Seeking sharks: North Sea biodiversity and habitat use revealed by Nanopore sequencing of environmental DNA

Reindert Nijland (Wageningen University, Netherlands) began his plenary talk by describing the North Sea: a small, shallow sea, most of which is <50 m deep. Its location, surrounded by countries including the UK, Norway, Denmark, and the Netherlands, means that the sea is busy with shipping, oil and gas drilling, and fishing. Together, these activities have resulted in a degraded ecosystem. Reindert described the ecological concept of the trophic cascade, in which the top predator eats its prey, reducing the prey’s abundance, and this cascades down the trophic levels. For example, a large fish preying on a smaller fish species would lead to a lower abundance of smaller fish, which in turn increases the abundance of zooplankton they feed on, which itself reduces the algae available for those zooplankton. If the top species, a predatory fish, is taken away, it’s prey fish species would become more abundant and consume more zooplankton, leading to an increase in algae abundance. This can lead to harmful algal blooms, but also in this example, the larvae of the top-predator are part of this zooplankton – meaning that the predatory fish are being eaten by their former prey, making population recovery more difficult. Restoration of this ecosystem would require finding a way of ensuring these predators and their young survive. This process – ‘fishing down the food-web’ – has been happening in the North Sea.

Fishing has reduced the population of sea bass in the North Sea, increasing the population of their prey, including herring. Reindert noted that, whilst an image he shared of a range of species in the North Sea looked like a healthy ecosystem, the baseline has already changed drastically. He showed images of fishing markets from ~100 years ago, displaying huge yields of large cod, halibut, and even bluefin tuna, the latter of which were relatively abundant until the 1960s. He quoted the ‘collective amnesia’ around changes in the species present going back more than a few decades. Reindert asked: how can we restore this ecosystem? Perhaps surprisingly, one potential option is the use of offshore wind farms which, if constructed and managed correctly, can provide shelter and habitat. Areas in which they are built are also closed for fishing. Reindert displayed planned wind farm sites: by 2050, it is planned that ~20% of the North Sea will be used for these. He showed an illustration of how the submerged part of wind turbines could be optimised to provide more habitats; this would require monitoring to ensure that it was effective.

‘Classical’ biodiversity modelling in the North Sea is achieved by fishing, analysis of sediment cores, or diving off wind farms and oil rigs to identify the animals and plants present. Samples are collected, then sent to the lab, where they are identified using a dissection microscope: a process requiring high levels of knowledge of the specimens. Reindert highlighted that this method is not scalable, and is also expensive. Alternatively, this analysis can be performed using DNA sequencing. Here, Reindert introduced the use of environmental DNA, or eDNA, for fish biomonitoring. eDNA can be found in seawater, originating from the epidermis, mucus, or faeces of fish present, providing a non-invasive method of collecting DNA for monitoring. As the eDNA degrades over hours or days after release, it also allows temporal analysis. eDNA is extracted from samples for sequencing; biodiversity can then be assessed via metabarcoding, which enables the identification of species from a mixed community. This process is similar to 16S rRNA gene analysis – a marker gene is PCR-amplified using universal primers. The sequenced amplicon is then compared to a database. Reindert described how he and his team chose a marker for this method. Classical metabarcoding uses the mitochondrial cytochrome oxidase I (COI) gene; however, the universal primers available are too universal to study only fish from eDNA samples with high levels of background. The same issue was present for a mitochondrial 16S marker. The short fish mitochondrial 12S marker presented a great choice, but its length limited species resolution. However, Reindert noted that long nanopore sequencing reads enable the sequencing of longer markers: instead of sequencing 300-600 nucleotides, amplicons of 2 kb or longer can be sequenced. Reindert noted that more information on this would be presented in Karlijn Doorenspleet’s talk at NCM 2020. Karlijn and her colleague, Lara Jansen, developed primers to enable a good level of fish species resolution via amplicon sequencing.

Reindert described how eDNA metabarcoding using nanopore sequencing is ‘a great combination’ as it is fast, relatively affordable, produces long reads, and enables real-time access to data and sequencing on location. Using the portable MinION and MinION Mk1C sequencing devices, Reindert and his colleagues have been assessing the biodiversity of the North Sea Borkum reef grounds. An artificial reef had been placed here; divers then collected water samples from the reef and, on board a ship, filtered the DNA from the water and extracted it from the filter. The marker of choice was then amplified from each sample and the libraries prepared for sequencing on MinION Flow Cells or the smaller Flongle Flow Cells, all whilst still at sea. He noted that this same technique had been used by his colleague Lukas Müller to study Bull sharks in Mozambique, on a small dinghy. Reindert noted that, when it comes to analysis, raw read accuracy is of low importance, with consensus accuracy much more important, as it allows for the analysis of features such as haplotype diversity. To obtain these consensus sequences, Saskia Oosterbroek developed Decona: from Demultiplexing to consensus for nanopore amplicon data (https://github.com/saskia-oosterbroek/decona).

Reindert noted that the lack of information present in DNA reference databases poses the current major bottleneck in DNA metabarcoding studies. He is involved in a project, Interreg, which aims to both develop new methods using DNA to study biodiversity and also sequence many animals from the North Sea to improve this database. They ideally plan to sequence a larger part of the mitogenome, or full mitogenome, where possible. Reindert pointed out that DNA biomonitoring can also enable analysis of diet, via sequencing of stomach contents or faeces, the constituents of which might not be otherwise possible to identify. Here, blocking primers are designed for the host organism. He also described the possibility for future autonomous sampling of eDNA, in which sampling automatically occurs at set timepoints, ideally coupled with automated library prep and sequencing, followed by data transfer via satellite. Reindert hopes that this will be possible using nanopore technology within the next few years. He suggested that, in future, a nanopore sequencer could be integrated into a sensor device with which marine animals are tagged, and perhaps deployed on underwater drones. He suggested that, in under ten years from now, it could be possible to perform online, real-time diet analysis in situ using eDNA sequencing.