Nitrogen Fixation by Symbiotic and Free-Living Spirochetes

See allHide authors and affiliations

Science  29 Jun 2001:
Vol. 292, Issue 5526, pp. 2495-2498
DOI: 10.1126/science.1060281


Spirochetes from termite hindguts and freshwater sediments possessed homologs of a nitrogenase gene (nifH) and exhibited nitrogenase activity, a previously unrecognized metabolic capability in spirochetes. Fixation of 15-dinitrogen was demonstrated with termite gut Treponema ZAS-9 and free-livingSpirochaeta aurantia. Homologs of nifH were also present in human oral and bovine ruminal treponemes. Results implicate spirochetes in the nitrogen nutrition of termites, whose food is typically low in nitrogen, and in global nitrogen cycling. These results also proffer spirochetes as a likely origin of certainnifHs observed in termite guts and other environments that were not previously attributable to known microbes.

Termites are important terrestrial decomposers of Earth's major form of biomass: lignocellulosic plant material and residues derived from it, e.g., humus (1). However, the carbon-rich but typically nitrogen-poor character of the termite diet has led many species into symbiotic interactions with gut microbes to augment their nitrogen economy. These interactions include the recycling of excretory (uric acid) nitrogen and the acquisition of new nitrogen through N2 fixation (2). In wood-feeding termites, whose food may contain as little as 0.05% nitrogen (dry weight basis), N2 fixation can supply up to 60% of the nitrogen in termite biomass (3). Unfortunately, our understanding of N2-fixing microbes in termites is meager: only a few strains have been isolated (Citrobacter freundii, Pantoea agglomerans, andDesulfovibrio spp.), and their contribution to N2 fixation in situ is questionable (2). Indeed, recent surveys of the nitrogenase iron-protein encoding gene (nifH) in termite guts implied that the diversity of N2-fixing microbes was far greater than that inferred by pure culture isolation (4–6), and most of the deduced amino acid sequences of NifH differed from those of known microbial taxa (7).

A long-recognized, major, and morphologically distinct component of the termite gut microbiota are spirochetes, whose cloned 16S rDNA gene sequences group them within the genusTreponema (8). Recently, the first pure cultures of these forms were obtained (9). Isolated strains ZAS-1, ZAS-2, and ZAS-9 were also phylogenetically affiliated with the treponemes (Fig. 1), and all three strains produced acetate as a major fermentation product (10). ZAS-1 and ZAS-2 could make acetate from H2 + CO2 (9), a mode of energy-yielding metabolism previously unknown in the phylumSpirochaetes (11). Hence, they are important to the nutrition of termites, which use microbially produced acetate as a major carbon and energy source (2). Having these spirochetes in culture prompted us to examine whether they might also fix N2 and thereby contribute to termite nitrogen economy as well. To do this, we examined their genomic DNA for the presence ofnifH (12) and their ability to fix N2(13).

Figure 1

Phylogenetic tree inferred by maximum likelihood analysis of near–full-length 16S rDNA sequences of termite gut Treponema strains ZAS-1, ZAS-2, and ZAS-9 (bold), representative known spirochetes, and spirochetal 16S rDNA clones obtained directly from gut contents of the termites (clone prefix): Zootermopsis angusticollis (ZAS),Reticulitermes flavipes (RFS), and Nasutitermes lujae (NL). The homologous sequence from E. coli was used as an outgroup. Scale bar represents units of evolutionary distance and is based on sequence divergence (40). Symbols are as follows: ▪ or □, nifH (present or not detected, respectively); • or ○, nitrogenase activity (present or not detected, respectively). Absence of a symbol indicates that the spirochete was not examined for the property.

Two nifH homologs were found in each termite gut treponeme. nifH homologs were also found in the bovine ruminal treponeme, Treponema bryantii; the human oral treponemes, Treponema denticola and Treponema pectinovorum; and the free-living spirochetes, Spirochaeta aurantia, Spirochaeta zuelzerae, and Spirochaeta stenostrepta. The deduced amino acid sequence of each NifH had motifs typically present in the nitrogenase iron-protein, including conserved cysteines at positions (Klebsiella pneumoniaenumbering) 86, 98, and 133 [and 39, for nifH clones obtained with the IGK forward primer for polymerase chain reaction (PCR)] and an arginine at position 101, which is a site for reversible inactivation by adenosine diphosphate–ribosylation in some bacteria (14). However, the NifHs were phylogenetically diverse and not congruent with spirochete phylogeny based on 16S rRNA sequences, which groups all spirochetes in a single phylum. This lack of congruence extended to multiple NifH homologs in the same spirochete (Fig. 2). One homolog from each termite gut treponeme was grouped in a deeply branching cluster (IV) that included NifH-like proteins from eight Euryarchaeota. However, it is not clear that proteins in cluster IV function only, or at all, in N2 fixation (15).

Figure 2

Unrooted maximum likelihood phylogenetic tree of deduced NifH sequences from spirochetes (bold, this study), from other representative prokaryotes, and from selected termite gut (T Gut) and other environmentalnifH clones. T Gut clones prefixed “NKN-RT” (highlighted) are known to be expressed in situ (6). Groups I through IV (shaded; composition adjacent) were observed in a larger “comprehensive tree” (40) and are in accord with previously published trees, but there was no support for group III when the smaller tree was inferred using maximum likelihood. From the comprehensive tree, environmental clones most closely related to spirochete NifHs, as well as sequences that illustrated the phylogenetic breadth of each cluster, were selected for inclusion in the smaller tree. Numbers to the right of selected nodes indicate support values for that node as estimated by quartet puzzling. The scale bar represents 0.1 expected substitution per amino acid position. Abbreviations are as follows: A, Azospirillum;Az, Azotobacter; C,Clostridium; Ch, Chlorobium;D, Desulfovibrio; Ms,Methanosarcina; S, Spirochaeta;T, Treponema. Source of T Gut clones: CFN,Coptotermes formosanus; GFN, Glyptotermes fuscus; NKN, Neotermes koshunensis; PNN, Pericapritermes nitobei; TDY, Reticulitermes speratus.

We could not demonstrate nifH in the halophile,Spirochaeta halophila, or in the swine pathogen,Brachyspira (Serpulina)hyodysenteriae. Furthermore, no structural genes for nitrogenase were identified in the completely sequenced genomes of the syphilis spirochete, Treponema pallidum, or the Lyme disease spirochete, Borrelia burgdorferi (16) (Fig. 1).

The presence of nifHs in S. aurantia, S. zuelzerae, and termite gut treponeme ZAS-9 was unambiguously correlated with N2 fixation, as shown by their exhibition of N2-dependent growth and NH4 +-repressible acetylene reduction (AR) activity (Fig. 3 and Table 1). For S. aurantia and ZAS-9, fixation of 15N2 was also demonstrated.S. aurantia grew in a chemically defined medium with N2 as nitrogen source (Fig. 3A), so the 15N content of cells (89.0129 atom % excess) was close to that of the15N2 used (99.3094 atom % excess), reduced only by the 14N in cells carried over with the 10% (v/v) inoculum pre-grown on unlabeled N2. The specific activity of nitrogenase in S. aurantia (Table 1) was sufficient to provide virtually all the nitrogen needed by cells [14.7 μg of N per (hour × mg protein)] during exponential growth on N2 [doubling time = 17.3 ± 2.2 hours (n = 4)], assuming that protein and nitrogen constitute 55% and 14%, respectively, of the cell dry mass (17). By contrast, ZAS-9 and S. zuelzeraecould not be grown without yeast autolysate (YA), which itself was a source of fixed nitrogen (18). Nevertheless, nitrogen-limited growth could be achieved by using media containing 2% (v/v) YA with no added NH4Cl. AR activity in these species commenced with the onset of N2-dependent growth, which was marked by the divergence in growth curves of cultures under N2/CO2 versus those under Ar/CO2(Fig. 3, B and C). Thus, the 15N content of ZAS-9 (6.3789 atom % excess) grown under15N2/CO2 was less than that forS. aurantia, because it was diluted by 14N assimilated from YA. On the basis of the difference in cell yield of ZAS-9 grown under N2/CO2 versus Ar/CO2 (Table 1), ZAS-9 should have contained about 38 atom % 15N if 15N2 were the sole nitrogen source during N2-dependent growth (19). However, the observed value of 6.3789 atom % excess implies that N2 fixation enabled cells to assimilate nitrogenous compounds in YA that would otherwise be utilized poorly or not at all. A similar situation may exist for S. zuelzerae, because nitrogenase activity [2.3 μg of N2 per (hour × mg protein); Table 1] during N2-dependent growth (doubling time ∼21 hours) (Fig. 3C) would supply only 20% of the nitrogen needed for each doubling in biomass. Therefore, nitrogenase activities reported in Table 1 for these two spirochetes are probably not the maximum attainable by cells.

Figure 3

N2-dependent growth optical density (OD) and rate of acetylene reduction to ethylene (C2H4) exhibited by S. aurantia in a chemically defined medium lacking NH4Cl (A), and by Treponema strain ZAS-9 (B) and S. zuelzerae (C) in media containing a growth-limiting amount of yeast autolysate, but no added NH4Cl (13). Incorporation of 5 to 10 mM NH4Cl into media resulted in increased growth yields, but complete suppression of acetylene reduction activity (not shown).

Table 1

N2 fixation by free-living spirochaetas and termite hindgut treponemes. Cell yields are the mean ± SD for cells grown in NH4 +-free medium (S. aurantia) or in media containing a growth-limiting amount of yeast autolysate, but no added NH4Cl (other strains). Nitrogenase activity is the mean of two determinations on cells: (i) growing exponentially in the N2-dependent phase of growth (S. aurantia, S. zuelzerae, Treponema strain ZAS-9) and (ii) growing under NH4 +-limitation under N2/CO2 (Treponema strain ZAS-1 or ZAS-2) (13). Asterisk indicates an activity < 0.1.

View this table:

O2 (0.01 atm) immediately and completely inhibited AR by S. aurantia (a facultative anaerobe) and byS. zuelzerae and ZAS-9 (strict anaerobes), implying that N2 fixation by ZAS-9 in situ might be inhibited if cells swam into the microoxic region near the hindgut epithelium (20). However, the central region of hindguts may not be an ideal refuge because concentrations of H2 [an inhibitor of N2 reduction, whose inhibitory constant (K i) typically ranges from 0.03 to 0.2 atm (21, 22)] can reach 50,000 parts per million volume (∼0.06 atm) (20). Termite gut spirochete nitrogenases may be relatively resistant to inhibition by H2. In this regard, it is noteworthy that several spirochetal NifHs, including those of ZAS-2 and ZAS-9, group in assemblage III (Fig. 2) with the conventional nitrogenase ofClostridium pasteurianum, which has a highK i for H2 inhibition of N2 reduction (0.5 atm) (22).

Analogous experiments with ZAS-1 and ZAS-2 (both of which required YA) revealed no enhancement of growth in the presence of N2 and only trace levels of nitrogenase (Table 1), which were nevertheless detectable even when ZAS-1 and ZAS-2 were grown in the presence of 10 mM NH4Cl. Omission of molybdenum from the medium or its replacement by 10 μM NaVO3, or inclusion of 1 mM homocitrate (23) with various trace metal mixtures, did not enhance growth under N2 or AR activity of ZAS-1 or ZAS-2. Nor was AR activity increased by resuspension of cells in YA-free (non–growth supporting) medium. A trace level of AR activity was also observed with S. stenostrepta and was accompanied by production of both ethylene and ethane, implying the activity of an alternative nitrogenase (24). This is consistent with the phylogenetic placement of S. stenostrepta NifH1 in group II (Fig. 2) among the iron-proteins of alternative nitrogenases of Azotobacter vinelandii, C. pasteurianum, and Methanosarcina barkeri. However, such activity was only observed occasionally, and then only from cells in stationary phase of NH4 +-limited cultures. No evidence for N2-dependent growth or nitrogenase activity was obtained with T. bryantii growing in a chemically defined medium under NH4 +-limitation, nor was AR observed with cells of T. denticola pre-grown in a complex medium and resuspended in a nitrogen-deficient, non–growth supporting medium for assay.

Our results reveal a new dimension to the metabolic diversity within the Spirochaetes and now extend to 6 (of 18) the number of phyla within the domain Bacteria that contain N2-fixing representatives (11,25). They also reveal a role for spirochetes in termite nitrogen nutrition. Two observations suggest that N2fixation by spirochetes is important to termite nitrogen economy. First, spirochetes are unusually abundant in termite guts, accounting for as much as 50% of all prokaryotes (26). Second, many of the spirochete NifHs characterized in this study were identical or nearly identical to NifH clones obtained from a variety of termites, including NifHs known to be expressed in termite guts (Fig. 2), suggesting a spirochete origin for the latter.

The potential contribution of spirochetes to the N2 fixation activity exhibited by termites can be estimated assuming that the spirochete population is about 2 × 106 cells per μl hindgut contents [this value is one that corresponds to half of the direct microscopic count of prokaryotes (26–28)] and that one out of every three spirochetes fixes N2 at a rate of 7.5 × 10−10 μg of N2 per (hour × cell), i.e., midway between the per-cell fixation rates observed for ZAS-9 [2.8 × 10−10 μg of N2 per (hour × cell)] and S. aurantia [12.1 × 10−10μg of N2 per (hour × cell)] (29,30). When calculated for worker larvae of Zootermopsis angusticollis (the species from which ZAS strains were isolated), which weigh 30 mg, have a gut volume of ∼10 μl, and exhibit fixation rates that range from 0.06 to 0.41 ng of N2 fixed per hour (2), the spirochete-specific contribution could be as much as 5 ng of N2 per hour. This is well above that needed to account for the rate exhibited by live insects. ForCoptotermes formosanus (∼3 mg fresh weight; gut volume ∼1 μl), a species exhibiting some of the highest recorded rates of N2 fixation [4.6 ng of N2 per hour; (2)], the contribution would still be substantial (0.5 ng of N2 per hour). This is probably another reason why elimination of spirochetes from guts decreases termite survival (31).

Our results also reveal a heretofore unrecognized role for free-living spirochetes in global N cycling. Spirochetes are ubiquitous in aquatic habitats (32), and considering the similarity of some spirochetal NifHs to environmental NifH clones from zooplankton, cordgrass rhizosphere, and Antarctic ice pools (Fig. 2), it is not unreasonable to expect that some of the latter clones will ultimately prove to be of spirochetal origin and that the spirochete-specific contribution to N2 fixation in such habitats will be substantial. Hence, the discovery of N2fixation in spirochetes adds a new “twist” to our appreciation of this important, uniquely prokaryote-mediated process.

  • * To whom correspondence should be addressed: E-mail: breznak{at}


View Abstract

Navigate This Article