Nitrification is one of the most essential steps in the nitrogen cycle, including the microbial oxidation of ammonia to nitrite and subsequently to nitrate. As the rate-limiting step in nitrification, ammonia oxidation is catalyzed by a series of phylogenetic and physiological microorganisms (Winogradsky, 1890; Venter et al., 2004; Könneke et al., 2005; Treusch et al., 2005). Before ammonia-oxidizing archaea (AOA) were found in Crenarchaeota in 2004, ammonia-oxidizing bacteria (AOB) were always considered the only contributors to ammonia oxidation. As a membrane-bound enzyme, ammonia monooxygenase (AMO) is an important functional protein for ammonia oxidation and can oxide ammonium in AOA and AOB (Rotthauwe et al., 1997; Könneke et al., 2005). By encoding the active site of ammonia monooxygenase, amoA can oxide ammonia and generate energy for the sequential oxidation reaction (Hooper et al., 1997; Junier et al., 2010). Thus, the amoA gene has been widely applied in molecular ecology research of AOA and AOB in different environments, such as lakes (Jiang et al., 2009; Hou et al., 2013b), wetlands (Ye et al., 2011; Sims et al., 2012a,b), soils (He et al., 2007; de Gannes et al., 2012; Wang et al., 2014; Zhao et al., 2015) and estuaries (Caffrey et al., 2007; Jin et al., 2011; Zheng et al., 2013). Compared with AOB, AOA predominates in many environments, indicating that the ammonia oxidation process promoted by AOA may be more intense (Shen et al., 2008; Jin et al., 2011; Hou et al., 2013a). The relative contribution of AOA and AOB to ammonia oxidation is one of the most important issues related to the nitrogen cycle in the ocean.

Studies have shown that salinity (Caffrey et al., 2007; Mosier and Francis, 2008; Santoro et al., 2008), ammonium concentration (Di et al., 2009; Niu et al., 2013; de Gannes et al., 2014) and dissolved oxygen concentration (Park et al., 2006; Santoro et al., 2008; Liu et al., 2011) are potential factors influencing the community structure of AOA and AOB. As electron acceptor and donor, oxygen and ammonia are essential for ammonia oxidation, but for AOA and AOB, the affinities of ammonia and oxygen differ. It has been found that AOA prefer areas with low ammonium concentrations while AOB prefer areas with high ammonium concentrations (Di et al., 2009), and several researchers have found that AOA show better survival than AOB under conditions of low dissolved oxygen (Kayee et al., 2011; Limpiyakorn et al., 2013). In addition to these factors, ammonia oxidizers’ communities were also affected by distinct seasonal variations. A previous study in Catalina Harbor showed that the AOA and AOB amoA gene copy numbers were greater in summer than in winter (Beman et al., 2012). In a seasonal study of AOA in the deep oligotrophic Lake Lucerne, the amoA gene copy numbers matched the seasonal dynamics within 16 months (Vissers et al., 2013). Seasonal changes in the amoA gene also reflected the dynamic changes in AOA community characteristics in oligotrophic high mountain lakes (Auguet et al., 2011).

Potential nitrification rates (PNR) are measured by the addition of excess ammonium under optimal conditions, and they can be applied to compare the different contributions of AOA and AOB to ammonia oxidation. Several studies have shown a correlation between PNR and amoA gene copy numbers. For example, PNR was positively related to the abundance of archaeal amoA gene, but it showed no significant correlation with bacterial amoA gene abundance in several southeastern regions of America (Caffrey et al., 2007). Wankel et al. (2011) were the first group to relate the amoA gene abundance to ¹⁵N isotope methods and found no obvious correlations between them in the Elkhorn Slough estuary. A positive correlation was also found between PNR and the archaeal amoA gene abundance in Yangtze estuarine sediments, suggesting that AOA might be more important for mediating ammonia oxidation in this area (Zheng et al., 2014).

Rushan Bay is a typical semi-enclosed sea and an important mariculture area in Shandong Peninsula, China. With the development of industry and aquaculture, large quantities of nitrogen (N) and phosphorous (P) nutrients have been transported into this area, contributing to a series of ecological problems such as eutrophication. Increasing nutrients and organic materials in estuarine and coastal waters can also result in an oxygen minimum zone (OMZ) or even hypoxia in the bottom water. An OMZ was previously found in Rushan Bay and its adjacent waters in 2009 (Cui et al., 2012; Ran et al., 2012). The presence of OMZ may destroy the oxidizing environment in surface sediment, which can impact the biogeochemical cycle of elements in marine environments, particularly for nitrification (Caffrey et al., 2007; Santoro et al., 2008; Qiu et al., 2013).

Sediments are habitats and can constitute enabling environments for microbes to flourish (Besaury et al., 2014). However, to date, there has been relatively little focus on the community structure and activity of AOA and AOB in surface sediments from adjacent waters of Rushan Bay. Moreover, a comprehensive understanding of the different contributions of AOA and AOB to ammonia oxidation in the studied sediments needs to be explored. In the present study, the composition, abundance and activity of AOA and AOB communities in sediments from adjacent waters of Rushan Bay were studied based on amoA gene sequences and potential nitrification incubation. Our main objectives were (1) to study the community structure and activity of AOA and AOB in surface sediments from adjacent waters of Rushan Bay, (2) to determine potential links between environmental factors and community structure, (3) to discuss the differential contributions of AOA and AOB to ammonia oxidation during summer and winter seasons in our studied area. We hypothesized that the (1) Nitrosopumilus group of AOA and Nitrosospira group of AOB would be more adaptable, exhibiting a marine influence to ammonia oxidizers’ communities, (2) AOA and AOB would contribute differentially to ammonia oxidation in different seasons owing to their different nitrification potentials and transcript abundance in different seasons.

Currently, most studies have used a cloning-sequencing method rather than a pyrosequencing method to examine AOA and AOB communities, and only a few studies have used pyrosequencing to describe amoA diversity (Pester et al., 2012; Mao et al., 2013; Hayashi et al., 2015; Tago et al., 2015; Zhang et al., 2016). In the present study, the community structure, diversity and activity of ammonia oxidizers were researched to understand the community information, the different contributions of AOA and AOB to ammonia oxidation and the impact of environmental factors on community composition in surface sediments from adjacent waters of Rushan Bay. AOB had a lower diversity than AOA, consistent with results from other coastal environments (Santoro et al., 2008; Cao et al., 2011b; Jin et al., 2011). Studies have identified that the diversity of AOA in estuaries might be greater than that in the adjacent open oceans, which may be due to nutrients in the river discharge (Llyd and Jody, 2003; Cao et al., 2011a). In the present study, the diversity of AOA at C0 was relatively greater than that at C2, consistent with the above view. Environmental factors, such as temperature, salinity, nitrate concentration and total nitrogen, were considered as driving forces affecting the diversity of AOA or AOB communities (Bernhard et al., 2005; Cao et al., 2011a; Yu et al., 2016). However, according to the correlation analysis, no noticeable environmental factors could be identified in relation to the diversity of AOA or AOB communities in our studied sediments (P > 0.05).

Phylogenetic analysis indicated that the AOA community was clustered into the Nitrosopumilus group and Nitrososphaera group (Figure 3A). A clear separation of AOA communities between marine and terrestrial environments was observed in our study, in accordance with a study of eastern intertidal sediments in Chongming Island (Zheng et al., 2013). The presence of soil-related AOA in our studied sediments could be explained by the deposition of microorganisms from upstream water with runoff, consistent with the greater relative abundance of Nitrososphaera group at C0 in both summer and winter (Figure 4A). The core microbiome of AOA communities exhibited greater similarity to the uncultured archaeal amoA gene sequences from other environments, such as the North Sea (Lipsewers et al., 2014), Tanoura Bay (Ando et al., 2009), Elkhorn Slough estuary (Wankel et al., 2011) and Eastern China Marginal Seas (Yu et al., 2016), showing that AOA obtained in this study was not representative of the special group in the studied sediments and their genotypes are ubiquitous worldwide.

The bacterial amoA sequences recovered from our studied areas were classified into the Nitrosospira group and Nitrosomonas group (Figure 3B). The Nitrosospira group accounted for most of AOB in our study, whereas the Nitrosomonas group showed lower relative abundance of 2.09, 0.48, 2.89, and 0.49% at SC0, SC2, WC0 and WC2, respectively (Figure 4B). The proportions of the Nitrosomonas group were greater at SC0 and WC0, suggesting that the upstream water has a significant impact on marine benthic microbes (Dang et al., 2008). Freitag et al. (2006) found that the Nitrosospira group dominated at stations with a marine influence, while the Nitrosomonas group had an advantage at stations with a freshwater influence. This evidence indicated that the Nitrosospira group may be more adaptable than the Nitrosomonas group in the coastal marine area, consistent with our results.

Various environmental factors can impact AOA and AOB communities, and their effects may be complicated. As the primary energy source of ammonia oxidizers, ammonium concentration is considered to be a vital factor influencing the ammonia oxidizers’ community structure. In a study of AOA communities in mangrove sediments, Li et al. (2010) demonstrated that ammonium concentrations might negatively impact the diversity of AOA. Ammonium concentrations were also found to be an important element for the community structure of AOB in intertidal sediments of Yangtze estuary (Zheng et al., 2014). Although AOB was sensitive to dissolved oxygen concentrations, studies have shown that dissolved oxygen concentrations can affect the activity of AOB rather than the AOB community (Park et al., 2008). Thus, no relationship between dissolved oxygen concentrations and the AOB community was found in the present study. RDA analysis indicated that temperature, salinity and ammonium concentration were the most important factors affecting the AOA community composition (Figure 5A), whereas temperature, chla and nitrite concentration appeared to play a pivotal role in the AOB community composition (Figure 5B). These factors have already been previously reported to be important for the determination of AOA and AOB communities (Sahan and Muyzer, 2008; Bouskill et al., 2012). Moreover, RDA indicated that no single environmental factor was likely to completely determine the AOA or AOB community compositions (P > 0.05, 499 Monte Carlo permutations), and the ammonia oxidizers’ community structure in the present study seemed to be influenced by several environmental factors in combination, possibly due to a variety of nature environmental gradients and multicollinearity among many abiotic factors in coastal estuarine regions.

The amoA gene obtained in the present study and other five pyrosequencing studies (sediments from Changjiang Estuary and Hangzhou Bay, soils from Minami-Daito Island and Ny-Ålesund) (Hayashi et al., 2015; Tago et al., 2015; Xia et al., 2015; Zhang et al., 2015, 2016) were trimmed as reported above. The remaining high-quality sequences were clustered into 280 OTUs (for AOA) and 201 OTUs (for AOB) at 97% similarity levels. UPGMA clustering of the samples from the five environments was conducted using the R program. The UPGMA cluster showed a clear separation of AOA and AOB communities in samples from these five environments. The community types of AOA were clustered into three different groups consisting of group I, group II and group III (Figure 9A). The environment samples from Rushan Bay (RSB), Changjiang Estuary (CJE) and Hangzhou Bay (HZB) showed marine characteristics and were grouped together, while the other two soil samples formed group II and group III. Therefore, we speculated that geography contributes significantly to the community structure of AOA, consistent with reports showing that diverse and distinct AOA communities are associated with each habitat (Francis et al., 2005; Dang et al., 2009). Furthermore, similar to the UPGMA clustering of AOA communities, the AOB communities were clustered into two different groups according to the sediment and soil habitats (Figure 9B). The environment samples from RSB, CJE and HZB, which showed marine characteristics, were grouped together. The AOB community from soils in Minami-Daito Island (MD) appeared to be more similar to that in Ny-Ålesund (NA), although they are far from each other geographically. Many studies have shown that AOB communities are dominated by Nitrosomonas in terrestrial environments, whereas in estuarine areas, Nitrosospira may be more adaptable (Bartosch et al., 2002; Dang et al., 2010; Cao et al., 2011a). Based on all of these results, we can speculate that geography also contributes significantly to the community structure of AOB. The difference in geography mainly reflects differences in environmental conditions. Thus, the role of geography in the ammonia oxidizers’ community structure essentially depends on different environmental factors, for instance, ammonium concentration, salinity and temperature.

Consistent with most reports, our study showed that AOA was more abundant than AOB (Liu et al., 2011; Yao et al., 2011; Sims et al., 2012a,b; Hou et al., 2013b). Environmental factors, for instance, salinity, temperature and ammonium concentration, have been identified as potential factors to determine the abundance of AOA and AOB. In our study, correlation analysis indicated that no single environmental factor was likely to completely determine the ammonia oxidizers’ abundance, which seemed to be influenced by a combination of factors. The abundance of amoA gene could provide information concerning the potential ammonia-oxidizing activity, while the transcript activity of the amoA gene is a measure of protein production. In the studied sediments, both the ratios of active AOA to total AOA and the ratios of active AOB to total AOB were less than 1%, which indicated that most of the AOA and AOB amoA genes are likely associated with inactive cells. Studies have shown that external ammonium concentrations can up-regulate the expression of the amo operon. For example, in Nitrosomonas europaea, the proximal amoCAB promoter is active only when recovery from ammonium starvation, while the distal amoCAB promoter is constitutively active in the presence of ammonium (Berube et al., 2007). Our finding that the transcript abundance of the archaeal amoA gene peaked in the lowest ammonium concentration region was consistent with previous findings. To regulate the archaeal amoA gene transcription, temperature is another critical element (Tourna et al., 2008). Greater archaeal amoA transcript abundances were observed in warmer seawaters in our studied sediments, but no significant relationship was observed between archaeal amoA transcripts and temperature (P > 0.05).

Potential nitrification rates can reflect potential ammonia-oxidizing activity. There may be some certain connections between PNR and the abundance of ammonia oxidizers; however, the relationships did not achieve significant levels in our study (P > 0.05). Many environmental factors, such as salinity, ammonia concentration, dissolved oxygen concentration and temperature, can impact PNR. In the present study, there was a strong negative correlation between the total PNR and salinity, indicating that a high level of salinity could inhibit the metabolic activity of ammonia oxidizers. In addition, the total PNR also showed a strong positive correlation with temperature, which could explain why the total PNR were higher in summer. In addition to the total PNR, the PNR of AOB and temperature also showed a significant positive correlation (P < 0.01). All previously identified AOB are mesophilic bacteria, and their optimum growth temperature usually ranges from 20°C to 30°C (Yu et al., 2015). Temperature can directly affect the nitrification activity of AOB, which peaks at 15°C to 25°C (Avrahami et al., 2003). Enzyme activity decreases with decreasing temperature, ultimately resulting in a decline in the nitrification activity of AOB in winter. In contrast to AOB, AOA has a wide tolerance to temperature (Schleper et al., 2005; Jiang et al., 2010), and thus AOA shows an obvious competitive advantage over AOB at a high or low temperature.

Although the abundance and transcript abundance of AOA were greater than that of AOB in both summer and winter, the PNR of AOB and AOA was greater in summer and winter, respectively, leading us to hypothesize that PNR is directly related to nitrification potentials and the transcript abundance of ammonia oxidizers. Studies have shown that the nitrification potentials of AOB are 1–3 times as many as AOA (Jia and Conrad, 2009). Temperature was one of the important factors affecting nitrification potentials in this study (P < 0.05). The seasonal variation of the nitrification potentials directly led to the seasonal differences in the relative contributions of AOA and AOB to ammonia oxidation. The nitrification potentials of AOB decreased during winter, and AOB’s transcript abundance was relative lower. Thus, AOA dominated ammonia oxidation in surface sediments from adjacent waters of Rushan Bay. In summer, AOA had a greater transcript abundance than AOB, while AOB had far higher nitrification potentials than AOA. Thus, AOB contributed significantly to ammonia oxidation in surface sediments from adjacent waters of Rushan Bay.

In general, in surface sediments from adjacent waters of Rushan Bay, the diversity of AOA was higher than that of AOB. Nitrosopumilus group of AOA and Nitrosospira group of AOB may be more adaptable in the studied sediments. Both the abundance and the transcript abundance of AOA were greater than those of AOB, and most AOA and AOB amoA genes are likely associated with inactive cells. Significant seasonal variations were observed in PNR, and AOA and AOB contributed differentially to ammonia oxidation in different seasons. AOB occupied the dominant position in mediating ammonia oxidation during summer, while AOA might be dominant in ammonia oxidation during winter.

HH, YZ, TM, and ZY designed the experiments. HH and LF performed the experiments and analyzed the data. HH and YZ wrote the paper.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.