Environmental DNA detection tracks established seasonal occurrence of blacktip sharks (Carcharhinus limbatus) in a semi-enclosed subtropical bay

Study species and area

The blacktip shark (Carcharhinus limbatus; Fig. 1B) is a cosmopolitan species, encountered globally in tropical and subtropical waters^34,35,36. It is a commercially and recreationally important species in the southeast U.S. and Gulf of Mexico shark fisheries, as well as being one of the top four species in the global shark fin trade^37,38. The species has moderately slow growth rates and low fecundity; producing pups in alternating years after a 10–12 month gestation period, giving birth to 4–7 pups^34,39. This leads to low natural recruitment abilities, rendering their populations susceptible to stock collapses from overexploitation. However, the U.S. C. limbatus population is well managed and sustainably fished. In this study we refer to four life-stage categories: neonate, young of the year, immature, and mature, following the definitions of Castro³⁵. Neonates are animals with an open umbilical scar, these scars close within the first 2–3 weeks of life. Young of the year (YOY) have closed umbilical scars, which are typically still visible. YOY can also be distinguished from older immature animals based on size⁴⁰. Mature males are distinguished from immature males based on claspers that are elongated and calcified, while for females, 50% size-at-maturity was used to distinguish between immature and mature animals³⁹. Along the Florida Gulf Coast, neonate, YOY, and immature C. limbatus seasonally use inshore nursery areas including Terra Ceia Bay^1,41,42. Terra Ceia Bay is a small (~ 5 km × 1.5 km), shallow (~ 4 m maximum depth) bay with a narrow (~ 1 km) opening into Tampa Bay (Fig. 1A). As C. limbatus is a migratory species that prefers temperatures above 21 °C, there is an influx of immature sharks into the bay during April after an absence for most of the winter. Gravid females enter the bay for parturition in May and June^1,33,34. Natural mortality of neonates is highest within the first 15 weeks of life, which is reflected in a sharp decrease in their abundance from July into August. The remaining young of the year and immature animals leave the bay and migrate southward during the fall months^1,33. These patterns have held with continued sampling in the bay until the present, although some immature animals now stay in the bay until later in the year and return earlier, possibly because waters have warmed in recent years (Gardiner, unpublished data).

Shark sampling

Shark surveys in Terra Ceia Bay were conducted as part of the GulfSPAN survey (www.fisheries.noaa.gov/southeast/endangered-species-conservation/shark-and-sawfish-surveys-and-tagging), a fishery-independent effort to assess shark diversity and relative abundance along the Florida Gulf Coast in order to inform stock assessments. All protocols for the handling and use of animals were approved by the Institutional Animal Care and Use Committee at the University of South Florida (protocol # W IS00004541) and sampling was conducted in accordance with Florida Fish and Wildlife Conservation Commission Special Activities Licenses SAL-18-1666-SRP and SAL-19-1666-SRP. The surveys are conducted each year from April to October. In order to conserve limited resources for this survey, sampling is not conducted from November to March when shark abundances are known to be very low due to decreased water temperatures during these winter months. The primary sampling method of this project for 2018 was monofilament gillnet. However, as this method primarily captures neonate and YOY sharks, it was considered not to be an appropriate comparison on its own to eDNA, which is shed by all life stages. Therefore, bottom longline gear was also deployed on a monthly basis in 2019 to sample a wider variety of life stages and to provide a better comparison to eDNA when combined with gillnets. For each month of sampling in 2019, catches were summed over 3–4 longline and gillnet stations and used to calculate monthly catch-per-unit-effort (CPUE), defined as the number of individuals caught per station per hour. All sampling coordinates and details of the fishing gear are summarized in Supplementary Material 1 and Supplementary Material 2, respectively.

eDNA collection and processing

Water samples were collected on a monthly basis concurrent with GulfSPAN surveys during June–October 2018 and April–October (except July) 2019. Four to six water samples were collected at each sampling location per monthly survey, immediately after gear is set in order to avoid contamination from captured animals. eDNA samples were collected in mid-water to avoid catching suspended sediment, ensuring that only recently released eDNA was collected^43,44. A total of 58 water samples of 2 L each (Supplementary Material 1) were collected with a Kemmerer type water sampler and vacuum filtration was carried out immediately after collection with a Pegasus Alexis peristaltic pump (www.fondriest.com). The cups containing the hydrophilic polyethersulfone (PES) filters (Pall Corporation; 47 mm diameter; 0.45 µm pore size) were placed in a clean, iced cooler (containing only eDNA samples) for transport to the laboratory. The filters containing sample filtrates were stored in sterile 5.0 mL cryogenic screw-cap vials containing silica beads. The silica beads function as a desiccator, drying out the filters, preventing the DNA from degradation²⁶. The filters were then stored at − 20 °C until extraction. DNA extraction from half of each filter was performed using the DNeasy PowerSoil Kit (www.qiagen.com), following the manufacturers’ protocol, with three additional specifications. As the outer circumference of the filter does not contain any DNA (due to placement of the filter cup), it was cut off and discarded. The remainder of the half filter was cut into small pieces prior to being added to the bead tubes for the first step of the extraction process. The vortexing (bead-beating) step of the protocol was performed in a shaker at 58 °C (for 10 min) instead of at room temperature, in order to maximize the DNA yield. Genomic DNA was eluted into 100 μL and frozen at − 20 °C until further processing.

Real-time PCR assay development

Primer design

Primers were designed to target a 149 bp fragment of the mitochondrial (mtDNA) nicotine adenine dinucleotide dehydrogenase subunit 2 (NADH2) gene within the C. limbatus genome, while excluding cross amplification with other elasmobranch species known to co-occur in the study area (Rhizoprionodon terraenovae, R. porosus, Carcharhinus brevipinna, C. acronotus, C. leucas, Sphyrna mokarran, S. lewini, S. zygaena, and S. tiburo). Primers were designed manually using Geneious Prime, version 2019.2.3 (https://www.geneious.com). NADH2 sequences from C. limbatus and the non-target shark species were downloaded from GenBank (Supplementary Material 3) and aligned using the Muscle 3.8.425 plugin in Geneious⁴⁵. Forward (588 F-limbatus-NADH2: 5′-TGCCCCCAATCTCACCTTAC-3′) and reverse (776 R-limbatus-NADH2: 5′-CCGGAAAGTGGGGGTAATCC-3′) primers were designed to amplify the fragment of interest exclusively in C. limbatus. This was accomplished by maximizing the number of mismatches in the ligation sites of both primers in the nine exclusion species. In order to confirm primer specificity, they were first tested by applying conventional PCR to total genomic DNA (gDNA) extracted from eight C. limbatus individuals and from eight individuals per each of the nine species known to co-occur in the study area. Genomic DNA was extracted from tissue samples using the DNeasy Blood & Tissue kit (www.qiagen.com) following the manufacturer’s protocol. Each amplification reaction was performed in a total volume of 15 µL and consisted of: 7.5 µL of Master Mix ‘Applied 2×’ (Applied Biosystems), 3.5 µL of DNase/RNase-free water (Fisher Scientific), 1 µL of each primer (10 µM) and 2 µL of genomic DNA (10 ng/µL). The PCR profile included an initial denaturing step of 95 °C for 2 min, 35 cycles of 95 °C for 30 s, 63 °C for 30 s and 72 °C for 30 s and a final extensions step of 72 °C for 5 min. The quality of all amplifications was assessed by electrophoresis, running the products through a 1.5% agarose gel stained with Gel Red (Gentaur, Kampenhout, Belgium) and visualized on a UV light platform. Subsequently, the primers were tested on the real-time PCR platform for amplification of tissue-derived gDNA from C. limbatus and the non-target species. PCR products were sequenced in both directions to confirm species identity, using an ABI 3730 genetic analyzer (Applied Biosystems). Sequences were checked using Geneious.

Real-time PCR

Concentrations of C. limbatus NADH2 fragments in our eDNA samples were checked by running six replicate reactions for each sample, using the Chai Open PCR System platform (Chai Biotechnologies, Santa Clara, CA, USA). The real-time PCR recipe was optimized to a final volume of 15 µL, containing 7.5 µL of PowerUp SYBR Green Master Mix (Applied Biosystems), 2.5 µL of DNase/RNase-free water (Fisher Scientific), 1 µL of each primer (5 µM) and 3 µL of eDNA template. The real-time PCR program consisted of 50 °C for 2 min, an initial denaturation step at 95 °C for 2 min, followed by 80 cycles of 95 °C for 30 s, 62 °C for 30 s, and 72 °C for 30 s. A final melting curve analysis step at 60 °C for 1 min was followed by an increasing ramp temperature of 5 °C per second to reach 95 °C for 5 min. Using 80 amplification cycles in real-time PCR is not standard protocol when working with tissue extractions, but with appropriate negative controls it has been found to be useful for detecting the presence of eDNA that which is usually present in very low concentrations¹⁶. The melt curve protocol was used to detect primer dimers, to check specificity by determining that a single amplicon was produced, and to obtain a melting temperature (Tm) value for that sample, which may then be compared with the standard curve of C. limbatus DNA. Each real-time PCR run consisted of 16 reactions (two 8-well PCR strips per run), including two positive (13.1 × 10⁻¹ ng/μL gDNA), and two negative (DNase/RNase-free water) controls. Cq values of the standard concentrations were compared between independent runs to check for consistency and replicability of results.

Primer efficiency, LOD and LOQ

Primer efficiency, limit of detection (LOD), and limit of quantification (LOQ) were assessed following the protocol and curve fitting method described in Klymus et al.⁴⁶. For the standard curve, tenfold dilutions (in ddH₂O) of C. limbatus gDNA derived from tissue extractions were used. The dilution series spanned seven orders of magnitude, ranging from 13.1 × 10⁻¹ to 13.1 × 10⁻⁷ ng/μL, measured with a Qubit 2.0 fluorometer (Invitrogen, Carlsbad, CA, USA). Each dilution was run ten times and reported Cq (cycle quantification) values were used to determine primer efficiency, LOD, and LOQ⁴⁶. LOD and LOQ were determined at a concentration of 13.1 × 10⁻⁶ ng/μL of gDNA (the conc. at which at least 95% of the replicates amplified with a CV of 54% or less, see “Results” section). Amplification efficiency was determined by plotting Cq values against gDNA dilutions and calculating the linear slope and the coefficient of determination (R²) value.

Contamination control

Strict adherence to contamination control procedures was followed during all field and laboratory stages. In the field, disposable gloves were used and changed between each sample. A section of the boat was allocated as a clean area for water collection and filtration. Work surfaces, water collection and filtration equipment were cleaned prior to and after each use with a 50% bleach solution. To prevent cross-contamination between samples, filtration was performed using disposable, single-use filter funnels, which were subsequently stored individually in two layers of double sealed plastic bags prior to being stored on ice. All eDNA sample processing and analyses were conducted in the Crustacean Genomics and Systematics Lab at Florida International University where there had never been any shark tissue or PCR product present. Extractions, pre-PCR preparations and (post)PCR procedures were physically and temporally separated. All laboratory equipment was cleaned with bleach solution and subsequently exposed to UV sterilization (including ddH₂O used for PCR reactions) for 15 min, before and after each use. Pertaining to filter extractions, bags containing the individual filters were cleaned on the outside with bleach prior to entering the lab. All work surfaces were cleaned with 50% bleach solution prior to and after each use, and between processing of individual samples. All equipment used for the extractions was likewise cleaned with bleach between each individual sample. Disposable gloves were used and changed between each sample or in addition, more often when deemed appropriate. Aerosol barrier pipette tips were used for all laboratory procedures. C. limbatus tissue extraction was performed months prior to the onset of the eDNA laboratory work, in a separate dedicated tissue extraction room. Separate laboratory rooms, located in different parts of the building, were used for PCR and real-time PCR experiments. No equipment was shared between the laboratories and commuting between the different laboratories within the same day was prohibited, this included colleagues who were not part of this study. In order to identify potential contamination, negative controls were added at multiple stages. Negative controls were collected in the field by filtering 2 L of tap water, processed as standard samples. In the laboratory, DNA extraction blanks (elution buffer from extraction kit) and PCR blanks were included.

Data analysis

Cq and melting temperature (Tm) values were used to determine the presence/absence of C. limbatus eDNA in a sample, by comparison with those of the standard curves and the positive controls. The Cq value indicates the cycle at which the fluorescence produced by the DNA amplification in a sample surpasses a preset threshold of fluorescence background noise, which is a proxy for the quantity of eDNA present in a sample. When a sample has a Cq value, it means that target DNA is present assuming there is no non-specific amplification. The Cq value represents the PCR cycle at which amplification is first detected and is positively correlated with the template DNA present in the sample. Filters were considered positive when at least two out of the six replicates produced a Cq value, in combination with an associated Tm value close (± 2 °C) to the mean Tm of the standard curve (78.2 °C). For each filter sample, the average Cq of all positive replicates were calculated to obtain a single value per filter.

Because many detections presented Cq values below LOQ and LOD (Cq>42.09, see “Results” section), they were treated as qualitative data (presence or absence of target DNA on the filter)⁴⁶. A Chi square proportion test was performed to test for a difference in the proportion of C. limbatus positive filters between the high (April–July) and low abundance (August–October) seasons over 2018 and 2019, and the catch data (CPUE) was also compared between the same two seasons in 2018 and 2019 using the Wilcoxon signed-rank test. In order to determine the correlation between the proportion of positive filters and catch data (CPUE) per month, a non-parametric Spearman correlation was performed. All statistical analyses were performed in R v 3.4.0 (https://www.R-project.org/).

Source: Ecology - nature.com