Community profiling of the intestinal microbial community of juvenile Hammerhead Sharks (Sphyrna lewini) from the Rewa Delta, Fiji

Fourteen juvenile scalloped hammerhead sharks (Sphyrna lewini; SHS) were captured between November and December 2014 in the Rewa Delta in Fiji, and assessed for intestinal microflora characterisation using 16S rRNA amplicon sequencing by Illumina Miseq. The microbial population revealed a fluctuating dominance between the Enterobacteriaceae and Vibrionaceae families, namely Citrobacter and Photobacterium spp. Other related marine operational taxonomic units were closely related to Afipia felis, Chloroflexus aggregans, Psychrobacter oceani, Pontibacter actiniarum and Shigella sonnei. Two sharks had distinctive profiles that were dominated by known pathogens, namely Aeromonas salmonicida and Klebsiella pneumonia. The presence of a Methanosaeta species, and of Shigella and Psychrobacter, would suggest sewage contamination because of a spill that occurred on the 6th of December 2014. This study successfully establishes a baseline for future research.

Detailed knowledge of SHS ecology, behaviour and habitat requirements, particularly during the first stage of their life, is still limited 10 . In addition, there are indications that contamination by human and animal waste from villages lining the riverbanks may have shifted the balance of the ecosystem of the Suva Lagoon by increasing nutrient and bacterial load in water and organisms [11][12][13] . The main contributors of nutrients in Laucala Bay associated with sewage are the effluent of the Kinoya sewage treatment plant released at the sea outfall located in Laucala Bay (Fig. 1), and the human and animal waste from villages on the banks of the Rewa River 12 . Fish and most shellfish are directly affected by degraded environments through their static feeding behaviour, while juvenile scallop hammerhead sharks experience low abundance of food or consume contaminated prey 14 .
Fish diseases caused by enteric bacteria have been reported in eutrophic waters associated with faecal pollution 15 , thus sharks feeding on prey living in sewage-polluted waters would reflect the bacterial load present in those waters. Studying the effect of the microbiome in conjunction to other factors is important in evaluating the environmental quality of critical habitats of endangered species. There are many ways to monitor pollution in the marine ecosystem, one of which is the use of indicator microorganisms. For example, faecal coliform such as Escherichia coli are indicators of contamination of water with faecal matter from humans or warm-blooded animals. This further implies that other pathogenic bacteria belonging to species of Salmonella, Shigella, Pseudomonas, and Streptococcus can also be present 14,16 . The intestine constitutes an ideal niche for microorganisms due to its readily available source of carbon, minerals, and solutes that are conducive to growth. Because, there is evidence that the microbial colonisation of the intestine of vertebrates starts after hatching or birth, it is influenced by the environmental factors that surround the habitat where a newly hatched or born individual begins to feed 17 . Consequently, investigation of bioindicators in aquatic animals usually involves the characterisation of the intestinal microbial profile at different life stages 17,18 .
For this bioindicator method to be successful, a baseline needs to be established to differentiate between normal colonisers and potential pathogens. Emerging molecular methods for analysing microbial communities allow high-resolution assessments of complex communities. Such protocols usually include culture-independent microbial profiling based on 16S ribosomal RNA (16S rRNA), which is not limited by cultivability, and can often detect even the least abundant members of the microbial community 19 . To the best of our knowledge, no data on the intestinal microbial profile of scallop hammerhead sharks are available in the literature, while studies on the gut microbiome of other shark species are limited. Exceptions are for striped burrfish spinner sharks (Carcharhinus brevipinna), atlantic sharp nose sharks (Rhizoprionodon terraenovae), and sandbar sharks (Carcharhinus plumbeus) [20][21][22][23] .
As part of ongoing projects to understand and protect critical shark habitats in Fiji, this study generates baseline data about the intestinal microbial profile of a representative sample of SHS from the Rewa Delta using 16S rRNA Illumina MiSeq amplicon sequencing (MiSeq). The database thus generated should contribute to the reference library of intestinal colonisers of juvenile SHS and possibly support further studies on trends in microbiological communities and the identification of bioindicator microorganisms as impacted by pollution of the waterways.

Figure 1.
Geographical location of sampling sites in the Rewa Delta, and location of sewage discharge during the spillage that occurred in December 2014, Viti Levu, Fiji (The map has been prepared by Mr Sione Kaituu using online resources from QGIS, version 3.6.0 https://www.qgis.org/en/site/). Grey dots reflect specific places where samples were collected, and the diameter is proportional to the number of sharks taken at that particular site.

Results
Bacterial diversity profile. Since ANOSIM indicated that the variability among the technical repeats was negligible (R = 0.602, P = 0.001), an average was used to compute the bacterial diversity of each shark, which will be referred to by their catch number. The major microbial communities were identified to the family level and analysed for percentage relative abundance, with greater focus given to those that make at least 10% of the community in a minimum of one shark (Fig. 2). Enterobacteraceae was detected in high abundance in most sharks, while Vibrionaceae was seen to be more dominant in most sharks collected after the 9 th of December 2014, namely SHS: 251 (44.2%), 265 (43.6%), 268 (17.5%) and 269 (76.0%). Samples from another shark had a distinctive profile: SHS 229 demonstrated an average of 68.8% relative abundance of Aeromonadaceae. The Shannon Index (1.07 ± 0.51) and Simpson index (0.39 ± 0.02) for SHS 229 was also lower in comparison with the other samples (Table 1), with a Chao 1 estimates of 4.67 ± 0.67, indicating less species richness and evenness. Similar observations were made for SHS 268 (Shannon Index:1.12 ± 0.61, Simpson Index: 0.37 ± 0.04 and Chao 1 estimate: 5.50 ± 0.50) and SHS 249 (Shannon Index:1.73 ± 0.73, Simpson Index: 0.50 ± 0.11 and Chao 1 estimate: 10.33 ± 2.18). A more diverse and evenly distributed microbial profile was attributed to SHS 251 (Shannon Index: 3.29 ± 0.10, Simpson Index: 0.85 ± 0.04 and Chao 1 estimate: 23.25 ± 5.41). In addition, members of the Moraxellaceae, Bradyrhizobiaceae, Pseudomonadaceae, Rhodobacteraceae, Staphylococcaceae, and Streptococcaceae were also detected among these samples.
Further comparison between the microbial profiles of the juvenile SHS sharks is illustrated in a Heatmap (Fig. 3). The Enterobacteriaceae microbes were further classified with major species of the genera Citrobacter, Shigella and Klebsiella. A bacterium closely related to the Citrobacter koseri (Accession No: HQ992945.1) was the most prevalent in most samples, ranging from a highest percentage relative abundance of 41.  www.nature.com/scientificreports www.nature.com/scientificreports/ Compilation of the closely related species to the top five OTUs, in order of highest percentage relative abundance for each SHS (     (Table 3) confirmed that no significant shift could be attributed to the age of the individual, estimated through the degree of healing of the umbilical scars (see Methods) and the diet at the time of capture. However, a trend was observed within the shark population (R 2 = 0.9, p = 0.001, Table 3) and with the date it was caught (R 2 = 0.5, p = 0.001, Table 3

Discussion
Our investigation on the intestinal microbiome of juvenile SHS from the Rewa Delta showed a diverse microbial community, with bacteria including members of the families Enterobacteraceae, Vibrionaceae, Propionibacteriaceae, Aeromonadaceae, Moraxellaceae, Bradyrhizobiaceae, Rhodobacteraceae, Staphylococcaceae, Streptococcaceae, Methanosaetaceae, Bradyrhizobiaceae, Fusobacteriaceae, Chlorothrixaceae, Moraxellaceae and Pseudomonadaceae. These are commonly known intestinal inhabitants of terrestrial and marine vertebrate species including humans. Major members of the groups identified in this study have also previously been found in Sharpnose, Spinner and Sand sharks, although their role as gut microbiota has yet to be confirmed 23,24 . While many have been associated with digestive physiology in other species, research in such areas for shark species is limited 25 , thus making accurate determination of their function difficult.
Other OTUs identified and previously seen in marine taxa were A. felis, C. aggregans, P. adeliensis (Accession No: AJ539105.1), P. actiniarum (Accession No: CP021235.1) and S. sonnei. While not directly linked to sharks, A. felis bacteria have been associated with infection in free-living amoeba and are often isolated from hospital water [28][29][30] . C. aggregans are phototrophic bacteria native to marine environment and are often associated with 'microbial mats' 31,32 . The genus Pontibacter belongs to the phylum Bacteroidetes, members of which are common colonisers of fish intestine and include the isolate P. actiniarum, first characterised from the Sea of Japan as a gram-negative, aerobic bacterium 33 . If not present since birth, published information would suggest that these bacteria originate from preys ingested and could thus colonised the gut of the juvenile SHS sharks.
Two bacterial species dominated the intestinal microbiota of two individual shark sampled. A closer analysis with BLAST related them to A. salmonisidas in SHS 229, and K. pneumoniae in SHS 248. Both microbes are described as opportunistic pathogens associated with nosocomial respiratory tract and urinary tract infections in humans and they were observed in lesions found in the gills and intestine of a dead black tip reef shark 15,34-36 . Being two isolated cases, the probability of these bacteria being indigenous to the intestinal microbial community   Table 3. ANCOVA table generated using the function "Anova" of package 'car' in R, to assess the significance of the impact of date of capture, shark individuality and diet, on the microbial community. Significant Codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '. ' 0.1 ' ' 1.
www.nature.com/scientificreports www.nature.com/scientificreports/ of juvenile SHS is quite low. A possible explanation could be that these sharks were diseased, with the said pathogens eventually dominating the gut microbial biota.
A temporal shift in community was observed in the PCA plot, towards Methanosaeta sp. and P. damsalea. Date of capture was identified as a significant factor affecting PC1. The emergence of Methanosaeta sp. was a major contributor to PC1. There is no indication of the presence of this methanogen as a native archaeon in the gut microbiome of any shark species studied so far. However, the role of these archaea, namely Methanosaeta concilii, in waste degradation, is well documented 37,38 and its presence has previously indicated sewage or effluent contamination of waterways 7,11 . It was later revealed that, midway through this study (6 th of December 2014), a major sewage spill occurred in the Cunningham River, Suva, which discharged about 200 Ls −1 of untreated waste water into Laucala Bay. This discharge continued unabated for 18 days, until temporary control measures were implemented that led to a Government Environmental Emergency Declaration that prohibited swimming and fishing in the affected waters. Even though this discharge released untreated sewage to Laucala Bay and not to the Rewa Delta, which is about 6 km away and in the opposite direction of the trade winds and prevalent currents, the fact that we find a shift in the microbial community after the spill could be explained by two mechanisms. First, contaminated prey may have moved from Laucala Bay to the Rewa Delta via the Vunidawa River that connects Laucala Bay or second, newly born and juvenile individuals move in search of prey and fed in areas reached by the sewage 12 .
With this incident in mind, a preliminary screening of the results was carried out for other potential evidence of sewage pollution. The presence of indicator species such as P. adeliensis and S. sonnei was considered. Originally isolated from fast ice around the Antarctica region, strains of Psychrobacter spp. have been proven effective indicators of pollution in sites with industrial, agricultural and urban effluents 39 . It has also been isolated from marine taxa and other aquatic environments contaminated with hydrocarbons 40,41 . S. sonnei requires specific pH and temperature ranges to survive, and its ideal host is the human gastrointestinal tract. Shigellosis has always been associated with contaminated water, as well as contaminated seafood, and is spread easily in crowded and unhygienic conditions [42][43][44] . The presence of these species in all of the samples might imply that they may be indigenous to this shark species, or could also be an indicator of constant contamination of the waters of the Rewa Delta with existing pollutants associated with nearby agricultural and sewage effluents. Constant contamination from untreated sewage has been reported previously in Laucala Bay and Rewa River and is associated with two main factors. First, the Kinoya sewage treatment plant was built for a population of 77,000 persons but nowadays it supports a population of about 120,000. Second, more than 40% of the main Suva population still uses septic tanks without the ability to remove nutrients and pathogens 12 . Furthermore, some microbial growth is known to be favoured by sewage effluents that increase organic content and nutrient concentrations, as well as decreases in salinity 16 , which might explain the shift in microbe community in sharks after the sewage spill observed in juvenile SHS caught after the 9 th of December 2014.

Conclusions
We have used culture-independent molecular techniques for the characterisation of the intestinal microbial community of fourteen juvenile SHS from a nursery site of the Rewa estuary in the Fiji Island of Viti Levu. This initial study provides baseline information previously lacking for this species in Fiji and in the South Pacific region. While most of the bacteria characterised have previously been identified in other shark or marine species, the www.nature.com/scientificreports www.nature.com/scientificreports/ bacterial community also included many known opportunistic pathogens. Determining whether these bacterial pathogens are part of the indigenous intestinal microbiome of SHS warrants further investigation. The unfortunate sewage spill that occurred during the sampling period could account for the presence of some known indicator microorganisms, namely Methanosaeta spp., Shigella spp. and Psychrobacter spp. It indicates that this technique can successfully identify bioindicator microorganisms associated with polluted environments.

Methods
Study site. The Rewa Delta (RD) (178.55°E, −18.15°S, Fig. 1) is the largest fluvial system in the Fiji Islands; it is found in the largest island of the country, Viti Levu, and it originates from the Rewa River, Fiji's longest river. The RD is characterised by strong currents and high wave actions because of the collision between river runoff and incoming waves/tides via the reef channel. This interaction gives the RD estuarine habitat conditions, such as large fluctuating salinities, a freshwater layer, high turbidity, and tidal waves 7,45 , which collectively make 45% of the RD inaccessible for sampling.

Sampling.
Sampling sites for the current study were located on the RD, and encompassed one third of total Rewa Delta (Fig. 1). Local licensed fishermen caught the sharks studied in this experiment accidentally as bycatch during their regular fishing trips, thus no animals were intentionally sacrificed for the purpose of the current study and no IACUC or equivalent was needed. Three sharks were collected within a day of each other on three different periods, as illustrated in Table 4, except for November 2014, when only 2 sharks were available. The sharks were preserved on ice during transport for a maximum of 3 hours before immediate deep freezing at −80 °C upon return to the laboratory. Prior to DNA extraction, the intestine of each shark, including the proximal, spiral and distal region, excluding the stomach, were isolated from the specimen and any food content carefully removed and visually inspected and recorded for another study. The intestine was further cut into pieces, mixed, and separated to make 3 technical repeats for DNA extraction.

DNA extraction. DNA extraction was performed according to the Council for Scientific and Industrial
Research (CSIR) protocol for the lysis of Corynebacterium species, with modifications [46][47][48] . In brief, approximately 2 g of the mixed intestinal sample was added to 500 μl of lysis buffer (20 mM Tris-HCL at pH 8.5, 2 mM EDTA at pH 8.0), with an additional 20 mg·ml −1 lysozyme (Thermofischer Scientific). The mixture was then incubated for at least an hour at 37 °C in a waterbath before 50 μg ml −1 of Proteinase K (Thermofischer Scientific) was added to the mixture and incubated for 30 mins. Sodium dodecyl sulphate (SDS) was added (10 ml; 20% v/v) and followed by further incubation at 65 °C for 90 mins. The supernatant was collected after centrifugation at 6 000 × g for 10 mins at room temperature, mixed with an equal volume of chloroform isoamyl alcohol (24:1 v/v), and incubated for 1 min at room temperature. The mixture was centrifuged again and the resulting supernatant was precipitated with 60% of its volume of isopropanol for 60 mins at room temperature. After centrifugation at 16 000 × g for 10 mins, the resulting pellet of crude nucleic acid was washed with 500 µl of 70% cold ethanol. The extracted genomic DNA was resuspended in Tris-EDTA buffer (10 mM Tris and 1 mM EDTA, pH 7.6) and stored at −20 °C. DNA extraction was carried out in triplicates for each individual shark sample. Extracted DNA was quantified by the Qubit ® 3.0 Fluorometer. Amplicon sequencing and data analysis. Amplification of the 16S rRNA gene region was verified using a universal primer set of 27 F (5′-AGAGTTTGATCMTGGCTCAG-3′) and 1392R (5′-ACGGGCGGTGTGTRC-3′) 37 before submission of DNA to the Australian Centre for Ecogenomics (ACE, University of Queensland) for paired-end  www.nature.com/scientificreports www.nature.com/scientificreports/ 16S rRNA amplicon sequencing by Illumina Miseq (Illumina Inc., USA). The amplification encompassed the V5 to V8 region of the 16S rRNA gene, using specific primers 803F (5′-TTAGANACCCNNGTAGTC-3′) and 1392Wr (5′-ACGGGCGGTGWGTRC-3′) containing Illumina adapter sequence as modified by ACE (University of Queensland, Australia).
Paired end sequencing data (ACE, University of Queensland), grouped to operational taxonomic units (OTUs) at 97% similarity, and aligned with the 16S rRNA identified sequences in the Greengenes database via the Quantitative Insights Into Microbial Ecology (QIIME) (Version 1.8.0) software package was received (ACE, University of Queensland). The resulting data were further processed in QIIME to calculate the alpha diversity including both the Shannon and Simpson indices as well as Chao 1 estimates, which is based on abundance. OTUs were further classified at the species level by BLAST analysis at 100% similarity (http://www.ncbi.nlm. nih.gov/). It is important to note that such identifications are subject to change over time should new sequences, with closer relationships, be uploaded in the database. A raw OTUs table was imported into R (v3. Statistical and correlation analysis. Analysis of variance in mixed categorical/continuous mode (ANCOVA) was done to test the significance of environmental and host biological parameters on microbial community using the function "Anova" of package 'car' in R. Specifically, PC values of overall communities were used as output, date of capture was treated as a coded continuous factor, and shark individuality, diet, umbilical scar as categorical factors and numerical variables, including weight as a continuous factor. The model was tested to identify the most parsimonious, by elimination of non-significant factors to the minimum parameter. A significance threshold of 0.05 was applied for rejection of the null hypothesis. Analysis of similarity (ANOSIM) was also conducted on microbial community profile between and within triplicate analysis of individual sharks with package 'vegan' in R (v3.2.3) (R Core Team, 2015).