Spatial variations of microbial communities in abyssal and hadal sediments across the Challenger Deep

Sample collection

Fourteen sediment samples were successfully collected from the Challenger Deep in three cruises R/V Dayang 37-II (DY37II), Tansuo01 (TS01) and Tansuo03 (TS03) during June–July of 2016, June–August of 2016 and January–March of 2017 (Fig. 1). Our sediment sampling was carried out on the slopes and the trench-axis bottom sites of the Challenger Deep from abyssal to hadal depths by “Jiaolong” manned submersible, a box sampler and a hadal lander (Fig. 1; Table S1). All these sediment cores were sliced into 2 cm subsamples except for three cores (T1L10, T3L08 and T3L11 above 10,000 m) that were sectioned into 3 cm subsamples. The ion concentrations of two sediment cores were measured with an Ion chromatography system (Dionex Corporation, Sunnyvale, CA, USA). All subsamples for community analysis were preserved at −80 °C until DNA extraction. The research has been permitted and secured from the Federated States of Micronesia.

DNA extraction and quantitative PCR (qPCR) analyses

Genomic DNA was extracted from 2 g of sediment subsamples using the PowerSoil DNA Isolation Kit (MO BIO Laboratories, Inc., Carlsbad, CA, USA). The quality and quantity of the genomic DNA were checked with a NanoDrop spectrophotometer (ND-1000, Nanodrop Technologies, Wilmington, DE, USA) and gel electrophoresis. The abundance of 16S rRNA was quantified as an average of three replicate analyses. The prokaryotic SSU rRNA gene was quantified using the primer Uni516F/Uni806R using the StepOnePlus^TM Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). The primer sequences and qPCR conditions were as described previously (Nunoura et al., 2018).

PCR amplification of the 16S rRNA genes

The 16S rRNA genes were amplified with a pair of universal primers: 341F (5′-CCTAYGGGRBGCASCAG-3′) (Zakrzewski et al., 2012) with a tagged six-nucleotide(nt) barcode and reverse primer 802R (5′-TACNVGGGTATCTAATCC-3′) (Nossa et al., 2010) that target V3-V4 variable regions. The PCR reaction was prepared according to the PrimerSTAR^® HS DNA Polymerase (TaKaRa, Dalian, China) with 2 μl of forward and reverse primers (10 μM) and 2 ng of template DNA. The PCR was performed on a thermal cycler (Bio-Rad, Hercules, CA, USA) in the following thermal cycles: 98 °C for 10 s, 26 cycles of 98 °C for 10 s, 50 °C for 15 s, 72 °C for 30 s and a final extension at 72 °C for 5 min. PCR products were purified using the TaKaRa Agarose Gel DNA Purification Kit (TaKaRa, Dalian, China) and Qubit 2.0 Fluorometer (Invitrogen, Carlsbad, CA, USA) for quantification.

Analysis of the barcoded amplicons and taxonomic assignment

An equal amount of PCR products from different samples was mixed together and subjected to sequencing on the Illumina Miseq platform (2 × 300 bp) in accordance with the manufacturer’s recommendation. The sequencing data were processed and analyzed using QIIME version 1.9.1 (Caporaso et al., 2010b). After qualification, the remaining reads were assigned to operational taxonomic units (OTUs) at 97% similarity level by UCLUST (Edgar, 2010). Chimeric reads were identified and excluded using ChimeraSlayer (Haas et al., 2011) and singleton OTUs with one sequence were removed. For the remaining OTUs, the most abundant read of each OTU was selected as a representative for subsequent taxonomic classification. Taxonomic assignment was conducted using the Ribosomal Database Project (RDP) classifier (version 2.2) (Wang et al., 2007) by referring to the SILVA132 database with a confidence level of 80%. The remaining representatives were aligned by PyNAST (Caporaso et al., 2010a) and a phylogenetic tree was built using FastTree (Price, Dehal & Arkin, 2009). Subsequently, alpha diversity was calculated with three parameters: observed OTUs, Chao1 and Shannon Index. Using percentages of the genera in the microbial communities, Bray–Curtis dissimilarity between the layers was calculated and used for Principal coordinate analysis (PCoA) with the OmicShare tools, a free online platform for data analysis (www.omicshare.com/tools).

Phylogenetic analysis of 16S rRNA genes

The representative reads of the most abundant OTUs (DatasetS1.fasta) were selected for construction of phylogenetic trees with reference sequences from the Silva and NCBI database. All the sequences were aligned using MAFFT 7.31 (Katoh & Standley, 2013) and the maximum-likelihood (ML) phylogenetic trees were constructed using RaxML (Silvestro & Michalak, 2012) with GTRGAMMA model. The bootstrap values were calculated based on 1,000 replications. The phylogenetic outputs were visualized and edited in iTOL (Letunic & Bork, 2016).

Microbial abundance

In the qPCR analysis, the abundance of the entire prokaryotic SSU rRNA gene ranged between 1.5 × 10⁵ and 5.5 × 10⁸ copies g^–1 sediment from sediment surface to 66 cmbsf. The results showed that, with a few exceptions, microbial abundance generally decreased with increasing depth of sediment layers. The SSU rRNA gene copy numbers of the trench-axis at the surface layer (0–2 cmbsf) were observed at least an order of magnitude higher than those of the slopes (Fig. S1). Moreover, the abundance of the entire prokaryotic SSU rRNA at the trench-axis shown in this study was higher than the result reported in a recent work (Nunoura et al., 2018).

Composition, diversity and species richness of microbial communities

After quality control and filtration of chimeras and singletons, a total of 624,821 reads of 16S rRNA gene amplicons for 95 sediment layers were obtained (Table S2). The rarefaction of the qualified reads resulted in the minimum 1,143 sequences per sample (Table S2). All the curves did not reach the plateau (Fig. S2), indicating that more species might be discovered with more sequencing reads. The qualified reads were grouped into 123,955 OTUs and were then classified into 69 phyla. The 11 most dominant bacterial phyla (>1%) were Proteobacteria, Chloroflexi, Actinobacteria, Planctomycetes, candidate phylum Patescibacteria, candidate phylum Marinimicrobia, Gemmatimonadetes, Bacteroidetes, Firmicutes, Acidobacteria and candidate phylum Zixibacteria (Fig. 2). The most abundant Proteobacteria in hadal sediment samples was Gammaproteobacteria composed of Pseudoalteromonas (similarity 100% to NR_152003), Halomonas (similarity 100% to NR_043299), Pseudomonas (similarity 100% to NR_159318) and Alteromonas (similarity 100% to NR_148755). These taxa varied in their percentages spanning 0–63% across all sediments layers. T1B08 was particularly enriched with Gammaproteobacteria highly represented by Alteromonas. Planctomycetes and Thaumarchaeota were significantly negatively correlated in their relative abundances (Correlation test with 95 samples; P < 0.00001). Two archaeal phyla were abundant in the samples: Thaumarchaeota and Nanoarchaeaeota. Nitrosopumilus of Thaumarchaeota was the major group in the Challenger Deep sediment layers, while the latter represented by “Ca. Woesearchaeota” was abundantly detected in deep layers of the deepest sediments (>10,000 m). The high abundance of Euryarchaeota archaea was exclusively shown in the layers of the T1B08 sample (Fig. S3).

The observed OTUs, Chao1 and Shannon index were normalized based on the minimal number of reads (Table S2). The sediments from <6,000-m depths were associated with the highest Chao1 and Shannon index such as in DMC02, DD114 and DD121. The diversity indexes for the different layers in the T1B08 were the lowest among the sediments. The communities in the bottom sediments >10,000 m depths were not more diversified than those in the slopes except for the T1B08. Moreover, from the surface to the bottom layers, the biodiversity indexes did not vary notably for all the samples.

Clustering relationships of microbial communities

A PCoA plot could separate the microbial communities into two groups (the surface 0–2 cmbsf layer samples were shown, simple IDs refer to Table S3), one composing of the slope samples (5,400–7,800 m) and another containing only trench-axis bottom samples (8,600–11,000 m). In the former group, the samples from the northern slope could be further split from those from the southern slope. The T1B08 sample was far from the two groups in the plot (Fig. 3A), reflecting its distinct composition of microbial community. Principally, the microbial communities were clustered based on the geographic locations and depths. The combination of PCO1 and PCO2 explained 53% of the differences among the communities (Fig. 3A). Hierarchical clustering of the samples using the percentages of genera in the corresponding communities also demonstrated that the microbial communities in trench-axis sediments differed from those in the slopes (Fig. 3B). The similarity of the microbial communities was determined by sampling sites, rather than depths to the surface layers, since the communities from the same cores tend to be grouped together (Fig. S4).

Phylogenetic analysis of 16S rRNA genes

The most abundant OTUs (DatasetS1.fasta) were used to determine their phylogenetic positions. The representative reads for a total of 35 OTUs were selected for the construction phylogenetic tree with their closest relatives in the NCBI. The most abundant one was within the same group with a neighbor of Nitrosopumilus (Fig. 4A). The Nitrosopumilus in the Challenger Deep sediments resembled those inhabiting other abyssal and hadal waters and sediments from the Ogasawara Trench, the Japan Trench and even the Puerto Rico Trench, with respect to their short phylogenetic distance. This has been observed and discussed in a recent work on the high homogeneity of Nitrosopumilus AOA in hadal zones (Wang et al., 2019b). The phylogenetic position of OTU100704 particularly enriched in T1B08 showed high affinity (100% identity) to mesophilic archaea of Marine Group II (MGII) (estimated optimum growth temperature: T_opt:= 40.76 (Kimura et al., 2013)) isolated from the Juan de Fuca hydrothermal flume (Anderson et al., 2013). The five OTUs affiliated with Nanoarchaeaeota are Ca. Woesearchaeota widely distributed in the Challenger Deep sediments from 5,481 to 10,953 m. Similar sequences were also obtained from saline water. The five OTUs in our results can split into two groups, and thus there would be more potentially novel species in this phylum.

Actinobacteria, Cyanobacteria, Firmicutes and Chloroflexi as members of superphylum Terrabacteria are almost ubiquitous in the sediments. The six OTUs of Chloroflexi in ML phylogenetic tree were split into four major groups (Fig. 4B), in which OTU124775 seems to represent a novel taxon. Interestingly, OTU344095 was approximate to Dehalococcoides spp., a known genus of strictly anaerobic bacteria with capacity of gaining energy from the reduction of chlorinated compounds (Tas et al., 2010). Furthermore, “Ca. Actinomarinales”, described as a new group of photoheterotrophic Actinobacteria with ubiquitous presence in the pelagic layer of the oligotrophic ocean (Ghai et al., 2013), was also found in sediment layers especially in surface layer of T1B08. Furthermore, a proportion of 16S rRNA OTUs in the surface layer of T1B08 belonged to the picophytoplankton Synechococcus (3%) and Prochlorococcus (7%) in Cyanobacteria phylum (Fig. 4B). The two Cyanobacteria genera and “Ca. Actinomarinales” originally present in the euphotic layer were rarely detected in the other sediment samples. The abundance of these euphotic bacteria in T1B08 surface layer was probably buried relicts from depositional material (Kirkpatrick, Walsh & D’Hondt, 2016) or might be derived from contamination during the operation of the box-core sampler.

FCB (Fibrobacteres, Chlorobi, Bacteroidetes) is another superphylum mainly containing Chlorobi, candidate phylum Marinimicrobia and Bacteroidetes (Bertagnolli et al., 2017; Getz et al., 2018; Hug et al., 2016). Nowadays, more new taxa within the FCB superphylum were recovered from sediment metagenomes (Castelle et al., 2013; Parks et al., 2017; Rinke et al., 2013), indicating their prevalence in various environments on Earth. In this study, OTU64116 was grouped with references of a newly defined phylum Zixibacteria (Castelle et al., 2013). Some of the Marinimicrobia OTUs were grouped with the sequences from previous studies, while the others represented novel groups that might have evolved in the hadal sediments (Fig. 4C).