Growth of nitrite-oxidizing bacteria by aerobic hydrogen oxidation

See allHide authors and affiliations

Science  29 Aug 2014:
Vol. 345, Issue 6200, pp. 1052-1054
DOI: 10.1126/science.1256985

Oxidizing hydrogen in place of nitrite

Microorganisms are important drivers of Earth's nitrogen cycle. Many of the organisms responsible for mediating the reactions of one phase of nitrogen to another are thought to be ecologic specialists. Using a combination of genomic and experimental analyses, Koch et al. show that Nitrospira moscoviensis, a member of a widely distributed genus of nitrite-oxidizing bacteria, can oxidize hydrogen instead of nitrite to support growth when oxygen is present. Not only does this ecologic flexibility suggest a broader distribution of these organisms in natural settings, but they may be important in engineered environments as well.

Science, this issue p. 1052


The bacterial oxidation of nitrite to nitrate is a key process of the biogeochemical nitrogen cycle. Nitrite-oxidizing bacteria are considered a highly specialized functional group, which depends on the supply of nitrite from other microorganisms and whose distribution strictly correlates with nitrification in the environment and in wastewater treatment plants. On the basis of genomics, physiological experiments, and single-cell analyses, we show that Nitrospira moscoviensis, which represents a widely distributed lineage of nitrite-oxidizing bacteria, has the genetic inventory to utilize hydrogen (H2) as an alternative energy source for aerobic respiration and grows on H2 without nitrite. CO2 fixation occurred with H2 as the sole electron donor. Our results demonstrate a chemolithoautotrophic lifestyle of nitrite-oxidizing bacteria outside the nitrogen cycle, suggesting greater ecological flexibility than previously assumed.

Nitrification, a key nitrogen cycling process, is prevalent in virtually all oxic habitats on Earth (1). This two-step process is carried out by chemolithoautotrophic aerobic ammonia-oxidizing bacteria (AOB) and archaea, and nitrite-oxidizing bacteria (NOB). Traditionally, nitrifying microorganisms (nitrifiers) have been regarded as obligate chemolithotrophs that grow only in the presence of ammonia or nitrite. Accordingly, their environmental distribution and population sizes are associated with nitrification, and NOB are thought to depend on ammonia-oxidizing or nitrate-reducing microorganisms as sources of nitrite.

Some nitrifiers, however, are physiologically versatile and can participate in denitrification, even in parallel to aerobic nitrification (2), or co-utilize various organic substrates if ammonia or nitrite is available (3, 4). Few Nitrobacter isolates (NOB) grow heterotrophically on simple organic compounds (5). Moreover, Nitrosomonas eutropha (AOB) and Nitrospira moscoviensis (NOB) utilize hydrogen (H2) as electron donor for anoxic respiration with the electron acceptors nitrite or nitrate, respectively [although growth under these conditions was reported only for N. eutropha (6, 7)]. Aside from the growth of Nitrobacter on organic compounds, these additional metabolic capabilities of nitrifiers are still linked to the nitrogen cycle and actually depend on nitrification as they require ammonia, nitrite, or nitrate as energy source or electron acceptor, respectively.

To understand the metabolic capabilities and flexibility of NOB, we partially sequenced the genome of Nitrospira moscoviensis, a member of Nitrospira lineage II that grows in pure culture (7). Nitrospira (8) represents the environmentally most widespread group of known NOB found in virtually all environments, including soils, fresh water and the oceans, sediments, subsurface aquifers, volcanic sites, hot springs, caves, iron pipes, drinking water treatment systems, and wastewater treatment plants (3). Biological knowledge on the mainly uncultured, slow-growing Nitrospira is scarce, and only one genome sequence has been obtained from “Candidatus N. defluvii,” a Nitrospira lineage I member enriched from a wastewater treatment plant (9).

The draft genome of N. moscoviensis contains a locus with genes coding for the small (HupS) and large (HupL) subunits of a putative group 2a [NiFe] hydrogenase (Hup) and homologs of proteins involved in hydrogenase maturation and transcriptional regulation in other bacteria (Fig. 1, table S1, and supplementary text). The hup locus likely explains the weak, anoxic, H2-oxidizing activity with nitrate of N. moscoviensis (7). However, in cyanobacteria and terrestrial actinomycetes, Hup functions as an uptake hydrogenase for the utilization of H2 as a source of electrons and energy under oxic conditions (10, 11). Consequently, we hypothesized that Hup may enable N. moscoviensis to gain energy by oxidizing H2 with O2 as the terminal electron acceptor, a Knallgas metabolism described for various non-nitrifying chemolithoautotrophic bacteria (12).

Fig. 1 Schematic illustration of the hydrogenase (hup) locus of N. moscoviensis.

Predicted structural features or domains are marked with colored vertical lines, TPR (tetratricopeptide repeats), or NHL (NHL repeats). Asterisks indicate significant gene up-regulation (DESeq Padj < 0.05, fold change ≥2) under H2-oxidizing conditions, as detected by whole transcriptome shotgun sequencing. Genes are drawn to scale. TMH, transmembrane helix.

To test this hypothesis, we used a pure culture of N. moscoviensis that was grown on nitrite, removed all residual nitrite and nitrate, and transferred the cells into nitrite-free mineral medium that contained ammonium as the nitrogen source for assimilation (13). Indeed, we observed growth in the presence of H2 and O2 as sole electron donor and acceptor, respectively, but not when either H2 or O2 was omitted (Fig. 2). Hydrogenotrophic growth continued when a fraction of a H2-grown batch culture was transferred into fresh mineral medium containing H2 (fig. S1). H2-dependent autotrophic carbon fixation of N. moscoviensis was confirmed by quantifying the uptake of [13C]HCO3 at the single-cell level by nanometer-scale secondary ion mass spectrometry (NanoSIMS) (Fig. 3). The simultaneous specific detection of N. moscoviensis by fluorescence in situ hybridization (FISH) in this experiment (Fig. 3A) ensured that the culture was free of contaminants, including other hydrogenotrophic organisms. Comparative transcriptome analyses of N. moscoviensis grown on nitrite or H2 (13) revealed that except for the upstream transcriptional regulator hoxA and gene NITMOv1_410085, the hup locus was up-regulated at least twofold and among the 110 most strongly transcribed genes in the presence of H2 (table S1 and Fig. 1). This up-regulation of Hup and its maturation factors is consistent with a key function in the growth on H2. Interestingly, the subunit genes hupL and hupS were also relatively strongly transcribed during growth on nitrite (table S1). A low constitutive level of the hydrogenase might function as a sensory mechanism for H2. When H2 becomes available, the resulting increased flux of reductant might trigger the up-regulation of the hydrogenase.

Fig. 2 Hydrogenotrophic growth of N. moscoviensis with O2 as terminal electron acceptor.

Data points represent means of three biological replicates for each incubation condition. Error bars represent 1 SD (not shown if smaller than symbols). Growth on H2 and O2 was linear, whereas growth of N. moscoviensis on nitrite and O2 is logarithmic (7). The slower growth on H2 indicates a yet unidentified limiting factor. In the incubation with O2 only, total protein decreased likely because of endogenous respiration in the absence of any external electron donor.

Fig. 3 Chemolithoautotrophic carbon fixation by N. moscoviensis with H2 as sole energy source.

(A) Representative FISH (left) and NanoSIMS (middle and right) images showing N. moscoviensis cells in the same field of view after 13 days of incubation with 13C-labeled bicarbonate, H2, and O2. Cells were detected by FISH with the Nitrospira lineage II–specific probe Ntspa1151 (control hybridizations with a nonsense probe gave no signals). The 32S- secondary ion signal intensity distribution image is shown to illustrate the distribution of biomass. Carbon fixation is demonstrated by a high 13C content of cells. (B) 13C labeling of N. moscoviensis quantified by NanoSIMS at the single-cell level at different time points during the same incubation as shown in (A) and control incubations. The label content is presented as the 13C/(12C + 13C) isotope fraction, given in atomic %. The box plots summarize the median and the quartiles of the 13C content of cells. Numbers of measured cells (n) are indicated.

N. moscoviensis can oxidize low levels of H2 injected in the culture headspace. The experiments described above were performed with 50 to 70% (v/v) H2, which corresponds to initial dissolved H2 concentrations of 0.37 to 0.52 mM in the medium at the beginning of the incubations (13). In additional incubations with only 0.7% (v/v) of H2 in the headspace (5.2 μM dissolved H2 in the medium), N. moscoviensis partially consumed the H2 and fixed carbon as shown by addition of [13C]HCO3 and isotope ratio mass spectrometry of the biomass (fig. S2, A to C). However, although H2 was used as an energy source for biosynthesis, net growth of N. moscoviensis was not observed (fig. S2D). We assume that only a fraction of the culture was metabolically highly active, whose growth during the incubation period was not detectable by the applied protein quantification assay. This explanation is consistent with the heterogeneous 13C-labeling at the single-cell level observed by NanoSIMS even after incubation under high-H2 conditions (Fig. 3). As shown by reverse transcription polymerase chain reaction (table S2), the hupS and hupL genes were also up-regulated at 5.2 μM H2 (fig. S3). Transcription of the only other potentially H2-oxidizing enzyme identified in the draft genome, a putative formate hydrogenlyase (supplementary text), was very weak at 0.37 to 0.52 mM H2 (table S1) or not detected at 5.2 μM H2 (fig. S3), suggesting that Hup catalyzed H2 oxidation at these H2 concentrations.

The concentration of 5.2 μM dissolved H2 is within the range found in gas bubbles in cyanobacterial mats (14). In prolonged incubations, N. moscoviensis still consumed H2 at 0.04% (v/v) in the culture headspace (293 nM dissolved H2) (fig. S4A). Levels of H2 in this range occur in hot springs (15) but are higher than steady-state H2 concentrations in lake and river sediments (16). The lowest concentration threshold of H2 utilization by N. moscoviensis was not determined here, but as shown recently, the high substrate affinity of group 2a [NiFe] hydrogenases of terrestrial actinomycetes allows even the scavenging of tropospheric H2 (11).

Phylogenetically, Hup of N. moscoviensis clusters with group 2a [NiFe] hydrogenases of Cyanobacteria and nonphototrophic Alpha- and Gammaproteobacteria (fig. S5), although Nitrospira belong to a distinct bacterial phylum (7). Thus, the hup locus was likely acquired by Nitrospira through lateral gene transfer. Whether the absence of hup in “Ca. N. defluvii” results from secondary loss of the hydrogenase locus in this organism, or whether hup was laterally acquired only by Nitrospira lineage II (including N. moscoviensis), will remain unclear until genomic sequences of other Nitrospira become available.

Aerobic H2 oxidation will be ecologically advantageous for NOB, as the amount of energy that can be gained (ΔG°' = –237 kJ mol−1 H2) is much larger than for nitrite oxidation (ΔG°' = –74 kJ mol−1 NO2). Low-potential electrons from H2 also reduce the energy requirement for reverse electron transport, which is needed to fix CO2 with nitrite as the electron donor. Addition of nitrite to H2-oxidizing cultures, which were incubated with low levels of H2, revealed that N. moscoviensis can oxidize both substrates simultaneously (fig. S4B). A lifestyle in which H2 and O2 are used as substrates, exclusively or in addition to aerobic nitrite oxidation, would increase the competitiveness of NOB in habitats where microbial processes provide H2, such as in cyanobacterial mats, at oxic-anoxic interfaces, and in hypoxic pockets of soils, sediments, or biofilms, or at hydrothermal sites where upwelling fluids contain H2 (17). Indeed, Nitrospira have been found in low-oxygen niches such as the basal zones of biofilms (18); in marine sediments, subsurface aquifers, and rice paddies; and also in deep-sea hydrothermal field sediment and hot springs (19, 20). In such habitats, aerobic H2 oxidation may provide extra energy and make (at least some) NOB independent of nitrite supplied by ammonia oxidizers or nitrate reducers.

Loci encoding putative hydrogenases of different types occur also in the genomes of NOB other than Nitrospira, indicating that H2 utilization may be a widespread feature of these organisms. The marine nitrite oxidizer Nitrospina gracilis (phylum Nitrospinae) possesses the genes of a cytoplasmic group 3b bidirectional [NiFe] hydrogenase (21). A group 5 uptake [NiFe] hydrogenase was identified in the genome of Nitrolancea hollandica (phylum Chloroflexi), a nitrite-oxidizing bacterium isolated from activated sludge (22) (fig. S5).

The ability to switch their lifestyle would enable NOB to colonize new ecological niches and may stabilize nitrification by maintaining NOB populations during periods of nitrite depletion. The aerobic growth on H2 of N. moscoviensis suggests that the contribution of NOB to chemolithoautotrophic CO2 fixation in microbial communities is potentially greater than expected from the low energy yield of nitrite oxidation (23). Thus, a reassessment of their functional roles will be essential to more fully understand the ecology of NOB and to determine the impact of these almost ubiquitous microorganisms on the biogeochemical cycles of nitrogen and carbon in natural and engineered ecosystems.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S5

Tables S1 and S2

References (2450)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: We thank the anonymous reviewers for their valuable comments and suggestions. The sequences of genes reported in this paper are available at NCBI GenBank under accession nos. KJ920237 to KJ920263. Other supporting data are available in the supplementary materials. This work was supported by the Vienna Science and Technology Fund (WWTF, grant LS09-40), the Austrian Science Fund (FWF, grants P25231-B21 and P24101-B22), the Danish Council for Strategic Research (EcoDesign), the Deutsche Forschungsgemeinschaft (DFG, grant SP667/3-1), and the European Research Council [Advanced Grant Nitrification Reloaded (NITRICARE) 294343].
View Abstract

Stay Connected to Science

Navigate This Article