Diversity of microbial community and its metabolic potential for nitrogen and sulfur cycling in sediments of Phu Quoc island, Gulf of Thailand

Abstract

Although Phu Quoc island, Gulf of Thailand possesses diverse marine and coastal ecosystems, biodiversity and metabolic capability of microbial communities remain poorly investigated. The aim of our study was to evaluate the biodiversity and metabolic potential of sediment microbial communities in Phu Quoc island. The marine sediments were collected from three different areas and analyzed by using 16S rRNA gene-based amplicon approach. A total of 1,143,939 reads were clustered at a 97% sequence similarity into 8,331 unique operational taxonomic units, representing 52 phyla. Bacteria and archaea occupied averagely around 86% and 14%, respectively, of the total prokaryotic community. Proteobacteria, Planctomycetes, Chloroflexi, and Thaumarchaeota were the dominant phyla in all sediments, which were involved in nitrogen and sulfur metabolism. Sediments harboring of higher nitrogen sources were found to coincide with increased abundance of archaeal phylum Thaumarchaeota. Predictive functional analysis showed high abundance prokaryotic genes associated with nitrogen cycling including nifA-Z, amoABC, nirA, narBIJ, napA, nxrAB, nrfA-K, nirBD, nirS, nirK, norB-Z, nlnA, ald, and ureA-J, based on taxonomic groups detected by 16S rRNA sequencing. Although the key genes involved in sulfur cycling were found to be at low to undetectable levels, the other genes encoding for sulfur-related biological processes were present, suggesting that alternative pathways may be involved in sulfur cycling at our study site. In conclusion, our study for the first time shed light on diversity of microbial communities in Phu Quoc island.

Supplementary Information

The online version contains supplementary material available at 10.1007/s42770-021-00481-8.

Introduction

Marine ecosystems host a variety of aquatic organisms with a myriad of contributions to a significant proportion of the world’s population, which have been threatened by human-induced effects of climate changes [1, 2]. Habitat loss, pollution, overexploitation, eutrophication, and invasions by non-native species consequently result in an imbalance currently affecting the marine diversity [3, 4]. There is an urgent need to reduce these pressures and to monitor fundamental parameters of biodiversity, including structure and function, in order to predict global warming scenarios [4].

Microorganisms present in marine sediments are considered a vital part of aquatic ecosystems, catalyzing many global biogeochemical cycles involved in carbon, nitrogen, sulfur, and oxygen dynamics [5, 6]. Although marine sediments are formed by a mixture of unconsolidated organic and inorganic particles deposited on the seafloor, about 90% of the total benthic biomass are generated in the presence of diverse microbial communities [7–9]. Microorganisms possess capacity to recycle nutrient, metabolize foreign compounds, colonize new ecological niches, and take part in biogeochemical cycling, which are greatly contributing to the health of marine ecosystems [10, 11]. Recently, the increasing interest in marine sediments generates unstoppable efforts to decipher microbial communities at different aspects. Microbial diversity in Pacific Ocean and China seas was well-studied by using 16S rRNA sequencing, showing bacterial abundance in surface sediments (0-5 cm), which clearly emphasized their importance in the biogeochemical cycles [10, 12, 13]. Bacteria predominantly occupy the surface sediments (0-5 cm) [12]. In addition, previous studies on surface marine sediments from East China Sea, Aarhus Bay (Denmark), and Svalbard (Artic Ocean) revealed Proteobacteria to be the most dominant phylum that distributed to over 50% of the microbial biomass [14–17]. Grammaproteobacteria was the significant class detected in the sediments as compared to other classes belonging to the phylum Proteobacteria [16, 18, 19]. The recent study showed a sharp shift in relative abundance of Gammaproteobacteria between the sediment surface and the bottom of the zone of more intense bioturbation, highlighting that Gammaproteobacteria characterized mainly surface sediments [17]. On the other hand, Taha et al. found that the phyla Actinobacteria, Chloroflexi, Firmicutes, and Proteobacteria inhabited equally in marine sediment cores collected at Okinawa Island, Japan [13]. Temperature-driven changes significantly decreased the relative abundance of all phyla such as Proteobacteria and Planctomycetes, while Bacteroidetes were resistant and became dominant, acting as the initial organic matter degraders in benthic carbon cycling [5]. Thus, the discovery of prokaryotic communities that are sensitive to environmental changes opens up new avenues for better understanding of how climate changes and human activities interfere with the original marine ecosystems.

Phu Quoc island is located at the Lower Gulf of Thailand, which has a rich number of natural resources and valuable coastal ecosystems [20]. However, global warming and human activities have resulted in beach erosion, aquatic resources depletion, and habitat degradation [21]. Until now, a few studies in coastal and reef waters have been conducted in the Gulf of Thailand but systematic study in microbial diversity associated with sediments is lacking [22, 23]. We carried out the first detailed 16S rRNA gene amplicon sequencing of microbial communities present in sediments collected along Phu Quoc island. This study provided new insights into the structure of microbial communities and their metabolic processes, laying a theoretical basis for environmental protection.

Materials and methods

Sample collection

Surface sediment samples were collected from 10 sites along Phu Quoc island, including 4 sites (DD1, DD2, DD3, and DD4) in Duong Dong (DD) area, 3 sites (DT1, DT2, and DT3) in Duong To (DT) area, and 3 sites (AT1, AT2, and AT3) in An Thoi (AT) area (Fig. S1). Samples from DD, DT, and AT were collected on 25-29 June 2019, 03-06 August 2019, and 18-22 September 2019 (the wet season), respectively (Table 1). Depth was determined from measured rope length. All the sediment samples (17-31 meters in depth from water surface and 0-15 cm below surface sediment) were collected using a sterile stainless steel grab sampler (Wildco, Florida, USA) and then left to settle for 48 h before the supernatant was removed. Following collection, each sample was stored at −20°C prior to DNA isolation.

Table 1.

Description of physicochemical characteristics of the sediment samples

Sample	Sampling date (dd.mm.yy)	Distance from shore (km)	Depth (m)	Salinity (‰)	Temperature (°C)	Organic matter (%)	Phosphorus (μg/g)	Nitrogen (μg/g)
DD1	25.06.2019	1.77	17	36.8	25.8	0.2	101.2	179.0
DD2	25.06.2019	5.47	23	38.6	25.5	0.2	117.7	174.9
DD3	27.06.2019	10.0	27	37.8	25.9	0.5	154.6	182.4
DD4	29.06.2019	16.1	31	38.3	25.7	0.2	180.1	250.6
DT1	03.08.2019	2.09	19	37.5	25.5	0.3	97.8	139.6
DT2	05.08.2019	5.6	26	38.5	25.6	0.1	124.8	124.1
DT3	06.08.2019	6.1	23	38.0	25.5	0.7	214.7	280.7
AT1	18.09.2019	4.7	24	37.4	26.3	0.6	272.6	341.8
AT2	19.09.2019	6.9	24	37	26.1	0.7	262.5	278.8
AT3	22.09.2019	4.0	25	36.5	25.9	0.4	209.3	250.5

Chemical analysis

Six environmental parameters, including depth, salinity, temperature, organic matter, phosphorus, and nitrogen, were investigated in this study. At each sampling site, the temperature and depth were measured by a conductivity-temperature-depth (CTD) oceanography instrument. Salinity was measured using a salinity meter (HANNA, Italy). Chemical analysis was performed at the Institute of Chemistry, Vietnam Academy of Science and Technology. Organic matter content in the sediments was obtained using the potassium dichromate method [24]. Total phosphorus was analyzed based on the perchloric acid digestion method [25]. Total nitrogen was determined by the potassium persulfate oxidation method [26].

DNA extraction and sequencing

Total DNA was isolated from each sample in three independent repetitions. For each isolation, 1 g of the sediment sample was extracted using DNeasy PowerMax Soil Kit (QIAGEN, Hilden, Germany) following the manufacturer’s instruction. The DNA was eluted by TE buffer. The quality and quantity of the DNA were determined by gel electrophoresis and NanoDrop 2000c spectrophotometer (Thermo Scientific, USA), respectively. The extracted DNA was stored at -20°C prior to analysis. The V4 hypervariable region of the 16S rRNA gene was amplified using primers 515F (5′-GTGCCAGCMGCCGCGGTAA-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) [10, 12], which subsequently was sequenced using an Illumina HiSeq X Five platform (San Diego, CA, USA).

16S rRNA gene-based amplicon analysis and accession numbers

The sequencing data was base called and demultiplexed using the BCL2Fastq software. The sequences were trimmed and assigned to individual samples based on their barcodes, and primer sequences were removed from the reads. Reads were then aligned with 16S rRNA (SILVA Release 132 [27]) database and inspected for chimeric errors using VSEARCH v2.6.2 [28]. After these quality assessment steps, reads of the marker genes (16S rRNA) were clustered de novo into different operational taxonomic units (OTUs) at 97% sequence similarity using the UPARSE v11.0.667 algorithm [29]. In addition, rare OTUs with less than 2 reads (doubleton) which are often spurious were deleted from downstream processing. A single representative sequence from each OTU was randomly chosen, and Pynast was used to align and construct a phylogenetic tree against the SILVA 132 16S rRNA database [30]. Taxonomic assignment of OTU was achieved using QIIME V1.9.1 against the SILVA 132 16S rRNA database (Release 132) [31]. For each sample, alpha diversity indices, including the richness (Chao) and diversity indices (Fisher Alpha, Shannon and Simpson), were also calculated in QIIME [31]. To compare variations in the richness and diversity across sediment groups (DD, DT, and AT), a one-way analyses of variance (ANOVA) followed by a pairwise t-test were conducted. Bray-Curtis distance was used for non-metric multidimensional scaling (nMDS). Diversity between sampling location (beta diversity) was determined based on Bray-Curtis distance and visualized by the nMDS. Furthermore, a Venn diagram was generated using Venn Diagram R package [32] to reveal the shared and unique OTUs among groups following to the occurrence of OTUs in a sample group subjected to their own relative abundance. A linear discriminant analysis effect size (LEfSe) was applied to the OTU table (non-parametric factorial Kruskal-Wallis (KW) sum-rank test P < 0.05, LDA > 2.0, linear discriminant analysis (LDA)) to identify the discriminant prokaryotic clade [33].

Raw sequence reads of the 16S rRNA gene amplicon for 10 samples were deposited in the NCBI Sequence Read Archive (SRA) database (accession number PRJNA615005).

Metabolic prediction of microbial communities

Prediction on functional composition of the microbial community from each sample was determined using the PICRUSt2 implementation tool (https://github.com/gavinmdouglas/q2-picrust2). The Kyoto Encyclopedia of Genes and Genomes (KEGG) and KEGG Orthology (KO) annotation were used for metabolic prediction of sediment samples. Predicted gene abundance was calculated as reads per kilobase of sequence per gigabase of mapped reads (RPKG).

Results

Environmental factor analysis of the sediments

The environmental parameters (temperature and salinity) and nutrients (organic matter, phosphorus, and nitrogen) concentrations of each sediment sample were measured and summarized in Table 1. The temperature and salinity fluctuated slightly from sample to sample. The chemical characteristics of each collected sample varied greatly depending on area and distance from shore. Among the three regions, sediments collected from DD region had the lowest average phosphorus and nitrogen content, of which the highest concentration of these elements were found in DD4 sample, being 180.1 μg/g and 250.6 μg/g, respectively (Table 1). The sediments of AT region were rich in organic matter, nitrogen, and sulfur with the highest values of 0.7%, 341.8 μg/g, and 272.6 μg/g, respectively. Noteworthy to mention, samples collected far away from shore such as DD4 and DT3 had higher phosphorus and nitrogen concentrations than those from near-shore sediments.

Richness and diversity of marine sediments

Amplicon sequencing using primers to target the V4 region of 16S rRNA on the Illumina sequencing platform resulted in a total of 1,143,939 reads. After data trimming and quality filtering, an average quality reads of 107,202; 120,659; and 112,099 were obtained from three groups of sediment samples DD, DT, and AT, respectively. The estimated OTU richness of AT samples was 5493, while the other two sample groups, DD and DT, possessed the average estimated OTU richness of 4686 and 4664, respectively. Chao index showed that AT group possessed slightly higher microbial community richness than either DD or DT (P > 0.05) (Fig. 1a). The rarefaction curves indicated similar profiles of all samples (Fig. S2). Neither the Fisher Alpha nor Shannon and Simpson indices revealed any significant difference among sampling areas, indicating the nearly similar species diversity in all regions (Fig. 1a). Furthermore, beta diversity showed a clear grouping of the total microbial community in AT regions and in three out of four DD sampling sites. Moreover, microbial compositions of AT regions were considerably separated from the DD region with an exception being sample designated DD4 (Fig. 1b).

Fig. 1.

Comparison of species richness and diversity of across sediments sampled at different areas. (a) The alpha-diversity measure was based on estimated OTU richness and diversity indices. Box plots for alpha diversity of the bacterial communities in different group of samples. (b) Beta diversity was represented by Bray–Curtis distance. Non-metric multidimensional scaling (nMDS) plot for all samples at the OTU phylogenetic level (stress value = 0.1)

Prokaryotic community composition

In total, 8331 unique OTUs were observed, representing 52 phyla. The prokaryotic community of marine sediments was dominated by bacteria accounting for 86% of the total community, while archaea corresponded to 14% (Table S2). At the phylum level, Acidobacteria, Bacteroidetes, Chloroflexi, Firmicutes, Nanoarchaeaeota, Gemmatimonadetes, Nitrospirae, Planctomycetes, Proteobacteria, and Thaumarchaeota were found to be the top ten ubiquitous phyla for all sediment samples (Fig. 2a). Among them, Proteobacteria was the most abundant bacterial phylum in all sediment samples, ranging between 37.7 and 49.5% of the overall microbial communities, followed by Planctomycetes (8.8-16.1%) and Chloroflexi (5.0-18.2%). Thaumarchaeota which was the fourth most abundant phyla (5.6–16.9% of the overall microbial communities) dominated the archaeal community. It constituted an average of 61.4% of all archaea in our study sites (Fig. 2a). Interestingly, archaea along with the sampling areas were found to be different (Table S2). Archaea community in AT areas were present in significantly higher proportions than those in DD areas (P = 0.0079), but not in DT areas (P > 0.05). It is consistent with the higher nitrogen concentration observed in AT areas as compared to DD areas (Table 1).

Fig. 2.

Taxonomic composition of dominant bacterial taxa at phylum (a) and class (b) levels in across ten sediments collected at DD, DT, and AT regions. Top ten taxa for each sediment were shown here

Bacterial taxa at class level observed in three different areas also showed similar compositions. The top ten taxa at the class level showed that the major classes were Gammaproteobacteria (19.1%-33.5%) and Deltaproteobacteria (19.9-22.6%) (Fig. 2b). Surprisingly, Gammaproteobacteria were found to be less abundant in farther offshore samples, i.e., DD3, DD4, and DT3, accounting for 20.7%, 21.4%, and 19.1%, respectively. Other major bacterial taxonomic groups in sediments included Nitrososphaeria (6.7-20.2%), Anaerolineae (5.4-20.6%), Planctomycetacia (5.8-11.3%), Bacteroidia (4.2-12.1%), Alphaproteobacteria (5.2-7.0%), Clostridia (1.9-7.0%), Thermoanaerobaculia (1.9-5.5%), and Phycisphaerae (2.2-4.9%) (Fig. 2b).

Similarities and differences in prokaryotic diversity

Core microbiota shared within the sampling sites and among the entire three sample groups were determined using the Venn diagram. The shared OTUs of three areas were 7,090, which took up 85% of the total OTUs (Fig. 3a). The unique OTUs of DD, DT, and AT were 138, 98, and 50, respectively, indicating that the vast majority of OTUs were common to the three sampling sites and very few were specific to each sampling site. AT, DT, and DD sample groups showed the most homogenous microbial community structure in which 60.2%, 42.1%, and 37.7% were present within the group, respectively (Fig. 3a). Apart from that, DD4 and DT3 showed heterogenous nature of the microbial community with the highest non-shared OTUs with other samples in the same group being 12.6% and 18.2% respectively (Fig. 3a).

Fig. 3.

The common and difference in bacterial communities of different sea areas. (a) Venn diagram denoting the number of unique and shared OTUs in the different libraries. (b) Linear discriminate (LDA) on effect size (LEfSe) identifying specific bacteria in the sampling areas

The linear discriminant analysis effect size (LEfSe) was used to compare bacterial communities and identify discriminative bacteria taxa among groups. The linear discriminant analysis (LDA) value distribution histogram exhibited 30 bacterial taxa with significant differences. Six groups, including Chloroflexi (phyla), Deltaproteobacteria (class), Gammaproteobacteria (class), Zeaxanthinibacter (genera), Gimesiaceae (family), and uncultured bacterium (species), were significantly found in the DD group (Fig. 3b). Eight groups of bacteria were significantly enriched in AT, including Acidobacteria (phyla), Flavobacteriales (order), Crocinitomicaceae (family), Halomonadaceae (family), Halomonas (family), Romboutsia (genera), Marinomonadaceae (family), and Marinomonas (genus). Only one taxa of bacteria, Moritellaceae (family), in the DT group was higher compared to both AT and DD groups (Fig. 3b).

Functional prediction of sediment prokaryotic communities

In order to gain deeper insight into the function of marine sediment microbial communities at Phu Quoc island, PICRUSt2 was used to infer microbial gene content from 16S rRNA gene data and relative abundance of functional genes. The most abundant sequences of each sediment were functionally assigned closer to metagenomes which have genes encoding N-cycling enzymes, including NifA-Z (nitrogen fixation), AmoABC (nitrification), NirA (assimilatory nitrite reduction), NarBIJ and NapA (nitrate reduction, dissimilatory), NxrAB (nitrification), NrfA-K and NirBD (dissimilatory nitrite reduction to ammonium), NirS, NirK, NorB-Z (denitrification), and especially GlnA, Ald, UreA-J (ammonia assimilation and ammonia production) (Fig. 4). However, Hao (nitrification) and HszA (anammox) were shown to have little or no presence. Among them, predicted genes involved in ammonia assimilation and ammonia production were the most predicted abundant sequences in the nitrogen metabolism. For example, sequences encoding for GlnA were about 3-fold and 6-fold higher on average than sequences from enzymes contributing to nitrogen fixation (NifA) and denitrification (NirK), respectively (Table S3).

Fig. 4.

Genes involved in nitrogen cycle as depicted by the width of lines. The bold solid lines indicate strong support for a pathway, while dashed line depicts a low abundance for a pathway. Abbreviations: RPKG, reads per kilobase of sequence per gigabase of mapped reads; nif, nitrogenase; gln, glutamine synthetase; ald, alanine dehydrogenase; ure, urease; amo, ammonia monooxygenase; hao, hydroxylamine oxidoreductase; nxr, nitrite oxidoreductase; nap, periplasmic nitrate reductase; nar, nitrate reductase; nir, nitrite reductase; nor, nitric oxide reductase; nos, nitrous oxide reductase; nrf, nitrite reductase

Genes associated with sulfur cycle also were identified in all sediment samples in which predicted gene abundances toward sediments were nearly identical. Genes (cysDHJNC, sir, fsr) responsive to assimilatory sulfate reduction were the most abundant sequences among microbial sulfur cycle (Fig. 5). For sulfur oxidation, genes related to thiosulfate oxidation (soxABCYZ), sulfide oxidation (fccAB), and sulfide-quinone oxidation (sqr) also were observed as the second-most abundant subprocess. Key enzymes that catalyze dissimilatory sulfate reduction as well as inorganic and organic sulfur transformations included anaerobic sulfite reductase (asrA), dissimilatory (bi)sulfite reductase (dsrAB), thiosulfate dehydrogenase (tsdA), tetrathionate reductase (ttrABC), thiosulphate-quinone oxidoreductase (doxD), sulfhydrogenase (hydA), cysteate sulfolyase, taurine dioxygenase (cuyA), and sulfoacetaldehyde acetyltransferase (xsc) (Fig. 5).

Fig. 5.

Sulfur cycling potential in Phu Quoc island. Genes labeled in black are found in 16S rRNA gene data, while colored red stands for having little or undetectable abundance. Abbreviations: asr, anaerobic sulfite reductase; dsr, dissimilatory (bi)sulfite reductase; cysH, phosphoadenosine phosphosulfate reductase; cysJ, sulfite reductase (NADPH) flavoprotein; cysNC, ATP sulfurylase/Adenylyl-sulfate kinase bifunctional enzyme; cysD, sulfate adenylyltransferase subunit; sir, sulfite reductase; fsr, F420-dependent sulfite reductase; sox, sulfur-oxidizing multienzyme complex; fcc, sulfide dehydrogenases; phs, thiosulfate reductase; tsd, thiosulfate dehydrogenase; ttr, tetrathionate reductase; dox, thiosulphate-quinone oxidoreductase; hyd, sulfhydrogenase/hydrogenase; pdo, persulfide dioxygenase; tst, thiosulfate oxidation; psr, polysulfide reductase

Discussion

Vietnam’s largest island, Phu Quoc is located in the Lower Gulf of Thailand, which has been affected drastically by not only recent climate change but also local anthropogenic activities. To evaluate and find the solutions to these threats, investigating the diversity of microorganisms plays an important role in understanding the ecological function of microorganisms in the local environment. Until now, knowledge of the microbial community structure and the biogeochemical cycles in the Lower Gulf of Thailand is still unexplored. This study analyzed for the first time the diversity of microbes in sediments from Phu Quoc island by using the culture-independent approach.

In the total community, the majority of OTUs were assigned to bacteria, which agreed with the results of previous studies [5, 12]. At the phylum level, sediment samples showed similar microbial community composition. Using alpha diversity tests, no significant differences were observed in the richness and diversity within three different areas, which were in agreement with environmental parameters (Fig. 1a and Table 1). Taking a look more closely in the microbial composition, Proteobacteria, the main contributor to the composition of marine environments, are the predominant bacterial group in the three areas (Fig. 2a). The major classes of the detected Proteobacteria OTUs are Gammaproteobacteria and Deltaproteobacteria (Fig. 2b), which formed large clusters in all of the marine sediments explored previously [5, 10, 12, 34]. Among those, the most abundant OTUs are those belonging to the families Pseudoalteromonaceae, Vibrionaceae, and Woeseiaceae of Gammaproteobacteria and Desulfobacteraceae of Deltaproteobacteria. Interestingly, Pseudoalteromonas belonging to the family Pseudoalteromonadaceae represented the second most detected OTUs (Table S5). To be known for their association with eukaryotic hosts and the ability to produce a wide range of secondary metabolites [35], Pseudoalteromonas may be a candidate for the production of novel active compounds. Woeseiaceae, having only one isolated strain, is a ubiquitous marine dweller, with up to 22% of bacteria amplicon in coastal sediments [36, 37]. This bacterial family exhibited a wide range of metabolic potentials involved in nitrogen and sulfur cycling. While Woeseiaceae may be involved in sulfide oxidation, the family Desulfobacteraceae of Deltaproteobacteria is likely responsible for the dissimilatory reduction of oxidized species of sulfur [37]. This finding is consistent with other studied sediments from the Bohai Sea, Yellow Sea, South China Sea [10], suggesting that Proteobacteria are the most important group of the offshore microbiota.

Besides, Planctomycetes as the second most abundant phyla are thought to be the key degrader of organic matter penetrating into marine sediment from the water column and nitrogen source. Moreover, they are presumably responsible for the degradation of recalcitrant organic matter due to their abilities to produce sulfatases [38]. Chloroflexi was the third most abundant phylum in all samples, represented mainly by the class Anaerolineae. Although Chloroflexi is one of the most dominant bacterial phyla in a wide variety of habitats [39], only a few can be maintained in axenic cultures [40]. Uncultured Chloroflexi might contribute to the degradation of the organic matter derived from anammox bacterial cells to low molecular weight carbon compounds such as acetate and ethanol [41, 42]. Sulfate-reducing bacteria can utilize these low molecular weight compounds as electron donor [43]. Environmental factor analysis and Venn diagram revealed that site DD4 showed significantly higher phosphorus and nitrogen levels and non-shared OTUs than the others collected close to coastline (Figs. 2 and 3). In addition, the strictly anaerobic order Anaerolineales belonging to Anaerolineae was strongly enriched in the DD4 sample (Table S2). As proved previously, the differences in bacterial diversity and community composition could be observed in sub-seafloor sedimentary environments that are separated by a few tens of kilometers [34]. These evidences supported that biogeochemical transformations may shape the microbial community composition of the DD4 sediment.

Based on the predicted functional analysis of the sediments, nitrogen cycling was shown to be the most dominant sink for metabolic potentials. A large number of genes involved in nitrogen cycling, including fixation, nitrification, denitrification, and especially ammonia assimilation and ammonia production, were predicted at higher levels in comparison to sulfur cycling (Figs. 4 and 5). The most dominant genes were glnA, ureA-J, and ald, indicating sediment microbial communities had the largest genomic potential for organic nitrogen via the glutamine synthase-glutamine oxoglutarate aminotransferase pathway for ammonia assimilation and ammonia production (Fig. 4). Ammonia assimilation plays an important role in maintaining the nitrogen cycle and the nutrient status of sediment in the marine environments, which are performed by the microbial community [44, 45]. Since urea is widely produced by microorganisms and human pollution, it is not surprising that urea degradation gene clusters were detected with high abundance [46]. Nitrification is a vital type of pathway involved in nitrogen metabolism, which converts the ammonia to hydroxylamine, nitrite, and nitrate [12, 46]. This process is encoded by genes including amoA, amoB, amoC [12, 19, 47], supported by a high abundance of Thaumarchaeota observed in this study. Autotrophic processes such as ammonia oxidation were supported by the high abundance of amoABC genes that assigned predominantly to the Nitrosopumilales lineage of Thaumarchaeota and slightly to Nitrosomonadaceae of Betaproteobacteria (Fig. 4). Of note, urea is required as the main energy and nitrogen source for the growth of Thaumarchaeota through intracellular conversion to ammonium [46, 47]. The main genes responsible for anammox such as nirK and nirS were found in all sediment samples, indicating that nitrogen might be recycled and stored in the environment at reduced conditions instead of being removed, leading to eutrophication effect via anaerobic dissimilatory nitrate reduction to ammonium. Another hypothesis is that the concentration and availability of urea were characterized by constantly high-supplied sources. Taken together, archaeal and bacterial ammonia oxidizers are of significant importance for nitrogen-cycle transformations at Phu Quoc island.

On the other hand, a metabolic potential for sulfur cycling was not fully represented in 16S rRNA gene sequence data. Overall, the most abundant genes belonged to the assimilatory sulfate reduction pathway, including cysDHJNC, sir, and fsr genes (Table S4). To utilize sulfate from the environment, CysDHJN or Sat enzymes are required for activation of sulfate (sulfur oxidation state +6) as the initial step of sulfate reduction [48]. The cysDN is only found in the anaerobic methanotrophic archaea such as Methanoregula, Methanococcoides, and Methanolobus [26, 49]. cysDN was present in high relative abundance eventhough only Methanococcoides was found in the archaeal community, which remains to be investigated. As described previously, sir and fsr genes mostly are found in the member of Alpha, Beta, Grammaproteobacteria, Actinobacteria, Acidobacteriia, and Acidimicrobiia [50, 51]. Alpha and Betaproteobacteria also harbor soxABCYZ genes responsible for thiosulfate oxidation that were abundant in our study. In contrast, sulfite-quinone oxidoreductase (soeAB), persulfide dioxygenase (pdo), thiosulfate oxidation (tst), and sulfur oxygenase reductase (sor) genes involved mainly in sulfide oxidation were not identified. Moreover, the sulfate reduction pathway lacked several important genes such as quinone interacting membrane-bound oxidoreductase complex (qmoABC), polysulfide reductase (psrB), adenylylsulfate reductase (aprAB), and sulfate adenylyltransferase (sat) (Fig. 5 and Table S4). Thus, we speculate that the other alternative enzymes found in less abundant or unidentified groups may compensate for the loss of these genes.

Conclusion

This study provided the first fundamental insight into microbial taxa inhabiting the sediments of Phu Quoc island at Gulf of Thailand that were dominated by heterotrophic Proteobacteria, Planctomycetes, Chloroflexi, and autotrophic Thaumarchaeota. Archaeal communities of AT areas have significant differences in comparison with DD, but not with DT, hypothesizing an influence of nutrient factors such as nitrogen. Functional prediction analysis suggested that the identified microbial communities contributed significantly to nitrogen cycling in Phu Quoc island. A large number of genes involved in nitrification, denitrification, fixation, and ammonia assimilation were predicted at high abundance in comparison to the sulfur pathway, showing their potentials to utilize nitrogen sources. This cataloging of microbial community structure acts as the reference microbial community of the healthy marine ecosystem.

Author contribution

NTQ, HTD, DMT, THNV, QVN, QHN, and PBC: experimental procedures, data preparation and interpretation, and writing the manuscript. HTD, DJHS, HK, WJL, THL, and DVT: experimental procedures and reviewing the manuscript. NTQ and QTP: writing the manuscript. DMT, KTN, and HHC: manuscript preparation and funding acquisition. All authors read and approved the final manuscript.

Funding

The study was financially supported by the Vietnam Academy of Science and Technology (VAST) and Ministry of Natural Resources and Environment (MONRE) under Grant number VAST.ĐA47.12/16-19 within the framework of Decision No. 47/2006/QD-TTg on “General plan for survey and management of marine resources and environment until 2010, with a vision to 2020.”

Declarations

Ethics approval

Not applicable.

Consent to participate

Not applicable.

Consent for publication

All authors consent for publication.

Conflict of interest

The authors declare no conflicts of interest.