Nitrogen Fixation at 92°C by a Hydrothermal Vent Archaeon

See allHide authors and affiliations

Science  15 Dec 2006:
Vol. 314, Issue 5806, pp. 1783-1786
DOI: 10.1126/science.1134772


A methanogenic archaeon isolated from deep-sea hydrothermal vent fluid was found to reduce N2 to NH3 at up to 92°C, which is 28°C higher than the current upper temperature limit of biological nitrogen fixation. The 16S ribosomal RNA gene of the hyperthermophilic nitrogen fixer, designated FS406-22, was 99% similar to that of non–nitrogen fixing Methanocaldococcus jannaschii DSM 2661. At its optimal growth temperature of 90°C, FS406-22 incorporated 15N2 and expressed nifH messenger RNA. This increase in the temperature limit of nitrogen fixation could reveal a broader range of conditions for life in the subseafloor biosphere and other nitrogen-limited ecosystems than previously estimated.

Hydrothermal fluids venting from unsedimented deep-sea mid-ocean ridges are low in nitrate and ammonia (1), indicating that the microbial community inhabiting the subseafloor (2) may be limited by fixed nitrogen. Dissolved N2 is abundant in hydrothermal vent fluids (3), and biological nitrogen fixation has been proposed to explain the depleted 15N/14N ratios of low–trophic level animals living around hydrothermal vents (4). The nitrogenase enzyme complex, encoded by the nifHDK genes, catalyzes nitrogen fixation, i.e., the reduction of N2 to NH3. Although nifH genes have been detected in hydrothermal vent fluid (5), no microorganism isolated from deep-sea vents has, to our knowledge, been reported to fix nitrogen. Awide range of diverse bacteria, and the methanogenic archaea, are diazotrophic, or capable of nitrogen fixation (6). The subseafloor at mid-ocean ridges is bathed in reduced, H2- and CO2-rich hydrothermal fluid and thus provides an ideal habitat for methanogens, strict anaerobes that produce methane as a by-product. The methanogen Methanothermococcus thermolithotrophicus is currently the most thermophilic microorganism known that fixes nitrogen, at up to 64°C (7). Synechococcus ecotypes from microbial mats also fix nitrogen at up to 63.4°C (8). Here we describe the isolation of a methanogen from deep-sea hydrothermal vent fluid that fixes nitrogen at up to 92°C, which extends the upper temperature limit of biological nitrogen fixation by 28°C.

Axial volcano is located in the northeast Pacific, at the intersection of the Cobb-Eikelberg hotspot and the Juan de Fuca Ridge. The upper 100 m of oceanic crust beneath Axial seamount is estimated to have high porosity (9, 10), and the diverse archaeal community associated with its subseafloor includes thermophilic and hyperthermophilic methanogens, with maximal growth rates from 45° to 80°C and above 80°C, respectively (11). During the 2004 New Millennium Observatory cruise to Axial volcano, fluid exiting the subseafloor was sampled from a diffuse vent named marker 113 (12). The fluid, which was measured to be 23°C at the point of sampling, was inoculated into a medium designed to select for diazotrophs and incubated anaerobically at 70° and 90°C. The enrichment cultures were positive at both temperatures and transferred into an antibiotic-containing medium to prevent the growth of bacteria. The amount of fixed nitrogen in the medium was reduced over time and then omitted completely (12).

The archaeal culture, named FS406-22, was capable of growth from 58° to 92°C with N2 as the sole source of nitrogen, in a medium containing marine salts and H2 plus CO2. Maximal growth occurred at 90°C, and no growth was detected at 55° and 95°C (Fig. 1). Agitation of the culture was necessary for growth on N2, and clusters of two or more cocci were visible by phase contrast microscopy during exponential growth. To determine the identity of FS406-22, we sequenced its 16S ribosomal RNA (rRNA) gene and found it to be 99% identical to the 16S rRNA gene from the methanogen Methanocaldococcus jannaschii (fig. S1). M. jannaschii was isolated from a deep-sea hydrothermal vent chimney on the East Pacific Rise in a nitrogen-rich medium, and grows in a temperature range of 50° to 86°C with an optimum near 85°C (13). Our hydrothermal vent isolate FS406-22 produced methane, determined by means of gas chromatography (GC). The specific growth rate of FS406-22 grown with nitrate and ammonium was 0.37 hour–1 at 90°C, which is lower than the rate reported for M. jannaschii at 85°C, which is ∼1.5 hour–1 (13).

Fig. 1.

Growth rate of FS406-22 grown with N2 as the sole source of nitrogen, as a function of temperature. The specific growth rate (μ) is the average of three determinations of growth rate during exponential phase. Growth was monitored by epifluorescence microscopy (12).

The nifH gene, which encodes dinitrogenase reductase, is highly conserved, and its phylogeny is mostly consistent with that of the 16S rRNA gene. The nifD and nifK genes encode a molybdenum-containing dinitrogenase that together with dinitrogenase reductase constitutes the nitrogenase enzyme complex. Cluster 2 of the nifH phylogenetic tree (Fig. 2A) includes molybdenum nitrogenases from methanogens, as well as alternative iron- or vanadium-containing nitrogenases from Methanosarcinales and bacteria (14). Cluster 4 of the nifH phylogeny consists of paralogous archaeal nitrogenases that are assumed not to fix nitrogen.

Fig. 2.

(A) NifH amino acid phylogenetic tree constructed by quartet puzzling maximum likelihood (12). Cluster 2 includes molybdenum dinitrogenase reductases from methanogens, as well as alternative vanadium and iron-only dinitrogenase reductases (VnfH, AnfH) from Methanosarcinales and bacteria. Cluster 4 includes paralogous dinitrogenase reductases that are probably not involved in nitrogen fixation. The scale bar indicates the number of amino acid substitutions per site. The tree is outgroup rooted with Plectonema boryanum frxC, a dinitrogenase reductase–like protein involved in the light-independent reduction of protochlorophyllide. GenBank ID numbers for tree sequences are listed in table S2, and the alignment is shown in fig. S2. (B) Lanes 1 and 2: the product of RT-PCR with nifH primers and 2 and 3 μl of RNA extracted from FS406-22 growing on N2 at 90°C; lane 3: RT-PCR without RNA; lane 4: Hi-Lo DNA ladder; lane 5: nifH PCR with 2 μl of RNA; lane 6: nifH PCR without RNA. The product in lanes 1 and 2 lies between the 400- and 500-bp bands of the ladder.

Hydrothermal vent isolate FS406-22 has two copies of the nifH gene (Fig. 2A), one similar to M. jannaschii's divergent copy (15) and another related to the nifH from M. thermolithotrophicus that is expressed at 64°C (16). Some of the nifH sequences previously recovered from deep-sea hydrothermal vent fluids (5) are more similar to nifH1 from FS406-22 than from M. thermolithotrophicus [Supporting Online Material (SOM) Text]. While growing at 90°C, FS406-22 expresses nifH1, which we determined by sequencing the product of the reverse transcription polymerase chain reaction (RT-PCR) (Fig. 2B). To confirm that FS406-22 fixes molecular nitrogen and to estimate its rate of nitrogen fixation, we performed 15N2 tracer assays at 90°C by incubating the cultures with 15N2 and determining the nitrogen isotope ratios of the microbial biomass with an isotope ratio mass spectrometer (17). The control culture without added 15N2 had a δ15N of –4.86 per mil (‰) (versus air), or 0.36 atom %, which reflects the negative isotopic fractionation that results from biological nitrogen fixation. The experimental cultures had δ15N values from 3650 to 4238‰ (versus air), or 1.68 to 1.89 atom %, and the estimated rate of nitrogen fixation at 90°C by FS406-22 was 0.30 μmol N2 liter–1 hour–1 (12). Although biological nitrogen fixation has often been invoked to explain the low δ15N values of deep-sea hydrothermal vent fauna compared to nonvent fauna (18), FS406-22 is the first microorganism isolated from the deep sea reported to exhibit diazotrophy.

To further characterize the nif operon in FS406-22, we designed PCR primers for nifD and sequenced the region downstream of the nifH gene (Fig. 3). The PII nitrogen sensor proteins, encoded by nifI1 and nifI2, are responsible for “ammonia switch-off” or posttranslational regulation of nitrogenase in Methanococcus maripaludis (19) and were present in FS406-22. Two additional primers designed for nifD as well as a primer designed for nifK, based on those genes from M. thermolithotrophicus, were not successful in FS406-22. A phylogenetic tree of NifI1 and NifI2 proteins from FS406-22 and other archaea is shown in Fig. 3.

Fig. 3.

(A) The nif genes sequenced from FS406-22: nifH (807 bp), nifI1 (318 bp), nifI2 (387 bp), and nifD (141 bp). (B) An unrooted NifI1 and NifI2 amino acid phylogenetic tree determined by quartet puzzling maximum likelihood (support values listed first) and by the distance-based Fitch-Margoliash method (bootstrap values listed second). The GlnB protein from M. maripaludis is displayed as an outgroup. The maximum-likelihood branch lengths are shown, and the scale bar indicates the number of amino acid substitutions per site. GenBank ID numbers for tree sequences are listed in table S3, and the alignment is shown in fig. S3.

Highly conserved genes such as 16S rRNA can differ by less than 3%, whereas the prokaryotes they originate from show much greater genomic and physiological variation (20). In the case of FS406-22 and M. jannaschii, whose 16 S rRNA genes diverge by only 1%, the former contains a functional nif operon and fixes nitrogen, whereas the latter does not. FS406-22 grown with N2 or fixed nitrogen can grow at 6°C hotter than M. jannaschii, but the optimal growth rate of M. jannaschii is four times as great as that of FS406-22. To perform a rigorous comparison, however, one would have to cultivate both organisms under identical conditions, which we have not done. We also sequenced a 1550–base pair (bp) region of the FS406-22 genome that included the riboflavin synthase gene (ribC) and a putative aldehyde ferredoxin oxidoreductase gene (aor) that was 85% similar to the same stretch of DNA in M. jannaschii. The cluster 4 nifH2 from FS406-22 was 87% similar to the only nifH from M. jannaschii at the nucleotide level and 96% similar at the amino acid level (Fig. 2A).

The revelation of a hyperthermophilic archaeal diazotroph may have implications for the evolutionary history of nitrogenase. Phylogenetic analysis of nitrogenase and chlorophyll iron proteins suggests that an ancestral iron protein duplicated and diverged into molybdenum and iron-only nitrogenases (nifH and anfH) before the separation of bacteria and methanogenic archaea (21). Subsequently, ancient duplications of nifH gave rise to the chlorophyll iron proteins, suggesting that nitrogenase predates photosynthesis and the rise of atmospheric oxygen (21). Raymond et al. detail two scenarios for the evolution of nitrogenase: that it was present in the last universal common ancestor and then lost in many lineages, or that it arose later within methanogenic archaea and was laterally transferred into bacteria (22). Given the lack of oxygen in the early Earth's atmosphere (23) and that nitrogenase is inactivated by O2, the first nitrogenase probably arose in an anaerobe.

All methanogenic diazotrophs examined, and several anaerobic bacteria like Chlorobium tepidum, Dehalococcoides ethenogenes, Heliobacterium chlorum, Clostridium acetobutylicum, and Desulfovibrio gigas, have the nifI1 and nifI2 genes (24). In all of the archaeal operons, they are located between nifH and nifD, and the NifI1 and NifI2 proteins form a complex that binds directly to dinitrogenase (19). Because NifI1, NifI2 and GlnB/GlnK constitute all three subfamilies of the PII nitrogen sensor family (25), the clusters provide a root for each other in an unrooted tree containing all three (26). Such phylogenetic analyses, as well as molecular evolutionary studies of other nif genes, all place M. thermolithotrophicus on a deep, basal branch (6, 2628). The unrooted NifI1 and NifI2 phylogenetic tree in Fig. 3 shows that the FS406-22 and M. thermolithotrophicus proteins form a deeply branching group that is basal to all other archaeal proteins, including those from alternative nitrogenases. However, the quartet puzzling support and bootstrap values are not high enough to rule out alternative topologies. The maximum-likelihood branch lengths in Fig. 3 suggest that FS406-22 NifI1 and NifI2 are the shortest distance to the internal node that represents the ancestral PII protein. A recent reconstruction of the tree of life with 31 universal gene families supports the hypothesis that the last universal common ancestor lived at high temperatures (29). We propose that among diazotrophic archaea, the nitrogenase from FS406-22 might have retained the most ancient characteristics, possibly derived from a nitrogenase present in the last common ancestor of modern life.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S4

Tables S1 to S4


References and Notes

View Abstract

Navigate This Article