Microbial Community Diversity of Coral Reef Sediments on Liuqiu Island, Southwestern Taiwan

Abstract

Microbes in coral reef sediments are thought to play an important role in organic matter remineralization and nutrient recycling. Microbial communities also reflect the environmental conditions, such as nutrient status, of an ecosystem. This study investigates the relationship between microbial community diversity in the reef sediments and environmental conditions at Liuqiu Island. We sampled sediments seasonally from four sites around the island, Beauty Cave, Geban Bay, Houshi Fringing Reef, and Lobster Cave, from 2015–2020. The V5–V6 hypervariable region of 16S rRNA was amplified and sequenced using the Illumina MiSeq platform to identify the microbial communities. The results showed that the high abundance of Pseudomonadota, Planctomycetota, and Bacteroidota might reflect the eutrophic environments of the sediments on Liuqiu Island. Second, the identification of putative pathogens and human-related genera suggests that human activities have affected the marine environment of Liuqiu Island. Third, the insignificant spatial differences and the significant temporal differences in the microbial communities of Liuqiu Island indicate that annual or periodical events, such as the Kuroshio Branch Current and South China Sea Surface Current, could shape the microbial communities of Liuqiu Island. Furthermore, the abundance of human-related genera—Cutibacterium, Herbaspirillum, Corynebacterium 1, Escherichia-Shigella, and Kocuria—increased dramatically in the Lobster Cave site in September 2015 and may have been induced by a strong climate event, such as a typhoon or heavy rainfall. Our results revealed that the microbial communities of Liuqiu Island are dynamic and sensitive to adjacent environmental conditions. The sedimented microbial communities could monitor the bacteria and pathogens related to human activities and even reveal the putative events that could affect the ecological environments.

1. Introduction

Tropical coral reefs are highly diverse ecosystems in which complex environments provide diverse habitats and niches over time and space [1,2]. Microbial communities, including bacteria, microbial eukaryotes (fungi), and small eukaryotes, play an important role in nutrient cycling in coral reef ecosystems, encompassing many pelagic and benthic processes [3,4,5,6,7,8]. Vertical community stratification caused by redox gradients enables highly diverse microbial communities in coral reef sediments [9]. The diverse microbial metabolisms remineralize the dissolved and particulate organic matter through the heterotrophic metabolism pathway in the oxygenated upper sediment layer [10,11,12,13]. Inorganic nitrogen is also fixed by bacteria and archaea present in reef sediments and can contribute significantly to the nitrogen budget of the coral reef [14]. The accompanying nutrient cycling is particularly important for coral reefs to maintain high levels of primary production and biomass in extremely nutrient-poor seawater [15,16]. Further, a nutritional link between seasonal dynamics and sediment-associated bacterial communities was demonstrated in a study conducted at the Red Sea [17]. Sediment-associated microbes may provide the seeds for coral to restore their mucus microbes [18,19]. Finally, in the Great Barrier Reef, the bacterial communities in surface sediments were different between inshore and offshore and could be used as a biological indicator of environmental perturbations in coral reef ecosystems [20,21].

The microbial community structures of sediment samples from different environments exhibit various characteristics. Microbial community examination of sediments from different regions—including the Sea of Japan, the South China Sea, the Sea of Okhotsk, the Marginal Sea of Peru, the Red Sea, the Bohai Sea, and Okinawa—has shown that the structure and diversity of their microbial communities are divergent [17,22,23,24,25,26,27]. For example, in some polluted areas, the petroleum-eating bacteria—which belong to the phyla Pseudomonadota, Bacillota, Actinomycetota, and Bacteroidota—are abundant in coastal sediments [28,29,30,31,32]. Several studies on coastal marine sediments further show that the abundance and diversity of pathogenic micro-organisms, such as the genera Vibrio, Pseudomonas, and Bacillus, correlate with the degree of pollution [33,34]. In many areas contaminated and disturbed by humans, the genera Firmicutes and Bacilli dominate the microbial community in coastal sediments [30,35]. In addition to urban pollution, mariculture also alters the microbial communities in surrounding sediments by accumulating organic matter and increasing the abundance of fish-gut-related bacteria [36]. The degree of community shift and diversity decline in sediment bacteria was inversely related to the distance and time of the mariculture sites [37]. Many bacterial biomarkers on the genus level, such as Sulfurovum, Desulfobacter, and Fusobacterium, are documented as being associated with mariculture.

Liuqiu Island, also known as Little Liuqiu, is located to the southwest of Taiwan Island and faces the mouth of the Gaoping River about eight nautical miles away. It is the only coral reef island in Taiwan. The whole island covers an area of 6.8 km² with a 12 km coastline and has a tropical monsoon climate. The waters around Liuqiu Island remain above 25 °C all year round. The average water temperature is 25.0 °C in January and is 29.4 °C in July. The sea area around Liuqiu Island is affected by the Kuroshio tributary, the southwest blowing current and the plume of the Gaoping River [38,39]. The environments of the coral reefs around the island are different. The reef check survey [40] showed that the coral coverage on the east side of the island is about 40–69%, while that on the west side is lower, about 6–30%. In addition to coral, the fish flocks were also different between west and east sides of the island. Furthermore, the alga was abundant in the benthic communities, and the scene of algae covering the coral can be observed around the island. The indicators of invertebrates, such as cherry blossom shrimp, devil sea urchin, sea cucumber, giant clam, etc., can be recorded in low numbers. On the other hand, the convenient location for travelers and pleasant climate make Liuqiu Island a popular tourist destination in Taiwan, much like Taiwan’s well-known Turtle Island, where numerous sea turtles reside. In the last decade, the number of tourists has risen, with about 3 million tourists visiting Liuqiu Island in 2018. Large numbers of tourists are known to bring with them environmental stress, ecological disturbance, and sewage and garbage pollution. The environment around Liuqiu Island is experiencing these effects as a result of the increased tourism.

In this study, the microbial communities of the coral reef sediments from four selected sites around Liuqiu Island from 2015 to 2020 were investigated to elucidate the microbial response to human activities and the role of the microbial communities as indicators of pollution. The goals of this study were: (1) to identify the spatial and temporal distribution of the microbial community structure in sediments among different coral reefs; (2) to understand the role of sediment micro-organisms on Liuqiu Island as indicators of environmental pollution associated with human activities; (3) to monitor the microbial community for signs of events affecting the ecological environments. The results of this study could provide valuable microbial indicators for monitoring the environmental changes of Liuqiu Island.

2. Materials and Methods

2.1. Samples Collection and DNA Extraction

The coral reef sediments were collected seasonally from 4 fixed sites around Liuqiu Island (Figure 1), from 2015–2020. They were Beauty Cave (BC, northwest), Geban Bay (GB, west), Houshi Fringing Reef (HFR, southeast), and Lobster Cave (LC, northeast), and the GPS information was listed as the following: BC, 22°21′15.8″ N, 120°22′22.3″ E; GB, 22°20′34.8″ N 120°21′38.2″ E; HFR, 22°19′37.6″ N 120°22′10.6″ E; LC, 22°20′52.4″ N 120°23′27.9″ E. The benthic communities of the four sites were different based on the reef check [40]. In two west sites, BC and GB, the benthic communities were dominated by algal, followed by coral. On the other hand, on the east side, the soft coral dominated in LC, and the reef coral dominated in HFR. We collected the surface sediments by SCUBA diving at a depth of about 10 m, pushed a 15 mL sterile plastic tube into the sediments (~3 cm), preserved the sediments with 95% ETOH immediately [41], and stored them in a −20 °C freezer until the analysis began. We collected 87 sediment samples in total during 2015–2020. We used a DNeasy Power Soil DNA kit (Qiagen, Germantown, MD, USA) to extract the total genomic DNA from 500 mg of sediments from each sample following the manufacturer’s instructions.

2.2. PCR Amplification and Sequencing

The 16S rRNA V5-V6 hypervariable region was chosen as the marker of microbial community analysis. We used the primers U789F (5′-TAG ATA CCC SSG TAG TCC-3′) and U1071R (5′-GAR CTG RCG RCR RCC ATG CA-3′) [42] to construct amplicon libraries using the following PCR program: 94 °C for 3 min and then 35 cycles of 94 °C for 30 s, 50 °C for 30 s, and 72 °C for 30 s. Each PCR reaction had a total volume of 20 μL and contained 1 μL (10 ng/μL) of total genomic DNA, 0.5 μL (5 mM) of each forward and reverse primer, 10 μL of 2× HotStarTaq Plus Master Mix Kit (Qiagen, Germantown, MD, USA), and 8 μL of PCR-grade water. We pooled five separate PCR reactions per sample to eliminate the potential for early cycle PCR induced error. All PCR reactions for sequencing produced bands of the correct size (~280 bp), whereas no bands or smears were observed for the negative controls. We sequenced the pooled amplicons using an Illumina MiSeq platform (Illumina, San Diego, CA, USA).

2.3. Data Analysis

The software package of CLC Genomic Workbench and Microbial Genomics Module (QIAGEN, Aarhus, Denmark, http://www.clcbio.com, accessed on 11 October 2017) was used to analyze the sequence data. High-throughput raw reads were imported into CLC Genomics Workbench version 11 according to Illumina pipelines 1.8. We used a sequence trim setting of 0.05 for the limit for quality scores and 2 for the maximum number of ambiguities. The trimmed reads were merged with the default settings (mismatch cost = 1, minimum score = 40, gap cost = 4, and maximum unaligned end mismatch = 1). CLC Microbial Genomics Module version 3.0 was used to analyze operational taxonomic unit (OTU) clustering and microbial community structure. The sequence similarity of OTU clustering and chimeric sequence detecting was set at 97%. The SILVA 132 rRNA database provided reference OTU data for this study [43]. The diversity indices (Shannon entropy and Simpson’s index) and heat map of abundance were also calculated using CLC Microbial Genomics Module. The formulas of the two indices were listed as the following:

Shannon entropy

$$\mathit{H}\mathbf{=}{\displaystyle \mathbf{\sum }}_{\mathit{i}\mathbf{=}\mathbf{1}}^{\mathit{n}}\mathit{p}\mathit{i} \mathit{l}\mathit{o}{\mathit{g}}_{\mathbf{2} }\mathit{p}\mathit{i}$$

Simpson’s index

$$SI=\mathbf{1}-{\displaystyle \mathbf{\sum }}_{\mathit{i}\mathbf{=}\mathbf{1}}^{\mathit{n}}\mathit{p}{\mathit{i}}^{\mathbf{2}}$$

where n is the number of features, and pi is the fraction of reads that belong to feature i [44,45]. A permutational multivariate analysis of variance (PERMANOVA) test based on Bray–Curtis [46] distance was used to test the differences of microbial community structures among the sampling sites, years, or seasons. The test of differential abundance at the genus level among years or sampling locations was conducted using the CLC Microbial Genomics Module, and the p-value was checked using false discovery rate (FDR). Non-metric multidimensional scaling (NMDS) analysis and an analysis of similarities (ANOSIM) test based on Bray–Curtis [46] distance were deployed to examine the similarities among samples and to calculate the significant differences among the microbial communities of the year or location groups, respectively. The tests were conducted using Paleontological Statistics (PAST4.0) software [47].

3. Results and Discussion

3.1. Diverse and Dynamic Microbial Community Structure on Liuqiu Island

The 16S rRNA amplicon of 87 sediment samples was seasonally collected from 10 m deep at the four sites (BC, GB, HRF, and LC) from 2015 to 2020. A total of 11,438 OTUs were identified from these samples. Table S1 summarizes the predicted OTUs, reads in OTUs, filtered reads, and Shannon entropy and Simpson’s index values of each sample. The two diversity indices, the Shannon entropy and Simpson’s indices, exhibited high values of 9.182 ± 0.525 and 0.992 ± 0.006, respectively (Figure 2). The boxplot of Shannon and Simpson indices of different years are shown in Figure 2b,c. The ANOVA test showed that the two diversity indices among years were different, while that among the locations were not different (Table S2). The Shannon indices showed that species richness was relatively low in 2015, 2017, and 2020, while the Simpson indices exhibited low species evenness in 2015 and 2017.

The OTU annotation based on the SILVA 132 database classified 11,438 OTUs into different levels of taxonomic precision, including 65 phyla, 204 classes, 558 orders, 1096 families, and 2122 genera. At the phylum level, the top ten ubiquitous phyla were Pseudomonadota, Planctomycetota, Bacteroidota, Nitrososphaerota, Acidobacteriota, Actinomycetota, Cyanobacteriota, Verrucomicrobiota, Gemmatimonadetes, and Chloroflexota (Figure 2 and Table S3). Pseudomonadota (43.5%) and Planctomycetota (19.2%) had the highest relative abundances on Liuqiu Island and dominated the microbial community structures of Liuqiu Island’s sediments across the period we sampled. This phenomenon may have resulted from the eutrophic environment. Previous studies in the Bohai, Yellow, South, and East China Seas [27,48] have shown that the high abundance of Pseudomonadota and Plancytomycetes could be due to the environmental eutrophication. Pseudomonadota is critical for the degradation of the organic nitrogen in sediments [49,50], and Planctomycetota play important biogeochemical roles in anaerobic ammonium oxidation [38], methane oxidation [51], and carbon recycling [42]. Both phyla are chemotrophic bacteria, and their high abundance might reflect the high concentration of relative chemicals in the environment [52]. Furthermore, the third and sixth abundant phyla, Bacteroidota and Actinomycetota, were suggested as degraders of complex organic matters. Bacteroidota are considered to be specialized in the degradation of complex organic matter, especially in polysaccharides and protein [53,54]. The mineralization of high molecular weight organic matter proceeded by Bacteroidota was reported in ubiquitous aquatic environments, such as the coastal sea and deep sea [55,56,57]. The Actinomycetota could breakdown and recycle organic compounds, such as cellulose and chitin [58], and were reported to play a significant role in organic matter mineralization, mineral nutrient immobilization, and nitrogen fixation [59]. The abundance of the Bacteroidota and Actinomycetota might reflect the concentration of Liuqiu Island’s organic compounds. The above results suggest that the sediments in Liuqiu Island might be eutrophic environments for the micro-organisms.

The major microbial communities of each sample were dynamic, and the differential abundance in phyla level was observed over different years. For example, the relative abundance of Actinomycetota and Patescibacteria in 2015, Cyanobacteriota in 2017, and Bacteriodiota and Planctomycetota in 2020 increased significantly. In 2015, the increase of Actinomycetota and Patescibacteria was observed in sample 201509LC, in which the relative abundance of Actinomycetota and Patescibacteria was 26.8% and 16.9%, respectively (Figure 2 and Table S3). Actinomycetota is thought to be a degrader of organic compounds [58], while Patescibacteria consists of ultramicrobacteria with limited metabolic capacities and anaerobic lifestyles [60,61]. The dramatic increase of the two phyla suggests that the environment of Lobster Cave in September 2015 could have been more suitable for Actinomycetota and Patescibacteria. This increase indicates an ecological event that may have changed the habitat environment in 2015 and might have induced the decrease of the diversity of the microbial communities in Liuqiu Island. Furthermore, the increase of Cyanobacteriota in 2017 and Bacteroidota in 2020 might have correlated with the low diversity of the microbial communities in 2017 and 2020, respectively.

At the genus level, the 25 genera with the highest coefficients of variation were Marinomonas, Aureibacter, Endozoicomonas, Zeaxanthinibacter, Woeseia, Blastopirellula, Rhodopirellula, Cutibacterium, Corynebacterium 1, Kocuria, Herbaspirillum, Escherichia-Shigella, and some no-rank bacteria (Figure 3). Figure 3 shows that samples from the same year seemed to share similar abundant microbial genera and had a tendency to cluster. For example, Marinomonas, Aureibacter, and Endozoicomonas were three highly abundant genera in the 2019 samples. The genus Marinomonas, a gammaproteobacterium that inhabits various marine environments [62,63,64,65], is involved in the metabolism of various chemicals, including dimethylsulfoniopropionate, phthalate, and oil compounds [66,67,68]. The genus Endozoicomonas, belonging to gammaproteobacteria, is a symbiont of marine animals in coral reefs, including coral, sponges, tunicates, sea slugs, and some mollusks [69,70,71,72,73,74,75]. The presence of Endozoicomonas is correlated with the health of coral reef systems and contributes to the nitrogen and sulfur cycles of organic molecules [72,76]. The genus Aureibacter, isolated from a coral reef sea squirt [77] and belonging to Bacteroidota, is abundant in the anaerobic digestion process [78]. Next, Zeaxanthinibacter and Woeseia were two highly abundant genera in 2017. The Zeaxanthinibacter, inhabiting seawater and sediments, can produce carotenoid pigments [79,80] and is dominant in oiled sediments [81]. The Woeseia could be widely involved in the cycling of detrital proteins in both coastal and deep-sea sediments [82,83]. Three no-rank bacteria genera belonging to the phylum Cyanobacteriota (f_uncultured diatom, o_Chloroplast, and f_uncultured cyanobacterium) were only abundant in 2017 (Figure 3), and they are thought to play crucial roles in the nitrogen and carbon cycles in marine ecosystems [84]. Furthermore, Blastopirellula and Rhodopirellula were two highly abundant genera in 2020. The genus Rhodopirellula is a globally distributed bacterium that can be found attached to sediment particles and is associated with phytoplankton, sponges, and macroalgae [85,86,87,88,89,90]. The genus Blastopirellula, a member of Planctomycetota, is chemoheterotrophic and uses carbohydrates as its main carbon and energy sources [91].

In 2015, Cutibacterium, Corynebacterium 1, Kocuria, Herbaspirillum, and Escherichia-Shigella were dominant in the microbial community, especially in the sample 201509LC. However, these genera were relatively rare in other years. The genus Cutibacterium primarily resides on the human skin and is recognized as an important human pathogen [92,93,94,95,96]. Corynebacteria, found on the human skin and the upper respiratory system, is a human and animal pathogen [97,98]. The genus Kocuria, inhabiting human skin and various environments, is not only considered a pathogen but also antibiotic-resistant bacteria [99,100,101]. Herbaspirillum is a plant-associated N₂-fixer, as well as an opportunistic pathogen and potential degrader of pollutants [102]. The Escherichia-Shigella genus, which comprises Escherichia and Shigella species, could induce human shigellosis and several foodborne diseases [103,104]. Escherichia coli, a member of the genus Escherichia-Shigella, is used as an indicator of pollution in food, water, and the environment [105]. All five aforementioned genera in 201509LC are human-associated bacteria. In short, our data showed that the microbial community of Liuqiu Island may exhibit different signature taxa each year. Most human-associated bacteria were found to be dominant in 2015, especially in the sample 201509LC. The above results were carried out based on the genera level annotation of the 16S amplicon; therefore, the accuracy issue may not be ignored. Further investigation on the species level may help to increase the resolution of the human-associated bacteria evaluation.

3.2. Temporal and Spatial Differences in Microbial Communities

Permutational multivariate analysis of variance (PERMANOVA) was performed to test statistical differences in the structure of microbial communities based on year, season, and location at the phylum and genus level. The results showed that the differences at both the phylum and genus level were significant across years, except between 2015 and 2016, but insignificant among seasons and locations (Table S4). The ANOSIM and NMDS analysis also showed that the temporal difference was significant but that the spatial difference was insignificant in the microbial communities of Liuqiu Island (Figure 4). The differences within sampling years were lower than the differences between sampling years based on ANOSIM analysis (Figure 4a). The within-group differences for the year 2020 were the lowest, followed by 2015 and 2016. This result was consistent with the NMDS analysis (Figure 4b). The NMDS plot also showed that the sample dots for years 2020, 2016, and 2015 were the most condensed, except for the sample dot 201509LC. Meanwhile, ANOSIM analysis showed that the differences between locations were similar to those within locations (Figure 4d). The NMDS plot also showed that the sample dots of the location groups were mixed together (Figure 4e). The above results indicate that the temporal difference was significant, and the spatial difference was insignificant in the microbial communities of Liuqiu Island. Thus, the microbial community structures of Liuqiu Island were further compared across different sampling years (Figure 4c) and locations (Figure 4f) at the genus level to reveal the signature genera for each sampling year and location. The temporal differences among the samples were significant in all of the top 25 genera (Figure 4c). For example, the abundances of genera Cutibacterium, Herbaspirillum, Corynebacterium 1, Escherichia-Shigella, and Kocuria in 2015 were significantly higher than those of other years. These five genera were abundant in sample 201509LC, with the relative abundances of 16.0%, 11.7%, 8.9%, 1.8%, and 0.7%, respectively. Furthermore, the genera Marinomonas, Aureibacter, and Endozoicomonas were the most abundant in 2019. The abundance of the genus Marinomonas was 1.5% in the sample 201912GB, and the abundance of genus Aureibacter and Endozoicomonas was 16.9% and 3.3%, respectively, in the sample 201912LC. The genus Candidatus Nitrosopumilus was the most abundant in 2017, and the most abundant sample was 201712LC with a relative abundance of 5.1%. The f_Thiohalorhabdaceae, Blastopirellula, Rhodopirellula, and Woeseia, were the most abundant genera from 2015 to 2020. The abundances of o_Chloroplast and uncultured cyanobacterium increased dramatically in 2017 but decreased in 2018 and 2020. In 2017, the abundances of o_Chloroplast, uncultured-cyanobacterium, and f_Thiohalorhabdaceae increased dramatically, while that of f_Thermoanaerobaculaceae decreased. In 2018 and 2020, o_Chloroplast, uncultured cyanobacterium, and CandidatusNitrosopumilus exhibited extra-low abundance. Furthermore, f_Thermoanaerobaculaceae and f_Nitrosopumilaceae also exhibited low abundance in 2020. Dynamic microbial community structures were exhibited at both the phylum and genus level. A study on marine micro-organisms showed that the increase or decrease of micro-organisms may reflect the conditions of the environment [52]. Therefore, the dynamic microbial communities of Liuqiu Island may reflect the changes of the environmental conditions.

In addition to the temporal difference, the spatial differences in genus abundances were also calculated. Only 13 genera exhibited differences among the sampling locations (Figure 4f). The genera Aureibacter, Endozoicomonas, Cutibacterium, Herbaspirillum, Corynebacterium 1, Escherichia-Shigella, and Kocuria were statistically more abundant in the Lobster Cave site than in the other three sites, while the genus Zeaxanthinibacter was significantly more abundant in Beauty Cave. The relative abundances of the genera Aureibacter and Endozoicomonas were 16.9% and 3.3%, respectively, in 201912LC, while those of Cutibacterium, Herbaspirillum, Corynebacterium 1, Escherichia-Shigella, and Kocuria were 16.0%, 11.7%, 8.9%, 1.8%, and 0.7%, respectively, in 201509LC. The genus Zeaxanthinibacter was 4.5% and 3.3% in 201812BC and 201709BC, respectively. The human-associated bacteriagenera, Cutibacterium, Herbaspirillum, Corynebacterium 1, Escherichia-Shigella, and Kocuria, were abundant in LC, while the oil-spill-related genus Zeaxanthinibacter was abundant in BC. The BC and LC were the closest sites from the mouth of Gaoping River and may be affected by the plume of Gaoping River easier than the other sites. The above results may suggest the northern Liuqiu Island might be affected by human activity easily.

Although the reef check showed the benthic communities were different among the four sites [40] where we collected samples, the differences among their microbial communities were insignificant. This may indicate that the main forces shaping the microbial communities in the coral reef sediments of Liuqiu Island were not local environmental factors but large-scale or global events. On the other hand, the significant temporal differences in microbial communities of Liuqiu Island may suggest that the influential factors may change annually or periodically. The near-shore environment of Liuqiu Island is affected not only by local coastal currents but also by large-scale events, such as the Kuroshio Branch Current (KBC), the monsoon-induced South China Sea Surface Current (SCSSC), and the internal-wave-induced upwelling [106,107,108,109]. The SCSSC brings warm and low-salinity water into the sea area surrounding Liuqiu Island in the summer, while the KBC carries the warm and high-salinity water into the sea area surrounding the island in winter. Furthermore, the upwelling provides Liuqiu Island with an extra flux of nutrient-rich water [109,110]. The aforementioned events could shape the near-shore environments of Liuqiu Island and are affected by El Niño–Southern Oscillation [106,107].

3.3. Putative Pathogen and Human-Related Bacteria

Many bacteria genera were reported as indicators of pathogenicity and pollution [111,112,113,114,115,116,117]. For example, in a study of the South China Sea, the genera Vibrio, Pseudoalteromonas, Escherichia-Shigella, Psychrobacter, and Pseudomonas were used as environmental indicators to evaluate the impact of human activity [34]. Another study in Hong Kong further showed the positive relationship between four genera of Desulfobacteraceae and the degree of pollution level [33]. Therefore, based on the taxonomic profiles from the literature [111,112,113,114,115,116,117], 20 opportunistic pathogen and human-related genera were identified from the microbial communities of all samples (Figure 5 and Table S5). In general, Vibrio was the most abundant (0.01%~1.06%) genus in most of the samples. Two typical opportunistic pathogenic species—V. vulnificus and V. harveyi—were found in all the samples. V. harveyi was the dominant species and was recognized as a significant pathogen of marine vertebrates/invertebrates [117]. The genera Cutibacterium and Corynebacterium 1 appeared in most samples, and both of them were pathogens of human skin [93,95,96,118]. Three desulfurized genera, Desulfatitalea, Desulfobacter, and Desulfococcus, that correlated with the degree of pollution level [33] were also identified in most samples. The accumulated abundances of the remaining 20 putative pathogens and human-related bacteria were relatively low (<1.3%) across all samples, except in samples 201509LC and 201707HFR. The relative abundance of all putative pathogens and human-related bacteria in most samples was not high in Liuqiu Island. However, their appearances may indicate that human activities have affected the marine environment of Liuqiu Island, even in the coral reef sediments, 10 m deep. On the other hand, human activities have had a very serious impact on Liuqiu Island over the past decade. The number of tourists grew from 2.3 million in 2015 to 2.9 million in 2019 (MOTC 2022), bringing a lot of pollutants such as garbage and wastewater. The excessive tourists may have contributed to the emergence of these possible pathogens. However, our data show that the abundance of the putative anthropogenic bacteria in the coral reef sediments remained at a low ratio. The abundance of several landmark species, such as E. coli, is even lower than that of coral reef islands in the South China Sea [34,119]. Our findings imply that the sediments around Liuqiu Island might be less prone to accumulating environmental pollutants. Rather than factors related to human activities, other physical and chemical factors could be the main forces that shape the microbial communities of the coral reefs in the Liuqiu Island. Further studies are required to reveal the key forces—such as currents, topography, and tides—that affect the sedimented microbial communities of Liuqiu Island.

Although most samples exhibited a low abundance of putative pathogens, two remarkable samples, 201509LC and 201707HFR, exhibited an extremely high abundance of the putative pathogens. In 201509LC, the accumulated abundance of these 20 putative pathogens and human-related genera was extremely high, accounting for 27.6% of the entire microbial community. The most accumulated abundant genus was Cutibacterium (16.0%), followed by Corynebacterium 1 (8.94%), Escherichia-Shigella (1.79%), and Kocuria (0.75%). The first genus, Cutibacterium, is associated with human disease [93,95,96], and the dominant OTU in this study was the pathogenic Cutibacterium ance (92.6%). The second genus, Corynebacterium 1, was dominated by a potential multi-drug-resistant bacterium (94.6%), Corynebacterium tuberculostearicum [120], in our analysis. In the third genus, Escherichia-Shigella, the dominant species was identified as E. coli, a pollution indicator [105]. The fourth genus, Kocuria, is an opportunist pathogen of humans and aquaculture [94,99], and Kocuriaarsenatis sp., associated with arsenic pollution [121], was the most abundant OTU (98%). In the sample 201707HFR, the total reads of the putative human-related genera accounted for 2.75%, and Desulfococcus was the most abundant genus (2.60%). Desulfococcus is usually found in anaerobic or polluted environments [33,35,122,123]. The unusually abundant putative pathogen and human-related bacteria in samples 201509LC and 201707HFR may imply a sudden outburst of ecological events occurred in 2015 and 2017. The above results were carried out based on the genera level annotation of 16S amplicon; therefore, the pathogenic prediction of micro-organisms may be misleading due to the accuracy issue. Further investigation based on genomic level may increase the resolution of the putative pathogenic analysis.

3.4. Prospect of Long-Term Monitoring

Many bacteria genera were reported as indicators of pathogenicity and pollution. The results of this study show that the microbial community structure of Liuqiu Island from 2015 to 2020 was dynamic and imply that the benthic ecology in the Liuqiu coral reef ecosystem is relatively unsteady in southwest Taiwan. Of all our samples taken from the changing microbial community of Liuqiu Island, those that showed unordinary microbial community structures may best reflect the massive alteration of habitat environment. For example, the sample with the greatest difference was 201509LC, in which the abundance of phyla Actinomycetota and Patescibacteria increased significantly (Figure 2). Figure 3 shows seven genera that were significantly abundant in 201509LC. They are Cutibacterium (class Actinomycetes), Herbaspirillum (class Betaproteobacteria), Corynebacterium 1 (class Actinomycetes), Escherichia-Shigella (class Gammaproteobacteria), Kocuria (class Actinomycetes), and two no-rank genera (class Saccharibacteria). Saccharibacteria (class Patescibacteria) are human-associated bacteria isolated from the oral microbiome and can be found in various environments, such as activated sludges [124,125], water treatment plant sludge [126], and rainforest soil [127]. These micro-organisms associated with human activities were only abundant in 201509LC and were very rare in the other samples/sites. Meanwhile, the Shannon entropy and Simpson’s index from the LC site dropped to 6.902 and 0.962, respectively, in September 2015. However, the two values later rose to 8.89 and 0.992 in March 2016 and 9.93 and 0.997 in June 2016 (Table S1). These dynamic changes imply that an unknown disturbance occurred in LC between March and September 2015. The disturbance in LC may have been influenced by the adjacent environments and resulted in these unusual changes in the microbial community. The microbiota was back to its normal rarity in site LC about six to twelve months after March 2016. The recovery of the microbial community may have been due to the resilience of the coral reef system [21] in Liuqiu Island.

In addition to the KBC, SCSSC, and upwelling, the plume of the Gaoping River induced by a typhoon or heavy rainfall also affected the near-shore environment of Liuqiu Island [38,39,109]. The plume brought abundant nutrients, sediments, freshwater, and pollutants from the Gaoping River Basin into Liuqiu Island and shaped its near-shore environment suddenly. In August 2015, Typhoon SOUDELOR passed near Liuqiu Island and brought high rainfall, causing the plume of the Gaoping River to surround Liuqiu Island. A month later, sample 201509LC was harvested for this study. This strong hurricane might be one of the factors that altered the microbial community of 201509LC. However, the local events, such as the heavy rain which induced the discharge of wastewater and sludge from Liuqiu Island, should not be excluded. This special case in the long-term monitoring study reveals that the coral reef of Liuqiu Island is resilient, not only in the fast recovery of the microbial community structure but also in the diversity index. In short, long-term monitoring of a microbial community not only establishes ecological background information and shows the dynamic changes in microbial indicators, but it reveals whether putative events might influence marine environments.

4. Conclusions

In this study, we sampled sediments seasonally from four sites around the island, Beauty Cave, Geban Bay, Houshi Fringing Reef, and Lobster Cave, from 2015–2020. The V5–V6 hypervariable region of 16S rRNA was amplified and sequenced using the Illumina MiSeq platform to identify the microbial communities. In total, 11,438 OTUs were classified into 65 phyla based on the SILVA 132 database. Through the analysis of the microbial community structures, several conclusions could be drawn. First, the dominant phylum was Pseudomonadota, followed by Planctomycetota, Bacteroidota, Nitrososphaerota, and Acidobacteriota. The high abundance of Pseudomonadota, Planctomycetota, and Bacteroidota might reflect the eutrophic environments of the sediments in Liuqiu Island. Second, the identification of putative pathogens and human-related genera suggested the human activities affected the marine environment of Liuqiu Island. Third, the insignificant spatial differences suggested the local environmental factors may not be the main force shaping microbial communities of Liuqiu Island; however, the differential abundance analysis suggested that the northern Liuqiu Island might be affected by human activity easily. On the other hand, the significant temporal differences in microbial communities in Liuqiu Island indicate that annual or periodical events, such as the KBC and SCSSC, could shape the microbial communities of Liuqiu Island. Finally, the dramatic alteration of the microbial community of 201509LC may have been induced by a strong climate event, such as a typhoon or heavy rainfall. This study has showed that long-term monitoring of a microbial community not only establishes ecological background information and shows the dynamic changes in microbial indicators but reveals whether putative events might influence marine environments.