Nitrosopumilus maritimus genome reveals uniquemechanisms for nitrification and autotrophy inglobally distributed marine crenarchaeaC. B. Walkera,b, J. R. de la Torrea, M. G. Klotzc, H. Urakawaa, N. Pinela, D. J. Arpd, C. Brochier-Armanete, P. S. G. Chainf,g,h,P. P. Chani, A. Gollabgirj, J. Hempk, M. Hüglerl,m, E. A. Karrn, M. Könnekeo, M. Shinf,g, T. J. Lawtonp, T. Lowei, W. Martens-Habbenaa, L. A. Sayavedra-Sotod, D. Langf,g, S. M. Sievertq, A. C. Rosenzweigp, G. Manningj, and D. A. Stahla,1

aDepartment of Civil and Environmental Engineering, University of Washington, Seattle, WA 98195; bGeosyntec Consultants, Seattle, WA 98101;cDepartment of Biology, University of Louisville, Louisville, KY 40292; dDepartment of Botany and Plant Pathology, Oregon State University, Corvalis, OR97331; eUniversité de Provence Aix-Marseille I, Laboratoire de Chimie Bactérienne, Centre National de la Recherche Scientifique Unité Propre de Recherche,Marseille, 13402 France; fBiosciences Division, Lawrence Livermore National Laboratory, Livermore, CA 94550; gMicrobial Program, Joint Genome Institute,Walnut Creek, CA 94598; hCenter for Microbial Ecology, Michigan State University, East Lansing, MI 48824; iDepartment of Biomolecular Engineering,University of California, Santa Cruz, CA 95064; jRazavi Newman Center for Bioinformatics, Salk Institute for Biological Studies, La Jolla, CA 92037; kSchool ofChemical Sciences, University of Illinois, Urbana, IL 61801; lLeibniz-Institut für Meereswissenschaften, Kiel, 24105 Germany; mWater Technology Center,Karlsruhe, 76139 Germany; nDepartment of Botany and Microbiology, University of Oklahoma, Norman, OK 73019; oInstitut für Chemie und Biologie desMeeres, Universität Oldenburg, Oldenburg, 26129 Germany; pDepartments of Biochemistry, Molecular Biology and Cell Biology, and Chemistry, NorthwesternUniversity, Evanston, IL 60208; and qBiology Department, Woods Hole Oceanographic Institution, Woods Hole, MA 02543

Edited by David Karl, University of Hawaii, Honolulu, HI, and approved April 2, 2010 (received for review December 6, 2009)

Ammonia-oxidizing archaea are ubiquitous in marine and terres-trial environments and now thought to be significant contributorsto carbon and nitrogen cycling. The isolation of Candidatus “Nitro-sopumilus maritimus” strain SCM1 provided the opportunity forlinking its chemolithotrophic physiology with a genomic inventoryof the globally distributed archaea. Here we report the 1,645,259-bp closed genome of strain SCM1, revealing highly copper-depen-dent systems for ammonia oxidation and electron transport thatare distinctly different from known ammonia-oxidizing bacteria.Consistent with in situ isotopic studies of marine archaea, thegenome sequence indicates N. maritimus grows autotrophicallyusing a variant of the 3-hydroxypropionate/4-hydroxybutryratepathway for carbon assimilation, while maintaining limited capac-ity for assimilation of organic carbon. This unique instance of ar-chaeal biosynthesis of the osmoprotectant ectoine and anunprecedented enrichment of multicopper oxidases, thioredoxin-like proteins, and transcriptional regulators points to an organismresponsive to environmental cues and adapted to handling reac-tive copper and nitrogen species that likely derive from its distinc-tive biochemistry. The conservation of N. maritimus gene contentand organization within marine metagenomes indicates that theunique physiology of these specialized oligophiles may play a sig-nificant role in the biogeochemical cycles of carbon and nitrogen.

ammonia oxidation | marine microbiology | archaea | nitroxyl

Marine Group I archaea are among the most abundantmicroorganisms in the global oceans (1–3). Originally dis-covered through ribosomal RNA gene sequencing (3, 4), recentmetagenomic, biogeochemical, and microbiological studiesestablished the capacity of these organisms to oxidize ammonia,thus linking this abundantmicrobial clade to oneof the key steps ofthe global nitrogen cycle (5–9). For a century following the dis-covery of autotrophic ammonia oxidizers, only Bacteria werethought to catalyze this generally rate-limiting transformation inthe two-step process of nitrification (10). Despite recent enrich-ment of mesophilic as well as thermophilic ammonia-oxidizingarchaea (AOA) (6, 11, 12), only a single Group I-related strain,isolated from a gravel inoculum from a tropical marine aquarium,has thus far been successfully obtained in pure culture (7).The isolation of Nitrosopumilus maritimus strain SCM1 ulti-

mately confirmed an archaeal capacity for chemoautotrophicgrowth on ammonia. More detailed characterization of this strainrevealed cytological and physiological adaptations critical for lifein an oligotrophic open ocean environment, most notably one of

the highest substrate affinities yet observed (13). Among charac-terized ammonia oxidizers, onlyN.maritimus is capable growing atthe extremely low concentrations of ammonia generally found inthe open ocean (7, 13). This strain therefore provided an excellentopportunity to investigate the core genetic inventory for ammonia-based chemoautotrophy by Group I crenarchaea.The gene content and gene order of N. maritimus is highly

similar to environmental populations represented in marine bac-terioplankton metagenomes, confirming on a genomic level itsclose relationship to many oceanic crenarchaea. Thus, an evalu-ation of the genomic inventory of N. maritimus should offera framework to identify features shared among ammonia-oxidizingGroup I crenarchaea, resolve physiological diversity amongAOA,and refine understanding of their ecology in relationship to thelarger assemblage of marine archaea—not all of which are am-monia oxidizers. In support of this expectation, the physiologicaland genomic profiles together show that many of the “non-extreme” archaea identified inmetagenomic studies, and currentlyassigned to the Crenarchaeota kingdom, are AOA that contributeto global carbon and nitrogen cycling, possibly determining ratesof nitrification in a variety of environments (6, 8, 9, 13).

Results and DiscussionPrimary Sequence Characteristics. N. maritimus strain SCM1 con-tains a single chromosome of 1,645,259 bp encoding 1,997 pre-dicted genes and no extrachromosomal elements or completeprophage sequences (Table 1). No unambiguous origin of rep-lication could be determined on the basis of local gene contentor GC skew, as commonly observed for other archaeal genomes(14). Approximately 61% of the N. maritimus open-reading

Author contributions: C.B.W., J.R.d.l.T., P.S.G.C., and D.A.S. designed research; C.B.W.,J.R.d.l.T., M.G.K., H.U., N.P., C.B-A., P.P.C., A.G., M.H., E.A.K., M.K., M.S., T.L., W.M-H.,M.S., D.L., S.M.S., A.C.R., G.M., and D.A.S. performed research; C.B.W. and J.R.d.l.T. con-tributed new reagents/analytic tools; C.B.W., J.R.d.l.T., M.G.K., H.U., N.P., D.J.A., C.B.-A.,P.P.C., A.G., J.H., M.H., E.A.K., M.K., T.J.L., T.L., W.M.-H., L.A.S.-S., S.M.S., A.C.R., G.M., andD.A.S. analyzed data; and C.B.W., J.R.d.l.T., M.G.K., H.U., N.P., C.B.-A., J.H., M.H., E.A.K.,M.K., T.L., S.M.S., A.C.R., G.M., and D.A.S. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Data deposition: The sequence reported in this paper has been deposited in the NCBIdatabase (accession no. NC_010085).1To whom correspondence should be addressed. E-mail: dastahl@u.washington.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental.

8818–8823 | PNAS | May 11, 2010 | vol. 107 | no. 19 www.pnas.org/cgi/doi/10.1073/pnas.0913533107

mailto:dastahl@u.washington.eduhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplementalhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplementalwww.pnas.org/cgi/doi/10.1073/pnas.0913533107

frames (ORFs) could be assigned to clusters of orthologousgroups of proteins (COGs), a lower percentage than for genomesof ammonia-oxidizing bacteria (AOB) (Table S1) but similar toCenarchaeum symbiosum (15). The genome possesses a relativelyhigh coding density (91.9%), with a larger fraction dedicated toenergy production/conservation, coenzyme transport/metabolism,and translation genes than other characterized Crenarchaeota, butsimilar to two common species of photoautotrophic marineBacteria, Prochlorococcus, and Synechococcus.

Energy Metabolism. The stoichiometry of ammonia oxidation tonitrite is similar to that of characterized aerobic, obligate che-molithoautotrophicAOB(13), yet the contributing biochemistry isdistinctly unique. All AOB share a common pathway where hy-droxylamine, producedby an ammoniamonooxygenase (AMO), isoxidized to nitrite by a heme-rich hydroxylamine oxidoreductase(HAO) complex; the oxidation of hydroxylamine supplies elec-trons to both the AMO and a typical electron transport chaincomposed of cytochrome c proteins. N. maritimus lacks genesencoding a recognizable AOB-like HAO complex and pertinentcytochrome c proteins, indicating an alternative archaeal pathway.The numerous copper-containing proteins, including multicopperoxidases and small blue copper-containing proteins (similar toplasto-, halo-, and sulfocyanins), suggest an alternative electrontransfer mechanism (Table S2). These predicted periplasmicproteins likely serve functionally similar roles to soluble cyto-chrome c proteins in other Archaea, including Natronomonaspharaonis and thermoacidophiles such as Sulfolobus (16). Thisapparent reliance on copper for redox reactions is a major di-vergence from the iron-based electron transfer system of AOB.The N. maritimus genome contains genes coding for six soluble

periplasmic multicopper oxidase (MCO) proteins: two nearlyidentical NO-forming nitrite reductase proteins (NirK;Nmar_1259 and -1667), each with three cupredoxin domains;two NcgA-like (nirK cluster gene A) MCOs (Nmar_1131 and-1663) with two cupredoxin domains; one MCO (Nmar_1136)with three cupredoxin domains; and one MCO (Nmar_1354)with two domains fused to a blue copper-containing protein(Table S2). Two (Nmar_1131 and Nmar_1663) of the threegenes that are classified as belonging to the emerging class oftwo-domain MCOs (2dMCOs) resemble the general architectureof the 2dMCO NcgA present in Nitrosomonas europaea. Al-though the overall sequence identity is low, clustering of NcgAwith a nitrite reductase suggests it may play a supporting role innitrite reduction (17). Genes Nmar_1131 andNmar_1663 are alsocolocatedwith amember of theDtxR family ofmetal regulators anda member of the ZIP metal transport family, suggesting a role inmetal homeostasis (see below). The third 2dMCO (Nmar_1354),possessing a fused blue copper-containing domain, has not beenfound in AOB and appears to be unique to AOA. Redox inter-

actions with the MCOs (and other predicted redox proteins) arelikely mediated by eight soluble and nine membrane-anchoredcopper-binding proteins containing plastocyanin-like domains(Table S2). The corresponding genes appear to be the result ofa series of duplications within the N. maritimus lineage (Fig. S1).A second family of predicted redox active periplasmic pro-

teins, composed of 11 thiol-disulfide oxidoreductases from thethioredoxin family (Nmar_0639, _0655, _0829, _0881, _1140,_1143, _1148, _1150, _1181, _1658, and Nmar_1670), show low(but recognizable) identity with the better characterized disulfidebond oxidases/isomerases found in Bacillus subtilis (BdbD) andEscherichia coli (DsbA, DsbC, and DsbG). The mean percentageof sequence identity between the N. maritimus proteins andBdbD is 21 ± 3%. The significantly lower mean percentage ofsequence identity to DsbA, DsbC, and DsbG (9.2, 10.7, and10.4%, respectively) is comparable to that shared between theE. coli proteins (10–11%). Although functional equivalencycannot be established, all but Nmar_0881 preserve the conservedthioredoxin-like active site FX4CXXC sequence (18–20). InE. coli, both DsbA and DsbC rectify nonspecific disulfide bondscatalyzed by copper (21), whereas up-regulation of dsbA by theCpx regulon occurs during copper stress (22, 23). Eukaryoticprotein disulfide isomerase (PDI) homologs sequester and/orreduce oxidized Cu(II), perhaps serving as copper acceptors/donors for copper-containing proteins (24). Another describedfunction of PDIs is the capture and transport of nitric oxide (25,26), a possible intermediate or by-product of ammonia oxidation.The related protein family in N. maritimus may function in partto alleviate copper and nitric oxide toxicity.

Pathways for Ammonia Oxidation and Electron Transfer. The threegenes (Nmar_1500, _1503, and _1502) annotated as amoA,amoB, and amoC and coding for a putative ammonia mono-oxygenase complex are the only recognizable genetic hallmarksof ammonia oxidation in the genome sequence. However, theN. maritimus sequences are no more similar (in either content ororganizational structure) to bacterial amo genes than they are tothe genes encoding bacterial particulate methane mono-oxygenases (pMMO), suggestive of functional differences be-tween the archaeal and bacterial versions of AMO (7, 27).Notably, mapping the sequence encoded by amoB onto thepmoB crystal structure of Methylococcus capsulatus (Bath) (28)reveals the conservation of the ligands to the pMMO metalcenters and the complete absence of both a transmembrane helixand a C-terminal cupredoxin domain predicted to be present inbacterial AMO (Fig. S2).The structural differences in the archaeal AMO, the lack of

genes encoding the hydroxylamine–ubiquinone redox module(29), and a periplasm enriched in redox active proteins togethersuggest significant divergence from the bacterial pathway of

Table 1. Genome features of N. maritimus SCM1, C. symbiosum, sequenced AOB, and crenarchaeal genome fragments

Nitrosopumilusmaritimus SCM1

Cenarchaeumsymbiosum

Nitrosococcusoceani

ATCC 19707

Nitrosomonaseuropaea

ATCC 19718Nitrosomonaseutropha C91

NitrosospiramultiformisATCC 25196

Fosmid4B7

CosmidDeepAnt-EC39

Fosmid74A4

Size (bp) 1,645,259 2,045,086 3,481,691 2,812,094 2,661,057 3,184,243 39,297 33,347 43,902Percent coding 91.90% 91.20% 86.80% 88.40% 85.60% 85.60% 89.10% 86.10% 84.00%GC content 34.20% 57.70% 50.30% 50.70% 48.50% 53.90% 34.40% 34.10% 32.60%ORFs 1,997 2,066 3,186 2,628 2,578 2,827 41 41 51ORF density (ORF/kb) 1.19 0.986 0.889 0.876 0.952 0.86 0.992 1.17 1.12Avg. ORF length (bp) 757 924 964 1009 890 980 898 737 753Standard tRNAs 44 45 45 41 41 43 0 0 0rRNAs 1 1 2 1 1 1 1 1 1Plasmids 0 0 1 0 2 3 NA NA NA

NA, not analyzed.

Walker et al. PNAS | May 11, 2010 | vol. 107 | no. 19 | 8819

MICRO

BIOLO

GY

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st01.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st02.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st02.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st02.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/pnas.200913533SI.pdf?targetid=nameddest=STXThttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/pnas.200913533SI.pdf?targetid=nameddest=STXT

ammonia oxidation. There are two hypothetical mechanisticalternatives (Fig. 1, Table S2): either a unique biochemistryexists for the oxidation of hydroxylamine or the divergent AMOdoes not actually produce hydroxylamine. If the former is true,hydroxylamine oxidation may occur via one of the periplasmicMCOs (CuHAO). Given the lack of cytochrome c proteins, thefour electrons would then be transferred to a quinone reductase(QRED) via small blue copper-containing plastocyanin-likeelectron carriers. The protein encoded by Nmar_1226, whichcontains four transmembrane-spanning regions and two plasto-cyanin-like domains, may serve as an analog of the membrane-bound cytochrome cM552 quinone reductase present in AOB(29) and is a good candidate for the QRED.In an alternative scenario, the archaeal AMO produces not

hydroxylamine, but the reactive intermediate nitroxyl (nitroxylhydride, HNO). Nitroxyl is a highly toxic and reactive compoundrecently recognized as having biological significance in a numberof systems (30, 31). During archaeal ammonia oxidation, nitroxylmight be formed by a unique monooxygenase function of ar-chaeal AMO. Alternatively, the archaeal AMO may act asa dioxygenase and insert two oxygen atoms into ammonia,producing nitroxyl from the spontaneous decay of HNOHOH.Both reaction sequences eliminate the requirement for reductantrecycling during the initial oxygenase reaction, a simplificationoffering significant ecological advantage (when compared withAOB) in nutrient poor environments. In this pathway, one of theMCO-like proteins may act as a nitroxyl oxidoreductase (NXOR)and facilitate the oxidation of nitroxyl to nitrite with the extrac-tion of two protons and two electrons in the presence of water.The proposed NXOR would relay the two extracted electronsinto the quinone pool via the QRED pathway described above.In this proposed model, the electrons extracted by either

a CuHAO or a NXOR (and transferred into the quinone pool)would generate a proton motive force (PMF) through complexesIII (plastocyanin-like subunit, Nmar_1542; Rieske-type subunit,Nmar_1544; transmembrane subunit, Nmar_1543) and IV(Nmar_0182-5), driving the generation of ATP by an F0F1-typeATP synthase (Nmar_1688–1693). The production of reductant(i.e., NADH) would require the reverse operation of complex I(NuoABCDHIJKMLN, Nmar_0276–286) as a quinol oxidasedriven by a PMF. The proposed biochemistry involving nitroxylproduces the same net gain as bacterial ammonia oxidation,providing two electrons for reduction of the quinone pool andsubsequent linear electron flow and the generation of a PMF.The presence of a copper-containing (versus heme) complex IIIand the unique evolutionary placement of terminal oxidase(complex IV) between two of the heme–copper oxygen reductasefamilies further distinguish this proposed ammonia oxidationpathway from that in AOB.

Carbon Fixation and Mixotrophy. N. maritimus, like all knownAOB, grows chemolithoautotrophically by using inorganic carbonas the sole carbon source (7, 32). However, whereas AOB use theCalvin–Bassham–Benson cycle with the CO2-fixing enzyme ribulose

bisphosphate carboxylase/oxygenase (RubisCO) as the key enzyme,the absence of genes inN. maritimus coding for RubisCO and otherenzymes of this cycle points to an alternative pathway for carbonfixation. The most likely mechanism supported by the genome se-quence is the 3-hydroxypropionate/4-hydroxybutyrate pathway elu-cidated in the thermophilic crenarchaote Metallosphaera sedula andsuggested as a potential pathway of carbon fixation in C. symbiosum(33). The pathway has two parts: a sequence including two carbox-ylation reactions transforming acetyl-CoA to succinyl-CoA anda multistep sequence converting succinyl-CoA into twomolecules ofacetyl-CoA. Genes identified in theN. maritimus genome coding forkey enzymes of the pathway (Fig. S3) include a biotin-dependentacetyl-CoA/propionyl-CoA carboxylase (Nmar_0272–0274), meth-ylmalonyl-CoA epimerase and mutase (Nmar_0953, _0954, and_0958), and 4-hydroxybutyrate dehydratase (Nmar_0207). With theexceptionofonegene (Nmar_1608), all of the genes implicated in the3-hydroxyproprionate/4-hydrobutyrate pathway for carbon assimila-tion in N. maritimus are present and show highest similarity to thegenes ofC. symbiosum (34, 35).AlthoughN.maritimus andM. sedulamost likely use the same CO2-fixation reaction sequences, not allindividual reactions appear to be catalyzed by identical enzymes. Inone instance, the stepwise reductive transformation of malonyl-CoAto propionyl-CoA involves five enzymes in M. sedula (33, 36, 37).Although the N. maritimus genome lacks any close homologs ofthe M. sedula genes, it contains alternative alcohol dehydrogenases,aldehyde dehydrogenases, acyl-CoA synthetases, and enoyl-CoAhydratases possibly fulfilling the same functions. Similarly, M. sedulacatalyzes theactivationof3-hydroxypropionate to3-hydroxypropionyl-CoA, using an AMP-forming 3-hydroxypropionyl-CoA synthetase(37). The N. maritimus genome lacks an obvious homolog, althoughdoescode foranADP-formingacyl-CoAsynthetase (Nmar_1309) thatsuggests a more energy efficient alternative.In addition to the genes coding for the 3-hydroxypropionate/4-

hydroxybutyrate pathway, the genome of N. maritimus containsa number of genes encoding enzymes of the tricarboxylic acid(TCA) cycle. No homologs for genes coding for a citrate-cleavingenzyme (ATP citrate lyase or citryl-CoA lyase) (38) were identi-fied, permitting exclusion of the reductive TCA cycle as a pathwayfor carbon fixation. The lack of these genes suggests that N. mar-itimus utilizes either an incomplete (or horseshoe-type) TCA cyclefor strictly biosynthetic purposes or possibly a complete oxidativeTCA cycle.N. maritimus grows on a completely inorganic medium, in-

dicating the genes coding for essential biosynthetic capacity (SIMaterials and Methods), yet its genomic inventory also suggestssome flexibility in the utilization of organic sources of phos-phorus and carbon. Two systems for phosphorous acquisition aresuggested: the high-affinity, high-activity phosphate transportsystem (pstSCAB, Nmar_0479, Nmar_0481–0483) and a phos-phonate transporter (Nmar_0873–0875). However, because thegenome lacks genes encoding known C-P lyases and hydrolases(39), and phosphate limitation is not alleviated by supplemen-tation with phosphonates common in the marine environment

Out

In

pcy

pcy

C I

NH3 + O2

H O + NH OHH O + HNO

QH2

Q

4H+ + HNO2H+ + HNO

AMOQH2

Q

pcy

QH2

QNDH I

2 H+

NO + H2OCuNIR

ATPaseaa3

Nitrite Reduction

Complex VComplex IVComplex IIIComplex I

Ammonia Metabolism

H+

NAD+ + H+ NADHH+

H+

H+ 2 H+

4 H+0.5 O2 H2O

4 H+ADP + P ATP

4 H+

H+

CuHAONXOR

QRED2 e-

4 e-2 e-

2 2

22

2

i

Fig. 1. Proposed AOA respiratory pathway. Text indi-cates the described possible hydroxylamine (blue textand arrows) and nitroxyl (green) pathways. Red arrowsindicate electron flow not involved in ammonia oxida-tion. Blue shading denotes blue copper-containing pro-teins. Pink box indicates possible alternative respiratoryelectron sink. Hexagons containing Q and QH2 representthe oxidized and reduced quinone pools, respectively.

8820 | www.pnas.org/cgi/doi/10.1073/pnas.0913533107 Walker et al.

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st02.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/pnas.200913533SI.pdf?targetid=nameddest=STXThttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/pnas.200913533SI.pdf?targetid=nameddest=STXThttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/pnas.200913533SI.pdf?targetid=nameddest=STXTwww.pnas.org/cgi/doi/10.1073/pnas.0913533107

(e.g., aminoethylphosphonate), there is as yet no support fora functional phosphonate utilization pathway. Numerous organictransport functions are also evident, broadly encompassing trans-porters for different amino acids, dipeptides/oligopeptides, sul-fonates/taurine, and glycerol. Additional physiological charac-terization will likely demonstrate some capacity for mixotrophicgrowth, as suggested by isotopic studies of natural populations(40–42). No genomic evidence exists for growth on urea, asN. maritimus lacks the homologs of the putative urease and ureatransporter genes identified in C. symbiosum (35).

Noncoding RNA Genes. The N. maritimus genome contains a fullcomplement of essential noncoding RNA (ncRNA) genes, in-cluding one copy each of 5S/16S/23S ribosomal RNAs, RNase P,SRP RNA, and 44 transfer RNAs (Table S3). In addition to sixnormally placed canonical tRNA introns, noncanonical intronswere found at different positions in six of the tRNAs (ValCAC,Met,Trp, ArgCCT, LeuTAA, and GluTTC), a phenomenon previouslyobserved only in thermophiles and C. symbiosum (34, 43). Allother sequenced crenarchaea (includingC. symbiosum) contain atleast 46 tRNAs. N. maritimus lacks tRNA sequences coding forProCGG or ArgCCG, perhaps resulting from (or related to) the lowG + C content of the genome and preference for protein codonsending in A/T. Other archaeal genomes with relatively low G+ Ccontent, such as the euryarchaeon Haloquadratum walsbyi, alsolack these tRNAs, while possessing the exact complement of 44tRNAs found in N. maritimus (43). This occurrence likely reflectsa difference in posttranscriptional modification of the wobble baseof tRNAs ProTGG and ArgTCG, allowing more efficient decodingof the rare codons CCG and CGG, respectively (44).Six candidates for C/D box small RNAs (sRNAs) were identified

(Nmar_sR1–sR6). Most C/D box sRNAs guide the precise posi-tioning of posttranscriptional 2′-O-methyl group addition to rRNAsor tRNAs, a process also occurring in eukaryotes, but notBacteria. InN. maritimus, predicted targets of 2′-O-methylation may include thewobble base of the LeuCAA tRNA and two different positions sep-arated by 26 nucleotides in the 23S rRNA. Before the N. maritimusand C. symbiosum genomes, multiple C/D box sRNAs were foundalmost exclusively in hyperthermophilic archaea (45). Detection ofthese conserved, syntenic guide sRNAs in the two mesophilic cren-archaeal genomes supports an RNA stabilization/chaperone func-tion not seen in other archaeal mesophiles and possiblymore similarto their predicted function(s) in eukaryotes (46).

Regulation of Transcription. The genome contains at least eighttranscription factor B (TFB) and two TATA-box binding protein

(TBP) (Table S3) genes required for starting site-specific tran-scription initiation, making N. maritimus among the densest andrichest archaeal genomes for these transcription factors. TFBand TBP are thought to serve functions similar to the bacterialsigma factors (e.g., modulating cellular function in response tofluctuating environmental conditions) in Archaea with genomescoding for multiple copies, with optimal TFB/TPB partners (47).Although many other archaeal genomes contain multiple copiesof these transcription factors, only the haloarchaea have morethan five TFB genes (47). The functional significance of thisexceptionally high density of regulatory factors in an apparentlymetabolically specialized organism will likely be informed byfuture transcriptional analyses of different growth states. Genesfor two widely distributed types of archaeal chromatin proteinsare present, an archaeal histone (Nmar_0683) and two Albagenes (Nmar_0255 and Nmar_0933). These are thought tomaintain chromosomal material in a state permitting polymeraseaccessibility, with differential expression possibly providing foraltered global transcription (48).

Unique Cell Division Machinery and Previously Uncharacterized Instanceof Archaeal Biosynthesis of Hydroxyectoine.TheN. maritimus genomecontains elements homologous with two systems of cell division: ftsZ(Nmar_1262) and cdvABC (Nmar_0700, _0816, and _1088). ThecdvABC operon, induced at the onset of genome segregation and celldivision, codes for machinery related to the eukaryotic endosomalsorting complex (49, 50). With the exception of N. maritimus, C.symbiosum, and the Thermoproteales (where the cell division ma-chinery remains uncharacterized), all available archaeal genomeshave either the FtsZ or theCdv cell divisionmachinery, but not both.The two cell division systems may comprise a hybrid mechanism orserve two distinct processes in marine Crenarchaeota.The genome of N. maritimus also encodes for synthesis of

the compatible solute hydroxyectoine: ectoine synthase (ectC, Nmar_1344), a L-2,4-diaminobutyrate transaminase (ectB, Nmar_1345),aL-2,4-diaminobutyrateacetyltransferase (ectA,Nmar_1346), andanectoine hydroxylase (ectD, Nmar_1343). Although widely distributedamong Bacteria (in particular the genome sequences of marineorganisms), the presence of these genes in N. maritimus representsa unique indication of archaeal biosynthesis.

Relationship to C. symbiosum. The genome of N. maritimus differssignificantly in G + C content and size from that of the closely re-lated sponge symbiont (∼97% 16S rRNA gene sequence identity).Despite the differences in overall genomic G + C content (34.2%forN.maritimus versus 57.7%forC. symbiosum), theG+Ccontent

A B

C

Fig. 2. Synteny plots comparing the N. maritimus genome with (A) the Cenarchaeum symbiosum A type genome, (B) crenarchaeal genome fragments, and(C) Sargasso Sea fosmids. Vertical gray bar indicates location of ribosomal RNA operon.

Walker et al. PNAS | May 11, 2010 | vol. 107 | no. 19 | 8821

MICRO

BIOLO

GY

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st03.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st03.doc

for the rRNAs is largely similar between both organisms (50–52%).The higher ORF density (1.19 ORF/kb) relative to C. symbiosum(0.986 ORF/kb) results principally from the 0.4 Mbp smaller ge-nome of N. maritimus, not from a large disparity in the number ofpredicted ORFs. The two genomes share 1,267 genes in common(when compared via reciprocal BLAST with expectation cutoffvalues of 10−4), yet there is little conservation of synteny (Fig. 2C,Table S4). Most of the increased size of the C. symbiosum genomeand much of the divergence in gene content are associated withdiscrete regions (“islands”), althoughnoobvious functionality couldbe assigned to individual islands. Homologs for a majority of genesimplicated in the archaeal ammonia oxidation pathway (51 of 69listed in Table S2) appear present in C. symbiosum.

Phylogeny and Evolution. The widely distributed Group I archaeallineage with which N. maritimus is affiliated was earlier assigned tothe hyperthermophilic Crenarchaeota (3). Questions regarding thisassociation arose through phylogenetic analysis of C. symbiosumribosomal proteins, indicating possible divergence before theCrenarchaeota–Euryarchaeota split and therefore deserving pro-visional assignment to a new archaeal kingdom, the Thaumarch-aeota (51). The basal position of the Group I archaea previouslyinferred from protein sequences encoded by the C. symbiosum ge-nome was reexamined by reanalysis of the combined dataset, usingpatterns of gene distribution (Table S5) and phylogeny inference.Maximum-likelihood analyses confirmed the basal branching withsignificant statistical support (bootstrap value = 90%, Fig. S4).Bayesian analysis of a selection of species from the same datasetproduced results linking C. symbiosum and N. maritimus as sistertaxa ofCrenarchaeota, albeit with nonsignificant support (posteriorprobability = 0.88). Although a definitive placement within theArchaea stillmust be confirmed by inclusion of genomes frommoredistantly related lineages, both analyses strongly support a lineagedistinct from all other cultivated Crenarchaeota.

High Similarity to the Metagenome of the Globally Distributed MarineGroup I Archaea. The genome of N. maritimus shares remarkable

conservation of gene content and gene order with numerous ar-chaeal sequences previously recovered in fosmid libraries and re-cent oceanic surveys. The Antarctic genome fragments DeepAnt-EC39 (taken from 500 m depth) (52) and cosmid 74A4 (from sur-facewaters) (53) both share veryhigh syntenywithportionsof theN.maritimus genome (Fig. 2B and Table S6) despite significant dif-ferences in rRNA sequence identity (93 and 98%, respectively).Sixteen Sargasso Sea contigs (93% 16S rRNA gene sequenceidentity with N. maritimus) also have high synteny (Fig. 2C andTable S6). Retrieval of DNA fragments from the Global OceanSampling (GOS) database using differential protein sequencesimilarity showed that N. maritimus-like sequences constituted anaverage of 1.15% of all sequences across a wide range of temper-atures (9–29 °C), salinities (freshwater to seawater, 0.1–63 practicalsalinity units), two open oceans, and several coastal environments(Fig. 3A). The Block Island, NY, coastal site and the Lake Gatun,Panama Canal, site (neither of which share any notable physical/chemical characteristics other than sample depth) both exhibitednotably large increases in density (>2.5%). The available GOSsequences provided almost complete and uniform coverage of theN. maritimus genome (Fig. 3B), although at least three significantgaps in coverage exist (possibly corresponding with uniqueN. maritimus genomic islands). Whereas some of the coverage mayresult in matches to bacterial sequences, particularly for very highlyconserved proteins, the majority of recruited proteins had >50%sequence identity to N. maritimus proteins. Together, these sharedgenomic features suggestN. maritimus is representative of many ofthe globally abundant marine Group I Crenarchaeaota and that itshould provide a useful model for developing an understanding ofthebasicphysiologyof theseabundantandcosmopolitanorganisms.A comparison of the sequence content and genome organi-

zation hints at functionally more divergent marine populationtypes. N. maritimus has limited syntenic similarity to a deep-water population represented by the North Pacific fosmid 4B7(93% 16S rRNA sequence identity), but shares proteins highlysimilar to most of those coded on this fosmid. Previous com-parison of marine crenarchaeal genomic fragments reportedchanges in genomic organization with sampling depths, sugges-tive of depth-related habitat types (54). Coupled with recentevidence indicating varied physiological lifestyles along depthand latitudinal gradients, distinct crenarchaeal ecotypes likelyexist, analogous to that observed in marine cyanobacteria (2, 55).However, no clear correlations currently exist between environ-mental parameters and the similarity of recruited fragments.The genome sequence presented here also offers further insight

into theecological successofAOA.Forexample,using the likelymoreenergy-efficient 3-hydroxypropionate/4-hydroxybutyrate pathway forCO2 fixation rather than the Calvin–Bassham–Benson cycle used byAOBcouldprovideagrowthadvantage.Furtherecological advantagemaybeconferredby theirpotential capacity formixotrophic growthorthe use of copper as a major redox-active metal for respiration ingenerally iron-limited oceans. However, a deeper understanding ofthe remarkable success of this archaeal lineage will come only frommore detailed physiological, biochemical, and genetic characteriza-tion of N. maritimus and additional environmental isolates.

Materials and MethodsGenome sequencing was performed on high-molecular-weight DNAextracted from two cultures of N. maritimus. Whole-genome shotgun se-quencing of 3-, 8-, and 40-kb DNA libraries by the Joint Genome Instituteproduced at least 8× coverage. Annotation of the closed genome was per-formed using Department of Energy (DOE) computational support at OakRidge National Laboratory and The Institute for Genomic Research (TIGR)Autoannotation Service in conjunction with Manatee visualization software.Complete details describing high-molecular-weight DNA purification andsequence analysis are found in SI Materials and Methods.

ACKNOWLEDGMENTS. The authors thank David Bruce and Paul Richardsonfrom the Joint Genome Institute for facilitating genome sequencing. This

Fig. 3. Comparison of N. maritimus genome to metagenomic sequence readsfrom GOS. (A) Percentage of reads from each GOS site that align to the N.maritimus genome by protein sequence similarity. (B) Number of GOS readshomologous to eachN.maritimus protein-coding gene. Counts of 40 aremostlydue to highly conserved proteins that include contaminants from other clades.

8822 | www.pnas.org/cgi/doi/10.1073/pnas.0913533107 Walker et al.

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st04.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st02.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st05.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/pnas.200913533SI.pdf?targetid=nameddest=STXThttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st06.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st06.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/pnas.200913533SI.pdf?targetid=nameddest=STXTwww.pnas.org/cgi/doi/10.1073/pnas.0913533107

work was supported by the Department of Energy Microbial GenomeProgram, byNational Science FoundationMicrobial Interactions andProcessesGrant MCB-0604448 (to D.A.S. and J.R.d.l.T.), by National Science FoundationMolecular andCellular BiosciencesGrantMCB-0920741 (toD.A.S.), byNationalScience Foundation Biological Oceanography Grants OCE-0623174 (to D.A.S.)

and OCE-0623908 (to S.M.S.), by National Science Foundation Grant EF-0412129 (to M.G.K.), by incentive funds from the University of Louisville VPResearch office (to M.G.K.), by the Deutsche Forschungsgemeinschaft (M.K.),by US Department of Agriculture Grant 2010-65115-20380 (to A.C.R.), and bya Salk Institute Innovation Grant (to G.M.).

1. Karner MB, DeLong EF, Karl DM (2001) Archaeal dominance in the mesopelagic zoneof the Pacific Ocean. Nature 409:507–510.

2. Agogué H, Brink M, Dinasquet J, Herndl GJ (2008) Major gradients in putativelynitrifying and non-nitrifying Archaea in the deep North Atlantic. Nature 456:788–791.

3. DeLong EF (1992) Archaea in coastal marine environments. Proc Natl Acad Sci USA 89:5685–5689.

4. Fuhrman JA, McCallum K, Davis AA (1992) Novel major archaebacterial group frommarine plankton. Nature 356:148–149.

5. Venter JC, et al. (2004) Environmental genome shotgun sequencing of the SargassoSea. Science 304:66–74.

6. Wuchter C, et al. (2006) Archaeal nitrification in the ocean. Proc Natl Acad Sci USA103:12317–12322.

7. Könneke M, et al. (2005) Isolation of an autotrophic ammonia-oxidizing marinearchaeon. Nature 437:543–546.

8. Leininger S, et al. (2006) Archaea predominate among ammonia-oxidizing prokaryotesin soils. Nature 442:806–809.

9. Prosser JI, Nicol GW (2008) Relative contributions of archaea and bacteria to aerobicammonia oxidation in the environment. Environ Microbiol 10:2931–2941.

10. Prosser JI (1989) Autotrophic nitrification in bacteria. Adv Microb Physiol 30:125–181.11. de la Torre JR, Walker CB, Ingalls AE, Könneke M, Stahl DA (2008) Cultivation of

a thermophilic ammonia oxidizing archaeon synthesizing crenarchaeol. EnvironMicrobiol 10:810–818.

12. Hatzenpichler R, et al. (2008) A moderately thermophilic ammonia-oxidizingcrenarchaeote from a hot spring. Proc Natl Acad Sci USA 105:2134–2139.

13. Martens-Habbena W, Berube PM, Urakawa H, de la Torre JR, Stahl DA (2009)Ammonia oxidation kinetics determine niche separation of nitrifying Archaea andBacteria. Nature 461:976–979.

14. Sernova NV, Gelfand MS (2008) Identification of replication origins in prokaryoticgenomes. Brief Bioinform 9:376–391.

15. Makarova KS, Sorokin AV, Novichkov PS, Wolf YI, Koonin EV (2007) Clusters oforthologous genes for 41 archaeal genomes and implications for evolutionarygenomics of archaea. Biol Direct 2:33.

16. Schäfer G (2004) Respiration in Archaea and Bacteria, ed Zannoni D (Springer,Dordrecht, The Netherlands), pp 1–33.

17. Beaumont HJ, Lens SI, Westerhoff HV, van Spanning RJ (2005) Novel nirK clustergenes in Nitrosomonas europaea are required for NirK-dependent tolerance tonitrite. J Bacteriol 187:6849–6851.

18. Andersen CL, Matthey-Dupraz A, Missiakas D, Raina S (1997) A new Escherichia coligene, dsbG, encodes a periplasmic protein involved in disulphide bond formation,required for recycling DsbA/DsbB and DsbC redox proteins.Mol Microbiol 26:121–132.

19. Meima R, et al. (2002) The bdbDC operon of Bacillus subtilis encodes thiol-disulfideoxidoreductases required for competence development. J Biol Chem 277:6994–7001.

20. Raina S, Missiakas D (1997) Making and breaking disulfide bonds. Annu Rev Microbiol51:179–202.

21. Hiniker A, Collet JF, Bardwell JC (2005) Copper stress causes an in vivo requirement forthe Escherichia coli disulfide isomerase DsbC. J Biol Chem 280:33785–33791.

22. Pogliano J, Lynch AS, Belin D, Lin EC, Beckwith J (1997) Regulation of Escherichia colicell envelope proteins involved in protein folding and degradation by the Cpx two-component system. Genes Dev 11:1169–1182.

23. Kershaw CJ, Brown NL, Constantinidou C, Patel MD, Hobman JL (2005) The expressionprofile of Escherichia coli K-12 in response to minimal, optimal and excess copperconcentrations. Microbiology 151:1187–1198.

24. Narindrasorasak S, Yao P, Sarkar B (2003) Protein disulfide isomerase, a multifunctionalprotein chaperone, shows copper-binding activity. Biochem Biophys Res Commun 311:405–414.

25. Sliskovic I, Raturi A, Mutus B (2005) Characterization of the S-denitrosation activity ofprotein disulfide isomerase. J Biol Chem 280:8733–8741.

26. Ramachandran N, Root P, Jiang XM, Hogg PJ, Mutus B (2001) Mechanism of transferof NO from extracellular S-nitrosothiols into the cytosol by cell-surface proteindisulfide isomerase. Proc Natl Acad Sci USA 98:9539–9544.

27. Nicol GW, Tscherko D, Chang L, Hammesfahr U, Prosser JI (2006) Crenarchaealcommunity assembly and microdiversity in developing soils at two sites associatedwith deglaciation. Environ Microbiol 8:1382–1393.

28. Lieberman RL, Rosenzweig AC (2005) Crystal structure of a membrane-boundmetalloenzyme that catalyses the biological oxidation of methane. Nature 434:177–182.

29. Klotz MG, Stein LY (2008) Nitrifier genomics and evolution of the nitrogen cycle.FEMS Microbiol Lett 278:146–156.

30. Fukuto JM, Switzer CH, Miranda KM, Wink DA (2005) Nitroxyl (HNO): Chemistry,biochemistry, and pharmacology. Annu Rev Pharmacol Toxicol 45:335–355.

31. Miranda KM, et al. (2003) A biochemical rationale for the discrete behavior of nitroxyland nitric oxide in the cardiovascular system. Proc Natl Acad Sci USA 100:9196–9201.

32. Arp DJ, Chain PS, Klotz MG (2007) The impact of genome analyses on ourunderstanding of ammonia-oxidizing bacteria. Annu Rev Microbiol 61:503–528.

33. Berg IA, KockelkornD, BuckelW, Fuchs G (2007) A 3-hydroxypropionate/4-hydroxybutyrateautotrophic carbon dioxide assimilation pathway in Archaea. Science 318:1782–1786.

34. Hallam SJ, et al. (2006) Genomic analysis of the uncultivated marine crenarchaeoteCenarchaeum symbiosum. Proc Natl Acad Sci USA 103:18296–18301.

35. Hallam SJ, et al. (2006) Pathways of carbon assimilation and ammonia oxidationsuggested by environmental genomic analyses of marine Crenarchaeota. PLoS Biol 4:e95.

36. Alber B, et al. (2006) Malonyl-coenzyme A reductase in the modified 3-hydroxypropionatecycle for autotrophic carbon fixation in archaeal Metallosphaera and Sulfolobus spp.J Bacteriol 188:8551–8559.

37. Alber BE, Kung JW, Fuchs G (2008) 3-Hydroxypropionyl-coenzyme A synthetase fromMetallosphaera sedula, an enzyme involved in autotrophic CO2 fixation. J Bacteriol190:1383–1389.

38. Hügler M, Huber H, Molyneaux SJ, Vetriani C, Sievert SM (2007) Autotrophic CO2fixation via the reductive tricarboxylic acid cycle in different lineages within thephylum Aquificae: Evidence for two ways of citrate cleavage. Environ Microbiol 9:81–92.

39. Quinn JP, Kulakova AN, Cooley NA, McGrath JW (2007) New ways to break an oldbond: The bacterial carbon-phosphorus hydrolases and their role in biogeochemicalphosphorus cycling. Environ Microbiol 9:2392–2400.

40. Herndl GJ, et al. (2005) Contribution of Archaea to total prokaryotic production in thedeep Atlantic Ocean. Appl Environ Microbiol 71:2303–2309.

41. Ingalls AE, et al. (2006) Quantifying archaeal community autotrophy in the mesopelagicocean using natural radiocarbon. Proc Natl Acad Sci USA 103:6442–6447.

42. Ouverney CC, Fuhrman JA (2000) Marine planktonic archaea take up amino acids.Appl Environ Microbiol 66:4829–4833.

43. Chan PP, Lowe TM (2009) GtRNAdb: A database of transfer RNA genes detected ingenomic sequence. Nucleic Acids Res 37 (Database issue):D93–D97.

44. Grosjean H, Marck C, de Crécy-Lagard V (2007) The various strategies of codondecoding in organisms of the three domains of life: Evolutionary implications. NucleicAcids Symp Ser (Oxf) 51:15–16.

45. Dennis PP, Omer A, Lowe T (2001) A guided tour: Small RNA function in Archaea. MolMicrobiol 40:509–519.

46. Maxwell ES, Fournier MJ (1995) The small nucleolar RNAs. Annu Rev Biochem 64:897–934.

47. Facciotti MT, et al. (2007) General transcription factor specified global generegulation in archaea. Proc Natl Acad Sci USA 104:4630–4635.

48. Sandman K, Reeve JN (2005) Archaeal chromatin proteins: Different structures butcommon function? Curr Opin Microbiol 8:656–661.

49. Lindås AC, Karlsson EA, Lindgren MT, Ettema TJ, Bernander R (2008) A unique celldivision machinery in the Archaea. Proc Natl Acad Sci USA 105:18942–18946.

50. Samson RY, Obita T, Freund SM, Williams RL, Bell SD (2008) A role for the ESCRTsystem in cell division in archaea. Science 322:1710–1713.

51. Brochier-Armanet C, Boussau B, Gribaldo S, Forterre P (2008) MesophilicCrenarchaeota: Proposal for a third archaeal phylum, the Thaumarchaeota. Nat RevMicrobiol 6:245–252.

52. Stein JL, Marsh TL, Wu KY, Shizuya H, DeLong EF (1996) Characterization ofuncultivated prokaryotes: Isolation and analysis of a 40-kilobase-pair genomefragment from a planktonic marine archaeon. J Bacteriol 178:591–599.

53. Béjà O, et al. (2002) Comparative genomic analysis of archaeal genotypic variants ina single population and in two different oceanic provinces. Appl Environ Microbiol68:335–345.

54. López-García P, Brochier C, Moreira D, Rodríguez-Valera F (2004) Comparativeanalysis of a genome fragment of an uncultivated mesopelagic crenarchaeote revealsmultiple horizontal gene transfers. Environ Microbiol 6:19–34.

55. Johnson ZI, et al. (2006) Niche partitioning among Prochlorococcus ecotypes alongocean-scale environmental gradients. Science 311:1737–1740.

Walker et al. PNAS | May 11, 2010 | vol. 107 | no. 19 | 8823

MICRO

BIOLO

GY

Supporting InformationWalker et al. 10.1073/pnas.0913533107SI Materials and MethodsHigh-Molecular-Weight Genomic DNA Preparation. Two cultures ofNitrosopumilus maritimus strain SCM1 were grown as previouslydescribed in 500 mL of media in 1-L flasks (1). Cells from bothcultures were harvested in late-exponential phase using Sterivexfilters for one culture and a 0.1-μm filter for the other. High-molecular-weight DNAwas isolated as previously described usingeither agarose plugs (2) or a modified guanadinium thiocynateprotocol (3). Cells from the Sterivex filter were resuspended in1 mL of 2× STE buffer [1 M NaCl, 0.1 M EDTA (pH 8.0), 10 mMTris (pH8.0)], extracted from the filter, andmixedwith 1 vol of 1%molten SeaPlaque LMP agarose (FMC). The mixture was cooledto 40 °C, immediately drawn into a 1-mL syringe, and placed on icefor 10 min. The agarose plug was mixed with 10 mL of lysis buffer,incubated at 37 °C for 1 h, and then transferred to 40 mL of ESPbuffer (1%Sarkosyl–1mg of ProteinaseKperml in 0.5MEDTA).After incubation at 55 °C for 16 h, the solution was replaced withfresh ESP buffer and incubated at 55 °C for another hour. DNAwas purified using phenol:chloroform:isoamyl alcohol (24:24:1)and recovered by precipitation with isopropanol.Cells collected on the 0.1-μm filter were resuspended in 100 μL of

Tris-EDTA (pH 8.0) and 100 mg/mL lysozyme before incubating at37 °C for 30 min. Then, 3.0 mL of a solution containing 5 M guani-dinium thiocyanate, 100 mM EDTA (pH 8.0), and 0.5% (vol/vol)sarkosyl was added. The solution wasmixed gently for 15min beforebeing cooledon ice for10min.After cooling, anequal volumeof cold7.5 M ammonium acetate was added, and the solution was mixedgently and cooledon ice. Purification andprecipitationofDNAwereperformed with chloroform:isoamyl alcohol (24:1) and isopropanol.

Genome Sequencing.A completely sequenced and closed genomeof N. maritimus was obtained through collaboration with theJoint Genome Institute. Whole-genome shotgun sequencing of3-, 8-, and 40-kb DNA libraries produced at least 8× coverage ofthe entire genome. Specifics of clone library generation, se-quencing, and assembly strategiesmay be found at theDOE JGIwebsite (www.jgi.doe.gov/sequencing/index.html).

Genome Sequence Analysis. Autoannotation of the closed genomesequence was performed by both the Computational Biologygroup at Oak Ridge National Laboratory (http://genome.ornl.gov/microbial/nmar/02jul07/) and the TIGR AutoannotationService (now hosted by JCVI; details available from http://www.jcvi.org/cms/research/projects/annotation-service/overview/). Thegenome visualization software Manatee (release 2.4.1; latestversion available from http://manatee.sourceforge.net/) was usedfor manual curation. Analysis of potential transporter genes wasperformed using the Transporter Automatic Annotation Pipe-line (TransAAP) through the TransportDB genomic comparisontool (membranetransport.org).Direct comparisons with theC. symbiosum genome and genome

fragments were performed using the Artemis Comparison Tool (3)

with a comparison library generated through WebACT (www.webact.org/WebACT/home). Orthologous genes shared betweenthese two organisms were identified through reciprocal BLASTsearches,with anexpectation cutoff valueof 10−4 and aminimumof75% alignable N. maritimus sequence. Comparisons with theGlobal Ocean Sampling (GOS) and Sargasso Sea metagenomicdatasets were performed using several single-copy universal ar-chaeal genes to determine a count of ∼15 archaeal genomes in theGOS dataset. An initial set of candidateN. maritimus-like proteinswas found by BLASTP searches with each N. maritimus protein-coding gene, using a cutoff of 100 hits and a maximum expectedvalue of e= 10. This set consisted of 125,326 peptides drawn from107,223 scaffolds. Neighboring ORFs on any scaffold with two ormore hits were added to make a total of 319,585 peptides,amounting to 5.2% of all ORFs in GOS. These sequences werefiltered by BLAST alignment to four sequence datasets, containing24 completeproteomes fromdiverse euryarchaeota, crenarchaeota,and bacteria. Sequences scoringmore highly toN.maritimus and/orC. symbiosum proteins than any other entry in these datasets wereretained, giving a filtered dataset of 21,278 ORFs. Coverage of theN. maritimus proteome was measured by bidirectional BLASTP ofN. maritimus proteins against the filtered dataset to assign putativeorthologs. The average coverage was ∼11×, although some highlyconservedgenes had amuch greater number of hits, probably due torecruitment of nonarchaeal homologs: 48 ORFs had >30 hits, ofwhich most were highly conserved. Searches for CRISPR regionswere performed using the Java-based CRISPR recognition tool(CRT) with least stringent settings (4).

Maximum-Likelihood and Bayesian Trees of the Archaeal Domain.The trees are based on the concatenation of ribosomal proteinsusedbyBrochier-Armanet et al. (5)but including sequences fromN.maritimus and from “Candidatus Korarchaeum cryptofilum” OPF8.Sequences were aligned using MUSCLE (6). Resulting alignmentswere visually inspected and improved with theMUST software (7).Regions where homology between sites was doubtful were removedfrom further phylogenetic analyses. A total of 6,142 positions werekept for the phylogenetic analyses. The maximum-likelihood treewas computed with PHYML, using the WAG model corrected bya gamma law to take into account evolutionary rate among sitevariations (8).Theparameteralphaof thegammadistributionas theproportion of invariable sites was estimated from the dataset. Therobustness of each branchwas estimated by thebootstrap procedureimplemented in PHYML.ABayesian tree analysis on a subset of 29taxa was performed using MrBayes 3.2 (9) with a mixed model ofamino acid substitution and a gamma distribution (eight discretecategories and an estimated proportion of invariant sites) to takeinto account among-site rate variation. MrBayes was run with fourchains for 1 million generations and trees were sampled every 100generations. To construct the consensus tree, the first 1,500 treeswere discarded as ‘‘burn-in.’’ The reduction of the taxonomic sam-pling was necessary to reduce the computation time.

1. Könneke M, et al. (2005) Isolation of an autotrophic ammonia-oxidizing marinearchaeon. Nature 437:543–546.

2. Stein LY, et al. (2007) Whole-genome analysis of the ammonia-oxidizing bacterium,Nitrosomonas eutropha C91: Implications for niche adaptation. Environ Microbiol 9:2993–3007.

3. Carver TJ, et al. (2005) ACT: The Artemis comparison tool. Bioinformatics 21:3422–3423.4. Bland C, et al. (2007) CRISPR recognition tool (CRT): A tool for automatic detection of

clustered regularly interspaced palindromic repeats. BMC Bioinformatics 8:209.5. Brochier-Armanet C, Boussau B, Gribaldo S, Forterre P (2008) Mesophilic Crenarchaeota:

Proposal for a thirdarchaeal phylum, theThaumarchaeota.NatRevMicrobiol6:245–252.

6. Edgar RC (2004) MUSCLE: Multiple sequence alignment with high accuracy and highthroughput. Nucleic Acids Res 32:1792–1797.

7. Philippe H (1993) MUST, a computer package of management utilities for sequencesand trees. Nucleic Acids Res 21:5264–5272.

8. Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm to estimate largephylogenies by maximum likelihood. Syst Biol 52:696–704.

9. Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference undermixed models. Bioinformatics 19:1572–1574.

Walker et al. www.pnas.org/cgi/content/short/0913533107 1 of 5

http://www.jgi.doe.gov/sequencing/index.htmlhttp://genome.ornl.gov/microbial/nmar/02jul07/http://genome.ornl.gov/microbial/nmar/02jul07/http://www.jcvi.org/cms/research/projects/annotation-service/overview/http://www.jcvi.org/cms/research/projects/annotation-service/overview/http://manatee.sourceforge.net/http://membranetransport.orghttp://www.webact.org/WebACT/homehttp://www.webact.org/WebACT/homewww.pnas.org/cgi/content/short/0913533107

0.5

Thermoplasma volcanium GSS1 [rusticyanin; gi13542283]Acidithiobacillus ferrooxidans [rusticyanin; gi3282064]

Nmar0167

Nmar0747

95

8399

61

62100

9999

96

8891

79

64

82

82

Thermobaculum terrenum ATCC BAA-798 [plastocyanin; gi227373287]Sphaerobacter thermophilus DSM 20745 [plastocyanin; gi229876615]

Mycobacterium avium subsp. avium ATCC 25291 [copper-binding protein; gi254776845]Mycobacterium avium 104 [copper-binding protein; gi118462975]

Mycobacterium intracellulare ATCC 13950 [hypothetical protein; gi254819204]Catenulispora acidiphila DSM 44928 [blue (type 1) copper domain protein; gi256391720]

Streptomyces sp. AA4 [hypothetical protein; gi256667890]Burkholderia cenocepacia PC184 [plastocyanin; gi254250279]

Methylocella silvestris BL2 [blue (type 1) copper domain protein; gi217976580]

Methylocella silvestris BL2 [blue (type 1) copper domain protein; gi217976580]Nitrobacter hamburgensis X14 [blue (type 1) copper domain protein; gi92118183]

Methylobacterium nodulans ORS [blue (type 1) copper domain protein; gi220920817]

Chromobacterium violaceum ATCC 12472 [hypothetical protein; gi34498952]

Methanosarcina mazei Go1 [copper-binding protein; gi21228445]Methanosarcina acetivorans C2A [plastocyanin/azurin family copper-binding protein; gi20090218]

Methanosarcina mazei Go1 [copper-binding protein; gi21228446]Methanosarcina acetivorans C2A [plastocyanin/azurin family copper-binding protein; gi20090219]

C. symbiosum A [copper-binding protein; gi118575216]

Nmar0621

Nmar0361Candidatus Koribacter versatillis [plastocyanin-like protein; gi94968815]

Nmar0902

Thermobaculum terrenum ATCC BAA-798 [plastocyanin; gi227373254]Deinococcus deserti VCD115 [copper-binding protein; gi226356528]Synechococcus sp. WH 8102 [plastocyanin; gi33866031]Prochlorococcus marinus MIT 9312 [plastocyanin; gi78778966]

Anabaena variabilisATCC 29413 [plastocyanin; gi75908957]Gloeobacter violaceus PCC 7421 [plastocyanin; gi35212909]

Jannaschia sp. CCS1 [blue (type 1) copper domain protein; gi89056161]Nmar0718

Bradyrhizobium japonicum USDA 110 [putative Amicyanin precursor; gi27378126]Paracoccus denitrificans PD1222 [amicyanin; gi119387457]

Methanosarcina mazei Go1 [hypothetical protein; gi21226177]Methanosarcina acetivorans C2A [hypothetical protein; gi20090537]

Nmar0190

Methanococcus maripaludis C7 [copper-binding protein; gi150402166]Methanococcus maripaludis S2 [copper-binding protein; gi45358560]

Nmar1913Nmar1789

Nmar0300Nmar0762

C. symbiosum A [copper-binding protein; gi118576966]C. symbiosum A [copper-binding protein; gi118576965]

Nmar0516Nmar1087

C. symbiosum A [hypothetical protein; gi118575848]C. symbiosum A [hypothetical protein; gi118575831]

uncultured marine Crenarchaeote [putative copper-binding protein; gi167044521]uncultured marine Crenarchaeote [putative copper-binding protein; gi167042919]

uncultured marine Crenarchaeote [putative copper-binding protein; gi167045254]

C. symbiosum A [copper-binding protein; gi118575244]Nmar0121

Nmar0732Nmar0324

C. symbiosum A [copper-binding protein; gi118576967]

C. symbiosum A [copper-binding protein; gi118575724]Nmar0273

100

98

100100

79

7980

100

A.

0.1

Bacillus subtilis subsp. subtilis str. 168 [BdbD thiol-disulfide oxidoreductase; gi 16080401]Bacillus amyloliquefaciens FZB42 [BdbD; gi 154687467]

Paenibacillus larvae subsp. larvae BRL-230010 [disulfide dehydrogenase D; gi 167461952]Deinococcus geothermalis DSM 11300 [DsbA oxidoreductase; gi 94984799]

Thermobaculum terrenum ATCC BAA-798 [protein-disulfide isomerase; gi 227374421]Gemmatimonas aurantiaca T-27 [putative oxidoreductase; gi 226228008]

Sphaerobacter thermophilus DSM 20745 [protein-disulfide isomerase; gi 229876989]Thermomicrobium roseum DSM 5159 [DsbA oxidoreductase; gi 221635547]

Mycobacterium kansasii ATCC 12478 [DsbA oxidoreductase; gi 240169146]Rubrobacter xylanophilus DSM 9941 [DsbA oxidoreductase; gi 108804655]

Streptomyces coelicolor A3(2) [hypothetical protein SCO5993; gi 21224330]Streptomyces ghanaensis ATCC 14672 [hypothetical protein SghaA1_08748; gi 239928300]

Thermus aquaticus Y51MC23 [DsbA oxidoreductase; gi 218295130]Thermobifida fusca YX [protein-disulfide isomerase; gi 72161925]

Stigmatella aurantiaca DW4/3-1 [conserved hypothetical protein; gi 115374845]Solibacter usitatus Ellin6076 [DsbA oxidoreductase; gi 116625220]

Myxococcus xanthus DK 1622 [putative lipoprotein; gi 108762810]Stigmatella aurantiaca DW4/3-1 [disulfide interchange protein; gi 115379912]

Anaeromyxobacter dehalogenans 2CP-1 [DsbA oxidoreductase; gi 220919173]

Solibacter usitatus Ellin6076 [DsbA oxidoreductase; gi 116624599]Syntrophobacter fumaroxidans MPOB [DsbA oxidoreductase; gi 116751066]

Archaeoglobus fulgidus DSM 4304 [hypothetical protein AF1354; gi 11498950]Cenarchaeum symbiosum A [protein-disulfide isomerase; gi 118575694]

Nmar0218Nmar0740

Nmar0175

Nmar0752Nmar0179

Natrialba magadii ATCC 43099 ]DsbA oxidoreductase; gi 224823141]Nmar1863

Cenarchaeum symbiosum A [protein-disulfide isomerase; gi 118576454]Nmar1805Cenarchaeum symbiosum A [protein-disulfide isomerase; gi 118576169]

Nmar0169Nmar1590

Cenarchaeum symbiosum A [protein-disulfide isomerase; gi 118575427]

Nmar1606Nmar0164

Symbiobacterium thermophilum IAM 14863 [hypothetical protein STH2058; gi 51893196]Roseiflexus castenholzii DSM 13941 [DsbA oxidoreductase; gi 156741642]

Chloroflexus aggregans DSM 9485 [DsbA oxidoreductase; gi 219849651] Chloroflexus aggregans DSM 9485 [DsbA oxidoreductase; gi 219847445]

Marinobacter algicola DG893 [DsbA oxidoreductase; gi 149377658]

Roseiflexus castenholzii DSM 13941 [DsbA oxidoreductase; gi 156743646]Chloroflexus aurantiacus J-10-fl [DsbA oxidoreductase; gi 163848707]

Roseiflexus castenholzii DSM 13941 [protein-disulfide isomerase-like protein; gi 156741356]

Herpetosiphon aurantiacus ATCC 23779 [DsbA oxidoreductase; gi 159896786]Chloroflexus aurantiacus J-10-fl [DsbA oxidoreductase; gi 163845898]

Chloroflexus aggregans DSM 9485 [DsbA oxidoreductase; gi 219850456]

Vibrio harveyi HY01 [DsbA oxidoreductase; gi 153832115]Vibrio parahaemolyticus RIMD 2210633 [hypothetical protein VPA0994; gi 28900849]

Moritella sp. PE36 [putative membrane protein; gi 149908466]Shewanella benthica KT99 [hypothetical protein KT99_10483; gi 163751466]

100

100

8275

100

87

82

100

93

7379

91

67

90

100

72

88

63

100

99

75

72100

100

99

C.

Nmar0121

Nmar0167

Nmar0190

Nmar0273

Nmar0300

Nmar0324

Nmar0361

Nmar0516

Nmar0621

Nmar0718

Nmar0732

Nmar0747

Nmar0762

Nmar1087

Nmar1789

Nmar1913

100 aa

B.

Fig. S1. Phylogeny of plastocyanin-like protein sequences. Sequences with significant matches to COG3794 (PetE: Plastocyanin [Energy production andconversion]) were used to query the non-redundant protein sequence database from NCBI. Sequences from the top 20–30 non-N. maritimus hits were re-trieved and their match to the above conserved domain model verified. Sequences from experimentally characterized proteins were obtained from theavailable literature and included, aligned with ClustalW and then curated manually. Distance-based phylogenies were inferred in Phylip using the Neighbor-Joining algorithm and 100 bootstrap replicates. Bootstrap support values >60% are displayed. Nodes with

Fig. S2. Archaeal ammonia monooxygenase AmoB sequence mapped onto the crystal structure of the particulate methane monooxygenase (PDB accessioncode 1YEW). The pmoA subunit is shown in dark gray, the pmoC subunit in light gray, and the pmoB subunit in red and pink. The red part represents theregion of pmoB conserved in the predicted N. maritimus AmoB. The transmembrane helix and C-terminal cupredoxin domain shown in pink are missing in thepredicted N. maritimus AmoB. Cyan spheres represent copper ions. The grey sphere represents a zinc ion.

Walker et al. www.pnas.org/cgi/content/short/0913533107 3 of 5

www.pnas.org/cgi/content/short/0913533107

Acetyl-CoA carboxylase(Nmar_0272, 0273, 0274)

Malonyl-CoA reductaseMalonate semialdehyde reductase(unknown)

3-Hydroxypropionyl-CoA synthetase3-Hydroxypropionyl-CoA dehydrataseAcryloyl-CoA reductase (unknown)

Propionyl-CoA carboxylase(Nmar_0272, 0273, 0274)

Methylmalonyl-CoA epimeraseMethylmalonyl-CoA mutase

(Nmar_0953, 0954, 0958)

Succinyl-CoA reductase(Nmar_1608)

Succinate semialdehyde reductase(Nmar_1110 or Nmar_0161)

4-Hydroxybutyryl-CoA synthetase(Nmar_0206)

4-Hydroxybutyryl-CoA dehydratase(Nmar_0207)

Crotonyl-CoA hydratase(Nmar_1308)

3-Hydroxybutyryl-CoA dehydrogenase(Nmar_1028)

Acetoacetyl-CoA -ketothiolase(Nmar_0841 or Nmar_1631)

Fig. S3. Proposed 3-hydroxypropionate/4-hydroxybutyrate cycle for autotrophic carbon fixation by N. maritimus.

Walker et al. www.pnas.org/cgi/content/short/0913533107 4 of 5

www.pnas.org/cgi/content/short/0913533107

Other Supporting Information Files

Table S1 (DOC)Table S2 (DOC)Table S3 (DOC)Table S4 (DOC)Table S5 (DOC)Table S6 (DOC)

0.1

Cenarchaeum symbiosumCandidatus Nitrosopumilus maritimus

Candidatus Korarchaeum cryptofilumThermofilum pendens

Caldivirga maquilingensisPyrobaculum calidifontis

Pyrobaculum islandicumPyrobaculum aerophilum

Pyrobaculum arsenaticumIgnicoccus hospitalisStaphylothermus marinus

Aeropyrum pernixHyperthermus butylicus

Metallosphaera sedulaSulfolobus solfataricus

Sulfolobus acidocaldariusSulfolobus tokodaii

Nanoarchaeum equitansThermococcus gammatoleransThermococcus kodakarensis

Pyrococcus furiosusPyrococcus abyssiPyrococcus horikoshii

Methanopyrus kandleriMethanosphaera stadtmanae

Methanothermobacter thermautotrophicusMethanocaldococcus jannaschii

Methanococcus aeolicusMethanococcus maripaludisMethanococcus vannielii

Picrophilus torridusFerroplasma acidarmanus

Thermoplasma acidophilumThermoplasma volcanium

Archaeoglobus fulgidusHalobacterium sp

Natronomonas pharaonisHaloarcula marismortui

Halorubrum lacusprofundiHaloquadratum walsbyi

Haloferax volcaniiMethanocorpusculum labreanum

Methanoculleus marisnigriCandidatus Methanoregula

Methanospirillum hungateiMethanosaeta thermophila

Methanococcoides burtoniiMethanosarcina barkeriMethanosarcina acetivoransMethanosarcina mazei

Giardia lambliaEntamoeba histolytica

Leishmania majorTrypanosoma cruziTrypanosoma brucei

Cryptosporidium parvumTheileria parvaPlasmodium yoelii

Plasmodium falciparumArabidopsis thaliana

Oryza sativaDictyostelium discoideum

Homo sapiensAnopheles gambiae

Saccharomyces cerevisiaeSchizosaccharomyces pombe

94

7061

90

52

55

71

69

8275

61

88

95

38

78

97

96

90

100

100

100100

100

100

100

100

100

100

100

100100

100

100

100100

100

100

100100

100100

100

100

100100

100100

100

100

8651

97

64

100100

100100

100

100

100

100

100

100

Eucarya

ThaumarchaeotaKorarchaeota

Sulfolobales

Desulfurococcales

Thermoproteales

Nanoarchaeota

Thermococcales

MethanopyralesMethanobacteriales

Methanococcales

Thermoplasmatales

Archaeoglobales

Halobacteriales

Methanomicrobiales

Methanosarcinales

Crenarchae ota

Eu r y arch aeota

0.1

Crenarcha eo t a

Eu ry archa eota

Saccharomyces cerevisiaeDictyostelium discoideum

Oryza sativa1.00

Candidatus Korarchaeum cryptofilumCandidatus Nitrosopumilus maritimusCenarchaeum symbiosum1.00

Pyrobaculum aerophilumThermofilum pendens1.00

Aeropyrum pernixHyperthermus butylicus

1.00

Sulfolobus acidocaldariusMetallosphaera sedula1.00

1.00

1.00

0.88

Nanoarchaeum equitansThermococcus kodakarensis

Pyrococcus abyssi1.00Methanopyrus kandleri

Methanothermobacter thermautotrophicusMethanosphaera stadtmanae1.00

1.00

Methanocaldococcus jannaschiiMethanococcus maripaludis1.00

Thermoplasma volcaniumFerroplasma acidarmanus1.00

Archaeoglobus fulgidusHaloarcula marismortuiNatronomonas pharaonis1.00

Methanosarcina mazeiMethanosaeta thermophila

1.00

Methanospirillum hungateiMethanocorpusculum labreanum1.00

0.70

1.00

1.00

1.00

1.00

1.00

1.00

1.00

0.88

1.00

Eucarya

Thaumarchaeota

Korarchaeota

Sulfolobales

Desulfurococcales

Thermoproteales

Nanoarchaeota

Thermococcales

Methanopyrales

Methanobacteriales

Methanococcales

Thermoplasmatales

Archaeoglobales

Halobacteriales

Methanomicrobiales

Methanosarcinales

A.

B.

Fig. S4. (A) Maximum-likelihood phylogeny of Group 1 Archaea. The phylogeny was inferred using an alignment of concatenated R-proteins (66 taxa, 6,142positions). WAG+Inv+Gamma (4 classes); 100 replicates. (B) Bayesian tree of mesophilic Group 1 Archaea inferred using an alignment of concatenated R-proteins (29 taxa, 6,142 positions). Mixed model + Gamma (4 classes); 100 replicates.

Walker et al. www.pnas.org/cgi/content/short/0913533107 5 of 5

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st01.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st02.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st02.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st02.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st03.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st03.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st03.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st04.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st04.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st04.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st05.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st05.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st05.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st06.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st06.dochttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.0913533107/-/DCSupplemental/st06.docwww.pnas.org/cgi/content/short/0913533107