Lucker 2010 Nitrospira.pdf

A Nitrospira metagenome illuminates the physiologyand evolution of globally importantnitrite-oxidizing bacteriaSebastian Lckera, Michael Wagnera, Frank Maixnera,1, Eric Pelletierb,c,d, Hanna Kocha, Benoit Vacherieb,Thomas Ratteie, Jaap S. Sinninghe Damstf, Eva Spieckg, Denis Le Paslierb,c,d, and Holger Daimsa,2

aDepartment of Microbial Ecology, Ecology Centre, University of Vienna, 1090 Vienna, Austria; bCommissariat lEnergie Atomique, Genoscope, 91000 Evry,France; cCentre National de la Recherche Scientique, UMR8030, 91000 Evry, France; dUniversit dEvry-Val-dEssonne, 91000 Evry, France; eDepartment forComputational Systems Biology, Ecology Centre, University of Vienna, 1090 Vienna, Austria; fDepartment of Marine Organic Biogeochemistry, RoyalNetherlands Institute for Sea Research, 1790 AB Den Burg, The Netherlands; and gBiozentrum Klein Flottbek, Mikrobiologie und Biotechnologie, University ofHamburg, 22609 Hamburg, Germany

Edited by Edward F. DeLong, Massachusetts Institute of Technology, Cambridge, MA, and approved June 18, 2010 (received for review March 25, 2010)

Nitrospira are barely studied and mostly uncultured nitrite-oxidizingbacteria, which are, according to molecular data, among the mostdiverse andwidespread nitriers in natural ecosystems and biologicalwastewater treatment. Here, environmental genomics was used toreconstruct the complete genome of Candidatus Nitrospira deuviifrom an activated sludge enrichment culture. On the basis of thisrst-deciphered Nitrospira genome and of experimental data, weshow that Ca. N. deuvii differs dramatically from other knownnitrite oxidizers in the key enzyme nitrite oxidoreductase (NXR),in the composition of the respiratory chain, and in the pathwayused for autotrophic carbon xation, suggesting multiple indepen-dent evolution of chemolithoautotrophic nitrite oxidation. Adapta-tions of Ca. N. deuvii to substrate-limited conditions include anunusual periplasmic NXR, which is constitutively expressed, andpathways for the transport, oxidation, and assimilation of simpleorganic compounds that allow a mixotrophic lifestyle. The reversetricarboxylic acid cycle as the pathway for CO2 xation and the lackof most classical defense mechanisms against oxidative stresssuggest that Nitrospira evolved from microaerophilic or even an-aerobic ancestors. Unexpectedly, comparative genomic analyses in-dicate functionally signicant lateral gene-transfer events betweenthe genus Nitrospira and anaerobic ammonium-oxidizing plancto-mycetes, which share highly similar forms of NXR and other pro-teins reecting that two key processes of the nitrogen cycle areevolutionarily connected.

environmental genomics | nitrication | Nitrospirae

Nitrication, the microbially catalyzed sequential oxidation ofammonia via nitrite to nitrate, is a key process of the bio-geochemical nitrogen cycle and of biological wastewater treatment.The second step of nitrication is carried out by chemo-lithoautotrophic nitrite-oxidizing bacteria (NOB), which are phy-logenetically heterogeneous (1) andoccur inawide rangeofaquaticand terrestrial ecosystems. Most studies on the physiology of NOBused pure cultures of Nitrobacter, which belong to the Alphapro-teobacteria (1), and complete genome sequences from NOB areavailable for three Nitrobacter strains (2, 3) and the marine gam-maproteobacterium Nitrococcus mobilis (GenBank accession no.NZ_AAOF00000000). However, cultivation-independent molec-ular methods revealed that Nitrospira, forming a deeply branchinglineage in the bacterial phylum Nitrospirae (4), are by far the mostdiverse and abundant NOB (5). In addition to their wide distribu-tion in natural habitats such as soils (6), sediments (7), the oceans(8), and hot springs (9), members of the genus Nitrospira are thepredominant NOB in wastewater treatment plants (5) and thusbelong to the microorganisms most relevant for biotechnology.The immense ecological and technical signicance of Nitrospira

contrasts with our scarce knowledge about these bacteria. As themajority ofNitrospira are uncultured, and the available cultures aredifcult to maintain, only a few studies have addressed their ecol-

ogy and physiology (e.g., 5, 10, 11). Furthermore, except for one137-kbp contig (12), genomic sequences from Nitrospira have notbeen obtained yet. This situation has been highly unsatisfactorybecause deeper insight into the biology of these elusive NOB iscrucial for a better understanding of nitrogen cycling in natural andengineered systems.Recently, a Nitrospira strain was enriched from activated

sludge and partly characterized (13). This organism, tentativelynamed Candidatus Nitrospira deuvii, belongs to Nitrospirasublineage I, which is most important for sewage treatment (5)but has no representative in pure culture. Here, the completegenome of Ca. N. deuvii was reconstructed from a meta-genomic library of the enrichment. More than two decades afterNitrospira were discovered (8), we provide an analysis ofa Nitrospira genome with previously unmatched insight into thebiology of Nitrospira, show striking differences in key metabolicpathways between Nitrospira and other NOB, and change thecurrent perception on the evolution of NO2

oxidation.

Results and DiscussionGenome Reconstruction. Quantitative FISH has shown that theNO2

-oxidizing enrichment consisted of 86%ofCa. N. deuvii anddid not contain other known NOB (13). The complete genome ofCa. N. deuvii was reconstructed from this enrichment by an en-vironmental genomics approach similar to that used for inferringthe genome sequence of the anaerobic ammonium-oxidizing bac-terium (anammox organism) Candidatus Kuenenia stuttgar-tiensis (14). The completeness and correct assembly of theNitrospira genomewas indicated by the retrieval of all 63 clusters oforthologous groups (COGs) of proteins, which are present in allgenomes in the current COG database (Fig. S1), by lack of suspi-cious redundancy in gene content, and by the presence of all es-sential genes in key biosynthetic pathways. The low frequency ofsingle nucleotide polymorphisms (about one per 500 kbp) stronglysuggests that the enrichment culture contained only one Nitrospirastrain. Key features of the genome are summarized in Table S1 andFig. S1. About 30% of the predicted coding sequences (CDS) have

Author contributions: S.L., M.W., D.L.P., and H.D. designed research; S.L., F.M., H.K., B.V.,and J.S.S.D. performed research; T.R. and E.S. contributed new reagents/analytic tools; S.L.,F.M., E.P., T.R., and H.D. analyzed data; and S.L., M.W., and H.D. wrote the paper.

The authors declare no conict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.

Data deposition: The sequence reported in this paper has been depositied in the GenBankdatabase (accession no. FP929003).1Present address: Institute for Mummies and the Iceman, EURAC Research, Viale Druso 1,39100 Bolzano, Italy.2To whom correspondence should be addressed. E-mail: [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1003860107/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1003860107 PNAS | July 27, 2010 | vol. 107 | no. 30 | 1347913484

MICRO

BIOLO

GY

no homologs in other organisms, reecting the distant relationshipof Nitrospira to other bacteria and the lack of genome sequencesfrom the genus Nitrospira in public databases. Furthermore, onlytwo lineages within the phylum Nitrospirae have been explored ona genomic level. The closest genome-sequenced relatives of Ca. N.deuvii belong to the genus Leptospirillum and are aerobic acido-philic iron oxidizers (1517). In addition, the genome sequence ofthe anaerobic sulfate reducer Thermodesulfovibrio yellowstonii(GenBank accession no. NC_011296), also belonging to theNitrospirae, is publicly available.

Nitrite Oxidation and Energy Metabolism. The key enzyme for NO2

oxidation by NOB is nitrite oxidoreductase (NXR), which shuttlestwo electrons per oxidized NO2

into the electron transport chain.InNitrobacter, NXR is an iron-sulfurmolybdoprotein (18) located atthe inner cell membrane and at the intracytoplasmic membranes(ICM). The reaction catalyzed by this NXR is reversible, so that theenzyme also reduces NO3

with electrons derived from organiccompounds. Depending on the applied purication method, thisNXR was found to consist of two (18) or three subunits witha supposed 221 stoichiometry (19). The -subunit (NxrA) isthought to contain the substrate-binding site with themolybdopterincofactor (Mo-co) (18, 19), whereas the -subunit (NxrB) with [Fe-S]clusters probably channels electrons from the - to the -subunit ordirectly to the membrane-integral electron transport chain (20).Nitrospira are Gram-negative bacteria lacking ICM (8). Al-

though no NO3-reducing activity has been demonstrated yet for

their nitrite-oxidizing system, the term NXR is used here to beconsistent with established terminology (2). The rst insight intothe nature of the Nitrospira NXR was obtained by studyinga pure culture of Nitrospira moscoviensis (21). Four major pro-teins were detected in membrane fractions showing a high NO2

-oxidizing activity in vitro. Antibodies originally raised againstNxrB of Nitrobacter bound to one of these proteins, which wasdesignated the NxrB of N. moscoviensis (21). Another proteinwith an apparent molecular mass of 130 kDa resembled theNxrA of Nitrobacter (115130 kDa). The other two proteins werenot further characterized. The NXR of N. moscoviensis was alsoshown to contain molybdenum and to be located at the inner cellmembrane, where it faces the periplasmic space (21).The genome of Ca. N. deuvii was screened for CDS with

a predicted molecular mass resembling the NxrA and NxrB of N.moscoviensis and similarity to known NO2

/NO3-binding molyb-

doenzymes, such as the NXR of Nitrobacter or bacterial nitratereductases (NARs). Two candidates were identied for each NxrAand NxrB (Table S2). The genes are colocalized in two clusters(nxrA1B1 and nxrA2B2), which are separated by 17 other CDSfrom each other. The amino acid identities are 86.6% for the twoNxrA and 100% for the two NxrB copies (the nxrB genes areidentical except for a synonymous single-base substitution). NxrA1and NxrA2 contain binding motifs for one [Fe-S] cluster and formolybdenum, which are indicative of the type II group in the di-methyl sulfoxide (DMSO) reductase family of Mo-cobindingenzymes (SI Results and Fig. S2 A and B). Five residues, which areconserved in the -subunits ofNARs and in theNxrAofNitrobacterand Nitrococcus, have been proposed to interact with NO2

/NO3

or to affect the conformation of the substrate entry channel (22).Except for one threonine, which is replaced by asparagine (Fig.S2B), these residues are conserved in both NxrA copies of Ca. N.deuvii, suggesting that the -subunit contains the substrate-binding site. Consistent with the periplasmic orientation of NXR inN. moscoviensis (21), NxrA1 and NxrA2 of Ca. N. deuvii containan N-terminal twin-arginine motif for export via the twin-arginineprotein translocation (Tat) pathway.Both NxrB copies of Ca. N. deuvii lack a predicted signal

peptide, but may be cotranslocated with NxrA into the periplasmby a hitchhiker mechanism as proposed for the -subunits ofother periplasmic Mo-cobinding enzymes (e.g., ref. 23). Fourcysteine-rich binding motifs for [Fe-S] clusters, which occur alsoin NxrB of Nitrobacter and Nitrococcus, were identied (Fig. S2 C

and D). Homologous [Fe-S] clusters mediate intramolecularelectron transfer in nitrate reductase A of Escherichia coli (24).All NxrA and NxrB copies of Ca. N. deuvii lack trans-

membrane helices, although NXR is membrane-associated inNitrospira (21). Theoretically, the /-complex might cluster witha membrane-bound terminal oxidase that receives electrons fromNXR. However, other enzymes in the DMSO reductase familycontain an additional membrane-integral -subunit, which is themembrane anchor of the holoenzyme and channels electronsbetween the -subunit and the electron transport chain via oneor two hemes (25). Four proteins encoded by Ca. N. deuviicould be heme-containing subunits of NXR (Table S2). Each hasone transmembrane domain and an N-terminal signal peptidefor translocation via the Sec pathway. The largest candidate(66.7 kDa) is a c-type cytochrome with two predicted heme-binding sites. The other three proteins are smaller (29.734.3kDa) and remotely similar to the -subunit of chlorate reductase,which contains one b-type heme (26). These genes are not indirect proximity of the nxrAB clusters, but the predicted molec-ular masses of their products resemble the two uncharacterizedmajor proteins from N. moscoviensis membrane extracts (62 and29 kDa) (21). Their biological functions and the exact compo-sition of NXR await experimental clarication.The sequenced Nitrobacter genomes encode a peptidyl-prolyl

cis-trans isomerase (NxrX) proposed to assist in the folding ofNXR (2, 3). Ca. N. deuvii lacks a homolog of NxrX, but oneCDS is similar to chaperones involved in the assembly of otherDMSO reductase-family enzymes (26). It is located directly up-stream of one putative membrane-integral NXR subunit (TableS2) and could play a role in NXR maturation.On the basis of biochemical (21) and genomic data forNitrospira,

a membrane-bound periplasmic NXR that consists of at least twosubunits is proposed (Fig. 1). High-potential electrons from NO2

are probably transferred to cytochrome (cyt.) c as in Nitrobacter(19) and then to a terminal cyt. c oxidase (Fig. 1). InNitrobacter, theterminal oxidase is of the aa3 type (3). The lack of detectable cyt.a in Nitrospira cultures (4, 8) and of genes coding for a-type cyto-chromes inCa. N. deuvii implies thatNitrospira possess a differenttype of terminal oxidase. Intriguingly, the genome does not encodeany known heme-copper oxidase, which could transfer electronsfrom cyt. c to O2. However,Ca. N. deuvii has a heterodimeric cyt.bd quinol oxidase (genes cydA and cydB; Table S2) that could re-ceive electrons derived from low-potential donors, such as organiccarbon, via the quinol pool (Fig. 1). The genome contains fouradditional CDS that resemble the CydA subunit of cyt. bd oxidases,but can be distinguished from the canonical proteins by phyloge-netic analysis (Fig. S3A). We refer to these uncharacterizedproteins as putative cyt. bd-like oxidases. They contain 14 pre-dicted transmembrane helices and several histidines that mayserve as heme ligands (Fig. S3B). Interestingly, one of these CDS(Nide0901) also contains a putative copper (CuB)-binding site (Fig.S3B). This motif is characteristic for the binuclear center of heme-copper cyt. c oxidases, and it is thus tempting to speculate thatNide0901 could replace the lacking canonical heme-copper oxi-dases in Nitrospira (Fig. 1). The proposed function of Nide0901 asterminal oxidase gains further support from transcriptional analy-sis. High levels of nide0901 mRNA were detected in the presenceof the electron donor NO2

and the terminal electron acceptor O2,whereas the transcription of this gene decreased markedly in theabsence of these substrates (Fig. S3D). An alternative to a mem-brane-bound terminal oxidase would be a soluble cytoplasmic O2reductase, but this is not supported by the genomic data.The genome-based model of energy metabolism in Ca. N.

deuvii composes a branched respiratory chain for NO2 oxi-

dation for the use of low-potential electron donors such as or-ganic substrates and for reverse electron transport (Fig. 1). Inaddition, two copper-containing nitrite reductases (NirK; TableS2) were identied. NirK forms NO from NO2

in denitrifyingorganisms, including other nitriers (e.g., 27). Although de-nitrication by Ca. N. deuvii has not been experimentally

13480 | www.pnas.org/cgi/doi/10.1073/pnas.1003860107 Lcker et al.

demonstrated, the nirK genes indicate that this organism maydenitrify NO2

, for example, by using organic substrates as theelectron donor. If NXR works reversibly in Nitrospira, de-nitrication could also start from NO3

. Other denitricationgenes were not found. In Nitrobacter, NO may function in reverseelectron transport (28) and in electron ux regulation (27). Itremains unclear whether NO plays similar physiological rolesin Nitrospira.

Expression of NXR. To test whether NO2 induces the expression

of NXR, RNA was extracted from enrichment biomass duringstarvation in NO2

-free medium and after addition of NO2, and

nxrB mRNA was analyzed by reverse transcription (RT)PCR.Interestingly, a low level of nxrB mRNA was detected afterstarvation for 11 d in NO2

-free medium (Fig. S2E). Addition ofNO2

led to an increased transcription of nxrB, whereas the levelof 16S rRNA from Ca. N. deuvii did not change markedly (Fig.S2E). NxrB protein was detected even after starvation in NO2

-free medium for 110 d, and its level increased markedly uponaddition of NO2

(Fig. S2E). These results support the annota-tion of NXR. The constitutive expression of NXR should enableCa. N. deuvii to use NO2

, whose concentration usually is lowand uctuates in natural habitats, immediately after this energysource becomes available.

Autotrophy. NOB of the genus Nitrobacter (2) and, on the basis ofgenomic data, also Nitrococcus use the CalvinBensonBassham(CBB) cycle for CO2 xation. The key enzymes of this pathway areribulose-1,5-bisphosphate carboxylase (RubisCO) and ribulose-5-phosphate kinase. Nitrospira also grow chemolithoautotrophicallyonNO2

andCO2 (4, 13), but their pathway for CO2 xationwas notidentied previously. Ca. N. deuvii encodes a form IV RubisCO-like protein (Fig. S4A) lacking functional key residues of canonicalRubisCO (Fig. S4B). In Bacillus subtilis, a form IV RubisCO-likeprotein has no bona de carboxylating activity (29). The absence ofother genes similar to RubisCO and of ribulose-5-phosphate kinasesuggests that the CBB cycle does not operate in Ca. N. deuvii.Instead, all genes of the reductive tricarboxylic acid (rTCA) cycle arepresent, including the key enzymes ATP-citrate lyase and 2-oxoglutarate:ferredoxin oxidoreductase (OGOR), and also pyru-vate:ferredoxin oxidoreductase (POR) (Table S2 and SI Results).Operation of the rTCA cycle in Ca. N. deuvii was conrmed

by the small carbon isotopic fractionation factor () betweenbiomass and CO2 of 26 (Table S3), typical for the rTCA cycle

(30). Furthermore, the abundant (80% of all fatty acids) andcharacteristic straight-chain fatty acid for Ca. N. deuvii, C16:15 (13), was 36 enriched relative to the biomass, whereasisoprenoid lipids were 4 depleted (Table S3). This trend ofmore enriched straight-chain lipids is unusual for almost allcarbon xation pathways except for the rTCA cycle (31).As POR and OGOR generally are O2-sensitive enzymes (32),

the rTCA cycle is found mainly in anaerobic organisms, and itspresence in an aerobic nitrier seems surprising. However, thispathway is functional in some microaerophilic autotrophs suchas Hydrogenobacter thermophilus (33), and it was identied inLeptospirillum genomes (16, 17).H. thermophilus has two isozymesof OGOR, a two-subunit enzyme needed under anoxic condi-tions and a more O2-tolerant unique ve-subunit form, whichsupports mainly aerobic growth (34), and it also has an unusualve-subunit POR (35). Highly similar ve-subunit OGOR andPOR in Ca. N. deuvii (SI Results) and Leptospirillum (16) mayallow the rTCA cycle to function in these aerobic members of theNitrospirae phylum. Thus, on the basis of genomic and isotopicdata, Nitrospira x CO2 via the rTCA cycle and represent the onlynitrier for which this pathway has been detected.

Use of Organic Substrates. Ca. N. deuvii and Nitrospira marinabenet from simple organic compounds in nitrite media (8, 13),and uncultured Nitrospira in sewage plants take up pyruvate (5).However, it is unknown whether Nitrospira use organic substratesonly as carbon sources or also for energy generation. Inter-estingly, the Ca. N. deuvii genome encodes pathways for thecatabolic degradation and for the assimilation of acetate, pyru-vate, and formate (Fig. S5 and SI Results), and candidate geneswere found for the degradation of branched amino acids. As theEmbdenMeyerhofParnas pathway is complete, Ca. N. deuviishould be able to metabolize hexose sugars. This is consistentwith carbon being stored as glycogen (SI Results). Two of thethree sequenced Nitrobacter genomes also contain the completeglycolysis pathway (3), but growth of Nitrobacter on sugars hasnot been reported. Whether Ca. N. deuvii can take up and usesugars should depend mainly on functional sugar transport sys-tems. The genome indeed contains putative sugar transporters(Table S1), but their function remains to be determined.The oxidative tricarboxylic acid (oTCA) cycle shares most

enzymes with the rTCA cycle except for citrate synthase and the2-oxoglutarate dehydrogenase complex (ODH). Ca. N. deuviiencodes citrate synthase but apparently lacks ODH, which may,

Fig. 1. Genome-based model of energymetabolism in Ca. N. deuvii. Orangearrows indicate electron ow in the oxi-dative branches of the electron transportchain; green arrows indicate reverse elec-tron transport from NO2

to NAD+.Dashed black lines point out that themembrane-integral subunit of NXR is un-certain. Dashed orange arrows show hy-pothetical possibilities for electron owfrom NXR to the putative cyt. c-oxidase.nH+ indicates that the number of trans-located protons is unknown because theH+/e ratioof the respective complexes hasnot been determined for Nitrospira. FRD,fumarate reductase; SDH, succinate de-hydrogenase. See Table S2 for a list of theinvolved proteins.

Lcker et al. PNAS | July 27, 2010 | vol. 107 | no. 30 | 13481

MICRO

BIOLO

GY

however, be replaced by OGOR (Table S2 and SI Results). Acomplete oTCA cycle was reported for Nitrobacter (28), in-dicating that this pathway is not unusual in NOB.Purely heterotrophic growth of Nitrospira has not been ob-

served yet. However, if all potentially involved genes are func-tional, Ca. N. deuvii benets from a mixotrophic lifestyle usingorganic compounds from sewage in addition to NO2

and CO2.

Stress Response and Defense. Ca. N. deuvii is exposed to a pleth-ora of potentially toxic substances in sewage. Accordingly, thegenome encodes multidrug efux systems and transporters forheavy metals, organic solvents, and antimicrobials (Table S1), andit contains genes for cyanate and arsenic resistance (Table S2 andSI Results). As shown previously (12), Ca. N. deuvii has a func-tional chlorite dismutase that could degrade ClO2

in pollutedenvironments, in chlorinated activated sludge, or in the proximityof chlorate-reducing microbes. Most intriguingly, Ca. N. deuviilacks key genes for protection from reactive oxygen species (ROS)present in most aerobic organisms. No catalase, superoxide dis-mutase, or superoxide reductase was found. Two cyt. c perox-idases and several thioredoxin-dependent peroxiredoxins couldfunction as H2O2 scavengers (Table S2 and SI Results). Protectionfrom O2

and H2O2 might be conferred by manganese [Mn(II)](36). Indeed, the required permease for manganese import wasidentied in the genome. Bacterioferritin and carotenoids (TableS2) could also contribute to protection from radicals and ROS.Moreover, the intracellular O2 level could be kept low by thecanonical cyt. bd oxidase. Homologs in other organisms havea high afnity to O2 and contribute to oxidative stress protection(37). Growth ofNitrospira in biolms and ocs (e.g., ref. 13) couldoffer additional protection from ambient O2.Ca. N. deuvii carries one region of clustered, regularly inter-

spaced short palindromic repeats (CRISPRs) and CRISPR-associated (cas) genes for phage defense (38). TheCRISPRrepeatsof Ca. N. deuvii show no sequence similarity to those of Lep-tospirillum groups II and III, which also differ in their Cas proteins(17), suggesting that this defense mechanism was independentlyacquired by different members of the Nitrospirae phylum.

Ecophysiology and Evolutionary History of Nitrospira. The NXRs ofNitrobacter, Nitrococcus, and Nitrospira differ in their subcellularlocalization and phylogenetic position within the DMSO re-ductase family. The NXRs of Nitrobacter and Nitrococcus areclosely related to NARs. They are associated with the cytoplasmicmembrane and ICMwith the active site facing the cytoplasm (39).The unique NXR of Nitrospira does not cluster with the NARs(Fig. 2). It is also attached to the cytoplasmic membrane, but isoriented toward the periplasmic space (21; this study). The peri-plasmic orientation should be energetically advantageous becauseproton release by NO2

oxidation in the periplasm and concom-itant proton consumption by O2 reduction in the cytoplasm con-tribute to the membrane potential (Fig. 1). Furthermore, onlya cytoplasmic NXR requires the transport of NO2

and NO3 in

opposite directions across the inner membrane. Accordingly,putative NO2

/NO3 transporters are found in all sequenced

Nitrobacter genomes (3) and in Nitrococcus. Their substrate af-nities and turnover rates could be limiting factors for NO2

ox-idation by these NOB. This and the catalytic properties of NXRcould explain the relatively high apparent Km(NO2

) value ofNitrobacter (11). In contrast, the predicted NO2

and NO3

transporters of Ca. N. deuvii (Table S1) most likely play no rolein nitrite oxidation but are required only for nitrogen assimilationand resistance against excess nitrite (SI Results).Consistent with the predicted advantages of their periplasmic

NXR, Nitrospira are better adapted to low NO2 concentrations

(10, 11), which also were key to the selection against coexistingNitrobacter during enrichment (13). As NO2

rarely accumulatesin natural environments, the highly efcient use of this substratemost likely is a main reason for the competitive success and widenatural distribution of Nitrospira.

The use of different key enzymes and pathways (e.g., CO2 xa-tion) by Nitrospira in contrast to the proteobacterial NOB Nitro-bacter and Nitrococcus suggests that chemolithoautotrophic NO2

oxidation evolved independently in these lineages. On the basis ofthe close phylogenetic afliation of Nitrobacter and Nitrococcus tophototrophic Proteobacteria, which also possess ICM, Teske et al.(1) hypothesized that these NOB were derived from phototrophicancestors. Indeed, a recently isolated anaerobic phototroph, whichuses NO2

as the electron donor, is closely related to Nitrococcus(40). A cytoplasmically oriented NXR would probably be no dis-advantage for phototrophic NOB where the membrane potential issustained mainly by light-driven cyclic electron ow. The orienta-tion of NXR may not easily be reversed, because it intimatelyaffects the interaction with downstream components of the electrontransport chain. Hence, the conservation of a cytoplasmic NXRduring the transition from phototrophy to chemolithotrophy couldexplain the orientation of NXR in Nitrobacter and Nitrococcus. Incontrast and consistent with the absence of ICM in Nitrospira, nophototrophic relative of Nitrospira is known and we hypothesizethat the capability to gain energy from NO2

oxidation has evolvedin this lineage from an anaerobic nonphototrophic ancestor. Ananaerobic or microaerophilic origin of Nitrospira would be consis-tent with the rTCA cycle, the presence of the anaerobic cobalaminbiosynthesis pathway (Table S2), and the lack of classic defensemechanisms against ROS. Additional support for this hypothesisstems from estimating genus divergence times within the Nitro-spirae phylum by using 16S rRNA as the molecular clock (SIResults). Extant Nitrospira are active at low dissolved O2 levels inbioreactors and might still prefer hypoxic conditions (41).Intriguingly, comparative genomics revealed an unexpected

evolutionary link between Nitrospira and anammox organisms.For example, the closest homolog of the NXR of Ca. N. deuviiwas found in Ca. K. stuttgartiensis (Fig. 2). NO2

oxidation is anintegral step of the anammox metabolism where it replenishes theelectron transport system (14), and this NXR-like protein is theonly candidate for a NO2

-oxidizing enzyme in the Kuenenia ge-nome. Its -subunit contains the signature residues of NO2/NO3

-binding molybdoenzymes (22) (Fig. S2 A and B). TheNXRs of Nitrospira and Kuenenia are highly similar (amino acididentities are 57.457.7% for the -subunit and 62.5% for the-subunit) and form a monophyletic lineage in the tree of type IIenzymes of the DMSO reductase family (Fig. 2). In addition, bothCa. N. deuvii and Ca. K. stuttgartiensis have the putative chap-erone for NXR assembly in analogy to NxrX of Nitrobacter. Ca. K.stuttgartiensis also has a putative cyt. bd-like oxidase, which is theclosest relative of the four cyt. bd-like oxidases of Ca. N. deuvii(Fig. S3A). Interestingly, its gene is located in close proximity tonxrA, nxrB, two putative membrane subunits of NXR, and thechaperone in the Kuenenia genome (Fig. 3). The same regioncontains a monoheme cyt. c-like protein and three proteins ofunknown function, which also have highly similar homologs inCa.N. deuvii (Fig. 3). Thus, both organisms share a set of highlysimilar proteins that function in NO2

oxidation and probably inelectron transport and respiration, and these genes are clusteredas a small metabolic island in the anammox genome. As anammoxorganisms are planctomycetes and consequently not closely re-lated to the Nitrospirae (14), these observations are strongly in-dicative of a horizontal gene transfer (HGT) that establishedNXR and the other proteins in both lineages. Consistent with thefundamental importance of the transferred genes for the basicmetabolism of Nitrospira and anammox, this HGT apparentlyoccurred early during the evolution of these lineages as no re-markable deviation in GC content or codon use of the respectivegenes was observed in either organism.To explore further the inuence of vertical gene transfer and

HGT on the evolutionary history of Ca. N. deuvii, we calculatedphylogenies for each protein of Ca. N. deuvii and identied theorganism encoding the respective most closely related homolog(Fig. S6). Most remarkably, in this analysis Ca. K. stuttgartiensiswas the single organism that shared the highest number of closest


homologs (71 hits) with Ca. N. deuvii and thus exchanged, com-pared with all other organisms for which genome sequences areavailable, the most genes withNitrospira via HGT. Surprisingly, the71 hits even exceed the number of best hits with members of theNitrospirae phylum, namely Thermodesulfovibrio (67 hits) and dif-ferent Leptospirillum strains (3966 hits). These ndings illustratea surprisingly small set of the most closely related homologs in theNitrospirae, most likely reecting the dramatically different eco-logical niches inhabited by the genera afliated with this phylum.Taken together, the metagenome sequence of Ca. N. deuvii

revealed that this globally important nitrite oxidizer differs fun-

damentally in its enzymatic repertoire (unusual NXR and putativeterminal oxidase) and metabolic pathways (rTCA for autotrophy)from all other known nitriers, but strikingly exploits almost thesame gene repertoire for NO2

oxidation as the anammox or-ganism Ca. K. stuttgartiensis. The unique genomic features ofNitrospira have already provided some well-supported hypothesesfor its competitive success in most nitrifying ecosystems and sug-gest that Nitrospira are well adapted to hypoxic environmentalniches where nitrite oxidation has rarely been studied until now.From an applied perspective, the lack of common protectionmechanisms against oxidative stress in Nitrospira implies that

Fig. 2. Maximum-likelihood treeshowing the phylogenetic posi-tioning of selected type II enzymesof the DMSO reductase family. Forphylogenetic analysis of the cata-lytic () subunits, 1,308 amino acidpositions were considered. Namesofvalidatedenzymesare indicated:Nxr, nitrite oxidoreductase; Nar,membrane-bound respiratory ni-trate reductase; Pcr, perchloratereductase; Ebd, ethylbenzene de-hydrogenase;Ddh,dimethylsuldedehydrogenase; Clr, chlorate re-ductase; Ser, selenate reductase.Parentheses contain the number ofsequences within a group or theaccession number, respectively.

Fig. 3. Schematic of the genomic regions in Ca. K. stuttgartiensis and Ca. N. deuvii, which contain shared genes coding for NXR, putative cyt. bd-likeoxidases and electron carriers, and proteins of unknown function. Solid lines connect genes that are the closest homologs on the basis of protein phylogeny.Their predicted functions are in boldface. Dashed lines connect similar genes that are not the closest relatives in the respective phylogenetic protein trees.Predicted CDS and connecting lines are colored according to functional classes. CDS and intergenic regions are drawn to scale.

Lcker et al. PNAS | July 27, 2010 | vol. 107 | no. 30 | 13483

MICRO

BIOLO

GY

a good aeration control is crucial for maintaining stable and activepopulations of these organisms in engineered systems.

Materials and MethodsGenomic Sequencing and Annotation. Metagenome sequencing and the re-construction of the whole Ca. N. deuvii genome were carried out byGenoscope (SI Materials and Methods). The MaGe software system (42) wasused for the prediction, automatic annotation, and manual annotation re-nement of all CDS as described in SI Materials and Methods.

Phylogenetic Analyses.Aminoacid sequencesof type IIDMSOreductase-familyenzymes, of RubisCO and RubisCO-like proteins, and of cyt. bd and cyt. bd-likeoxidases were aligned, and phylogenetic trees were computed by using ARB(43). For the calculation of phylogenetic trees for each protein in the pro-teome, PhyloGenie (44) was used. For details, see SI Materials and Methods.

Expression Analysis of NxrB and the Putative Terminal Cyt. c Oxidase (Nide0901).Ca. N. deuvii enrichment biomass was incubated in mineral media with orwithout NO2

and, for Nide0901, also under oxic or anoxic conditions as

described in SI Materials and Methods. Following total RNA extraction, 16SrRNA of Nitrospira and nxrB or nide0901 transcripts were detected by RTPCR (SI Materials and Methods). Translation of NxrB was shown by Westernblotting with a monoclonal antibody that binds to the NxrB of Nitrospira(21) (SI Materials and Methods).

Stable Carbon Isotopic Fractionation. The isotopic fraction of Ca. N. deuvii wasmeasured followingmethods published earlier (45) (SI Materials andMethods).

ACKNOWLEDGMENTS. We thank Lisa Stein for analyses of nirK genes andPeter Bottomley, Jim Hemp, Jim Prosser and Andreas Schramm for helpfuldiscussions. Christiane Dorninger, Alexander Galushko, Christian Baranyi, JanDolinek, Patrick Tischler, Irene Rijpstra, and Michiel Kienhuis are acknowl-edged for technical support. This work was supported by the Vienna Scienceand Technology Fund (Wiener Wissenschafts-, Forschungs-, und Technolo-giefonds, Grant LS 216 to H.D., S.L., and F.M. and by Grant LS09-40 to H.D.,H.K., and S.L.), the Austrian Research Fund (Fonds zur Frderung der Wis-senschaftlichen Forschung, Grant S10002-B17 to H.D., M.W., S.L., and F.M.),and the German Research Foundation (Deutsche Forschungsgemeinschaft,Grant SP 667/3-1 to E.S.).

1. Teske A, et al. (1994) Evolutionary relationships among ammonia- and nitrite-oxidizing bacteria. J Bacteriol 176:66236630.

2. Starkenburg SR, et al. (2006) Genome sequence of the chemolithoautotrophic nitrite-oxidizing bacterium Nitrobacter winogradskyi Nb-255. Appl Environ Microbiol 72:20502063.

3. Starkenburg SR, et al. (2008) Complete genome sequence of Nitrobacterhamburgensis X14 and comparative genomic analysis of species within the genusNitrobacter. Appl Environ Microbiol 74:28522863.

4. Ehrich S, Behrens D, Lebedeva E, Ludwig W, Bock E (1995) A new obligatelychemolithoautotrophic, nitrite-oxidizing bacterium, Nitrospira moscoviensis sp. nov.and its phylogenetic relationship. Arch Microbiol 164:1623.

5. Daims H, Nielsen JL, Nielsen PH, Schleifer KH, Wagner M (2001) In situcharacterization of Nitrospira-like nitrite-oxidizing bacteria active in wastewatertreatment plants. Appl Environ Microbiol 67:52735284.

6. Freitag TE, Chang L, Clegg CD, Prosser JI (2005) Inuence of inorganic nitrogenmanagement regime on the diversity of nitrite-oxidizing bacteria in agriculturalgrassland soils. Appl Environ Microbiol 71:83238334.

7. Altmann D, Stief P, Amann R, De Beer D, Schramm A (2003) In situ distribution andactivity of nitrifying bacteria in freshwater sediment. Environ Microbiol 5:798803.

8. Watson SW, Bock E, Valois FW, Waterbury JB, Schlosser U (1986) Nitrospira marinagen. nov. sp. nov.: A chemolithotrophic nitrite-oxidizing bacterium. Arch Microbiol144:17.

9. Lebedeva EV, et al. (2005) Moderately thermophilic nitrifying bacteria from a hotspring of the Baikal rift zone. FEMS Microbiol Ecol 54:297306.

10. Maixner F, et al. (2006) Nitrite concentration inuences the population structure ofNitrospira-like bacteria. Environ Microbiol 8:14871495.

11. Schramm A, de Beer D, van den Heuvel JC, Ottengraf S, Amann R (1999) Microscaledistribution of populations and activities of Nitrosospira and Nitrospira spp. alonga macroscale gradient in a nitrifying bioreactor: Quantication by in situhybridization and the use of microsensors. Appl Environ Microbiol 65:36903696.

12. Maixner F, et al. (2008) Environmental genomics reveals a functional chloritedismutase in the nitrite-oxidizing bacterium Candidatus Nitrospira deuvii. EnvironMicrobiol 10:30433056.

13. Spieck E, et al. (2006) Selective enrichment and molecular characterization ofa previously uncultured Nitrospira-like bacterium from activated sludge. EnvironMicrobiol 8:405415.

14. Strous M, et al. (2006) Deciphering the evolution and metabolism of an anammoxbacterium from a community genome. Nature 440:790794.

15. Tyson GW, et al. (2004) Community structure and metabolism through reconstructionof microbial genomes from the environment. Nature 428:3743.

16. Levicn G, Ugalde JA, Ehrenfeld N, Maass A, Parada P (2008) Comparative genomicanalysis of carbon and nitrogen assimilation mechanisms in three indigenousbioleaching bacteria: Predictions and validations. BMC Genomics 9:581.

17. Goltsman DS, et al. (2009) Community genomic and proteomic analyses ofchemoautotrophic iron-oxidizing Leptospirillum rubarum (group II) andLeptospirillum ferrodiazotrophum (group III) bacteria in acid mine drainagebiolms. Appl Environ Microbiol 75:45994615.

18. Meincke M, Bock E, Kastrau D, Kroneck PMH (1992) Nitrite oxidoreductase fromNitrobacter hamburgensis: Redox centers and their catalytic role. Arch Microbiol 158:127131.

19. Sundermeyer-Klinger H, Meyer W, Warninghoff B, Bock E (1984) Membrane-boundnitrite oxidoreductase of Nitrobacter: Evidence for a nitrate reductase system. ArchMicrobiol 140:153158.

20. Kirstein K, Bock E (1993) Close genetic relationship between Nitrobacterhamburgensis nitrite oxidoreductase and Escherichia coli nitrate reductases. ArchMicrobiol 160:447453.

21. Spieck E, Ehrich S, Aamand J, Bock E (1998) Isolation and immunocytochemicallocation of the nitrite-oxidizing system in Nitrospira moscoviensis. Arch Microbiol 169:225230.

22. Martinez-Espinosa RM, et al. (2007) Look on the positive side! The orientation,identication and bioenergetics of Archaeal membrane-bound nitrate reductases.FEMS Microbiol Lett 276:129139.

23. McDevitt CA, Hugenholtz P, Hanson GR, McEwan AG (2002) Molecular analysis ofdimethyl sulphide dehydrogenase from Rhodovulum suldophilum: Its place in thedimethyl sulphoxide reductase family of microbial molybdopterin-containingenzymes. Mol Microbiol 44:15751587.

24. Blasco F, et al. (2001) The coordination and function of the redox centres of themembrane-bound nitrate reductases. Cell Mol Life Sci 58:179193.

25. Rothery RA, Workun GJ, Weiner JH (2008) The prokaryotic complex iron-sulfurmolybdoenzyme family. Biochim Biophys Acta 1778:18971929.

26. Thorell HD, Stenklo K, Karlsson J, Nilsson T (2003) A gene cluster for chloratemetabolism in Ideonella dechloratans. Appl Environ Microbiol 69:55855592.

27. Starkenburg SR, Arp DJ, Bottomley PJ (2008) Expression of a putative nitrite reductaseand the reversible inhibition of nitrite-dependent respiration by nitric oxide inNitrobacter winogradskyi Nb-255. Environ Microbiol 10:30363042.

28. Bock E, Koops HP, Harms H, Ahlers B (1991) The biochemistry of nitrifying organisms.Variations in Autotrophic Life, eds Shively JM, Barton LL (Academic Press, London), pp171200.

29. Ashida H, et al. (2003) A functional link between RuBisCO-like protein of Bacillus andphotosynthetic RuBisCO. Science 302:286290.

30. Quandt L, Gottschalk G, Ziegler H, Stichler W (1977) Isotope discrimination byphotosynthetic bacteria. FEMS Microbiol Lett 1:125128.

31. van der Meer MTJ, Schouten S, Damste JSS (1998) The effect of the reversedtricarboxylic acid cycle on the C-13 contents of bacterial lipids. Org Geochem 28:527533.

32. Campbell BJ, Engel AS, Porter ML, Takai K (2006) The versatile -proteobacteria: Keyplayers in sulphidic habitats. Nat Rev Microbiol 4:458468.

33. Shiba H, Kawasumi T, Igarashi Y, Kodama T, Minoda Y (1985) The CO2 assimilation viathe reductive tricarboxylic-acid cycle in an obligately autotrophic, aerobic hydrogen-oxidizing bacterium, Hydrogenobacter thermophilus. Arch Microbiol 141:198203.

34. Yamamoto M, Arai H, Ishii M, Igarashi Y (2006) Role of two 2-oxoglutarate:ferredoxinoxidoreductases in Hydrogenobacter thermophilus under aerobic and anaerobicconditions. FEMS Microbiol Lett 263:189193.

35. Ikeda T, et al. (2006) Anabolic ve subunit-type pyruvate:ferredoxin oxidoreductasefrom Hydrogenobacter thermophilus TK-6. Biochem Biophys Res Commun 340:7682.

36. Horsburgh MJ, Wharton SJ, Karavolos M, Foster SJ (2002) Manganese: Elementaldefence for a life with oxygen. Trends Microbiol 10:496501.

37. Das A, Silaghi-Dumitrescu R, Ljungdahl LG, Kurtz DM, Jr (2005) Cytochrome bdoxidase, oxidative stress, and dioxygen tolerance of the strictly anaerobic bacteriumMoorella thermoacetica. J Bacteriol 187:20202029.

38. Barrangou R, et al. (2007) CRISPR provides acquired resistance against viruses inprokaryotes. Science 315:17091712.

39. Spieck E, Bock E (2005) The lithoautotrophic nitrite-oxidizing bacteria. BergeysManual of Systematic Bacteriology, eds Staley JT, et al. (Springer Science+BusinessMedia, New York), Vol 2, 2nd Ed, pp 149153.

40. Grifn BM, Schott J, Schink B (2007) Nitrite, an electron donor for anoxygenicphotosynthesis. Science 316:1870.

41. Park HD, Noguera DR (2008) Nitrospira community composition in nitrifying reactorsoperated with two different dissolved oxygen levels. J Microbiol Biotechnol 18:14701474.

42. Vallenet D, et al. (2006) MaGe: A microbial genome annotation system supported bysynteny results. Nucleic Acids Res 34:5365.

43. Ludwig W, et al. (2004) ARB: A software environment for sequence data. NucleicAcids Res 32:13631371.

44. Frickey T, Lupas AN (2004) PhyloGenie: Automated phylome generation and analysis.Nucleic Acids Res 32:52315238.

45. Schouten S, et al. (2004) Stable carbon isotopic fractionations associated withinorganic carbon xation by anaerobic ammonium-oxidizing bacteria. Appl EnvironMicrobiol 70:37853788.


Supporting InformationLcker et al. 10.1073/pnas.1003860107SI Results[Fe-S] and Molybdenum Ligands in NxrA.Both nitrite oxidoreductase(NXR) -subunits ofCandidatusNitrospira deuvii (Ca. N. deuvii)(NxrA1 and NxrA2) contain, close to the N terminus, one cyste-ine-rich [Fe-S] binding motif (C-X3-D-X3-C-X39-C) (Fig. S2A).This motif resembles the consensus [Fe-S] bindingmotif (H/C-X3-C-X3-C-Xn-C) of the type II group in the dimethyl sulfoxide(DMSO) reductase family of molybdopterin-binding enzymes(1, 2). In Ca. N. deuvii, the second cysteine residue of the con-sensus motif is replaced by aspartate, which can also function asa [Fe-S] ligand as shown for a ferredoxin of Pyrococcus furiosus(3). This aspartate residue also occurs in the phylogeneticallyclosely related NXR-like proteins of Ca. K. stuttgartiensis, Hydro-genobaculum, and Beggiatoa (Fig. S2A). On the basis of thesedata, we propose a consensus [Fe-S] binding motif (H/C-X3-D/C-X3-C-Xn-C) for the type II group in the DMSO reductase family.A highly conserved aspartate residue, which functions as amolybdenum ligand in nitrate reductase A (subunit NarG) ofEscherichia coli and most likely also in the other type II DMSOreductase-family enzymes (4), is present in both NxrA copies ofCa. N. deuvii (Asp278 in NxrA1) (Fig. S2B).

Key Enzymes of the Reductive and Oxidative Tricarboxylic Acid Cycles.Key enzymes for CO2 xation via the reductive tricarboxylic acid(rTCA)cycle are oxoglutarate:ferredoxinoxidoreductase (OGOR),pyruvate:ferredoxin oxidoreductase (POR), fumarate reductase(FRD), and ATP-citrate lyase (ACL). OGOR and POR usuallyconsist of one to four distinct subunits (5, 6). The Ca. N. deuviigenome contains three gene clusters encoding 2-oxoacid:ferredoxinoxidoreductases that could have POR or OGOR activity. Onecluster consists of the -, -, and fused /-subunits of a putativefour-subunit POR, which is similar to homologs in Pelobacter,Desulfotalea (both Deltaproteobacteria), and Ca. K. stuttgartiensis(Planctomycetes). Alternatively, these three coding sequences(CDS) might represent a 2-oxoisovalerate:ferredoxin oxidoreduc-tase, which could function in the degradation of branched-chainamino acids. Each of the other two gene clusters consists of veCDS, which are highly similar to ve-subunit forms of POR andOGORfound recently inmembers of the phylumAquicae, such asAquifex aeolicus andHydrogenobacter thermophilus (5, 7), andalso inLeptospirillum (8). One of these clusters inCa. N. deuvii is slightlymore similar to OGOR, whereas the other cluster is more likely tobe POR (Table S2), suggesting that both enzymes are present inNitrospira.Four CDS in the Ca. N. deuvii genome code for subunits of

FRD or the highly similar counterpart of this enzyme in the oxi-dative tricarboxylic acid (oTCA) cycle, succinate dehydrogenase(SDH) (Table S2). Two of these CDS are homologous to thehighly conserved fumarate- or succinate-binding avoproteinsubunit FrdA/SdhA, one is homologous to the iron-sulfur subunitFrdB/SdhB, and one is similar to subunit FrdE/SdhE. To date, vetypes (AE) of FRD/SDH that differ in their subunit compositionand distribution among bacteria, archaea, and eukaryotes areknown (9). Subunit A is too conserved for a classication of thesetypes, but FrdB/SdhB and FrdE/SdhE of Ca. N. deuvii resemblethe respective components of the four-subunit type E enzymes.This type was described in archaea such as Sulfolobus spp. andAcidianus ambivalens, but occurs also in various bacteria (10). Nohomolog of the fourth subunit (named SdhF in type E enzymes)was found in Ca. N. deuvii, indicating that Nitrospira has a non-canonical form of FRD/SDH that is similar to the type E enzymesknown from other organisms. A unique type E-like FRD/SDH

exists also in Leptospirillum, but it seems to lack a homolog ofFrdE/SdhE (11). The type E SDH of A. ambivalens is reversibleand catalyzes both the oxidation of succinate and the reduction offumarate (12). If the type E-like enzyme of Ca. N. deuvii alsooperates in either direction, it could function in the rTCA and theoTCA cycles. As two CDS encode FrdA/SdhA-like subunits, it istempting to speculate that in Ca. N. deuvii the substrate speci-city and catalytic properties of the holoenzyme depend on therespective version of subunit A.Although both Leptospirillum and Ca. N. deuvii belong to the

phylum Nitrospirae and x CO2 via the rTCA cycle, they differ inone critical step of this pathway: the cleavage of citrate to acetyl-CoA and oxaloacetate. Ca. N. deuvii employs ACL and encodesboth subunits of ACL at one aclBA locus in close proximity tothe ve-subunit OGOR gene cluster. In contrast, Leptospirillumlacks ACL but uses two enzymes, citryl-CoA synthetase and citryl-CoA lyase, for cleaving citrate (11).In the oTCA cycle, the 2-oxoglutarate dehydrogenase complex

(ODH) irreversibly catalyzes the oxidative decarboxylation of 2-oxoglutarate to succinyl-CoA and CO2. Ca. N. deuvii possessesthree gene clusters coding for the E1 and E2 components and twogenes encoding the E3 component of 2-oxoacid dehydrogenasecomplexes (Table S2). These CDS most likely represent twocopies of pyruvate dehydrogenase and one 2-oxoisovalerate de-hydrogenase, but probably not ODH.Despite the apparent lack ofODH, the oTCA cycle may operate if ODH is replaced byOGOR, which is present inCa. N. deuvii (see above). In contrastto ODH, OGOR catalyzes a reversible reaction and thus canfunction in both the rTCA and oTCA cycles. For example, inHelicobacter pylori, which also lacks ODH, a four-subunit form ofOGOR functionally replaces ODH in the oTCA cycle (13, 14).Further studies are needed to clarify whether the complete oTCAcycle is functional in Nitrospira and how the reductive and oxi-dative versions of the pathway are regulated in vivo and underdifferent growth conditions.

Use of Organic Substrates. In a previous study, FISH combined withmicroautoradiography showed that Nitrospira from a sewagetreatment plant used pyruvate, but not acetate, as an organiccarbon source (15). Consistent with these results, no canonicalacetate permease was identied in the genome of Ca. N. deuvii.However, the genome encodes a putative member of the GPR1/YaaH protein family (Nide1910). In yeast, one protein from thisfamily has been identied as a candidate acetate transporter (16).If the remote homolog in Ca. N. deuvii indeed facilitates acetateuptake, acetate can be metabolized by activation to acetyl-CoA inthe acetyl-CoA synthetase reaction and subsequent carboxylationto pyruvate, which is catalyzed by POR. Because pyruvate isa precursor for sugar biosynthesis via gluconeogenesis, carbonfrom exogenous acetate or pyruvate can be stored in glycogendeposits within Ca. N. deuvii cells.The genome encodes a soluble formate dehydrogenase for the

oxidation of formate to CO2 with NAD+ as electron acceptor,

suggesting that Ca. N. deuvii can use formate as a substrate. Inaddition, a cluster of six CDS seems to code for a six-subunit,membrane-bound [NiFe]-hydrogenase that might be part ofa formate hydrogenlyase complex (Table S2). However, thefunction of this putative hydrogenase remains unclear. It lacks theamino acid signatures of all known [NiFe]-hydrogenase groups(17) and the cysteine ligands of the [NiFe] center, suggesting thatthis enzyme does not have hydrogenase activity or belongs toa unique class of hydrogenases. Interestingly, a highly similar

Lcker et al. www.pnas.org/cgi/content/short/1003860107 1 of 13

enzyme found in the genome of Ca. K. stuttgartiensis has thecysteine ligands and the signature of group 4 membrane-associ-ated, energy-converting, and H2-evolving hydrogenases (17).

Uptake, Secretion, and Storage. About 56% of the Ca. N. deuviigenome consists of genes involved in transport and secretion,which comprise diverse transporter families (Table S1). Trans-porters for various organic nutrients (Fig. S5) support the notionthat Ca. N. deuvii is not conned to pure autotrophy. Ca. N.deuvii has uptake systems for PO4

3, NH3/NH4+, and NO2

(Table S1). The predicted NO3/NO2

antiporter (Nide1382)could alternatively function as H+/NO2

antiporter and thus beimportant for resistance against elevated cytoplasmic nitriteconcentrations (18). A gene coding for polyphosphate kinase wasidentied, which is consistent with observed polyphosphategranules in Nitrospira moscoviensis cells (19). The genome enc-odes a ferredoxin-nitrite reductase for the reduction of NO2

toNH4

+, indicating that NO2 also serves as a nitrogen source for

biosyntheses. Carbon is stored in glycogen as suggested by genesfor the complete gluconeogenesis pathway, glycogen synthase,glycogen phosphorylase, and two putative glycogen-debranchingenzymes (Table S2). Indeed, glycogen deposits have been ob-served by electron microscopy in Nitrospira cells (20).The genome encodes complete type I and VI protein secretion

systems and the Sec and Tat systems for protein transport to theinnermembrane and periplasmic space. A number of proteins thatare conserved in Gram-negative bacteria and that function in typeII protein secretion or type IV pilus assembly were identied.Ca. N. deuvii has a high demand for iron needed in key

components of the respiratory chain, including NXR, whichcontains several [Fe-S] clusters and is constitutively expressed. Aregion consisting of 38 CDS (77.5 kbp) is dedicated to the syn-thesis of siderophores and the import of iron. It includes genes ofve nonribosomal peptide synthetases, a polyketide synthase,a class III aminotransferase, a type II thioesterase, and a putativecyclic peptide transporter (Table S2). Recently, a gene cluster ofsimilar composition was shown to be involved in siderophoreproduction in the cyanobacterium Anabaena PCC 7120 (21). InCa. N. deuvii, this region also contains genes for the ferriccitrate sensor FecR, the iron uptake regulators FecI and Fur,and TonB-dependent uptake systems for Fe3+, ferric dicitrate,and ferrichrome-type siderophores (Table S2). An additionalputative Fe3+ transporter of the ATP binding cassette (ABC)type I family is encoded at a different locus. At another locationof the Ca. N. deuvii genome, two adjacent genes encode twohighly similar (52.9%) subunits of a bacterioferritin that mostlikely functions in the intracellular storage of iron, but may alsoplay roles in iron and oxygen detoxication (22). A third ferritin-like CDS was identied, which is not in proximity to these twobacterioferritin genes, and a CDS that is remotely similar toa small [2Fe-2S] ferredoxin (BFD) of E. coli (Table S2). BFD isthought to be involved in iron-dependent gene regulation and inthe release of iron from bacterioferritin (23).Consistent with molybdenum being another cofactor of NXR,

an ABC type I transporter for molybdate was also identied.

Stress Response and Defense. Ca. N. deuvii possesses a cyanase forcyanate detoxication and a genomic locus with several genes forarsenic resistance, including an arsenite efux transporter (ArsB),arsenate reductase (ArsC), the arsenic resistance operon regulatorArsR, and a putative arsenite S-adenosylmethyltransferase.Interestingly, this genomic region also encodes both subunits ofarsenite oxidase (AOX), a member of the DMSO reductase familyof molybdenum proteins. AOX could function in arsenite de-toxicationorenableCa.N.deuvii tousearsenite aselectrondonor.The annotation of several -lactamaselike CDS is consistent

with the previously observed resistance of enrichedCa. N. deuviito moderate concentrations of ampicillin (24), whereas two pu-

tative tetracycline efux transporters are contrary to the observedtetracycline sensitivity of Ca. N. deuvii (24).The thioredoxin-dependent peroxiredoxins, which may be in-

volved in H2O2 protection in Ca. N. deuvii, include glutathioneperoxidase, thiol peroxidase, and a putative alkylhydroperoxidase(Table S2). Thioredoxin reductase, which is important for the re-generation of reduced thioredoxin as a prerequisite for H2O2 de-toxication by peroxiredoxins, was also identied in the genome.

Evolutionary History of Nitrospira. Additional support for our hy-pothesis thatNitrospira evolved from anaerobic or microaerophilicancestors stems from estimating genus divergence times within theNitrospirae phylum by using 16S rRNA as a molecular clock. Weare aware of the limitations of this approach (25), but noted aninteresting correlation of the predicted emergence time of thegenus Nitrospira and geochemical data. By analyzing the currentsequence dataset, we found a minimal 16S rRNA similarity of83.4% within the genus Nitrospira, which is considered to containexclusively NO2

-oxidizing bacteria. On the basis of an estimatedrate of 16S rRNA divergence of 1% per 50 million years (Myr)(26), the radiation of Nitrospira took place approximately duringthe past 830 Myr. The 16S rRNA similarity between theNitrospiraand Leptospirillum lineages ranges from 75.8% to 82.8% on thebasis of current datasets. Almost identical values (75.882.6%)were determined for Nitrospira and Thermodesulfovibrio. Usingthese values, we estimate that the three lineages shared a commonancestor about 8701,210 million years ago (Mya). Geochemicaldata indicate that a signicant increase of the atmospheric andoceanic O2 levels began in the late Proterozoic about 850 Mya,whereas Earth was only mildly oxygenated in the preceding 109

years (27). Thus, ancient members of the phylumNitrospirae mostlikely existed under conditions favoring an anaerobic or micro-aerophilic lifestyle. The sharp increase in O2 must have resulted innew ecological niches for those chemolithotrophs that also evolveda sufcient O2 tolerance. We assume that this environmentalchange gave rise to the lineages Nitrospira and Leptospirillum andled to their separation from the still anaerobicThermodesulfovibriolineage. It is interesting to note that the minimal 16S rRNA sim-ilarity among all known anammox lineages (83.6%) indicates thatancestral anammox bacteria and Nitrospiramight have lived in thesame era (about 830 Mya).

SI Materials and MethodsGenome Sequencing and Annotation. The same DNA extractionprotocolwasused for all genomic libraries.Biomass from theCa.N.deuvii enrichment was harvested by centrifugation, and DNAwas extracted from the biomass pellet in agarose plugs as describedin ref. 28. Shotgun randomly sheared DNA libraries were con-structed using a fosmid vector (pCC1FOS; Epicentre Biotechno-logies) and low- or high-copy plasmids [pCNS (3 kb insert) andpCDNA2.1 (6 kb insert), respectively]. Terminal clone end se-quences were determined using BigDye terminator chemistry andcapillary DNA sequencers (model 3730XL; Applied Biosystems)according to standard protocols established at Genoscope. A totalof 99,899 Sanger reads (12,565 fosmid ends, 65,702 pCDNA2.1,and 21,632 pCNS plasmid ends) were assembled using Phrap(version 0.960731; http://www.phrap.org) and produced 39 contigsorganized into one scaffold. Gap closure and manual nishing wascarried out by (i) transposon mutagenesis of two regions and (ii)PCR amplication and sequencing of specic targeted regions.The complete genome sequence of Ca. N. deuvii contains 51,095Sanger reads, achieving an average of 8.2-fold sequence coverageper base.Only 374 additional Sanger reads were needed during thenishing step. Genome assembly robustness was validated by fos-mid coverage coherence (relative orientation and fosmid insertsize of about 3,000 fosmids).The automated analysis pipeline of the MaGe software system

(29) was used for the prediction and annotation of CDS. CDSwere


predicted using the software AMIGene (30) and then submittedto automatic functional annotation (29). Subsequently, the anno-tation of the entire genomewas renedmanually on the basis of thecomprehensive set of data collected automatically for each CDS inthe relational database NitrospiraScope (https://www.genoscope.cns.fr/agc/mage/wwwpkgdb/MageHome/index.php). CDS were as-signed to functional categories according to the MultiFun (31) andTIGRFAM (32) functional role catalogs. Proteins with an aminoacid identity 35% (over at least 80% of the sequence lengths) tocharacterized proteins in the SwissProt or TrEMBLdatabases wereannotated as homologous to proteins with a known function. Es-pecially in ambiguous cases, information on orthologous relation-ships retrieved from the clusters of orthologous groups (COG)database, protein signatures collected from the InterPro database,and enzyme prole data provided by PRIAM and HAMAP wereused for a tentative functional assignment of annotated genes.CDS with an amino acid identity 25% (over at least 80% of thesequence lengths) to characterized proteins or signatures in theaforementioned databases were annotated as putative homologs ofthe respective database entries. The relatively low thresholds of35% and 25% sequence identity were chosen to account for thelarge phylogenetic distance betweenCa. N. deuvii and most othergenome-sequenced microorganisms. CDS with an amino acididentity 25% (over at least 80% of the sequence lengths) to un-characterized proteins were annotated as conserved proteins ofunknown function. In the absence of any signicant database hit,CDS were annotated as proteins of unknown function and, in thecase of uncertain CDS prediction, as doubtful CDS. Finally, CDSwith an amino acid identity 25% to any database entry over lessthan 50% of the length of the longer sequence were annotated asmodular proteins or protein fragments, respectively. The genomiccontext of CDS and the functions of anking genes, as predictedeither in Ca. N. deuvii or in reference genomes from the PkGDBand NCBI RefSeq databases, were considered during CDS anno-tation on the basis of the synteny information and visualization thatis provided by the MaGe software. Metabolic pathways were re-constructed with the help of the KEGG (33) and MetaCyc (34)pathway tools implemented in MaGe. The 63 COGs, which arepresent in all genomes in the current COG database (50 bacterial,13 archaeal, 3 eukaryotic genomes), were identied by using thesoftware EPPS (35) via the online interface (http://web.dmz.uni-wh.de/projects/protein_chemistry/epps/index.php).

Phylogenetic Analyses. Amino acid sequence databases of type IIDMSO reductase-familymolybdopterin cofactor-binding enzymes,of forms IIV RubisCO and RubisCO-like proteins, and of cyt. bdand cyt. bd-like oxidases were established using the software ARB(36). Multiple protein sequence alignments were created auto-matically by ClustalW2 (37) andMUSCLE (38) and weremanuallyrened by using the sequence editor included in the ARB software.Phylogenetic analyses of these proteins were performed by apply-ing distance-matrix, maximum-parsimony, and maximum-likeli-hood methods: neighbor-joining (with 1,000 bootstrap iterationsusing the Dayhoff PAM 001 matrix as amino acid substitutionmodel and the implementation in the ARB software package),protein parsimony (PHYLIP version 3.66 with 100 or 1,000 boot-strap iterations), and protein maximum-likelihood [PHYLIP ver-sion 3.66, PhyML (39), and TREE-PUZZLE (40) with theDayhoffPAM 001, Whelan-Goldman, or the JTT substitution model and1,000 bootstrap iterations using PhyML]. If applicable, N-terminalsignal peptide sequences were excluded from the analyses andmanually created indel lters were used. To determine theminimal16S rRNA sequence similarities within the Nitrospirae phylum andamong the anammox organisms, pairwise similarity matrices weregenerated, by using ARB, from all high-quality Nitrospira (n =206), Leptospirillum (n = 203), Thermodesulfovibrio (n = 28), andanammox planctomycete (n = 140) sequences in the SILVA 10016S rRNA database (released in August 2009) (41). High-quality

sequences were longer than 1,399 nucleotides, had a Pintail (42)score greater than 79, and fell into the monophyletic lineagesformed by each of the aforementioned groups. For sequencesimilarity calculations, the alignment positions 91,507 (E. colinumbering) were considered. The similarity matrices were ex-ported to spreadsheet software (Microsoft Excel) and the minimalvalues were extracted for each phylogenetic lineage.For the calculation of phylogenetic trees for each protein in the

proteome, the fully automated software PhyloGenie (43) wasused. The reference database for PhyloGenie was generatedfrom the National Center for Biotechnology Information non-redundant protein database NCBI nr (44), in which taxon nameswere edited to remove characters that control the structure oftree les in the Newick format. The NCBI taxonomy databasename le was adapted in a similar manner. The PhyloGeniesoftware was executed for each query protein using default pa-rameters with the following modication: -blammerparams=-taxid f. For the BLAST (45) calculations in PhyloGenie, NCBIBLAST (version 2.2.19) was used. Protein phylogenies werecalculated on the basis of full or partial automatic alignmentsproduced by the BLAMMER program included in PhyloGenie.All trees were postprocessed by an in-house script, which sortedall operational taxonomic units according to their distances inthe tree to the query protein.

Incubation of Ca. N. deuvii for Expression Analyses of NxrB. Ca. N.deuvii enrichment biomass was starved for 11 (for mRNAanalysis) or 110 (for protein analysis) d in mineral medium (24)lacking any energy source. After removing a biomass aliquot forlater analysis, 300 M NO2 was added to the medium and theremaining biomass was further incubated for 3 (mRNA analysis)or 8 (protein analysis) d. Biomass from all samples was harvestedby centrifugation and stored at 80 C until further processing.

Quantication of Ca. N. deuvii Cells in the Enrichment. For the im-munological detection of NxrB, total protein had to be extractedfrom similar numbers of starved or NO2

-oxidizingCa. N. deuviicells to ensure that the results were comparable between thesetreatments. For this purpose, the large cell clusters formed byCa. N. deuvii were disintegrated by bead-beating of biomasswith a Fastprep Bead-beater (BIO 101) at level 4 for 5 s. Subse-quently, the biomass was harvested by centrifugation (10,000 g,20 min), and the pellet was resuspended in 1 PBS. As con-rmed by FISH with a Nitrospira-specic probe (15), this treat-ment resulted in a cell suspension containing mainly planktonicNitrospira cells and only a few small cell clusters. An aliquot ofthis suspension was stored at 4 C for protein extraction. Theremaining cell suspension was used for determining the Nitro-spira cell density by quantitative FISH. It was diluted, para-formaldehyde-xed according to ref. 46, and dened volumeswere ltered onto polycarbonate lters (pore size 0.2 m, di-ameter 47 mm, type GTTP; Millipore). The lters were washedtwo times in 1 PBS and double-distilled water, air dried, andstored at 20 C. FISH of the Ca. N. deuvii cells on the lterswas performed according to ref. 47 with the Nitrospira-specic16S rRNA-targeted probes Ntspa1431 (48), Ntspa662, andNtspa712 (15), which were 5-labeled with Cy3 and applied si-multaneously to increase the signal-to-background ratio. Fol-lowing FISH, 28 images of each ltered cell suspension wererecorded using a confocal laser scanning microscope (LSM 510Meta; Zeiss), and the average Nitrospira cell number per imagewas determined by visual counting of the probe-labeled cells ineach image. The Nitrospira cell density in the original (undiluted)cell suspension was then calculated from the average number ofcells per image, the known area of one image in square micro-meters as reported by the Zeiss imaging software, the knownarea of the polycarbonate lter, the volume of ltered cell sus-pension, the dilution factor, and a correction factor. The cor-


rection factor was introduced to account for the possible loss ofcells from the lters during FISH. To determine this factor, lterpieces containing biomass were embedded, before or after FISH,in a mixture of the antifadent Citiour (Citiuor) with a 1:500diluted Sybr Green II solution (Cambrex) for uorescent stain-ing of the total bacterial biomass. Subsequently, 17 images of thestained total biomass were recorded (by confocal microscopy)per lter piece. This was done separately with lter pieces thathad been embedded before or after FISH. The median area (inpixels) of the biomass in each set of images was measured byusing the image analysis software DAIME (49). The ratio of themedian biomass areas before and after FISH informed on theextent of cell loss during FISH and was the aforementionedcorrection factor for the calculation of cell density.

Transcriptional Analysis of nxrB. Total RNA was extracted fromstarved or NO2

-oxidizing Ca. N. deuvii enrichment biomass byusing TRIzol (Invitrogen) according to the protocol recom-mended by the manufacturer and with the modications de-scribed by Hatzenpichler et al. (50). After DNA digestion usingDNase (Fermentas), reverse transcription of 3 g total RNA fromeach treatment was carried out by using the RevertAID rst-strand cDNA synthesis kit (Fermentas) according to the manu-facturers instructions. The primers Ntspa1158R (48), specic forthe 16S rRNA gene of the genusNitrospira, and nxrBR1237 (GTAGAT CGG CTC TTC GAC CTG), targeting both nxrB genes,were used for cDNA synthesis. For cDNA amplication, reactionmixtures with the primer combinations 907F/Ntspa1158R (48, 51)for the 16S rRNA gene and nxrBF916 (GAG CAG GTG GCGCTC CCG C)/nxrBR1237 for the nxrB genes, respectively, wereprepared according the manufacturers recommendations ina total volume of 50 L with 2 mM MgCl2 and 1.25 U of Taqpolymerase (Fermentas). For both primer combinations, thermalcycling comprised initial denaturation at 95 C for 4 min followedby 40 cycles of denaturation at 95 C for 40 s, annealing at 58 C for40 s, and elongation at 72 C for 60 s. Cycling was completed bya nal elongation step at 72 C for 10 min.

Translational Analysis of NxrB.Dened volumes of starved orNO2-

oxidizingCa. N. deuvii cell suspensions were centrifuged (10,000 g for 20 min) to harvest the biomass. On the basis of the resultsof quantitative FISH (see above), these aliquots contained approx-imately the same numbers of Nitrospira cells in all experiments.The cell pellet was resuspended in 5 lysis buffer [7 M urea, 2 Mthiourea, 20 mg/mL amberlite, 4% (wt/vol) 3-[(3-Cholamidopropyl)dimethylammonio]-1-propanesulfonate, 40 mM Tris, 2% (vol/vol)IPGbuffer, 0.2% (wt/vol) bromophenol blue, 1% (wt/vol) dithio-threitol, 10% (wt/vol) glycerol], heated for 3 min at 90 C, and thecrude extract was loaded onto a SDS PAGE gel (12.5% poly-acrylamide) with a molecular weight marker (PageRuler PrestainedProtein Ladder #SM0671; Fermentas). All subsequent steps of theimmunological detection of NxrB by Western blotting were per-formed as described for chlorite dismutase in ref. 52 with the fol-lowing modications. The polyvinylidene uoride membrane wasincubated for 30 min with the NxrB-specic monoclonal antibodyHyb 153.3 (53), which had been diluted 1:1,000 in TBS buffer(20 mM Tris, 150 mM NaCl, pH 7.5) containing 0.1% (vol/vol)Tween 20. The secondary antibody (peroxidase-conjugated goatantimouse IgG; Dianova Germany) was diluted 1:5,000 in TBS-Tween buffer. Crude cell extract from E. coli BL21 (DE3) ex-pressing recombinant NxrB was used as a positive control in theWestern blot experiments.

Cloning and Heterologous Expression of NxrB.The two identical nxrBgenes of Ca. N. deuvii were PCR-amplied by using the HighFidelity PCR enzyme mix (Fermentas) according to the protocolrecommended by the manufacturer. Instead of extracted genomicDNA, 2 L of precooked Ca. N. deuvii enrichment biomass was

added directly to the PCR mix. The applied nxrB-specic primerswere the forward primer NXRV2 (CGA GCG CAT ATG CCAGAA GTC TAT AAC TGG), which contains an NdeI restrictionsite upstream of the NxrB start codon, and the reverse primerNXRR (TTA CGAGAA TTC CCC AGCCAG TTCACG CGCTC), which contains a 5-EcoRI restriction site. Thermal cyclingcomprised an initial denaturation step at 94 C for 5 min followedby 35 cycles of denaturation at 94 C for 30 s, annealing at 56 C for40 s, and elongation at 68 C for 90 s. Cycling was completed bya nal elongation step at 68 C for 10 min. The amplicon wascloned into the vector pCR-XL-TOPO by using the TOPO XLcloning kit (Invitrogen) as recommended by the manufacturer.For heterologous expression, the amplicon was cloned into theexpression vector pET21b(+), which contains a promoter forT7 RNA polymerase and a C-terminal His-tag (Novagen), bydigestion of amplicon and vector with the restriction endonu-cleases NdeI and EcoRI followed by ligation with T4 DNA Ligase(Invitrogen) according to the manufacturers protocol. The ex-pression vector with the nxrB gene was rst transformed by elec-troporation into E. coli XL1 blue cells (Stratagene). Sangersequencing conrmed that the cloned nxrB gene used for heter-ologous expression was identical to the nxrB genes in the Ca. N.deuvii genome. For the expression of NxrB, the vector pET21b(+) containing the nxrB gene was transformed into E. coli BL21(DE3) cells (Stratagene). The recombinant cells were grownat 37 C under agitation (225 rpm) in liquid Luria Bertani me-dium. After growth up to an optical density (600 nm) of 0.8,the expression of NxrB was induced by adding isopropyl--D-thiogalactopyranosidase to a nal concentration of 1 mM. Cellswere harvested after 4 h by centrifugation (5,000 g for 10 min),and the cell pellets were stored at 20 C.

Transcriptional Analysis of the Putative Cytochrome bd-Like Oxidase.Ca. N. deuvii enrichment biomass was kept under oxic con-ditions with NO2

or was starved for 14 d in mineral medium (24)lacking any energy source. NO2

test strips (Merckoquant;Merck) were used to conrm the absence of residual NO2

in thestarved cultures. Biomass was then harvested by centrifugation,the supernatant was discarded, and the biomass was resuspendedin anoxic mineral medium lacking NO2

(starved biomass) orcontaining 3 mM NO2

(nonstarved biomass). The anoxic me-dium had been prepared in accordance with basic principles ofmedium preparation for strict anaerobes as described by Widdeland Bak (54). Following one additional centrifugation andwashing step, the starved and nonstarved biomass was transferredinto two separate 150-mL asks containing 40 mL of anoxicmineral medium and air-free headspace (ushed with N2). Sub-sequently, biomass aliquots were transferred into 300-mL askscontaining 100mL of mineral medium. For anoxic treatments, theasks contained anoxic medium and air-free headspace. For oxictreatments, the headspace of the asks contained air. All askswere plugged with butyl rubber stoppers that were xed with screwcaps. To all treatments, 5 cm3 of an N2:CO2 (80:20) gas mixturewas added to provide CO2 as carbon source, and all working stepswere performed using strictly anoxic techniques. NO2

was addedto nonstarved cultures to a nal concentration of 3 mM, whereasno NO2

was added to the starved cultures. On the basis of theseprocedures, four different incubation conditions were realized:oxic with nitrite, oxic without nitrite, anoxic with nitrite, and an-oxic without nitrite. For each treatment, two replicate askscontaining biomass were incubated for 5 d at 30 C. During in-cubations with NO2

, the consumption of NO2was monitored by

using test strips, and consumed NO2 was replenished. After the

incubations, the biomass from each ask was harvested by cen-trifugation. Total RNAwas extracted according to the protocol ofLueders et al. (55) with the modication that samples were keptfor 2 h on ice after the addition of polyethylene glycol. After DNAdigestion using DNase (Fermentas), PCRwas carried out with the


primers 341F and 518R (51, 56), which target bacterial 16S rRNAgenes. All these test PCR runs were negative, conrming theabsence of residual DNA in the RNA extracts. Subsequently,reverse transcription of 250 ng total RNA from each treatmentwas carried out by using the RevertAID rst-strand cDNAsynthesis kit (Fermentas) according to the manufacturers in-structions. The following primers were used for multiplex cDNAsynthesis: Nide0901R1 (CTCGGAAGCATCGGCCTCAGG),specic for the putative cyt. bd-like terminal oxidase of Ca.N. deuvii (gene nide0901), and 1431R, specic for the 16S rRNAof sublineage I Nitrospira (48). For cDNA amplication, reactionmixtures with the primer combinations 1158Fa (modied fromref. 48; ACT GCC CAG GAT AAC GGG)/1431R for the 16SrRNA gene and Nide0901F (GGT GTC TGG GGT TAC TTCGTT)/Nide0901R2 (ACCGTAGATGTG CCAGTGAAC) forgene nide0901, respectively, were prepared according the manu-facturers recommendations in a total volume of 50 L with 2 mMMgCl2 and 1.25 U of Taq polymerase (Fermentas). For bothprimer combinations, thermal cycling comprised initial de-naturation at 95 C for 5 min followed by 10 cycles of denaturationat 95 C for 30 s, annealing at 7065 C for 30 s (0.5 C in eachcycle), and elongation at 72 C for 40 s. This was followed by 25cycles of denaturation at 95 C for 30 s, annealing at 65 C for 30 s,and elongation at 72 C for 40 s. Cycling was completed by a nalelongation step at 72 C for 10 min. The specic reverse tran-scription of the target RNAs was conrmed by Sanger sequencingof the obtained amplicons.

Stable Carbon Isotopic Fractionation.Highly enriched cultures ofCa.N. deuvii were grown in batchmode in mineral medium (24) in 5-L bottles at 28 C in the dark. Cell suspensions were moderatelystirred, and NO2

was replenished from a 2.5 M stock solution.Incubation was performed for about 4 wk until the suspension was

turbid. In the second batch culture, 13CDIC was monitored overtime. Duplicate samples of 25 mL were taken at different stages ofgrowth and transferred to gas-tight tubes, xed with one drop of35% (vol/vol) formaldehyde, and stored at 4 C. Biomass washarvested by centrifugation at 10,000 rpm (15,600 g), washed,suspended in 0.9% NaCl, and frozen at 20 C.Analysis of the 13C of total dissolved inorganic carbon (DIC) in

the medium was performed by headspace analysis of 0.51 mL ofwater that had reacted with H3PO4 for at least 1 h at room tem-perature. The headspace was subsequently analyzed 10 times byusing a Thermonnigan Gas Bench II coupled to a DeltaPLUS

irmMS system with typical SDs of 0.3. Stable carbon isotoperatios were determined relative to laboratory standards calibratedon NBS-18 carbonate [International Atomic Energy Agency(IAEA)]. Differences in 13CDIC of the duplicate samples werealways

30. Bocs S, Cruveiller S, Vallenet D, Nuel G, Mdigue C (2003) AMIGene: Annotation ofMIcrobial Genes. Nucleic Acids Res 31:37233726.

31. Serres MH, Riley M (2000) MultiFun, a multifunctional classication scheme forEscherichia coli K-12 gene products. Microb Comp Genomics 5:205222.

32. Haft DH, et al. (2001) TIGRFAMs: A protein family resource for the functionalidentication of proteins. Nucleic Acids Res 29:4143.

33. Kanehisa M, Goto S (2000) KEGG: Kyoto encyclopedia of genes and genomes. NucleicAcids Res 28:2730.

34. Caspi R, et al. (2006) MetaCyc: A multiorganism database of metabolic pathways andenzymes. Nucleic Acids Res 34 (Database issue):D511D516.

35. Reichard K, Kaufmann M (2003) EPPS: Mining the COG database by an extendedphylogenetic patterns search. Bioinformatics 19:784785.

36. Ludwig W, et al. (2004) ARB: A software environment for sequence data. NucleicAcids Res 32:13631371.

37. Larkin MA, et al. (2007) Clustal W and Clustal X version 2.0. Bioinformatics 23:29472948.

38. Edgar RC (2004) MUSCLE: A multiple sequence alignment method with reduced timeand space complexity. BMC Bioinformatics 5:113.

39. Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm to estimate largephylogenies by maximum likelihood. Syst Biol 52:696704.

40. Schmidt HA, Strimmer K, Vingron M, von Haeseler A (2002) TREE-PUZZLE: Maximumlikelihood phylogenetic analysis using quartets and parallel computing. Bioinfor-matics 18:502504.

41. Pruesse E, et al. (2007) SILVA: A comprehensive online resource for quality checkedand aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Res 35:71887196.

42. Ashelford KE, Chuzhanova NA, Fry JC, Jones AJ, Weightman AJ (2005) At least 1 in 2016S rRNA sequence records currently held in public repositories is estimated tocontain substantial anomalies. Appl Environ Microbiol 71:77247736.

43. Frickey T, Lupas AN (2004) PhyloGenie: Automated phylome generation and analysis.Nucleic Acids Res 32:52315238.

44. Sayers EW, et al. (2010) Database resources of the National Center for BiotechnologyInformation. Nucleic Acids Res 38 (Database issue):D5D16.

45. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local alignmentsearch tool. J Mol Biol 215:403410.

46. Daims H, Stoecker K, Wagner M (2005) Fluorescence in situ hybridisation for the

detection of prokaryotes. Molecular Microbial Ecology, eds Osborn AM, Smith CJ

(Bios-Garland, Abingdon, United Kingdom), pp 213239.47. Glckner FO, et al. (1996) An in situ hybridization protocol for detection and

identication of planctonic bacteria. Syst Appl Microbiol 19:403406.48. Maixner F, et al. (2006) Nitrite concentration inuences the population structure of

Nitrospira-like bacteria. Environ Microbiol 8:14871495.49. Daims H, Lcker S, Wagner M (2006) daime, a novel image analysis program for

microbial ecology and biolm research. Environ Microbiol 8:200213.50. Hatzenpichler R, et al. (2008) A moderately thermophilic ammonia-oxidizing

crenarchaeote from a hot spring. Proc Natl Acad Sci USA 105:21342139.51. Lane DJ, et al. (1985) Rapid determination of 16S ribosomal RNA sequences for

phylogenetic analyses. Proc Natl Acad Sci USA 82:69556959.52. Maixner F, et al. (2008) Environmental genomics reveals a functional chlorite

dismutase in the nitrite-oxidizing bacterium Candidatus Nitrospira deuvii. Environ

Microbiol 10:30433056.53. Aamand J, Ahl T, Spieck E (1996) Monoclonal antibodies recognizing nitrite

oxidoreductase of Nitrobacter hamburgensis, N. winogradskyi, and N. vulgaris. Appl

Environ Microbiol 62:23522355.54. Widdel F, Bak F (1992) Gram-negative mesophilic sulfate-reducing bacteria. The

Prokaryotes, eds Balows A, Trper HG, Dworking M, Harder W, Schleifer K-H

(Springer, New York), 2nd Ed, pp 33523378.55. Lueders T, Maneeld M, Friedrich MW (2004) Enhanced sensitivity of DNA- and rRNA-

based stable isotope probing by fractionation and quantitative analysis of isopycnic

centrifugation gradients. Environ Microbiol 6:7378.56. Edwards U, Rogall T, Blcker H, Emde M, Bttger EC (1989) Isolation and direct

complete nucleotide determination of entire genes. Characterization of a gene

coding for 16S ribosomal RNA. Nucleic Acids Res 17:78437853.57. Schouten S, et al. (1998) Biosynthetic effects on the stable carbon isotopic

compositions of algal lipids: Implications for deciphering the carbon isotopic

biomarker record. Geochim Cosmochim Acta 62:13971406.58. Rohmer M, Bouvier-Nav P, Ourisson G (1984) Distribution of hopanoid triterpenes in

prokaryotes. J Gen Microbiol 130:11371150.


Protein function:

Metabolism Energy

Cell cycle /DNA processing

Transcription

Protein synthesisProtein fate

Cellular transport Regulatoryfunctions

Signal transductionCell envelope

Cellular processesMobile/extrachrom.element functions

Unknown function

RNA genes:

rRNA tRNA

Universal gene set:

COGs present in all genomes in the COG database

GC bias (grid spacing = 2%)

GC skew (grid spacing = 0.05)

Protein phylogeny (closest relative):

AlphaProteobacteria:

Beta Gamma Delta

Firmicutes

Leptospirillum

Nitrospirae:Thermodesulfovibrio

Cyanobacteria

Kuenenia

Planctomycetes:other Planctomycetes

Chloroflexi Verrucomicrobia

Acidobacteria Actinobacteria

Euryarchaeota Bacteroidetes

Aquificae Deinococcus-Thermus

Chlorobi others

others

Mobile elements:

Phage-related Insertion sequence elements

Plasmid functions Putative transposases

Ca. Nitrospiradefluvii

4,317,083 bp

Fig. S1. Circular representation of the Ca. N. deuvii chromosome. Annotated coding sequences (rings 1 + 2), genome-wide protein phylogeny (ring 3), RNAgenes (ring 4), universal clusters of orthologous groups (ring 5), mobile genetic elements (rings 6 + 7), and local nucleotide composition measures (rings 8 + 9)are shown. Very short features were enlarged to enhance visibility. Clustered genes, such as several tRNA genes, may appear as one line due to space limi-tations. The most closely related homolog in other sequenced genomes was determined for each protein of Ca. N. deuvii (ring 3) by using a genome-widephylogenetic approach (SI Materials and Methods). Three larger clusters of phage-related genes (in rings 6 + 7) probably represent prophages or remnantsthereof. The image was created by using the software Circos (1).

1. Krzywinski M, et al. (2009) Circos: An information aesthetic for comparative genomics. Genome Res 19:16391645.


55

43

kDa + No nitrite (110 days) Nitrite added to medium

III

250 bp

500 bp

I

250 bp

500 bp

II

M MM+- - + 1 2 1 1 2 2 + RT + RT - RT - RT

1 2

AICa. Nitrospira defluvii NxrA1 (NIDE3237) 60 RYDSSFT-----WCCSPNDTHG--CRVRAFVRNGVVMRVEQNYDHQTYEDLYGNRGTFAHNPRMCLKGFTFH 124 Ca. Nitrospira defluvii NxrA2 (NIDE3255) 59 RYDSSFT-----WVCSPNDTHA--CRIRAFVRNGVVMRVEQNYDHQTYEDLYGNRGTFAHNPRMCLKGFTFH 123 Ca. K. stuttgartiensis NxrA (CAJ72445) 58 QYDRTFT-----YCCSPNDTHA--CRIRAFVRNNVMMRVEQNYDHQNYSDLYGNKATRNWNPRMCLKGYTFH 122 Hydrogenobaculum sp. (YP_002121006) 55 KPDETTV-----ILCNPNDTHG--CYLNAYVKNGIITRLEPTYKYGDATDIYGIKASHRWEPRCCNKGLALV 119 Beggiatoa sp. (ZP_02000390) 48 HYDSSFT-----FLCAPNDTHN--CLLRGYVKNGVVTRIAPTYGYQKAKDLDGNQSTQRWDPRCCQKGLALV 112 IINitrobacter hamburgensis NxrA (YP_578638) 24 SHDNVFR-----STHGVNCTGG--CSWAIYVKDGIITWEMQQTDYPLLER-----SLPPYEPRGCQRGISAS 83 Nitrococcus mobilis NxrA (ZP_01125872) 24 QHDNIFR-----STHGVNCTGG--CSWAIYVKDGIITWEMQQTDYPLLGRGEGGRGIPPYEPRGCQRGISAS 88 IIIEscherichia coli NarG (1Q16_A) 41 QHDKIVR-----STHGVNCTGS--CSWKIYVKNGLVTWETQQTDYPRTRP-----DLPNHEPRGCPRGASYS 100 Haloarcula marismortui NarG (YP_135852) 70 DWDSVAR-----STHSVNCTGS--CSWNVYVKDGQVWREEQAGDYPTFDES-----LPDPNPRGCQKGACYT 129 IVAzoarcus sp. EB1 EbdA (AAK76387) 73 KWDKVN---WGS--HLNICWPQGSCKFYVYVRNGIVWREEQAAQTPACNVD-----YVDYNPLGCQKGSAFN 134 Rhodovulum sulfidophilum DdhA (Q8GPG4) 57 TWDY-----VGKAAHCINCLGN--CAFDIYVKDGIVIREEQLAKYPQISPD-----IPDANPRGCQKGAIHS 116 Ideonella dechloratans ClrA (P60068) 60 TWDS-----VGVMTHSNGCVAG--CAWNVFVKNGIPMREEQISKYPQL-PG-----IPDMNPRGCQKGAVYC 118 Dechloromonas agitata PcrA (AAO49008) 51 SWDKKTR-----GAHLINCTGA--CPHFVYTKDGVVIREEQSKDIPPMPN------IPELNPRGCNKGECAH 109 Thauera selenatis SerA (Q9S1H0) 64 TWDS-----TGFITHSNGCVAG--CAWRVFVKNGVPMREEQVSEYPQL-PG-----VPDMNPRGCQKGAVYC 122 Archaeoglobus fulgidus (NP_069015) 37 GEIKTVR-----SVCSPNCTGA--CGFDALVYNGRIETLTQAADYPEPEYN----------PRGCLRGQSMM 91 Moorella thermoacetica (YP_430751) 68 KEVKYVR-----TTCSPNCTLA--CGIRAMVVDGQIKALLPSNDYPEPEYG----------PRGCLRGLSFI 122

B aa.L Id.%ICa. Nitrospira defluvii NxrA1 (NIDE3237) 236 VGKMMNTRFNGGCLPLLDSWIRKVDAEKAQGGKYYSNYTWHGDQDPSHPFWNGTQNCDVD 295 1146 100 Ca. Nitrospira defluvii NxrA2 (NIDE3255) 235 MGKHANTRFNNCVLPLLDSWIRKVNPDQAQGGRYWNNYTWHGDQDPSQPWWNGTQNCDVD 294 1147 87 Ca. K. stuttgartiensis NxrA (CAJ72445) 234 IGKYGMYRFNN-CLAIVDAHNRGVGPDQALGGRNWSNYTWHGDQAPGHPFSHGLQTSDVD 292 1148 58 Hydrogenobaculum sp. (YP_002121006) 230 LRYVGQYRMSN-MMALLDSNIRKVEPDKALGGVGWDNYSEHTDLPPAHTLVTGQQTVDFD 289 1154 37 Beggiatoa sp. (ZP_02000390) 223 TRVFAQYREAN-AMALLDDKIRGKGTD-SLGGRGFDNYSWHTDLPPGHPMVTGQQTVDFD 281 1138 39 IINitrobacter hamburgensis NxrA (YP_578638) 181 LSFAAGTRFLS-----------------LMGGSLMSFYDWYADLPTSFPEIWGDQTDVCE 223 1214 16 Nitrococcus mobilis NxrA (ZP_01125872) 186 LSYAAGSRFLQ-----------------LFGGVNMSFYDWYADLPNSFPEIWGDQTDVCE 228 1218 17 IIIEscherichia coli NarG (1Q16_A) 198 VSYASGARYLS-----------------LIGGTCLSFYDWYCDLPPASPQTWGEQTDVPE 240 1247 16 Haloarcula marismortui NarG (YP_135852) 193 VSFASGSRLVN-----------------LLGGVSHSFYDWYSDLPPGQPITWGTQTDNAE 235 952 21 IVAzoarcus sp. EB1 EbdA (AAK76387) 198 IAWGAGFRMTY-----------------LMDGVSPDINVDIGDTYMGAFHTFGKMHMGYS 240 976 19 Rhodovulum sulfidophilum DdhA (Q8GPG4) 180 MRMAAPYRFAS-----------------LVGGVQLDIFTDVGDLNTGAHLAYGNALESFT 222 910 17 Ideonella dechloratans ClrA (P60068) 180 ITNTAYTRMTK-----------------LLGAISPDATSMTGDLYTGIQTVRVPASTVST 222 914 19 Dechloromonas agitata PcrA (AAO49008) 173 VSFSAGHRFAH-----------------YIGAHTHTFFDW

Documents

Lucker 2010 Nitrospira.pdf