Introduction

Nitrogen is a limiting nutrient in much of the world’s ocean. Thirty to fifty percent of the fixed nitrogen available for marine organisms is lost from the ocean due to the biological production of dinitrogen gas (N2) in oxygen minimum zones (OMZs) (Codispoti et al., 2001; Galloway et al., 2008). Nitrogen loss in these regions is attributed to two microbially mediated processes, heterotrophic denitrification (Jayakumar et al., 2004; Castro-Gonzalez et al., 2005; Ward et al., 2009; Babbin et al., 2014) and anaerobic ammonia oxidation (anammox) (Kuypers et al., 2005; Thamdrup et al., 2006; Hamersley et al., 2007; Lam et al., 2009). Genomic data have identified an abundant group of sulfur-oxidizing marine chemoautotrophs (SUP05) that are assumed to contribute to either denitrification or anammox (Walsh et al., 2009; Canfield et al., 2010; Zaikova et al., 2010; Ulloa et al., 2012; Wright et al., 2012; Mattes et al., 2013; Hawley et al., 2014; Murillo et al., 2014). Members of the SUP05 clade are hypothesized to contribute directly by sequential reduction of nitrate (NO3−) to nitrogenous gases (N2O or N2), or indirectly by dissimilative NO3− reduction to ammonia (DNRA), which can in turn fuel anammox (Walsh et al., 2009; Canfield et al., 2010; Hawley et al., 2014; Murillo et al., 2014).

SUP05 also have the genetic potential to produce nitrite (NO2−), which is a critical intermediate in denitrification and a necessary reductant for anammox. The sources of NO2−, and in particular the secondary NO2− maximum, in OMZs are poorly understood (Lam et al., 2009). Accumulation of NO2− in OMZs has been attributed to heterotrophic denitrification leading to N-loss (Lam et al., 2009). The current paradigm is that NO2− in OMZs is produced at the oxycline by aerobic ammonia oxidizing archaea and bacteria (AOA, AOB) (Hawley et al., 2014) and within the OMZ by heterotrophs (Codispoti et al., 2001; Francis et al., 2007; Lam et al., 2009). Chemoautotrophic SUP05 also have the genetic potential to respire NO3− and produce NO2− (Walsh et al., 2009; Hawley et al., 2014; Murillo et al., 2014). Here we test the hypothesis that a chemoautotrophic bacterium from the SUP05 clade respires NO3− and produces NO2−.

We isolated a representative from the SUP05 clade to elucidate the effects of these sulfur-oxidizing bacteria on the marine nitrogen cycle. We describe the isolation, genetic potential and growth requirements of a sulfur-oxidizing chemolithoautotroph that produces NO2− and is limited by ammonium (NH4+). We provide insights into the hypothesized roles of SUP05 in the marine nitrogen cycle that could not be inferred from environmental sequence data alone. Our primary finding suggests that SUP05 produce NO2− in OMZs, a critical intermediate required for nitrogen removal processes (anammox and denitrification).

We propose the following name for the first isolate from the SUP05 clade:

Thioglobus gen. (Marshall and Morris, 2013).

“Candidatus Thioglobus autotrophicus” sp. nov

Etymology: au.to.tro'phi.cus. Gr. n. autos self; Gr. adj. trophikos nursing, tending or feeding; N.L. masc. adj. autotrophicus autotroph.

Materials and methods

Cultivation

Water was collected from Effingham Inlet (49°04.2685’ N, 125°09.4270’ W) during a research cruise aboard the R/V Thomas G. Thompson in February 2013 using a rosette with 10 liter Niskin bottles (General Oceanics, Miami, FL, USA) and equipped with a Seabird conductivity, temperature and density (CTD) device and dissolved oxygen (DO) sensors calibrated by Winkler titrations. Live cells were collected from the top of the suboxic zone (defined as the minimum value obtained by the Seabird DO sensor) (Supplementary Figure 1). Cell numbers were estimated based on prior results and diluted to extinction (5 cells/well) in two 96 well Teflon plates (SonomaTesting, Santa Rosa, CA, USA) using a high throughput culture method with on-site filter sterilized seawater (30 kD) used as media. One 96 well plate was enriched with 1 mM thiosulfate (S2O32−) while the other was used as a control and contained only filter sterilized seawater. Growth in each well was measured every 7 days in incubations at in situ temperatures (10 °C) by first transferring 150 μl aliquots of culture into 96 well polycarbonate plates (Millipore, Billerica, MA, USA) and then staining each well with SYBR Green I (Invitrogen, Carlsbad, CA, USA) diluted in TRIS buffer pH 7.4 at a final concentration of 1/2000. Cell concentrations in 96 well polycarbonate plates were measured using an Easyflow Guava flow cytometer equipped with a 96 well plate reader (Millipore, Billerica, MA, USA).

Isolate identification

Cultures were identified as previously described (Marshall and Morris, 2013), with the following modifications. Cells were transferred to 250 ml acid washed polycarbonate flasks (10% HCl) and grown to early stationary phase (~2.0 × 106 cells/ml), then collected on sterile Supor-200 0.2 μm polyethersulfone filters (Pall, Port Washington, NY, USA). We identified cultures by amplifying and sequencing the 16S rRNA gene with bacterial primers (27F, 519F/R, 926F/R and 1492R). Sequences were subsequently aligned with nearly complete 16S rRNA gene sequences and a maximum likelihood tree was constructed using RAxML. The Methylococcus capsulatus (Texas strain) 16S rRNA gene was selected as an out-group and 100 bootstrap replicates were used to evaluate clusters (Figure 1a). Culture purity was verified using terminal restriction fragment length polymorphism (TRFLP) analysis. Briefly, universal primers 27F and 1492R were used to amplify 16S DNA that was then restricted with either MboI or HaeIII (New England Biolabs, Ipswich, MA, USA) as previously described (Marshall and Morris, 2013).

Figure 1
figure 1

Identity and genome of “Ca. T. autotrophicus” strain EF1. (a) Maximum likelihood tree of marine gamma sulfur-oxidizing bacteria 16S rRNA sequences. The tree was constructed using RAxML. Bootstrap values >40 are labeled at nodes (100 iterations). Red are symbionts, green are clones and black are isolates. (b) Circular representation of the strain EF1 genome. Innermost circle - GC skew (purple is negative skew and green is positive skew). Bars within solid lines indicate predicted coding regions colored by metabolic categories. Orange (S-oxidation and reduction genes): sox (sulfur oxidation), apr (adenosine-5-phosphate reductase), fcc (flavocytochrome c sulfide dehydrogenase), and dsr (dissimilatory sulfite reductase). Navy (NO3− and NO reduction genes): nap (periplasmic nitrate reductase), nar (nitrate reductase), nor (nitric oxide reductase). Teal (N assimilation): GOGAT (glutamate synthase) and ammonia, amino acid, and peptide transporters. Red (CO2 fixation): rbcS (RuBisCO small subunit), rbcL (RuBisCO large subunit).

Full size image

Genome sequencing and annotation

DNA was extracted as previously described (Marshall and Morris, 2013), with the following modifications. Genomic DNA was extracted from a total of 62 pure cultures grown in 100 ml aliquots. Cells were grown to early stationary phase (~2.0 × 106 cells/ml) and then collected on sterile Supor-200 0.2 μm polyethersulfone filters (Pall, Port Washington, NY, USA). Clone library preparation for genome sequencing was performed at the University of Washington’s Genome Sciences Department using Pacific Bioscience’s Single Molecule Real Time (SMRT) sequencing technology. De novo assembly of the “Ca. T. autotrophicus” strain EF1 genome was conducted using the Hierarchical Genome Assembly Process (HGAP) as previously described (Koren et al., 2012). This method has been found to produce highly accurate de novo assemblies of small prokaryotic genomes (Roberts et al., 2013). HGAP assembly of the “Ca. T. autotrophicus” strain EF1 genome resulted in a single contig that was closed using one PCR reaction (Supplementary Methods). With an average coverage of 106X, the final assembly supported a single circular chromosome 1,512,449 bp in length (Figure 1b, Supplementary Table 1). Annotations obtained by NCBI’s automatic Prokaryotic Genome Annotation Pipeline were checked and corrected by comparisons with RAST annotations, IMG annotations, and in some cases by phylogenetic inference. The complete genome sequence of strain EF1 was announced (Shah and Morris 2015), and is available under the GenBank nucleotide accession no. CP010552.

Growth conditions

T. autotrophicus EF1 cells were grown at in situ temperature (10 °C) in acid-washed (10% HCl) and autoclaved bottles containing copiotrophic seawater media from Puget Sound or oligotrophic seawater media from the Sargasso Sea (Supplementary Table 2). Aerobic cultures were maintained in acid washed polycarbonate bottles. Anaerobic cultures were maintained in acid washed and autoclaved 125 ml serum bottles, sealed with 20 mm butyl rubber stoppers (Wheaton, Millville, NJ, USA), then bubbled with an N2:CO2 gas mix (1000 ppm CO2, Praxair, Danbury, CT, USA, Specialty Gas Mix) for 10 min, and headspace sparged for an additional 5 min. Complete removal of oxygen inside serum bottles was confirmed using BD (Franklin Lakes, NJ, USA) GasPak Anaerobic strips added to an un-inoculated control serum bottle. Experiments (aerobic and anaerobic) were started by transferring 1000 cells in early exponential growth phase to new media. Culture purity was checked before and after each physiology study by sequence analysis and TRFLP.

Results and Discussion

Isolation of a representative from the SUP05 clade

Six out of 192 culture wells inoculated with water from a redox gradient in Effingham Inlet were positive for growth after 21 days. Four of the cultures had identical 16S rRNA gene sequences and were identified as members of the SUP05 clade. One culture was identified as a closely related gamma-proteobacterium just outside the clade and one culture was identified as an epsilon-proteobacterium related to Arcobacter sp. associated with marine sponges. All are suspected sulfur-oxidizing bacteria. “Ca. T. autotrophicus” strain EF1 was selected for further study. The remaining cultures were cryopreserved.

Phylogenetic analysis and genome sequencing further confirmed the identity and genetic potential of “Ca. T. autotrophicus” strain EF1 (Figure 1). Strain EF1 is most closely related to sequences from the original SUP05 clade described by Walsh et al. (2009). These include environmental clones recovered from a broad range of OMZs and symbionts of deep-sea mollusk Bathymodiolus sp. Related sequences derived from the whole genomes of symbionts, and from environmental clones from the northeast Pacific ridge (Huber et al., 2006), the Namibian upwelling system (Lavik et al., 2009), Suiyo Seamount (Sunamara et al., 2004), Saanich Inlet (Walsh et al., 2009), the Eastern North Pacific and Eastern South Pacific (Stevens and Ulloa, 2008), the South Atlantic and North Pacific Gyres (Swan et al., 2011) and Puget Sound (Marshall and Morris, 2013). The complete 16S rRNA gene sequences obtained from the “Ca. T. singularis” strain PS1 and “Ca. T. autotrophicus” strain EF1 genomes were also analyzed using the SILVA high quality ribosomal RNA database (Quast et al., 2013). “Ca. T. singularis” strain PS1 was most closely related to sequences in the Arctic96BD-19 subclade (ZD0405 in SILVA). “Ca. T. autotrophicus” strain EF1 was most closely related to sequences in the SUP05 subclade.

The purity of “Ca. T. autotrophicus” strain EF1 was confirmed by several methods. TRFLP analyses identified a single 266 bp fragment using the restriction enzyme MboI and a 193 bp fragment using the restriction enzyme HaeIII (Supplementary Figure 2). These exactly match the fragments predicted from the 16S rRNA gene sequence. No other fragments were observed. Purity was further confirmed by quantitative fluorescence in situ hybridization (FISH) analyses with a SUP05 specific probe (GSO-1032) that exactly matches the 16S rRNA (Glaubitz et al., 2013) (Supplementary Figure 2). All of the DAPI stained objects observed in three images (117/115/118) also hybridized to the SUP05 probe (total=350/350). Cultures were subsequently cryopreserved and revived from glycerol stocks several times and under both aerobic and anaerobic growth conditions. Sequence and restriction analyses were used to check purity every time a culture was revived from glycerol stocks and before and after every physiology experiment. In every case, these analyses produced the same 16S rRNA gene sequences and the same restriction patterns. Transmission electron microscopy images revealed a single morphology, indicating that strain EF1 is a small (~0.3–0.4 μm) cocci shaped bacterium that produces extracellular globules resembling those produced by “Ca. T. singularis” PS1 (Supplementary Figure 3).

Genetic potential of “Ca. T. autotrophicus” strain EF1

The complete genome of “Ca. T. autotrophicus” strain EF1 has a GC content of 39.1%. It codes for 1,506 proteins, 92 pseudogenes, 3 rRNAs (5S, 16S and 23S) and 35 tRNAs. It has the genetic potential to grow as a facultatively anaerobic chemolithoautotroph that oxidizes sulfur and can reduce O2, NO3− and NO (Figure 1b). The genome codes for key enzymes for carbon fixation via the Calvin-Benson-Bassham (CBB) cycle, including cbbYCOQR, carbonic anhydrase and cytochrome cbb3, one copy of the small subunit of RuBisCO (form I) and two copies of the large subunit of RuBisCO (form I and form II) (Badger and Bek, 2008). RuBisCO form I is composed of large and small subunits and is present in most chemoautotrophic bacteria, cyanobacteria, red and brown algae and all plants. Form II is a dimer of large subunits and is present in purple non-sulfur bacteria, some chemoautotrophic bacteria and in dinoflagellates. Some non-sulfur phototrophic bacteria contain both forms of RuBisCO. Strain EF1 is facultatively anaerobic and has likely adapted to use form IA RuBisCO or form II RuBisCO, depending on the ratio of CO2 to O2. Strain EF1 codes for complete glycolytic and phosphogluconate pathways (nonoxidative), and has most genes encoding the tricarboxylic acid (TCA) cycle. “Ca. T. autotrophicus” strain EF1 does not code for α-ketoglutarate dehydrogenase, a key TCA enzyme that is also missing from closely related symbiont genomes and a planktonic SUP05 population genome from Saanich Inlet (Walsh et al., 2009). The absence of α-ketoglutarate dehydrogenase suggests that “Ca. T. autotrophicus” strain EF1 is an obligate autotroph (Wood et al., 2004). Cytochrome c oxidase was also identified, along with a suite of genes for oxidative phosphorylation, further indicating the potential for strain EF1 to grow under aerobic conditions.

Genes for chemoautotrophic energy conservation were also identified. These include genes for inorganic sulfur oxidation (fccAB, dsrABCH, aprABM, soxABXYZ and rhodanese sulfurtransferase) and for aerobic and anaerobic respiration on O2, NO3− and NO (narQGHIJ, napABGD and norBCD) (Figure 1b). Sulfur oxidation genes confer the ability to oxidize a broad range of reduced sulfur compounds, including hydrogen sulfide (H2S), elemental sulfur (S0) and thiosulfate (S2O32−). Similar to Saanich Inlet SUP05, strain EF1 is also missing soxCD sulfur dehydrogenase genes, suggesting that they store S0. The absence of soxCD has coincided with the ability of a closely related symbiont, Ruthia magnifica, to store sulfur in the form of extracellular globules (Newton et al., 2007).

Genes for anaerobic respiration confer the ability to carry out two of the four steps associated with sequential denitrification (NO3− → NO2− → NO → N2O → N2). Strain EF1 has the genetic potential to reduce nitrate to nitrite (NO3− → NO2−) and to reduce nitric oxide to nitrous oxide (NO → N2O) (Figure 1b), but lacks genes to reduce nitrite to nitric oxide (NO2− → NO) or to reduce nitrous oxide to nitrogen gas (N2O → N2). Although the ability to use NO2− and N2O as terminal electron acceptors has been observed in environmental datasets (Walsh et al., 2009; Hawley et al., 2014; Murillo et al., 2014). We also found that strain EF1 is missing key genes required to use hydrogen gas (H2) as an electron donor, as previously reported for environmental SUP05 in the Guaymas Basin (Anantharaman et al., 2013)

“Ca. T. autotrophicus” strain EF1 has key genes required to use either NH4+ or organic nitrogen for biosynthesis (Figure 1b). These include genes that confer the ability to regulate intracellular nitrogen and to assimilate NH4+ and amino acids. Strain EF1 codes for two ammonia transporters, NADPH-dependent glutamate synthase (GS) and glutamine oxoglutarate aminotransferase (GOGAT), as well as components of amino acid (AA) and peptide ABC-transporters.

Growth requirements

Physiology experiments confirmed that strain EF1 is a facultatively anaerobic chemolithoautotroph that requires inorganic sulfur as a source of electrons (Figure 2). Batch cultures grew to an average final cell density of 3.6 × 106 cells/ml under aerobic conditions in seawater media containing 1 mM S2O32− and to an average final cell density of 2.8 × 106 cells/ml under anaerobic conditions in seawater media amended with 1 mM S2O32−, 100 μm of NO3− and sparged with a mixture of N2:CO2 (Figures 2a and b, respectively). Cells were unable to grow in aerobic seawater media that lacked S2O32− or in anaerobic media that lacked either S2O32− or CO2. This indicates that “Ca. T. autotrophicus” strain EF1 requires a reduced form of inorganic sulfur for electrons and CO2 for biosynthesis. We have found that strain EF1 cells survive two transfers with no additional sulfur (Supplementary Figure 3). This is likely due to their potential to store sulfur in extracellular globules (Walsh et al., 2009; Marshall and Morris, 2013; Hawley et al., 2014). No growth was observed when “Ca. T. autotrophicus” strain EF1 was grown with H2 as an electron donor (Supplementary Figure 4).

Figure 2
figure 2

Chemoautotrophic growth of “Ca. T. autotrophicus” strain EF1 under aerobic and anaerobic growth conditions. (a) Aerobic growth on copiotrophic seawater media from Puget Sound amended with 1mM S2O32−. (b) Anaerobic growth on seawater media bubbled with either N2 only or N2:CO2 gas mix and amended with 1 mM S2O32− and 100 μm NO3−. The controls were not amended with S2O32−, NO3− or CO2. Experimental treatments were conducted in triplicate.

Full size image

We grew strain EF1 on copiotrophic media (Supplementary Table 2) enriched with NO3− to further evaluate its potential for dissimilatory NO3− reduction under anaerobic conditions (Figure 3). Cultures grew to similar cell densities at in situ concentrations of NO3− (32 μm) and when 100 μm of additional NO3− was added to the media (Figure 3a). Strain EF1 was not limited by NO3− on copiotrophic seawater media. In both cases there was strong evidence for dissimilatory NO3− reduction, as indicated by a 1:1 conversion of NO3− to NO2− (Figure 3b), as well as uptake of NH4+ (Figure 3c) and an increase in the production of N2O (Figure 3d). Although NO was not added to the media, some NO may have been present in the seawater used to make the media or introduced into as a contaminant via the N2:CO2 gas mix used to sparge the media. Regardless, growth experiments support genomic predictions and previously published results from the field indicating that SUP05 produce N2O (Walsh et al., 2009; Hawley et al., 2014). EF1 cells did not produce N2 gas (Supplementary Figure 4).

Figure 3
figure 3

Nitrogen utilization by “Ca. T. autotrophicus” strain EF1 on copiotrophic media. (a) Anaerobic growth on seawater media from Puget Sound with in situ concentrations of NO3− (32 μM) and amended with 100 μM of additional NO3− (total=132 μM). (b) Concentrations of NO3−, NO2−, and NH4+. (c) Concentrations of only NH4+. (d) Concentrations of N2O measured by gas chromatography (Supplementary Materials and Methods). Experimental treatments were conducted in triplicate and nitrogen was measured at initial (T0 hour) and final (T192 hour) time points.

Full size image

The potential for “Ca. T. autotrophicus” strain EF1 to respire NO3− and assimilate NH4+ was evaluated further under nitrogen limitation using oligotrophic seawater media (Figure 4, Supplementary Table 2). Cultures grew to the highest final cell densities (average 1.2 × 106 cells/ml) in media that was amended with 1 mM S2O32−, 100 μm NO3− and 5 μm NH4+ (Figure 4a). Cells grew to lower cell densities (average 6.2 × 105 cells/ml) and had slower growth rates when only S2O32− and NO3− were added to the media. There was a 1:1 conversion of NO3− to NO2− in treatments amended with S2O32− and NO3−, or with S2O32−, NO3− and NH4+ (Figure 4b). The amount of NO3− converted to NO2− increased four-fold (average increase from 5 μm to 22 μm) when NH4+ was added to the media and there was a decrease in NH4+ concentration over time (Figure 4b). Some growth was observed in S2O32− only controls. This is likely due to the low concentrations of NO3− (0.15 μM) and NH4+ (0.04 μm) present in oligotrophic seawater media (Supplementary Table 2). This experiment further confirmed that “Ca. T. autotrophicus” strain EF1 was unable to use NO2− as a terminal electron acceptor (Figures 4c and d). When cells were amended with NO2− instead of NO3−, no difference in growth was observed between amendments and controls. NO2− and NH4+ concentrations remained constant throughout the experiment. The potential to assimilate amino acids (Promega, Madison, WI, USA, Amino Acid Mixture) in oligotrophic seawater was also tested. Cultures were grown in media amended with 1 mM S2O32, 100 μM NO3− and an amino acid mixture (5 μM of N) (Supplementary Figure 4). There was no discernable difference between cultures amended with amino acids and unamended controls, indicating that strain EF1 prefers NH4+ for biogenic nitrogen.

Figure 4
figure 4

Nitrogen utilization of “Ca. T. autotrophicus” strain EF1 on oligotrophic media. Anaerobic growth on seawater media from the Sargasso Sea amended with (a) NO3− (100 μM), NH4+ (5 μM) and with only S2O32− (1 mM) and (b) NO2− (100 μM), NH4+ (5 μM) and with only S2O32− (1 mM). (c and d) Concentrations of NO3−, NO2− and NH4+ measured at initial (T0 hour) and final (T168 hour) time points in (a) and (b), respectively. All experimental treatments and nutrient measurements were conducted in triplicate.

Full size image

Roles in the marine nitrogen cycle

Evidence that cultured SUP05 produce NO2− suggests that related strains are a potential source of NO2− in OMZs. There is strong molecular evidence indicating that environmental SUP05 are capable of mediating sequential steps in denitrification. Because these environmental sequence data provide a fragmented view of a population of cells, it is also possible that different SUP05 cells carry out different steps in denitrification, depending on the diversity of SUP05 in a population, on the concentrations of substrates, and on the range of interactions within a community. Genetic and physiology data from this study suggest that a single strain of SUP05 carries out two non-sequential steps in denitrification. Both cultivation-dependent and cultivation-independent data indicate that the first step in denitrification (NO3− reduction to NO2−) is highly conserved (Canfield et al., 2010; Walsh et al., 2009; Hawley et al., 2014; Murillo et al., 2014), while the potential to carry out subsequent steps in denitrification are not always identified (Murillo et al., 2014). In addition, SUP05 are often most abundant in areas where NO2− accumulates (Canfield et al., 2010; Hawley et al., 2014; Murillo et al., 2014). Glaubitz et al. (2013) reported a positive correlation between SUP05 and nitrite concentrations in the Black Sea. Although heterotrophic NO3− respiration is currently considered the primary process leading to NO2− accumulation within OMZs, there is ample genetic and physiological data suggesting that sulfur-oxidizing chemoautotrophs from the SUP05 clade are a potential source of NO2−, fueling competing nitrogen removal processes in the ocean.

“Ca. T. autotrophicus” strain EF1 growth was also limited by NH4+. In the mid 1950s, Baalsrud and Baalsrud, and van Niel found that the S-oxidizing and NO3−-reducing bacterium Thiobacillus denitrificans required NH4+ or amino acids for biosynthesis (Baalsrud and Baalsrud, 1954; van Niel, 1955). Genes for NH4+ transport and amino acid assimilation were identified in the strain EF1 genome and in a SUP05 metagenome assembled from Saanich Inlet (Walsh et al., 2009). They were also expressed by SUP05 in Saanich Inlet and the Southern Ocean (Wilkins et al., 2013; Hawley et al., 2014). The biogenic nitrogen requirement of strain EF1 is low. For example, if “Ca. T. autotrophicus” strain EF1 cells have 10 fg of carbon/cell and are at Redfield ratios for carbon and nitrogen (106:16), we estimate that they required ~0.22 μM of nitrogen for biosynthesis in cultures grown on natural seawater media. By comparison, they reduced ~30 μM of NO3− to ~30 μM NO2− during respiration. There was no evidence that they respired NO2− or that they used amino acids instead of NH4+ for biosynthesis (Supplementary Figure 4B). These data support the conclusion that strain EF1 requires relatively high concentrations of NO3− for respiration and relatively low concentrations of NH4+ for biosynthesis.

“Ca. T. autotrophicus” strain EF1 also produced N2O. Members of the SUP05 bacteria expressed genes to produce N2O in regions of significant N2O cycling and emission (Codispoti et al., 2001). A recent study by Babbin and colleagues (Babbin et al., 2015) suggests that rapid N2O cycling in the suboxic ocean could lead to future increases in N2O emissions. Environmental sequence data indicate that SUP05 are broadly distributed and abundant in these regions. Growth experiments support field expression data, suggesting that N2O producing chemoautotrophic SUP05 have important roles in biologically driven nitrogen loss from the ocean (Walsh et al., 2009; Hawley et al., 2014; Murillo et al., 2014).

Conclusions

Data resulting from the cultivation of “Ca. T. autotrophicus” strain EF1 have expanded the roles of SUP05 in the marine nitrogen cycle. They suggest that SUP05 are a potential source of NO2− and sink for NH4+ in anoxic marine waters. In the conceptual metabolic coupling model of an OMZ proposed by Hawley et al. (2014), ammonia-oxidizing archaea are the source of NO2− in the upper and lower oxycline and SUP05 are the hypothesized source of NH4+ at and below the lower oxycline. If under the right conditions SUP05 produce NO2−, then elevated concentrations of NH4+ are not required to account for the elevated NO2− concentrations at and below the lower oxycline. There is strong evidence indicating that NO3− reduction is an independent process that can account for a significant fraction of NO2− accumulation in OMZs (Lam et al., 2009). We hypothesize that SUP05 contribute to the secondary NO2− maxima when NO3− fluxes are relatively high, ammonia concentrations are relatively low and oxygen concentrations fall below 4 μM. Our physiological data suggest that SUP05 cells require between 0.03 to 0.06 μM of NH4+ for every μM of NO2− produced (Figures 3 and 4). This is in stark contrast to ammonia-oxidation, which is 1:1. It suggests that chemoautotrophic members of the SUP05 clade are an important source of NO2− in OMZs where NH4+ is limiting.