Introduction

Nitrogen is a key element for living organisms but is also one of the main causes of eutrophication in aquatic ecosystems (Bowen and Valiela 2001). There has been growing concern about the migration and transformation of nitrogen in recent years (Galloway et al. 2008). Because of the strong interaction between land and sea, the gradients of various physical, chemical, and biological factors have changed sharply (Kirwan and Megonigal 2013), and the processes of the nitrogen cycle in estuary areas have become more complex. Estuaries are typically environmentally vulnerable and sensitive zones in which the stability of ecosystems is easily disturbed and destroyed by natural and artificial factors. In addition, estuarine areas are rich in various nutrients and pollutants and are mixed with seawater and freshwater currents. Thus, environmental parameter gradients such as salinity are formed; as a result, the microbial community structure of estuaries is unique, as these areas contain abundant information about unknown genetics and ecology (Delgado-Baquerizo et al. 2016). Therefore, studying the distribution and community composition of microbial flora in estuarine water is of great significance for the nitrogen cycle in ecotones.

The degradation of organic nitrogen contamination can be accomplished by the processes of ammonia oxidation, nitrification, denitrification, and anaerobic ammonia oxidation (Holland 1998; Zehr and Ward 2002; Trimmer et al. 2003). The whole process, including organic nitrogen, entails oxidation to ammonia (ammonification), which is subsequently oxidized to nitrite and then to nitrate (nitrification). Then, nitrate is reduced to nitrite and then to dinitrogen gas (denitrification) (Kuypers et al. 2018). In the first stage, the corresponding functional microorganisms are ammonia-oxidizing bacteria (AOB) and ammonia-oxidizing archaea (AOA). Both AOB and AOA have been proven to be capable of oxidizing ammonia under aerobic conditions. These reactions are catalyzed by ammonia monooxygenase and coded by amoA (Konneke et al. 2005). Baptista (2014) adopted a reliable quantitative polymerase chain reaction (qPCR) technique to quantify the amoA gene and found an order-of-magnitude differences in AOB cell abundances, which indicated that the amoA gene is generally used as a specific marker to provide a scientific explanation for the mechanism of ammoxidation. Denitrifying microbes dominate the denitrification stage. Genes of interest include nitrite reductase genes (nirS and nirK), the nitric oxide reductase gene (norB), and the nitrous oxide reductase gene (nosZ) (Zhou et al. 2011). Among these, the nitrite reductase genes have been studied more extensively (Ligi et al. 2014; Zhi and Ji 2014). Previous studies have found that the nirS-type functional gene could reveal patterns of abundance and community structure (Yin et al. 2012; Deslippe et al. 2014) and the dominant species (Yu et al. 2018) of denitrifying organisms.

In addition, the ammoxidation process that proceeds under anaerobic conditions was also found to be a nonnegligible part of the global nitrogen cycle, namely, anaerobic ammonia-oxidizing (anammox) bacteria, which were theoretically proposed by Broda (1977). Anammox bacteria exist in anoxic environments (Mulder et al. 1995) and have the ability to reduce nitrite into N2 with ammonium as the electron donor, playing an important role in the nitrogen cycle and nitrogen pollution control (Ding et al. 2013). A continental surface may offer great possibilities and environmental conditions for anammox bacteria in freshwater wetland ecosystems (Zhu et al. 2010). Hydrazine synthase is a unique biomarker for studying the presence and biodiversity of anammox bacteria (Harhangi et al. 2012). One of its subunits is encoded by the functional gene hzsB; thus, researchers generally use the hzsB gene to characterize the community structure of anammox bacteria (Zhu et al. 2015; Zhou et al. 2017). In recent years, fluorescence in situ hybridization (FISH), 16S rRNA-based phylogeny analysis, and PCR have been used for the quantitation and identification of anammox bacteria and have indicated that the most common anammox bacteria in the environment are Candidatus Scalindua, Candidatus Brocadia, and Candidatus Kuenenia (Schmid et al. 2005; Yoshinaga et al. 2011).

Because of the important role of microorganisms in nitrogen cycling, many studies have focused on the relationship between microbial characteristics and the living environment or its niches. Previous studies mostly focused on gene abundance and community structure diversity in different environments, such as natural soil (Tourna et al. 2008) and aquarium biofiltration systems (Urakawa et al. 2008), and the effects of different soil fertility on populations of AOA and AOB (Bai et al. 2015; Wang et al. 2016a). Salinity has generally been considered an influencing parameter on archaeal and bacterial cells (Yang et al. 2015) as well as on anammox bacteria (Wei et al. 2016) and denitrifying organisms (Yoshie et al. 2004). Researchers have suggested that because of changes in salinity gradients (Bernhard et al. 2005; Zheng et al. 2013) and the influence of habitat ecotones or the transition zones of ecosystems (Huang et al. 2016), the structure and distribution of the nitrogen-transformation-related microbial communities have obvious regularity, but analyses of the effect of environmental parameters on the four microbes in the same location have been rare.

The Yangtze River is one of the main channels in China. Large amounts of sediment and seawater flood into this river every year. In addition, the salinity in the Yangtze Estuary is also affected by the intrusion of saltwater downstream of open seas (Chen et al. 2016). The aim of this study was to explore the impacts of salinity on nitrogen transformation-related microbes and reveal the phylogenetic characteristics of microorganisms in the Yangtze Estuary ecosystem. For this purpose, distribution regulation associated with environmental factors and the phylogenetic community structure of AOB, AOA, denitrifying microbes (nirS-type), and anammox in the estuary of the Yangtze River was assessed through a specific molecular biology method. Classification of different levels was accomplished, and correlations with the water quality indices were analyzed. The following questions were proposed before the study:

1. Is salinity the environmental factor that restrains the gene abundances of the four bacteria?

2. How do the diversity and community structure of microorganisms react when salinity gradients are regulated?

3. What is the dominant form of nitrogen transformation in the Yangtze River Estuary?

Materials and methods

General situation

The Yangtze River Estuary starts at the Xuliujing Wharf and extends to the East China Sea, spanning a total length of approximately 181.8 km. The estuarine section is fan-shaped and is divided into the north tributary and south tributary. The latter is divided into the north port and south port, as shown in Fig. 1. According to the salinity gradient, four regions were included in the sampling area, as shown in Additional file 1: Table S1 and Fig. 1. The four regions are S1, the freshwater upstream with sampling sites 1–3; S2, the south tributary transition area with sampling sites 4–7; S3, the area between the south port and north port with sampling sites 8–11; and S4, the north branch of the estuary area with sampling sites 12–14. Sediment samples were collected from the junction of water and sediments at each sampling site by a grab sampler and then sealed and stored at 4 °C. After being transported back to the laboratory, the samples were stored at 4 °C and − 80 °C for chemical analysis and DNA extraction, respectively.

Fig. 1
figure 1

Distribution of sampling sites in the Yangtze River Estuary. Sampling region S1 represents sampling sites 1–3 from freshwater upstream, sampling region S2 represents sampling sites 4–7 in the south tributary transition area, sampling region S3 represents sampling sites 8–11 in the area between the south port and north port, and sampling region S4 represents sampling sites 12–14 in the north branch of the estuary area

Chemical analysis

The NH4+-N and NO3-N concentrations in the sediment samples from the four sampling areas were measured using the potassium chloride solution-spectrophotometric method (China National Soil Quality Standard, GB 15618-1995) (He et al. 2015). Briefly, samples were extracted with a 2 mol/L KCl solution and filtered through a 0.45-μm filter membrane. NH4+-N was subsequently determined through Nessler’s reagent spectrometry, and NO3-N was determined using an ultraviolet spectrophotometer. The salinity, pH, and temperature measurements were detected in situ with an inline YSI EXO1 monitor.

Biological analysis

We lyophilized a 0.33-g sediment sample and used the Fast DNA Spin Kit for Soil DNA (MP Biomedicals, USA) to extract the total DNA. The mass of total DNA was detected using 1% gel electrophoresis, and the concentration was measured using a NanoDrop 2000 UV-Vis Spectrophotometer (Thermo Fisher Scientific, USA). PCR amplification was conducted with a Promega Agarose Gel DNA Purification Kit (Promega, Madison, WI). The PCR product was gel-purified and ligated to pGEM-T Easy Vector (Promega, Madison) and then transferred to JM109 competent cells (Takara, Dalian). Then, the combination was coated on a Luria-Bertani (LB) medium. Finally, after incubation, the blue-white selection was conducted, and five white clones were selected randomly and then amplified. The positive clones were sequenced (QingKexinye Biotechnology Limited Company, Beijing), and the obtained sequences were analyzed using the BLAST tool to aid in the selection of the closest matches (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Next, the obtained sequences were aligned using the Clustal_X program (Wang et al. 2018). Operational taxonomic units (OTUs) with 98% similarity were selected using UCLUST (Cole et al. 2009; Luria et al. 2014). The α-diversity indices were then calculated using Mothur to characterize the richness of each bacterium. Phylogenetic trees of the four bacteria were constructed using MEGA 5 software, with the neighbor-joining method based on different levels. The functional gene sequences were submitted to the National Center for Biotechnology Information (NCBI) under the accession number MN602086-MN602141.

Quantitative analysis of the gene abundances of the four microbial groups mentioned above, i.e., amoA, nirS (Zheng et al. 2015), and hzsB, in the sediment samples was conducted through real-time qPCR. All primers and protocols (Zheng et al. 2017; Zhu et al. 2018; Wang et al. 2012) that were used for amplification are summarized in Additional file 1: Table S2. The quantitative mixture totaled 20 μL in volume, including 10 μL of SYBR Premix Ex Taq enzyme, 0.4 μL of ROX 50, 0.4 μL of BSA (Takara Bio Company, DaLian), and 0.5 μL of each primer. Moreover, 2 μL of the DNA template was used, and the final volume was adjusted to 20 μL with H2O. An ABI 7300 Real-Time PCR system (CA, USA) was used for the quantitative amplification, and the SYBR Green method was used for quantitation. The plasmid DNA with known copies was diluted using 10 gradients. Thus, seven standard samples were obtained and were amplified at the same time, and a standard curve was obtained. The quantitative PCR results of gene abundance were calculated using 7300 SDS System software. The amplification efficiency between 90% and 110% and correlation coefficients above 0.98 was applied. A negative control was set up in the experiment, and each sample was determined three times.

Data analysis

Statistical analysis was performed using SPSS software. Pearson correlations and Student’s t tests were used to analyze the abundances of the four functional genes and the water quality indices. Plotting was conducted using Origin 8.5 and Microsoft Excel. Furthermore, a phylogenetic tree was constructed using MEGA version 5.0.

Results and discussion

Environmental parameters in the sediments

The values of the environmental parameters of temperature, pH, salinity, NH4+-N, and NO3-N in the samples in the four sampling areas with the ratio of NH4+-N and NO3-N are shown in Additional file 1: Table S3. The proportions of NH4+-N and NO3-N are shown in the pie chart in Fig. 1. The concentrations of both ammonia and nitrate nitrogen fluctuated greatly. In addition, the concentration of NH4+-N was higher than that of NO3-N at sampling points 5 and 7 in S2 and 11 in S3; at other sites, the nitrate concentrations were higher than the ammonia concentrations.

Identification of the sequence

Significant sequences were detected, in which 64 copies of AOA and 75 copies of AOB were obtained. In addition, sequences of denitrifying microbes (based on nirS gene, 53 copies) and anaerobic ammonia-oxidizing bacteria (26 copies) were also obtained. All sequences at different sampling sites were classified based on the flora of microorganisms. Moreover, Fig. 2 shows the proportion of four microbial sequences at species levels (the proportion of the four microbial sequences at species levels in the four sampling regions are shown in Additional file 1: Fig. S2). As exhibited, the majority of AOA were generally identified as the phyla Thaumarchaeota and Crenarchaeota and the genera Nitrosopumilus, Nitrososphaera, and Candidatus Nitrosocosmicus. AOB were identified as the phylum Proteobacteria and the genera Nitrosomonas and Nitrosospira (the identification of the sequences in S1–S4 at the phylum, class, order, family, and genus levels is shown in Additional file 1: Fig. S1). The sequence identification was similar to those in the adjacent waters of Rushan Bay (He et al. 2018), terrestrial environments (Oton et al. 2016), soil microorganisms (Verhamme et al. 2011), tropical stream sediments (Reis et al. 2015), and the Pacific Ocean (Church et al. 2010). In addition, the denitrifying microbes were identified as the phylum Proteobacteria and classes Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria, which was similar to Kim’s research findings (2016). Sequence analysis showed that the anammox bacteria were identified as the phylum Planctomycetes and the order Candidatus Brocadiales, which was similar to Jetten’s research findings (2001), and genera Candidatus Brocadia, Candidatus Jettenia, Candidatus Kuenenia, and Candidatus Scalindua. To date, the marine anammox bacteria that have been found in marine water columns and sediments worldwide are almost exclusively Ca. Scalindua species, which are capable of reducing nitrate via nitrite to ammonium under limited electron acceptor conditions by using organic matter and the uptake of small organic compounds (van de Vossenberg et al. 2013).

Fig. 2
figure 2

Identification of AOA, AOB, anammox bacteria, and denitrifying microbes on species levels at sampling regions S1–S4. Species names in bold represent the highest proportion in all sampling regions

Proportions of each taxon are shown in Fig. 2. The dominant species varied in the different sampling regions from S1–S4. The shift and differentiation pattern are common because of the environmental parameters of nitrogen content (van der Star et al. 2008) and salinity (Zhou et al. 2017). However, in this research, the dominant species of AOA were basically crenarchaeote enrichment cultures,thaumarchaeote enrichment cultures, and Nitrosopumilus maritime cultures. N. maritimus cultures could be utilized to determine whether putative gene coding for novel crenarchaeal metabolisms was functional (Prosser and Nicol 2008; Martin-Cuadrado et al. 2008). AOB were Betaproteobacterium enrichment cultures and Nitrosomona sp., which have a high tolerance for elevated ammonia concentrations and are believed to contribute significantly to the global production of nitrous oxide (Koops et al. 1991; Chain et al. 2003). Denitrifying organisms included Thauera selenatis, which has been found to obligate autotrophic denitrifiers and dominate in hydrogenotrophic denitrification (Mao et al. 2013; Xing et al. 2018). Furthermore, the anammox bacteria were Candidatus Brocadia sp. Among the five genera identified to date (Narita et al. 2017), members of Ca. Brocadia are considered as growth-rate strategists with a high growth rate and low substrate affinity (van der Star et al. 2008). The functions of the identified dominant species indicated that nitrogen transformation activities caused by the four microbes might coexist.

Alpha diversity index and OTU

In order to investigate the richness and community distribution of the four microbes, variance analysis was conducted between alpha diversity indices, and the results are shown in Table 1. Among all sampling sites, the diversity of the denitrifying microbes (based on nirS gene) was significantly different from the AOA, denitrifying microbes at the Shannon, Chao 1, and Ace indices, indicating that the diversity of denitrifying microbes is the highest. Since no significant difference of any indices was found in the anammox bacteria and AOA, and both of them were lower than AOB, it could be inferred from the table that AOB is the second most diverse microbe.

Table 1 α-diversity indices of four microbial groups in all sampling regions. Correlated microbe pairs are highlighted in light gray

The results of operational taxonomic units (OTUs) for AOA, AOB, denitrifying microbes, and anammox bacteria showed that denitrifying microbes had the highest OTU number, 46, followed by AOB (23), AOA (15), and anammox bacteria (14), and the order was consistent with the α-diversity indices. The highest number of OTUs for the denitrifying microbes was in S2 (15 OTUs), AOB was in S2 (16 OTUs), AOA was in S3 and S4 (9 OTUs), and anammox bacteria was in S2 and S3 (6 OTUs), suggesting that the moderate salinity regions of S2 and S3 are suitable for the growth of four groups (Biller et al. 2012). No overlaps of OTUs of these microbes were found in the four sampling regions except AOA, which had 2 OTUs in the overlaps of all sampling regions (shown in Additional file 1: Fig. S3), indicating that AOA has a strong adaptability to different salinity gradients. A previous study confirmed that the AOA ecotype might tolerate elevated salinity (Blainey et al. 2011) and could outcompete AOB in terms of growth and ammonia oxidation activity under moderate- and high-salinity conditions (Zhang et al. 2015).

Phylogenetic tree construction

The reference species were obtained by matching with the BLAST tool in the GenBank database. The sequence closest to the reference species in each sampling region was optimized, and the simplified phylogenetic tree of each microbe was drawn, as shown in Figs. 3, 4, 5, and 6. The phylogenetic tree of AOA suggested that the species in S4 shows an obvious affiliation with intertidal sediment (JQ345772) and Nitrosopumillaceae archaeon MY1 (HQ33117), while the species in S2 might be similar to those in intertidal sediments (KC735320). In S1, the species were similar to those in sediments in the Beiyun River Nitrosoarchaeum limmia SFBIN (AECG01000066). The species in S3 had no specific affiliations.

Fig. 3
figure 3

Neighbor-joining phylogenetic tree of archaeal amoA gene sequences in all sampling sites in the Yangtze River Estuary, where the white up-pointing triangle represents the sampling sites in S1, the black up-pointing triangle represents the sampling sites in S2, the white down-pointing triangle represents the sampling sites in S3, the black down-pointing triangle represents the sampling sites in S4, and next to these symbols are the numbers of sampling sites; the black circle represents the reference sequences gained from GenBank, and the two capital letters and six numbers in brackets are the accession numbers of the reference sequences. Bootstrap values close to the nodes are greater than 50% of 1000 replicates. The scale indicates 0.05 nucleotide substitutions per site

Fig. 4
figure 4

Neighbor-joining phylogenetic tree of bacterial amoA gene sequences in all sampling sites in Yangtze River Estuary, where the white up-pointing triangle represents the sampling sites in S1, the black up-pointing triangle represents the sampling sites in S2, the white down-pointing triangle represents the sampling sites in S3, the black down-pointing triangle represents the sampling sites in S4, and next to these symbols are the numbers of sampling sites; the black circle represents the reference sequences gained from GenBank, and the two capital letters and six numbers in brackets are the accession numbers of the reference sequences. Bootstrap values close to the nodes are greater than 50% of 1000 replicates. The scale indicates 0.05 nucleotide substitutions per site

Fig. 5
figure 5

Neighbor-joining phylogenetic tree of hzsB gene sequences in all sampling sites in Yangtze River Estuary, where the white up-pointing triangle represents the sampling sites in S1, the black up-pointing triangle represents the sampling sites in S2, the white down-pointing triangle represents the sampling sites in S3, the black down-pointing triangle represents the sampling sites in S4, and next to these symbols are the numbers of sampling sites, the black circle represents the reference sequences gained from GenBank, and the two capital letters and six numbers in brackets are the accession numbers of the reference sequences. Bootstrap values close to the nodes are greater than 50% of 1000 replicates. The scale indicates 0.05 nucleotide substitutions per site

Fig. 6
figure 6

Neighbor-joining phylogenetic tree of nirS gene sequences in all sampling sites in Yangtze River Estuary, where the increment represents the sampling sites in S1, the black up-pointing triangle represents the sampling sites in S2, the white down-pointing triangle represents the sampling sites in S3, the black down-pointing triangle represents the sampling sites in S4, and next to these symbols are the numbers of sampling sites; the black circle represents the reference sequences gained from GenBank, and the two capital letters and six numbers in brackets are the accession numbers of the reference sequences. Bootstrap values close to the nodes are greater than 50% of 1000 replicates. The scale indicates 0.05 nucleotide substitutions per site

For AOB, it is worth noting that the species in the four sampling regions showed obvious regional characteristics. In S1, the species were similar to soil (EU625169). In S2, the species were close to Nitrosospira (AY123821), Nitrosovibrio (AY123824), and Nitrosomonas (AF272406). In S3, the species were similar to species in paddy soil (FJ755469). Moreover, the species in S4 were similar to Nitrosospira sp. N15 (AY12382). The regional distribution of AOB suggested that variation in the environmental factors may influence the patterns of microorganisms. The sediments at sampling sites 1, 5, 6, 9, 12, 13, and 14 were collected far from shore, whereas those from 2, 3, 4, 7, 8, 10, and 11 were collected near the shore. It was obvious that species in the near-shore samples 2, 3, 7, 8, and 11 were similar to species in soil sediments (EU625169, FJ755469, and EU625023), whereas the far-shore samples were less regular. This finding might be attributed to the edge effect in ecotones (Lei et al. 2011), which is generally characterized by acute variation in the composition and structure of species (Miao et al. 2006). Located in the freshwater ecotone of the community, S1 might have marginal effects on the microbial community providing biogeographical habitats (Zhu et al. 2013).

The anammox bacteria, except the species in S4, were similar to Candidatus Scalindua. sp. (JN704474) and Candidatus Brocadia sp. (JN704470), and in S3, the species were similar to Candidatus Kuenenia sp. (JN704475); no obvious similar strains were found for the species in S1 and S2.

For the denitrifying microbes (based on nirS gene), the species in S1 were similar to those in a Tibetan Plateau wetland (KC468943), intertidal sediments of the Yangtze Estuary (KM891947), and the sediments of the Jiulongjiang and Minjiang clones (JX941962); the species in S2 were similar to those in the San Francisco Bay (KR060797) and the Yellow River Estuary (KX952167); the species in S3 were similar to those in intertidal marshes (KT444066) and Yangtze lakes (MF776477); and the species in S4 were similar to those in sediments of the Yangtze Estuary and its adjacent area in the East China Sea (MF420548). Overall, the species in S4 were less complex than the others, which might be attributed to the stress of salinity.

Gene abundances at the sampling sites

The order of gene abundance of the four functional genes in the sampling regions was S1 > S4 > S2 > S3. The specific values are shown in Additional file 1: Table S4. Among all sampling regions, the gene abundance of nirS was significantly higher than that of bacterial amoA (Student’s t test, P = 0.011), archaeal amoA (P = 0.011), and hzsB (P = 0.031). Additionally, the gene abundance of bacterial amoA was significantly higher than hzsB (P = 0.008). The results in Fig. 7 show that the abundance of bacterial amoA was (by dry weight) 4.80 × 104 (S2) − 1.93 × 107 copies g−1 sediments (S4). The range in abundance of archaeal amoA was (by dry weight) 2.88 × 104 (S3) − 4.78 × 106 copies g−1 sediments (S1). In addition, the hzsB gene abundance was (by dry weight) 2.97 × 105 (S2) − 6.77 × 106 copies g−1 sediments (S1). Finally, the abundance of the nirS gene was (by dry weight) 2.61 × 106 (S3) − 9.50 × 108 copies g−1 sediments (S1).

Fig. 7
figure 7

Copy numbers of four microbial genes in four sampling regions. Values are presented as means of the triplicate samplings, and the error bars are standard deviations of triplicate PCR assays

Obviously, except for bacterial amoA, the greatest abundance of the other genes was found in S1, indicating that AOB bacteria are adaptable to relatively higher salinity levels, whereas the others are suitable to relatively lower salinity conditions (Blainey et al. 2011; Mosier et al. 2012). However, the AOB in the estuary of Hangzhou Bay were not sensitive to salinity variation, whereas the AOA were abundant under moderate salinity (Zhang et al. 2015). In addition, the bacterial amoA was highest at sampling site 14 in S4, which might be because the highest salinity level was measured there (4.17 psu). The highest hszB gene abundance of anammox bacteria was at site 2 in S1. Moreover, the abundance of archaeal amoA was highest at site 3 in S1, and there was no variation trend in the other sampling regions. Hence, it could be expected that S1 might be the ecotone for the growth of anammox bacteria and AOA microorganisms. Although the conditions at S1 promoted a higher abundance of archaeal amoA, hszB, and nirS genes, not all bacteria were rich in diversity. One explanation for the higher abundances of bacterial amoA and nirS and richer diversity of AOA and denitrifying microbes (based on nirS gene) in S4 is that the exchange of salt and freshwater in the intersection zone resulted in unequal salinity, which is one of the important factors affecting the structure of the microbial community (Awata 2012). Compared to the other bacteria, the gene abundance of nirS was one or two orders of magnitude higher, especially in S1 and S4, and the diversity was highest, especially in S2. Salinity did not seem to restrict the growth and richness of denitrifying microbes in this research. In the Yangtze Estuary, Zheng found that the diversity of the nirS gene was significantly richer at the lower salinity sites than at the higher salinity sites, and the abundance was merely related to salinity in the intertidal sediments (Zheng et al. 2015).

Regarding the near-shore and far-shore samples mentioned above, an obvious trend was found; namely, all the far-shore sediment samples had a higher bacterial amoA gene abundance than did archaeal amoA, whereas in the sediments from the near-shore area, the former showed no superiority. This finding was similar to the results of Kim, who found a higher gene abundance and activities in near-shore sediments than in stream sediments (Kim et al. 2016). This finding offers an explanation for the higher abundance of the archaeal amoA gene at near-shore sampling site 3.

Correlation analysis

Anammox, denitrification, and aerobic ammonia oxidation processes have synergistic effects (Zhou 2006). Anammox, which uses NH4+-N as the electron donor, nitrate or nitrite as the electron acceptor, and CO2 as the carbon source, produces a certain amount of N2 (Mulder 1995). Denitrification converts nitrate or nitrite into N2, consuming chemical oxygen demand (COD), producing CO2, and providing inorganic carbon sources for anammox. Aerobic ammonia oxidation converts NH4+-N to nitrate or nitrite by AOB and AOA, providing reactors for the anammox and denitrification processes (Xu et al. 2017). It could thus be inferred that nitrogen-related microbes were correlated with each other because of the interaction within the nitrogen cycle. The correlation analysis in Table 2 showed that bacterial amoA with nirS, archaeal amoA with nirS, archaeal amoA with hzsB, and nirS with hzsB were significantly correlated, suggesting that aerobic ammonia oxidation is correlated with denitrification (defined by nirS gene in this work) and anammox. Moreover, some correlations were found in S2 and S3, and the aerobic ammonia oxidation was closely related to the denitrification process and anammox, implying that there are interactions within the moderate-salinity region. Zhi et al. (Zhi and Ji 2014) confirmed an interaction and ecological association among nitrogen-related microbial communities due to their similar environmental conditions and interrelated but distinct ecological niches (Ji et al. 2012).

Table 2 Correlation analysis between the abundance of the four functional genes. Correlated microbe pairs are highlighted in light gray

Environmental parameters, such as pH, temperature, salinity, and the concentrations of ammonium, nitrate, and phosphorus, affect the growth and metabolism of N-related microbes, which are commonly found in terrestrial and marine sediments (Zhang et al. 2015; Zhou et al. 2016; Lee and Francis 2017). The correlation analysis of the environmental factors with the gene abundances of the four functional genes are shown in Table 3. There was a significant correlation between bacterial amoA gene abundance and salinity in every sampling region, but there were no correlations between the other microbial genes and the ammonia or nitrate contents. The significant correlation suggested that salinity is the dominant limiting factor for AOB, which was consistent with the regional pattern exhibited in the phylogenetic tree. Salinity appeared to be an influential factor for archaeal and bacterial cells (Yang et al. 2015). In this study, it promoted abundance while restricting the diversity of AOB and had reverse impacts on AOA. Specifically, the abundance of the bacterial amoA gene was highest in S4, whereas the abundance of archaeal amoA was the highest in S1; the diversity of AOB was richer in S2, whereas that of AOA was richer in S3 and S4, implying that salinity is conducive to the richness and diversity of some microbial groups. Since the activity of anammox bacteria, represented by Candidatus Brocadia, could be inhibited by the presence of nitrite (Oshiki et al. 2011), and the anammox process was likely limited by both nitrite and O2 but not by the ammonium concentration (Verhamme et al. 2011); this finding indicates that the nitrogen content is potentially irrelevant.

Table 3 Correlation analysis between the gene abundances of the four functional genes and the environmental parameters in the four sampling sites. Correlated microbe pairs are highlighted in light gray

Suspects regarding nitrogen transformation

NH4+-N, NO2-N and NO3-N are important reactive substrates in N transformation. When nitrite, which is produced from nitrate denitrification, is inadequate in oxidizing NH4+-N, part of NH4+-N may be nitrified by O2 (Lai and Zhou 2010). However, the presence of excess nitrite could irreversibly inhibit nitrogen removal and restrict synthetization processes because of protein deficiency (Wang et al. 2016b). The accumulated nitrite might participate in the anammox process and be transformed into N2 (Engström et al. 2005). Previous research on the effects of functional gene groups on the N cycle transformation rate (Xu et al. 2018) found that the ammonia oxidation executed by amoA, anammox, and nitrite oxidoreductase (nxrA) contributed 93% to NH4+-N oxidation and 73% to nitrite; in addition, nxrA and nirS contributed 90% to denitrification. Anammox bacteria have been proven to contribute up to 50% of N2 production in marine sediments (Zhu et al. 2010). In mangrove sediments distributed along estuaries, the anammox process accounted for 57.5% of N2 production, and denitrification accounted for 30–58% in summer (Cao et al. 2017). Li’s study found that the ratio of anammox and denitrification in a lateral subsurface wetland was 0.75 to 0.9:1 (Li 2015). In this study, the NH4+-N/NO3-N ratio at most sampling points ranged from nearly 1:3 to 2:3; at a few sampling sites, NH4+-N was higher than NO3-N. However, in Zheng’s study in the Yangtze River Estuary (Zheng et al. 2014), NH4+-N was significantly higher than NO3-N. The contrasting results in a similar study region indicate that the nitrogen content is dynamic and does not reflect the transformation process directly. Another study found that the gene abundance of archaeal amoA was closely correlated with potential nitrification rates (Zheng et al. 2014), which aligned with the lower potential nitrification rate in AOA compared with AOB in the ammonia oxidation process at most sampling sites in this study. In all sampling regions, the nirS gene was significantly richer than the hzsB gene; however, the anammox process was negligible in summer (Cao et al. 2017). Therefore, these research findings confirmed the point at which the nitrogen-transformation activities of the four microbes might coexist and react simultaneously.

Conclusion

In this study, the abundance, diversity, and distribution patterns of four nitrogen transformation-related microbes were determined along with salinity in the sediments of the Yangtze Estuary. The abundance and diversity of the nirS gene were the highest among all detected genes, whereas salinity seemed to have no correlation with the abundance. However, for AOB, it was adaptable to relatively higher salinity. As the correlated factor for AOB in this study, salinity promoted its abundance while restricting its diversity. In contrast, the resistance of archaeal characteristics allowed for the maintenance of the abundance and diversity and thus afforded strong adaptability regardless of the salinity gradients. Moreover, the spatial distribution patterns and dominant species of each microbial group varied and shifted along the sampling regions. When subjected to salinity gradients, a shift toward the distribution pattern and abundance of AOB was observed. Denitrifying microbes (nirS-type) were considered to interact with the other three microbial groups during nitrogen migration. In addition, based on the functions of identified dominant species and the nitrogen transformation in a previous study, four nitrogen transformation-related microbes were considered to coexist and might react simultaneously, which provided further insights regarding nitrogen transformation in the estuarine zone.