Fish species are commonly used as experimental models in the laboratory. DNA is routinely collected from these animals to permit identification of their genotype. The current standard procedure to sample DNA is fin clipping, which involves anaesthetising individuals and removing a portion of the caudal fin. While fin clipping reliably generates good quality DNA samples for downstream applications, there is evidence that it can alter health and welfare, leading to infection and impacting on the fish’s behaviour. This in turn can result in greater variation in the data collected. In a recent study we adapted a skin swabbing protocol to collect DNA from small-bodied fish, including sticklebacks and zebrafish, without the use of anaesthetics or sharp instruments. A rayon-tipped swab was used to collect mucus from the flank of the fish which was then used for DNA extraction. We subsequently demonstrated that – compared to fin clipping – skin swabbing triggered fewer changes in stress axis activation and behaviour. We also found that swabbing had a smaller impact on subsequent data collection, potentially allowing smaller sample sizes in experimental groups after using this technique, so reducing animal use. Here we provide a detailed protocol explaining how to collect DNA samples using skin swabs from small laboratory fish. Skin swabbing is a refined technique to collect DNA from fish, with the potential to reduce the number of animals used in experiments as well.

Fin clipping sticklebacks and zebrafish

To collect DNA by fin clipping, a single fish was removed from its home tank using a small hand net. The fish was pre-treated with an anaesthetic by placing it in a tank containing 168 mg/L ethyl 3-aminobenzoatemethanesulfonate (MS-222) buffered to pH 7.2 with sodium bicarbonate dissolved in fresh system water. Once the fish was no longer response to touch, it was gently caught in a net and placed into a Petri dish containing a small amount of water. Fins were clipped using a sterile razor blade, taking care to only remove about one third of the caudal fin. The excised fin tissue was placed into a sterile labelled Eppendorf tube, and the fish was moved to a recovery tank containing fresh system. The fish’s behaviour was monitored until it had recovered consciousness, so that it swam around in the tank freely. Fin clip DNA was extracted using DNA extraction buffer containing 15 µl of 20 mg/ml Proteinase K. This was incubated at 57˚C for 30 min, followed by addition of 400 µl chilled isopropanol. The solutions were mixed and the DNA solution was then chilled at -80˚C for 30 min. The solution was centrifuged for 10 min at 10,625 g, the supernatant decanted, and the remaining pellet washed with 190 µl 70% EtOH. After a further centrifugation step (2 min at 10,625 g) the DNA pellet was air dried and resuspended in 30 µl ddH2O.

Skin swabbing sticklebacks and zebrafish

To collect DNA by skin swabbing, a single fish was removed from its home tank using a small hand net. The fish was pre-treated with an analgesic, by placing it in a tank containing 2 mg/L lidocaine dissolved in fresh system water. The fish was immersed in this solution for 45 min immediately before skin swabbing was carried out. The fish was then re-caught in the net, and gently restrained on top of a wetted sponge. The uppermost surface of the fish was exposed to the air to permit DNA collection. A sterile swab was gently stoked along the flank of the fish, from head to tail, four or five times. The swab was placed into a clean labelled Eppendorf tube, and the fish was placed into a holding tank until DNA extraction and identification was complete. The DNA was extracted from the swab by adding DNA extraction buffer warmed to 55˚C and letting it incubate for 2 min. The swab was the removed, taking care to squeeze out as much liquid as possible. The DNA was precipitated by addition of 400 µl chilled isopropanol. The solutions were mixed and the DNA solution was chilled at -80˚C for 30 min. The solution was centrifuged for 10 min at maximum speed (10,625 g), the supernatant decanted, and the remaining pellet washed with 190 µl 70% EtOH. After a further centrifugation step (2 min at 10,625 g) the DNA pellet was air dried and resuspended in 30 µl ddH2O.

Water sample collection.

Cortisol measurements were made following protocols from Ellis et al. (2004), adapted by Sebire et al. (2007, 2009). One hour post-manipulation fish were placed into individual 200 ml glass beakers containing 100 ml of freshly prepared reverse osmosis water. This water contained Instant Ocean salts (Aquarium Systems, UK) ensuring the same conductivity as the system water without background cortisol traces. The fish were left undisturbed in beakers, visually isolated by white dividers, for 30 min during which they released cortisol into the water. After 30 min the 100 ml water sample was collected using a serological pipette and split into two 50 ml Falcon tubes. Blank samples of 100 ml of RO water were also collected each week during the experimental period. 0.5 ml of methanol was added to each water sample prior to snap freezing in dry ice and storage at − 80 °C for extraction at a later date. The fish were euthanised by immersion in 168 mg/L MS-222 until they no longer touch responsive, followed by decapitation. Their dry weight was recorded to within 0.0001 g to allow cortisol concentration to be expressed as ng cortisol/g fish per h.

Cortisol extraction and quantification.

Cortisol was extracted from the water samples by pumping it through Sep-pak Plus C18 solid phase extraction cartridges (Waters Ltd., UK) following the protocol developed by the Cefas Weymouth Laboratory (Ellis et al., 2004). Cartridges were primed with 5 ml of methanol followed by 5 ml of distilled water (dH20) and water samples were pumped through the cartridges at 5 ml/min. Each cartridge was washed with 5 ml of dH₂0, air-dried, wrapped in Parafilm® and stored at −80 °C until elution with 5 ml ethyl acetate. For the quantification by radioimmunoassay (RIA, Sebire et al., 2007; Ellis et al., 2004; Scott et al., 1982) the eluted extracts were evaporated at 45°C under nitrogen and each residue was reconstituted in 500 µl of RIA buffer until assayed. The elution and the quantification were carried out in a blind manner to reduce bias when analysing the data.

Experimental design

All experiments presented here were approved by the Animal Welfare and Ethical Review Board at the University of Leicester, and were carried out under the authority of a UK Home Office project licence (PPL P8F9CCE8B). The experimental design and fish numbers for each group were determined by the NC3Rs Experimental Design Assistant EDA and power analysis. The raw data used in power calculations were taken from previous publications from the Norton laboratory. For example, the cortisol experiment was based upon published data from the Norton lab comparing wild-type zebrafish to a linc5 mutant. Mean control = 20.0±3.9 pg/g, mean treatment group = 9.9±3.2 pg/g. (1-α) is 0.05 and power (1-β) is 80 giving an n value of 4. The outcome measure was radioimmunoassay recordings of cortisol concentrations in water as described in Breacker et al., 2017. The DNA concentrations presented here are based upon data presented in Breacker et al., 2017. The outcome measure was readings of concentration and purity using a spectrophotometer. The PCRs show representative data from experiments published in Breacker et al., 2017. The outcome measure was a photograph of stained bands on the agarose gel. No animals were excluded from the final data analysis. Animals were haphazardly selected for inclusion in each experimental group, and the experimenter was aware of the identity of the samples processed, apart from the cortisol experiment which was conducted blind.

One week prior to experimentation, fish were caught and haphazardly distributed into 13.4 L tanks for sticklebacks (maximum capacity 27 sticklebacks) and 3.5 L tanks for zebrafish (maximum capacity 30 zebrafish). The tanks were located in central positions in large racks within the respective species’ aquarium rooms, ensuring that all tanks received similar illumination and were surrounded by other tanks on the sides. No enrichment was provided, as is standard procedure in our fish facility. Fish were fed ad libitum each afternoon at the end of the experiments (sticklebacks: defrosted bloodworms, zebrafish: Zebrafeed (Sparos)). No fish of either species died during these experiments, and all animals were killed by a Schedule 1 procedure at the end of this study.

Statistical design.

Statistical analyses were carried out using GraphPad Prism7. The DNA concentration was compared using a one-way ANOVA followed by Tukey’s post hoc test. The cortisol data were tested for normality using the using the Shapiro–Wilk test. Since it was not distributed normally, we analysed it using the non-parametric Kruskal–Wallis test followed by a Dunn’s multiple comparisons test comparing each treatment to the control group. This means that the data was not transformed in order to satisfy parametric assumption. All tables were prepared in Excel (Microsoft).

Please contact Dr. Norton if you require any further information.