Community Structure and In Situ Activity of Nitrifying Bacteria in Phragmites Root-Associated Biofilms

The amount of oxygen released by Phragmites roots and the community structure and in situ activity of nitrifying bacteria in the root biofilms were analyzed by the combined use of 16S rRNA gene-cloning analysis, quantitative PCR (qPCR) assay and microelectrodes. Axial and radial O2 microprofiles were obtained for individual roots of Phragmites in a horizontal flow reactor fed with artificial medium continuously. Axial O2 profiles revealed that O2 was released at a rate of 0.21 μmol O2 cm−2 (root surface area) h−1 only in the apical region (up to ca. 40 mm from the root apex), where there was a high abundance (107 to 108 copies g−1 biomass) of Nitrosomonas-like AOB and Nitrospira-like NOB. This abundance, however, sharply declined to the detection limit at positions more basal than 80 mm. Phylogenetic analysis based on 16S rRNA gene identified strains related to Nitrosomonas oligotropha and Nitrosomonas cryotolerans as the predominant AOB and strains related to Nitrospira marina and Nitrospira moscoviensis as the predominant NOB in the root biofilms. Based on radial O2 microprofiles, the oxic region only extended about 0.5 mm into the surrounding sediment due to a high rate of O2 consumption in the rhizosphere. The net NH4+ and O2 consumption rates in the apical region were higher than those determined at the oxic sediment surface in which the abundance of AOB and NOB was one order of magnitude lower than in the rhizosphere. These results clearly indicated that Phragmites root biofilms played an important role in nitrification in the waterlogged anoxic sediment.

Nitrification, the microbially mediated oxidation of NH4 + to NO3 − , plays a central role in the N-cycle of aquatic environments, especially with respect to the growing eutrophication of aquatic environments. Since waterlogged environments (e.g., river banks and sediments) become anoxic almost immediately beneath the sediment-water interface, nitrification is expected to occur only in narrow surface oxic zones (44). It is, however, recognized that the immediate root surface of emergent plants, such as Phragmites, offers a supply of oxygen through leakage from aerenchymatous tissue (3,6,10), which may promote nitrification. Oxygenreleasing plants might therefore have a significant impact on nitrogen cycling (i.e., nitrification) in otherwise anoxic sediment layers. Plant-induced nitrification can subsequently lead to enhanced denitrification, thus increasing nitrogen removal from aquatic environments.
Despite the longstanding realization that emergent plant roots release oxygen and affect microbial communities in the immediate rhizosphere, the relationship between oxygen release from roots and structure and in situ activity of nitrifying bacteria in the rhizosphere is not fully understood because of difficulty in the direct measurement of nitrifying activity and oxygen release from plant roots and the limitations of traditional culture-dependent techniques. There appears to have been no published work reporting combined information about the diversity, abundance, and fine-scale distribution of the in situ activity of nitrifying bacteria in the immediate rhizosphere in aquatic sediments. Therefore, it is imperative to directly quantify the available oxygen concentration and the in situ activity and abundance of nitrifying bacteria in the immediate rhizosphere.
Cultivation-independent molecular techniques have provided better understanding of the diversity and distribution of nitrifying bacteria in various aquatic environments, in which nitrifying bacteria were usually found in small numbers (1,8,28,43). Recently, Briones et al. (7) analyzed ammoniaoxidizing bacterial (AOB) communities in the rhizosphere of rice by using 16S rRNA gene-based denaturing gradient gel electrophoresis (DGGE) and fluorescence in situ hybridization (FISH) and found that the detection frequencies of Nitrosomonas and Nitrosospira varied with plant variety. The diversity and abundance of nitrite-oxidizing bacteria (NOB) in freshwater sediments and rhizospheres were, however, neglected in most of the studies. The molecular approach has been applied successfully in the rhizospheres (14,35,46) and phytosphere (41,50) of various plants, demonstrating plant-dependent enrichment or the diversity of microorganisms. In addition, the combination of molecular techniques with microelectrodes has been proven to be a powerful tool to analyze the structure and in situ activity of microbial communities in freshwater sediments (1,28,44), biofilms (21,(31)(32)(33)(34)45), microbial granules (43), and rhizospheres (7).
The goal of this study was to investigate the relationship between oxygen release from the Phragmites roots and structure and in situ activity of nitrifying bacterial populations in the root biofilms. In this study, we directly determined the microprofiles of substrates (O 2 , NH 4 + , NO 3 − and pH) in the rhizosphere of Phragmites by using microelectrodes to quantify radial oxygen release and the in situ activity of nitrifying bacteria. We also applied 16S rRNA gene-based techniques (16S rRNA gene-cloning and quantitative PCR (qPCR)) to analyze the community structure and abundance of nitrifying bacteria (both AOB and NOB) in the root biofilms.

Plant material and reactors
Phragmites communis with horizontal rhizomes and roots was carefully collected from the Sosei River, Sapporo, Japan, which receives treated domestic wastewater. A horizontal flow reactor (20 [L]×5 [W]×5 [H] cm) was constructed using acrylic boards. Total working volume was 400 cm 3 . Approximately 20 cm tall Phragmites with roots 10-15 cm long were selected, transferred, and positioned in the reactor, as shown in Fig. 1. Some roots were gently but firmly fixed onto supporting material using toothpicks. The roots were positioned horizontally ca. 5 mm below the surface of the sediment and close to the acrylic sidewall for microelectrode measurements. The positions of the root apex were marked with inset needles. The reactor was carefully filled with fine silt (after passing through a 2 mm sieve), which was taken from the sampling site of Phragmites communis.

Microelectrode measurements
Steady-state concentration profiles of O2, NH4 + , NO3 − , and pH in the rhizosphere were measured using microelectrodes (6,40). A Clark-type O2 microelectrode was prepared and calibrated as previously described (39). LIX-type microelectrodes for NH4 + , NO3 − , and pH were constructed, calibrated, and used according to the protocol described by Okabe et al. (31,32). The reactor was then acclimated in the light condition for at least a few hours before measurements to ensure that steady-state profiles were obtained.
The microelectrodes that were attached to a motor-driven micromanipulator were positioned manually near the root, and then the micromanipulator was operated in all three directions for fine positioning. We first ensured that the microelectrode tip touched the surface of the root at a given distance from the apex by monitoring changes in the readings carefully (the reading jumped when the microelectrode tip touched the root surface). The position of the microelectrode was defined as the starting point (0, 0) of the microprofiles. The concentration profiles in the rhizosphere were then measured in steps of 50 μm from the root surface to the bulk sediment or the overlaying water. The sediment surface was determined by using a dissection microscope model Stemi 2000 (Carl Zeiss). For radial profile measurements, the microelectrodes were positioned vertically above the root so that the microelectrodes penetrated the center of the roots. Microprofiles were also measured axially along the root surface. To ensure reproducibility of the measurements, the microprofiles were measured again by tracing the original path toward the root surface. Both of the results were usually in very close agreement; therefore, the concentration profile data reported in this paper were obtained as the microelectrodes were moved away from the root surface.
Based on the concentration profiles measured, the total O2, NH4 + and NO3 − consumption rates in the rhizosphere were calculated using Fick's first law of diffusion (32). The molecular diffusion coefficients used for the calculations were 2.06×10 −5 cm 2 s −1 for O2, 1.93×10 −5 cm 2 s −1 for NH4 + , and 1.89×10 −5 cm 2 s −1 for NO3 − in water at 20°C (11). The radial oxygen loss (ROL, [mole O2 cm −2 root surface h −1 ]) from a root to the rhizosphere was calculated from the gradient of the O2 microprofile, as described elsewhere (4). The plant oxygen release in the horizontal flow reactor and the field was also estimated as described previously (6). The active root surface area was calculated based on the root diameter and the active root length (i.e., root length over which ROL could be measured). The number of roots per unit sediment surface area was directly measured in the horizontal flow reactor and the field.

DNA extraction and PCR amplification
After microelectrode measurements, the Phragmites were removed from the horizontal flow reactor. The roots were shaken gently to remove soil that loosely adhered to the roots. We found that the root surface was covered with brown thin biofilm (usually about 100-μm thick) (Fig. S1). The biofilm firmly attached to the roots was subjected to the following molecular analyses of microbial community structure.
All selected roots were cut into 5-mm fragments from the root apex with a sterilized razor. The fragments were then grouped into subsamples: 0-5 mm, 5-10 mm, 10-15 mm, 15-20 mm, 20-25 mm, 25-30 mm, 30-35 mm, 35-40 mm, 45-55 mm, 55-65 mm, 65-75 mm, 75-85 mm, and 155-165 mm. Surface sediment samples were also included in the analysis for comparison. Prior to DNA extraction, for each root fragment, 3 g roots with firmly adhering biofilm were placed in 9 mL pure water and vortexed for 3 min. The roots were removed, placed in 9 mL pure water again, and treated in the same way. This treatment was repeated once more. The resulting supernatants of three treatments were combined. Biomass pellets were harvested by centrifugation (10,000×g, 30 min) and kept at −80°C until DNA extraction.
Total DNA was extracted from each biomass pellet using a Fast DNA spin kit (MP Biomedicals, Tokyo Japan) as described in the manufacturer's instructions. The extracted DNA samples (from 0 mm to 85 mm) were combined for the following 16S rRNA genecloning analysis for betaproteobacterial ammonia-oxidizing bacteria (AOB), nitrite-oxidizing bacteria (NOB), and ammonia-oxidizing archaea (AOA) communities. To reduce the possible bias caused by PCR amplification, the 16S rRNA gene in the combined subsamples was amplified in triplicate tubes and then combined for the next cloning step. The 16S rRNA gene fragments from the combined total DNA were amplified with Taq DNA polymerase (TaKaRa Bio, Ohtsu, Japan) using previously described protocols and the AOB specific primer set of CTO189f and CTO654r (24), the Nitrospira-like NOB specific primer sets of NTSPAf (28) -universal 1492r (51) and Ntspa685 (20) -NTSPAr (28), and the archaeal specific primer set of Arch21F (13) -1492r (51). Comparative investigation of the microbial community other than AOB and NOB in the root biofilm and bulk sediment was conducted using the universal primer set of 8F-1492r (51). All PCRs were performed for 35 cycles using the PCR conditions previously reported for each primer set. PCR products were checked by electrophoresis on a 1% (w/v) agarose gel. In addition, since the archaeal 16S rRNA gene was not amplified from the root biofilms, the archaeal amoA gene was also amplified using previously described protocols and the primer set of Arch-AmoAF/Arch-AmoAR (16); however, the archaeal amoA gene was also not amplified probably due to low abundance of AOA in the root biofilms.
Cloning and sequencing of the 16S rRNA gene and phylogenetic analysis PCR products were ligated into a pCR-XL-TOPO vector and transformed into ONE SHOT Escherichia coli cells (TOPO XL PCR cloning; Invitrogen) following the manufacturer's instructions. Partial sequencing of the 16S rRNA gene inserts (about 470 bp for AOB and 410 bp and 500 bp for Nitrospira-like NOB) was performed using an automatic sequencer (ABI Prism 3100 Avant Genetic Analyzer; Applied Biosystems) with a BigDye Terminator Ready Reaction kit (Applied Biosystems). All sequences were checked for chimeric artifacts by the CHECK_CHIMERA program in the Ribosomal Database Project (25) and compared with similar sequences of the reference organisms by a BLAST search (2). Sequence data were aligned with the CLUSTAL W package (49). Clone sequences with 97% or greater similarity were grouped into the same operational taxonomic unit (OTU) by a SIMILARITY_MATRIX program, and their representative sequences were used for phylogenetic analysis. Phylogenetic trees were constructed using the neighbor-joining method (42). Tree topology was also tested using the maximum-parsimony method. Bootstrap resampling analysis of 1,000 replicates was performed to estimate the confidence of the tree topologies.

Quantification of AOB and NOB by qPCR
Quantitative PCR (qPCR) assays were performed to quantify copy numbers of betaproteobacterial AOB and Nitrospira-like and Nitrobacter-like NOB-specific 16S rRNA genes in root biofilms at different locations along the root (as mentioned above). Since no PCR products were obtained using the archaeal 16S rRNA gene and archaeal amoA gene specific primer sets, their quantification was not performed in this study. The 16S rRNA genes of betaproteobacterial AOB were quantified using the primer set of CTO 189fA/B/C and RT1r and TaqMan probe TMP1, as previously described by Hermansson and Lindgren (18). The primers CTO 189fA/B and CTO 189fC were used at a 2:1 ratio. Standardization was performed with a dilution series (10 to 10 8 copies per well) of genomic DNA of Nitrosomonas europaea (NBRC 14298). The amplification efficiency was calculated to be 87-91% from the equation, εc=10 −1/S −1. The 16S rRNA genes of Nitrospira-like NOB were quantified using the primer set of NTSPAf and NTSPAr and TaqMan probe NTSPATaq, as described by Nakamura et al. (28). Standardization was performed with a dilution series (10 to 10 7 copies per well) of the standard vector plasmid carrying ca. 400 bp of 16S rRNA gene of uncultured bacterium A-4 (AF033559) related to Candidatus Nitrospira defluvii (28). The amplification efficiency was 91 to 95%.
The Nitrobacter-like NOB 16S rRNA gene was quantified as previously described by Degrange and Bardin (12) in a total volume of 25 μL with SYBR Green PCR master mix (PE Applied Biosystems), 300 nM of each of the forward primer FGPS872f and reverse primer FGPS1269r (12), 100 nM of bovine serum albumin (Sigma) and either 0.1 pg sample DNA or 10 to 10 5 copies per well of the standard genomic DNA of Nitrobacter winogradskyi (NBRC 14297). The specificities of PCR products were confirmed by melting curve analysis. qPCR assays for AOB, Nitrospira-like NOB, and Nitrobacter-like NOB were performed in MicroAmp Optical 96-well reaction plates with an optical cap (PE Applied Biosystems). The template DNA in the reaction mixtures was amplified and monitored with an ABI prism 7000 Sequence Detection System (PE Applied Biosystems). The detection limits for AOB, Nitrospira-like NOB and Nitrobacter-like NOB in this study were 2.7×10, 1.6×10 2 , and 5.4×10 copies per well, respectively, which corresponded to 6.7×10 4 , 4.0×10 4 , and 1.4×10 4 copies/cm 3 when the sample volume and DNA extraction step were taken into account.

Analytical methods
Total bacterial cell counts in root biofilms taken at different locations along the root (as mentioned above) and in surface sediment samples were also performed after the diluted sediment samples on 0.2-μm-membrane filters were stained with 6-diamidino-2-phenylindole (DAPI). At least 20 replicate analyses were performed for each sample. The NH4 + , NO2 − and NO3 − concentrations in the influent and effluent of the reactors were determined using an ion chromatograph (HIC-6A; Shimadzu) equipped with a Shimpack CS16 column and CS6A. The samples for NH4 + , NO2 − and NO3 − were filtered through 0.2-μm pore membrane filters (DISMIC-13CP; Advantec) before analysis. The O2 concentration and pH in the overlaying water were determined using an O2 and a pH electrode, respectively. Light intensity was measured by a quantum meter (Fujiwara Scientific Company, Tokyo Japan), which senses only in the 400-700 nm region. All measurements were made above the water surface.

Nucleotide sequence accession numbers
The GenBank/EMBL/DDBJ accession numbers for the 16S rRNA gene sequences of the clones used for phylogenetic tree analysis are AB255039 to AB255048 and AB373105 to AB373109.

Water quality in the horizontal flow reactors
The Phragmites were cultured for 8 weeks under 12-h light and dark cycles in the laboratory at 20°C. Steady-state conditions were attained after about one month. NH4 + concentration in the effluent decreased along with pH decline while NO3 − concentration increased in the horizontal flow reactors with and without Phragmites, indicating the occurrence of nitrification. The net NH4 + consumption rate in the reactor with Phragmites was 0.039±0.016 µmol/cm 2 /h (average±SD), which was higher than that obtained in the reactor without the plant (0.020±0.007 µmol/cm 2 /h) (p<0.05) ( Table 1).

O2 concentration profiles in the rhizosphere
Changes in O2 concentrations within the cortex and at the root surface with distance along an intact 100-mm long Phragmites root are shown in Fig. 2. Microelectrode measurements were made five times, and the average is presented. The cortical oxygen concentration gradually increased toward the root base. In contrast, the O2 concentration at the root surface was highest (62 µM) at ca. 25 mm from the root apex and declined markedly at positions more basal than 25 mm, indicating the overall decline in radial oxygen loss (ROL). We could not insert microelectrodes into the root base (more basal than 30 mm from the apex) due to high penetration resistance. Figure 3 shows a two-dimensional O2 concentration profile around the 100-mm long Phragmites root, which was immersed about 5 mm below the sediment surface. The root cross-sectional position corresponded to ca. 25 mm from the root apex. The root diameter was approximately 1 mm. O2 penetration depth was about 2 mm from the sediment surface, below which anoxic conditions prevailed. More than 150 µM O2 was detected inside the root, and O2 diffused and radiated out from the root surface, creating an oxic microenvironment around the root. Figure 4 indicates axial distributions of the total bac-terial cell counts and the specific 16S rRNA gene copy numbers of the betaproteobacterial AOB and Nitrospiraand Nitrobacter-like NOB in the biofilm sampled along the root. The total bacterial cell numbers gradually increased from 8.9±2.5×10 7 cells g −1 biomass near the root apex to 3.5±0.8×10 9 cells g −1 biomass at 160 mm from the root apex. The betaproteobacterial AOB specific 16S rRNA gene copy number was highest (2.3±1.8×10 8 copies g −1 biomass) at ca. 20 mm from the root apex. Nitrospira-like NOB-specific 16S rRNA gene copy numbers were evenly distributed (10 7 to 10 8 copies g −1 biomass) from the root apex to 40 mm. The copy numbers of both AOB and Nitrospira-like NOB decreased markedly between 40 mm and 80 mm from the root apex and were under the detection limit (6.7×10 4 and 4.0×10 4 copies g −1 biomass, respectively) at 80 mm from the root apex. The Nitrobacter-like NOB-specific 16S rRNA gene copy numbers were one to two orders of magnitude lower (10 5 to 10 6 copies g −1 biomass) than those of Nitrospira-like NOB. The total bacterial cell count at the oxic sediment surface was 4.7±0.5×10 9 cells g −1 sediment, which was comparable with the root biofilm. Both archaeal amoA gene and 16S rRNA gene were not amplified by PCR from the combined DNA samples (from 0 mm to 85 mm of the root), suggesting that the abundance of AOA in the root biofilm was under the detection limit.

Community structure of AOB and NOB in the root biofilm
The 16S rRNA gene clone libraries were constructed from the biofilm attached to the root surface to investigate the betaproteobacterial AOB and Nitrospira-like NOB community structures. Forty-four clones were randomly selected from the betaproteobacterial AOB clone library, and the partial sequences of approximately 470 bp were analyzed. In total, 20 clone sequences were affiliated with the betaproteobacterial AOB and were grouped into 5 OTUs based on more than 97% sequence similarity among the OTUs. Their representative sequences were used for phylogenetic analysis (Fig. S2). We classified the betaproteobacterial AOB into seven stable lineages according to Purkhold et al. (38). The distribution of the 20 clones affiliated with known betaproteobacterial AOB is shown in Table 2. Oxygen concentrations within the cortex ( ) and at the root surface ( ) along roots. The roots were immersed in anoxic sediment; the shoot was exposed to air.  Twelve clones were related to the Nitrosomonas oligotropha lineage with 98.5-99.1% sequence similarity, seven clones were related to the Nitrosomonas cryotolerans lineage with 95.7-96.3% sequence similarity, and one clone was related to the Nitrosomonas briensis lineage with 95.7% sequence similarity.
Another 16S rRNA gene clone library was constructed from the root biofilm for the Nitrospira-like NOB community. Partial sequences of approximately 410 bp and 500 bp were analyzed from 47 and 23 clones retrieved using primer sets of NTSPAf-universal 1492r and Ntspa685-NTSPAr, respectively (Fig. S3). The distribution of all clones retrieved from the root biofilm using two different sets of PCR primers is shown in Table 3. The clones were mainly affiliated with the Nitrospira moscoviensis and Nitrospira marina lineages with more than 96.5% sequence similarity.
Microbial communities other than AOB and NOB in the root biofilm and bulk sediment were compared, showing higher microbial diversity in the root biofilm than in the bulk sediment (Table S1).

Concentration profiles in the sediment
The representative steady-state concentration profiles of O2, NH4 + , NO3 − and pH in the sediment and rhizosphere of a Phragmites root at a position 25 mm from the root apex were measured with microelectrodes (Fig. 5A). O2 concentration was ca. 55 µM at the sediment surface (at 0 mm) and completely depleted at a depth of 1.1 mm from the sediment surface; however, O2 was detected again at ca. 0.5 mm above the root surface (at 1.5 mm from the sediment surface) and the O 2 concentration increased up to ca. 80 µM at the root surface. The maximum O2 concentration inside the root was about 150 µM. NH4 + concentration decreased in the upper 0.5 mm of the oxic sediment surface whereas it was unchanged in the anoxic zone of the sediment (ca. 0.5 mm to 1.5 mm from the sediment surface). NH4 + concentration decreased again in the rhizosphere. NO 3 − concentration gradually increased toward the depth of the sediment. pH slightly decreased in the rhizosphere.
Similarly, concentration profiles of O2, NH4 + , and NO3 − were also measured at points ca. 5, 20, and 35 mm from the root apex, and net consumption and production rates of O2, NH4 + , and NO3 − are summarized in Table 4. Net NH4 + and O 2 consumption rates were high 20-25 mm from the root apex, which is consistent with the radial oxygen loss (ROL) ( Figure 2) and abundance of AOB and NOB (Figure 4) along the root. We could not detect significant NH 4 + consumption 50 mm from the root apex (data not shown).

Radial oxygen release from roots
Several studies have shown that emergent plants such as Phragmites, Scirpus validus, and rice released O2 from their roots (4-7, 10, 40). Several analytical methods have been used to evaluate the radial oxygen release from roots. Among which, measurement using O2 microelectrodes provides high spatial (several 10 micrometers) and temporal resolution (4,6,7,40). We demonstrated that O2 release was high in the apical region (up to ca. 40 mm from the root apex) and declined basipetally. The O2 release sufficiently supported the abundance and in situ activity of nitrifying bacteria in the biofilm attached to the roots of Phragmites.
Based on the O2 microprofiles that were determined in the anoxic sediment, the average radial oxygen loss (ROL) was determined to be 0.21 µmol O 2 cm −2 root surface h −1 in this study, which was comparable with the previously reported values for Phragmites (0.15 µmol O2 cm −2 root surface h −1 at 25 mm from the root apex (5) and 0.11 µmol O 2 cm −2 -root surface h −1 at 7 mm from the root apex (4) and for Scirpus  (4)(5)(6). Plant O2 release is highly dependent upon the plant (species, age, and health), sediment conditions (e.g., oxidation reduction potential (ORP), biological and chemical oxygen demand, availability of nutrients, and other factors), and microbial activity in the rhizosphere (4-6).

Abundance of nitrifying bacteria along the root
Very little is known about the in situ distribution of AOB, NOB, and AOA in the root biofilm. Cultivation-independent molecular analyses revealed that a mixed community of AOB and NOB was present in the biofilm attached to Phragmits communis roots and their levels of abundance were about one order of magnitude higher on a weight basis than the densities usually encountered in fresh water oxic sediment surfaces (28,44) (Fig. 4).
A high abundance (10 7 to 10 8 copies g −1 biomass) of Nitrosomonas-like AOB and Nitrospira-like NOB was present in the apical region (up to ca. 40 mm from the root apex). The higher abundance in the apical region could be partly explained by higher oxygen release from the roots, which consequently resulted in the promotion of nitrification rates determined with microelectrodes (Fig. 5). Although the phenomenon of oxygen release by Phragmites roots has long been recognized, such oxygen release has never been convincingly linked to the ecology of AOB and NOB in the root-associated biofilm. Assuming that AOB and NOB have one 16S rRNA gene copy on average, since N. europaea NBRC 14298 carries one 16S rRNA gene copy (22), the betaproteobacterial AOB and Nitrospira-like NOB cell densities were approximately 10 7 to 10 8 cells g −1 biomass in the apical region (up to 40 mm from the root apex) of the root biofilm (Fig. 4). These values were comparable with the abundance of Nitrosomonas spp. in the root biofilm of rice (5.0-6.8×10 7 cells cm −3 root) detected by FISH with an Nsm156 probe (7).
The 16S rRNA gene-cloning analysis for AOB shows that since the CTO primer used in this study were not completely specific to AOB of the Betaproteobacteria, several clone sequences were not affiliated with known AOB ( Table 2). None of the published primers that are intended to target all AOB of the Betaproteobacteria show both 100% sensitivity (targeting all AOB) and 100% specificity (excluding all non-AOB) (9,37). The same limitation inherent in PCR-based techniques also applies to the analysis of NOB in sediment and soil samples (Table 3). Thus, the possibility of underestimating and/or overestimating AOB-and NOB-specific 16S rRNA gene copy numbers cannot be excluded in this study; however, this does not negate the general trend that the abundance and activity of AOB and NOB was high in the apical region (up to ca. 40 mm from the root apex) due to high O 2 release.

Microbial community structure in the root biofilm
The results of 16S rRNA gene-cloning analysis revealed that the clones related to the Nitrosomonas oligotropha and Nitrosomonas cryotolerans lineages were numerically dominant in the root biofilms among AOB. N. oligotropha and N. cryotolerans have been isolated from freshwater and wastewater habitats and characterized by their ability to cope with low ammonia concentrations (i.e., K m (NH 3 )=ca. 3 µM) (23,36,47). Although no kinetic data with respect to oxygen are available for N. oligotropha-like AOB, it could be speculated that they also have high affinity toward oxygen (17). Briones et al. (7) have shown that the community structure of root-associated AOB populations was dependent on the microscale O2 availability in the rhizosphere. For the NOB population, the clones related to Nitrospira marina and Nitrospira moscoviensis lineages dominated in the clone library. Dominance of Nitrospira-like NOB in freshwater sediment has been reported by several previous studies (1,20,28,44). N. moscoviensis was originally isolated from the corroded iron pipe of a heating system (15). The N. moscoviensis lineage consists of uncultured bacterium clones retrieved from a non-chlorinated model drinking water distribution system (26) and from freshwater environmental metal-rich sediments from Green Bay (48). Inorganic and organic nitrogen concentrations in the overlaying water at the sampling site (n=25) were NH4 + =not detected, NO2 − =not detected, NO3 − =193±11 µM, and total nitrogen 212±16 µM, which might explain the nitrifying bacterial communities in the root biofilm in this study.
Comparative investigation of the microbial community in the root biofilm and bulk sediment clearly showed higher microbial diversity in the root biofilm (Table S1). This is probably because the roots excrete organic matter as well as oxygen. This result clearly indicated that the oxic microenvironment developed near the immediate root environment by O 2 leakage from the roots; otherwise, the anoxic bulk sediment prevailed.

Potential contribution of an emergent plant
Changes in NH4 + and/or NO3 − concentrations in the immediate root environment could be attributed to microbial activities (nitrification, denitrification, and/or assimilation) and plant uptake (Fig. 5); however, we could not evaluate the relative contribution of each process to the overall net consumption of NH 4 + and NO 3 − . Net NH 4 + and O 2 consumption rates at the root surface were 0.023±0.007 µmol NH4 + cm −2 h −1 and 0.42±0.06 µmol O2 cm −2 h −1 , which were both higher than those at the oxic sediment surface (0.016±0.006 µmol NH4 + cm −2 h −1 and 0.085±0.004 µmol O2 cm −2 h −1 ). As a consequence, approximately 8% and 28% of O 2 were used for ammonia oxidation at the root surface and the oxic sediment surface, respectively. This may suggest the occurrence of high competition for oxygen with heterotrophic bacteria, especially in the rhizosphere (30), because the roots excrete organic matter. Furthermore, oxygen could be used for oxidation of potential phytotoxins (e.g., Fe 2+ , Mn 2+ , and sulfide) (52). Total NO3 − production rates at the root surface and at the oxic sediment surface (0.016±0.016 µmol NO3 − cm −2 h −1 and 0.007±0.001 µmol NO 3 − cm −2 h −1 ) were about 70% (0.016/0.023) and 40% (0.007/0.016) of the total NH4 + consumption rates, respectively. This might suggest that a proportion of NO 3 − produced by nitrifying bacteria was taken up by the plant in the rhizosphere and/or simultaneously denitrified.
We evaluated the potential oxygen contribution of an emergent plant, Phragmites, to anoxic sediment at the sampling site (Sosei River, Sapporo, Japan). The amount of oxygen released by Phragmites roots per unit sediment surface area can be estimated based on the radial oxygen loss (ROL), the number and length of active roots, and the field root density, which was quantified at the sampling site. The total number of the roots was determined to be ca. 309±69 roots per 100 cm −2 -sediment surface. Assuming that only the 40 mm tip of a root of 0.5 mm average diameter can release O2 (Fig. 2), total O2 release (oxic) surface area of the root was then calculated to be 194 ± 44 cm 2 root surface per 100 cm −2 sediment surface, indicating that about 2 times more oxic root surface area exists in the sediment. Based on oxygen microprofiles, the average amount of O 2 released by the Phragmites, namely ROL, was calculated to be 0.21 µmol O2 cm −2 root surface h −1 (Fig. 2); therefore, we now can calculate the amount of oxygen released by Phragmites roots per 100 cm 2 of the sediment surface to be 0.21×194=40.7 µmol O2 per 100 cm −2 sediment surface h −1 , corresponding to 3.14 g-O 2 m −2 sediment surface d −1 . This value is within the range of the amount of oxygen released by wetland plants reported by various researchers (summarized by Bezbaruah and Zhang (6)). This result indicated that Phragmites roots played a significant role in the oxygen supply to freshwater sediments.
In conclusion, we successfully determined oxygen release by Phragmites roots using a oxygen microelectrode and then related this to the in situ abundance and activity of AOB and NOB in the rhizosphere environment by the combined use of 16S rRNA gene-cloning analysis, qPCR assay and microelectrodes. O 2 was released from the apical region (up to ca. 40 mm from the root apex) with a high abundance of nitrifying bacteria (10 7 to 10 8 copies g −1 -biomass). The total NH 4 + and O 2 consumption rates in the apical region were higher than those at the oxic sediment surface. These results suggest that nitrifying bacteria associated with Phragmites roots played an important role in NH 4 + consumption in the waterlogged anoxic sediment.