1

Supplementary Information Appendix 1

Nitrifier adaptation to low energy flux controls inventory of reduced 2

nitrogen in the dark ocean 3

4

Yao Zhanga,b,1

, Wei Qinc, Lei Hou

a,b, Emily J. Zakem

d, Xianhui Wan

a, Zihao Zhao

e, Li 5

Liua,b

, Kristopher A. Huntf, Nianzhi Jiao

a,b, Shuh-Ji Kao

a,b, Kai Tang

a,b, Xiabing Xie

a, 6

Jiaming Shena,b

, Yufang Lia,b

, Mingming Chena,b

, Xiaofeng Daia,b

, Chang Liua,b

, 7

Wenchao Denga,b

, Minhan Daia,b

, Anitra E. Ingallsc, David A. Stahl

f, and Gerhard J. 8

Herndle,g

9

aState Key Laboratory of Marine Environmental Sciences, Xiamen University, 10

Xiamen 361101, China; bCollege of Ocean and Earth Sciences, Xiamen University, 11

Xiamen 361101, China; cSchool of Oceanography, University of Washington, Seattle, 12

WA 98195, USA; dDepartment of Biological Sciences, University of Southern 13

California, Los Angeles, CA 90089, USA; eDepartment of Limnology and Bio-14

Oceanography, Center of Functional Ecology, University of Vienna, Althanstrasse 14, 15

A-1090 Vienna, Austria; fDepartment of Civil and Environmental Engineering, 16

University of Washington, Seattle, WA 98195, USA; gNIOZ, Department of Marine 17

Microbiology and Biogeochemistry, Royal Netherlands Institute for Sea Research, 18

Utrecht University, 1790 AB Den Burg, The Netherlands. 19

1To whom correspondence should be addressed. Email: yaozhang@xmu.edu.cn 20

www.pnas.org/cgi/doi/10.1073/pnas.1912367117

2

This PDF file includes: 21

Materials and Methods 22

Supplementary Text 23

Figs. S1 to S10 24

Tables S1 to S5 25

References for SI Appendix citations 26

3

Materials and Methods 27

Study sites and sampling 28

A 492 km oceanographic transect with six sites (S1–S6) across the continental shelf 29

and slope and the central basin of the South China Sea (SCS) was sampled for 30

molecular analysis of the prokaryotic community during a research cruise on board of 31

RV Shi Yan 1 in September 2014. Three basin sites (S6, S7, and S8) were sampled for 32

nitrification rate measurements in May 2016 on board of RV Dong Fang Hong 2. Site 33

S6 was sampled again in June 2017 for nitrification kinetics on board of RV Tan Kah 34

Kee. In addition, two sites (W1 and W2) were sampled in the Western Pacific Ocean 35

(WP) during a spring and summer research cruise on board of RV Dong Fang Hong 2 36

and Ke Xue in April and August 2015, respectively. Water from ten sites (Fig. S1) was 37

collected along a vertical profile with two to 16 depth layers (see Fig. S2 for depth 38

layers of each site) using a conductivity-temperature-depth (CTD)-rosette sampling 39

system with Go-Flo bottles mounted in the rosette (SBE 9/17 plus; SeaBird Inc, 40

USA). 41

A total of 54 samples was collected for gene analyses while, 34 samples were 42

collected for transcript analyses at sites S6, W1, and W2. Three samples from 5, 200, 43

and 3000 m depth from site S6 were subjected to metatranscriptomics analyses. In 44

addition, 33 samples from sites S6, S7, S8, and W2 were amended with 15

N-labeled 45

ammonium and 15

N-labeled nitrite to measure the corresponding oxidation rates. One 46

sample from 150 m of site W2 and five samples collected between 75 and 200 m at 47

site S6 were used for ammonia (hereafter defined as combined ammonia and 48

4

ammonium) oxidation and nitrite oxidation kinetic experiments, respectively. 49

50

DNA and RNA extraction 51

Two to 4 L water samples were filtered through 0.22 m pore-size polycarbonate 52

membranes (47 mm diameter; Millipore) for DNA and RNA extraction. Samples for 53

RNA analyses were filtered within 30 min and stored instantly in 2 mL RNase-free 54

tubes containing 1 mL RNAlater RNA stabilizer (Ambion). All membranes were 55

flash-frozen in liquid nitrogen and then transferred to −80°C until further analysis. 56

DNA samples from the SCS were extracted using the UltraClean Soil DNA kit 57

(MoBio, San Diego, CA, USA) following the manufacturer’s protocols. DNA samples 58

from the WP were extracted using the phenol-chloroform-isoamyl alcohol method as 59

described by Massana et al. (1). Concentrations and purity of the genomic DNA were 60

determined using a NanoDrop spectrophotometer (Thermo Scientific 2000/2000c). 61

RNA was extracted with TRIzol reagent (Invitrogen, Carlsbad, CA) according to 62

Simms et al. (2) with minor modifications. DNA was digested with Turbo DNase 63

(Life Technologies). DNA contamination was checked by amplifying the bacterial 64

16S rRNA genes with the universal primers 27F and 1492R. Total RNA without DNA 65

contamination was reverse transcribed to synthesize single-strand complementary 66

DNA (cDNA) using the SuperScript RT-PCR system with random primers 67

(Invitrogen, Carlsbad, CA, USA). 68

69

Quantitative PCR amplification 70

5

Abundances of the archaeal and β-proteobacterial amoA genes, thaumarchaeal (MGI) 71

16S rRNA genes, Nitrospira and Nitrospina 16S rRNA genes, and archaeal accA 72

genes were quantified using the qPCR method and a CFX 96™ real-time system 73

(BIO-RAD, Singapore). Standard curves were constructed for archaeal and β-74

proteobacterial amoA genes and the archaeal accA gene using plasmid DNA from 75

clone libraries. For Thaumarchaeota, Nitrospira, and Nitrospina 16S rRNA genes, the 76

target DNA fragments of the pure-culture strains were used. qPCR reactions were 77

performed in triplicate and analyzed against a range of standards (100 to 10

7 copies 78

per l). Primer pair sequences, qPCR mixtures and conditions for each gene are listed 79

in Table S3. The efficiencies of qPCR amplification ranged from 90% to 104% with 80

R2 >0.99. The specificity of the qPCR reactions was checked by melting curve 81

analysis and agarose gel electrophoresis analysis. Ambiguous products were 82

sequenced to confirm their veracity. Inhibition tests were performed by 2-fold and 5-83

fold dilutions of all samples. Based on these tests, we concluded that our samples 84

were not inhibited. 85

86

Metatranscriptomics analysis 87

About 100 L samples were collected from the surface layer (5 m), the bottom of the 88

euphotic zone (200 m), and the bathypelagic zone (3000 m) of site S6 using in situ 89

pump filtration (McLane Large Volume Water Transfer System Sampler, USA) 90

through a 100 m Nitex screen, and 20 m and 0.22 m pore-size polycarbonate 91

filters (142 mm diameter; Millipore) within 30 min. The 0.22 m pore-size filters 92

6

were preserved in RNAlater (Ambion), flash-frozen in liquid nitrogen and transferred 93

to −80°C until further analysis. Total RNA was extracted with TRIzol reagent 94

(Invitrogen) (2). The residual DNA was removed using a Turbo DNA-free kit 95

(Ambion) and the RNA was purified using a RNeasy MinElute Cleanup kit (Qiagen). 96

mRNA was enriched by removing the bacterial and archaeal 16S and 23S ribosomal 97

RNA (rRNA) transcripts in the total RNA with both MICROB Express™ (Ambion) 98

and Ribozero (Epicentre) kits and by removing small RNA and tRNA using the 99

MEGAclearTM

kit (Ambion). Subsequently, mRNA was converted into cDNA using 100

the SuperScript III First Strand Synthesis System (Invitrogen) with random hexamer 101

primers. The cDNA was treated with RiboShredderTM

RNase Blend (Epicentre) to 102

remove trace RNA contaminants. To improve the yield of cDNA, the single-stranded 103

cDNA was amplified once using the GenomiPHITM

V2 method (GE Healthcare) (3) 104

and the amplified samples were then treated with S1 nuclease (4). Three ‒ 11 g 105

cDNA was purified using the QIAquick PCR purification kit (Qiagen) and sequenced 106

(one sequencing library per sample) on an Illumina HiSeq 2000 at the Beijing 107

Genomic Institute, Shenzhen, China following the manufacturer’s instructions. 108

From the raw sequences, the reads with >10% ambiguous bases or 20% low 109

quality bases (<Q20) were removed or those contaminated by the adapter. rRNA 110

transcripts were identified (>90% identity) using the SOAP software by blasting 111

against the Silva database and then removed. Subsequently, de novo assembly of the 112

clean reads was performed using Trinity with default settings (--jaccard_clip; --113

SS_lib_type RF; --min_kmer_cov 1; --min_glue 2; --group_pairs_distance 500; --114

7

path_reinforcement_distance 75) and only transcripts of 200 bp or longer were 115

reported (5). The produced unigenes were taken for further processing, such as 116

sequence splicing and redundancy removal, with sequence clustering software 117

(Illumina) to acquire non-redundant unigenes as long as possible. Subsequently, the 118

unigenes were blasted against public databases including NCBI non-redundant protein 119

(NR), Swiss-Prot, KEGG, COG, and GO (blastx, E-value <10−5

). Summary statistics 120

are given in Table S4. The protein with the highest sequence similarity was retrieved 121

and annotated to each unigene. For annotated unigenes, protein coding sequence 122

(CDS) information was retrieved from the blast results. For unannotated unigenes, 123

ESTscan (6) was used to perform CDS prediction. The predicted CDSs were then 124

blasted (blastp, E-value <10−5

) against databases, including KEGG, eggNOG, CAZy, 125

and ARDB, to obtain further functional information. To assess the quality of 126

assembly, reads of each sample were mapped back to the merged unigenes using 127

Bowtie2 version 2.2.5 with the setting of one mismatch in the seed alignment (7). 128

129

Ammonia and nitrite oxidation rate measurements 130

Incubations to determine ammonia and nitrite oxidation rates with water collected at 131

the S6‒8 and W2 stations were conducted on deck using the 15

N-labelling technique. 132

Briefly, for determining ammonia oxidation rates, 1 mL of 15

N-NH4+ tracer (98% of 133

15N atom, Sigma-Aldrich) was injected into 250 mL samples in Nalgene HDPE 134

bottles to obtain a final tracer concentration of 20 nmol L−1

at W2 and 30 nmol L−1

at 135

S6‒8, and 1 mL of 14

N-NO2− was added as carrier with a final concentration of 1 136

8

mol L−1

to avoid underestimating ammonia oxidation rates due to the oxidation of 137

15N-NO2

− to

15N-NO3

− (8). Immediately after tracer and NO2

− injection, 40 mL of 138

sample was filtered through a 0.2 m syringe filter for determining 15

N-labeling 139

percentages of initial substrate pools. The residual water was kept in the dark at near 140

in-situ (±1°C) temperature for 12 h and terminated by filtering through a 0.2 m 141

syringe filter. Samples were stored at −20°C immediately after collection. Similarly, 142

nitrite oxidation rates were obtained by using 15

N-NO2− tracer (98% of

15N atom, 143

Sigma-Aldrich) with a final tracer concentration of 20 nmol L−1

at W2 and 30 nmol 144

L−1

at S6‒8. All incubations were carried out in the dark and at in-situ (±1°C) 145

temperature. Each of the incubations was performed in triplicate. 146

δ15

N of NO2− was determined using the azide reduction method (9). In brief, a 147

fresh 1:1 (v:v) mixture of 2 mol L−1

sodium azide and 20% acetic acid was prepared 148

and purged with helium gas for 30 min to remove any pre-existing N2O, and then 0.8 149

mL of the mixture was added to each sample. NO2− was quantitatively converted into 150

N2O. Five NO2− stable isotope standards were made by combining different quantities 151

of 98% Na15

NO2 (Sigma-Aldrich) and 99% Na14

NO2 (Merck). The δ15

N of NaNO2 152

standards was calibrated against NO3− isotope standards USGS 34, IAEA N3, and 153

USGS 32 using the bacterial method. δ15

N of NO3− was determined using the 154

bacterial method (10). Briefly, samples were first treated with sulfamic acid for 12 h 155

at room temperature (22‒26°C) in the dark to remove NO2−

(11), and then neutralized 156

with NaOH and stored at −20°C before isotope analysis. NO3− was quantitatively 157

converted to N2O using the bacterial strain Pseudomonas aureofaciens (ATTC no. 158

9

13985). Three NO3− international reference materials (USGS 34, IAEA N3, and 159

USGS 32) were used to calibrate the δ15

N of NO3−. The N2O was then introduced to a 160

gas chromatography isotope ratio mass spectrometry (GC-IRMS, Thermo Delta V 161

Advantage) coupled with an on-line N2O cryogenic extraction and purification 162

system. Accuracy (pooled standard deviation) was better than ±0.2‰ for bacterial 163

method and ±0.4‰ for azide reduction method according to the analyses of these 164

standards at an injection concentration of 20 nmol N. 165

Ammonia oxidation and nitrite oxidation rates were primarily determined by the 166

accumulation of 15

N in the product pool relative to the initial. We used Equation (1) to 167

quantify the transformation rate of bulk substrate (8). 168

𝑅 =𝐶×(𝑛𝑡−𝑛0)

𝑡×𝑓15 (1) 169

where R represents the bulk reaction rates for all substrates after tracer enrichment 170

(nmol N L−1

h−1

), C denotes the product concentrations (nmol N L−1

), f15

is the 171

fraction of 15

N of the substrate pool at the beginning of incubation, nt and n0 are at% 172

15N of the product pool at the end and beginning of the incubation (%), respectively, t 173

is the incubation time (h). The detection limits for ammonia oxidation rates and nitrite 174

oxidation rates were 0.001–0.08 nmol N L−1

d−1

and 0.0003–0.085 N L−1

d−1

, 175

respectively, which were estimated by taking three times the pooled standard 176

deviation as a reliable enrichment of 15

N in the product pool. 177

178

Kinetics experiments 179

Ammonia oxidation kinetics experiments were performed with water collected at 150 180

10

m depth at site W2. The dependence of NH4+ oxidation rate on substrate concentration 181

was investigated using five different concentrations (0.03, 0.048, 0.096, 0.4, and 2 182

M) of 15

N-NH4+ (98% of

15N atom, Sigma-Aldrich).

14NO2

− at a final concentration 183

of 0.5 M was added as a carrier. For each set, the tracer was added separately into 184

duplicate 250 mL Nalgene HDPE bottles. Immediately after tracer injection, around 185

40 mL of sample was filtered through a 0.2 m syringe filter to represent the initial 186

condition. The remaining samples were incubated in a thermostat incubator at in-situ 187

(±1°C) temperature in the dark for 12 h and terminated by filtration. δ15

N of NO2− was 188

measured as described above. 189

Nitrite oxidation kinetics experiments were performed with water collected at 75, 190

95, 110, 150, and 200 m depths at site S6. 15

N-NO2− tracer (98% of

15N atom, Sigma-191

Aldrich) was injected into five Nalgene HDPE bottles with 250 mL samples to reach a 192

final tracer concentration of 0.03, 0.1, 0.2, 0.5, and 2 M. For samples from 75 m, 0.5 193

mL of 14

N-NO3− was additionally added as carrier to obtain a final concentration of 1 194

M. Each of the incubations was performed in duplicate. After tracer addition, 40 mL 195

sample was immediately filtered through a 0.2 m syringe filter to represent initial 196

condition (t0). The remaining samples were kept in a thermostat incubator at in-situ 197

(±1°C) temperature in the dark for time-series incubations (26 (t1) and 52 (t2) hours), 198

and at each time interval, incubation was terminated by filtration. δ15

N of NO3− was 199

measured as described above. Ammonia and nitrite oxidation rates were calculated 200

from the slopes of linear regression with 15

N-production as a function of time. 201

The kinetics constants (Vmax and Ks) were estimated using Equation (2): 202

11

𝑉 =𝑉max×[𝑆]

𝐾s+[𝑆] (2) 203

Where Vmax is the potential maximum rate of ammonia or nitrite oxidation and Ks 204

denotes the half-saturation constant. S represents the substrate concentration. 205

206

Laboratory experiments with specific bacterial strains 207

Nitrosopumilus maritimus SCM1 (at 24°C) Nitrospira moscoviensis NSP M-1 (at 208

26°C), and Nitrospina gracilis 3/211 (at 26°C) were cultured according to previous 209

reports (12−14) with minor modifications. Additionally, we also performed 210

incubation experiments with Nitrococcus mobilis Nb-231 (at 26°C) inhabiting mainly 211

oxygen-deficient zones, with initial cell abundances of ~4.0 × 105 cell mL

−1. 212

213

Estimating global oceanic dark DIC fixation 214

The depth profiles of ammonia and nitrite oxidization are consistent with a rapid 215

decrease in sinking organic matter because ammonia is only supplied by 216

mineralization in the meso- and bathypelagic zones of the ocean (15). Thus, we fitted 217

the depth profiles below the euphotic zone (100 m in the SCS and 200 m in the WP) 218

of ammonia and nitrite oxidation rates into the Martin curve using a Power Law 219

Equation (3) (16). 220

𝐹 = 𝐹100 (𝑍

100)𝑏

(3) 221

Integrated ammonia and nitrite oxidation rates were calculated and then integrated 222

DIC fixation rates were calculated according to the stoichiometric relations (Type I 223

regression) between inorganic nitrogen oxidization and DIC fixation rates obtained in 224

12

the laboratory experiments (see above). Extrapolating this integrated value to the 225

entire volume of the ocean below the euphotic zone based on a mean ocean depth of 226

3700 m and area of 3.6 × 1014

m2 (17) gives an estimate of global oceanic dark DIC 227

fixation. 228

229

Statistical analysis 230

Since normality in the distribution of the individual data sets was not always met, we 231

used the non-parametric Wilcoxon tests for comparing two related variables. 232

Polynomial and logarithm models (Sigmaplot) were used to determine the 233

relationships between variables. 234

235

Supplementary Text 236

Text S1. Abundance patterns of marine AOA and NOB 237

Archaeal and β-proteobacterial genes encoding the subunit A of the key enzyme for 238

ammonia oxidation, ammonia monooxygenase (amoA), as well as the 16S rRNA 239

genes of Thaumarchaeota, Nitrospira and Nitrospina (two major NOB groups in the 240

marine environment) were quantified via qPCR. As observed in many other studies, 241

the archaeal amoA was orders of magnitude more abundant than the β-proteobacterial 242

amoA gene (Table S5) with the highest abundance near the bottom of the euphotic 243

zone (Fig. S2 A‒C). Depth profiles of thaumarchaeotal 16S rRNA gene abundance 244

were generally similar to that of archaeal amoA genes and the two gene abundances 245

were positively correlated (R = 0.85‒0.91, P <0.01) with an average amoA : 16S 246

13

rRNA gene ratio of 1.6±0.4 (n = 48), suggesting all Thaumarchaeota are capable of 247

ammonia oxidation, with limited exceptions (18, 19). 248

Higher gene abundances of Thaumarchaeota 16S rRNA than archaeal amoA 249

mainly occurred in the WP, suggesting that some Thaumarchaeota may not be 250

capable of ammonia oxidation (19). Higher gene abundances of archaeal amoA than 251

Thaumarchaeota 16S rRNA occurred in the shelf waters and above 200 m of the slope 252

of the SCS. Possibly, there were other AOA groups (e.g. pSL12, also coined Marine 253

Benthic Group A) in the area close to the estuary. Hu et al. (20) have detected pSL12 254

16S rRNA genes in the SCS by qPCR. Similar discrepancies between 255

Thaumarchaeota 16S and archaeal amoA gene abundances were also observed in the 256

Pacific such as at Monterey Bay and the North Pacific Subtropical Gyre and the 257

pSL12 gene copy numbers agreed well with the difference between archaeal amoA 258

and thaumarchaeotal 16S rRNA gene abundance (21). 259

We also determined the abundance of the acetyl-coenzyme A carboxylase gene 260

(accA), involved in thaumarchaeotal carbon fixation. The ratio of archaeal accA : 261

amoA gene abundance increased along the SCS transect from the shelf to the basin, as 262

well as with depth (Fig. S2 D). A positive correlation was observed between archaeal 263

accA and amoA gene abundances in mesopelagic waters below 100 m with a slope of 264

1.03 (Fig. S2 E), suggesting that AOA biomass production is primarily 265

chemoautotrophic (22). 266

Nitrospina and Nitrospira 16S rRNA gene abundances varied from 2.71 × 102 to 267

4.93 × 106 copies L

−1 and 22 to 7.60 × 10

4 copies L

−1 in the SCS water column, 268

14

respectively, as well as from 62 to 3.77 × 105 copies L

−1 and below detection limit to 269

5.25 × 104 copies L

−1 in the WP, respectively (Fig. S8). Thus, AOA was one to two 270

orders of magnitude more abundant than NOB (Wilcoxon, P <0.01) (Fig. S2 A‒C). 271

Nevertheless, AOA and NOB abundances were correlated, as indicated by the 272

distribution pattern of archaeal amoA and NOB 16S rRNA gene abundances (R = 273

0.83‒0.89, P <0.01). 274

275

Text S2. Gene copy number per genome 276

All genes targeted in the present study are typically present as single copy per genome 277

(21, 23‒25). Only the β-proteobacterial amo operon was found in multiple (2‒3) 278

nearly identical copies in the representative strains (26). We downloaded a total of 23 279

Thaumarchaeota and five NOB (four Nitrospira and one Nitrospina) publically 280

available complete genome sequences from the NCBI database. Alignment results 281

also indicated that the Thaumarchaeota amoA, accA, and 16S rRNA genes and the 282

Nitrospina and Nitrospira 16S rRNA genes are present as a single copy in the 283

respective genome. We also noted that, different from our alignment results, Mincer et 284

al. (21) suggested two copies of the rRNA operon per genome in Nitrospina. Further 285

studies on Nitrospina, which has rarely been investigated, are required to resolve this. 286

Nevertheless, two copies per genome would not affect our conclusion that there is a 287

major difference in abundance between the low abundance of NOB and the one to two 288

orders of magnitude more abundant AOA. 289

290

15

Text S3. Coverage of the qPCR primers targeting NOB 291

The primers Nspra675f and Nspra746r were used in our qPCR approach for targeting 292

NOB Nitrospira 16S rRNA genes. Alignments of the primer sequences within the 293

SILVA rRNA database using TestPrime indicated a coverage of 92.6%. The primers 294

NitSSU_130F and NitSSU_282R were used for detecting the NOB Nitrospina 16S 295

rRNA gene. Among 29 OTU sequences of Nitrospina 16S rRNA genes available in 296

the databases listed by Levipan et al. (27), only one sequence could not be targeted by 297

the primers NitSSU_130F and NitSSU_282R due to >3 mismatching bases for either 298

primer, indicating a ~97% coverage of the primers. Alignments of the sequences of 299

the primer NitSSU_130F and NitSSU_282R within the SILVA rRNA database using 300

TestPrime indicated a 91.4% coverage. 301

302

Text S4. Activities measured on representative AOA and NOB strains 303

Physiological investigations were performed on Nitrosopumilus maritimus SCM1, 304

Nitrospira moscoviensis NSP M-1, Nitrospina gracilis 3/211, and Nitrococcus mobilis 305

Nb-231. SCM1 is the first isolated marine AOA strain (12) and has been widely used 306

as a model organism for developing an understanding of marine AOA physiology and 307

biochemistry (28). NSP M-1 is widespread in a diverse range of habitats. Strain 3/211 308

is the only isolated NOB Nitrospina strain from the oxygenated ocean (29). Nb-231 is 309

the only isolated NOB Nitrococcus strain, which originates from the surface waters of 310

Pacific Ocean collected from a foamy surface slick (29). The strains were cultivated 311

in batch cultures with a similar initial cell abundance. The abundance of SCM1 (107 312

16

cells mL−1

) was two times higher than that of NSP M-1 and one order of magnitude 313

higher than those of strains 3/211 and Nb-231 (106 cells mL

−1) in the stationary phase 314

(Fig. S3 A, E, I, and M). In contrast, ammonia oxidation rates of SCM1 were one to 315

two orders of magnitude lower than nitrite oxidation rates of NSP M-1, 3/211, and 316

Nb-231 in the exponential phase. The concomitantly measured DIC fixation rates of 317

SCM1 on a volume base were similar to strains NSP M-1 and 3/211 and 318

approximately three times lower than that of Nb-231 (Fig. S3 B, F, J, and N). Thus, 319

the cell-specific nitrite oxidation rates of NSP M-1, 3/211, and Nb-231 were 2-fold, 320

one and two orders of magnitude higher, respectively, than cell-specific ammonia 321

oxidation rates of SCM1. Cell-specific DIC fixation rates of SCM1 were similar to 322

those of NSP M-1 and 3/211, and one order of magnitude lower than those of Nb-231 323

(Fig. S3 C, G, K, and O). Taken together, the measured activities confirmed the 324

expected higher DIC fixation efficiency (per-N oxidized) of SCM1 than all NOB 325

strains (Fig. S3 D, H, L, and P). Nitrococcus mobilis Nb-231 had an extremely low 326

efficiency, which was not included in Fig. 1 since it abundantly occurs in the oxygen-327

deficient zones. In addition, we also performed physiological investigations on 328

Nitrosopumilus maritimus SCM1 at 30°C and Nitrospira moscoviensis NSP M-1 at 329

35°C. The results (Fig. S9) were generally consistent with the findings shown in Fig. 330

S3 while the time reaching the stationary phase was reduced greatly. 331

332

Text S5. Thermodynamic analysis 333

Ammonia and nitrite oxidizers have different carbon fixation yields and therefore 334

17

different biomass production rates, which is hypothesized to be due to the available 335

free energy of their redox reactions. Convoluting matters further, the production and 336

consumption of the shared metabolite nitrite creates an interdependence of relative 337

free energy on the relative activity of the two metabolisms. Two scenarios were 338

modeled to constrain this interdependence that represent 1) surface water, where the 339

ammonia concentration is high and nitrate concentration low, and 2) deep water where 340

the ammonia concentration is low and nitrate concentration high. Using the 341

commonly accepted overall metabolism for ammonia and nitrate oxidizers, nitrite 342

concentration was varied to quantify the response of relative free energy availability 343

between the two metabolisms on the basis of relative joule per mol of nitrogen reacted 344

(Fig. S4). As nitrite concentration increased, the energy available from ammonia 345

oxidation per energy available from nitrite oxidation decreased, shifting from 3.7 to 346

3.5 J/J for high ammonia and 4.5 to 4.2 J/J for high nitrate at the minimum and 347

maximum observed nitrite concentrations, respectively (Fig. S4). These relative free 348

energies provide an estimate for the relative carbon fixation efficiencies of the two 349

chemolithoautotrophs since they assume the machinery needed for energy 350

conservation has been tuned to the relevant concentrations. 351

352

Text S6. Relationship of transcriptional levels to measured activities 353

Since NOB are one to two orders of magnitude lower in abundance than AOA, they 354

exhibit one to two orders of magnitude higher cell-specific oxidation rates than AOA 355

throughout the water column except for the surface waters (Wilcoxon, P <0.05) (Table 356

18

1). This is consistent with activity differences between NOB and AOA inferred from 357

the 16S rRNA:gene ratio as an activity proxy (30, 31). The ratio of 16S rRNA to 16S 358

rRNA gene abundance was significantly higher (Wilcoxon, P <0.05) for the low 359

abundance NOB than for the more abundant ammonia-oxidizing Thaumarchaeota, but 360

their 16S rRNA:gene ratios were positively related (R = 0.90, P <0.01) (Fig. S10) 361

supporting a coupling of relative activities of the two dominant oceanic nitrifying 362

groups. Higher relative activity of the NOB was also supported by metatranscriptomic 363

analysis of samples from S6, showing high representation of transcripts of genes 364

encoding the key enzyme for nitrite oxidation, nitrite oxidoreductase (comprising 365

nxrA and nxrB subunits) at 200 m and 3000 m depths. Notably, nxrAB transcript 366

abundance was higher than that for archaeal amo genes (comprising amoA, amoB, and 367

amoC subunits) at 200 m depth (Fig. 2). Transcripts for genes associated with carbon 368

fixation by the two groups were also compared. Transcripts encoding enzymes 369

involved in the rTCA cycle of NOB (Nitrospina and Nitrospira) were more abundant 370

than transcripts for the thaumarchaeotal HP/HB cycle at 200 m depth, and expressed 371

at comparable levels at 3000 m depth (Fig. 2 and Table S2). Therefore, although NOB 372

are present at lower abundance than AOA, they are highly active throughout the 373

mesopelagic and bathypelagic zones and even exhibit higher relative transcript 374

abundances than AOA in the upper mesopelagic ocean. 375

19

376

377

Fig. S1. Site locations and bathymetry. Gene analysis was performed on samples 378

from all sites except for sites S7 and S8. The purple asterisks indicate the location 379

where samples were collected for metatranscriptomes/transcript analyses. The yellow 380

solid squares indicate locations where ammonia and nitrite oxidation rates were 381

measured, and underlines indicate stations where ammonia and nitrite oxidation 382

kinetics were determined. This figure was produced using Ocean Data View (32). 383

CHINA

Pearl R.

Lu

zo

n

JAPANL

ati

tud

e

Longitude

20

384

385

Fig. S2. Distribution of ammonia-oxidizing archaea (AOA) and nitrite-oxidizing 386

bacteria (NOB) based on gene abundance. (A) Depth profiles of the archaeal amoA 387

gene and Thaumarchaeota (error bars representing standard deviations of three 388

technical replicates are not visible because they are smaller than the symbols) and 389

NOB (sum of Nitrospira and Nitrospina, no error bars) 16S rRNA gene abundances in 390

the South China Sea (SCS) shelf (sites S1‒3), (B) slope and basin (sites S4‒6), and 391

(C) the western Pacific Ocean (WP) (sites W1 and 2). The black lines indicate the 392

depths of the sea bottom. The dashed lines indicate the depth of the euphotic zone 393

Ocean

Data

Vie

w /

D

Ratio of archaeal accA:amoA genes

114°E 114.5°E

Longitude

115°E 116°E115.5°E

0

1000

2000

3000

4000

Depth

(m

)

S1 S2 S3 S4 S5 S6

6

5

4

3

2

1

0

DIV

A

Genes (copies L 1

)

101102103104105 106 107 108

Genes (copies L 1

)

101102103 104 105 106 107 108

3000

2000

1000

800

500

200

100

75

5

101

Genes (copies L 1

)

101 103 104 105 106 107 108

De

pth

(m

)

90

75

50

25

5

3000

2000

1000

800

750

650

500

450

400

350

300

250

200

150

SCS Shelf

SCSSlope & Basin WP

A B C

100

101 10

3

AOA amoA

S1 S2 S3 S4 S5 S6

NOB 16S rRNA

Thaumarchaeota 16S rRNA

W1 W2

5

100

103

S1

S2

S3

_ _ _

Genes (copies L 1

)

101102103104105 106 107 108

Genes (copies L 1

)

101102103 104 105 106 107 108

3000

2000

1000

800

500

200

100

75

5

101

Genes (copies L 1

)

101 103 104 105 106 107 108

De

pth

(m

)

90

75

50

25

5

3000

2000

1000

800

750

650

500

450

400

350

300

250

200

150

SCS Shelf

SCSSlope & Basin WP

A B C

100

101 10

3

AOA amoA

S1 S2 S3 S4 S5 S6

NOB 16S rRNA

Thaumarchaeota 16S rRNA

W1 W2

5

100

103

S1

S2

S3

_ _ _

AOA amoA genes

(107 copies L 1)

0.0 0.6 1.2 1.8 3.04.0

Arc

haeal accA

genes (

10

7 c

opie

s L

1)

0.0

0.4

0.8

1.2E

Slope = 1.03R = 0.66P < 0.05

<100 m≥100 m

_

_

21

(0.1% of PAR). (D) Ratio of archaeal accA : amoA gene abundance along the transect 394

with six sites across the SCS continental shelf (S1–S3), slope (S4–S5), and the central 395

basin (S6). The filled grey area is bathymetry. There is a statistically significant 396

correlation between (E) archaeal amoA and accA gene abundance in the water column 397

below 100 m depth. The error bars for gene abundance represent standard deviations 398

of three technical replicates. 399

22

400

0 3 6 9 12 15 18 21 24 27 30 33N-n

utr

ient concentr

ation

(m

M)

0.0

0.4

0.8

1.2

Abundance (

10

7 c

ells

mL

1)

0

1

2

3

4

5

6

7

0 3 6 9 12 15 18 21 24 27 30 33

DIC

fix

ation r

ate

(μM

C d

1)

0

3

6

9

12

15

18

Am

moniu

m o

xid

ation

rate

(10

2 μ

M N

d 1

)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0 3 6 9 12 15 18 21 24 27 30 33

DIC

fix

ation r

ate

per

cell

(fm

ol C

cell

1 d

1)

0

1

2

3

Am

moniu

m o

xid

ation

rate

per

cell

(10

2 fm

ol N

cell

1 d

1)

0.0

0.1

0.2

0.3

0 3 6 9 12 15 18 21 24 27 30 33

Carb

on y

ield

(10

-2 m

ol C

fix

ed p

er

mol N

oxid

ized)

0

5

10

15

20

Ratio o

f 16S

rR

NA

:rD

NA

0

10

20

30

40

50

NH

4＋

NO

2N

H4＋

+ N

O2－

0 2 4 6 8 10 12 14 16 18

N-n

utr

ient concentr

ation

(m

M)

0.0

1.1

2.2

3.3

Abundance (

10

7 c

ells

mL

1)

0

1

2

3

A B C D

E

Time (days)

0 1 2 3 4 5 6 7 8 9 10N-n

utr

ient concentr

ation

(m

M)

0

5

10

15

20

25

Abundance (

10

6 c

ells

mL

1)

0

1

2

3

4

NO

2

NO

3

NO

2 +

NO

3

M

Time (days)

0 1 2 3 4 5 6 7 8 9 10

DIC

fix

ation r

ate

(μM

C d

1)

0

10

20

30

40

50

Nitrite

oxid

ation r

ate

(10

2 μ

M N

d 1

)

0

10

20

30

40

50N

Time (days)

0 1 2 3 4 5 6 7 8 9 10

Ratio o

f 16S

rR

NA

:rD

NA

0

100

200

300

400

Carb

on y

ield

(10

-2 m

ol C

fix

ed p

er

mol N

oxid

ized)

0

1

2

3

4P

Time (days)

0 1 2 3 4 5 6 7 8 9 10

DIC

fix

ation r

ate

per

cell

(fm

ol C

cell

1 d

1)

0

10

20

30

40

50 N

itrite

oxid

ation r

ate

per

cell

(10

2 fm

ol N

cell

1 d

1)

0

10

20

30

40

50O

NOB Nitrococcus mobilis Nb-231 (26 C)

NOB Nitrospira moscoviensis NSP M-1 (26 C)

NOB Nitrospina gracilis 3/211 (26 C)

AOA Nitrosopumilus maritimus SCM1 (24 C)

NO

2

NO

3

NO

2 +

NO

3

0 2 4 6 8 10 12 14 16 18

DIC

fix

ation r

ate

(μM

C d

1)

0

3

6

9

12

15

18

Nitrite

oxid

ation r

ate

(10

2μM

N d

1)

0

1

2

3

4

5

6

7

0 2 4 6 8 10 12 14 16 18

DIC

fix

ation r

ate

per

cell

(fm

ol C

cell

1 d

1)

0

1

2

3

Nitrite

oxid

ation r

ate

per

cell

(10

2 fm

ol N

cell

1 d

1)

0.0

0.2

0.4

0.6

0 2 4 6 8 10 12 14 16 18

Carb

on y

ield

(10

-2 m

ol C

fix

ed p

er

mol N

oxid

ized)

0

1

2

3

4

Ratio o

f 16S

rR

NA

:rD

NA

0

100

200

300

400F G H

0 1 2 3 4 5 6 7 8 9 101112

N-n

utr

ient concentr

ation

(m

M)

0

1

2

3

4

Abundance (

10

6 c

ells

mL

1)

0

3

6

9I

NO

2

NO

3

NO

2 +

NO

3

0 1 2 3 4 5 6 7 8 9 101112

DIC

fix

ation r

ate

(μM

C d

1)

0

3

6

9

12

15

18

Nitrite

oxid

ation r

ate

(10

2μM

N d

1)

0

1

2

3

4

5

6

7

0 1 2 3 4 5 6 7 8 9 101112

Nitrite

oxid

ation r

ate

per

cell

(10

2 fm

ol N

cell

1 d

1)

0

1

2

3

DIC

fix

ation r

ate

per

cell

(fm

ol C

cell

1 d

1)

0

1

2

3

0 1 2 3 4 5 6 7 8 9 101112

Carb

on y

ield

(10

-2 m

ol C

fix

ed p

er

mol N

oxid

ized)

0

1

2

3

4J K L

°

°

°

°

_

_

__

_ _

_

_

_

_

_

__

__

__

__

__

_ _

_

__

_

_ _

__

__

__

__

__

__

23

Fig. S3. Physiological characteristics of the ammonia-oxidizing archaeal and 401

nitrite-oxidizing bacterial strains. Physiological experiments of Nitrosopumilus 402

maritimus SCM1 (A–D), Nitrospira moscoviensis NSP M-1 (E–H), Nitrospina 403

gracilis 3/211 (I–L), and Nitrococcus mobilis Nb-231 (M–P) from batch cultures 404

show variations in (A, E, I, M) cell abundance and N-nutrient concentration, (B, F, J, 405

N) total and (C, G, K, O) cell-specific ammonium/nitrite oxidation and dissolved 406

inorganic carbon (DIC) fixation rates, (D, H, L, P) the ratios of DIC fixation rate to 407

ammonium/nitrite oxidation rate and 16S rRNA:rDNA during batch growth. The error 408

bars for SCM1 and NSP M-1 represent standard deviations of biological replicates (n 409

= 3) except for gene (three technical replicates); some are not visible because they are 410

smaller than the symbols. Twice experiments were performed for strain 3/211. For the 411

first experiment (solid lines), biological replicates for strain 3/211 were not taken 412

because of the very small culture volumes; only two nitrite oxidation and one DIC 413

fixation rates were measured. Grey symbols in J and K panels represent the estimate 414

values of nitrite oxidation rate based on nitrate concentration changes over time. The 415

error bars for cell abundance of 3/211 represent standard deviations of technical 416

replicates (n = 3). For the second experiment (dashed lines), biological replicates were 417

performed and the error bars represent standard deviations of triplicate biological 418

replicates except for gene (three technical replicates); DIC fixation rates were 419

measured only at two time-points. For strain Nb-231, the error bars for cell abundance 420

and nutrient concentration represent standard deviations of biological replicates (n = 421

3) and for gene abundance standard deviations are from technical replicates (n = 3); 422

24

the rate values are means from two biological replicates. 423

25

424

425

Fig. S4. The favorability of ammonia oxidation over nitrite oxidation decreases 426

with increasing nitrite concentration. 427

26

428

429

Fig. S5. Fitting the depth profiles below the euphotic zone of ammonia and nitrite 430

oxidation rates into the Martin curve using a power law equation. (A) Ammonia 431

oxidation (black symbols are from a previous study at the same stations (33)) and (B) 432

nitrite oxidation rates at sites S6 (circles), S7 (triangles), and S8 (squares) in the South 433

China Sea (SCS). (C) Ammonia oxidation and (D) nitrite oxidation rates at site W2 434

(diamonds) in the Western Pacific Ocean (WP). Euphotic zone depth is 100 m in the 435

Ammonia oxidation rate (nM N d－1

)

0 5 10 15 20

De

pth

(m

)

0

500

1000

1500

2000

2500

3000

Ammonia oxidation rate (nM N d－1

)

0 10 20 30 40 50D

epth

(m

)

0

500

1000

1500

2000

2500

3000

Nitrite oxidation rate (nM N d－1

)

0 5 10 15 20 250

500

1000

1500

2000

2500

3000

Nitrite oxidation rate (nM N d－1

)

0 5 10 15 20 250

500

1000

1500

2000

2500

3000

WP

Rate = 17.44 × (

)－3.84

R2 = 0.99 P < 0.01

WP

Rate = 21.7 × (

)－3.5

R2 = 0.95 P < 0.01

SCS

Rate = 21.03 × (

)－3.13

R2 = 0.9 P < 0.01

SCS

Rate = 21.83 × (

)－3.1

R2 = 0.88 P < 0.01

A B

C D

27

SCS and 200 m in the WP. Error bars represent standard deviations of biological 436

replicates (n = 3). Some are not visible because they are smaller than the symbols. 437

28

438

439

Fig. S6. Michaelis-Menten kinetics of ammonia oxidation. Ammonia oxidation 440

rates were measured at different substrate concentrations in the South China Sea 441

(colored lines and dots; data are from a previous study (34)) and the Western Pacific 442

Ocean (black lines and diamonds; data are from 150 m of site W2 in this study). 443

Measured rates were obtained from the slope of the linear regression of four 444

independent time course bottles (see Materials and Methods). Error bars represent the 445

standard error of the regression coefficient. The solid lines were fitted using the 446

Michaelis-Menten equation. R2, coefficients (Vmax and Ks) of the best fit and their 447

standard errors are shown in the figure. 448

29

449

Fig. S7. Simple estimate of the steady state concentration of nitrite incorporating 450

observed uptake kinetic parameters and an exponential fit of these constants that 451

extend the measurements to depth. (A) Maximum uptake rate Vmax as measured 452

(dots) and with a simple extension of the values to depth (line). (B) Half-saturation 453

constant Ks as measured (dots) and with a simple extension to depth (line). (C) Two 454

estimates of population level specific loss rates, assuming constant (blue line) and 455

exponentially decreasing (red line) rates with depth. (D) Observed and calculated 456

nitrite concentrations (calculated using Equation (1) for subsistence concentration 457

R*). The calculation uses the yield of nitrite oxidation of y = 1/334 mol biomass N 458

synthesized per mol NO2− (35) and two estimates of loss rate L: a constant loss rate 459

(0.005 d−1

) and an exponentially decreasing loss rate with depth (𝐿 = 𝐿0𝑒−(𝑧−𝑧0)

𝑧∗⁄ 460

with L0 = 0.01 d−1

, z0 = 75 m, and z* = 50 m), and converts the bulk uptake rate to a 461

specific uptake rate using a cell abundance of 2.4×105 cells mL

−1 as estimated from 462

the observations and a cell quota of 6 fmol N per cell. With the exception of the most 463

shallow data point, the calculations illustrate how the decrease in both Vmax and Ks 464

with depth compensates for one another to give relatively similar steady state 465

concentrations of nitrite. 466

30

467

468

Fig. S8. Abundance of subgroups of 16S rRNA genes of nitrite-oxidizing bacteria 469

along the water column obtained by quantitative PCR. The error bars represent 470

standard deviations of technical replicates (n = 3). Some are not visible because they 471

are smaller than the symbols. 472

Genes (copies L 1

)

101 102 103 104 105 106 107D

ep

th (

m)

5

25

50

75

90

100

Genes (copies L 1

)

101 102 103 104 105 106 107

5

75

100

200

500

800

1000

2000

3000

Genes (copies L 1

)

101 102 103 104 105 106 107

150

200

250

300

350

400

450

500

650

750

800

1000

2000

3000

5

100

Nitrospira 16S rRNA

Nitrospina 16S rRNA

S1 S2 S3 S4 S5 S6 W1 W2

SCS Shelf

SCSSlope & Basin WP

A B C

_ _ _

31

473

474

Fig. S9. Physiological characteristics of the ammonia-oxidizing archaeal and nitrite-oxidizing bacterial strains. Physiological 475

experiments of Nitrosopumilus maritimus SCM1 (A–D) and Nitrospira moscoviensis NSP M-1 (E–H) from batch cultures show variations in 476

(A, E) cell abundance and N-nutrient concentration, (B, F) total and (C, G) cell-specific ammonium/nitrite oxidation and477

0 2 4 6 8 10 12 14

N-n

utr

ien

t co

nce

ntr

atio

n (

mM

)

0.0

0.4

0.8

1.2

Ab

un

da

nce

(1

07 c

ells

mL

1)

0

2

4

6

8

10

NO

2

NH

4+

NH

4+

+ N

O2

0 2 4 6 8 10 12 14

Ra

tio

of 1

6S

rR

NA

:rD

NA

0

20

40

60

80

100

Ca

rbo

n y

ield

(10

-2 m

ol C

fix

ed

pe

r m

ol N

oxid

ize

d)

0

5

10

15

20

0 2 4 6 8 10 12 14

Am

mo

niu

m o

xid

atio

n r

ate

pe

r ce

ll (1

02 fm

ol N

ce

ll 1 d

1)

0.0

0.1

0.2

0.3

DIC

fix

atio

n r

ate

pe

r ce

ll (f

mo

l C

ce

ll 1 d

1)

0

1

2

3A C D

0 2 4 6 8 10 12 14

DIC

fix

atio

n r

ate

(μM

C d

1)

0

5

10

15

20

Am

mo

niu

m o

xid

atio

n r

ate

(10

2 μ

M N

d 1

)

0

1

2

3B

Time (days)

0 1 2 3 4 5 6 7 8 9 101112N-n

utr

ien

t co

nce

ntr

atio

n (

mM

)

0.0

0.5

1.0

1.5

2.0

2.5A

bu

nd

an

ce

(1

07 c

ells

mL

1)

0.0

0.5

1.0

1.5

2.0

Time (days)

0 1 2 3 4 5 6 7 8 9 101112

Nitrite

oxid

atio

n r

ate

pe

r ce

ll (1

02 fm

ol N

ce

ll 1 d

1)

0.0

0.4

0.8

1.2

DIC

fix

atio

n r

ate

pe

r ce

ll (f

mo

l C

ce

ll 1 d

1)

0

2

4

6

NO

2

NO

3N

O2 +

NO

3

Time (days)

0 1 2 3 4 5 6 7 8 9 101112

Ca

rbo

n y

ield

(10

-2 m

ol C

fix

ed

pe

r m

ol N

oxid

ize

d)

0

2

4

6

8

Time (days)

0 1 2 3 4 5 6 7 8 9 101112

Nitrite

oxid

atio

n r

ate

(10

2 μ

M N

d 1

)

0

2

4

6

8

DIC

fix

atio

n r

ate

(μM

C d

1)

0

5

10

15

20

E

F G H

AOA Nitrosopumilus maritimus SCM1 (30 C)

NOB Nitrospira moscoviensis NSP M-1 (35 C)°

°

__

_

__

_ _

_

_

__

_

__

__

__

__

32

dissolved inorganic carbon (DIC) fixation rates, (D, H) the ratios of DIC fixation rate 478

to ammonium/nitrite oxidation rate and 16S rRNA : rDNA during growth in batch 479

cultures. The error bars for cell abundance and nutrient concentration represent 480

standard deviations of biological replicates (n = 3) and for gene abundance standard 481

deviations are from technical replicates (n = 3). All rate values are means from two 482

biological replicates.483

33

484

485

Fig. S10. The ratio of Thaumarchaeota 16S rRNA to rRNA gene abundance 486

versus the ratio of nitrite-oxidizing bacteria (NOB) 16S rRNA to rRNA gene 487

abundance. The dashed line is the 1:1 line. Linear regression is on the log-log scale. 488

Ratio of Thaumarchaeota 16S rRNA:rDNA

10-5 10-4 10-3 10-2 10-1 100 101 102

Ra

tio

of

NO

B 1

6S

rR

NA

:rD

NA

10-5

10-4

10-3

10-2

10-1

100

101

102

R = 0.90P < 0.01

1:1S6 W1 W2

34

Table S1. Taxonomic breakdown for target groups in metatranscriptomes (Please refer 489

to the excel file). 490

491

Table S2. Detected sequences for nitrogen, carbon, and sulfur metabolism related 492

genes in metatranscriptomes (Please refer to the excel file). 493

35

Table S3. Primer pair sequences, qPCR mixtures and conditions for each gene. 494

Target gene Primer Sequence (5'-3') PCR mixture PCR conditions Efficiency Detection

limita

References

β-proteobacterial

amoA

amoA-1F GGGGHTTYTACTGGTGGT 25 l reaction mixture:

SYBR® Premix Ex Taq™

(TakaRa, Dalian, China) 12.5

l, BSA 5 g, primers 0.4

M, DNA template 1 l

94 °C for 30 s;

45 × (94 °C for

15 s, 60 °C for

60 s, and 72 °C

for 90 s).

96-104 % 2

Hu et al. (20);

Mincer et al.

(21)

amoA-2R CCCCTCKGSAAAGCCTTCTTC

Archaeal amoA Arch-amoAF STAATGGTCTGGCTTAGACG

25 l reaction mixture:

SYBR® Premix Ex Taq™

(TakaRa) 12.5 l, BSA 5 g,

primers 0.4 M, DNA

template 1 l

95 °C for 30 s;

40 × (95 °C for

30 s, 53 °C for

60 s, and 72 °C

for 45 s).

91-98% 3

Hu et al. (20);

Francis et al.

(36)

Arch-amoAR GCGGCCATCCATCTGTATGT

Archaeal amoAb

Arch-

amoAFB CATCCRATGTGGATTCCATCDTG

25 l reaction mixture:

SYBR® Premix Ex Taq™

(TakaRa) 12.5 l, BSA 5 g,

primers 0.4 M, DNA

template 1 l

95 °C for 240 s;

40 × (95 °C for

30 s, 55 °C for

45 s, and 72 °C

for 60 s).

91-100% 2 Beman et al.

(37)

Arch-amoAR GCGGCCATCCATCTGTATGT

Thaumarchaeota

16S rRNA

GI_751F GTCTACCAGAACAYGTTC

20 l reaction mixture:

SYBR® Premix Ex Taq™

(TakaRa) 10 l, BSA 4 g,

primers 0.4 M, DNA

template 1 l

95 °C for 30 s;

40 × (95 °C for

15 s, 58 °C for

30 s, and 72 °C

for 30 s).

92-98 % 1

Hu et al. (20);

Mincer et al.

(21)

GI_956R HGGCGTTGACTCCAATTG

Nitrospira

16S rRNA

Nspra-675f GCGGTGAAATGCGTAGAKATCG

25 l reaction mixture:

SYBR® Premix Ex Taq™

(TakaRa) 12.5l, BSA 15 g,

primers 0.2 M, DNA

template 1 l

95 °C for 10 min;

45 × (94 °C for

30 s, 64 °C for

30 s, 72 °C for

60 s).

92-98 % 2 Attard et al.

(38)

Nspra-746r TCAGCGTCAGRWAYGTTCCAGAG

36

Nitrospina

16S rRNA

NitSSU_130F GGGTGAGTAACACGTGAATAA

25 l reaction mixture:

SYBR® Premix Ex Taq™

(TakaRa) 12.5l, BSA 1 g,

primers 0.4 M, DNA

template 1 l

94 °C for 15 min;

45 × (94 °C for

15 s, 57.5 °C for

15 s, 72 °C for

30 s, 77 °C for 1

s).

90-100 % 3 Mincer et al.

(21)

NitSSU_282R TCAGGCCGGCTAAMCA

Archaeal accA Crena_529F GCWATGACWGAYTTTGTYRTAATG

25 l reaction mixture:

SYBR® Premix Ex Taq™

(TakaRa) 12.5 l, BSA 30 g,

primers 1.6 M, MgCl2 2mM,

DNA template 1 l

40 × (95 °C for

30 s, 53 °C for

45 s, and 72 °C

for 45 s).

91-97 % 1

Yakimov et al.

(39); Hu et al.

(20)

Crena_981R TGGWTKRYTTGCAAYTATWCC

Bacterial 16S

rRNAc

EUB338 ACTCCTACGGGAGGCAGCAG 30 l reaction mixture:

SYBR® Premix Ex Taq™

(TakaRa) 15 l, BSA 0.38 g,

primers 0.2 M, DNA

template 1 l

95 °C for 3.25

min; 40 × (95 °C

for 15 s, 55 °C

for 30 s, 72 °C

for 30 s).

98-100 % 4 Castro et al.

(40) EUB518 ATTACCGCGGCTGCTGG

aThe unit of detection limit of the qPCR reaction: gene copies per PCR reaction. 495

bThe primer set was used in the samples from waters column below 450 m at the sites in the West Pacific. 496

cThe primer set was used in the samples from the culture of Nitrococcus mobilis Nb-231. 497

37

Table S4. Summary statistics for metatranscriptome sequencing, assembly, 498

annotation, and read mapping. 499

5m 200m 3000m

Sequences quality control

Raw data (Mbp) 6,844.90 6865.70 5,314.06

Clean data (Mbp) 6,193.43 6,090.97 5,000

Clean data rate (%) 90.48 88.72 94.09

High quality data (Mb) 6,186.30 6,090.04 4,990.64

High quality data rate (%) 90.38 88.70 93.91

Assembly Resultsa

Number of Contig 3,079,143

Mean Length (nt) of Contig 203

Number of Unigene 419,084

Mean Length (nt) of Unigene 649

Assessment of assembly

Number of Total Reads Pairs 30,931,478 30,450,216 27,725,767

Alignment rate (%) 72.71 49.36 65.70

Number of annotated Unigene

NR 321,088

KEGG 240,237

GO 205,814

ALL 322,138

Number of mapped readsb 9,784,207 10,597,309 11,149,154

aThe clean data from the three samples were pooled for assembly. 500

bAlignment of the clean data to the merge-Unigene was carried out using Bowtie. 501

38

Table S5. Archaeal and β-proteobacterial amoA gene abundances in the South China 502

Sea and Western Pacific Ocean. 503

Station Depth

(m)

β-proteobacterial amoA Archaeal amoA

Gene abundance

(copies L−1

) SD

Gene abundance

(copies L−1

) SD

S1 5 0

2.23 × 10

3 452

25 3.15 × 103 32 1.10 × 10

6 1.68 × 10

4

S2

5 0

1.92 × 103 89

25 0

6.34 × 104 2.47 × 10

3

50 613

1.81 × 107 4.97 × 10

5

S3

5 44 15 3.23 × 103 130

25 0

1.51 × 106 4.82 × 10

4

50 76 18 5.92 × 106 3.47 × 10

5

75 0

9.02 × 106 4.04 × 10

5

90 1.09 × 103 259 3.40 × 10

7 1.08 × 10

6

S4

5

Not detected (ND)

6.05 × 103 171

75 1.51 × 107 8.11 × 10

5

200 5.53 × 106 2.16 × 10

5

500 4.22 × 106 3.23 × 10

5

S5

5

ND

2.07 × 103 3

75 4.17 × 106 3.65 × 10

5

200 4.74 × 106 9.40 × 10

4

500 3.35 × 106 1.86 × 10

5

800 1.46 × 106 1.27 × 10

5

1000 9.83 × 105 9.11 × 10

4

2000 2.39 × 105 3.77 × 10

3

S6

5

ND

70 8

75 4.15 × 105 3.22 × 10

4

100 1.30 × 106 1.22 × 10

5

200 3.57 × 106 6.34 × 10

5

500 1.87 × 106 8.43 × 10

4

800 8.24 × 105 1.03 × 10

5

3000 1.29 × 105 2.06 × 10

4

W1

5 0

0

25 0

0

50 384 38 0

75 57 10 0

100 29 5 3.62 × 103 143

200 92 27 1.33 × 106 7.48 × 10

4

500 402 50 1.34 × 106 1.66 × 10

5

800 42

4.67 × 105 1.56 × 10

4

2000 141 63 5.43 × 104 4.62 × 10

3

3000 26

1.89 × 104 243

39

W2

25 110 5 2.09 × 103 87

50 0

6.05 × 105 4.19 × 10

4

55 106 92 3.06 × 105 2.68 × 10

4

75 260 30 3.51 × 106 8.82 × 10

4

150 79 21 3.42 × 106 3.03 × 10

5

200 66 20 1.84 × 106 4.68 × 10

4

250 53 6 2.61 × 106 1.29 × 10

5

300 217 101 2.68 × 106 8.17 × 10

4

350 563

3.29 × 106 2.33 × 10

5

400 167 22 1.24 × 106 1.52 × 10

4

450 279 144 2.53 × 105 1.60 × 10

4

500 314 102 6.23 × 105 1.15 × 10

4

650 122 18 1.62 × 106 8.38 × 10

4

750 391 78 1.46 × 106 1.39 × 10

4

800 266 176 1.36 × 106 1.19 × 10

5

1000 97 41 4.67 × 105 1.48 × 10

4

504

References 505

1. R. Massana, E. F. DeLong, C. Pedrós-Alió, A few cosmopolitan phylotypes 506

dominate planktonic archaeal assemblages in widely different oceanic provinces. 507

Appl. Environ. Microbiol. 66, 1777–1787 (2000). 508

2. D. Simms, P. E. Cizdziel, P. Chomczynski, TRIzol: a new reagent for optimal 509

single-step isolation of RNA. Focus 15, 99–102 (1993). 510

3. J. A. Gilbert et al., Detection of large numbers of novel sequences in the 511

metatranscriptomes of complex marine microbial communities. PloS ONE 3, 512

e3042 (2008). 513

4. K. Zhang et al., Sequencing genomes from single cells by polymerase cloning. 514

Nat. Biotechnol. 24, 680–686 (2006). 515

5. M. G. Grabherr et al., Trinity: reconstructing a full-length transcriptome without a 516

genome from RNA-Seq data. Nat. Biotechnol. 29, 644–652 (2011). 517

6. C. Iseli, C. V. Jongeneel, P. Bucher, ESTScan: a program for detecting, 518

40

evaluating, and reconstructing potential coding regions in EST sequences. Proc. 519

Int. Conf. Intell. Syst. Mol. Biol. 99, 138–148 (1999). 520

7. B. Langmead, C. Trapnell, M. Pop, S. L. Salzberg, Ultrafast and memory-521

efficient alignment of short DNA sequences to the human genome. Genome boil. 522

10, R25 1–10 (2009). 523

8. X. Peng et al., Ammonia and nitrite oxidation in the Eastern Tropical North 524

Pacific. Global Biogeochem. Cy. 29, 2034–2049 (2015). 525

9. M. R. McIlvin, M. A. Altabet, Chemical conversion of nitrate and nitrite to 526

nitrous oxide for nitrogen and oxygen isotopic analysis in freshwater and 527

seawater. Anal. Chem. 77, 5589–5595 (2005). 528

10. D. M. Sigman et al., A bacterial method for the nitrogen isotopic analysis of 529

nitrate in seawater and freshwater. Anal. Chem. 73, 4145–4153 (2001). 530

11. J. Granger, D. M. Sigman, Removal of nitrite with sulfamic acid for nitrate N and 531

O isotope analysis with the denitrifier method. Rapid Commun. Mass SP. 23, 532

3753–3762 (2009). 533

12. M. Könneke et al., Isolation of an autotrophic ammonia-oxidizing marine 534

archaeon. Nature 437, 543–546 (2005) 535

13. S. Ehrich, D. Behrens, E. Lebedeva, W. Ludwig, E. Bock, A new obligately 536

chemolithoautotrophic, nitrite-oxidizing bacterium, Nitrospira moscoviensis sp. 537

nov. and its phylogenetic relationship. Arch. Microbiol. 164, 16-23 (1995). 538

14. E. Spieck, A. Lipski, “Cultivation, growth physiology, and chemotaxonomy of 539

nitrite-oxidizing bacteria” in Methods in enzymology, A. E. Hargrove, Eds. 540

41

(Elsevier, 2011), pp. 109–130. 541

15. N. A. Ahlgren, C. A. Fuchsman, G. Rocap, J. A. Fuhrman, Discovery of several 542

novel, widespread, and ecologically distinct marine Thaumarchaeota viruses that 543

encode amoC nitrification genes. ISME J. 13, 618 (2018). 544

16. J. H. Martin, G. A. Knauer, D. M. Karl, W. W. Broenkow, VERTEX: carbon 545

cycling in the northeast Pacific. Deep-Sea Res. 34, 267–285 (1987). 546

17. M. A. Charette, W. H. F. Smith, The volume of Earth's ocean. Oceanography 23, 547

112–114 (2010). 548

18. C. Wuchter et al., Archaeal nitrification in the ocean. Proc. Natl. Acad. Sci. U.S.A. 549

103, 12317–12322 (2006). 550

19. M. Pester, C. Schleper, M. Wagner, The Thaumarchaeota: an emerging view of 551

their phylogeny and ecophysiology. Curr. Opin. Microbiol. 14, 300–306 (2011). 552

20. A. Hu, N. Jiao, C. L. Zhang, Community structure and function of planktonic 553

Crenarchaeota: changes with depth in the South China Sea. Microb. Ecol. 62, 554

549–563 (2011). 555

21. T. J. Mincer et al., Quantitative distribution of presumptive archaeal and bacterial 556

nitrifiers in Monterey Bay and the North Pacific Subtropical Gyre. Environ. 557

Microbiol. 9, 1162–1175 (2007). 558

22. M. M. Yakimov et al., Contribution of crenarchaeal autotrophic ammonia 559

oxidizers to the dark primary production in Tyrrhenian deep waters (Central 560

Mediterranean Sea). ISME J. 5, 945–961 (2011). 561

23. I. A. Berg, D. Kockelkorn, W. Buckel, G. Fuchs, A 3-hydroxypropionate/4-562

42

hydroxybutyrate autotrophic carbon dioxide assimilation pathway in archaea. 563

Science 318, 1782–1786 (2007). 564

24. H. M. Dionisi et al., Quantification of Nitrosomonas oligotropha-like ammonia-565

oxidizing bacteria and Nitrospira spp. from full-scale wastewater treatment plants 566

by competitive PCR. Appl. Environ. Microbiol. 68, 245–253 (2002). 567

25. V. La Cono et al., Partaking of Archaea to biogeochemical cycling in oxygen-568

deficient zones of meromictic saline Lake Faro (Messina, Italy). Environ. 569

Microbiol. 15, 1717–1733 (2013). 570

26. J. M. Norton, J. J. Alzerreca, Y. Suwa, M. G. Klotz, Diversity of ammonia 571

monooxygenase operon in autotrophic ammonia-oxidizing bacteria. Arch. 572

Microbiol. 177, 139–149 (2002). 573

27. H. A. Levipan, V. Molina, C. Fernandez, Nitrospina-like bacteria are the main 574

586 drivers of nitrite oxidation in the seasonal upwelling area of the Eastern 575

South Pacific (Central Chile ~36°S). Environ. Microbiol. Rep. 6, 565–573 (2014). 576

28. D. A. Stahl, J. R. de la Torre, Physiology and diversity of ammonia-oxidizing 577

archaea. Ann. Rev. Microbiol. 66, 83–101 (2012). 578

29. S. W. Watson, J. B. Waterbury, Characteristics of two marine nitrite oxidizing 579

bacteria, Nitrospina gracilis nov. gen. nov. sp. and Nitrococcus mobilis nov. gen. 580

nov. sp. Arch. Microbiol. 77, 203–230 (1971). 581

30. D. L. Kirchman, Growth rates of microbes in the oceans. Annu. Rev. Mar. Sci. 8, 582

285–309 (2016). 583

31. T. S. Lankiewicz, M. T. Cottrell, D. L. Kirchman, Growth rates and rRNA content 584

43

of four marine bacteria in pure cultures and in the Delaware estuary. ISME J. 10, 585

823-832 (2015). 586

32. R. Schlitzer, Ocean Data View. odv.awi.de (2017). 587

33. M. Xu et al., Enhanced ammonia oxidation caused by lateral Kuroshio intrusion 588

in the boundary zone of the northern South China Sea. Geophys. Res. Lett. 45, 589

6585–6593 (2018). 590

34. X. S. Wan et al., Ambient nitrate switches the ammonium consumption pathway 591

in the euphotic ocean. Nat. Commun. 9, 915 (2018). 592

35. E. J. Zakem et al., Ecological control of nitrite in the upper ocean. Nat. Commun. 593

9, 1206–1218 (2018). 594

36. C. A. Francis, K. J. Roberts, J. M. Beman, A. E. Santoro, B. B. Oakley, Ubiquity 595

and diversity of ammonia-oxidizing archaea in water columns and sediments of 596

the ocean. Proc. Natl. Acad. Sci. U.S.A. 102, 14683–14688 (2005). 597

37. J. M. Beman, B. N. Popp, C. A. Francis, Molecular and biogeochemical evidence 598

for ammonia oxidation by marine Crenarchaeota in the Gulf of California. ISME 599

J. 2, 429–441 (2008). 600

38. E. Attard et al., Shifts between Nitrospira- and Nitrobacter-like nitrite oxidizers 601

underlie the response of soil potential nitrite oxidation to changes in tillage 602

practices. Environ. Microbiol. 12, 315–326 (2010). 603

39. M. M. Yakimov, V. La Cono, R. Denaro, A first insight into the occurrence and 604

expression of functional amoA and accA genes of autotrophic and ammonia-605

oxidizing bathypelagic Crenarchaeota of Tyrrhenian Sea. Deep-Sea Res. Pt. II 56, 606

44

748–754 (2009). 607

40. H. F. Castro, A. T. Classen, E. E. Austin, R. J. Norby, C. W. Schadt, Soil 608

microbial community responses to multiple experimental climate change drivers. 609

Appl. Environ. Microbiol. 76, 999–1007 (2010). 610