Copepods enhance nutritional status, growth and development in Atlantic cod (Gadus morhua L.) larvae — can we identify the underlying factors?

Fish material and experimental design

This study was carried out within the Norwegian animal welfare act guidelines, in accordance with the Animal Welfare Act of 20th December 1974, amended 19th June, 2009, at a facility with permission to conduct experiments on fish (code 93) provided by the Norwegian Animal Research Authority (FDU, www.fdu.no). The first trial was assumed to be a nutrition trial not expected to harm the animals, no specific permit was required under the guidelines. The second experiment was approved by FDU, FOOTS ID 5448.

Fertilized eggs for Exp-1 were obtained from Atlantic cod broodstock of coastal origin, western Norway, which were held at the Institute of Marine Research (IMR), Austevoll Research Station. The naturally spawning fish, 45 females and 25 males, were kept in a tank of seawater (yearly means; temperature: 7.9±0.3 °C, salinity: 34.7±0.2 g L⁻¹). The eggs were collected overnight March 15, 2012, using an egg collector as described by van der Meeren & Lønøy (1998). Mean egg diameter was 1.29 mm, and fertilization rate was 79%.

Eggs for Exp-2 were obtained from siblings of the broodstock used in Exp-1, kept at sea in net pens at the IMR facility in Parisvannet, Øygarden. Eggs were stripped on April 4, 2013 from three females (1.2 L in total), and fertilized with mixed milt from 10 males. Sea temperature at time of stripping was 4.0 °C. The fertilized eggs were immediately transported for about three hours by car to IMR-Austevoll in a 15 l closed plastic bag with an air space, held in a polystyrene box. On arrival at Austevoll, the temperature had risen to 4.7 °C. The eggs were disinfected using 400 mg L⁻¹ glutaraldehyde for 8 min according to Harboe, Huse & Øie (1994). Mean egg diameter was 1.30 mm, and fertilization rate was 86%.

Water for the Austevoll station is pumped from a depth of 168 m, sand-filtered, temperature-adjusted and aerated before it enters the tanks. The eggs were held in 70 L black polyethylene incubators with conical bases; 1 L of eggs per tank in Exp-1, and 0.20 L in Exp-2. The tanks had gentle aeration (van der Meeren & Lønøy, 1998), water supplied at 0.5 L min⁻¹, 5.8–6.1 °C, and with continuous light. Dead eggs were removed daily and measured volumetrically. In Exp-1, 50% hatching (day 0) was reached on 8th of April while in Exp-2, 50% hatching on the 18th of April. In both experiments, cod larvae were transferred to start-feeding tanks on 4 dph.

In Exp-1, 50,000 cod larvae were stocked in each of the six black PEH 500 L start-feeding tanks with gentle aeration, in order to prevent feed organisms and larvae from clogging the sieves. Larvae in three of the tanks (randomized) were fed copepods, while in the other three tanks they were first-fed enriched rotifers followed by enriched Artemia. Water flow was increased with age, from 1 L min⁻¹ initially to 6 L⋅ min⁻¹ at weaning, and further to 10 L min⁻¹ at experiment termination, giving a water exchange rate of 32 tank volumes day⁻¹ (van der Meeren et al., 2011). The temperature was gradually increased from 8 °C at transfer to 11.6 °C at 11 dph (Fig. 1), where it was kept for the rest of the experiment. Salinity was constant at 34.7 ± 0.2 g L⁻¹. The tanks were equipped with air skimmers to remove any biofilm on the water surface. An automated cleaning arm (van der Meeren et al., 1998), rinsed the bottom of the tanks as needed from 12 dph. To modify the visual feeding environment and enhance feeding (Naas, Næss & Harboe, 1992; van der Meeren, Mangor-Jensen & Pickova, 2007), 10 ml algal paste (Marine microalgae concentrate, Nannochloropsis sp., Instant Algae^®, Nanno 3600; Reed Mariculture, Campbell, California, USA) was added to each tank 15 min prior to each meal until 36 dph (Fig. 2). A 16L:8D photoperiod was employed, with 30 min simulated twilight at each dusk and dawn. The light source consisted of two 20 W tungsten halogen light bulbs (12 V) over each tank that provided 300–500 µW/cm² at the water surface (IL 1400A photometer; International Light Inc., Boston, Massachusetts, USA).

In Exp-2, 1900 cod larvae were distributed randomly to each of 12 green PEH 50 L startfeeding tanks at 4 dph. Three feeding regimes were employed: larvae in four tanks received enriched rotifers followed by enriched Artemia, another four tanks received copepod nauplii and copepodids with prey whose size increased with larval age, and the last four tanks were given copepod nauplii without increases in prey size through the larval period (Fig. 3).

The tank design and setup employed in Exp-2 were miniatures of the tanks in Exp-1, but without cleaning arm and surface skimming. Cleaning was done manually by siphoning the bottom when necessary. Water quality was similar to Exp-1, except that the water was not temperature-adjusted. The temperature therefore ranged between 7.1 and 8.6 °C (Fig. 1). The water flow was kept stable at 0.42 L min⁻¹ during the experiment. Oxygen saturation was not measured as the daily water-exchange rate was high relative to tank size (Fig. 1). Illumination was provided by broad-spectrum fluorescent light tubes (Philips 965 TL-D 90 De Luxe Pro; Philips, Amsterdam, Netherlands) providing 200–300 µW/cm² at the water surface. Algal paste (3.3 mL tank⁻¹, Nannochloropsis sp.; Reed Mariculture, Campbell, California, USA) was added 15 min prior to every feeding from 4 to 33 dph (Fig. 3).

Larval feed and feeding regimes

In Exp-1, the cod larvae were fed three times a day (09:45, 15:15 and 19:00). The rotifer cultures were seven-day-old batch cultures, kept in 2 m³ conical tanks at 24 °C at densities ranging between 850 and 2600 rotifers mL⁻¹, and fed four times h⁻¹. The diet consisted of dry baker’s yeast (0.11–0.18 g million rotifers⁻¹) and Rotifer Diet ^® (0.3–1.5 g million rotifers⁻¹; Reed Mariculture Inc., Calfornia, USA). The rotifers were enriched before they were fed to the larvae. The enrichment protocol was performed daily by moving the desired number of rotifers into an enrichment tank at densities in the range of 1000–2000 rotifers ml⁻¹. The culture was given 0.6 mg Sel-Plex^® (Alltech, Vejle, Denmark) and 0.2 g Larviva Mulitigain (Biomar, Trondheim, Norway) per million rotifers over a 20 min period each hour from 12:00 to 08:00 the following day. After enrichment the culture was rinsed in clean 24 °C seawater, then cooled to the same temperature as the start-feeding tanks, and finally stored under moderate aeration and continued cooling until given to the cod larvae. From 32 to 35 dph the rotifers were gradually replaced with enriched Artemia, which were fed until 63 dph. The Artemia were hatched from SepArt cysts (INVE Aquaculture, Dendermonde, Belgium), and enriched at densities of maximum 500 Artemia mL⁻¹ using 0.2 g Larviva multigain L⁻¹ tank volume, given four times from 20:00 to 08:00. After enrichment, the Artemia were rinsed in clean seawater, cooled, and stored like the rotifers. From 58 to 63 dph, Artemia were fed to the cod larvae only once a day at 15:15. Weaning to the formulated diet AgloNorse ^® 400–600 µm (Tromsø Fiskeindustri AS, Tromsø, Norway) started by hand feeding on 55 dph, and continued using feeding automats from 57 dph (Fig. 2A).

Copepods were collected from “Svartatjern,” a nearby 25 000 m³ sea-water pond system (Naas, van der Meeren & Aksnes, 1991). Pond operation, hydrographical and biological monitoring, and copepod filtration system and harvest procedures are described in detail by van der Meeren et al. (2014). A UNIK-900 wheel filter (Unik Filtersystem AS, Os, Norway) was used to fraction, concentrate and harvest the copepods. The filter fractions used in Exp-1 (Fig. 2B) were 80–150 µm (4–11 dph), 80–180 µm (11–18 dph), 80–212 µm (18–23 dph), 80–250 µm (23–36 dph), and 80–350 µm (37–44 dph). The collected copepods were concentrated under aeration by a 80 µm plankton net and transported in a 10 L bucket with lid and aeration for 7 min by boat to the larval rearing facility. On arrival, the concentrated mixture of copepods was diluted to 30–60 L with 7–8 °C seawater and stored under the same conditions as the enriched and cooled rotifers and Artemia. Viability was checked under a 40× magnification binocular microscope. Copepods were harvested from the pond one to three times a day, depending on larval needs and plankton availability in the pond. The quantities of copepods added to the larval tanks are shown in Fig. 2B. The larvae were weaned to a formulated diet from 37 dph as described above. At 70 dph, AgloNorse was switched to Gemma Diamond^® 1.0 (Skretting, Stavanger, Norway).

In Exp-2, larvae in four tanks were fed rotifers and Artemia, which were produced as in Exp-1. Rotifers were used from 4 to 34 dph, while Artemia were used from 34 to 47 dph (Fig. 3A). The copepods were also obtained as described in Exp-1. The four tanks with larvae fed copepods which increased in size with larval age were given the 80–150 µm fraction from 4 to 15 dph, 80–180 µm from 16 to 27 dph, 80–212 µm on 28 to 34 dph, and 212–250 µm from 35 dph and onwards (Fig. 3B). Finally, larvae in the four tanks given small sized prey only were fed copepod nauplii collected from the 80–150 µm fraction from 4 to 15 dph and 80–180 µm for the rest of the experiment (Fig. 3C). Exp-2 was terminated on 47 dph. To measure copepod size, samples were fixed in 1:50 Lugol’s solution. Samples taken in the morning and afternoon on four days were photographed with two images per sample (Leica MS5 with an Olympus DP70), and the total length of all nauplii and the prosome length of all copepodids on the images were measured using ImageJ (v2013-2, National Institute of Health, Bethesda, Maryland, USA).

Larval sampling and staging

Due to the unequal growth of the larvae in the different dietary groups, the dates of sampling in Exp-1 were adjusted according to larval size rather than age. Before the trial, a system of stages was developed, which was based on the sequence of mineralisation of craniofacial bones and is correlated with certain ranges of standard lengths (Dr. Ø Sæle, NIFES, pers. comm., 2012). Here we report data from larval stages 0–5 and juveniles. Sampling at stages 0, 1, and 2 corresponds to 4, 11 and 22 dph for both groups since the growth rate was similar in the early stages. For stages 3, 4, 5 and juveniles the time for sampling corresponds to 28, 36, 52 and 73 dph in the copepod groups and 30, 53, 70 and 84 dph in the rotifer groups (Table 1). Additional samples for growth measurements were collected in between these ages. The samples were collected two hours after the first meal of the day in order to standardise gut filling conditions. The reason for sampling larvae with a full gut was that parallel samples were taken for studies of digestive physiology, the results of which are reported elsewhere. The time of feeding prior to sampling was adjusted to keep a fixed standardized time between feeding and sampling for all tanks. Larvae used for length and dry weight analysis were individually sedated, photographed, killed with an overdose of MS 222, rinsed in filtered distilled water, laid in pre-weighed aluminium beakers, and freeze-dried. The beakers with larvae were weighed, and larval weight determined by subtraction. Larval standard length, from the tip of the snout to the end of the notochord, was measured on the photographs. At the two last samplings for each of the groups, larval length and wet weight were measured. Larvae used for analysis of nutrient composition were sampled at stages 3 (at the end of the rotifer period) and 5 (both groups were weaned after sampling at stage 4). Pooled larvae (one sample per tank) were killed by an overdose of MS 222 and sieved through a plankton filter which was subsequently patted dry from underneath with a paper towel. Samples were distributed to tubes for the different nutrient analyses, immediately frozen on dry ice, and then transported to National Institute of Nutrition and Seafood Research (NIFES) and stored at −80 °C until analyses.

Exp-2 was basically a growth and survival study, and larval samples were taken for weight and length only, using the protocol described above. The larvae were not staged and the samples were taken at the same ages for all groups.

Data analysis and statistical analysis

Specific growth rate (SGR) and daily length increment (DLI) were calculated according to Ricker (1979):

SGR (% day⁻¹) = 100(e^g − 1), where g = (lnW₂ − lnW₁)(t₂ − t₁)⁻¹

DLI (mm day⁻¹) = (L₂ − L₁)(t₂ − t₁)⁻¹

W₂ and W₁ are dry weights, while L₂ and L₁ are lengths at times t₂ and t₁, respectively.

Differences in survival between the two treatments in Exp-1 were compared using Mann–Whitney U test, while the three treatments in Exp-2 were compared using Kruskal–Wallis ANOVA. Size (length, weight) was compared using nested two-way ANOVA (treatment, time) with tanks nested under treatment, followed by a Tukey HSD test if significant if p < 0.05. Growth rates (SGR and DLI) were compared for separate stages. In Exp-1 compared by means of Student’s t-test, and in Exp-2 by a one-way ANOVA and if significant (p < 0.05) followed by Tukey HSD.

The data on the nutrient composition of the feeds were first checked for homogeneous variances using Levene’s test and log transformed if significant, using the Statistica software package (ver. 12, StatSoft Inc. Tulsa, Oklahoma, USA). They were then analysed by ANOVA and Tukey HSD test for unequal sample sizes. Nutrient composition of larvae in Exp-1 on stages 3 and 5 were analysed by t-tests.

Survival

In Exp-1, between 4223 and 6210 larvae were collected from each of the tanks at the end of the trial. Estimates of survival, calculated as the proportion of the larvae counted out of the tanks at termination, divided by the number of larvae added to the tanks at start (after sampling at 4 dph, N = 50 000) corrected for samplings (N = 10 000), ranged between 11 and 16%. Mean (± SD) survivals were 14 ± 2% and 12 ± 1% in the copepod and the rotifer/Artemia fed groups, respectively, and were not significantly different between the treatments (Mann–Whitney U-test, p = 0.190).

Estimated survival at the end of Exp-2, calculated as above, ranged from 10 to 47%. Mean survival was lower in the group fed rotifers/Artemia (13%, range 10–20%) than in the small copepod (39%, range 35–47%) and large copepod (37%, range 31–43%) groups (Kruskal–Wallis ANOVA, p < 0.001). The large and small copepod groups did not differ (Mann–Whitney U-test, p = 0.837).

Growth

In Exp-1 the larvae were 4.13 ± 0.12 mm (N = 25) at hatching, and 4.46 ± 0.18 mm (N = 24) at 4 dph (Stage 0). The growth in weight and length, DLI and SGR of larvae in Exp-1 were not significantly different between the groups at stages 0–2, from 4 to 22 dph (Figs. 5A, 5C and Table 3A). The copepod group displayed significantly improved DLI and SGR compared to the intensive group between stages 2 and 4 (t-test, p < 0.01), while there were no significant differences from stages 4 to 5 or 5 to juvenile in length increment or growth rate (Figs. 5A, 5C and Table 3A)

In Exp-2, the larval dry weight at 4 dph was 0.083 ± 0.024 mg and length 4.51 ± 0.21 mm (Figs. 5B and 5D). Mean length and weight did not differ between the treatments until 20 dph (nested ANOVA, p > 0.935). Thereafter, both copepod groups were significantly longer and heavier than the intensive group (nested ANOVA, p < 0.005). No significant differences in length were observed between larvae of the two copepod treatments (nested ANOVA, p > 0.398) except at 40 dph, when the group fed large copepods was significantly longer than the group fed small copepods (p < 0.001). Similarly, there were no significant difference in dry weights between the two copepod groups except at days 40 and 47, where the larvae fed the large copepods had higher dry weights than those fed the small copepods (p < 0.015, Figs. 5B and 5D).

The DLI and SGR in Exp-2 were not significantly different between the treatments from 4 to 20 dph (Table 3B). Thereafter (days 20–34 dph and days 34–47 dph), growth was slower in the intensive group, particularly during the last period, when the SGR of the intensive group fell to only half of the previous period (from 8.9 to 4.4% day⁻¹) (Table 3B). There were no significant differences in DLI or SGR between the two copepod groups (Mann–Whitney U-test, p > 0.15).

Prey size and nutrient composition of live feed organisms and cod larvae

The nauplii had overall lengths of 60 to 225 µm, while the copepodids had prosome lengths of 225 to 660 µm with little overlap (Fig. 4). Rotifers typically have a lorica length of 100–200 µm and Artemia instar I about 4–600 µm.

The nutrients present at lower levels in rotifers than in copepods in Exp-1 were protein (p < 0.001), taurin (p < 0.001), astaxanthin (p < 0.001), iodine (p < 0.05) and zinc (p < 0.001) (Table 4). There were also differences in fatty acid composition, in that the rotifers’ ARA fraction (p < 0.0001) was higher than in copepods and EPA was lower (p < 0.01). Furthermore, copepods had a higher ratio of phospholipids (PL) to total lipid (TL) than rotifers (p = 0.01). DHA, the lipid level measured as total fatty acids, all the vitamins analysed, as well as manganese, copper and selenium were higher or similar in rotifers than in copepods. The Artemia had a lower protein content than copepods (p < 0.001), similar to that in rotifers, and were characterized by taurine levels between those in rotifers and copepods (different to both, p < 0.01), high levels of cantaxanthin, no astaxanthin, and higher average levels of vitamin C, E and K (p < 0.01) compared to copepods. Iodine (p = 0.02) and zinc (p < 0.001) levels were lower in Artemia than in copepods. The fatty acid composition was characterized by similar DHA levels, but a much higher EPA:ARA ratio in copepods compared to Artemia. The level of total fatty acids and the ratio of PL to total lipids were similar to that in copepods.

The levels of selected nutrients in cod larvae fed rotifers/Artemia or copepods in Exp-1 were analysed at stages 3 and 5. Dry matter (DM) and protein did not differ between larvae fed the two diets (Table 6). At stage 3, taurin was almost 20 times as high in larvae fed copepods as in those fed rotifers/Artemia (40 and 2.1 µmol g⁻¹ DM, respectively, p < 0.001). This difference became smaller, but was still present at stage 5 (46 and 39 µmol g⁻¹ DM, respectively, p < 0.05). Vitamin levels were either similar in the two groups, or higher in larvae fed rotifers/Artemia than in those fed copepods. The same relationship applied to the minerals manganese, copper, zinc and selenium. Iodine was lower in larvae fed rotifers than in copepods at stage 3 (p < 0.05), but not at stage 5, where the groups had similar levels of iodine. The fatty acid composition at stage 3 partially mirrored the fatty acid composition of the feed, with a much higher EPA:ARA ratio in larvae fed copepods than in those fed rotifers (8.4 and 1.1, respectively, p < 0.01). Larvae fed rotifers had a higher DHA level than larvae fed copepods (37 and 31% of total fatty acids, respectively, p < 0.001). Although there were still statistical differences, the fatty acid composition appeared similar between the larval groups at stage 5. The lipid level measured as total fatty acids was similar in the two groups at stage 3, but higher in whole body of larvae fed rotifers/Artemia than in those fed zooplankton at stage 5 (p < 0.05). This was followed by a lower ratio of phospholipids to total lipid in cod larvae fed rotifers/Artemia in stage 5 (p < 0.01).

The differences in nutrient composition of the live feed in Exp-2 resembled those observed in Exp-1, where rotifers contained significant lower concentration of protein (p < 0.001), taurine (p < 0.001), iodine (p < 0.05) and zinc (p < 0.001) than copepods (Table 4B). The small copepod fraction is dominated by nauplii, the large copepod by juvenile and adult copepods (Table 5). In this experiment, the selenium level in rotifers was also lower than in copepods (p < 0.001), but above the requirement in fish of 0.3 mg kg⁻¹ DM (NRC, 2011). There were also differences in fatty acid composition, where ARA was again higher and EPA lower in rotifers (p < 0.01). DHA was similar between the feeds. Copepods had a higher ratio of phospholipids to total lipids than rotifers (p < 0.01), corresponding to a lower total lipid level (p < 0.001). Vitamin A was not detected in any of the feed organisms, while vitamin D was absent in copepods. Vitamin E (p < 0.01) and thiamine (p < 0.05) were higher in rotifers than in copepods, while vitamin C, Mn, and Cu levels in the two feeds were similar. Only one replicate sample of the Artemia used in Exp-2 was analysed, so statistical treatment was not possible. However, the Artemia seemed to have more protein and taurine than in Exp-1, but still less than the copepods. The Artemia was also low in thiamine, vitamin A was not detected, while the content of other vitamins was similar to that in the rotifers. The analyses in Exp-2 indicate that iodine, manganese, and zinc levels were lower in Artemia than in copepods. The fatty acid composition was characterized by lower DHA and EPA levels in Artemia, but higher ARA levels. Lipid level and ratio of phospholipids to total lipid was similar to that in rotifers, namely with more lipid and a lower proportion of phospholipids to total lipid than copepods (Table 4B). With the exception of a higher Se content and a lower iodine content in large copepods, there were no significant differences in nutrient composition between small and large copepods, but the variation in some of the nutrients was wide (Table 4B).