:: Biblioteca Digital de Ecopsicologia (BDE)

Hadal biosphere: Insight into the microbial ecosystem in the deepest ocean on Earth

Hadal biosphere: Insight into the microbial ecosystem

in the deepest ocean on Earth

Takuro Nunouraa,1, Yoshihiro Takakia,b, Miho Hiraia, Shigeru Shimamurab, Akiko Makabec,d,2, Osamu Koidea,

Tohru Kikuchie, Junichi Miyazakib, Keisuke Kobac, Naohiro Yoshidad,3, Michinari Sunamuraf, and Ken Takaib

bDepartment of Subsurface Geobiological Analysis and Research and aResearch and Development Center for Marine Biosciences, Japan Agency for

Marine-Earth Science & Technology, Yokosuka 237-0061, Japan; cInstitute of Agriculture, Tokyo University of Agriculture and Technology, Fuchu, Tokyo

183-8509, Japan; dDepartment of Environmental Science and Technology, Tokyo Institute of Technology, Midori-ku, Yokohama 226-8502, Japan; eGraduate

School of Nanobioscience, Yokohama City University, Kanazawa-ku, Yokohama 236-0027, Japan; and fDepartment of Earth and Planetary Science,

University of Tokyo, Bunkyo-ku, Tokyo 113-0033, Japan

Edited by David M. Karl, University of Hawaii, Honolulu, HI, and approved January 26, 2015 (received for review November 17, 2014)

Hadal oceans at water depths below 6,000 m are the least-explored

aquatic biosphere. The Challenger Deep, located in the western

equatorial Pacific, with a water depth of ∼11 km, is the deepest

ocean on Earth. Microbial communities associated with waters

from the sea surface to the trench bottom (0 ∼10,257 m) in the

Challenger Deep were analyzed, and unprecedented trench micro-

bial communities were identified in the hadal waters (6,000

∼10,257 m) that were distinct from the abyssal microbial commu-

nities. The potentially chemolithotrophic populations were less

abundant in the hadal water than those in the upper abyssal

waters. The emerging members of chemolithotrophic nitrifiers in

the hadal water that likely adapt to the higher flux of electron

donors were also different from those in the abyssal waters that

adapt to the lower flux of electron donors. Species-level niche sep-

aration in most of the dominant taxa was also found between the

hadal and abyssal microbial communities. Considering the geomor-

phology and the isolated hydrotopographical nature of the Mariana

Trench, we hypothesized that the distinct hadal microbial ecosystem

was driven by the endogenous recycling of organic matter in the

hadal waters associated with the trench geomorphology.

traits, a significantly higher microbial carbon turnover rate has been

identified in hadal sediments than that in adjacent abyssal plain

sediments (13). Furthermore, Jamieson et al. indicated that the

potential trench-specific distribution of eukaryotes is likely a result

of the hydrostatic pressure and hydrotopographical isolation (7).

The Challenger Deep in the Mariana Trench is the deepest

ocean on Earth, and its geology, current patterns, water chem-

istries, and benthic microbial communities have been relatively

well studied (8, 9, 13–17). However, the hadal aquatic microbial

communities in the Challenger Deep remain completely unchar-

acterized. This study aims to characterize the microbial commu-

nity compositions and functions of the hadal water microbial

ecosystem.

Results

Geochemical and Physical Environments. Water samples were taken

in a total of three dives of the remotely operated vehicle (ROV)

ABISMO at the same station (11–22.25′N, 142–42.75′E) on the

Challenger Deep. The temperature and salinity profiles were

| | | | hadal trench niche separation nitrification Challenger Deep

Significance

Microbial life in the dark ocean below mesopelagic water

(corresponding to 200- to 1,000-m depth range) is thought

to be primarily supported by the sinking organic carbon from

surface waters. However, it has recently been revealed that the

deep-sea biogeochemical cycles are more complex than pre-

viously expected and that the mismatch between the organic

carbon supply and microbial heterotrophic demand has led to

imbalances in some oceans (1–4). Currently, the contribution of

chemolithoautotrophy and mixotrophy to the biogeochemical

cycle (e.g., dark carbon fixation coupled with nitrification and

sulfur oxidations) is also recognized as another significant or-

ganic carbon source in the dark ocean (3, 5, 6). It has been es-

timated that the dissolved inorganic carbon fixation in the dark

ocean by these organisms could be on the same order-of-mag-

nitude as heterotrophic biomass production (3, 5).

Hadal oceans at water depths below 6,000 m are comprised

almost exclusively of trenches and are the least-explored aquatic

biosphere on Earth. Trench environments are differentiated from

upper abyssal oceans by their elevated hydrostatic pressure and

their hydrotopographically isolated nature, whereas other physical

and chemical conditions, such as temperature, salinity, dissolved

oxygen, and nutrients are comparable to those in abyssal oceans

(7–9). The microbiological and geochemical investigations of ha-

dal waters have been limited (10), in contrast to the long history

of hadal benthic microbiological studies occurring since the

1950s (11). To date, many piezophiles and piezotolerant bacteria

have been isolated from hadal benthic habitats, and their phe-

notypical and genomic features have been characterized to be

distinct from those of the close relatives obtained from shallow

marine environments (12). As one of the major biogeochemical

Although many microbial explorations for hadal sediments

began in the 1950s, the hadal water is the least-explored mi-

crobial biosphere. In this study, unexpected microbial ecosys-

tems associated with the hadal trench water were discovered

down to a 10,257-m water depth in the Challenger Deep of the

Mariana Trench, which is the deepest ocean on Earth. We

found the enrichment of heterotrophic population in the hadal

water (6,000 ∼10,257 m) microbial communities, whereas the

chemolithotrophic populations were more abundant in the up-

per abyssal waters. This observation suggested that the hadal

microbial biosphere was supported by the endogenous recycling

of organic matter in the hadal waters associated with the trench

geomorphology.

Author contributions: T.N. and K.T. designed research; T.N., M.H., A.M., O.K., T.K., J.M.,

K.K., and M.S. performed research; T.N., Y.T., S.S., A.M., and N.Y. analyzed data; and T.N.,

Y.T., A.M., and K.T. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.

Data deposition: The sequences reported in this paper have been deposited in the GenBank

database (accession nos. AB703684–AB703973 and AB703974–AB704001) and Short Read Ar-

chive database (accession no. DRA002518).

1To whom correspondence should be addressed. Email: takuron@jamstec.go.jp.

2Present address: Project Team for Development of New-generation Research Protocol for

Submarine Resources, Japan Agency for Marine-Earth Science & Technology, Yokosuka

237-0061, Japan.

3Present address: Earth-Life Science Institute, Tokyo Institute of Technology, Meguro-ku,

Tokyo 152-8550, Japan.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.

1073/pnas.1421816112/-/DCSupplemental.

E1230–E1236 | PNAS | Published online February 23, 2015

www.pnas.org/cgi/doi/10.1073/pnas.1421816112

similar to previous observations, and the salinity slightly increased

at depths greater than 9,000 m below the surface (mbs) (8, 9, 18).

The sea surface temperature was 29.1 °C; the potential tempera-

ture decreased to 1.02 °C at 10,257 mbs. The sea surface salinity

was 34.2, and the maximum (34.8) and minimum (34.4) salinities

were found at ∼150 and 400 mbs, respectively (Fig. 1A). The sa-

linity was constant at ∼34.6 below 1,500 mbs and slightly increased

below 9,000 mbs, as reported in previous studies (9, 18). The

dissolved oxygen concentration (DO) at the sea surface was

219 μM and decreased to 62.5 μM at 500 mbs (Fig. 1B and

Dataset S1). Then, the DO increased to 156 μM at 4,000 mbs and

presented relatively constant concentrations between 156 μM

and 172 μM below 4,000 mbs. The nitrate concentration at the

sea surface was 0.22 μM, and it drastically increased along the

thermo- and chemoclines (200–400 mbs) (Fig. 1B and Dataset

S1). The maximum nitrate concentration (41.2 μM) was detected

at 1,500 mbs, and constant nitrate concentrations (36.2–36.5 μM)

were found below 5,000 mbs. The nitrite concentrations were less

than 0.07 μM (detection limit <0.01 μM) throughout the water

column. The ammonium concentrations were under the de-

tection limit (detection limit <0.1 μM). The natural abundances

of 15N and 18O of nitrate were 4.7–7.4‰ and 2.1–4.5‰, re-

spectively (Fig. 1C and Dataset S1). The δ15N and δ18O values of

deep nitrate were similar to previously reported data in other

oceanic regions (19). The temperature and salinity profiles sug-

gest the presence of at least five types of water masses throughout

the depth profile: surface water (above 50 mbs), high salinity

water (∼150 mbs), low salinity water (∼400 mbs), mesopelagic

water (above 2,000 mbs), and deep water (below 2,000 mbs) (Fig.

1A). No significant change in the inorganic nutrients and isotopic

signatures of nitrate between the hadal (below 6,000 mbs) and

abyssal (below 4,000 m) waters was found, which was similar to

a previous study (18) (Fig. 1B).

Cellular and Viral Abundance. The maximum cell abundance was

found at 51 mbs (6.4 × 105 cells/mL) and the abundance de-

creased to a depth of 1,499 mbs (1.5 × 104 cells/mL) (Fig. 1D and

Dataset S1). Below 2,000 mbs the cell abundance was relatively

constant but gradually decreased with increasing depth, and the

cell abundance of the bottom waters was ∼6 × 103 cells/mL. The

abundance of virus-like particles (VLPs) increased from the sea

surface to 101 mbs (5.8 × 106 particles/mL) and then decreased

with increasing depth to 1,499 mbs (3.4 × 104 particles/mL). The

VLP abundance increased at 2,000 mbs (2.4 × 105 particles/mL),

and relatively constant VLP abundance was observed below

2,000 mbs (2.2–3.6 × 105 particles/mL) (Fig. 1D). The maximum

and minimum VLP-prokaryote ratio (VPR) was found at 147

and 1,499 mbs, respectively. The VPR below 2,000 mbs slightly

increased with increasing depth; however, there were a few

exceptions. The transition of microbial and viral abundance

profiles between 1,500 and 2,000 mbs is likely equivalent to the

water–mass interface.

SSU rRNA Gene Community Structures. Both tag sequencing and

clone analyses were applied to investigate the prokaryotic small

subunit (SSU) rRNA gene community structure along the water

column on the Challenger Deep (Fig. 2 and Dataset S2). Overall,

the niche separation at the species to phylum levels along the

water column was identified. The tag sequence populations were

grouped along the water column at the species level, as shown by

the Jaccard and Bray–Curtis similarity matrixes: the euphotic

zone (above 101 mbs), the mesopelagic to abyssopelagic waters

(202–5,000 mbs), and the trench waters (below 6,001 mbs); a

similar trend was also found at higher taxonomic levels (Fig. 3

and Figs. S1 and S2).

In the euphotic zone, obligately phototrophic Prochlorococcus

and potentially photoheterotrophic bacterial lineages, such as

SAR11 and Bacteroidetes (20, 21), dominated the prokaryotic

SSU rRNA gene communities. The high abundance of Actino-

bacteria only occurred above 147 mbs, which implies the occur-

rence of phototrophic metabolism (22). SAR11 also dominated

the SSU rRNA gene communities in aphotic waters above 2,000

mbs, and its abundance decreased drastically at 3,000 mbs. Below

147 mbs, thaumarchaeal phylotypes became the predominant pop-

ulations; its relative abundance decreased with increasing trench

water depth (Fig. 2). The abundance of SAR324, a potential

chemolithoautotrophic deltaproteobacterial subgroup (6, 23, 24),

also increased at 147 mbs and was detected as one of the

0 34

100

Salinity

34.5 35

NO3- (µM)

0 10 20 30 40 50

15N (‰)

35.5 7.5

8.5 4 5 6 7

Cell & VLP abundance

(cells or particles ml-1)

8 103 104 105 106 107

pH NO2-

NO3-

PO4

15N

cell

18O

VLP

VPR

1000

2000

4000

6000

8000

10000

Salinity

Temp

0 5 10 15 20 25 30 0 50 100 150 200 250 2

Potential temperature (˚C)

DO (µM)

NO2- , PO4 (µM)

18O (‰)

100

Virus prokaryote ratio

(VPR)

Fig. 1. Temperature and salinity (A); oxygen, nitrate, nitrite, and phosphate concentrations and pH (B); the δ15N and δ18O of nitrate (C); and the abundance

of prokaryotic cells and VLPs and VPR (D) along the water column in the Challenger Deep. The CTD profile of temperature and salinity was obtained in dive

AB#11. Other profiles were obtained by a total of three dives at the same location.

Nunoura et al.

PNAS | Published online February 23, 2015 | E1231

Relative abundance in tag sequences (%)

100

(24812)

(17443)

100

(24273)

(16285)

(17334)

(15254)

(22389)

(14740)

1000

2000

(17276)

(17509)

(8109)

4000

(11695)

6000

(12348)

(11546)

8000

(11679)

(24669)

(35710)

10000

(15679)

Cyanobacteria

chloroplast

SAR11 (Alphaproteobacteria)

Alphaproteobacteria (others)

SAR324 (Deltaproteobacteria)

Nitrospina

Deltaproteobacteria (others)

Gammaproteobacteria

Chloroflexi

Actinobacteria

Acidobacteria

Bacteroidetes

SAR406

Gemmatimonadetes

Planctomycetes

other bacteria

Thaumarchaeota

DHVEG 6

other archaea

Fig. 2. Prokaryotic SSU rRNA gene community composition along the water

column in the Challenger Deep. Numbers in parentheses indicate the num-

ber of tag sequences.

predominant populations above 5000 mbs. In contrast, the potential

heterotrophic SAR406 (Marine Group A) and Bacteroidetes

(10, 25) dominated the prokaryotic SSU rRNA gene communities

below 6,000 mbs, whereas they were found as only a minor pop-

ulation above 5,000 mbs. Intriguingly, the predominance of the

tag sequences closely related to Halomonas sp. and Pseudomonas

stutzeri occurred at the bottom of the trench waters (9,000 and

10,241 mbs) (Fig. 2 and Dataset S2).

Among the potential nitrifier populations in the SSU rRNA

gene analyses, we found niche separation for both nitrite and

ammonia oxidizers along the water column. A relatively high

abundance of the nitrite-oxidizing Nitrospina was detected in

waters between 147 and 2,000 mbs, whereas the abundance of

Nitrospira increased in the trench waters (Fig. 2 and Dataset S2).

Among the potential ammonia oxidizers, the niche separation of

thaumarchaeal subgroups was found along the water column,

and the Nitrosomonas population was detected only in the trench

waters. Subgroup β of the marine group I (MGI) thaumarch-

aeote dominated the thaumarchaeal population at the bottom of

the photic zone (147–300 mbs), the δ subgroup was detected

above 2,000 mbs, the α subgroup dominated the thaumarchaeal

population in the trench waters below 6,000 mbs, and the γ

subgroup was detected throughout the water column below 147 mbs

(Figs. S3 and S4A).

Detection and Quantification of Nitrifiers. To clarify the niche

separation of ammonia oxidizers suggested by the SSU rRNA

gene-sequencing analyses, a clone sequencing analysis was con-

ducted for the archaeal ammonia monooxygenase α subunit

(amoA) gene, and subsequent quantitative PCR analyses were

applied for other bacterial nitrifiers to characterize the niche

separation of nitrifiers in further detail. For the preparation

of sufficient template DNA, amplified environmental DNA

was used for the analyses. In fact, the composition of envi-

ronmental DNA must be biased during the genome amplifi-

cation; however, the technique has advantages in environmental

molecular studies using insufficient amounts of DNA (26, 27).

The archaeal amoA gene clone analysis suggested the distinctive

distribution of four major amoA subgroups (A, Ba, Bb, and D)

along the water column (Figs. S4B and S5). The niche separa-

tion of ammonia-oxidizing archaea (AOA) above the abyssal

waters that suggested by the amoA gene clone analysis resem-

bles the previously observed pattern that showed the distinctive

distribution of the “high ammonia cluster” (HAC) and “low

ammonia cluster” (LAC) of archaeal amoA along the water

column (28) (Fig. S5). The HAC was comprised of the groups A,

C, and D, and the LAC was identical to groups Ba and Bb in

this study.

Based on the amoA sequences obtained in this study, group-

specific primers for quantification were constructed for each

dominant amoA group to clarify the niche separation of the

AOA along the water column at this site. The distribution and

abundance patterns obtained in the quantitative PCR were more

similar to the MGI SSU rRNA gene communities than that in

the archaeal amoA gene clone analysis (Fig. 4 and Fig. S4).

Group A and Bb of amoA dominated in the photic zone above

200 mbs and in the mesopelagic waters above 1,000 mbs, re-

spectively; however, amoA group Ba was found below 300 mbs

and predominated between the mesopelagic and bathypelagic

zone (300–6,000 mbs). Group D predominantly occurred in the

hadal waters. The sum of the archaeal amoA copy number below

147 mbs correlated with archaeal SSU rRNA gene copy number

below 147 mbs in the ratio of 0.9 (R = 0.91). The archaeal SSU

rRNA gene communities below 147 mbs are dominated by the

MGI thaumarchaeotes (Fig. 2), and the genomes of the MGI

thaumarchaeotes in the publically accessible database are known

to encode only one copy of the amoA and SSU rRNA genes.

Thus, most of the MGI thaumarchaeotes distributed in the water

column would harbor amoA, as suggested by Sintes et al. (28).

These results also suggest the coordination between the thau-

marchaeal SSU rRNA and amoA gene clusters: group α in the

SSU rRNA gene and D in amoA, β and A, γ and Ba, and δ and

Ba, respectively.

In addition, conventional and quantitative PCR were exam-

ined for genes from gamma- and betaproteobacterial ammonia

oxidizers and nitrite-oxidizing Nitrospira, Nitrospina, and alpha-

and gammaproteobacteria. The gammaproteobacterial ammonia

oxidizers and alpha- and gammaproteobacterial nitrite oxidizers

(e.g., Nitrobacter and Nitrococcus, respectively) were not detected

throughout the water column. The betaproteobacterial amoA

genes were revealed to be abundant in the amoA composition of

the photic zone and the trench bottom waters, but they were

absent in most depths of the mesopelagic to abyssal waters

(between 400 and 5,000 mbs). Among the nitrite oxidizers, niche

separation between Nitrospina and Nitrospira was also revealed.

The Nitrospira SSU rRNA gene population outcompeted the

Nitrospina SSU rRNA gene population in waters at 50 and

100 mbs, as well as in the hadal waters. The Nitrospina SSU

rRNA gene abundance was higher than that of the Nitrospira

E1232 | www.pnas.org/cgi/doi/10.1073/pnas.1421816112

Nunoura et al.

Jaccard

50 m

100 m

147 m

200 m

300 m

406 m

500 m

1000 m

1499 m

2000 m

3000 m

5000 m

6001 m

7000 m

7998 m

9000 m

10257 m

Distance

B Bray-Currtis

Distance

C Unifrac

0 0.2 0.4 0.6 0.8 1.0

50 m

100 m

147 m

200 m

300 m

406 m

500 m

1000 m

1499 m

2000 m

3000 m

5000 m

6001 m

7000 m

7998 m

9000 m

10257 m

Value

0 0.2 0.4 0.6 0.8 1.0

Photic

Mesopelagic to

abyssal

Hadal

Photic

Mesopelagic to

abyssal

Hadal

Photic

Mesopelagic to

abyssal

Fig. 3. Similarity matrixes of Jaccard (A), Bray–Curtis (B), and UniFrac (C) of the SSU rRNA gene-tag sequences.

Hadal

population below the bottom of the euphotic zone to the bathyal

waters (150–5,000 mbs) (Fig. 4). Gammaproteobacterial nxrA

was not detected in this study.

Dilution Counting of Heterotrophic Bacteria. The culturable het-

erotrophic microbial population was estimated on board by using

serial dilution cultivation of heterotrophic bacteria at 4 °C; the

partial SSU rRNA gene sequences of the strains obtained from

the most diluted inocula with microbial growth were determined.

Surprisingly, a quite high culturable population was found (at > 2 ×

103 cells/mL) in the trench bottom waters from 9,003 and 10,243 m,

whereas the microbial cell abundances at these depths were only 5.5

and 6.3 × 103 cells/mL, respectively (Dataset S1). The predominant

culturable heterotrophic microbes at these depths were very similar

to P. stutzeri (99% similarity), which was consistent with the envi-

ronmental SSU rRNA gene analyses at these depths (Fig. 2 and

Dataset S2). Conversely, the culturable population in the upper

waters (0–1,500 mbs) ranged from > 0–2 × 103 cells/mL and the

population in deep waters (2,000–8,000 mbs) ranged from > 0–2 ×

10 cells/mL.

Discussion

Niche Separation of Nitrifiers. The composition of nitrifiers could be

indexes of carbon and nitrogen cycles in oceanic ecosystem because

NH4+ is provided by nitrogenous organic matter decomposition. In

addition, it has been noted that the niche separation of nitrifiers is

regulated by the amount of available electron donors such as am-

monia and nitrite (28–30). Thus, the niche separation of nitrifiers

could be a signature of NH4+ flux from organic matter de-

composition that cannot be identified from NH4+ concentration

measurements. In fact, the SSU rRNA gene-sequencing analyses

and quantitative PCR analyses of nitrifiers suggested the possible

niche separation along with the transition of the entire microbial

community structure and NH4+ flux.

Sunlight, pH, and salinity are additional significant factors that

affect the niche separation of ammonia oxidizers besides the

NH4+ concentration (31–33). The salinity change is less than 1 in

the water column on the Challenger Deep and is negligible in

this case. The higher abundance of ammonia-oxidizing bacteria

(AOB) than AOA in the upper euphotic zone (above 50 mbs) is

likely consistent with the high sensitivity of AOA to photo-

inhibition than the sensitivity of AOB shown by the previous

cultivation experiments (31). The predominance of group A

amoA (possibly group β MGI) in the bottom of the euphotic

zone may be associated with the relatively higher pH zone along

the water column (Figs. 1A and 4). The co-occurrence of AOB

and archaeal amoA group D HAC) in hadal water and the

predominance of archaeal amoA groups Ba and Bb (LAC) in

abyssal water are consistent with the previous observation that

AOA prefers lower ammonia concentrations than AOB (30).

The distinctive distribution of potential nitrite oxidizers, such

as the Nitrospina and Nitrospira SSU rRNA genes, was also

revealed. The Nitrospira SSU rRNA gene population overtook

the Nitrospina SSU rRNA gene population in waters at 50 and

100 mbs, as well as the hadal waters. The Nitrospina SSU rRNA

gene population overcame that of the Nitrospira population

below the bottom of the euphotic zone to the bathyal waters

(150–5,000 mbs) (Fig. 4). This distribution pattern is generally

consistent with the abundance of the Nitrospina and Nitrospira

populations in the SSU rRNA gene-tag sequences. The significance

of Nitrospina in the bottom of the euphotic zone to mesopelagic

water is in accordance with the previous reports for (sub)tropical

oceans (34–36). In contrast, the significance of Nitrospira in oceanic

waters has not yet been revealed, although their contribution in

nitrification was reported in seafloor environments (37–39). These

observations could be interpreted as a result of kinetic-dependent

niche separation, but several uncertainties cannot be ruled out (e.g.,

technical difficulties in the detection of nitrite oxidizers) (39). In

general, Nitrospira and Nitrospina have typically been found in eu-

trophic and oligotrophic marine environments, respectively. Thus,

Nitrospira likely adapts to a higher nitrite flux than Nitrospina.

Factors Controlling the Hadal Biosphere. The distribution and abun-

dance patterns of the entire microbial and nitrifiers communities

in the hadal waters were found to be distinct from those in abyssal

waters. In the hadal water, the potential chemolithoautotrophs

decreased in relative abundance with increasing depth and were

likely replaced by the heterotrophic populations. In addition, most

likely responding to the potentially elevated ammonia supply by

the heterotrophic activity with increasing depth, the dominant

groups of both ammonia and nitrite oxidizers were different.

These results suggest that the formation of a unique hadal bio-

sphere in the Challenger Deep may be driven by the input of

organic matter and the following heterotrophic degradation;

however, the distribution and abundance patterns of microbial

communities cannot be explained by the vertical flux of sinking

organic particles.

Nunoura et al.

PNAS | Published online February 23, 2015 | E1233

100

B Ammonia

oxidizers

Nitrite oxidizers

100

1000

2000

4000

6000

8000

10000

10 102 103 104 105 106

Gene abundance (copies/ml)

All prokaryotic SSU rRNA gene

Archaeal SSU rRNA gene

amoA from archaeal and bacterial ammonia oxidizers

SSU rRNA gene from nitrite oxidizers

Group A

Group Ba

Group Bb

Group D

Betaproteobacteria

Nitrospina

Nitrospira

Fig. 4. The abundance of whole prokaryotic, archaeal and nitrite oxidizer SSU rRNA and amoA genes (A), and the abundance of subgroups of amoA genes

and SSU rRNA genes of nitrite-oxidizing bacteria (B) along the water column in the Challenger Deep obtained by quantitative PCR. Grouping of amoA are

shown in Fig. S5.

The Challenger Deep is located in the oligotrophic ocean

region and is geographically and hydrotopographically isolated

from other trenches in the Western Pacific. Thus, endogenous

organic carbon sources are required to support the hadal water

heterotrophic microbial communities. Recent studies have pointed

to the importance of suspended organic matter, including both

sinking and suspended organic matter in bathypelagic waters

(40, 41). In addition, Kawagucci et al. reported the impacts of

suspended sediment associated with huge earthquakes that

affected the microbial communities in the bathyal and abyssal

water ecosystems (42). The microbial heterotrophic popula-

tions influenced by the suspended organic matter were similar

to those found in the Challenger Deep hadal biosphere that

were enriched by heterotrophs. A steep slope, narrow geo-

morphology, slow trench current (8, 9), and earthquakes may

supply a steady state or the occasional input of sinking and

suspended organic matter. Because suspended particulate matters

are transported vertically as well as horizontally (42, 43), the

suspended organic matter from slopes likely influences the

geochemical cycle in the entire trench waters. The higher

sediment deposition rate in the trench bottom compared with

the adjacent abyssal plain (13), as well as the clear stratification

of trench bottom sediment under the sampling site (Fig. S6),

also suggest the occasional input of sediment supply from the

trench slope. Moreover, the slightly elevated salinity of the

bottom water could induce the density-driven stratification of

hadal water mass (9). The stratification of hadal water mass may

promote the development of isolated and unique biogeochemical

cycles and microbial communities. This study hypothesizes that

the unique microbial ecosystem in one of the deepest accessible

biospheres on this planet is primarily driven by the geomorphology

of the Mariana Trench.

Experimental Procedures

Site Description and Sampling. A total of three dives for the ROV ABISMO was

conducted in the Challenger Deep of the Mariana Trench (11°22.25′N,

142°42.75′E, 10,300 m) during the Japan Agency for Marine-Earth Science &

Technology (JAMSTEC) R/V Kairei KR08-05 cruise (June 2008) (44) (Dataset

S1). Temperature, depth, and salinity were measured using a conductivity,

temperature, and depth (CTD) sensor SBE49 (Sea-Bird Electronics). Waters

from near the bottom of the trench (water depth of approx. 10,300 m) to

the surface were taken using Niskin bottles (5 L) (General Oceanic) that were

settled on the ROV ABISMO. Sea-surface water was collected using a bucket. The

water samples used in this study are summarized in Dataset S1, and the sub-

samples for geochemical and microbiological analyses were taken from the same

bottles. Samples for cell counting were fixed by formaldehyde [final concen-

tration 3% (vol/vol)], and filtered on a polycarbonate membrane filter (0.2 μm).

The filters were stored at −80 °C. The water samples for counting virus-like

particles were filtered with cellulose nitrate membrane filter (0.2 μm), fixed by

formaldehyde [final concentration 3% (vol/vol)], frozen in liquid nitrogen, and

then stored at −80 °C. Microbial cells in each 2–3 L water for molecular ecological

analyses were collected on cellulose nitrate membrane filters, and stored at

−80 °C. Trench bottom sediment was obtained by a gravity corer of the ROV

ABISMO in the same dive. A sediment core was split onboard and stored at 5 °C.

E1234 | www.pnas.org/cgi/doi/10.1073/pnas.1421816112

Nunoura et al.

Geochemical Analyses. DO was measured using an oxygen CHEMets kit

(CHEMeytrics) onboard. Samples for nutrient analysis were filtrated with

0.2-μm cellulose acetate filter and stored at −20 °C until further analysis.

Concentrations of NO3−, NO2−, PO4−, and NH4+ were analyzed spectropho-

tometrically using an automated QuAAtro 2-HR analyzer (BL TEC).

Nitrogen and oxygen isotopic compositions of the nitrate were determined

using the denitrifier method (45–47). Nitrate was converted to N2O by the

strain Pseudomonas chlororaphis (JCM20509 = ATCC13985), a denitrifying

bacterium lacking the capability of N2O reduction (46). Then, the produced

nitrous oxide was extracted, purified, and measured for nitrogen and oxygen

isotopic ratios using a 20–22 continuous flow isotope ratio mass spectrometer

(Sercon) at Tokyo University of Agriculture and Technology. The amount of

nitrous oxide introduced to CF-IRMS ranged from 3 to 15 nmol. International

isotopic reference materials, USGS32 (δ15N = 180‰, δ18O = 25.7‰), USGS34

(δ15N = −1.8‰, δ18O = -27.9‰), USGS35 (δ18O = 57.5‰) and IAEA-NO3 (δ15N =

4.7‰, δ18O = 25.6‰) (47–49), were used for the calibration. The δ15N values

were reported relative to atmospheric N2, and the δ18O values were reported

relative to standard mean ocean water (V-SMOW). The analytical precision of

in-house material was typically less than 0.2‰ for δ15N and 0.3‰ for δ18O.

DNA Extraction and Amplification. Environmental DNA was extracted from

the cells on cellulose nitrate membrane filters using a Soil DNA Isolation Kit

(Mo-Bio Lab) with minor modifications. A portion of environmental DNA was

amplified using a REPLI-g Mini Kit (Qiagen) for the molecular analyses de-

scribed below. Amplified DNA assemblages were digested by S1 nuclease

(Invitrogen) before the following studies.

Diversity Analyses for SSU rRNA and Archaeal amoA Genes. Prokaryotic SSU

rRNA gene fragments were amplified with a primer set of 530F and 907R (50)

from the original environmental DNA assemblages using LA Taq polymerase

with GC buffer (Takara Bio) as previously described (50). For tag sequencing,

primers with 10 bp of extended tag sequences in 5′-termini were used for the

SSU rRNA gene amplification. For archaeal amoA clone analysis, gene frag-

ments were obtained using EX Taq polymerase (Takara Bio) from the ampli-

fied environmental DNA assemblages. The amplification conditions and primer

sequences for each of the PCR amplifications are summarized in Table S1.

For the clone analyses, amplified DNA fragments were cloned into pCR2.1

vector (Invitrogen), and clone libraries were constructed. The inserts were

directly sequenced with the M13M4 primer using an ABI3730xl genetic an-

alyzer with Big Dye v3.1. SSU rRNA gene sequences with >97% identity were

assigned as the same clone type (phylotype) based on FastGroup II (51) and

similarity analysis in GENETYX-MAC v15 (GENETYX). SSU rRNA gene ampli-

cons for tag sequencing were analyzed by 454 FLX Titanium sequencer

(Roche). All of the raw tag sequences were treated with shhh.flows pipeline

in MOTHUR 1.31.1 (52–54), and the primer sequences in either or both ends

of the tags were eliminated. Tag sequences shorter than 300 bp were re-

moved from the downstream analyses. Potential chimera sequences were

surveyed using UCHIME (55). Next, phylogenetic assignment and statistical

analyses for the tag sequences were conducted. Sequencing tags were

aligned using the partial order algorithm (SINA; www.arb-silva.de/aligner/)

with the reference multiple alignment SILVA SSU Ref NR (56). All of the

aligned sequences were then clustered into operational taxonomic units

(OTUs) by 97% sequence identity using MOTHUR 1.31.1, with default

parameters according to the average-clustering algorithm. Output was then

parsed to produce occurrence tables of each OTU in each sample. The tax-

onomic position of each OTU was automatically assigned based on Blast

analysis in the QIIME software package (57) using SILVA Ref NR as a refer-

ence dataset of SSU rRNA gene sequences. The sequences were excluded

that are closely related to the potential contaminants belonging to genera

that inhabit soil and the human body and have been detected from negative

control experiments of environmental microbiology in the laboratory,

such as Bradyrhizobium, Brevundimonas, Burkholderiaceae, Delftia, Eryth-

robacter, Lactococcus, Legionella, Methylobacterium, Mycobacterium, Neis-

seria, Novosphingobium, Propionibacterium, Sphingobium, Sphingomonas,

Sphingopyxis, Staphylococcus, Stenotrophomonas, and Streptococcus. Se-

quences presenting with relatively high E-values (>1.0E-30) or low identity

(<90%) to the best match sequence were designated as other archaea or

bacteria, and sequences that did not present significant similarity to any ref-

erence sequences were also excluded from the analysis. α-Diversity indices

(rarefaction curves, Chao1, ACE, Shannon, Shannon evenness, and Simpson) in

each library and taxa/divisions were also calculated using MOTHUR 3.6. The

Jaccard and Bray–Curtis dissimilarity indices among each library were esti-

mated using the vegan package in the R-environment (vegan.r-forge.r-project.

org). Phylogenetic trees were constructed from the curated alignment of

representative sequences using the Clustal W program (www.clustal.org).

A weighted UniFrac distance matrix among core samples was constructed

from the phylogenetic tree and a sample mapping files that showed fre-

quency of the sequence tags within OTUs. Sequencing tags for each phy-

lum/class were collected from the entire dataset in accordance with the

taxonomic position of each OTU. The subdatasets were analyzed as in the

case of the entire dataset.

All SSU rRNA gene sequences obtained in this study were compared using

the UniFrac program (58) after omitting potential chimera sequences and

potential experimental contamination sequences. The alignment of each SSU

rRNA gene clone library was constructed using the SINA alignment service

(www.arb-silva.de/aligner/) (59). The phylogenetic tree of all SSU rRNA gene

sequence obtained in this study was constructed by Clustal X (www.clustal.

org), and the principal component analysis was carried out using the tree by

UniFrac. Representative SSU rRNA gene sequences were aligned and phylo-

genetically classified into certain taxonomic units using ARB (56). The phy-

logenetic tree of thaumarchaeal SSU rRNA genes was constructed by Clustal X

based on the unambiguous residues. Representative amoA sequences were

automatically aligned with closely related nucleotide sequences, and the

phylogenetic tree was then constructed using Clustal X v2.0 (60).

Quantitative PCR Analyses. Primers, probes and components of standard

mixture used for the quantitative PCR analyses are summarized in Table S1.

The abundance of each gene was quantified as an average of the duplicate or

triplicate analyses. Original DNA assemblages were only used for the quan-

tification of archaeal and prokaryotic SSU rRNA genes, and the amplified DNA

assemblages were used for the other genes. The abundance of nitrifier genes

in each water mass was estimated from the relative abundance of archaeal

SSU rRNA gene and a respective gene in the amplified DNA assemblages.

The 7500 Real Time PCR System (Applied Biosystems) was used for all of

the quantitative PCR analyses in this study. Quantification of the archaeal

and all prokaryotic SSU rRNA genes was performed using both the original

and amplified environmental DNA (Table S1). Detection and quantification

of nitrifiers were assessed using the amplified environmental DNA assemb-

lages. Detection of the alpha- and betaproteobacterial nxrA and alphapro-

teobacterial amoA was examined using Ex Taq polymerase (Takara Bio) with

a Mg2+ buffer as described previously (39) (Table S1). The abundance of

Nitrospina and Nitrospira SSU rRNA genes was also examined according to

methods described previously (39, 61). To identify group specific distribution

of each archaeal amoA group, novel primer sets were constructed based on

archaeal amoA gene sequences obtained in this study as follows: The nu-

cleotide alignments of the archaeal amoA gene were constructed by Clustal

X v2.0, and primers were designed that individually matched with the spe-

cific amoA sequences in groups A, Ba, Bb, and D (Table S1).

For the preparation of quantitative PCR mixtures, qPCR Quick GoldStar

Mastermix Plus (Eurogentec) was applied for SSU rRNA genes of archaea, all

prokaryotes and Nitrospira, and a SYBR Premix Ex Taq II (Takara Bio) for

amoA genes and the Nitrospina SSU rRNA gene. Amplified products from

quantitative PCR using SYBR Premix reagent were confirmed by agarose gel

electrophoresis. Amplification specificity was confirmed by clone analysis for

amplicons from several depths, particularly in the cases of amoA genes.

Dilution Counting for Heterotrophs. The abundance of culturable hetero-

trophs at each depth was evaluated onboard by serial dilution cultivation

methods with 1×, 1/100×, and 1/10,000× Marine Broth (Difco) using 96-well

microtiter plates at 5 °C for 2 mo under atmospheric pressure. For the di-

lution of Marine Broth, MJ synthetic seawater (62) was used. Then, SSU rRNA

gene sequences from the cultures obtained from the most diluted wells

were amplified with a primer set of B27F and U1492R (Table S1), and directly

sequenced using the ABI3730xl genetic analyzer with Big Dye v3.1 (Applied

Biosystems) according to the manufacturer’s recommendations.

ACKNOWLEDGMENTS. We thank the captain, crew, and science party of the

R/V Kairei (Japan Agency for Marine-Earth Science & Technology) during the

KR08-05 cruise; and the development and operational teams of the remotely

operated vehicle ABISMO. T.N. was supported in part by a Grant-in-Aid for

Scientific Research (B) (24370015) from the Japan Society for the Promotion

of Science.

1. Guidi L, et al. (2009) Effects of phytoplankton community on production, size and export

of large aggregates: A world-ocean analysis. Limnol Oceanogr 54(6):1951–1963.

2. Bochdansky AB, van Aken HM, Herndl GJ (2010) Role of macroscopic particles in deep-

sea oxygen consumption. Proc Natl Acad Sci USA 107(18):8287–8291.

Nunoura et al.

PNAS | Published online February 23, 2015 | E1235

3. Herndl GJ, Reinthaler T (2013) Microbial control of the dark end of the biological

pump. Nat Geosci 6(9):718–724.

4. Yokokawa T, Yang Y, Motegi C, Nagata T (2013) Large-scale geographical variation in

prokaryotic abundance and production in meso-and bathypelagic zones of the cen-

tral Pacific and Southern Ocean. Limnol Oceanogr 58(1):61–73.

5. Alonso-Sáez L, Galand PE, Casamayor EO, Pedrós-Alió C, Bertilsson S (2010) High bi-

carbonate assimilation in the dark by Arctic bacteria. ISME J 4(12):1581–1590.

6. Swan BK, et al. (2011) Potential for chemolithoautotrophy among ubiquitous bacteria

lineages in the dark ocean. Science 333(6047):1296–1300.

7. Jamieson AJ, Fujii T, Mayor DJ, Solan M, Priede IG (2010) Hadal trenches: The ecology

of the deepest places on Earth. Trends Ecol Evol 25(3):190–197.

8. Taira K, Kitagawa S, Yamashiro T, Yanagimoto D (2004) Deep and bottom currents in

the challenger deep, measured with super-deep current meters. J Oceanogr 60(6):

919–926.

9. Taira K, Yanagimoto D, Kitagawa S (2005) Deep CTD casts in the Challenger Deep,

Mariana Trench. J Oceanogr 61(3):447–454.

10. Nagata T, et al. (2010) Emerging concepts on microbial processes in the bathypelagic

ocean—Ecology, biogeochemistry, and genomics. Deep Sea Res Part II Top Stud

Oceanogr 57(16):1519–1536.

11. Bartlett DH (2009) Microbial life in the trenches. MTS J 43(5):128–131.

12. Tamburini C, Boutrif M, Garel M, Colwell RR, Deming JW (2013) Prokaryotic responses

to hydrostatic pressure in the ocean—A review. Environ Microbiol 15(5):1262–1274.

13. Glud RN, et al. (2013) High rates of microbial carbon turnover in sediments in the

deepest oceanic trench on Earth. Nat Geosci 6(4):284–288.

14. Takami H, Inoue A, Fuji F, Horikoshi K (1997) Microbial flora in the deepest sea mud of

the Mariana Trench. FEMS Microbiol Lett 152(2):279–285.

15. Kato C, et al. (1998) Extremely barophilic bacteria isolated from the Mariana

Trench, Challenger Deep, at a depth of 11,000 meters. Appl Environ Microbiol

64(4):1510–1513.

16. Fujioka K, Okino K, Kanamatsu T, Ohara Y (2002) Morphology and origin of the

Challenger Deep in the Southern Mariana Trench. Geophy Res Let 29(10):10-1–10-4.

17. Yoshida M, Takaki Y, Eitoku M, Nunoura T, Takai K (2013) Metagenomic analysis of

viral communities in (hado)pelagic sediments. PLoS ONE 8(2):e57271.

18. Mantyla AW, Reid JL (1978) Measurements of water characteristics at depths greater

than 10 km in the Marianas Trench. Deep-Sea Res 25(2):169–173.

19. Sigman DM, et al. (2009) The dual isotopes of deep nitrate as a constraint on the cycle

and budget of oceanic fixed nitrogen. Deep Sea Res Part I Oceanogr Res Pap 56(9):

1419–1439.

20. Giovannoni SJ, et al. (2005) Proteorhodopsin in the ubiquitous marine bacterium

SAR11. Nature 438(7064):82–85.

21. Gómez-Consarnau L, et al. (2007) Light stimulates growth of proteorhodopsin-con-

taining marine Flavobacteria. Nature 445(7124):210–213.

22. Ghai R, Mizuno CM, Picazo A, Camacho A, Rodriguez-Valera F (2013) Metagenomics un-

covers a new group of low GC and ultra-small marine Actinobacteria. Sci Rep 3:2471.

23. Chitsaz H, et al. (2011) Efficient de novo assembly of single-cell bacterial genomes

from short-read data sets. Nat Biotechnol 29(10):915–921.

24. Sheik CS, Jain S, Dick GJ (2014) Metabolic flexibility of enigmatic SAR324 revealed

through metagenomics and metatranscriptomics. Environ Microbiol 16(1):304–317.

25. Wright JJ, et al. (2014) Genomic properties of Marine Group A bacteria indicate a role

in the marine sulfur cycle. ISME J 8(2):455–468.

26. Chen Y, et al. (2008) Revealing the uncultivated majority: Combining DNA stable-

isotope probing, multiple displacement amplification and metagenomic analyses of

uncultivated Methylocystis in acidic peatlands. Environ Microbiol 10(10):2609–2622.

27. Bodelier PLE, Kamst M, Meima-Franke M, Stralis-Pavese N, Bodrossy L (2009) Whole-

community genome amplification (WCGA) leads to compositional bias in methane-

oxidizing communities as assessed by pmoA-based microarray analyses and QPCR.

Environ Microbiol Rep 1(5):434–441.

28. Sintes E, Bergauer K, De Corte D, Yokokawa T, Herndl GJ (2013) Archaeal amoA gene

diversity points to distinct biogeography of ammonia-oxidizing Crenarchaeota in the

ocean. Environ Microbiol 15(5):1647–1658.

29. Blackburne R, Vadivelu VM, Yuan Z, Keller J (2007) Kinetic characterisation of

an enriched Nitrospira culture with comparison to Nitrobacter. Water Res 41(14):

3033–3042.

30. Martens-Habbena W, Berube PM, Urakawa H, de la Torre JR, Stahl DA (2009) Am-

monia oxidation kinetics determine niche separation of nitrifying Archaea and Bac-

teria. Nature 461(7266):976–979.

31. Merbt SN, et al. (2012) Differential photoinhibition of bacterial and archaeal am-

monia oxidation. FEMS Microbiol Lett 327(1):41–46.

32. Gubry-Rangin C, et al. (2011) Niche specialization of terrestrial archaeal ammonia

oxidizers. Proc Natl Acad Sci USA 108(52):21206–21211.

33. Mosier AC, Francis CA (2008) Relative abundance and diversity of ammonia-oxidizing

archaea and bacteria in the San Francisco Bay estuary. Environ Microbiol 10(11):

3002–3016.

34. Mincer TJ, et al. (2007) Quantitative distribution of presumptive archaeal and bac-

terial nitrifiers in Monterey Bay and the North Pacific Subtropical Gyre. Environ Mi-

crobiol 9(5):1162–1175.

35. Santoro AE, Casciotti KL, Francis CA (2010) Activity, abundance and diversity of ni-

trifying archaea and bacteria in the central California Current. Environ Microbiol

12(7):1989–2006.

36. Füssel J, et al. (2012) Nitrite oxidation in the Namibian oxygen minimum zone. ISME J

6(6):1200–1209.

37. Hoffmann F, et al. (2009) Complex nitrogen cycling in the sponge Geodia barretti.

Environ Microbiol 11(9):2228–2243.

38. Off S, Alawi M, Spieck E (2010) Enrichment and physiological characterization of a

novel Nitrospira-like bacterium obtained from a marine sponge. Appl Environ Mi-

crobiol 76(14):4640–4646.

39. Nunoura T, et al. (2013) Molecular biological and isotopic biogeochemical prognoses

of the nitrification-driven dynamic microbial nitrogen cycle in hadopelagic sediments.

Environ Microbiol 15(11):3087–3107.

40. Baltar F, et al. (2009) Evidence of prokaryotic metabolism on suspended particulate

organic matter in the dark waters of the subtropical North Atlantic. Limnol Oceanogr

54(1):182–193.

41. Aristegui J, Gasol JM, Duarte CM, Herndl GJ (2009) Microbial oceanography of the

dark ocean’s pelagic realm. Limnol Oceanogr 54(5):1501–1529.

42. Kawagucci S, et al. (2012) Disturbance of deep-sea environments induced by the M9.0

Tohoku Earthquake. Sci Rep 2:270.

43. Otosaka S, Noriki S (2000) REEs and Mn/Al ratio of settling particles: Horizontal

transport of particulate material in the northern Japan Trench. Mar Chem 72(2-4):

329–342.

44. Yoshida H, et al. (2009) The ABISMO mud and water sampling ROV for surveys at

11,000 m depth. MTS J 43(5):87–96.

45. Sigman DM, et al. (2001) A bacterial method for the nitrogen isotopic analysis of

nitrate in seawater and freshwater. Anal Chem 73(17):4145–4153.

46. Casciotti KL, Sigman DM, Hastings MG, Böhlke JK, Hilkert A (2002) Measurement of

the oxygen isotopic composition of nitrate in seawater and freshwater using the

denitrifier method. Anal Chem 74(19):4905–4912.

47. McIlvin MR, Casciotti KL (2011) Technical updates to the bacterial method for nitrate

isotopic analyses. Anal Chem 83(5):1850–1856.

48. Böhlke JK, Coplen TB (1995) Reference and Intercomparison Materials for Stable

Isotopes of Light Elements. IAEA TECDOC 825 (International Atomic Energy Agency,

Vienna), pp 51–66.

49. Böhlke JK, Mroczkowski SJ, Coplen TB (2003) Oxygen isotopes in nitrate: New ref-

erence materials for 18O:17O:16O measurements and observations on nitrate-water

equilibration. Rapid Commun Mass Spectrom 17(16):1835–1846.

50. Nunoura T, et al. (2012) Microbial diversity in deep-sea methane seep sediments

presented by SSU rRNA gene tag sequencing. Microbes Environ 27(4):382–390.

51. Yu Y, Breitbart M, McNairnie P, Rohwer F (2006) FastGroupII: A web-based bioinformatics

platform for analyses of large 16S rDNA libraries. BMC Bioinformatics 7:57.

52. Quince C, et al. (2009) Accurate determination of microbial diversity from 454 py-

rosequencing data. Nat Methods 6(9):639–641.

53. Schloss PD, et al. (2009) Introducing MOTHUR: Open-source, platform-independent,

community-supported software for describing and comparing microbial communities.

Appl Environ Microbiol 75(23):7537–7541.

54. Schloss PD, Gevers D, Westcott SL (2011) Reducing the effects of PCR amplification

and sequencing artifacts on 16S rRNA-based studies. PLoS ONE 6(12):e27310.

55. Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R (2011) UCHIME improves sensi-

tivity and speed of chimera detection. Bioinformatics 27(16):2194–2200.

56. Yilmaz P, et al. (2014) The SILVA and “All-species Living Tree Project (LTP)” taxonomic

frameworks. Nucleic Acids Res 42(Database issue, D1):D643–D648.

57. Caporaso JG, et al. (2010) QIIME allows analysis of high-throughput community se-

quencing data. Nat Methods 7(5):335–336.

58. Lozupone C, Hamady M, Knight R (2006) UniFrac—An online tool for comparing microbial

community diversity in a phylogenetic context. BMC Bioinformatics 7(1):371.

59. Pruesse E, Peplies J, Glöckner FO (2012) SINA: Accurate high-throughput multiple

sequence alignment of ribosomal RNA genes. Bioinformatics 28(14):1823–1829.

60. Larkin MA, et al. (2007) Clustal W and Clustal X version 2.0. Bioinformatics 23(21):

2947–2948.

61. Graham DW, et al. (2007) Experimental demonstration of chaotic instability in bi-

ological nitrification. ISME J 1(5):385–393.

62. Sako Y, Takai K, Ishida Y, Uchida A, Katayama Y (1996) Rhodothermus obamensis sp.

nov., a modern lineage of extremely thermophilic marine bacteria. Int J Syst Bacteriol

46(4):1099–1104.

E1236 | www.pnas.org/cgi/doi/10.1073/pnas.1421816112

Nunoura et al.