Upper-Layer Bacterioplankton Potentially Impact the Annual Variation and Carbon Cycling of the Bathypelagic Communities in the South China Sea

Abstract

Pelagic bacterioplankton exhibit biogeographical patterns linked with exporting organic carbon and energy fluxes into the deep ocean. However, knowledge of the mechanisms shaping deep-sea bacterial communities remains largely elusive. In this study, we used high throughput sequencing of the 16S rRNA gene to reveal significant annual bacterioplankton community dynamics in the South China Sea during three summer cruises (2016–2018). As we expected, the epipelagic–bathypelagic connective amplicon sequence variants (ASVs, mostly belonging to Actinobacteriota, Firmicutes, and Cyanobacteria) suggested that they not only affect the community structure but also influence the carbon cycling functions of bathypelagic bacterioplankton in different years. However, the microbial source tracking (MST) analysis indicated that the directly linked proportions between the bathypelagic and epipelagic samples were minimal. Thus, the epipelagic bacteria communities may form “seeds” rather than directly sinking into the deep ocean to influence bathypelagic bacteria. This study provides a new perspective on the mechanisms shaping the deep ocean bacterioplankton communities.

1. Introduction

The deep ocean is of great biogeochemical interest, as it is the site for the global storage of ocean carbon [1,2] and nitrogen [3] as well as sulfur cycling [4]. Approximately 4 gigatons of carbon, which is roughly equivalent to the total standing carbon stock in marine biomass, are annually stored in the dark ocean [5]. As the key player in marine biogeochemical cycling [6], bacterioplankton also regulates the critical processes of ocean carbon cycling via diverse microbial processes. On the one hand, phototrophic bacterioplankton is the significant producers of primary production; conversely, their heterotrophic counterparts are major remineralizers and decomposers of particulate organic carbon (POC) in the ocean. Most POC is remineralized to CO₂ via microbial respiration rather than flowing to the higher trophic level by the consumption and modification; thus, only 5–25% of the fixed carbon leaves comes from the euphotic zone, but about 1% of that reaches the sediments [7,8]. Furthermore, bacterioplankton can transfer labile-dissolved organic carbon (LDOC) to refractory-dissolved organic carbon (RDOC) and finally form an RDOC pool, which can store the carbon in the deep sea over thousands of years [9,10]. Overall, the communities and ecological functions of bacterioplankton in the dark ocean remain largely unknown.

Traditionally, the deep ocean is considered to be a relatively uniform environment [11], and the shift of the deep-sea bacterioplankton communities is reduced by the limited mixing between water masses [12] or modulated via advection [13]. Until recently, emerging lines of evidence suggested the deep-sea bacterial communities also have significant temporospatial variations [14,15,16]. Primarily, organic particles are readily colonized by diver bacterioplankton and other micro-eukaryotes; upper-layer microbes were suggested to largely determine the structure, annual variations, functioning, and biogeography of the deep-sea bacterioplankton communities [17,18]. However, current knowledge of the dynamic mechanism of temporal variations in deep-sea bacterioplankton communities and their potential connection with their upper-layer counterparts is still minimal.

The South China Sea is one of the largest marginal seas in the world, possessing high biodiversity and essential carbon cycling functions [19,20]. In this study, we examined the community dynamics of bacterioplankton from surface to bathypelagic in the South China Sea during three summer cruises (2016–2018). The simultaneous analyses of surface and deep-layer bacterioplankton were carried out to primarily evaluate the impact of surface bacteria on pelagic bacterioplankton communities. Overall, the ultimate goal of this study was to uncover the succession patterns and driving factors of the deep-sea bacterioplankton communities.

2. Materials and Methods

2.1. Seawater Sampling

Samples were collected from four stations (O7: 117° E, 17° N; O8: 116° E, 18° N; O9: 115.5° E, 18.83° N; and O12: 115° E, 17° N) during three South China Sea cruises in May–June 2016 (O8 and O9 stations), June–August 2017 (O7 and O12 stations), and June–August 2018 (O7 and O12 stations). Using the Sea Bird CTD rosette sampler, seawater samples were taken from the surface (SUF, 5 m), the deep chlorophyll maximum (DCM, ~75 m), the bottom of the euphotic (BOE, 200 m), the mesopelagic (MP, 500 m), and the bathypelagic (BP, 1000–2000 m) layers. For assessing the molecular community of bacteria, 4 L of seawater of each sample was filtered using 0.22 μm polycarbonate Ispore membrane filters (Millipore, Billerica, MA, USA) and stored at the −80 °C until the DNA extraction. The samples have been named in the following format: station + cruise + layer, such like O8T16BP. In each water layer, there were 5–8 seawater samples used for analysis.

2.2. DNA Extraction and High-Throughput Sequencing

The total DNA of all the samples was extracted using the E.Z.N.ATM Water DNA Kit (OMEGA, Norcross, GA, USA). The resulting DNA was suspended in 100 μL of sterile Milli-Q water and then measured using a NanoDrop (Thermo Scientific, Waltham, MA, USA). The barcode primer site for the bacterial 16S rRNA genes (V3-V4 hypervariable regions), 338F (5′-ACTCCTACGGGAGGCAGCA-3′), and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) were used for the high-throughput sequencing analysis [21]. Approximately 10 ng of DNA was used as a template in a 25 μL reaction volume containing 12.5 μL of the Taq PCR mix (Takara Bio, Otsu, Japan) and 10 μM of each primer. The PCR reactions were thermo-cycled using the following protocol: 95 °C for 3 min, 35 cycles of 30 s at 95 °C, 30 s at 55 °C, and 45 s at 72 °C, with a final extension of 72 °C for 10 min [22]. Triplicate PCR products for each sample were combined, purified with the TIANquick Midi Purification Kit (Tiangen, Beijing, China), then quantified with a NanoDrop and sequenced on the Illumina Miseq platform according to the standard protocols in Biomarker Technologies Company (Beijing, China).

2.3. High-Throughput Sequencing Analysis

Raw sequences were trimmed to remove primer and adapters sequences with CUTADAPT (Version 1.3). The plugin DADA2 was used for further quality trimming, denoising, paired-end read merging, and chimera removal. Singleton and non-bacteria ASVs were then excluded from the feature table. Taxonomy was assigned to the resulting amplicon sequence variants (ASVs) using the SILVA SSU rRNA database (Version 138). FAPROTAX (Version 1.1) was used to annotate the potential functional taxa [22].

2.4. Statistical Analysis

The alpha-diversity indices (Observed ASVs numbers, Shannon index) were calculated using the diversity plugin of QIIME 2, with an even subsampling depth of 5124 (the least sequence number across all samples). NMDS analysis based on the Bray–Curtis distance was performed with the ‘vegan’ package using R software (Version 4.1.2). Adonis analysis and pairwise permutation MANOVAs were applied to test for community differences across depths. The Kruskal–Wallis test was used to calculate the significant differences in the diversity indexes among different depths. Based on the assumption that the bathypelagic bacterial communities are affected by the upper-layer bacterial communities, a Bayesian community-based microbial source-tracking algorithm, ‘SourceTracker’ (Version 0.9), was applied to the ASV abundance table [23]. All samples were randomly sub-sampled to 5124 sequences before the SourceTracker analysis. The alpha1 and alpha2 values of the SourceTracker analysis were tuned using cross-validation and finally set at 0.001 and 0.01, respectively.

3. Results

3.1. Temporospatial Variations in Deep-Sea Bacterioplankton Diversity and Community Structures

After the sequence processing, 941,515 bacterioplankton sequences and 1541 ASVs were obtained from the 31 samples of the South China Sea. The richness and the Shannon diversity of bacterioplankton among the samples varied between 66–416 and 4.1–7.4, respectively. A significant difference in the bacterioplankton richness and the Shannon diversity (p < 0.05, Pairwise.wilcox test) was found between different depths. As shown in Figure 1, the bottom of the euphotic (BOE) layer and the surface (SUF) layer represented the highest and lowest diversity indices of the five depths, respectively. Notably, bacterioplankton diversity in the mesopelagic (MP) and bathypelagic (BP) layers were comparable (p > 0.05, Pairwise.wilcox test) to that in the surface water layers (e.g., SUF and DCM) with high productivity, suggesting unneglectable high bacterioplankton diversity occurred in the deep ocean.

Non-metric Multidimensional Scaling (NMDS) analysis showed a clear separation of the bacterioplankton communities depending on their sampling depths. Specifically, the bacterioplankton communities below the euphotic layer, including the BOE, MP, and BP layers, were significantly different (p < 0.05, permutation MANOVAs test) from that in the epipelagic (SUF and DCM) layers (Figure 2a), suggesting that numerous unique bacterioplankton are living in deep-sea waters. Notably, the NMDS plot and further Betadisper analysis indicated that the annual variation in bacterioplankton communities in the deep-sea (MP and BP) layers was even higher than that of the epipelagic layers (e.g., SUF, DCM, and BOE) (Figure 2b), indicating an unneglectable annual succession of the deep-sea bacterioplankton communities.

3.2. The Connection between Epipelagic-Layer and Deep-Sea Bacterioplankton ASVs

To uncover the mechanisms shaping the community of the deep-ocean bacterioplankton, we assessed the connection between the deep-sea (MP and BP) and epipelagic layers (SUF and DCM) of bacterioplankton communities in different cruises. In the vertical direction, many deep-sea bacterioplankton ASVs can be found in the epipelagic layers. Indeed, in each cruise, the number of ASVs that co-occurred in the deep sea and epipelagic layers were always more remarkable than that of deep-ocean-specific ASVs. For example, the number of epipelagic–BP connective ASVs was 194, 122, and 208 in Cruise 2016, 2017, and 2018, respectively. Comparatively, the BP-specific ASVs’ number was 151, 79, and 190 in the three cruises, respectively. Overall, in the three cruises, the number of epipelagic–MP and epipelagic–BP connective ASVs was 265 and 387, respectively, which was also higher than the number of MP/BP-specific ASVs (BP-specific and MP-specific ASVs were 194 and 284, respectively) (Figure 3). Thus, these lines of evidence suggested upper-layer bacterioplankton may impose an import effect to shape the bathypelagic bacterioplankton communities.

To further evaluate the influence of upper-layer bacterioplankton on the bathypelagic communities, we applied a microbial source tracking (MST) method to assess the statistically probable links between these bacterioplankton communities. The samples from the SUF, DCM, BOE, and MP layers were set as the ‘source’ and the BP samples as the ‘sink’ in the analysis. In most samples, the BP sequences were estimated to mainly originate from the MP samples (generally exceeding 60% of the bathypelagic sequences), indicating a solid communication between the BP and MP bacterioplankton communities (Figure 4). Comparatively, the direct link proportions between the bathypelagic and epipelagic samples were limited (generally less than 10% of the total bathypelagic sequences) (Figure 4).

3.3. Taxonomic and Functional Annotation of Epipelagic–Bathypelagic Connective ASVs

To understand the ecological characters within the epipelagic–BP connective bacterioplankton communities, we annotated these ASVs at the phyla level. As shown in Figure 5, Delta-proteobacteria and Alpha-proteobacteria dominated all the epipelagic–bathypelagic connective ASVs. Furthermore, some dominant epipelagic–bathypelagic connective ASVs were annotated to some specific phyla in different years. For example, Actinobacteriota, Firmicutes, and Cyanobacteria dominated the epipelagic–bathypelagic connective ASVs in 2016, 2017, and 2018, respectively. Notably, these ASVs contributed significantly to the interannual changes in bacterioplankton communities (Figure 5). Compared to the epipelagic–BP connective ASVs, the phyla constituted of BP-specific ASVs were more stable. Actinobacteriota, Alpha-proteobacteria, Chloroflexi, and Delta-proteobacteria were identified as the dominant phyla of BP-specific ASVs, which were also different from that of epipelagic–bathypelagic connective ASVs.

FAPROTAX analysis was further used to predict the function of the epipelagic–BP connective and BP-specific bacterioplankton. The functions of most epipelagic–BP connective sequences (with an average of 92.5%) were focused on carbon cycling (Figure 6). Among them, aerobic chemoheterotrophy was the most prevalent metabolic function and accounted for 39.8 to 87.9% of all the annotated epipelagic–BP connective sequences. Cyanobacteria was another identified carbon cycling function, ranging from 0.41% to 26.4% in different samples. Except for them, some carbon cycling functions provided significant advantages in some samples. For example, in the sample O12T17BP, more than 7.8% of the epipelagic–BP connective sequences belonged to the members of cellulolysis. Still, 9.2–12.0% and 6.5–7.9% of the sequences derived from samples O9T16BP and O7T17BP, respectively, were classified into the groups of fermentation and animal parasites/symbionts. In addition, a few epipelagic–BP connective sequences were annotated to the members of sulfur (e.g., dark oxidation of sulfur compounds) and nitrogen cycling functions (e.g., denitrification). The functions of BP-specific sequences were also primarily (in an average of 92.5%) ascribed to the members of carbon cycling (Figure 6). However, although aerobic chemoheterotrophy was still a critical annotated component of BP-specific sequences (in an average of 25.7%), their relative abundance was significantly lower than that of the intracellular parasites’ function (in an average of 40.9%). Moreover, in terms of the nitrogen cycling function, the BP-specific sequences of numerous samples (e.g., O8T16BP, O9T16BP, and O7T18BP) have been annotated to the nitrification, which was rarely detected in the epipelagic–BP connective results. These results suggested significantly different compositions and ecological functions between epipelagic–BP connective and BP-specific bacterioplankton.

4. Discussion

Previous studies have reported that the community structure of deep-sea bacterioplankton varied significantly each year and was linked with seasonal sinking particles [15]. In this study, our results indicated that there were also clear annual variations in bathypelagic bacterioplankton communities. Compared with the BP-specific ASVs, epipelagic–BP connective ASVs were likely to make more impact on the yearly changes of deep-sea bacterioplankton, especially some ASVs belonging to Actinobacteriota, Firmicutes, and Cyanobacteria. Notably, these phyla have consistently been detected with a positive relationship with sinking particles [24,25,26,27]. For example, in the Atlantic Subarctic Upper and Puerto Rico Trench water, the members of Actinobacteriota are observed to dominate the particle-associated bacterioplankton populations [24,25]. The taxa within Firmicutes are identified as the leading particle-attached indicators in the Northern Adriatic Sea [26]. In the Baltic Sea, Cyanobacteria ASVs are also found with high abundance in the particles rather than free-living in the water [27]. Thus, these lines of evidence suggest that sinking particles may be one of the main drivers for the annual variations in the deep-sea bacterioplankton. Nevertheless, some other environmental factors (e.g., temperature, pH, and DO), which largely depend on water stratification, were also performed in the annual variations, and more importantly, affected the microbial diversity in the South China Sea [16,28]. Thus, the quantitative description of the effects of the sinking particles on the annual variations in the deep-sea bacterioplankton still needs further investigation. Based on the FAPROTAX analysis, we uncovered that epipelagic–BP connective ASVs held significantly higher carbon cycling functions than BP-specific ASVs. It was consistent with previous observations that particle-associated bacterioplankton carry more abundant active genes than free-living bacteria [26,29,30]. For example, in the Northern Adriatic Sea, one study indicated that most of the functional genes related to substance metabolisms, such as amyA (starch degradation), dmdA (DMSP degradation), ureC (organic nitrogen degradation), and arsC (metal resistance), are more abundant in the particle-associated samples than in the free-living samples [26]. Furthermore, the particle-associated communities often demonstrate a broader range of carbon cycling enzyme (e.g., polysaccharide hydrolase) activities than the bulk water community [29,30]. Thus, along with other studies, our findings also supported that the epipelagic–BP connective ASVs have a positive relationship with the particles and suggested their essential roles in oceanic biogeochemical cycles.

Furthermore, previous studies indicated that sinking particles may be a relevant “seed bank” of viable taxa for the deep ocean [17,31]. In this study, although epipelagic–BP connective ASVs were found in high proportion in the bathypelagic communities, based on the MST analysis, the direct link proportions between the bathypelagic and epipelagic samples still need further investigation. Therefore, our results seem to support that the upper-layer bacterioplankton in particles can only form some “seeds”, rather than sinking into the deep sea to influence the structure of bathypelagic bacterioplankton. However, although our results exhibited some potential “seed” ASVs, their colonizing, living, and sinking characteristics still require more research.

5. Conclusions

This study applied high throughput sequencing analysis to reveal the significant annual variations in bacterioplankton communities and their function in the bathypelagic layers. Some epipelagic–bathypelagic connective ASVs belonging to members of Actinobacteriota, Firmicutes, and Cyanobacteria were found to contribute significantly to the changes in the annual deep-sea bacterioplankton. Interestingly, the ASVS of these phyla were always detected with a positive relationship with sinking particles, suggesting sinking particles may be one of the main drivers for the annual variation in deep-sea bacterioplankton communities. Further, the MST analysis suggested that the upper-layer bacterial communities could serve as a “seed bank” for influencing the bathypelagic bacterioplankton structure and carbon cycling functions. However, the character and biogeochemical role of these “seed” bacterioplankton ASVs still need further elucidation.