Corresponding author: L. Hou, State Key Laboratory of Estuarine and Coastal Research, East China Normal University, 3663 North Zhongshan Road, Shanghai 200062, China. (email@example.com)
 Anaerobic ammonium oxidation (anammox) as an important process of nitrogen cycle has been studied in estuarine environments. However, knowledge about the dynamics of anammox bacteria and their interactions with associated activity remains scarce in these environments. Here we report the anammox bacterial diversity, abundance, and activity in the Yangtze Estuary, using molecular and isotope-tracing techniques. The phylogenetic analysis of 16S rRNA indicated that high anammox bacterial diversity occurred in this estuary, including Scalindua, Brocadia, Kuenenia, and two novel clusters. The patterns of community composition and diversity of anammox bacteria differed across the estuary. Salinity was a key environmental factor defining the geographical distribution and diversity of the anammox bacterial community at the estuarine ecosystem. Temperature and organic carbon also had significant influences on anammox bacterial biodiversity. The abundance of anammox bacteria ranged from 2.63 × 106 and 1.56 × 107 gene copies g−1, and its spatiotemporal variations were related significantly to salinity, temperature, and nitrite content. The anammox activity was related to temperature, nitrite, and anammox bacterial abundance, with values of 0.94–6.61 nmol N g−1 h−1. The tight link between the anammox and denitrification processes implied that denitrifying bacteria may be a primary source of nitrite for the anammox bacteria in the estuarine marshes. On the basis of the 15N tracing experiments, the anammox process was estimated to contribute 6.6%–12.9% to the total nitrogen loss whereas the remainder was attributed to denitrification.
 Global nitrogen overload in aquatic ecosystems is an important environmental issue, due mainly to excessive input of anthropogenetic nitrogen [Seitzinger, 2008]. An improved understanding of nitrogen removal pathways and associated controlling mechanisms is required on local and global scales to develop strategies to protect the water quality and health of aquatic ecosystems [Gardner and McCarthy, 2009; Hou et al., 2012]. For decades, heterotrophic denitrification was considered as the only important process for bacterial removal of fixed nitrogen from aquatic environments into the air. However, this view has been changed since the discovery of anaerobic ammonium oxidation (anammox) which oxidizes ammonium via reducing nitrite into dinitrogen under anaerobic conditions. This process was observed first in a wastewater treatment plant in the Netherlands in 1995 [Mulder et al., 1995]. Subsequently, the anammox process was observed in various natural ecosystems including freshwater sediments [Penton et al., 2006; Zhang et al., 2007; Hu et al., 2012; Wang et al., 2012; Zhu et al., 2013], marine sediments [Thamdrup and Dalsgaard, 2002; Kuypers et al., 2003; Trimmer and Nicholls, 2009; Dang et al., 2010; Jetten et al., 2010], and anaerobic water columns [Schubert et al., 2006; Lam et al., 2007; Hamersley et al., 2009; Jensen et al., 2011]. At present, this process has been identified as a nonignorable pathway for removing nitrogen from these aquatic environments.
 The anammox process is mediated by chemolithoautotrophic bacteria affiliated with the phylum Planctomycetes [Jetten et al., 2010]. At present, five genera of anammox bacteria, including Candidatus Brocadia, Kuenenia, Scalindua, Anammoxoglobus, and Jettenia, were identified in different wastewater treatment systems [Strous et al., 1999; Schmid et al., 2000; 2003; Kartal et al., 2007b; Quan et al., 2008]. Anammox bacteria are also ubiquitous in natural environments [Schmid et al., 2007; Humbert et al., 2010]. So far, most studies of anammox bacteria have focused on marine, riverine, and lacustrine ecosystems [Penton et al., 2006; Zhang et al., 2007; Schmid et al., 2007; Hu et al., 2012; Wang et al., 2012]. Estuaries, as the ecotones between riverine and marine ecosystems, are important sites of nitrogen biogeochemical cycling [McClain et al., 2003; Dahnke et al., 2008]. These land-sea transitional zones are therefore hypothesized to be anammox hot spots, due to their biogeochemical features. However, diversity and distribution of anammox bacterial community are not described fully in estuarine ecosystems [Dale et al., 2009]. Few data are available to draw conclusions on the dynamics of anammox bacteria and their interactions with associated activity in these complex estuarine environments.
 The Yangtze Estuary was selected as our study area to examine anammox bacterial dynamics and related activity. The Yangtze River is the largest river in Euro-Asian continent and is ranked third in length, fourth in sediment discharge, and fifth in freshwater discharge in the world, and thus plays an important role in global biogeochemical cycles [Hou et al., 2008]. The Yangtze Estuary is located in one of the most densely populated and industrialized areas of China. Annually, it receives a high load of anthropogenic nitrogen from increased agricultural activities, fish farming, and domestic and industrial wastewater discharge in the Yangtze River Basin [Hou et al., 2006], consequently leading to severe eutrophication and frequent occurrences of harmful algal blooms in the estuary and adjacent coastal areas [Dai et al., 2011]. Hence, the microbial nitrogen transformations are of major concern in the Yangtze Estuary. Although the importance of anammox has been identified, few studies have as yet examined the dynamics of anammox bacteria and associated activity in the nitrogen-enriched estuary. In this paper, we investigated the distribution, biodiversity, and abundance of anammox bacterial communities in the sediments of the Yangtze Estuary using molecular methods. Slurry experiments were also conducted to measure the activity of anammox bacteria with a nitrogen isotope-tracing technique. Furthermore, environmental factors were determined and compared to elucidate their correlations with anammox bacteria and related activity in the estuarine ecosystem. Another goal was to describe the inherent link between the dynamics of anammox bacteria and rates of the anammox reaction. This work may provide novel insights into the microbial nitrogen cycle in the Yangtze Estuary.
2 Materials and Methods
2.1 Study Area
 The Yangtze Estuary located in the subtropical monsoon climate zone is China's largest estuary. It is over 250 km long from the tidal current boundary to its mouth and on average about 90 km wide, which covers an area of about 8500 km2. The tide in the estuary is semidiurnal and irregular. The tidal amplitude reaches its maximum at the river mouth and decreases landward and seaward from there, with the average tidal amplitudes of 2.4–4.6 m in the estuarine system. At present, over 2 × 108 t of suspended sediment is carried annually into the estuary and its adjacent areas by the Yangtze River, although the flux of suspended sediment has recently decreased due mainly to the construction of the Three-Gorge Dam [Ying et al., 2005]. Consequently, extensive tidal marshes have developed along the Yangtze estuarine and coastal zones. Eutrophication and harmful algal blooms are the most serious environmental issues in this ecosystem, mainly due to excessive enrichment of inorganic nitrogen. In the last four decades, the mean concentration of inorganic nitrogen in pelagic water has increased from about 10 μM to over 130 μM [Chai et al., 2006]. Nowadays, about 1.1 × 106 t of inorganic nitrogen is annually transported into the estuary and its adjacent coastal areas [Dai et al., 2011].
2.2 Collection of Sediment Samples
 In this study, seven representative sampling sites were selected in the tidal marshes along the Yangtze Estuary (Figure S1). Field surveys were conducted in January and August 2011, respectively. Triplicate surface sediments (0–5 cm) at each site were collected with stainless steel tubes and shovels during the ebb periods. The sediment samples were stored in sterile plastic bags, sealed, and transported to the laboratory on ice within 4 h. In the laboratory, triplicate surface sediments from each site were mixed under a helium atmosphere to form one composite sample. Then, one part of the sediment samples from each site was incubated to measure anammox activity, and another portion was examined for sediment physiochemical characteristics. Meanwhile, subsamples were preserved at −80°C for DNA extraction and molecular analysis.
2.3 Analysis of Sediment Characteristics
 Sediment temperature was measured in situ with a portable electronic thermometer at each site. Exchangeable ammonium (NH4+-N), nitrite (NO2−-N), and nitrate (NO3−-N) were extracted from fresh sediments with 2 M KCl and measured spectrophotometrically on a continuous-flow analyzer (SAN plus, Skalar Analytical B.V., the Netherlands) with detection limits of 0.5 μM for NH4+-N and 0.1 μM for NO2−-N and NO3−-N. Organic carbon was determined by a Carbon-Hydrogen-Nitrogen elementary analyzer (VVarioELIII) after removing carbonate by leaching with 0.1 M HCl [Hou et al., 2008]. Total sedimentary phosphorus (TP) was measured colorimetrically by the ascorbic acid-molybdate blue method after 2 h of combustion (500°C) and 16 h of extraction with 1 M HCl [Hou et al., 2011]. Sediment salinity was measured using a Yellow Springs Instrument Model 30 salinity meter, after sediments mixed with deionized water free of CO2 at a ratio (sediment/water) of 1:2.5. Sediment grain size was measured using a LS 13 320 Laser grain sizer. All physiochemical parameters of sediments were analyzed in triplicate. Detailed information on sediment characteristics are given in Table S1.
2.4 DNA Isolation, Polymerase Chain Reaction (PCR), Cloning, Sequencing, and Phylogenetic Analysis
 Total genomic DNA was extracted from 1 g of sediment samples using UltracleanTM soil DNA Isolation Kits (Mobile Biometry, USA) according to the manufacturer's protocol. A nested PCR assay was carried out to amplify anammox 16S rRNA gene. The initial amplification was conducted using the PLA46f-1390r primers with a thermal profile of 94°C for 4 min followed by 30 cycles of 45 s at 95°C, 50 s at 59°C, and 1 min 40 s at 72°C, and a final 5 min extension cycle at 72°C [Schmid et al., 2000]. After the first step, 1 µl of the PCR product was used as a template for the second amplification with Amx368f-Amx820r primers using a thermal profile of 94°C for 4 min followed by 30 cycles of 45 s at 95°C, 50 s at 59°C, 1 min at 72°C, and a final 5 min extension cycle at 72°C [Schmid et al., 2005]. PCR was performed in a total volume of 50 µl containing dNTP (each 10 mM, Sangon, China) 1 µl, 10 × PCR buffer 5 µl (without MgCl2, Sangon), MgCl2 (25 mM, Sangon) 4 µl, each forward and reverse primer (10 μM, Sangon) 1 µl, template DNA 2 µl, Taq DNA Polymerase(5 U µl−1, Sangon) 1 µl.
 Appropriately sized fragments of the nested PCR products were separated by electrophoresis in 1% agarose gels and purified using Gel Advance-Gel Extraction system (Viogene, China). The purified fragments were cloned using the TOPO-TA cloning kit (Invitrogen, USA) in accordance with the manufacturer's instructions. Clones were selected randomly for further analysis. The sequences obtained in this study for anammox bacteria are available in NCBI under Accession numbers of JX243121-JX243743. All the sequences and their relatives obtained from the NCBI were aligned by using the ClustalX program (version 2.1) [Thompson et al., 1997]. The sequences with 97% identity were grouped into one operational taxonomic unit (OTU) using Mothur program (version 1.23.0, USA) (http://www.mothur.org/wiki/Main_Page) by the furthest neighbor approach. Phylogenetic trees were constructed by neighbor-joining (NJ) method using the Molecular Evolutionary Genetics Analysis software (version 5.03) [Kumar et al., 2004]. The relative confidence of the tree topologies was evaluated by performing 1000 bootstrap replicates [Tamura et al., 2007].
2.5 Quantitative PCR Assay
 Q-PCR was performed in triplicate with the primers AMX-808-F and AMX-1040-R with an ABI 7500 Sequence Detection System (Applied Biosystems, Canada) using the Synergy Brands green method [Hamersley et al., 2007]. Plasmids carrying anammox 16S rRNA genes were generated by cloning 16S rRNA genes using a Plasmid Mini Preparation Kit (Tiangen, China). Concentrations of plasmid DNA were measured using a Nanodrop-2000 Spectrophotometer (Thermo, USA). The 25 µl Q-PCR mixture contained 12.5 µl of Maxima SYBR Green/RoxqPCR Master Mix (Fermentas, Lithuania), 1 µl of each primer (10 μM), and 1 µl template DNA. All reactions were performed in eight-strip thin-well PCR tubes with ultraclear cap strips (ABgene, United Kingdom). The PCR protocols were performed as follows: 50°C for 2 min, 95°C for 10 min, followed by 45 cycles of 30 s at 95°C, 30 s at 55°C, and 30 s at 72°C. The specificity of the Q-PCR amplification was determined by the melting curve and by gel electrophoresis. Cycle thresholds were determined by comparison with standard curves constructed using serial dilution of the quantified standard plasmids. In all experiments, negative controls containing no template DNA were subjected to the same qPCR procedure to detect and exclude any possible contamination or carryover.
2.6 Measurement of Anammox and Denitrification Rates
 Slurry experiments were conducted to measure the activities of anammox and denitrifying bacteria with a nitrogen isotope-tracing method. Detailed information on slurry experiments was also described by Risgaard-Petersen et al.  and Engström et al. . In brief, slurries were made with fresh sediments and helium-purged tidal waters from each site at a ratio (sediment/water) of 1:5. The resulting slurries were then transferred into a series of 12 mL glass vials (Exetainer, Labco, High Wycombe, Buckinghamshire, UK) under a helium atmosphere. Subsequently, they were preincubated for 12 h to eliminate residual nitrate, nitrite, and oxygen at in situ temperature. After preincubation, these vials were divided into three groups, which were spiked through the septa of each vial with helium-purged stock solutions of (1) 15NH4+ (15N at 99.6%), (2) 15NH4+ + 14NO3−, and (3) 15NO3− (15N at 99%), respectively. The final concentrations of 15N in each vial were about 100 μM. Incubation of slurries was stopped by injecting 300 µl of 50% ZnCl2 solution after 8 h. The incubation time was determined based on a test experiment that showed a linear increase in N2 production within 8 h of incubation (Figure S2). The concentrations of 29N2 and 30N2 produced over the incubation were measured by membrane inlet mass spectrometry as described by Kana et al.  and An and Gardner . The rates of both anammox and denitrification, and their respective contribution to total N2 production were calculated using the methods developed by Thamdrup and Dalsgaard  and Trimmer et al. .
2.7 Statistics Analysis
 The biodiversity indicators (Shannon and Chao 1) were calculated for each constructed gene library with Mothur program (version 1.23.0). The coverage of each clone library was estimated by the percentage of the observed number of OTUs divided by Chao 1 estimate. The ecological distribution of anammox bacterial communities and their correlations with environmental factors were determined, respectively, by the Weighted UniFrac PCoA analysis and canonical correspondence analysis (CCA) using the Software for Canonical Community Ordination [ter Braak and Šmilauer, 2002]. Pearson correlation analyses were conducted with Statistical Package for Social Sciences program (version 11.5) to examine correlations among the anammox bacterial diversity, abundance, and environmental factors [Hu et al., 2012]. In addition, Student's t test was also performed to compare spatial and seasonal differences in the nitrogen transformation rates measured in the present study.
3.1 Identification of Anammox Bacteria in the Estuarine Marshes
 Anammox bacterial communities in the marsh sediments of the Yangtze Estuary were detected successfully with the nested PCR technique. During the initial PCR with the primers Pla46F and 1390R, the amplified PCR products with around 1300 bits/s size was not generated in some samples. However, in the subsequent nested PCR with the primers Amx368-Amx820, the amplified products of 477 bits/s were obtained from all the samples (Figure S3). In this study, a total of 1069 clones from the 14 libraries were sequenced. Cloning and sequence analysis confirmed that all retrieved sequences were related to the anammox bacterial 16S rRNA gene, compared with the related sequences deposited in GenBank.
3.2 Community Composition, Distribution, and Diversity of Anammox Bacteria
 The anammox bacterial 16S rRNA gene sequences recovered from the samples of the estuarine marshes were applied to examine the community structure, biodiversity, and distribution pattern of anammox bacteria along the Yangtze Estuary. The phylogenetic analysis of 16S rRNA genes indicated that three known anammox bacterial genera including Scalindua, Brocadia, and Kuenenia were detected at the study area (Figure 1). In the phylogenetic tree, five distinctive Scalindua clusters were retrieved. Scalindua clusters I and II were affiliated with Candidatus Scalindua sorokinii with 92.5%–97.7% and 92.5%–95.4% sequence identity, respectively (Table S2). Sequences of Scalindua clusters III and IV showed 92.2%–93.5% and 96.6%–100% identity to the 16S rRNA gene of Candidatus Scalindua brodae, respectively. Sequences of Scalindua cluster V had 92.7%–98.7% identity to the 16S rRNA gene of Candidatus Scalindua wagneri. However, this cluster was related most closely to the 16S rRNA gene sequences recovered from the sediments of the Cape Fear Estuary with 99% sequence identity [Dale et al., 2009]. Two distinctive Brocadia clusters were identified in the phylogenetic tree. Both clusters had 94.5%–98.5% and 93.3%–94.0% identity to the 16S rRNA gene sequences of Candidatus Brocadia fulgida, respectively. Brocadia cluster I was affiliated closest with the clones retrieved from the sediments of the Pear River Estuary with 98.9% 16S rRNA gene identity [Wang et al., 2012]. Sequences of Kuenenia clusters I and II had 95.2%–97.3% and 92.7%–94.3% identity to the 16S rRNA gene sequences of Candidatus Kuenenia stuttgartiensis, respectively. Kuenenia cluster I was affiliated most closely with the clones retrieved from the sediments of the Cape Fear Estuary with 98% 16S rRNA gene identity [Dale et al., 2009]. Kuenenia cluster II was related most closely to the 16S rRNA gene sequences recovered from the marine sponges of Florida coastal sea with 99% sequence identity [Mohamed et al., 2010]. Additionally, two clusters of sequences were not identified with any known anammox bacterial genus. They may have formed new assemblages in the anammox bacteria group. Sequences of novel cluster I were related less to the 16S rRNA gene of Candidatus Jetteniaasiatica with 93.4%–94% identity. This new cluster was related most closely with the 16S rRNA gene sequences recovered from the pasture subsoil with 99.5% sequence identity (E. A. Dell et al., unpublished data, 2008; direct GenBank submission). Novel cluster II was related distantly to the 16S rRNA gene sequences of Candidatus Kuenenia stuttgartiensis with 92.7%–93.9% identity. The closest relatives of this novel cluster were the 16S rRNA gene sequences observed in the paddy soil of Southern China with 99.4% identity [Zhu et al., 2011].
 Weighted UniFrac PCoA analysis with 16S rRNA gene sequences showed high anammox bacterial spatiotemporal heterogeneity in the community compositions of anammox bacteria (Figure 2). The first two PCoA principal coordinates (P1 and P2) explained 75.1% of the community changes of anammox bacteria among all the sampling sites. The study area anammox bacterial assemblage was divided into three distinctive groups. Anammox bacteria in group I were recovered in freshwater sediments whereas those in group III were retrieved in higher-salinity sediments. The remaining anammox bacteria in group II were detected mainly in sediments with low salinity. The potential relationships between the anammox bacterial spatiotemporal changes and environmental factors in the estuarine marshes were examined by weighted CCA analysis (Figure 3). The first two CCA axes (CCA1 and CCA2) explained 79.5% of the cumulative variation of the anammox bacterium-environment relationship. Of all the environmental factors investigated, only salinity had a significant influence on the anammox bacterial community distribution (P = 0.002).
 The biodiversity of anammox bacteria was investigated via rarefaction analysis of the 16S rRNA gene sequences, Chao1 estimator and Shannon index calculations. The operational taxonomic unit (OTU) cutoff was examined by the 97% sequence similarity to reflect phylotype diversity of anammox bacteria detected at each sampling site (Figure 4 and Table 1). The number of OTUs, Chao 1 estimator, and Shannon index showed high seasonal and spatial differences in diversity of anammox bacteria at the study area. Compared with summer, the community of anammox bacteria in winter had relatively high diversity at most sites. In the cool season, high diversity of anammox bacteria was observed at sites DX, YY, and LC with more than 16 OTUs. However, in the warm season, the community of anammox bacteria had higher diversity at sites XP, WS, and LC where at least 11 OTUs were observed. Compared with environmental factors, the diversity of anammox bacteria at the study area was related significantly to salinity, temperature, and organic carbon contents (Table S3).
Table 1. Summary of Anammox Bacterial Biodiversity in the Marsh Sediments of the Yangtze Estuary
Numbers of OTUs and Sequences
Chao 1 Estimate
3.3 Quantification of Anammox Bacteria
 Melting-curve analyses of the anammox bacterial 16S rRNA revealed only one peak appearing at 84.3°C, confirming that fluorescent signals were derived from specific PCR products in the process of the qPCR quantifications. With plasmids containing cloned 16S rRNA gene fragments, standard curves were generated by plotting the threshold cycle (Ct) versus the log10 value of gene copy number. A significantly linear relationship (R2 = 0.99) was obtained over 6 orders of magnitude of the standard plasmid DNA concentration (2 × 101 to 2 × 107 copies µl−1), indicating the high primer hybridization and extension efficiency. In this study, the numbers of anammox bacterial 16S rRNA genes were observed between 2.63 × 106 and 1.56 × 107 copies g−1 dry sediment in the mashes of the Yangtze Estuary (Figure 5). The anammox bacterial abundance was characterized by great spatiotemporal heterogeneity at the study area (Student's t test, P < 0.05). In general, the abundance was higher in summer (4.54 × 106 to 1.56 × 107 copies g−1 dry sediment) than in winter (2.63 × 106 to 9.48 × 107 copies g−1 dry sediment). Also, the anammox bacteria were more abundant at the saline sites than at the freshwater-dominated sites. Pearson correlation analyses revealed that the anammox bacterial abundance was related significantly to the changes of salinity, nitrite, and temperature, as compared with other environmental factors (Table S3).
3.4 Rates of Anammox and Denitrification
 Slurry incubations were conducted with a 15N tracing technique to estimate activities of anammox and denitrifying bacteria in the estuarine environment. During the incubations spiked with 15NH4+ only, no significant production of 15N-labeled gases (29N2 and/or 30N2) was measured at all sampling sites (Figure S2a), showing that ambient nitrite and nitrate in slurries were consumed within the 12 h preincubation. During the incubations spiked with both 15NH4+ and 14NO3−, the production of 29N2 was detected at all sites whereas no production of 30N2 was observed (Figure S2b). These incubations implied that anammox occurred at the study area. During the incubations spiked with 15NO3− only, the significant production of both 29N2 and 30N2 was determined (Figure S2c). The activities of anammox and denitrifying bacteria could thereby be obtained within the 15NO3− incubations.
 The calculated results indicated that the anammox bacterial activity in the estuarine marshes ranged from 0.94 to 6.61 nmol N g−1 dry sediment h−1 (Figure 6), with distinct spatiotemporal heterogeneity (Student's t test, P < 0.05). In this study, the anammox bacteria were more active in summer than in winter, with respective rates of 3.92–6.61 nmol N g−1 dry sediment h−1 and 0.94–2.33 nmol N g−1 dry sediment h−1. Also, higher activity of anammox bacteria was recorded at the saline sites than at the freshwater-dominated sites, with average values of 4.18 and 3.02 nmol N g−1 dry sediment h−1, respectively. The cell-specific anammox activity was calculated using the data on the anammox bacterial abundance and activity, assuming that each cell had equal activity and each genome contained one gene copy. The estimated cell-specific rates varied between 5.18 and 27.49 fmol d−1 (Figure S4), which are comparable to the reported values (2–21 fmol d−1) [Strous et al., 1999; Kuypers et al., 2003; Zhu et al., 2011]. Remarkably higher cell-specific anammox rates occurred in summer (9.92–27.48 fmol d−1) than in winter (5.18–14.63 fmol d−1).
 The rates of denitrification in marsh sediments were also measured to examine its potential link to the anammox process. The denitrification rates varied between 41.82 and 67.13 nmol N g−1 dry sediment h−1 in summer sediments (Figure 7). In contrast, significantly lower rates of denitrification were observed in winter sediments (Student's t test, P < 0.05), with values of 12.04–20.03 nmol N g−1 dry sediment h−1. Meanwhile, the denitrification rates was found to strongly correlate to the anammox rates at the study area (R = 0.951, P < 0.0001).
 Spatial and seasonal variations of anammox bacterial community structure, biodiversity, abundance, and related activity were examined in the marsh sediments of the Yangtze Estuary. The community of anammox bacteria at the study area had distinctive spatial heterogeneity along the estuary (Figure S5). At the high-salinity sites, the anammox bacterial community was dominated by Scalindua which on average occupied 79% of total detected sequences. At freshwater sites, it was dominated by Brocadia which accounted for almost all of the sequences. In contrast, the community of anammox bacteria at the low-salinity sites was codominated by Brocadia and Scalindua which occupied 55% and 30% of total sequences, respectively. These results imply that salinity was a key environmental factor defining the biogeographical distribution of the anammox bacterial community structure in this estuary. The distribution pattern was supported by the CCA analysis (Figure 3). Scalindua genus often has a high tolerance for salinity whereas the growth of Brocadia generally is favored in freshwater and (or) brackish environments [Jetten et al., 2003; Schmid et al., 2007; Hu et al., 2012; Wang et al., 2012]. The difference in the ecophysiology of these anammox bacteria may be the main mechanism for the geographical distribution pattern of the anammox bacterial community in the estuarine environment.
 Relatively low biodiversity of anammox bacteria has been reported for coastal seas, rivers, and lakes [Penton et al., 2006; Schmid et al., 2007; Zhang et al., 2007; Hu et al., 2012; Wang et al., 2012]. So far, the anammox bacterial biodiversity reported in marine, riverine, and lacustrine environments has been restricted mainly to Scalindua or Brocadia [Schmid et al., 2007; Zhang et al., 2007; Hamersley et al., 2009; Yoshinaga et al., 2011; Hu et al., 2012]. However, we observed that the anammox bacteria at the study area are related closely to Scalindua, Brocadia, and Kuenenia, except for two novel clusters based on the 16S rRNA genes. High biodiversity of anammox bacteria was also observed in the Cape Fear River Estuary [Dale et al., 2009]. Therefore, it was hypothesized that high anammox bacterial biodiversity always occurs in the estuarine ecosystems, which may be attributed to land-sea interaction effects. During the interaction of land and sea, the anammox bacteria may be transported and blended by river runoff and tidal current, thus enhancing the diversity of anammox bacterial populations. Also, the diversity of anammox bacteria changed seasonally in the marsh sediments of the Yangtze Estuary. In summer, the anammox bacteria consisted mainly of Scalindua and Brocadia. However, it shifted to Scalindua, Brocadia, and Kuenenia in winter. In addition, anammox bacterial diversity was affected significantly by the sediment organic carbon content. Generally, relatively high biodiversity was observed at sites with low organic carbon content (Table S3). Thus, low organic carbon content may favor the coexistence of diverse anammox bacteria. A similar correlation occurred in Qiantang River sediments [Hu et al., 2012]. These results reflected that the microscale environmental factors may play an important role in shaping the diversity of anammox bacterial community in the estuarine ecosystem.
 The numbers of anammox bacteria quantified in the marsh sediments of the Yangtze Estuary were comparable to those measured in the Cape Fear River Estuary [Dale et al., 2009], the paddy soil of Southern China [Zhu et al., 2011], Barents Sea [Schmid et al., 2007], and the Pear River [Wang et al., 2012]. However, distinct spatiotemporal changes in the anammox bacterial abundance were observed at the study area. The spatial distribution may be controlled partly by the pattern of sediment salinity in the marshes (Table S3). The correlation between salinity and the numbers of anammox bacteria was also observed in the Cape Fear River Estuary [Dale et al., 2009]. The response of anammox bacterial abundance to salinity may result from the distribution of Scalindua-like bacteria which may adapt better in higher-salinity systems and thus propagate quickly in mesohaline and polyhaline environments [Schmid et al., 2007; Dale et al., 2009]. In contrast, the seasonal fluctuation of anammox bacterial abundances may result from temperature changes (Table S3). In the high temperature season, the growth of anammox bacteria may be stimulated by the nitrite substrate produced from enhanced denitrification rates (Figure S6). This explanation would imply that nitrite is a limiting environmental factor for anammox bacterial population development in the marsh sediments of the Yangtze Estuary. This hypothesis is supported by the significant relationship between the anammox bacterial numbers and nitrite content in sediments (Table S3). The stoichiometry of the anammox reaction supports this conclusion. Although previous studies indicate that anammox bacteria may have a low affinity constant for nitrite [Thamdrup and Dalsgaard, 2002; Trimmer et al., 2005], ammonium and nitrite are needed at an almost equimolar ratio during anammox bacteria metabolism [Strous et al., 1998]. However, the molar ratios of ammonium to nitrite reached between 282 and 1050 at the study area (Table S1), and consequently, nitrite substrate as a limiting factor affected the anammox bacterial abundance in the environments. This nitrite-dependent correlation also occurred in marine sediments [Rich et al., 2008; Dang et al., 2010], river sediments [Hu et al., 2012], and water columns [Lam et al., 2007; Woebken et al., 2007].
 Our isotopic tracing experiments showed that the measured anammox rates varied between 0.94 and 6.61 nmol N g−1 h−1 in the marsh sediments of the Yangtze Estuary, which were comparable to the values reported in other terrestrial soils, and estuarine and coastal sediments [Trimmer et al., 2003; Risgaard-Petersen et al., 2004; Dale et al., 2009; Zhu et al., 2011; Hou et al., 2012]. The anammox activity at the study area was characterized by great spatial and seasonal variations. Correlation analyses indicated that the changes of anammox activity were mainly controlled by nitrite availability, temperature, and the abundance of anammox bacteria (Table S3 and Figure S7). Interestingly, although the diversity and abundance of anammox bacteria were affected by salinity fluctuations, no statistically significant relationship between the anammox activity and salinity was observed in this study. These relationships implicated that salinity might have different impacts on the diversity, abundance, and activity of anammox bacteria. In contrast, the anammox activity at the study area might be more sensitive to the changes of nitrite and temperature (Table S3). In addition, compared with denitrification, it is estimated that the anammox process contributed approximately 6.6%–12.9% to total nitrogen loss in the marshes of the Yangtze Estuary. If the measured anammox rates were extrapolated to the entire estuary, about 1.35 × 104 t N can be removed from the aquatic ecosystem by this process on an annual basis, assuming that the bulk density of the sediments from the study area is 2.68 g cm−3 [Hou et al., 2006]. This removal amount of nitrogen accounts for approximately 1.23% of the total terrigenous inorganic nitrogen transported annually into the estuary.
 In the present study, the anammox rates were also compared to the potential nitrification rates (Figure S6; Y. Zheng et al., Community dynamics and activity of ammonia oxidizing prokaryotes in marsh sediments of the Yangtze Estuary, submitted to Applied and Environmental Microbiology, 2013) and denitrification rates to elucidate the links among these nitrogen transformation processes. It is found that there was no significant relationship between the anammox and nitrification rates at the study area. In contrast, a close correlation between anammox and denitrification rates was observed in the estuarine marshes. These correlation analyses reflected that anammox may be coupled closely to denitrification in the estuarine environment. During the interaction between anammox bacteria and denitrifiers, the denitrification process was likely a primary source of nitrite for the anammox reaction in the marsh sediments of the estuarine ecosystem [Meyer et al., 2005]. The tight linking of both processes is also supported by the significant correlations of the anammox rates with nitrite, temperature, and the anammox bacterial abundance. In summer, more nitrite was generated as an intermediate product by enhanced denitrification (Table S1 and Figure S8). As a result, the development of anammox bacteria was stimulated, and enhanced the anammox rates in this warm season (Table S3 and Figure S7). In contrast, denitrifying bacteria competed for nitrite with anammox bacteria in winter, probably because less nitrite was produced by decreased denitrification under low temperature conditions. Therefore, low availability of nitrite may inhibit anammox bacterial growth and activities in this season. In addition, the process of dissimilatory nitrate reduction to ammonium (DNRA) can also produce nitrite as an intermediate product [Philippot and Højberg, 1999] and thereby act as an alternative source of substrate for anammox bacteria. Indeed, the coupling between the anammox and DNRA processes has been reported in the oxygen minimum zones (OMZs) where denitrification was not detectable [Kartal et al., 2007a; Lam et al., 2009; Jensen et al., 2011]. Compared with the OMZs, the DNRA process was likely to make a minor contribution to production of nitrite for metabolic activity of anammox bacteria in the nitrogen-enriched estuarine marshes where the denitrification process was highly active [Roberts et al., 2012]. However, further work is still required to confirm this hypothesis.
 This study first investigated the anammox bacterial diversity, abundance, and activity in the marsh sediments of the Yangtze Estuary. Our results showed high biodiversity of anammox bacteria at the estuarine area, including Candidatus Scalindua, Brocadia, Kuenenia, and two novel clusters. Also, salinity was a key environmental factor regulating the geographical distribution and biodiversity of the anammox bacterial community across the estuary. The anammox process made a significant contribution to the loss of nitrogen from the aquatic ecosystem. Meanwhile, its rates were not related to the diversity of anammox bacteria but to their abundance. The coupling between the anammox and denitrification processes was strong in the estuarine marshes, perhaps indicating that denitrification is an important source of nitrite substrate for anammox bacterial metabolism.
 This work was funded by the National Natural Science Foundations of China (41130525, 41021064, 41071135, and 41271114) and the State Key Laboratory of Estuarine and Coastal Research (2010RCDW07). It was also supported by the Fundamental Research Funds for the Central Universities and the Marine Scientific Research Project for Public Interest (200905007). We thank B. Xie and S. P. Wang for sharing their analytical expertise on detection of anammox bacteria and Wayne S. Gardner for his comments on the manuscript. The anonymous reviewers are thanked for their constructive comments on a preliminary draft of the manuscript.