Diversity of NC10 bacteria associated with sediments of submerged Potamogeton crispus (Alismatales: Potmogetonaceae)

Background The nitrite-dependent anaerobic methane oxidation (N-DAMO) pathway, which plays an important role in carbon and nitrogen cycling in aquatic ecosystems, is mediated by “Candidatus Methylomirabilis oxyfera” (M. oxyfera) of the NC10 phylum. M. oxyfera-like bacteria are widespread in nature, however, the presence, spatial heterogeneity and genetic diversity of M. oxyfera in the rhizosphere of aquatic plants has not been widely reported. Method In order to simulate the rhizosphere microenvironment of submerged plants, Potamogeton crispus was cultivated using the rhizobox approach. Sediments from three compartments of the rhizobox: root (R), near-rhizosphere (including five sub-compartments of one mm width, N1–N5) and non-rhizosphere (>5 mm, Non), were sampled. The 16S rRNA gene library was used to investigate the diversity of M. oxyfera-like bacteria in these sediments. Results Methylomirabilis oxyfera-like bacteria were found in all three sections, with all 16S rRNA gene sequences belonging to 16 operational taxonomic units (OTUs). A maximum of six OTUs was found in the N1 sub-compartment of the near-rhizosphere compartment and a minimum of four in the root compartment (R) and N5 near-rhizosphere sub-compartment. Indices of bacterial community diversity (Shannon) and richness (Chao1) were 0.73–1.16 and 4–9, respectively. Phylogenetic analysis showed that OTU1-11 were classified into group b, while OTU12 was in a new cluster of NC10. Discussion Our results confirmed the existence of M. oxyfera-like bacteria in the rhizosphere microenvironment of the submerged plant P. crispus. Group b of M. oxyfera-like bacteria was the dominant group in this study as opposed to previous findings that both group a and b coexist in most other environments. Our results indicate that understanding the ecophysiology of M. oxyfera-like bacteria group b may help to explain their existence in the rhizosphere sediment of aquatic plant.


INTRODUCTION
The region known as the rhizosphere of submerged aquatic plants is generally several millimeters in thickness and surrounds the plant roots. It is a "hotspot" for many biogeochemical interactions between the plant root system and microbial processes (Hartmann, Rothballer & Schmid, 2008). Environmental factors related to sediment processes and plant metabolism, including oxygen consumed or produced (Bodelier, 2003) and nutrient exudates secreted by plants (Haichar et al., 2014), influence the community structure of the rhizosphere and make microbial diversity of this region distinct from that of the surrounding sediments. Microbial activity related to nitrogen transformations is especially active in the plant rhizosphere. It is mediated by nitrogen cycling bacteria which exist in close association with plant roots.
The diversity and abundance of ammonia-oxidizing bacteria have been reported to be higher in the plant rhizosphere than in the bulk soils in calcareous regions (Ai et al., 2013). The abundance of denitrifying bacteria in the rhizosphere of rice and wheat have also been found to be higher than in the non-rhizosphere soil (Hussain et al., 2011;Hamonts et al., 2013). Ativities of the rhizosphere microbial community, notably microbial nitrogen cycling, are often influenced by the species of host plant (Ruiz-Rueda, Hallin & Bañeras, 2009;Hu et al., 2015). For example, Iris pseudacorus and Typha orientalis have been found to support higher diversity of anaerobic ammonium oxidizing bacteria in rhizosphere sediments than in Thalia dealbata (Chu et al., 2015). Nitrogen cycling microbes are the main drivers of denitrification in lakes (Zhao et al., 2015) and therefore understanding microbial diversity in rhizosphere sediments could assist with enhancing remediation of lake eutrophication.
Candidatus Methylomirabilis oxyfera, or M. oxyfera bacteria, belong to the NC10 phylum and are not currently available from pure cultures (Ettwig et al., 2010). Some M. oxyfera-like bacteria can couple the oxidation of methane with nitrite reduction under anaerobic conditions using a process called nitrite-dependent anaerobic methane oxidation (N-DAMO) (Raghoebarsing et al., 2006;Ettwig et al., 2010). These organisms have an internal mechanism of oxygen generation, which is considered to be the fourth major biological pathway for oxygen production, and potentially has a substantial impact on global cycling of carbon and nitrogen (Wu et al., 2011;Ettwig et al., 2012). Numerous studies have been undertaken on N-DAMO microbes inhabiting aquatic ecosystems and the existence of M. oxyfera-like bacteria. The 16S rRNA genes of M. oxyfera-like bacteria have been detected in marine (Chen, Zhou & Gu, 2014), estuarine (Shen et al., 2014b;Chen, Zhou & Gu, 2015;Yan et al., 2015), lake (Deutzmann & Schink, 2011;Kojima et al., 2012;Wang et al., 2017), river (Shen et al., 2014a), natural wetland , peatland (Zhu et al., 2012), and constructed wetland ecosystems . Ettwig et al. (2009) divided the 16S rRNA genes of NC10 bacteria into four groups: a, b, c, and d. Group a and group b are considered to be the dominant branches. Currently known M. oxyfera-like bacteria with N-DAMO function belong to group a (Welte et al., 2016). However, group b were found in most natural environments and occupy the majority of the M. oxyfera-like bacteria community in many freshwater environments Kojima et al., 2012;Shen et al., 2015). Although M. oxyfera-like bacteria of the NC10 phylum are widespread, they have not been reported in the rhizosphere of aquatic plants.
Many studies have shown that anaerobic microbes, such as denitrifying bacteria and anaerobic ammonia-oxidizing bacteria, existed in the rhizosphere of aquatic plants (Lu et al., 2014;Zheng et al., 2016), therefore we speculated that M. oxyfera-like bacteria may also exist in this region. In this study, a rhizobox approach was used to experimentally cultivate submerged plants from a freshwater lake to detect the occurrence and diversity of M. oxyfera-like bacteria in the rhizosphere and non-rhizosphere zones of the submerged plants.

Test sediments and plants
Sediment samples were collected from Liangzi Lake (114 38′23″N, 30 14′28″E), located in Hubei Province of China. The basic characteristics of the sediments were 72.5 ± 1.3% water content, 8.85 ± 0.86 g kg -1 of organic matter (OM), 0.50 ± 0.06 g kg -1 of total nitrogen and 0.45 ± 0.05 g kg -1 of total phosphorus. The chemical composition of the interstitial water was 4.39 ± 1.29 mg l -1 of ammonium-nitrogen (NH 4 + -N), 0.22 ± 0.07 mg l -1 of nitrate-nitrogen (NO 3 --N) and pH of 6.99 ± 0.08. Potamogeton crispus (Potmogetonaceae), a common, perennial, herbaceous, submerged plant, was selected as the plant host in this experiment. This plant is also a native of lakes of southern China.

Design of rhizobox
Potamogeton crispus was cultivated using a three-compartment rhizobox with multiple interlayers. The rhizobox was modified under the design of He et al. (2005) (Fig. 1). Each plant was fully submerged into the bottom of a water tank (height of 350 mm) during cultivation. Air-dried sediment was used to fill the three compartments. The rhizobox had three sections: a root compartment (20 mm in width), near-rhizosphere compartment (five mm), and non-rhizosphere compartment (>5 mm). The near-rhizosphere compartment was further separated into five sub-compartments (one mm thick) by nylon mesh (<25 mm) in order to prevent root hairs entering the adjacent sediment interlayers and to keep microbial and root exudates separated in the sediment interlayers.

Plant cultivation and sample collection
Six P. crispus turions showing similar germination times and growth characteristics were selected to plant in the root compartment. Plants were cultivated for 6 months, from November 2014 to May 2015. In order to simulate three nutritional conditions (low, intermediate and eutrophic nutrient status) of freshwater lakes, three regimes of slow-release urea (Luxi Chemical Co., Ltd, Liaocheng, Shandong, China) as a nutrient input were used at concentrations of 0, 400, and 600 mg urea per kg sediment. Each concentration nutrient input experiment consists of three replicate rhizoboxes. After 5 months, the root systems of the plants filled the root compartments and slow-release urea was injected into the sediments. Three parallel sediment samples for each nutritional condition were collected from the root compartment (R), five near-rhizosphere sub-compartments (N1-N5), and the non-rhizosphere compartment (Non) at 14-day intervals, starting at day 0, for 56 days after urea injection. The sediment samples collected from the same interlayer were combined and then seven mixed samples were used for DNA extraction, cloning and sequencing to detect whether M. oxyfera-like bacteria exist in every interlayer.

Physicochemical measurement and genomic DNA extraction
Before the sediments were collected, the dissolved oxygen (DO) and pH of the rhizosphere sediment were measured in situ using a microelectrode system (Unisense, Aarhus, Denmark). The measured depth was 48 mm due to limited length of probe of the microelectrode system. Depth of oxygen permeation (Dop, DO > 6.25 mmol l -1 ) in each near-rhizosphere sub-compartment was identified based on the measured DO values. The collected sediments were centrifuged for 10 min at 4,000 rpm to obtain interstitial water. The concentration of NH 4 + -N and NO 3 --N of interstitial water was measured using flow injection analysis (SEAL Analytical AA3; SEAL Analytical, Norderstedt, Germany). The genomic DNA of the sediments was extracted using a Fast DNA Spin Kit for Soil (MP Biomedicals, Fountain Parkway Solon, OH, USA) according to the manufacturer's instructions. Approximately 0.5 g of sediment was used for DNA isolation. The extracted DNA were then stored at -20 C until further analysis. Concentrations of the DNA were determined using a NanoDrop 2000 UV-Vis Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), and the quality was checked by electrophoresis on a 1.2% agarose gel.

Statistical analysis
Analysis of variation (Least-Significant Difference (LSD) and Tukey's test) was used to detect significant differences amongst physicochemical indices of sediments and interstitial water. Operational taxonomic unit (OTU) cut-off values of 3% were applied to determine the 16S rRNA genetic diversity of M. oxyfera-like bacteria, and furthest neighbor method was used in sequences clustering using Mothur program (v.1.34.4) (Schloss et al., 2009). Multiple sequence alignment was conducted with ClustalW 1.6 program. Phylogenetic analyses of the 16S rRNA gene sequences were conducted with Mega 6 software using the neighbor-joining method. The calculation model was Jukes-Cantor model. Bootstrap analysis with 1,000 replicates was applied to examine the confidence levels of the clustering of the trees (Shen et al., 2014a). The Chao1 estimator and the Shannon index were generated using Mothur software (v.1.34.4) to assess diversity. Sequences obtained in this study were BLAST with M. oxyfera bacteria sequence (FP565575) in the NCBI database to get their similarities and uploaded to GenBank under accession numbers MH092300-MH092623.

Diversity of M. oxyfera-like bacteria 16S rRNA genes
A total of 324 sequences of NC10 phylum bacteria were obtained in this study (Table 2). All sequences were divided into 16 OTUs based on 97% similarity. Approximately 45-50 sequences were detected in each rhizobox interlayer, representing four to six separate OTUs (Table 2). A maximum of six OTUs was found in the N1 near-rhizosphere subcompartment while a minimum of four OTUs was found in the root compartment (R) and N5 near-rhizosphere sub-compartment. The library coverage values ranged from 0.94 to 1.00, indicating that the 16S rRNA gene sequences of M. oxyfera-like bacteria in rhizosphere sediment of P. crispus were sufficiently over-represented in these clone libraries. The N2 near-rhizosphere sub-compartment had the highest diversity, with the Shannon index and Chao1 richness estimators of 1.16 and 5, respectively. The lowest diversity was observed in the N3 near-rhizosphere sub-compartment, with Shannon index and Chao1 richness estimators of 0.73 and 6, respectively.

Phylogenetic diversity of M. oxyfera-like bacterial 16S rRNA genes
Representative sequences of OTU1-11 were classified into group b according to Ettwig et al. (2009), with 88-93% identity to the 16S rRNA gene of M. oxyfera (Table 3). The group b sequences were sorted into sub-clades designated clusters from 1 to 5. OTU12 represents a new cluster of NC10 bacteria and does not belong to group a, b, c, or d (Fig. 2). The sequence in cluster 1 showed 92% identity with the 16S rRNA gene of M. oxyfera from the non-rhizosphere compartment. Cluster 2 comprised OTU2, OTU8, OTU10, OTU11, and OTU15. This cluster contained 73 sequences from all interlayers except

Notes:
R and Non represent sediment samples in root compartment and non-rhizosphere compartment, respectively. N1-N5 represent sediment samples taken from one to five mm of near-rhizosphere sub-compartments.
-Indicates that the root system hindered microelectrode measurement in the root compartment. Different letters after values in the same column indicate significant difference (P < 0.05). * Indicates that the non-rhizosphere was not measured. the N5 near-rhizosphere sub-compartment and had 90-92% similarity to M. oxyfera. The cluster 3 was the dominant in group b, which consisted of the largest number of sequences (205 of the total found of 324). Cluster 3 members recorded in all interlayers and showed 91-92% identity with the M. oxyfera. Cluster 4 had the highest similarity to M. oxyfera (93%), and it contained 42 sequences from all interlayers except the root compartment. Cluster 5 and OTU12 were detected from N1 and N3 near-rhizosphere sub-compartments, showing 88-89% sequence similarity with M. oxyfera (Table 3).

DISCUSSION
In this study, we used a rhizobox approach to investigate the occurrence and diversity of M. oxyfera-like bacteria in the rhizosphere and non-rhizosphere of a common submerged plant. The results show that these bacteria are present in both the rhizosphere and the nonrhizosphere sediments. Most of the OTUs were classified into group b. There were four to six OTUs in every rhizobox interlayer. And indices of bacterial community diversity (Shannon) and richness (Chao1) were 0.73-1.16 and 4-9, respectively. The diversity for the compartments ranked near-rhizosphere > root compartment > non-rhizosphere (Table 2), indicating that the near-rhizosphere are more favorable for the M. oxyfera-like bacteria community. Some studies have found that denitrifying bacteria (Ruiz-Rueda, Hallin & Bañeras, 2009), including anaerobic ammonium oxidation bacteria  and anaerobic methane-oxidizing bacteria (Vaksmaa et al., 2016) distributed around the root systems of submerged aquatic plants. In our study, both the   Figure 2 Neighbor-joining phylogenetic tree of M. oxyfera-like bacterial 16S rRNA gene sequences in rhizosphere and non-rhizosphere sediments of P. crispus. R and Non represent sediment samples in root compartment and non-rhizosphere, respectively. N1-N5 represent sediment samples taken from one to five mm of near-rhizosphere sub-compartments. The numbers in the brackets represent the ratio of the number of sequences out of the total number in the corresponding interlayer. The numbers at the nodes are percentages that indicate the levels of bootstrap support based on 1,000 replicates, and only percentages more than 50% are shown. The scale bar represents 1% sequence divergence.
Full-size  DOI: 10.7717/peerj.6041/ fig-2 methane (CH 4 ) by methanogenic archaea (Kerdchoechuen, 2005). Furthermore, both partial denitrification and nitrification in the rhizosphere may produce nitrite (Nie et al., 2015). These substrates are conducive for the growth of M. oxyfera-like bacteria in the near-rhizosphere sediments. Sequences obtained in the seven interlayers of this study had high similarity to the M. oxyfera-like bacterial 16S rRNA gene sequences recovered from other aquatic ecosystems. Clusters in our study were closely related to the sequences in the sediments of an estuary (Yan et al., 2015), wetlands (Wang et al., 2012Hu et al., 2014), lakes (Wang et al., 2015(Wang et al., , 2017Kojima et al., 2012) and anaerobic sewage sludge (Ma et al., 2017). The sequences of group b dominated M. oxyfera-like bacteria in the rhizosphere and non-rhizosphere of P. crispus (Fig. 2). Functional gene pmoA was not amplified in this study. Similarly, only group b of M. oxyfera-like bacterial 16S rRNA genes sequences has been found in sediments of two freshwater lakes in China, Lake Dongchang, and Lake Dongping (Wang et al., 2017). In the sediments of the Three Gorges reservoir, most of the sequences (65/67) were group b  and similarly in the sediments of Lake Biwa (5/6 OTUs belonged to group b) (Kojima et al., 2012) and an urban wetland (8/11 OTUs belonged to group b) (Shen et al., 2015). In contrast, more than 87% M. oxyfera-like 16S rRNA genes sequences in sediments of the Jiaojiang estuary (Shen et al., 2014b) were classified into group a. There is no evidence that communities consisting primarily of group b bacteria can dominate N-DAMO without group, a and the two appear to coexist in the majority of natural environments. The functions attributed to group b need to be further studied to improve our understanding of NC10 bacterial communities in rhizospheres.
The diversity indices of M. oxyfera-like bacteria of this study were similar to that in other freshwater sediments from rivers (Shen et al., 2014a) and wetlands . However, in natural environments, a higher diversity of M. oxyfera-like bacteria is often found in marine and estuarine sediments (Chen, Zhou & Gu, 2014) compared with freshwater sediments from rivers and wetlands (Yan et al., 2015;Shen et al., 2014a;Wang et al., 2016) and particularly lakes (Deutzmann & Schink, 2011;Kojima et al., 2012;Wang et al., 2017). Recently, a global analysis on the distribution pattern of the M. oxyferalike bacteria demonstrated that a significant community difference was found between the marine and freshwater habitats (Zhang, Liu & Gu, 2018). Therefore, community difference might be a major reason why the diversity of this bacteria is higher in ocean than in freshwater.
In this experiment, samples of the same interlayer were mixed and sequenced. Although the mixture was able to detect all M. oxyfera-like bacteria, this method could not statistically analyze the significant differences of OTUs in different interlayers. Therefore, the spatial distribution of this bacteria was not analyzed in this paper.

CONCLUSION
Methylomirabilis oxyfera-like 16S rRNA genes were detected in sediment of all the three compartments of P. crispus. There were less OTU in rhizoshpere and only four to six OTUs existed in each interlayer. The Shannon and Chao1 indices were similar to that in sediments of freshwater wetlands and rivers. Phylogenetic analysis showed that all the OTUs were classified into group b of M. oxyfera-like 16S rRNA genes, except for one OTU into a new cluster of NC10 bacteria, which suggests that the group b bacteria may be important for nitrogen biogeochemical cycles and may play an important role in regulating eutrophication in freshwater systems.