerobic environments
declined with the oxygenation of the atmosphere, some organisms maintained or remained specialized to function in
low oxygen environments. Plectonema boryanum is an example of a well studied filamentous cyanobacteria containing nitrogenase in all cells in equal amounts (Fig. 2) [57].
Plectonema will only fix nitrogen under microaerobic conditions (under 1.5% oxygen in darkness and 0.5% oxygen in
light) [73]. Plectonema is unable to fix nitrogen when grown
anaerobically, reflecting the need of oxygen for respiratory
processes that are a critical source of ATP and reductant [46].
Under a microaerobic environment, Plectonema temporally
separates oxygenic photosynthesis and nitrogen fixation; but
will only do so under the experimentally contrived condition
of continuous illumination [46,49]. Moreover, transcription
of nitrogenase and photosynthetic genes are temporally separated within the photoperiod [50]. Changes in the intracellular ratio of carbon to nitrogen regulate the redox levels of
the secondary quinone acceptor, QB , leading to impairment
of electron transfer between QA and QB [47] and to direct
energy transfer from the phycobilisomes to photosystem I
(PSI) during the period of nitrogen fixation [48]. Other examples of species fixing only under anaerobic or microaerobic conditions (i.e., Plectonema type) include Phormidium,
and some species of Pseudoanabaena [59]. Under contemporary oxygen levels, all of these organisms are relegated to
narrow environmental niches.
A temporal separation without the need for a microaerobic environment is observed for some other non-heterocystous cyanobacteria. These include the filamentous Symploca [42] and Lyngbya majuscula (Lundgren et al., in
press), and the unicellular Gloeothece and Cyanothece
[18,30,42,44] (Fig. 2). These organisms fix nitrogen at night
and nitrogenase is typically found in all cells. In Cyanothece (ATCC51142) nitrogen is fixed during the scotoperiod
when grown under a light:dark cycle, or under the subjec-
Fig. 2. Morphological and behavioral adaptations enabling nitrogen fixation
for different cyanobacteria. Gray shaded areas indicate the localization of
nitrogenase and black solid lines in the graphs designate nitrogen fixation
rates. Photosynthesis is symbolized by the double dashed-line. Efficiency
refers to a comparison of values given in the literature for activity measured
under aerobic conditions (except for group I) as nmol ethylene reduced chl
a −1 h−1 . For some cyanobacteria the ability of nitrogen fixation has been
lost, or was never present (bottom). Others (e.g., Plectonema) are capable of
nitrogen fixation only under microaerobic conditions (I). Other cyanobacteria fix nitrogen during the dark period of a light/dark cycle (II). For 2 groups
a cell specialization occurs (III and IV). Group III includes only two genera: Trichodesmium and Katagnymene. For these, nitrogen fixation occurs
during the light period, and nitrogenase is sequestered into only a fraction
of the cells, often occurring consecutively [6,41,42]. Group IV compartmentalizes all heterocystous cyanobacteria, where nitrogen fixation occurs
mostly during the light period, and where nitrogenase under aerobic conditions is found only in the heterocysts [2,8]. Although, under microaerobic
conditions, nitrogenase can also be present in the other cells [64,66].
tive dark phase when grown under continuous light [72].
High nitrogenase activity coincides with high respiration
rates, and with a phase difference of 12 h from the peak
of photosynthetic activity. This pattern is also reflected at
the transcriptional level, and observed under either continuous light or darkness, implicating circadian control [18,62].
Interestingly, psaB transcription, translation, and PSI activity increase towards the end of the photoperiod and at the
beginning of the scotoperiod, indicating a higher capacity
for cyclic photophosphorylation in the light (supplying more
ATP) at the end of the photoperiod [72]. The marine origins
of Cyanothece may be traced to the recently described marine unicellular isolates that expressed nifH genes and were,
phylogenetically, most closely related to nifH sequences
from Cyanothece, Myxosarcina, and Gloeothece [82]. The
distribution and abundance of these marine unicellular diazotrophs are presently poorly characterized. Yet accruing
evidence indicates that this fraction of cyanobacteria (cell
sizes 3–10 µm) may be more important contributors to global
N2 fixation than previously recognized, supplying up to
100 Tg N y−1 (Falcon and Carpenter pers. comm.) [12,82].
Yet another adaptation is seen for the marine nonheterocystous filamentous cyanobacteria Trichodesmium and
Katagnymene (Katagnymene has been reaffiliated to Trichodesmium [44]) (Fig. 2). Unlike all other non-heterocys-
I. Berman-Frank et al. / Research in Microbiology 154 (2003) 157–164
tous species of cyanobacteria, these species fix nitrogen during the day and their mechanism of protection from photosynthetically evolved oxygen has puzzled researchers for
many years [13]. In Trichodesmium, nitrogenase is compartmentalized in a fraction of the cells (typically between 10 and 20%) that are often arranged consecutively
along the trichome [43]. However, active photosynthetic
components (such as PSI and PSII complexes, Rubisco,
carboxysomes) are found in all cells, even those harboring
nitrogenase [6,28,29,34]. Hence, spatial aggregation of nitrogenase into certain zones/cells is not sufficient to protect
against oxygen evolution. Indeed, there is evidence that a
significant fraction of nitrogenase may be inhibited by O2
at any moment in time. Protection against oxygen in Trichodesmium was found recently to be a complex interaction
between spatial and temporal segregation of the photosynthetic, respiratory and nitrogen fixation processes [6,17,43].
In Trichodesmium, a circadian clock controls the transcription of nitrogenase and the expression of photosynthetic
genes essential to the activation of photosystems I and II
(PSII) [16]. These genes (psaA, psbA) undergo a temporal phase difference with nifHDK expression of ∼6 h, with
nifHDK expression followed by psbA psaA [17]. Parallel
measurements of changes in the activity of PSII, oxygen
production, carbon, and N2 fixation reveal a temporal separation between N2 fixation and photosynthesis during the
photoperiod. The period of high N2 fixation is characterized by a decline in gross photosynthetic production as well
as enhanced light-dependent oxygen consumption and scavenging of oxygen by PSI via the Mehler reaction [36,37]
(Gerchman et al., in preparation), which results in a negative
net production of oxygen [6]. Cells can turn photosynthetic
activity (seen as changes in cellular variable fluorescence)
on or off within 10–15 min, illustrating that the photosynthetic machinery is not physically damaged. In contrast to
fully evolved heterocystous cyanobacteria, all cells in a trichome are photosynthetically competent but individual cells
modulate oxygen production and consumption during the
photoperiod. Moreover, the increased occurrence of inactive
photosynthetic zones during the hours of high N2 fixation
provides evidence of both a temporal and spatial segregation
of the two processes [6].
A highly refined specialization is found in heterocystous cyanobacteria (Fig. 2). Here, nitrogenase is confined
to a micro-anaerobic cell, the heterocyst, which differentiates completely and irreversibly 12–20 h after combined nitrogen sources are removed from the medium. These cells
are characterized by a thick membrane that slows the diffusion of O2 , high PSI activity, the absence of PSII, and
loss of division capability. Heterocystous organisms cannot obtain reductant directly from non-cyclic photosynthetic
electron flow and rely on the supply of fixed carbon from
adjacent vegetative cells for reducing equivalents. Cyclic
electron flow around PSI supplies ATP (see recent comprehensive reviews on the formation and regulation of the heterocysts [1,2,8]). It was long thought that (a) heterocysts were
161
required for nitrogen fixation, and that (b) the trigger for
the formation of these specialized cells was N-deficiency.
However, under anaerobic conditions some heterocystous
cyanobacteria, such as Anabaena variabilis, can synthesize
a different Mo-dependent nitrogenase (Nif2) in the vegetative cells [61,68,69]. Nif2 is expressed shortly after nitrogen
depletion but prior to heterocyst formation, and can support
the fixed N needs of the filaments independently of Nif1 nitrogenase in the heterocysts [67,69]. Nif2 is also found in
vegetative cells of non-heterocystous species [61]. Interestingly, a gene that is part of the nif2 cluster in Anabaena variabilis ATCC 29413, fdxH2 (encoding the [2Fe–2S] ferredoxin known as the direct electron donor to nitrogenase
in heterocysts), has more residues in common and shares
its oxygen sensitivity with the single FdxH from the nonheterocystous, filamentous cyanobacterium Plectonema boryanum PCC 73110 [65].
4. Evolutionary history of adaptive strategies
for nitrogen fixation
The diverse strategies to overcome the oxygen dilemma
reflect the wide-ranging flexibility and niches occupied by
cyanobacteria: From anaerobic sediments and pore-waters
to pelagic waters saturated with oxygen. It is also plausible to assume that the evolution of different strategies
has been brought on by the increased oxygen pressure.
While some paleontological evidence suggested that nonheterocystous cyanobacteria were present at 3.5 Ga bp [64],
the validity of these results is questioned [9]. The earliest evidence of cyanobacteria is based on fossil lipid biomarkers, [66] that date cyanobacteria to at least 2.85 Ga bp.
Evidence for heterocystous forms is found in akinete fossils dated at 1.5 to 2 Ga bp from cherts in West Africa
and Siberia and silicified carbonates from Australia [63].
This fossil reconstruction is reflected in phylogenetic trees
using functional genes involved in nitrogen fixation, such
as nifH [76,80]. We constructed an unrooted phylogenetic
tree of cyanobacterial representatives from the different
adaptive strategies (Fig. 3). This tree suggests that adaptations evolved from anaerobic diazotrophs (i.e., Phormidium, Lyngbya lagerheimii, Pseudanabaena PCC7403) to microaerobic diazotrophs such as the filamentous Plectonema
boryanum in which all cells can fix nitrogen with strict temporal separation from photosynthetic oxygen evolution only
under continuous illumination. A full temporal separation,
where nitrogen is only fixed at night, developed in unicellular cyanobacterial diazotrophs (Cyanothece ATC51142,
Gloeothece, Synechocystis sp. WH8501) and some nonheterocystous filamentous diazotrophs (Symploca, see Ref.
above). Further diversity in adaptation strategies also exists
within the unicellular diazotrophs. In the genus Cyanothece,
for example, the seven described strains could fix nitrogen
either aerobically or anaerobically [75]. Recent molecular
phylogenies, for 33 strains of unicellular N2 fixers using
162
I. Berman-Frank et al. / Research in Microbiology 154 (2003) 157–164
small subunit ribosomal RNA sequences, showed no correlation between the phylogenetic relationships and the type
of N2 fixation [74]. Moreover, the authors hypothesize that
the results of the phylogeny are indicative of multiple gains
and/or losses of N2 fixation abilities among the unicellular cyanobacteria [74]. Thus, the loss of cyanobacterial nif
genes could suggest that different strategies existed early,
where some organisms were able to adapt to an oxic world,
while others were not (e.g., Oscillatoria).
Tracing the phylogeny of cyanobacteria using nifH is
problematic. The reconstructed trees are very sensitive to
small changes in sequence. For example, Trichodesmium’s
unique biological solution of combining a semi-temporal
separation of N2 fixation and photosynthesis with spatial
heterogeneity appears in several trees as an early branch,
suggesting a very ancient past [76,80,82]. Yet, in other trees
(Fig. 3) it shows a later divergence. Moreover, organisms
with the same strategy (e.g., Lyngbya and Cyanothece) do
not always cluster together (Fig. 3). Whether this is due
to parallel evolution or is an artifact of the phylogenetic
reconstruction remains to be elucidated.
It is noteworthy, however, that the most recent phylogenetic branching (both for nifH and RNA trees) is observed for filamentous species where complete segregation
of N2 fixation and photosynthesis was achieved with the
evolution of heterocystous cyanobacteria [79]. Heterocystous cyanobacteria are predominantly terrestrial, fresh wa-
Fig. 3. Phylogenetic analysis of nifH. A maximum-likelihood tree was
constructed using fast DNAml [27]. The scale represents the expected
number of changes per sequence position. The numbers depict bootstrap
values obtained for a bootstrap sampling of 100. Numbers at the side
correspond to the different N2 -fixation strategies as described in Fig. 2.
nifH sequences were aligned using the Genetic Data Environment (GDE)
multiple sequence editor. 324 homologous nucleotides were used to infer
the phylogenetic position of the nifH genes. Evolutionary distances were
computed from pair-wise similarities by using the correction of Jukes
and Cantor [35]. Distance trees were constructed by the least-squares
algorithm of DeSoete [19] from a normal evolutionary distance matrix.
Maximum-likelihood trees were constructed using fast DNAml, which uses
the generalized two-parameters model of evolution [41], and using jumbled
orders for the addition of taxa to avoid potential bias introduced by the order
of sequence addition. Bootstrap analysis was used to provide confidence
estimates for phylogenetic tree topologies.
ter and coastal species, inhabiting eutrophic or brackish
environments (e.g., Baltic Sea), with a few epiphytic and
symbiotic representatives in the marine environment [52].
Very few heterocystous free-living species are found in the
pelagic oceans though recently a novel species was found,
designated Anabaena gerdii [14]. Relative to oceanic environments, fresh-water and coastal systems undergo more
frequent and more extreme environmental changes so that
the selective pressure on organisms in those environments
would be higher and evolutionary adaptations more rapid,
leading to the dominance of heterocystous cyanobacteria in
these systems. In the large, slowly undulating world of the
oceans, the ancient pathway adapted by Trichodesmium has
persisted throughout the present time. This persistence suggests the tempo of evolution in marine cyanobacteria is extremely slow. Moreover, the relative evolutionary stability of
nitrogenase across species reflects the flexibility of the enzyme in appropriating a variety of trace-metals as prosthetic
groups, and its response to the inhibitory effects of molecular oxygen.
5. Future challenges—prospects
The nitrogen cycle of Earth is one of the most critical yet
poorly understood biogeochemical cycles. Current estimates
of global N2 fixation are ∼240 Tg N y−1 with a marine
contribution of 100–190 Tg N y−1 . Of this, a single nonheterocystous genus, Trichodesmium sp. contributes approximately ∼100 Tg N y−1 (Capone pers. comm.). Geochemical
evidence suggests that, on a global scale, nitrogen fixation
does not always keep pace with denitrification on time scales
of centuries to millenia [23], yet it remains unclear what
process(es) limits nitrogen fixation in the oceans. More importantly, given the potential for heterocystous cyanobacteria to outcompete organisms such as Trichodesmium, it is
unclear why the apparent tempo of evolution of marine diazotrophic cyanobacteria is so slow. Diazotrophic cyanobacteria have effectively become the “gate keepers” of oceanic
productivity, yet despite the rapid radiation of eukaryotic
oxygenic photoautotrophs throughout the Phanaerozoic eon,
marine cyanobacteria seem like living fossils. While molecular analyses and functional genomics may shed some light
on this apparent paradox, we suggest that ultimately the answer lies in an understanding of cyanobacterial physiological ecology, which must itself be viewed from a paleoecological perspective. We are only just beginning to understand
how biogeochemical cycles and microbes co-evolved on the
planet.
Acknowledgements
Many thanks to John Reinfelder for help with the
MINEQL+ calculations of iron speciation in Table 1, to
I. Berman-Frank et al. / Research in Microbiology 154 (2003) 157–164
Felisa Wolfe and Costantino Vetriani for generating and analyzing the phylogenetic tree (Fig. 3), Y. Ziv for graphic help,
and J. Kasting and anonymous reviewers for critical comments on the manuscript. Support for this research was provided by Rutgers University as a postdoctoral research fellowship (to IBF), and by the National Science Foundation
through a Biocomplexity grant to PGF (OCE 0084032).
References
[1] D.G. Adams, Heterocyst formation in cyanobacteria, Curr. Opin.
Microbiol. 3 (2000) 618–624.
[2] D.G. Adams, P.S. Duggan, Heterocyst and akinete differentiation in
cyanobacteria, New Phytol. 144 (1999) 3–33.
[3] A.D. Anbar, A.H. Knoll, Proterozoic ocean chemistry and evolution:
A bioinorganic bridge, Science 297 (2002) 1137–1142.
[4] B. Bergman, et al., N2 fixation by non-heterocystous cyanobacteria,
FEMs Microbiol. Rev. 19 (1997) 139–185.
[5] I. Berman-Frank, et al., Iron availability, cellular iron quotas, and
nitrogen fixation in Trichodesmium, Limnol. Oceanogr. 46 (2001)
1249–1260.
[6] I. Berman-Frank, et al., Segregation of nitrogen fixation and oxygenic photosynthesis in the marine cyanobacterium Trichodesmium,
Science 294 (2001) 1534–1537.
[7] C.J. Bjerrum, D.E. Canfield, Ocean productivity before about 1.9 Gyr
ago limited by phosphorus adsorption onto iron oxides, Nature 417
(2002) 159–162.
[8] H. Bohme, Regulation of nitrogen fixation in heterocyst-forming
cyanobacteria, Trends Plant Sci. 3 (1998) 346–351.
[9] M.D. Brasier, et al., Questioning the evidence for Earth’s oldest fossils,
Nature 416 (2002) 76–81.
[10] E. Broda, G.A. Peschek, Nitrogen fixation as evidence for the reducing
nature of the early atmosphere, Biosystems 16 (1983) 1–8.
[11] B.K. Burgess, D.J. Lowe, Mechanism of molybdenum nitrogenase,
Chem. Rev. 96 (1996) 2983–3011.
[12] D.G. Capone, J.P. Montoya, Nitrogen fixation and denitrification,
Meth. Microbiol. 30 (2001) 501–515.
[13] D.G. Capone, et al., Trichodesmium, a globally significant marine
cyanobacterium, Science 276 (1997) 1221–1229.
[14] E.J. Carpenter, S. Janson, Anabaena gerdii sp. nov., a new planktonic
filamentous cyanobacterium from the South Pacific Ocean and Arabian Sea, Phycologia 40 (2001) 105–110.
[15] D.C. Catling, K.J. Zahnle, C. McKay, Biogenic methane, hydrogen
escape, and the irreversible oxidation of early Earth, Science 293
(2001) 839–843.
[16] Y.B. Chen, et al., Circadian rhythm of nitrogenase gene expression
in the diazotrophic filamentous nonheterocystous Cyanobacterium
Trichodesmium sp. strain IMS101, J. Bacteriol. 180 (1998) 3598–
3605.
[17] Y.B. Chen, et al., Expression of photosynthesis genes in relation to
nitrogen fixation in the diazotrophic filamentous nonheterocystous
cyanobacterium Trichodesmium sp. IMS 101, Plant Mol. Biol. 41
(1999) 89–104.
[18] M. Colon-Lopez, D.M. Sherman, L.A. Sherman, Transcriptional and
translational regulation of nitrogenase in light-dark- and continuouslight-grown cultures of the unicellular cyanobacterium Cyanothece sp.
strain ATCC 51142, J. Bacteriol. 179 (1997) 4319–4327.
[19] G. DeSoete, A least squares algorithm for fitting additive trees to
proximity data, Psychometrica 48 (1983) 621–626.
[20] R.R. Eady, Structure function relationships of alternative nitrogenases,
Chem. Rev. 96 (1996) 3013–3030.
[21] P.G. Falkowski, Biogeochemical cycles, in: Encyclopedia of Biodiversity, Academic Press, 2001, pp. 437–453.
163
[22] P.G. Falkowski, Evolution of the nitrogen cycle and its influence on
the biological sequestration of CO2 in the ocean, Nature 387 (1997)
272–275.
[23] P.G. Falkowski, J.A. Raven, Aquatic Photosynthesis, 1 Edition, Blackwell Science, Oxford, 1997, p. 375.
[24] R. Fani, R. Gallo, P. Lio, Molecular evolution of nitrogen fixation: The
evolutionary history of the nifD, nifK, nifE, and nifN genes, J. Mol.
Evol. 51 (2000) 1–11.
[25] J. Farquhar, H. Bao, M. Thiemens, Atmospheric influence of Earth’s
earliest sulfur cycle, Science 289 (2000) 756–758.
[26] P. Fay, Oxygen relations of nitrogen fixation in cyanobacteria, Microbiol. Rev. 56 (1992) 340–373.
[27] J. Felsenstein, PHYLIP, University of Washington, Seattle, 1993.
[28] C. Fredriksson, B. Bergman, Ultrastructural characterisation of cells
specialised for nitrogen fixation in a non-heterocystous cyanobacterium, Trichodesmium spp., Protoplasma 197 (1997) 76–85.
[29] C. Fredriksson, et al., Aerobic nitrogen fixation is confined to a subset
of cells in the non-heterocystous cyanobacterium Symploca PCC 8002,
New Phytol. 140 (1998) 531–538.
[30] J.R. Gallon, N2 fixation in phototrophs: Adaptation to a specialized
way of life, Plant Soil 230 (2001) 39–48.
[31] J.R. Gallon, Reconciling the incompatible: N2 fixation and oxygen,
New Phytol. 122 (1992) 571–609.
[32] H.D. Holland, Geochemistry—evidence for life on Earth more than
3850 million years ago, Science 275 (1997) 38–39.
[33] J.B. Howard, D.C. Rees, Structure of the nitrogenase protein components, in: E.W. Triplett (Ed.), Prokaryotic Nitrogen Fixation: A Model
System for the Analysis of a Biological Process, Horizon Scientific
Press, Norfolk, 2000, p. 800.
[34] S. Janson, E.J. Carpenter, B. Bergman, Compartmentalization of
nitrogenase in a non-heterocystous cyanobacterium—Trichodesmium
contortum, FEMS Microbiol. Lett. 118 (1994) 9–14.
[35] T.H. Jukes, C.R. Cantor, Evolution of protein molecules, in: H.N.
Munro (Ed.), Mammalian Protein Metabolism, Academic Press, New
York, 1969, pp. 21–132.
[36] T.M. Kana, Oxygen cycling in cyanobacteria with specific reference
to oxygen protection in Trichodesmium spp., in: E.J. Carpenter (Ed.),
Marine Pelagic Cyanobacteria: Trichodesmium and other Diazotrophs,
Kluwer Academic, 1992, pp. 29–41.
[37] T.M. Kana, Rapid oxygen cycling in Trichodesmium thiebautii,
Limnol. Oceanogr. 38 (1993) 18–24.
[38] J.F. Kasting, The rise of atmospheric oxygen, Science 293 (2001) 819–
820.
[39] J.F. Kasting, J.L. Siefert, Life and the evolution of Earth’s atmosphere,
Science 296 (2002) 1066–1068.
[40] J.F. Kasting, J.L. Siefert, The nitrogen fix, Nature 412 (2001) 26–27.
[41] H. Kishino, M. Hasegawa, Evaluation of the maximum likelihood
estimate of the evolutionary tree topologies from DNA sequence data,
and the branching order in Hominoidea, J. Mol. Evol. 29 (1989) 170–
179.
[42] S. Kumazawa, S. Yumura, H. Yoshisuji, Photoautotrophic growth of
a recently isolated N2 -fixing marine non-heterocystous filamentous
cyanobacterium, Symploca sp. (Cyanobacteria), J. Phycol. 37 (2001)
482–487.
[43] S.J. Lin, et al., Whole-cell immunolocalization of nitrogenase in marine diazotrophic cyanobacteria, Trichodesmium spp., Appl. Environ.
Microbiol. 64 (1998) 3052–3058.
[44] P. Lundgren, et al., Katagnymene: Characterization of a novel marine
diazotroph, J. Phycol. 37 (2001) 1052–1062.
[45] R.W. Miller, R.R. Eady, Molybdenum and vanadium nitrogenases of
Azotobacter chroococcum. Low temperature favours N2 reduction by
vanadium nitrogenase, Biochem. J. 256 (1988) 429–432.
[46] H.S. Misra, Oxygen implication in the diazotrophic growth of Plectonema boryanum in dark-light cycles, Plant Sci. 143 (1999) 135–142.
[47] H.S. Misra, T.S. Desai, Involvement of acceptor side components of
PSII in the regulatory mechanism of Plectonema boryanum grown
photoautotrophically under diazotrophic condition, Biochem. Biophys. Res. Commun. 194 (1993) 1001–1007.
164
I. Berman-Frank et al. / Research in Microbiology 154 (2003) 157–164
[48] H.S. Misra, S.K. Mahajan, Excitation energy transfer from phycobilisomes to photosystems: A phenomenon associated with the temporal separation of photosynthesis and nitrogen fixation in a cyanobacterium, Plectonema boryanum, Biochim. Biophys. Acta 1459 (2000)
139–147.
[49] H.S. Misra, R. Tuli, Uncoupling of photosystems during lightdependent dinitrogen fixation by a nonheterocystous cyanobacterium
Plectonema boryznum, Ind. J. Biochem. Biophys. 31 (1994) 310–314.
[50] H.S. Misra, R. Tuli, Differential expression of photosynthesis and
nitrogen fixation genes in the cyanobacterium Plectonema boryanum,
Plant Physiol. 122 (2000) 731–736.
[51] H.W. Paerl, L.E. Prufert-Bebout, C. Guo, Iron-stimulated N2 fixation
and growth in natural and cultured populations of the planktonic marine cyanobacteria Trichodesmium spp., Appl. Environ. Microbiol. 60
(1994) 1044–1047.
[52] H.W. Paerl, J. Zehr, Marine nitrogen fixation, in: D.L. Kirchman (Ed.),
Microbial Ecology of the Oceans, Wiley–Liss, 2000, pp. 387–427.
[53] A.A. Pavlov, et al., Greenhouse warming by CH4 in the atmosphere of
early Earth, J. Geophys. Res. 105 (2000) 11981–11990.
[54] A.A. Pavlov, J.F. Kasting, Mass-independent fractionation of sulfur
isotopes in archean sediments: Strong evidence for an anoxic archean
atmosphere, Astrobiology 2 (2002) 27–41.
[55] J.R. Postgate, Nitrogen Fixation, 3 Edition, Cambridge University
Press, Cambridge, 1998, p. 112.
[56] J.R. Postgate, R.R. Eady, The evolution of biological nitrogen fixation,
in: H. Bothe, F.J. DeBruijn, W.E. Newton (Eds.), Nitrogen Fixation:
One Hundred Years After, Fischer, Stuttgart, 1988, pp. 31–40.
[57] A.N. Rai, M. Borthakur, B. Bergman, Nitrogenase derepression,
its regulation and metabolic changes associated with diazotrophy
in the non-heterocystous cyanobacterium, Plectonema boryanum
PCC73110, Gen. Microbiol. 138 (1992) 481–491.
[58] D.C. Rees, J.B. Howard, Nitrogenase: Standing at the crossroads, Curr.
Opin. Chem. Biol. 4 (2000) 559–566.
[59] R. Rippka, J.B. Waterbury, The synthesis of nitrogenase by nonheterocystous cyanobacteria, FEMS Microbiol. Lett. 2 (1977) 83–86.
[60] R. Rye, H.D. Holland, Paleosols and the evolution of atmospheric
oxygen: A critical review, Am. J. Sci. 298 (1998) 621–672.
[61] W. Schiefer, et al., Molecular cloning and characterization of hetR
genes from filamentous cyanobacteria, BBA—Gene Struct. Expr. 1577
(2002) 139–143.
[62] M.A. Schneegurt, et al., Metabolic rhythms of a diazotrophic
cyanobacterium, Cyanothece sp. strain ATCC 51142, heterotrophically grown in continuous dark, J. Phycol. 36 (2000) 107–117.
[63] J.W. Schopf, The fossil record: Tracing the roots of the cyanobacterial
lineage, in: B.A. Whitton, M. Potts (Eds.), The Ecology of Cyanobacteria, Kluwer Academic, The Netherlands, 2000, pp. 13–35.
[64] J.W. Schopf, Microfossils of the early Archean Apex chert: New
evidence of the antiquity of life, Science 260 (1993) 640–646.
[65] B. Schrautemeier, U. Neveling, S. Schmitz, Distinct and differently
regulated Mo-dependent nitrogen-fixing systems evolved for hetero-
[66]
[67]
[68]
[69]
[70]
[71]
[72]
[73]
[74]
[75]
[76]
[77]
[78]
[79]
[80]
[81]
[82]
cysts and vegetative cells of Anabaena variabilis ATCC 29413: Characterization of the fdxH1/2 gene regions as part of the nif1/2 gene clusters, Mol. Microbiol. 18 (1995) 357–369.
R.L. Summons, J.H. Jahnke, J.M. Hope, G.A. Logan, 2-Methylhopanoids as biomarkers for cyanobacterial oxygenic photosynthesis,
Nature 400 (1999) 554–557.
T. Thiel, E.M. Lyons, J.C. Erker, Characterization of genes for a
second Mo-dependent nitrogenase in the cyanobacterium Anabaena
variabilis, J. Bacteriol. 179 (1997) 5222–5225.
T. Thiel, et al., A 2nd nitrogenase in vegetative cells of a heterocystforming cyanobacterium, Proc. Natl. Acad. Sci. USA 92 (1995) 9358–
9362.
T. Thiel, B. Pratte, Effect on heterocyst differentiation of nitrogen
fixation in vegetative cells of the cyanobacterium Anabaena variabilis
ATCC 29413, J. Bacteriol. 183 (2001) 280–286.
K.M. Towe, Environmental oxygen conditions during the origin and
early evolution of life, Adv. Space Res. 18 (1996) 7–15.
K.M. Towe, Evolution of nitrogen fixation, Science 295 (2002) 798.
D.L. Tucker, et al., The manganese stabilizing protein (MSP) and
the control of O2 evolution in the unicellular, diazotrophic cyanobacterium, Cyanothece sp. ATCC 51142, BBA—Bioenergetics 1504
(2001) 409–422.
R. Tuli, S. Naithani, H.S. Misra, Cyanobacterial photosynthesis and
the problem of oxygen in nitrogen-fixation: A molecular genetic view,
J. Sci. Indust. Res. 55 (1996) 638–657.
S. Turner, T.-C. Huang, S.-M. Chaw, Molecular phylogeny of
nitrogen-fixing unicellular cyanobacteria, Bot. Bull. Acad. Sinica 42
(2001) 181–186.
J.B. Waterbury, R. Rippka, Order Chroococcales Wettstein 1924,
emend. Rippka et al. 1979, in: J.T. Staley, et al. (Eds.), Bergey’s
Manual of Systematic Bacteriology, Williams & Wilkins, Baltimore,
1989, pp. 1728–1746.
F. Widmer, et al., Analysis of nifH gene pool complexity in soil and
litter at a Douglas Fir forest site in the Oregon cascade mountain range,
Appl. Environ. Microbiol. 65 (1999) 374–380.
R.J.P. Williams, J.J.R. Frausto da Silva, The natural selection of the
chemical elements, Oxford University Press, Oxford, 1996, p. 646.
C. Woese, The universal ancestor, Proc. Natl. Acad. Sci. USA 95
(1998) 6854–6859.
C.P. Wolk, A. Ernst, J. Elhai, Heterocyst metabolism and development,
in: E.D.E. Bryant (Ed.), The Molecular Biology of Cyanobacteria,
Kluwer Academic, Dordecht, 1994, pp. 769–823.
J. Zehr, M.T. Mellon, S. Zani, New nitrogen-fixing microorganisms
detected in oligotrophic oceans by amplification of nitrogenase (nifH)
genes, Appl. Environ. Microbiol. 64 (1998) 3444–3450.
J.P. Zehr, E.J. Carpenter, T.A. Villareal, New perspectives on nitrogenfixing microorganisms in tropical and subtropical oceans, Trends
Microbiol. 8 (2000) 68–73.
J.P. Zehr, et al., Unicellular cyanobacteria fix N2 in the subtropical
North Pacific Ocean, Nature 412 (2001) 635–638.