DNA and RNA-SIP Reveal Nitrospira Spp. As Key Drivers of Nitrification in 2 Groundwater-Fed Biofilters

Home , Nitrite oxidoreductase, Pedomicrobium

bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

1 DNA and RNA-SIP reveal Nitrospira spp. as key drivers of nitrification in 2 groundwater-fed biofilters

4 Running title: Nitrospira drives nitrification in groundwater-fed biofilters

5 Authors: Arda Gülay1,4*, Jane Fowler1, Karolina Tatari1, Bo Thamdrup3, Hans-Jørgen Albrechtsen1, 6 Waleed Abu Al-Soud2, Søren J. Sørensen2 and Barth F. Smets1*

7 1 Department of Environmental Engineering, Technical University of Denmark, Building 113, Miljøvej, 2800 8 Kgs Lyngby, Denmark. Phone: +45 45251600. FAX: +45 45932850. e-mail: [email protected], jfow@ 9 env.dtu.dk, [email protected], [email protected]* 10 2 Department of Biology, University of Copenhagen, Universitetsparken 15, Building 1, 2100 Copenhagen, 11 Denmark. Phone: +45 35323710. FAX: +45 35322128. e-mail: [email protected], [email protected] 12 3 Nordic Center for Earth Evolution, Department of Biology, University of Southern Denmark, Campusvej 55, 13 5230 Odense, Denmark. Phone: +45 35323710. FAX: +45 35322128. e-mail: [email protected] 14 4 Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, MA, United States, 15 26 Oxford St, Cambridge, MA 02138, Phone: +1 (617)4951564. e-mail: [email protected] 16

17 *Corresponding authors

18 Keywords: Nitrification, comammox, Nitrospira, DNA SIP, RNA SIP

19 bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

20 Abstract

21 Nitrification, the oxidative process converting ammonia to nitrite and nitrate, is driven by microbes 22 and plays a central role in the global nitrogen cycle. Our earlier metagenomics, amoA-amplicon, and 23 amoA-qPCR based investigations of groundwater-fed biofilters indicated a consistently high 24 abundance of comammox Nitrospira, and we hypothesized that these non-classical nitrifiers drive 25 ammonia-N oxidation. Hence, we used DNA and RNA stable isotope probing (SIP) coupled with 26 16S rRNA amplicon sequencing to identify the active members in the biofilter community when + - 13 - 12 - 27 subject to a continuous supply of NH4 or NO2 in the presence of C-HCO3 (labelled) or C-HCO3 28 (unlabelled). Allylthiourea (ATU) and sodium chlorate were added to inhibit autotrophic ammonia- 29 and nitrite-oxidizing bacteria, respectively. Our results confirmed that lineage II Nitrospira 30 dominated ammonium oxidation in the biofilter community. A total of 78 (8 in RNA-SIP and 70 in 31 DNA-SIP) and 96 (25 in RNA-SIP and 71 in DNA-SIP) Nitrospira phylotypes (at 99% 16S rRNA 32 sequence similarity) were identified as complete ammonia- and nitrite-oxidizing, respectively. We - 33 also detected significant HCO3 uptake by Acidobacteria subgroup10, Pedomicrobium, Rhizobacter, 34 and Acidovorax under conditions that favoured ammonium oxidation. Canonical Nitrospira alone 35 drove nitrite oxidation in the biofilter community, and activity of archaeal ammonia oxidizing taxa 36 was not detected in the SIP fractions. This study provides the first in-situ evidence of ammonia 37 oxidation by comammox Nitrospira in an ecologically relevant complex microbiome. 38 bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

39 Introduction

- - 40 Nitrification, the stepwise oxidation of ammonia (NH3) to nitrite (NO2 ) and nitrate (NO3 ), supplies

41 the substrates for processes that initiate the loss of reactive nitrogen from the biosphere as N2. 42 Understanding the organisms and environmental controls that drive nitrification is important as it 43 controls global homeostasis of the N cycle. In engineered environments, complete nitrification is 44 often desired: this is essential when waters are prepared and distributed for human consumption. - 45 Residual NH3 or NO2 - the result of incomplete nitrification - renders the water biologically unstable 46 and unsafe for human consumption. Hence, biological systems for source water treatment are 47 contingent on nitrifying prokaryotes. Based on evolutionarily conserved taxonomic (small subunit, 48 16S rRNA) and functional (e.g. ammonia monooxygenase, amoA) gene surveys, Nitrosomonas (1– 49 4), Nitrosoarchaeum, and Nitrososphaera have been identified as the abundant ammonium oxidizing 50 prokaryotes (AOP) and Nitrospira (5, 6) as the abundant nitrite oxidizing prokaryotes (NOP) in 51 drinking water treatment systems, consistent with the classical assumption of division of labor in the 52 two nitrification steps.

53 Our previous studies on rapid gravity sand filters (RGSF), used in potable water preparation from 54 groundwater, revealed nitrifying microbial communities with Nitrospira far more abundant than 55 Nitrosomonas (7),with several Nitrospira genomes containing genes for ammonia oxidation (8), and 56 with an abundance of comammox (complete ammonia oxidizing) amoA over AOB amoA genes (9). 57 Together with the concurrent discovery of comammox Nitrospira strains by others (10–12), this 58 suggested that comammox Nitrospira may drive ammonia oxidation in the examined groundwater- 59 fed RGSFs. In addition, like Nitrospira, several Acidobacterial, and γ- and α-proteobacterial taxa 60 were at consistently higher abundance than Nitrosomonas questioning their potential role in

61 nitrification, as NH3 is the primary growth substrate entering the filters (7, 13–15). Identifying the 62 active ammonia and nitrite oxidizing organisms is essential not only for making predictions for 63 engineering purposes, but also for understanding the niches and biodiversity of nitrifiers. There has 64 been a rapidly increasing documentation of global comammox Nitrospira occurrence across a myriad 65 of habitats ranging from the subsurface, to soils to sediments, and from groundwaters to source and 66 residual water treatment plants, but apparently excluding open oceanic waters (9, 16–19); with 67 occasional abundances that exceed those of canonical AOB (9, 20, 21). Nonetheless, it is yet to be 68 shown whether comammox Nitrospira truly drives ammonia oxidation in open oligotrophic 69 freshwater and soil environments, their presumed preferred habitat based on genomic and 70 physiological evidence (22, 23). bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

71 Here, we sought to identify the active ammonia and nitrite oxidizers in a groundwater-fed RGSF 72 using RNA and DNA stable isotope probing (SIP) coupled to 16S rRNA amplicon sequencing. Lab- 73 scale columns packed with filter material from a full-scale RGSF were fed with effluent water + - 13 74 amended with NH4 or NO2 and with C labelled or unlabelled HCO3- for 15 days in the presence 75 or absence of inhibitors of autotrophic ammonia and nitrite oxidation (24). Our findings indicate that 76 Nitrospira drive both ammonia and nitrite oxidation. In addition, several other taxa take up - 77 substantial HCO3 and their DNA and RNA increases in relative abundance when ammonia was the 78 only energy source provided. This study provides the first in-situ evidence of ammonia oxidation by 79 comammox Nitrospira in an ecologically relevant complex microbiome. 80 Materials and methods 81 Sampling sites and procedure 82 Filter material samples were collected from a rapid gravity sandfilter (biofilter) at Islevbro 83 waterworks (Rødovre, Denmark) in May 2013. The influent and effluent water quality is reported 84 elsewhere (13, 14, 25). Filter material was collected from three random horizontal locations of the 85 biofilter using a hand-pushed core sampler. From the extracted filter material core, the top 10 cm was 86 aseptically segregated on site and stored on ice for further use. A portion was frozen on-site in liquid 87 nitrogen for RNA extraction.

88 Column experiments and stable isotope labelling 89 Experiments were conducted using a continuous-flow lab-scale system consisting of glass columns 90 (2.6 cm diameter, 6 cm long) filled with parent filter material (26.5 cm3) as described previously 91 (14). Effluent water from the investigated waterworks was used as the medium in all experiments to 92 avoid interference of other autotrophic processes and approximate full scale conditions.

93 The experimental design consisted of 4 treatments applied to labeled and unlabeled control columns. 94 The experiments were organized in two phases of 4 columns each; filter material was sampled for 95 DNA and RNA extraction just before the onset of each experimental phase. In the 4 treatments, the + 96 influent waters were spiked with: (i) NH4 (NH4Cl at 1 mg/L N (71 M); Sigma-Aldrich, 254134), + 97 (ii) NH4 and ATU (N-Allylthiourea at 100 µM, Merck chemicals, 808158) , (iii) NO2ˉ (NaNO2 at 1 + 98 mg/L N (71 M)) Sigma-Aldrich, S2252), (iv) NH4 and ClO3ˉ (KClO3ˉ at 1 mM; 99%, Sigma- + 99 Aldrich, 12634) (Table 1) (26).(26). The applied flow rates (40 mL/hr) and influent (NH4 or NO2ˉ) + 3 100 concentrations were set to match the volumetric NH4 -N loading rates (approx. 1.5 g N/m /hr) 101 experienced by the full-scale parent biofilter (14). Test and control columns were operated for 15 102 days with continuous feeding to allow sufficient 13C label incorporation. Further details are given in 103 SI Appendix, SI Materials and Methods. bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

104 Analytical methods 105 Column effluents were sampled daily, filtered (0.2 µm cut-off), frozen and analyzed colorimetrically + 106 for NH4 and NO2ˉ as described in Tatari et al. (26). Colorimetric analysis of ammonium in samples + + 107 containing ATU underestimated the NH4 concentration (26) and thus NH4 in these samples was

108 quantified by flow injection analysis (27). NO3ˉ was quantified by Ion Chromatography (Dionex, 109 ICS 1500) fitted with a guard column (Dionex, AG 22) and an analytical column (Dionex, ION PAC + + 110 AS22). NH4 removal (%) was calculated by subtracting effluent from influent NH4 concentration + 111 and normalizing for the influent NH4 concentration. NO2ˉ removal (%) was calculated as the

112 difference between produced NO2ˉ concentration and effluent NO2ˉ concentration, after correcting

113 for trace NO2ˉ present in the water (ca. 0.3 μM NO2ˉ) and normalization for the produced

114 NO2ˉconcentration. The NO2ˉ produced by ammonia oxidation was estimated as the difference + 115 between influent and effluent NH4 concentrations. NO3ˉ accumulation (in %) was calculated from

116 the difference between the effluent NO3ˉ and influent NO3ˉ concentration, normalized for the

117 produced NO3ˉ concentration. The NO3ˉ produced was estimated as the difference between influent + - 118 and effluent NH4 (or NO2 in the case of Col. 5 and 6) concentrations. .

119 Nucleic acid extraction and stable isotope probing (SIP)

120 Filter material samples collected from the full-scale biofilter and the sacrificed columns were subject 121 to DNA and RNA extraction. Genomic DNA was extracted from 0.5 g of drained filter material 122 using the MP FastDNA™ SPIN Kit (MP Biomedicals LLC., Solon, USA) according to 123 manufacturer’s instructions. The concentration and purity of extracted DNA was checked by 124 spectrophotometry (NanoDrop Technologies, Wilmington, DE, USA). RNA was extracted from 125 frozen filter material samples (-80 °C) with a MoBio PowerSoil Total RNA Isolation Kit (#12866- 126 25) according to manufacturer’s instructions. The RNA was further purified with a Qiagen AllPrep 127 DNA/RNA Mini Kit (Hilden, Germany) and quantified with a Ribogreen RNA-quantification kit 128 (Invitrogen, Eugene, OR, USA). Extracted rRNA (approximately 650 ng) was mixed well with 129 cesium trifluoroacetate solution to achieve an initial density of 1.790 g/mL before ultracentrifugation 130 at 38,400 rpm for 72 h at 20 °C in Beckmann with a VTi65.2 rotor (28). Centrifuged rRNA gradients 131 were fractionated into 250 µL fractions, the buoyant density of each fraction measured with 132 refractometry and rRNA precipitated from fractions as described previously by Whiteley et al. (29). 133 The concentration of purified RNA was determined using a Ribogreen RNA-quantification kit.

134 Density gradient ultracentrifugation of DNA isolated from columns and full-scale was performed 135 according to Neufeld et al. (30). Briefly, 1.6 µg of DNA in CsCl with a final density of 136 approximately 1.725 g/mL was subject to ultracentrifugation at 44,800 rpm for 44 h, 20°C in an bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

137 ultracentrifuge (Beckmann) with a VTi65.2 rotor (Beckmann). Gradients were fractionated into 250 138 µL fractions, density was determined by refractometry and DNA was recovered by precipitation with 139 PEG. DNA concentration was determined using a Picogreen high sensitivity dsDNA quantification 140 kit (Invitrogen).

141 PCR amplification and tag sequencing

142 RNA purified from density gradient fractions, and from sacrificed column experiments and full scale 143 biofilters were reverse transcribed using reverse primer 1492R with Sensiscript RT kit (Qiagen) 144 according to the manufacturer’s protocol. 10 ng of cDNA or DNA from direct DNA extracts (Table 145 1) were used to amplify the V3-V4 regions of bacterial 16S rRNA genes using the Phusion (Pfu) 146 DNA polymerase (Finnzymes, Finland) and 16S rRNA gene targeted (rDNA) modified universal 147 primers PRK341F and PRK806R (31). PCR was performed as described in (7). All fractions (a total 148 of 145 fractions; Fig.S1.a) from DNA-SIP and selected fractions (a total of 62 fractions; 149 Supplementary Fig.S1.b) from RNA-SIP experiments were sequenced on an Illumina MiSeq and GS 150 FLX pyrosequencing platform, respectively. Pyrosequencing was applied in a two-region 454 run on 151 a 70-75 GS PicoTiterPlate using a Titanium kit (7); paired-end 16S rRNA amplicon sequencing was 152 done on the Illumina MiSeq platform with MiSeq Reagent Kit v3 (2 × 301 bp; Illumina). All 153 sequencing was performed at the National High-throughput DNA Sequencing Center (Copenhagen, 154 DK).

155 Bioinformatic and statistical analysis

156 All bioinformatic and statistical analyses are described in detail in SI Appendix, SI Materials and 157 Methods. Briefly, raw 454 sequence data from RNA-SIP samples were quality-checked (denoised) 158 with Ampliconnoise (32) and chimeras were removed with UCHIME (33) using default settings. 159 Raw Miseq Illumina sequence data from DNA-SIP samples were quality-controlled with MOTHUR 160 (34) and chimeras were removed with UCHIME (33) using a reference dataset. Sequence libraries 161 were combined and trimmed to 418 bp. All further sequence analyses were performed in QIIME 162 1.9.1(35).

163 A total of six filter steps were applied to identify ammonia and nitrite oxidizing phylotypes (Fig.S2). 13 - 164 Detailed steps are described in SI Appendix, SI Materials and Methods. OTUs incorporating H CO3

165 were determined by 1. (Filter 1) comparing the mean buoyant density of each OTU in columns with 166 and without 13C amendment (36), 2. (Filter 2) identifying OTUs affiliated with genera that are 167 present in both DNA and RNA SIP, and 3. (Filter 3) selecting OTUs with buoyant density shifts 168 higher than genus-specific 90% CIs for buoyant density shifts. The remaining filter steps were bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

169 applied to assess ammonia and nitrite oxidizing phylotypes. Cross-feeders and taxa performing

170 heterotrophic CO2 fixation were largely removed by 4. (Filter 4) excluding OTUs with lower 171 buoyant density shift than the maximum buoyant density shift value of labelled Nitrosomonas and 172 Nitrospira OTUs, respectively, 5. (Filter 5) selecting the genera which contained OTUs in both RNA + + 173 and DNA-SIP, and 6. (Filter 6) comparing the labelled OTUs between treatments (NH4 fed, NH4 - - + - 174 ATU fed, NO2 fed, NH4 -ClO3 fed): To identify ammonia oxidizing phylotypes, labelled OTUs in + 175 all treatments excluding the one fed only with NH4 , were removed from the labelled OTU library of + 176 NH4 fed treatment. To identify nitrite oxidizing phylotypes, labelled OTUs in the treatment fed with - - 177 ClO3 were removed from the labelled OTU library of the only NO2 fed treatment. 178 As an additional step, detected genera were ranked according to the increase in their relative 179 abundance in both total DNA and RNA from the beginning (day 0) to the end (day 15) of the 180 experimental runs. 181 R codes for all bioinformatics and statistics including the detection of labelled OTUs in DNA and 182 RNA –SIP can be found in https://github.com/ardagulay. 183 All sequence data have been deposited at NCBI GenBank under Biosample accession numbers from 184 SAMN12227610 to SAMN12227705.

185 Results

186 Four different experimental treatments were designed to identify the microbes involved in ammonia + + 187 and nitrite oxidation (Table 1): (i) 71 μM of NH4 (Cols 1-2), (ii) 71 μM of NH4 and 100 μM - + 188 allylthiourea (ATU) (Cols 3-4), (iii) 71 μM of NO2 (Cols 5-6) and (iv) 71 μM of NH4 and 1 mM 13 189 NaClO3 (Cols 7-8). All columns were operated with an influent containing 100% C-labelled or 190 100% unlabelled bicarbonate for 15 days.

13 ˉ ˉ 191 In Col. 1 (H CO3 ) and Col. 2 (HCO3 ), the full-scale conditions were mimicked, with the aim to ˉ 13 - 192 elucidate the complete in situ food web related to nitrification. In Col. 3 (HCO3 ) and Col. 4 (H CO3 + 193 ), ATU was used to suppress bacterial ammonia oxidation while feeding at the same NH4 loading as 194 in Col. 1 and 2 (26). Complete inhibition of bacterial ammonia oxidation with ATU has been 195 previously observed at ATU concentrations of 8-86 µM (37), while archaeal ammonia oxidation is 196 less sensitive to ATU (38). The mechanism of ATU inhibition in AOB is proposed to be chelation of 197 the Cu2+ from the active site in the AMO enzyme (24). To identify taxa associated with nitrite - 13 ˉ ˉ 13 198 oxidation, NO2 was fed to Col. 5 (H CO3 ) and Col. 6 (HCO3 ). In Col. 8 (H CO3) and Col. 7 ˉ + 199 (HCO3), ClO3 was used to inhibit nitrite oxidation under NH4 feeding, with the aim to identify the + 200 taxa solely associated with NH4 oxidation (26). Chlorate is commonly used as a selective inhibitor bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

201 for nitrite oxidation, as it is reduced by reverse activity of the nitrite oxidoreductase to the toxic ˉ 202 chlorite (ClO2 )(39, 40).

203 Table 1 Summary of experimental design, bulk 13C incorporation, substrate utilization and 204 accumulation levels, and sequenced samples

N C NH + (a) NO - (a) NO - (a) Total Runs Inhibitor 13C/12C 4 2 3 Total RNA SIP source source removal removal accretion DNA

+ 13 - - 99%±1 100%±0

Col.1 NH4 H CO3 279 88%±32 + + +

+ - Col.2 NH4 HCO3 - - 98%±3 100%±0 82%±28 + + + scale - (c) Run1 + - - 19%±15 101%±2 ND

Full Col.3 NH4 HCO3 ATU + + + Inoculum 1 (c)

+ 13 - Col.4 NH4 H CO3 ATU 54 11%±15 99%±6 ND + + +

Col.5 NO - H13CO - - 89 NA (b) 88%±1 99%±34 + + + 2 3

- - (b) Col.6 NO2 HCO3 - - NA 92%±3 62%±36 + + + scale scale - (c) Run 2 Run2 + - - - 11%±5 70%±44 ND

Full Col.7 NH4 HCO3 ClO3 + + + Inoculum 2 + 13 - - (c) Col.8 NH4 H CO3 ClO3 63 6%±7 97%±28 ND + + +

(a) + - - Removal and accumulation rates were estimated from daily NH4 , NO2 , NO3 measurements - -NO2 removal was calculated based on ammonium removed (except for Col. 5 & 6 where it is based on influent nitrite) - -NO3 accretion was calculated based on ammonium removed (except for Col. 5 & 6 where it is based on nitrite removed) (b) NA: Not Applicable (c) - ND: Differences between in and effluent NO3 concentrations were not detectable and accretion could not be calculated 205 206 Physiological activity

+ + 207 In the 71 μM NH4 fed treatments, complete NH4 removal (99%) was observed without inhibitor + 208 addition, while NH4 removal ranged from 11 to 19% with ATU-amendment. (Table 1, Fig.S3.a). 13 + 209 Inhibitor addition also significantly reduced overall C incorporation. Columns fed with NH4 , + - - 210 NH4 -ATU and NO2 all had similarly high degrees of NO2 removal ranging from 88% to 100%. In - + 211 the 1 mM ClO3 amended columns, NH4 removal was severely inhibited (6% to 11%); removal of - - 212 formed NO2 continued (from 70 to 54%), although accumulation of NO3 could not be detected. + - - 213 Nitrogen mass balances, based on influent and effluent NH4 , NO2 , and NO3 concentrations closed 214 for most experimental runs minimizing the possibility of additional nitrogen cycling; N loss was only 215 observed in the ATU supplemented columns (columns 3,4) with ongoing treatment (Table 1; 216 Fig.S3.b-c)

217 Detection of 13C-labelled taxa from DNA- and RNA- SIP

218 DNA and RNA, extracted from column samples taken at the end of the experiments, were subject to 219 equilibrium density centrifugation, gradient fractionation and 16S rRNA gene amplification. A total bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

220 of 147 and 65 gradient fractions from DNA-SIP and RNA-SIP were sequenced using Illumina-Miseq 221 and 454-pyrosequencing platforms, respectively (Fig.S1.a-b). OTUs were defined at 99% similarity, 222 to minimize the effect of microdiversity, as 98.7% and lower similarities represent the taxonomic 223 levels of species, genus and higher (41).

224 We first examined the incorporation of 13C in OTUs in all treatments by comparing replicate 12 - 13 - 225 columns with H CO3 versus H CO3 amendment. In DNA-SIP, where all SIP fractions were 226 sequenced, we calculated the average shift in buoyant density of each OTU based on its relative 227 sequence abundance and buoyant density in all fractions (SI appendix, Eq. 2). As only selected

228 fractions were sequenced in RNA-SIP, the mean buoyant density of each OTU in treatments with 12 - 13 - 229 H CO3 versus H CO3 amendment was calculated using the standard deviation of the RNA 230 distribution across the buoyant density gradient (Fig.S4), as described in Zemb et al.(36). The

231 buoyant density shift of each OTU was then determined from the calculated mean buoyant density in 232 the replicate columns of each treatment.

+ + 233 Among all detected OTUs (3,364,425), 4,075 and 5,045 in NH4 treatment, 4,133 and 5,155 in NH4 - + - 234 plus ATU treatment, 4,183 and 706 in NO2 treatment, and 44,916 and 52 in the NH4 plus ClO3 235 fed treatment showed a buoyant density shift (after Filter 1; Fig.S2) in the DNA- and RNA-SIP 236 experiment, respectively. Only those OTUs that belonged to genera that contained OTUs that were 237 detected as 13C-labelled in both RNA and DNA-SIP were retained (Filter 2; Fig.S2). A bootstrap 238 resampling of labelled OTUs within each genus was then used to estimate taxon-specific 90% CIs 239 for the buoyant density shift of a labelled genus (Filter 3; SI Appendix, M.M and Fig.S2).

rd 13 + 240 Hence, after the 3 filter step, 676 (57% DNA, 43% RNA of the total C-labelled OTUs – NH4 fed + 241 treatment), 735 (67% DNA, 33% RNA – NH4 plus ATU fed treatment), 529 (65% DNA, 35% RNA - + - 242 – NO2 treatment) and 43 (83% DNA, 16% RNA – NH4 plus ClO3 fed treatment) OTUs were 243 retained as significantly labeled. The fractional 13C uptake of labelled OTUs was calculated by 244 dividing the DNA and RNA buoyant density shift for each OTU by the total observed buoyant 245 density shift for DNA and RNA, respectively. The abundance of labeled OTUs in the total 246 community were estimated based on total (i.e. non-fractionated) DNA and rRNA extracts collected 247 on day 15 (Fig.S5.a-d).

+ 13 248 In the NH4 -only fed treatment, C-labelled OTUs affiliated with 17 genera of the α-, β-, and γ- 249 Proteobacteria, Nitrospira, Actinobacteria, Latescibacteria, and Acidobacteria (Fig.S5a). Among 250 them, the genus Nitrospira had the highest fraction of 13C uptake (32% and 1.1% for DNA-SIP and 251 RNA-SIP respectively), and highest relative abundance in the total DNA (26%). Nitrosomonas spp. bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

252 OTUs were also labelled but displayed low levels of 13C uptake (0.07% and 0.8% for DNA-SIP and 253 RNA-SIP respectively) and were at low abundance (0.15% and 0.18% in total DNA and RNA, 254 respectively). Labelled ribosomal RNA, an approximation of metabolic activity, was distributed 255 evenly between 5 different 13C-labelled genera including Woodsholea, Blastocella, Subgroup 10 256 Acidobacteria, Pedomicrobium, and Sphingomonas (Fig.S5.a) Although ammonium oxidation was + 13 257 severely inhibited in the NH4 plus ATU fed column (Fig.1), OTUs in 15 genera incorporated C. + 258 These were identical to labelled OTUs in the NH4 fed treatment with the exception of OM27, 259 Rhizobacter, Variovorax and uncultured representatives of the order Xanthomonadales, which were 13 - 260 not labelled in the presence of ATU (Fig.S5.b). Azospira incorporated H CO3 only in the presence + 261 of ATU. In the NH4 plus ATU fed treatment, Nitrospira (10% DNA-SIP, 4.7% RNA-SIP), 262 Pseudomonas (3% DNA-SIP, 1% RNA-SIP), Methyloglobulus (2.7% DNA-SIP, 2.6% RNA-SIP) 263 and Blastocatella (2.6% DNA-SIP, 3.1% RNA-SIP), incorporated the highest fraction of label, while 264 Sphingomonas (1.8%) and Woodsholea (1.4%) were dominant in the total RNA pool (Fig.S5.b). In + - 265 the NH4 plus ClO3 amended columns, where both ammonium and nitrite oxidation were 266 suppressed, OM27 (2.7% DNA, 2.6% RNA) and Woodsholea were the only taxa that assimilated - 267 significant amounts of HCO3 (Fig.S5.c).

- 13 - 268 In the NO2 fed columns, 12 genera, belonging to α-Proteobacteria (10% of H CO3 uptake in DNA-

13 - 269 SIP, 23% of H CO3 uptake in RNA-SIP), -Proteobacteria (1% DNA-SIP, 21% RNA-SIP) and γ- 270 Proteobacteria (6% DNA-SIP, 11% RNA-SIP), Nitrospira (71% DNA-SIP, 15% RNA-SIP), 271 Actinobacteria (3% DNA-SIP, 7% RNA-SIP), Latescibacteria (0.5% DNA-SIP, 5% RNA-SIP), and 272 Acidobacteria (6% DNA-SIP, 11% RNA-SIP), were labelled (after Filter 3; Fig.S5.d). Nitrospira 13 - 273 had the highest number of labelled OTUs (96) and was responsible for the majority of the H CO3 274 uptake. After application of Filters 4 through 6, only Nitrospira OTUs were retained and hence 275 identified as the sole nitrite oxiders.

276

277 bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

278 279 Fig.1. (a) Approach to identify ammonia and nitrite oxidizing Nitrospira (b) Heatmap of all + - 280 identified labelled Nitrospira OTUs in columns fed with NH4 and NO2 , (c) 16S rRNA based 281 phylogenetic tree of all identified labelled Nitrospira OTUs and published Nitrospira strains with 282 known lineages. Open and filled triangles represent Nitrospira OTUs identified by RNA and DNA- bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

283 SIP, respectively. Comammox Nitrospira sequences are indicated by filled circles. Bootstrap values 284 >60% are shown by orange dots. The scale bar represents 0.01 substitutions per nucleotide position.

285 Hypothesis: Nitrospira is an active ammonium oxidizer

13 - + + - 286 H CO3 was incorporated by Nitrospira in treatments fed with NH4 , NH4 plus ATU and NO2 287 (after Filter 3, Fig.S5.a-d). The observed labelling in individual treatments does not indicate whether 288 labelled Nitrospira OTUs are capable of ammonia oxidation because both ammonium and nitrite + 289 oxidation occur in NH4 treatment. We therefore performed a binary comparision between labelled + - 290 Nitrospira OTUs detected in the NH4 versus NO2 fed treatments (Fig.1a). We assume that the + 291 labelled Nitrospira OTUs in NH4 fed treatment would include both comammox and nitrite oxidizing - 292 Nitrospira, while the NO2 fed treatment would exclude comammox Nitrospira based on observation 293 that comammox Nitrospira growth is not supported by oxidation of environmental nitrite in the 294 absence of ammonia(11).

295 Heatmaps of labelled Nitrospira OTUs (Fig.1.b) reveal that 8 (24%) and 70 (51%) OTUs are + 296 uniquely labelled in NH4 fed treatment at the RNA and DNA levels, respectively, indicating that 13 -. 297 several ammonium oxidizing Nitrospira strains actively assimilating H CO3 A high number of + 298 labelled Nitrospira OTUs are unique to the NH4 fed column, and few Nitrospira OTUs are shared + - 299 between the NH4 and NO2 fed columns suggesting that most comammox Nitrospira do not readily 300 switch from ammonium oxidation to solely nitrite oxidation. A large number of OTUs were also - 301 uniquely labelled in the NO2 amended columns (25 in RNA and 66 in DNA),which suggests that, in + - 302 the NH4 fed treatment, the produced NO2 was not sufficient to achieve labelling of nitrite-oxidizing 303 Nitrospira due to complete nitrification by commamox Nitrospira.

13 + - 304 Based on their C labelling in NH4 and NO2 fed treatments, 78 (8 in RNA-SIP and 70 in DNA-SIP) 305 and 96 (25 in RNA-SIP and 71 in DNA-SIP) Nitrospira OTUs were identified as complete 306 ammonia- and nitrite-oxidizing, respectively (Fig.1c). All labelled Nitrospira belonged to lineage II, 307 which comprises both comammox and non-comammox types. No clear branching between 308 comammox and nitrite oxidizing phylotypes was observed from the tree topology. bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

309 13 ˉ 310 Fig.2. 16S rRNA based phylogenetic tree showing phylotypes incorporating H CO3 in DNA-SIP and RNA-SIP experiments selective for putatively (a) 311 ammonium and (b) nitrite oxidation. Peak heights on circles represent (i) relative abundance in total DNA (purple) and (ii) total RNA (red) after 15 days, and 312 (iii) 13C label percentage (orange). The outer ring represents the OTUs retrieved from DNA-SIP (filled circles) or RNA-SIP (no circles). The scale bar 313 represents 0.10 substitutions per nucleotide position. bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

13 - 314 High H CO3 incorporation and growth by other bacteria

315 In our previous 16S rRNA amplicon-based analysis of the same and related RGSF communities, 316 members of the Rhizobiales (α-proteobacteria), and Acidobacteria were consistently more abundant + 317 than Nitrosomonas where NH4 is thought to be the largest source of energy available for microbial 13 - 318 growth (7). We observed that some of these taxa incorporated HCO3 both in RNA- and DNA-SIP 319 (after Filter 6; Fig. 2a). In addition, several OTUs displayed higher buoyant density shifts than + 320 Nitrosomonas in NH4 -fed columns (Fig.S5.a). Finally, these OTUs also increased in relative 321 abundance in both DNA and RNA over the course of the experiment (Fig.3c).

322 Only Nitrospira was associated with both ammonium and nitrite oxidation (after Filter 6; Fig.2a-b). 13 - + 323 Among HCO3 incorporating taxa in the presence of NH4 , Subgroup 10 (Acidobacteria), 324 Nitrospira (Nitrospira), Pedomicrobium, (α-Proteobacteria), Rhizobacter and Acidovorax (β- 325 Proteobacteria) displayed higher shifts in relative RNA and DNA abundance compared to 326 Nitrosomonas (Fig 3c). With the exception of Pseudomonas (Fig.S6 A), 13C-labelled genera in SIP 327 columns (Fig.S5.a-d) differed from the dominant genera in the feedwater, itself the effluent from 328 the full-scale biofilter. Hence, invasion from feed water communities did not cause the increased 329 relative abundance of Subgroup 10 Acidobacteria, Pedomicrobium, Rhizobacter, and Acidovorax.

330 A metagenome, obtained from the same parent biofilter within the same year (2013, (Palomo et al. 331 2016)) was further examined for its content of genes that are not canonical AOB amoA, canonical 332 AOA amoA, or comammox Nitrospira amoA, nor methanotrophic pmoA (Palomo et al. 2016; 333 Fig.S7) but showed homology with genes encoding AMOA protein family fragments from putative 334 heterotrophic nitrifiers (PF05145, IPR017516 Pfam and InterPro database, respectively). Twenty- 335 nine unique amoA gene fragments matching the PF05145 model were aligned with reference 336 putative heterotrophic amoA gene fragments (Fig. S7). However, no amoB and amoC genes were 337 found on any of the contigs carrying these atypical putative amoA gene fragments (Fig.S8); 338 furthermore no similar gene synteny was detected between the other genes of PF05145 amoA 339 containing contigs and our metagenome amoA containing contigs (Fig.S8).

340 The genus Nitrospira was the only taxon associated with nitrite oxidation (Fig. 2.b). bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

341

342 343 Fig.3. (a) Approach (Filter 6; SI Methods, Fig. S3) to identify putative ammonia and nitrite + - + 344 oxidizing phylotypes: (b) Heatmap of OTUs significantly labelled under NH4 , NO2 , NH4 plus + - 345 ATU, and NH4 plus ClO3 treatment (c) Fold-change in relative abundance in community DNA 346 and RNA for taxa identified as putative ammonia oxidizers (as shown also on Fig 2A). bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

347 Discussion

348 Stable isotope probing has previously been used to identify active nitrifiers in sediments (42–44) 349 and soils (45–48). In most studies, either DNA-SIP (45) or RNA-SIP (49) are applied individually; 350 yet both are important to identify key catalysts(48). While DNA-SIP detects isotope incorporation 351 into dividing cells, RNA-SIP detects active potentially slow- or non-growing cells (50). By 352 coupling SIP with next generation sequencing (NGS) we improved taxonomic resolution and 353 differentiated phylotypes in taxa with high microdiversity. 13 - 354 Here, we examined the assimilation of H CO3 coupled to nitrification in a RGSF using both RNA- 355 and DNA-SIP. OTUs incorporating 13C isotope label in the different treatments were 356 unambiguously identified as those displaying significant buoyant density shifts between the 12 - 13 - 357 H CO3 and the H CO3 replicates, and were detected at high phylogenetic resolution (>99% 358 pairwise identity (41)). A total of 200 gradient fractions were processed with sample size 359 equalization. + 360 Our results provide the first in-situ physiological evidence of ecologically relevant NH4 oxidation 361 by comammox Nitrospira in any environment. Other reports of in situ activity are inferred from 362 bulk observations (ammonium removal when comammox Nitrospira are more abundant than AOBs 363 or AOAs (20, 51)) or from comammox-specific amoA transcript analysis (52). Our former 364 observations that Nitrospira was more abundant than Nitrosomonas and the discovery that 365 Nitrospira harbours the complete nitrification pathway in a full-scale RGSF microbiome (8, 10, 11) 366 is thus directly linked to the ammonia oxidizing activity of Nitrospira in this environment. 13 - + - 367 Nitrospira was the only genus incorporating H CO3 in both NH4 fed and NO2 fed treatments, + - 368 indicating that Nitrospira is the only genus oxidizing both environmental NH4 and NO2 in this 369 system. 370 Ammonia and nitrite oxidizing phylotypes of Nitrospira were compared phylogenetically; the 371 resulting 16S rRNA tree topology shows no clear evolutionary separation of comammox and 372 canonical Nitrospira. This is in line with previous studies that show that comammox Nitrospira are 373 not evolutionarily distant from known canonical Nitrospira (>99% 16S rRNA nucleotide identity) 374 (10). Furthermore, our phylogenetic analysis shows that the labelled – both ammonia oxidizing and 375 nitrite oxidizing - Nitrospira OTUs, branch within Nitrospira sublineage II, as reported in previous 376 studies (8, 10, 11, 53). We did not identify the amoA clade affiliation of the active comammox 377 Nitrospira phylotypes in this study, although separate investigations on this and related RGSF have 378 indicated a predominance of amoA clade B comammox Nitrospira (9, 23). 379 It remains unclear whether comammox Nitrospira can switch between modes of ammonia and 380 nitrite oxidation. However, the large numbers of Nitrospira phylotypes which were exclusively bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

+ - 381 labelled in the NH4 versus NO2 fed columns, respectively, suggests that comammox Nitrospira - 382 may not prefer to oxidize external NO2 alone, in agreement with observations in Can. N. inopinata 383 (10, 11). - 384 Although ClO3 is a well-known competitive inhibitor of nitrite oxidoreductase (54), strong + - 385 inhibition of both ammonium and nitrite oxidation was observed in NH4 plus ClO3 fed columns. 386 We expect that the inhibition of nitrite oxidation in comammox Nitrospira would be caused by - - 387 ClO3 reduction to the toxic ClO2 , which would negatively affect overall metabolism in these - 388 organisms including ammonia oxidation. Hence, it appears that the inhibitory effect of ClO3 on 389 ammonium oxidation provides preliminary support for the contribution of comammox Nitrospira to - 390 ammonia oxidation, as we observed before (26). In addition, ClO2 may contribute to inhibition of 391 other ammonia oxidizers as observed before (55). PTIO has also been documented as a potent 392 inhibitor of ammonia oxidation in N. inopinata, although its selectivity is unclear (56). ATU + 393 significantly suppressed ammonia oxidation in NH4 -ATU fed treatments, although the taxa 13 - + 394 assimilating HCO3 did not change significantly compared to the NH4 fed treatment, excluding 395 Azospira. This similarity in labelled taxa in the ATU-fed treatment may indicate that the AMO of 396 the major ammonia oxidizers in this environment may be less sensitive to ATU than AOB at the 397 given concentrations, as previously observed for AOA (57). 398 No archaeal taxa were 13C-labelled in any of the columns although archaeal ammonium oxidizers 399 (AOA) are present, be it at much lower abundance than Nitrospira (100 to 1000 fold), in the RGSF + 400 used in this study (8, 13, 58). Columns were fed 71 μM NH4 to mimic full-scale conditions (14), 401 while bottom layers of the full-scale biofilter receive very low ammonium concentrations due to 402 removal in the top layers (14, 15). The absence of AOA in the 13C-labelled taxa may be due to their 403 low initial abundances, or the elevated concentrations applied during the experiment, as AOA may 404 thrive better in conditions of reduced energy supply consistent with their elevated abundance at 405 bottom layers of the examined RGSF (59–62); even though a recently isolated comammox strain N. + 406 inopinata displayed higher NH4 affinity than many of the characterized AOAs (22).

407 After strong filters were applied to remove heterotrophic OTUs, we retain the following taxa with 13 - + 408 substantial H CO3 incorporation in the NH4 fed treatments: Subgroup10, Pedomicrobium, 409 Rhizobacter, and Acidovorax; these taxa also show a greater fold-change in relative abundance in 410 DNA and RNA during the experiment than Nitrosomonas and Nitrospira OTUs (Fig.3c; Fig.S9).

411 Most heterotrophic microbes can engage in CO2 fixation via carboxylation reactions (63, 64).

412 However, CO2 fixation via anaplerotic metabolism (65) typically results in only 3-8% of the cellular 413 carbon assimilated by heterotrophs, which would be insufficient for label detection by DNA-SIP bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

13 - 414 (63, 66). Thus, heterotrophic carbon fixation would not explain the higher H CO3 incorporation 415 extents (higher density shifts, See Fig.2, TableS1, S2) relative to Nitrosomonas OTUs, known 416 ammonium oxidizers. Furthermore, the cellular mass and activity supported by cross-feeding decay 417 products from autotrophs (67) would be significantly less than the chemolithoautotrophic biomass 418 and activity itself. Thus, the observation of 13C-labelled genera with higher buoyant density shifts 419 and higher fold-changes in DNA and RNA abundance shifts (Fig.2, Fig.3) compared to 420 Nitrosomonas and Nitrospira, are difficult to explain by cross-feeding alone. 13 - 421 Can a plausible explanation for the high H CO3 assimilation of these taxa be nitrification? An 422 earlier metagenome from the same parent material revealed the presence of putative amoA genes 423 that could not be classified as amoA from canonical AOB, canonical AOA, or comammox 424 Nitrospira ((8); Fig.S7); yet were phylogenetically related to the PF05145, purported to contain 425 AMOA encoding genes in heterotrophic bacteria (68). In addition, the phylogeny of 10/30 of these 426 aberrant amoA genes indicated their presence, among others, in Hyphomicrobiaceae and + 427 Comamonadaceae . Of the highly labeled taxa (in both RNA-SIP, and DNA-SIP) in the NH4 fed 428 treatment, Pedomicrobium, and Acidovorax (but not Subgroup10 or Rhizobacter) belong to the 429 Hyphomicrobioaceae and Comamondaceae. While it is tempting to speculate that we have 430 identified novel ammonia oxidizing bacteria, we are unable to identify additional amo genes that 431 would constitute a complete amo operon on any of the metagenomic contigs. In addition, recent 432 doubt has been cast on the assignment of PF05145 as encoding a putative ammonia monooxygenase 433 (69), and careful physiological or genomic evidence of heterotrophic nitrification remains elusive 434 (70). On the other hand, some of the Acidobacterial MAGs that were retrieved from the studied

435 RGSF metagenome, contained CO2 fixation pathways (8).

436 The second step of nitrification, the oxidation of nitrite to nitrate, is known to be performed by 437 nitrite-oxidizing chemolithoautotrophs such as Nitrotoga, Nitrospina, Nitrobacter, Nitrolancea and 438 Nitrospira (71–73), which use nitrite oxidoreductase (NXR) as the key enzyme. The known 439 autotrophic nitrite oxidizer Nitrospira was identified as the only active nitrite oxidizer in the studied 440 system. 441 In summary, comammox Nitrospira and Nitrosomonas are the chemolithoautotrophic drivers of 442 ammonia oxidation in the groundwater fed biofilter and comammox Nitrospira make the greatest 443 contribution to ammonia oxidation. This finding raises the question of whether comammox 444 Nitrospira is an equally important ammonia oxidizer in other environments. AOA did not 445 significantly contribute to nitrification and Nitrospira were the only nitrite oxidizers identified in 446 this environment. Hence, we provide the first in-situ evidence of ecologically relevant ammonia bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

447 oxidation by comammox Nitrospira in a complex microbiome and document an unexpectedly high 13 - 448 H CO3 uptake and growth of Proteobacterial and Acidobacterial taxa under ammonium 449 selectivity. 450

451 Acknowledgments

452 We thank Dr. Hannah Sophia Weber and Heidi Grøn Jensen for support with SIP, and Alex Palomo 453 for feedback on the manuscript. This research was supported by The Danish Council for Strategic 454 Research (Project DW Biofilter), the European Commission (MERMAID-ITN; An initial training 455 network funded by the PeopleProgramme - Marie Skodowska-Curie Actions- of the European 456 Union's Seventh Framework Programme FP7/2007-2013/ under REA grant agreement n1607492) 457 and the VILLUM foundation (Project Expa-N, 13391).

458

459 References

460 1. van der Wielen PWJJ, Voost S, van der Kooij D. 2009. Ammonia-oxidizing bacteria and 461 archaea in groundwater treatment and drinking water distribution systems. Appl Environ 462 Microbiol 75:4687–95.

463 2. de Vet WWJM, Kleerebezem R, van der Wielen PWJJ, Rietveld LC, van Loosdrecht MCM. 464 2011. Assessment of nitrification in groundwater filters for drinking water production by 465 qPCR and activity measurement. Water Res 45:4008–18.

466 3. Niu J, Kasuga I, Kurisu F, Furumai H, Shigeeda T. 2013. Evaluation of autotrophic growth 467 of ammonia-oxidizers associated with granular activated carbon used for drinking water 468 purification byDNA-stable isotope probing. Water Res 47:7053–7065.

469 4. White CP, Debry RW, Lytle D a. 2012. Microbial survey of a full-scale, biologically active 470 filter for treatment of drinking water. Appl Environ Microbiol 78:6390–4.

471 5. Martiny AC, Albrechtsen HJ, Arvin E, Molin S. 2005. Identification of bacteria in biofilm 472 and bulk water samples from a nonchlorinated model drinking water distribution system: 473 detection of a large nitrite-oxidizing population associated with Nitrospira spp. Appl Env 474 Microbiol 71:8611–8617.

475 6. de Vet WWJM, Dinkla IJT, Muyzer G, Rietveld LC, van Loosdrecht MCM. 2009. Molecular 476 characterization of microbial populations in groundwater sources and sand filters for drinking 477 water production. Water Res 43:182–94. bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

478 7. Gülay A, Musovic S, Albrechtsen H-J, Al-Soud WA, Sørensen SJ, Smets BF. 2016. 479 Ecological patterns, diversity and core taxa of microbial communities in groundwater-fed 480 rapid gravity filters. ISME J 10:2209–2222.

481 8. Palomo A, Fowler SJ, Gülay A, Rasmussen S, Sicheritz-Ponten T, Smets BF. 2016. 482 Metagenomic analysis of rapid gravity sand filter microbial communities suggests novel 483 physiology of Nitrospira spp. ISME J 10:2569–2581.

484 9. Fowler SJ, Palomo A, Dechesne A, Mines PD, Smets BF. 2018. Comammox Nitrospira are 485 abundant ammonia oxidizers in diverse groundwater-fed rapid sand filter communities. 486 Environ Microbiol 20:1002–1015.

487 10. van Kessel MAHJ, Speth DR, Albertsen M, Nielsen PH, Op den Camp HJM, Kartal B, Jetten 488 MSM, Lücker S. 2015. Complete nitrification by a single microorganism. Nature 528:555– 489 559.

490 11. Daims H, Lebedeva E V, Pjevac P, Han P, Herbold C, Albertsen M, Jehmlich N, Palatinszky 491 M, Vierheilig J, Bulaev A, Kirkegaard RH, von Bergen M, Rattei T, Bendinger B, Nielsen 492 PH, Wagner M. 2015. Complete nitrification by Nitrospira bacteria. Nature 528:504–9.

493 12. Pinto AJ, Marcus DN, Ijaz UZ, Bautista-de lose Santos QM, Dick GJ, Raskin L. 2016. 494 Metagenomic Evidence for the Presence of Comammox Nitrospira -Like Bacteria in a 495 Drinking Water System. mSphere 1:e00054-15.

496 13. Gülay A, Tatari K, Musovic S, Mateiu R V., Albrechtsen H-J, Smets BF. 2014. Internal 497 porosity of mineral coating supports microbial activity in rapid sand filters for groundwater 498 treatment. Appl Environ Microbiol 80:7010–20.

499 14. Tatari K, Smets BF, Albrechtsen HJ. 2013. A novel bench-scale column assay to investigate 500 site-specific nitrification biokinetics in biological rapid sand filters. Water Res 47:6380– 501 6387.

502 15. Lee CO, Boe-Hansen R, Musovic S, Smets B, Albrechtsen H-J, Binning P. 2014. Effects of 503 dynamic operating conditions on nitrification in biological rapid sand filters for drinking 504 water treatment. Water Res 64:226–236.

505 16. Pjevac P, Schauberger C, Poghosyan L, Herbold CW, van Kessel MAHJ, Daebeler A, 506 Steinberger M, Jetten MSM, Lücker S, Wagner M, Daims H. 2017. AmoA-Targeted 507 Polymerase Chain Reaction Primers for the Specific Detection and Quantification of 508 Comammox Nitrospira in the Environment. Front Microbiol 8:1508. bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

509 17. Xia F, Wang J-G, Zhu T, Zou B, Rhee S-K, Quan Z-X. 2018. Ubiquity and Diversity of 510 Complete Ammonia Oxidizers (Comammox). Appl Environ Microbiol 84:1–14.

511 18. Hu H-W, He J-Z. 2017. Comammox—a newly discovered nitrification process in the 512 terrestrial nitrogen cycle. J Soils Sediments 17:2709–2717.

513 19. Poghosyan L, Koch H, Lavy A, Frank J, Kessel MAHJ, Jetten MSM, Banfield JF, Lücker S. 514 2019. Metagenomic recovery of two distinct comammox Nitrospira from the terrestrial 515 subsurface. Environ Microbiol 00:1462-2920.14691.

516 20. Bartelme RP, McLellan SL, Newton RJ. 2017. Freshwater Recirculating Aquaculture System 517 Operations Drive Biofilter Bacterial Community Shifts around a Stable Nitrifying 518 Consortium of Ammonia-Oxidizing Archaea and Comammox Nitrospira. Front Microbiol 519 8:101.

520 21. Wang M, Huang G, Zhao Z, Dang C, Liu W, Zheng M. 2018. Newly designed primer pair 521 revealed dominant and diverse comammox amoA gene in full-scale wastewater treatment 522 plants. Bioresour Technol 270:580–587.

523 22. Kits KD, Sedlacek CJ, Lebedeva E V., Han P, Bulaev A, Pjevac P, Daebeler A, Romano S, 524 Albertsen M, Stein LY, Daims H, Wagner M. 2017. Kinetic analysis of a complete nitrifier 525 reveals an oligotrophic lifestyle. Nature.

526 23. Palomo A, Pedersen AG, Fowler SJ, Dechesne A, Sicheritz-Pontén T, Smets BF. 2018. 527 Comparative genomics sheds light on niche differentiation and the evolutionary history of 528 comammox Nitrospira. ISME J 12:1779–1793.

529 24. Bédard C, Knowles R. 1989. Physiology, biochemistry, and specific inhibitors of CH4, 530 NH4+, and CO oxidation by methanotrophs and nitrifiers. Microbiol Rev 53:68–84.

531 25. Gülay A, Musovic S, Albrechtsen H-JH-JH-J, Smets BF. 2013. Neutrophilic iron-oxidizing 532 bacteria: occurrence and relevance in biological drinking water treatment. Water Sci Technol 533 Water Supply 13:1295.

534 26. Tatari K, Gülay A, Thamdrup B, Albrechtsen HJ, Smets BF. 2017. Challenges in using 535 allylthiourea and chlorate as specific nitrification inhibitors. Chemosphere 182:301–305.

536 27. Hall P, Aller R. 1992. Rapid, small-volume, flow-injection analysis for sigma-CO2 and 537 NH4+ in marine and fresh-waters. Limnol Oceanogr 37:1113–1119.

538 28. Vandieken V, Pester M, Finke N, Hyun J-H, Friedrich MW, Loy A, Thamdrup B. 2012. bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

539 Three manganese oxide-rich marine sediments harbor similar communities of acetate- 540 oxidizing manganese-reducing bacteria. ISME J 6:2078–90.

541 29. Whiteley AS, Thomson B, Lueders T, Manefield M. 2007. RNA stable-isotope probing. Nat 542 Protoc 2:838–844.

543 30. Neufeld JD, Vohra J, Dumont MG, Lueders T, Manefield M, Friedrich MW, Murrell JC. 544 2007. DNA stable-isotope probing. Nat Protoc 2:860–866.

545 31. Yu Y, Lee C, Kim J, Hwang S. 2005. Group-specific primer and probe sets to detect 546 methanogenic communities using quantitative real-time polymerase chain reaction. 547 Biotechnol Bioeng 89:670–9.

548 32. Quince C, Lanzen A, Davenport RJ, Turnbaugh PJ. 2011. Removing noise from 549 pyrosequenced amplicons. BMC Bioinformatics 12:38.

550 33. Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R. 2011. UCHIME improves sensitivity 551 and speed of chimera detection. Bioinformatics 27:2194–200.

552 34. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister EB, Lesniewski R a, 553 Oakley BB, Parks DH, Robinson CJ, Sahl JW, Stres B, Thallinger GG, Van Horn DJ, Weber 554 CF. 2009. Introducing mothur: open-source, platform-independent, community-supported 555 software for describing and comparing microbial communities. Appl Environ Microbiol 556 75:7537–41.

557 35. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, Fierer N, 558 Peña AG, Goodrich JK, Gordon JI, Huttley GA, Kelley ST, Knights D, Koenig JE, Ley RE, 559 Lozupone CA, Mcdonald D, Muegge BD, Pirrung M, Reeder J, Sevinsky JR, Turnbaugh PJ, 560 Walters WA, Widmann J, Yatsunenko T, Zaneveld J, Knight R. 2010. QIIME allows 561 analysis of high- throughput community sequencing data. Nat Methods 7:335–336.

562 36. Zemb O, Lee M, Gutierrez-Zamora ML, Hamelin J, Coupland K, Hazrin-Chong NH, Taleb I, 563 Manefield M. 2012. Improvement of RNA-SIP by pyrosequencing to identify putative 4-n- 564 nonylphenol degraders in activated sludge. Water Res 46:601–610.

565 37. Ginestet P, Audic J-M, Urbain V, Block J-C. 1998. Estimation of nitrifying bacterial 566 activities by measuring oxygen uptake in the presence of the metabolic inhibitors 567 allylthiourea and azide. Appl Environ Microbiol 64:2266–2268.

568 38. Taylor AE, Zeglin LH, Dooley S, Myrold DD, Bottomley PJ. 2010. Evidence for different 569 contributions of archaea and bacteria to the ammonia-oxidizing potential of diverse oregon bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

570 soils. Appl Environ Microbiol 76:7691–7698.

571 39. SurmaczGorska J, Gernaey K, Demuynck C, Vanrolleghem P, Verstraete W, Surmacz- 572 Gorska J, Gernaey K, Demuynck C, Vanrolleghem P, Verstraete W. 1996. Nitrification 573 monitoring in activated sludge by oxygen uptake rate (OUR) measurements. Water Res 574 30:1228–1236.

575 40. Belser LW, Mays EL. 1980. Specific inhibition of nitrite oxidation by chlorate and its use in 576 assessing nitrification in soils and sediments. Appl Environ Microbiol 39:505–510.

577 41. Yarza P, Yilmaz P, Pruesse E, Glöckner FO, Ludwig W, Schleifer K-H, Whitman WB, 578 Euzéby J, Amann R, Rosselló-Móra R. 2014. Uniting the classification of cultured and 579 uncultured bacteria and archaea using 16S rRNA gene sequences. Nat Rev Microbiol 580 12:635–645.

581 42. Wu Y, Ke X, Hernández M, Wang B, Dumont MG, Jia Z, Conrad R. 2013. Autotrophic 582 growth of bacterial and archaeal ammonia oxidizers in freshwater sediment microcosms 583 incubated at different temperatures. Appl Environ Microbiol 79:3076–3084.

584 43. Freitag TE, Chang L, Prosser JI. 2006. Changes in the community structure and activity of 585 betaproteobacterial ammonia-oxidizing sediment bacteria along a freshwater-marine 586 gradient. Environ Microbiol 8:684–696.

587 44. Seyler LM, McGuinness LM, Kerkhof LJ. 2014. Crenarchaeal heterotrophy in salt marsh 588 sediments. ISME J 1–10.

589 45. Zhang L-M, Hu H-W, Shen J-P, He J-Z. 2012. Ammonia-oxidizing archaea have more 590 important role than ammonia-oxidizing bacteria in ammonia oxidation of strongly acidic 591 soils. ISME J 6:1032–45.

592 46. Zhang L-M, Offre PR, He J-Z, Verhamme DT, Nicol GW, Prosser JI. 2010. Autotrophic 593 ammonia oxidation by soil thaumarchaea. Proc Natl Acad Sci U S A 107:17240–17245.

594 47. Xia WW, Zhang CX, Zeng XW, Feng YZ, Weng JH, Lin XG, Zhu JG, Xiong ZQ, Xu J, Cai 595 ZC, Jia ZJ. 2011. Autotrophic growth of nitrifying community in an agricultural soil. Isme J 596 5:1226–1236.

597 48. Pratscher J, Dumont MG, Conrad R. 2011. Ammonia oxidation coupled to CO2 fixation by 598 archaea and bacteria in an agricultural soil. Proc Natl Acad Sci U S A 108:4170–5.

599 49. Moreno AM, Matz C, Kjelleberg S, Manefield M. 2010. Identification of ciliate grazers of bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

600 autotrophic bacteria in ammonia-oxidizing activated sludge by RNA stable isotope probing. 601 Appl Environ Microbiol 76:2203–2211.

602 50. Dumont MG, Pommerenke B, Casper P, Conrad R. 2011. DNA-, rRNA- and mRNA-based 603 stable isotope probing of aerobic methanotrophs in lake sediment. Environ Microbiol 604 13:1153–67.

605 51. Roots P, Wang Y, Rosenthal AF, Griffin JS, Sabba F, Petrovich M, Yang F, Kozak JA, 606 Zhang H, Wells GF. 2019. Comammox Nitrospira are the dominant ammonia oxidizers in a 607 mainstream low dissolved oxygen nitrification reactor. Water Res 396–405.

608 52. Zheng M, Wang M, Zhao Z, Zhou N, He S, Liu S, Wang J, Wang X. 2019. Transcriptional 609 activity and diversity of comammox bacteria as a previously overlooked ammonia oxidizing 610 prokaryote in full-scale wastewater treatment plants. Sci Total Environ 656:717–722.

611 53. Pinto AJ, Marcus DN, Ijaz Z, Bautista-de los Santos QM, Dick GJ, Raskin L. 2015. 612 Metagenomic Evidence for the Presence of Comammox Nitrospira-Like Bacteria in a 613 Drinking Water System. mSphere 1:e00054-15.

614 54. Straat PA, Nason A. 1964. Characterization of a Nitrite Reductase from the Chemoautotroph 615 Nitrobacter agilis. J Biol Chem 240:1412–1426.

616 55. Hynes RK, Knowles R. 1983. Inhibition of Chemoautotrophic Nitrification by Sodium 617 Chlorate and Sodium Chlorite : a Reexamination. Appl Environ Microbiol 45:1178–1182.

618 56. Kits KD, Jung MY, Vierheilig J, Pjevac P, Sedlacek CJ, Liu S, Herbold C, Stein LY, Richter 619 A, Wissel H, Brüggemann N, Wagner M, Daims H. 2019. Low yield and abiotic origin of N 620 2 O formed by the complete nitrifier Nitrospira inopinata. Nat Commun 10:1–12.

621 57. Santoro AE, Casciotti KL. 2011. Enrichment and characterization of ammonia-oxidizing 622 archaea from the open ocean: phylogeny, physiology and stable isotope fractionation. ISME J 623 5:1796–808.

624 58. Tatari K, Musovic S, Gülay A, Dechesne A, Albrechtsen H-J, Smets BF. 2017. Density and 625 distribution of nitrifying guilds in rapid sand filters for drinking water production: 626 Dominance of Nitrospira spp. Water Res 127:239–248.

627 59. Tatari K, Smets BF, Albrechtsen H-J. 2016. Depth investigation of rapid sand filters for 628 drinking water production reveals strong stratification in nitrification biokinetic behavior. 629 Water Res 101:402–410. bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

630 60. Leininger S, Urich T, Schloter M, Schwark L, Qi J, Nicol GW, Prosser JI, Schuster SC, 631 Schleper C. 2006. Archaea predominate among ammonia-oxidizing prokaryotes in soils. 632 Nature 442:806–809.

633 61. Verhamme DT, Prosser JI, Nicol GW. 2011. Ammonia concentration determines differential 634 growth of ammonia-oxidising archaea and bacteria in soil microcosms. ISME J 5:1067–1071.

635 62. Wuchter C, Abbas B, Coolen MJ, Herfort L, van Bleijswijk J, Timmers P, Strous M, Teira E, 636 Herndl GJ, Middelburg JJ, Schouten S, Sinninghe Damste JS. 2006. Archaeal nitrification in 637 the ocean. Proc Natl Acad Sci U S A 103:12317–12322.

638 63. Feisthauer S, Wick LY, Kästner M, Kaschabek SR, Schlömann M, Richnow HH. 2008. 639 Differences of heterotrophic 13CO2 assimilation by Pseudomonas knackmussii strain B13 640 and Rhodococcus opacus 1CP and potential impact on biomarker stable isotope probing. 641 Environ Microbiol 10:1641–1651.

642 64. Hesselsoe M, Nielsen JL, Roslev P, Nielsen PH. 2005. Isotope labeling and 643 microautoradiography of active heterotrophic bacteria on the basis of assimilation of 644 14CO(2). Appl Env Microbiol 71:646–655.

645 65. Burnap RL, Vermaas WFJ. 2012. Functional Genomics and Evolution of Photosynthetic 646 Systems. Springer Netherlands, Dordrecht.

647 66. Roslev P, Larsen MB, Jørgensen D, Hesselsoe M. 2004. Use of heterotrophic CO2 648 assimilation as a measure of metabolic activity in planktonic and sessile bacteria. J Microbiol 649 Methods 59:381–93.

650 67. Okabe S, Kindaichi T, Ito T. 2005. Fate of C-14-labeled microbial products derived from 651 nitrifying bacteria in autotrophic nitrifying biofilms. Appl Environ Microbiol 71:3987–3994.

652 68. Daum M, Zimmer W, Papen H, Kloos K, Nawrath K, Bothe H. 1998. Physiological and 653 molecular biological characterization of ammonia oxidation of the heterotrophic nitrifier 654 Pseudomonas putida. Curr Microbiol 37:281–8.

655 69. EMBL-EBI. 2019. AbrB duplication(IPR017516). Available 656 at: https://www.ebi.ac.uk/interpro/entry/IPR017516 (Accessed: 04 July 2019).

657 70. Stein LY. 2011. Heterotrophic nitrification and nitrifier denitrification, p. 95–114. In Ward, 658 BB, Arp, DJ, Klotz, MG (eds.), Nitrification2. ASM Press, Washington D. C.

659 71. Abeliovich A. 2006. The Nitrite Oxidizing Bacteria, p. 861–872. In Balows, A, Trüper, GG, 660 Dworkin, M, Harder, W, Schleifer, K-H (eds.), The Prokaryotes, 2nd ed. Springer New York, bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

661 New York, NY.

662 72. Sorokin DY, Lücker S, Vejmelkova D, Kostrikina NA, Kleerebezem R, Rijpstra WIC, 663 Damsté JSS, Le Paslier D, Muyzer G, Wagner M, van Loosdrecht MCM, Daims H. 2012. 664 Nitrification expanded: discovery, physiology and genomics of a nitrite-oxidizing bacterium 665 from the phylum Chloroflexi. ISME J 6:2245–56.

666 73. Alawi M, Lipski A, Sanders T, Spieck E. 2007. Cultivation of a novel cold-adapted nitrite 667 oxidizing betaproteobacterium from the Siberian Arctic. ISME J 1:256–264.

668 bioRxiv preprint doi: https://doi.org/10.1101/703868; this version posted July 16, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.