Introduction
The European lobster (Homarus gammarus, L.) is an economically important decapod crustacean geographically ranging from Northern Norway to Morocco and the Eastern Mediterranean(Reference Triantafyllidis, Apostolidis and Katsares1). The increased fishing pressure from 1930 to 1970 led to the collapse of several European lobster populations throughout Europe(Reference Agnalt, Van der Meeren, Jørstad, Howell, Moksness and Svåsand2). To counteract the decline, multiple experimental stock enhancement programmes have been launched to release the large numbers of juveniles to lobster grounds in France(Reference Latrouite and Lorec3), the United Kingdom(Reference Cook4), Ireland(Reference Browne and Mercer5), Norway(Reference Agnalt, Leber, Kitada and Blankenship6) and Germany(Reference Schmalenbach, Mehrtens and Janke7). Commercial aquaculture is gaining interest as an additional approach. However, in comparison with other aquaculture sectors, lobster farming still faces several constraints for their commercial production to be economically sustainable. Among them are the high mortality rates, energy and feeding costs(Reference Powell8). Significant advances in automated rearing systems and water quality have been made in recent years that improve European lobster hatchery production and cost-effectiveness(Reference Daniels, Wills and Ruiz-Perez9–Reference Middlemiss, Daniels and Urbina13). Nevertheless, the lack of a balanced artificial feed that is affordable, with better nutritional value, and easier to handle and store compared to live, fresh or frozen foods remains a major drawback in the sustainable farming of H. gammarus (Reference Powell, Hinchcliffe and Sundell14,Reference Gamble, Pirozzi and Hall15) .
Fresh, live mussels(Reference Floreto, Bayer and Brown16,Reference Crear and Forteath17) and frozen krill(Reference Goncalves, Lund and Gesto18,Reference Burton19) are commonly used as preferred reference diets, as they appear to perform consistently well as food for several lobster species. However, the future success of lobster aquaculture will depend on the ability to replace live, fresh or frozen diets by more practical and cost-effective artificial feeds. The transition to specifically formulated, dry pelleted feeds has been applied with relative good success to several crustacean species within the Penaeidae(Reference Wickins and Lee20). While considerable research efforts in the last three decades have been devoted to spiny lobster species (Palinuridae), the transition to extruded feeds has proven to be challenging(Reference Williams21,Reference Perera and Simon22) . Results indicate that protein is the most important macronutrient for optimal growth. Carbohydrates are readily digested and absorbed but their subsequent fate is poorly understood, while lipids appear to be poorly digested, absorbed and used. These findings are in agreement with the chemical composition of the preferred prey of several spiny lobster species which presents high protein, low lipid and moderate to high carbohydrate content(Reference Williams23).
The feasibility of using artificial feeds for the production of juvenile American lobster (Homarus americanus) was extensively investigated(Reference Floreto, Bayer and Brown16,Reference Castell and Budson24–Reference Tlusty, Fiore and Goldstein27) , but growth obtained was not as high as animals that are fed live/fresh food. The efforts on the development of artificial feeds for the European lobster have been less and only a few published studies refer to the potential application of formulated feeds for juvenile H. gammarus farming(Reference Goncalves, Lund and Gesto18,Reference Hinchcliffe, Powell and Langeland28) . Reported results are similar to what has been observed for the H. americanus, i.e., lower growth performance in lobsters fed the experimental artificial diets. The relatively limited success of formulated feeds for lobster juveniles can be partially explained by the lack of information on how lobsters digest and assimilate formulated feeds(Reference Perera and Simon22). Additionally, the higher growth and feeding rates of lobsters fed natural diets has been generally related to: (a) their superior water stability and attractiveness and (b) a nutrient profile approximating closely to their body composition(Reference Tacon29).
Most nutritional studies performed with crustacean species address weight and length gains as growth performance indicators. However, maximum size per se does not necessarily imply the best condition. This is particularly relevant in decapod crustacean species since their body weight is highly influenced by the water content and moulting stage(Reference Albalat, Johnson and Coates30). The use of biochemical indicators can aid in a more comprehensive evaluation of the nutritional status of the reared animals. The ratio of RNA:DNA concentrations has been previously used as an indicator of recent growth and nutritional condition in a variety of crustacean species including the H. gammarus (Reference Yun, Shahkar and Katya31–Reference Schoo, Aberle and Malzahn34). The main premises are that the amount of RNA in a cell varies in proportion to protein synthesis, while the amount of DNA remains fairly constant(Reference Chícharo and Chícharo35). Thus, theoretically, poor nutritional condition contributes to low protein synthesis and slow growth, resulting in a low RNA:DNA ratio(Reference Labh, Prasad and Ranjan36).
Increasing the knowledge on the activity of digestive enzymes of European lobsters is essential to establish a better understanding of their digestive ability. Analysing the digestive response might help revealing why some diets show better performance than others might. There are few published studies referring to the digestive capacity of H. gammarus. A range of carbohydrases(Reference Glass and Stark37) and proteases(Reference Glass and Stark38) have been detected in the hepatopancreas and gastric juice of wild-caught adults. However, changes in the digestive enzyme activity with the type of food or nutritional composition have not yet been determined in European lobster. To the best of our knowledge, only one such study on lobsters has been published(Reference Simon39) but on a different species. In this cited work, the rearing of juvenile spiny lobsters (Jasus edwardsii) on a formulated diet for 6 months resulted in a pronounced decrease in the enzymatic activity (protease, trypsin, α-amylase and α-glucosidase) of the foregut and digestive gland compared with juveniles reared on fresh mussels.
In the present study, we aimed to evaluate the effect of protein levels and the ratios between lipid and carbohydrates in formulated extruded feeds in terms of survival, growth, feed utilisation, nutritional condition and enzymatic activity to explore the potential nutritional requirements of European lobster juveniles. The experimental extruded feeds were tested against a conventional control diet (Antarctic krill, Euphausia superba).
Material and methods
Experimental animals
Juveniles of H. gammarus were obtained from a wild European lobster female caught in the Skagerrak coast of North Jutland, Denmark. Animals used in this experiment were procured from the same female to reduce phenotypic variation. The fertility/fecundity of the chosen female, as well as the survival and apparent quality of her progeny, was observed to be as usual in our facilities. From hatching and until settling, larvae were reared in 500-l rectangular tanks installed at the North Sea Oceanarium, Hirtshals. Each tank was equipped with a top seawater inflow (50 l/h) and a vertical outflow filter (1 mm mesh diameter). Temperature and salinity were kept at 18 ± 1°C and 34 ± 1 PSU, respectively. Strong aeration was provided from the bottom of the tank using airstones to maintain larvae in the water column. Larvae were stocked sequentially into tanks over three consecutive days after hatching at an initial density of 5–7 larvae/l and daily fed thawed Antarctic krill, Euphausia superba (Akudim A/S, Denmark). The bottom of the tanks was syphoned every week to remove debris. Upon metamorphosis to stage IV(Reference Rötzer and Haug40), post-larvae were transferred to the aquaculture facilities at the Technical University of Denmark, Section for Aquaculture, Hirtshals. Animals were individually reared in 3D printed polylactic acid (PLA) bioplastic cassette systems of 200 ml compartments with perforated grids for water in- and outlet. Cassettes were distributed and immersed in raceway tanks of 80-l capacity supplied by the recirculation seawater system at a constant flow rate of 330 l/h (18 ± 0⋅5°C temperature, 34 ± 1 PSU salinity, >90 % dissolved oxygen, <0⋅1 mg/l ammonia-N) subjected to a photoperiod cycle of 8 h light : 16 h dark. Lobsters were adapted to the trial conditions for 1 month, during which they were daily fed thawed Antarctic krill.
Experimental extruded feeds and control diet
Details of the experimental feeds and control diet are reported in Ref. (Reference Goncalves, Lund and Gesto18). Briefly, experimental extruded feeds were formulated to contain 400 or 500 g/kg protein, with lipid contents ranging from 86 to 233 g/kg and carbohydrates from 210 to 347 g/kg resulting in three ratios of lipid to carbohydrate: low (1:3), medium (1:2) and high (approx. 1:1). This resulted in six experimental feeds referred to as 400LOW, 400MED, 400HIGH, 500LOW, 500MED and 500HIGH. The different protein, carbohydrate and lipid contents were achieved by altering squid meal, wheat gluten, wheat starch and fish oil inclusion levels. Experimental feeds were extruded as 4 mm pellets and were manufactured by SPAROS, Lda. (Olhão, Portugal). Thawed Antarctic krill was used as control feed (CTRL). Proximate analysis of the Antarctic krill and extruded feeds were performed in duplicate. Feeds and krill were finely ground using a Krups Speedy Pro homogenizer and analysed for crude protein (i.e. Kjeldahl N × 6⋅25(41)), total lipid(Reference Bligh and Dyer42), dry matter (DM) and ash(43). Likewise, amino acid analyses were performed in duplicate by use of hydrolysed feed samples(Reference Rutherfurd44). The amino acid content was determined by HPLC(Reference Larsen, Dalsgaard and Pedersen45). The ingredients and proximate chemical composition of the experimental extruded feeds and control diet are shown in Table 1. The amino acid profiles of the extruded feeds and control diet are listed in Table 2.
Formulation and chemical composition of experimental extruded feeds and Antarctic krill (adapted from(Reference Goncalves, Lund and Gesto18))
a Micronorse: 70⋅9 % CP, 8⋅7 % CF, Tromsø Fiskeindustri AS, Norway.
b Squid meal: 83 % CP, 4 % CF, Sopropêche, France.
c Krill meal: 61⋅1 % CP, 17⋅4 % CF, Aker Biomarine, Norway.
d VITAL: 80⋅4 % CP, 5⋅6 % CF, Roquette, France.
e Wheat meal: 11⋅7 % CP, 1⋅6 % CF, Molisur, Spain.
f Meritena 200: 0⋅4 % CP, 0⋅1 % CF, 90 % starch, Tereos, France.
g Fish oil: 98⋅1 % CF, 16 % EPA, 12 % DHA, Sopropêche, France.
h P700IPM, Lecico GmbH, Germany.
i Vitamins (IU or mg/kg diet): DL-alpha tocopherol acetate, 200 mg; sodium menadione bisulphate, 50 mg; retinyl acetate, 40 000 IU; DL-cholecalciferol, 4000 IU; thiamine, 60 mg; riboflavin, 60 mg; pyridoxine, 40 mg; cyanocobalamin, 0⋅2 mg; nicotinic acid, 400 mg; folic acid, 30 mg; ascorbic acid, 1000 mg; inositol, 1000 mg; biotin, 6 mg; calcium pantothenate, 200 mg; choline chloride, 2000 mg, betaine, 1000 mg. Minerals (g or mg/kg diet): copper sulphate, 18 mg; ferric sulphate, 12 mg; potassium iodide, 1 mg; manganese oxide, 20 mg; sodium selenite, 0⋅02 mg; zinc sulphate, 15 mg; sodium chloride, 800 mg; excipient wheat gluten, Premix Lda., Portugal.
j Carophyll Pink 10 % CWS, 10 % astaxanthin, DSM Nutritional Products, Switzerland.
x Estimated by difference: Carbohydrates (%) = 100 − (Crude protein + Total lipid + Ash).
y Gross energy (MJ/kg) = Protein content × 21⋅3 kJ/g + Lipid content × 39⋅5 kJ/g + Carbohydrate content × 17⋅6 kJ/g)/1000 kJ/MJ(Reference Cuzon, Guillaume, D'Abramo, Conklin and Akiyama46).
Amino acid profile (g per 100 g as is) of experimental extruded feeds and Antarctic krill
Growth trial
Prior to the start of the experiment, 24 h unfed lobster juveniles (stage V to VI) were individually weighed and measured (carapace length). Homogeneous groups of 30 individuals (initial weight of 90⋅0 ± 2⋅8 mg per lobster; carapace length of 6⋅5 ± 0⋅6 mm, mean (sd)) were randomly allocated to each dietary treatment. As each lobster was held separately, the experimental unit in the present study was the individual.
Individual lobsters were hand-fed in excess a pre-weighed amount of thawed krill or extruded pellet each morning, and allowed to feed for 4 h. Thus, all lobsters had equal access to feed for a limited amount of time per day. Additionally, lobsters were permitted to consume their shedded exoskeletons. After each meal, the uneaten feed was siphoned off and stored in 50 ml tubes at −20°C. To estimate feed intake (FI), each treatment was divided into subgroups of ten individuals. The uneaten feed fraction of each subgroup (minus sampled or dead animals, when applicable) was collected daily, accumulated over 6 weeks and stored in a single tube, allowing triplicate values to be obtained for FI estimation. At the end of the trial, each tube content was filtered, dried (24 h at 60°C) and weighed for FI estimation using the following formula(Reference Nguyen, Chim and Lemaire47):
where dF is distributed feed, uF is unconsumed feed, L is leaching after 4 h, BWi is the initial body weight and Δt is the number of days during which uneaten feed was collected (42 d). Leaching was estimated by placing a pre-weighed quantity of each diet in the cassette compartments under the same conditions as in the trial but, in this case, without animals. FI was expressed in the percent of DM ingested per initial body weight per day. Energy intake was calculated by multiplying the daily individual intake (mg) by the gross energy content of each diet.
The presence of shedded exoskeletons and mortality were recorded daily. The experiment was conducted for 8 weeks and lobsters were individually weighed and measured every second week. Significant mortality was gradually observed in the group of animals fed the 400HIGH diet. Consequently, at week 6, samples for enzyme activity (nine individual lobsters per treatment) were collected and the growth trial ended for the 400HIGH group. The growth trial continued for two more weeks for the other groups. Body wet weight was recorded to the nearest 0⋅001 g after gently blotted dry each individual lobster with a paper towel. Carapace length was recorded to the nearest 0⋅1 mm with a Vernier calliper from the base of the eye socket to the posterior edge of the cephalothorax. The following formulas were used to determine growth performance:
where FCR is the feed conversion ratio; FI is feed intake (dry weight, mg); BG is the biomass gain (wet weight, mg).
where SGR is the specific growth rate; BWf is the final wet body weight (at week 6); BWi is the initial wet body weight and Δt is the number of growth trial days (42) considered for SGR estimation.
where iCL is the increment in carapace length; CLf is the final carapace length (at 42nd day) and CLi is the initial carapace length.
Juvenile lobsters proximate chemical composition
At the end of the trial, lobsters were lethally anaesthetised in ice-cold seawater, weighed, measured, rinsed in distilled water and stored at −80°C for proximate chemical composition analysis. Proximate analysis of the juveniles was performed in analytical triplicates per dietary treatment. Briefly, the pool of six to twelve individuals per treatment were finely ground using a Krups Speedy Pro homogenizer and analysed for DM and ash(43). Protein was determined spectrophotometrically at 750 nm using a commercial Lowry-based, micro-protein determination kit (BIO-RAD 500-0112). Lipids were extracted with chloroform–methanol (2:1 by volume) according to the Folch method(Reference Christie and Han48) and lipid content determined gravimetrically.
Nucleic acid determinations
At week 4, six juveniles per dietary treatment were collected for nucleic acid analysis 24 h after being fed. Lobsters were lethally anaesthetized in ice-cold seawater, weighed, measured and rinsed in distilled water before being frozen at −80°C. Lobsters were freeze-dried and weighed (±1 μg dry weight (DW)) on an electronic microbalance (Sartorius M5P). The concentration of nucleic acids (RNA, DNA) was quantified in approximately 1 mg of dry abdominal muscle tissue from each individual lobster following procedures described previously(Reference Caldarone, Wagner and St Onge-Burns49). Briefly, the muscle samples were chemically (cold sarcosyl Tris-EDTA extraction buffer) and mechanically homogenised in an ultrasonic homogeniser unit (4710 Series, Cole Parmer Instruments Co.). The homogenate was centrifuged for 15 min at 1200g, 4°C, and the supernatant extract was used for the analysis of RNA and DNA. The supernatants (30 μl) were placed into fluorescent microplate wells with Tris buffer (140 μl). Finally, a specific nucleic acid fluorochrome dye GelRED (30 μl) was added into each well for the fluorescent reading of nucleic acids. Fluorescence was measured on a microplate reader (Biotek Synergy HT model SIAFRTD – BioTek Instruments, Inc., Vermont, USA) with an excitation wavelength of 365 nm and an emission wavelength of 590 nm. Following the first scan to determine the total fluorescence of RNA and DNA, a ribonuclease A (Type-II A) solution (30 μl) was added to each well and activated by incubating the fluorescent plates at 37°C for 30 min. The plates were read again, and the concentration of DNA was calculated directly from the standard curve described later. The RNA fluorescence was calculated subtracting the DNA fluorescence (second scan) from total fluorescence (first scan). Concentrations were determined by running standard curves of DNA-GelRED and RNA-GelRED with known concentrations of ƛ-phagus DNA (0⋅25 μg/μl) and 16S-23S E. coli RNA (4 μg/μl) (Roche). The average ratio of DNA and RNA slopes (average and sd) was 5⋅85 ± 0⋅01. The RNA/DNA ratios were standardised (sRD) using DNA and RNA slope ratios and the reference slope ratio of 2⋅4(Reference Caldarone, Clemmesen and Berdalet50). The physiological condition of the juveniles was estimated with nucleic acid derived indices: RNA (μg/mg DW), DNA (μg/mg DW) and sRD (standardised RNA/DNA ratios).
Digestive enzyme activities
At week 6, nine lobsters per treatment were collected between 4 h to 5 h after being fed for the analysis of digestive enzyme activities. Individuals were rinsed in distilled water and stored at −80°C. To prepare the enzyme extracts, juveniles were previously freeze-dried. Rostrum, chelipeds, legs, pleopods, uropods and telson were removed and the abdominal section of each individual was mechanically homogenised in 1 ml distilled water. The homogenate was centrifuged for 10 min at 15 800g, 4°C, and the supernatant extract was used for the analysis of trypsin, amylase and lipase activities. All samples were kept in ice all times to avoid enzymes denaturation or damage. Enzyme extract aliquots were stored at −80°C until analysis.
For trypsin analysis, the fluorogenic substrate Boc-Gln-Ala-Arg-7-methylcoumarin hydrochloride (BOC-SIGMA B4153) was diluted in dimethyl sulfoxide (DMSO), to a final concentration of 20 μM. For analysis, 5 μl of this substrate, 190 μl of 50 mm Tris + 10 mm CaCl2 buffer (pH 8⋅5) and 15 μl of the enzyme extract were added to the microplate(Reference Rotllant, Moyano and Andrés51). Fluorescence was measured at 355 nm (excitation) and 460 nm (emission).
Ultra Amylase Assay Kit (E33651) from Thermo Scientific was used for amylase analysis. The kit contains a starch derivate labelled with a fluorophore dye as a substrate. This substrate was diluted in 3-(N-morpholino) propane sulfonic acid (MOPS; pH 6⋅9) and substrate solvent (sodium acetate; pH 4⋅0), to a final concentration of 200 μg/ml. For analysis, 50 μl of the substrate solution and 15 μl of the enzyme extract were added to the microplate. Fluorescence was measured at 485 nm (excitation) and 538 nm (emission).
Lipase activity was assayed using 4-methylumbelliferyl heptanoate (M2514, Sigma-Aldrich). The substrate was dissolved in phosphate buffer pH 7⋅0 to a final concentration of 0⋅4 mm, modified method from Ref. (Reference Rotllant, Moyano and Andrés51). Fifteen microlitres of the lobster extract was added to the microplate and mixed with 250 μl of 0⋅4 mm substrate for the analysis. Fluorescence was measured at 355 nm (excitation) and 460 nm (emission).
All enzyme activities were expressed as RFU (Relative Fluorescence Units) per mg lobster dry weight before dissection.
Statistical analyses
The amino acid profiles of control diet and experimental feeds were analysed by principal component analysis performed using R version 3.5.1 software(Reference Core Team52) and the factoextra version 1.0.7 package(Reference Kasambara and Mundt53). Survival and moult occurrence data were analysed by a Kaplan–Meier procedure. Log-rank (Mantel–Cox) test was used to determine significance (P < 0⋅05). Whenever significance was detected, a χ 2 table with multiple comparisons was generated to identify differences among dietary treatments. All other parameters shown are expressed as means (sem) unless otherwise specified. Before analyses, the ANOVA assumptions of normality of residuals and homogeneity of variances were tested using the Shapiro–Wilk and Levene's test, respectively. In instances where assumptions were not met, data were square root or log-transformed. A second-order polynomial regression with time was used as an estimate of the average growth, both in terms of carapace length and body weight. The models were validated via residual plot analysis and generated using R version 3.5.1 software(Reference Core Team52). Curve coefficients were compared in a one-way ANOVA and whenever significant differences were detected, the coefficients from experimental feed curves were tested against the CTRL using the Dunnett t-test. For the remaining estimated growth parameters, RNA and DNA concentrations, sRD and digestive enzyme activities, the dietary treatments were subjected to a one-way ANOVA to test the experimental formulated dry feeds against the CTRL diet. Whenever significant differences were identified, comparisons against the CTRL diet were conducted using the Dunnett t-test. Data from experimental extruded feed treatments were subsequently subjected to a two-way ANOVA, considering the protein level and the L:CHO ratio as explanatory variables. Following the two-way ANOVA and, whenever significant differences were identified, means were compared by the Holm-Sidak post hoc test. A standardised RNA/DNA ratio was expected to reflect changes in growth performance parameters(Reference Grimm, Lehmann and Clemmesen54). Thus, a simple linear regression was calculated to predict sRD based on SGR. Apart from the principal component analysis and the second-order polynomial regression for carapace length and body weight, all statistical tests were performed using the IBM SPSS Statistics 25.0 and graphics were generated by GraphPad Prism version 5.0 software package.
Results
Diets amino acid composition
The CTRL and 500MED feed contained higher levels of both essential (EAA) and nonessential (NEAA) amino acids, while the 400MED and 400HIGH contained the lowest levels for EAA and NEAA (Table 2). A principal component analysis (Fig. 1) revealed that the 500LOW and 500MED were more similar to the CTRL than any other experimental feeds. The majority of the variance was explained by PC1 (84⋅1 %) which mostly identified differences related to the concentration of the total amino acids that was lower in all the three 400 and in the 500HIGH compared with the other groups. PC2 explained 14⋅3 % of the variability and showed that methionine and taurine were the amino acids with higher contribution for differences between the CTRL and the extruded 500MED and 500LOW feeds.
Graphical representation of principal components analysis of amino acid profiles from tested diets. The two panels are complementary to one another. (a) Biplot of the first two principal components (PC1 and PC2) of tested diets amino acid profiles. PC1 separated the amino acid profiles horizontally and explained 84⋅1 % of the variance. The amino acid profiles of 500MED, 500LOW and CTRL diets formed a succinct group to the right of the plot and were positively correlated to PC1. The 400HIGH, 400MED, 400LOW and 500HIGH diets were negatively correlated to PC1. PC2 separated the amino acid profiles vertically and explained 14⋅3 % of the variance. The 500LOW and 500MED diets were positively correlated, while the CTRL was negatively correlated to PC2. (b) Variables (amino acids) used to construct the principal components. The circle in this plot is the correlation circle, the stronger the correlation of an amino acid to PC1 or PC2 the closer its arrowhead to the circle. The arrows indicate how the amino acids contributed to the formation of PC1 and PC2 and thus the formation of plot (a).
Survival, growth and feed utilisation
Survival was not significantly affected by dietary treatment (χ 2 8⋅2, df 6, P = 0⋅22) varying between 65 and 90 % (Fig. 2). Although no statistically significant difference was detected in the survival rate of the 400HIGH group, the projected curve indicates a marked divergence from the other dietary groups. In order to secure enough sample size for subsequent analysis, we decided to end the growth trial for this treatment 2 weeks earlier. Therefore, despite the survival curves were calculated for an 8-week period in all treatments other than the 400HIGH, their comparison was performed only for the first 6-week period (Fig. 2).
Survival of H. gammarus juveniles (% of initial numbers) fed on the different diets. The dashed vertical line indicates the rearing time limit (42 d) considered in the statistical comparison of the curves.
As for survival, moulting curves comparison was restricted to the first 6 weeks. Moulting occurrence was significantly affected by dietary treatment (Fig. 3) and the rate of a first moult varied between 62 and 100 % among diets. Mean cumulative first moult was significantly higher for lobsters offered the 500MED feed compared with juveniles fed the 400HIGH, 400MED and 500LOW feeds (χ 2 14⋅4, df 6, P = 0⋅03). Juvenile lobsters offered the 400HIGH diet did not succeed a second moult. In the remaining groups, the occurrence rate of the second moult varied between 21 and 60 %. Mean cumulative successful second moult was significantly higher for the 500MED than the other groups, except for CTRL and 400LOW (χ 2 32⋅3, df 6, P < 0⋅001).
Cumulative moults of H. gammarus juveniles (% of initial numbers) fed on different diets. Panel (a) refers to the first moult occurred after the beginning of the growth trial and panel (b) refers to the second moult occurrence. Dietary treatments are indicated in the legend in ascending order of cumulative moults. The dashed vertical line indicates the rearing time limit (42 d) considered in the statistical comparison of the curves. Different superscript letters indicate significant differences between dietary treatments.
Changes in weight and carapace length throughout the trial were evaluated by second-order polynomial regression with time. Regression curves are represented for all dietary treatments shown in Fig. 4. However, because the trial ended at week 6 for the 400HIGH treatment, the curve coefficients from this group were not statistically compared with the CTRL. During the 8-week experimental period, lobsters grew from an initial mean weight of 90 mg (6⋅5 mm carapace length) to mean weights ranging from 134 to 254 mg (7⋅9–9⋅5 mm carapace length). The juveniles fed the 500MED and CTRL diets sustained an increase in carapace length and body weight over the 8-week growth trial. The remaining groups reached a growth plateau, both in terms of length and weight, after 6 weeks (Fig. 4). The 400MED and 500LOW experimental feeds had a significant lower carapace length increase over time compared with the CTRL (Fig. 4(a) and 4(b)). In terms of body weight gain over the trial period, all the experimental dietary groups except the 500MED and 400LOW showed a significantly slower growth compared with the reference group (CTRL) (Fig. 4(c) and 4(d)).
Second-order polynomial model fit to average carapace length (upper panels (a) and (b)) and wet body weight (lower panels (c) and (d)) of H. gammarus juveniles fed on different diets throughout the growth trial period. Data points represented as mean (sem). Dietary treatments that were significantly different from CTRL are marked with an asterisk (*). # indicates that curve coefficients from the 400HIGH group were not statistically compared with CTRL.
The growth performance indices – SGR and iCL – were determined for a 6-week period in all treatments (Table 3). A significantly higher SGR was observed for krill-fed lobsters in comparison with lobsters fed the extruded feeds, except for the 500MED diet (Table 3). Within the extruded feeds, SGR was significantly affected by the protein level, the L:CHO ratio and the interaction of protein × L:CHO ratio. Results showed that for the 400-protein group, juveniles fed the 400LOW diet grew faster than the ones fed the 400MED and 400HIGH feeds. Within the 500-protein level, a significantly faster growth was observed in the 500MED treatment. The protein content did not affect the SGR in the LOW L:CHO ratio category, while in the MED and HIGH groups, juveniles fed the 500 g/kg protein feeds grew significantly faster (Table 3). Juveniles fed the 400MED, 400HIGH and 500HIGH feeds had a significantly lower iCL than the CTRL group. The protein content, L:CHO ratio and protein × L:CHO all had a significant effect on the iCL. Among the 400-protein level, the highest iCL was achieved for the 400LOW and the lowest for the 400HIGH group. Within the 500 level, the highest iCL was observed for the 500MED and the lowest for the 500HIGH treatment. The protein content only had a significant effect on iCL at the intermediate level of L:CHO for which the 500MED were larger than the 400MED group (Table 3).
Growth performance, feed efficiency and whole body composition of juvenile H. gammarus fed the experimental extruded feeds and Antarctic krill
SGR, specific growth rate; iCL, increment in carapace length; FI, dry feed intake; FCR, dry feed intake/wet weight gain; EI, energy intake, N, number of replicates per treatment.
Values are means (sem).
#indicates dietary groups significantly different from CTRL (Krill).
Means in the same row with a different superscript ‘a, b, c’ or ‘A, B, C’ are significantly different within the 400- or 500-protein level, respectively.
Means in the same row with a different superscript ‘x or y’ are significantly different within the same L:CHO ratio category.
† Estimated by difference: Carbohydrates (%) = 100 − (Soluble protein + Total lipid + Ash).
* P < 0⋅05, **P < 0⋅01, ***P < 0⋅001.
As for SGR and iCL, the feed efficiency indices – energy intake, FI and FCR – were calculated for the first 6-week period in all treatments (Table 3). FI was significantly higher for all experimental feeds compared with the CTRL except for the 400HIGH. Within the experimental feeds, the L:CHO ratio and the interaction of protein × L:CHO had a significant effect on the FI. A higher FI in extruded feeds with a higher carbohydrate content was observed for both protein levels examined (400 and 500 mg/kg) with most pronounced differences in feeds with a lower protein content. As such, lobsters fed LOW feeds presented higher FI than HIGH with intermediate levels observed in the MED treatments (Table 3). Intake expressed in terms of energy followed the same trend as FI. The FCR was significantly lower for the CTRL diet compared with the experimental extruded feeds except for the 500MED group. There was a significant effect on FCR from the L:CHO ratio among the experimental dry feeds, for which FCR in juveniles reared on MED feeds was significant lower compared with LOW feeds (Table 3).
Proximate composition
No statistical analysis was performed on the proximate composition data set since only analytical replicates were used in the analysis of a single pooled sample. The DM content varied from 20 to 29 %, with a minimum observed for the 400HIGH and maximum for the CTRL group. The inverse trend was detected for the ash content, with a minimum of 32 % observed for the CTRL and a maximum of 43 % for the 400HIGH treatment. The protein fluctuated between 41 and 50 %, lipids between 4 and 7 % and carbohydrates between 9 and 18 %. Juveniles reared on the 500LOW experimental diet had the highest lipid and the lowest protein content, while the lobsters fed the 400LOW diet had the highest protein and lowest carbohydrate contents. Minimum lipid content was observed for the 400HIGH group and maximum carbohydrate content for the 500MED group (Table 3).
Nucleic acid derived indices
The changes in the weight-specific RNA and DNA concentrations and sRD are shown in Fig. 5. The RNA content was significantly lower (F 6,41 5⋅61, P < 0⋅001) in animals fed the 400MED, 400HIGH and 500HIGH feeds when compared with the krill-fed lobsters. Within the experimental feeds, the RNA content was affected by the L:CHO ratio (F 2,35 7⋅61, P < 0⋅002). Feeds in the HIGH category triggered a significant reduction in the RNA content of the abdominal muscle (Fig. 5(a)).
Changes in RNA concentration (a), DNA concentration (b) and standardised RNA/DNA ratio, sRD (c) of abdominal muscle tissue of H. gammarus juveniles fed on different diets (n 6). The box includes observations from the 25th to 75th percentiles, and the whiskers above and below the box indicate the 10th and 90th percentiles. The horizontal line within the box represents the median value and the symbol (+) indicates the mean. Dietary treatments that were significantly different from CTRL are marked with an asterisk (*). Different letters ‘a, b’ or ‘A, B’ indicate significant differences within the 400- or 500-protein level, respectively. Different ‘x or y’ indicate significant differences within the same L:CHO ratio category.
In comparison with the CTRL group, the 400HIGH treatment presented a significant increase (F 6,41 5⋅08, P = 0⋅001) in DNA content. Among the experimental extruded feeds, both protein (F 1,35 10⋅41, P = 0⋅003) and L:CHO ratio (F 2,35 3⋅75, P = 0⋅035) significantly affected the DNA concentration in the muscle. Feeds within the 400-protein group supported a significant increase in DNA content compared with the 500-protein group. The DNA content of the abdominal muscle of juveniles reared on the L:CHO MED category feeds was significantly lower compared with the HIGH groups (Fig. 5(b)).
The standardised ratio (sRD) was significantly affected by dietary treatment. The sRD was significantly lower (F 6,41 7⋅87, P < 0⋅001) in animals fed the 400MED, 400HIGH and 500HIGH feeds when compared with the CTRL. Within the experimental feeds, sRD was significantly affected by protein (F 1,35 7⋅63, P = 0⋅010), L:CHO ratio (F 2,35 7⋅09, P = 0⋅003) and the interaction of protein × L:CHO (F 2,35 5⋅83, P = 0⋅007). For the 400-protein level, juveniles fed the 400MED and 400HIGH feeds showed a significantly lower ratio compared with the lobsters fed the 400LOW diet. For the 500 level, the 500MED showed higher sRD than the 500HIGH group. Protein level had a significant effect within the MED L:CHO category, for which the 500 resulted in a higher sRD than the 400 group (Fig. 5(c)). A significant regression equation was found between sRD and SGR (F 1,40 49⋅05, P < 0⋅001), with an R 2 of 0⋅55. The predicted sRD for juvenile lobsters was equal to − 0.143 + 0.644 (SGR) when SGR is estimated in % d−1 (Fig. 6).
Fitted linear regression model of standardised RNA/DNA ratio (sRD) vs. specific growth rate (SGR). Data points per dietary treatment represented as mean (sem). The solid line displays the average estimates of the predicted sRD and the dashed lines display the 95 % confidence limits. Simple linear regression model: y = 0.644x − 0.143 (R 2 = 0.55).
Digestive enzyme activities
Trypsin activity was significantly higher for the animals fed the experimental feeds than for the CTRL group, except for the 400LOW treatment (F 6,62 8⋅37, P < 0⋅001). Within the experimental feeds, trypsin activity was affected by the L:CHO ratio (F 2,53 9⋅86, P < 0⋅001) and the interaction of protein × L:CHO (F 2,53 6⋅49, P = 0⋅003). No differences were observed among the three feeds in the 500-protein level, but in the 400 group, the 400HIGH and 400MED treatments promoted a significant increase in trypsin activity compared with the 400LOW. The dietary protein content significantly affected the trypsin activity for the LOW (higher for the 500 level) and HIGH (higher for the 400 level) categories (Fig. 7(a)).
Changes in trypsin (a), amylase (b) and lipase (c) activity of H. gammarus juveniles fed on the different tested diets (n 9). The box includes observations from the 25th to 75th percentiles, and the whiskers above and below the box indicate the 10th and 90th percentiles. The horizontal line within the box represents the median value and the symbol (+) indicates the mean. Dietary treatments that were significantly different from CTRL are marked with an asterisk (*). Different letters ‘a, b’ or ‘A, B’ indicate significant differences within the 400- or 500-protein level, respectively. Different ‘x or y’ indicate significant differences within the same L:CHO ratio category.
As for trypsin, a similar trend was observed for the activity of amylase. When compared with the CTRL, amylase activity was significantly higher in the 400MED and 400HIGH treatments (F 6,62 5⋅00, P < 0⋅001). Protein (F 1,53 28⋅91, P < 0⋅001), L:CHO ratio (F 2,53 12⋅79, P < 0⋅001) and the interaction of both factors (F 2,53 4⋅97, P = 0⋅011) had a significant effect on amylase activity. No significant differences were observed within the 500-protein level. For the 400-protein group, amylase activity was significantly higher in juveniles fed the 400MED and 400HIGH feeds. Within the MED and HIGH categories, amylase activity increased for the 400-protein level (Fig. 7(b)).
Juveniles fed the 500LOW and 500MED feeds had a higher lipase activity in comparison with juveniles fed the CTRL diet (F 6,62 8⋅33, P < 0⋅001). Within the experimental dietary groups, both the protein level (F 1,53 23⋅36, P < 0⋅001) and L:CHO ratio (F 2,53 8⋅39, P = 0⋅001) significantly affected the lipase activity. The activity of this enzyme was significantly higher for the 500-protein content feeds. Within the L:CHO ratio, juveniles fed the LOW feeds had a significantly higher lipase activity than the ones fed the MED and HIGH feeds (Fig. 7(c)).
Discussion
Growth performance
Results obtained indicate that the overall performance of European lobster juveniles reared on the 500MED extruded feed for an 8-week period was comparable to the performance of juveniles fed thawed Antarctic krill (CTRL). Previous nutritional studies carried out on H. gammarus (Reference Hinchcliffe, Powell and Langeland28), H. americanus (Reference Tlusty, Fiore and Goldstein27) and several Palinuridae species(Reference Perera and Simon22) showed that the growth rate was generally poorer in lobsters fed formulated compared with fresh diets. A recent study(Reference Goncalves, Lund and Gesto18) investigated the effect of the same feeds that are used here on the performance, metabolic rate and nitrogen retention in European lobster juveniles. The results indicated that H. gammarus grew faster when fed the 500 g/kg compared with the 400 g/kg protein feeds. The results also suggested a potential protein sparring capacity in the formulated feeds, when compensated by the inclusion of carbohydrates, given that the growth obtained for the 400LOW was comparable to that of the three 500 feeds. In the present study, we found that the growth rate of juveniles fed the 400LOW, 500LOW and 500HIGH feeds was lower than previously projected. This is most likely due to the too short growth trial period in the previous study. We, thus, confirm what had already been hypothesised in the previous work, i.e., a period of 4 weeks is not long enough to assess the effects of experimental feeds on the growth of juvenile European lobsters. Even the present study is arguably a little too short. As growth in crustaceans, as opposed to fish, rely on sequential moultings(Reference Simon, Fitzgibbon and Battison55), nutritional trials performed in crustacean species aiming growth performance as the primary outcome must consider a minimum duration that allows individuals to complete two moults.
The poorest overall performance observed in the group of lobsters reared on the 400HIGH diet is most likely related to the low FI. The reason for such low intake in this dietary group remains unclear, but it might be associated with a lower palatability, different smell or that the higher lipid content caused faster satiation. Results from a feeding trial carried out with European lobster juveniles of similar size(Reference Mente, Houlihan and Smith56) showed that animals fed 5 % dry feed of body mass grew significantly less than juveniles fed 10 %. In the present study, only lobsters fed the 400LOW diet were close to the 10 % level recommended(Reference Mente, Houlihan and Smith56), but a direct comparison between the current and the mentioned study is not possible since both test diets and feeding duration were different. In the cited study(Reference Mente, Houlihan and Smith56), juveniles were allowed to feed from the afternoon until the next morning, while only 4 h in our set-up. Nonetheless, the fact that low intake compromised growth was validated in both studies. In animals fed the 400HIGH diet, not only growth but also survival was compromised. Results suggest that juveniles were not taking enough feed to support growth, nor even enough to sustain the minimum standard metabolic rate (SMR). In our previous study using the same feeds(Reference Goncalves, Lund and Gesto18), a significant decrease was observed in the SMR of European lobster juveniles reared on the 400MED and 400HIGH feeds for a 32-d period. Another likely reason for such a low performance on this diet may be related to a low tolerance to high lipid diets by crustaceans(Reference Nelson, Bruce, Nichols and Phillips57). It has been previously suggested(Reference Cuzon, Guillaume and Cahu58) that high dietary lipids reduce the capacity for amino acid absorption due to an increase in bile acid concentration in response to a high-fat diet.
Feed efficiency
Although the CTRL group grew faster with the lowest FI among all treatments, the low FI observed for this group is masked by the fact that calculations were performed on a dry weight basis. The CTRL diet composed of frozen Antarctic krill had a much lower DM content (approx. 11 %) than the experimental extruded feeds (DM approx. 90 %) which complicates FI comparisons. Also, the energy intake was lower in lobsters fed with krill compared with the extruded feeds. Moreover, the form in which krill is presented is so different from an extruded feed, which may cause differences in the overall nutrient absorption and digestibility. The highest intake observed for the experimental feeds in the LOW L:CHO category suggests that feeds with the high carbohydrate content are more attractive to European lobster juveniles. This is supported by previous studies showing that the composition of the preferred prey of spiny lobster is generally high in protein content, moderate to high in carbohydrate content and low in lipid content(Reference Williams23). In the present study, we observed that the CTRL group was the most efficient in terms of FCR but highly influenced by the low DM content of the thawed Antarctic krill. When FCR is estimated on a wet weight basis, the trend inverts, for example, FCR increases from 2⋅0 to 2⋅2 and from 0⋅7 to 6⋅0 in the 500MED and CTRL groups, respectively. More meaningful is the comparison of FCR values among the experimental extruded feeds. Results indicate a lower FCR in diet 500MED and points to a better utilisation of this diet. The general best performance of animals fed the 500MED when compared with the other extruded feeds could also be related to a more balanced amino acid composition. This diet, together with the 500LOW, showed the most similarity to the Antarctic krill among the experimental extruded feeds. Still, the apparent lower level of taurine in all the extruded feeds might have limited the growth of juveniles. This amino acid is known to elicit a feed-attractant stimulus in crustaceans(Reference Perera and Simon22). A potential solution to increase the performance of lobsters fed on artificial feeds could be the incorporation of taurine to formulated feeds to increase FI of lobsters(Reference Floreto, Bayer and Brown16). The 500 mg/kg protein level in the experimental feeds was achieved by increasing the proportion of squid meal. This ingredient is often used in crustaceans’ diets to increase growth and attractiveness of artificial feeds and the replacement of fish by squid meal has been proven beneficial for growth in several crustacean species(Reference Mente59). However, since the amino acid profiles of fish and squid meal are similar(Reference Nelson, Bruce, Nichols and Phillips57), it is unlikely that the dissimilarities observed among the experimental extruded feeds were caused by the different inclusion levels of this ingredient.
Nucleic acid derived indices
The decrease in sRD for the 400MED and 500HIGH dietary groups were mainly caused by a reduction in the RNA content, while for the 400HIGH group, results show that the low sRD was a combined effect of decreased RNA and increased DNA concentrations in the muscle tissue.
The quantified RNA includes ribosomal RNA (rRNA), messenger RNA (mRNA) and transfer RNA (tRNA) which respond in different ways and have different functions. However, since rRNA – responsible for building protein – makes up the majority of total RNA, changes in total RNA were assumed to primarily reflect changes in rRNA(Reference Buckley, Caldarone and Ong60). Thus, it was presumed that the reduction in RNA reflected a decrease in protein synthesis, while the relative increase in DNA was associated with an increase in the number of cells per tissue portion, however, of smaller dimensions(Reference Pilar Olivar, Diaz and Alexandra Chícharo61). Thus, the results suggest that juveniles reared on the 400HIGH diet were not only affected by a decreased capacity for protein synthesis but also by the mobilisation of proteins from the abdominal muscle to obtain energy, as previously suggested in starving fish larvae(Reference Catalán, Berdalet and Olivar62). Overall, results showed that the sRD was positively correlated with growth performance and feed efficiency indicators. Specifically, we demonstrated that sRD estimated from samples collected at week 4 were significantly correlated with SGR calculated at week 6. These observations suggest that sRD is a sensitive indicator of the nutritional condition of European lobster juveniles. The use of this estimate in crustaceans is of great applicability as it allows the assessment of recent growth and nutritional condition(Reference Parslow-Williams, Atkinson and Taylor33,Reference Lemos, Garcia-Carreño and Hernández63,Reference Moss64) . Most of the nutritional studies performed on crustaceans considered growth rate as gains in wet weight and length. However, the growth process in crustacean species has an interrupted character because is connected to the moult cycle, whereas somatic growth and the accumulation of energy reserves in tissues are a continuous process(Reference Nguyen, Chim and Lemaire47). We conclude that the nucleic derived indices are a useful tool in future nutritional studies with European lobster juveniles as it allows a faster evaluation when compared with conventional growth performance estimators.
Digestive capacity
The lowest activity levels of trypsin and amylase enzymes were detected in European lobster juveniles reared on Antarctic krill while activities increased in animals offered the 400HIGH and 400MED extruded feeds. This is in contradiction to what was previously reported(Reference Simon39). In that study, was observed that spiny lobsters (Jasus edwardsii) reared on a formulated diet had a marked decrease in trypsin and α-amylase activities of the foregut and digestive gland compared with juveniles fed on fresh mussels. A main difference, however, may be that in the present study, enzymatic activity was estimated from the entire abdominal section and not from the foregut and/or hepatopancreas. Thus, the activity detected in tissue homogenates may not necessarily represent the activity of enzymes that will be secreted into the digestive lumen. In crustaceans, proteases and carbohydrases activity was found in tissues outside the gut(Reference Mattson and Mykles65,Reference O'brien and Skinner66) . It may also be that the secretion of large amounts of enzymes may maximise the use of limiting nutrients. Such elevated enzymatic activity could maximise hydrolysis and the resulting extraction of a dietary substrate that was ingested in small amounts(Reference Lovett and Felder67). Results from(Reference Simon and Jeffs68) point to the same hypothesis. In the cited study, the authors reported a significant increase in the specific activity of α-amylase in J. edwardsii juveniles fed diets with a low starch inclusion level. Thus, based on this, the results suggest that trypsin and amylase activities increased to compensate for the low levels of protein and carbohydrates lobsters could obtain when fed on the 400MED and 400HIGH feeds. For lipase, the activity was lower for juveniles fed the 400HIGH and 400MED than for individuals reared on the 500LOW and 500MED extruded feeds. While carbohydrates are the primary source of energy for crustaceans(Reference Casillas Hernández, Magallón and Portillo69), lipids are the main energy reserve. European lobsters store the metabolised lipids in the R-cells of the hepatopancreas(Reference Barker and Gibson70). Overall, the results point toward a combined effect of low protein, high lipid and low FI in the activity of trypsin and amylase. The increased activity of these enzymes in the group of animals fed the 400MED and 400HIGH feeds suggests the activation of a mechanism to optimise the use of protein and carbohydrates as they seem to be the most important nutrients for European lobster juveniles. In contrast, lipase activity increased with protein level in the diet, with significantly higher activity for the 500-protein lobster groups than the 400-protein groups. This suggests that the required supply of protein for growth is covered for the 500 group and that energy reserves to support growth potential are ensured for the lobsters reared on the 500LOW and 500MED feeds.
Concluding remarks
In the present study, we demonstrated that European lobster juveniles fed the 500MED experimental diet achieved a growth performance and nutritional condition statistically comparable with that observed when animals were fed thawed Antarctic krill. This extruded feed contained 500 g/kg protein, 237 g/kg carbohydrates and 119 g/kg lipids (as fed basis). The results suggest that this composition is the most balanced to meet the overall nutritional requirements of the European lobster juveniles among the tested extruded feeds. Results also support what was previously observed for spiny lobster species: formulated feeds with a very similar L:CHO ratio (1:2) have been proven to provide the best balance of carbohydrate and lipid when rearing juvenile J. edwardsii (Reference Johnston, Calvert and Crear71). Therefore, based on the results presented, we recommend that high protein and moderate inclusion of carbohydrates must be considered when formulating feeds for H. gammarus juveniles. While relative growth benefits of lobsters fed the 500MED experimental diet did not exceed that of lobsters fed the control diet, it is worth considering that replacing fresh diets with extruded feeds can be advantageous in a cost-effective point of view. However, from an economic point of view, the ingredients used in the formulation of experimental extruded feeds were not the most sustainable. Future studies, following some of the recommendations presented here and using sustainable alternative ingredients (e.g. insect meal, industrial by-products), will be of extreme importance for the establishment of European lobster farming on a commercial scale.
Acknowledgements
The authors are grateful to Gelmar Bechara for language editing. The authors also thank Ole Larsen, Rasmus Jensen and Jens Nedergaard for their assistance in taking care of the lobster systems, to Ulla Sproegel and Brian Møller for their support with the laboratory analysis and Tilo Pfalzgraff for his help in sampling the weight and length of the lobsters.
This study was financed by a Fisheries local action group (FiskeriLAG NORD, Denmark), the ENV Fonden (Nord Energi, Denmark) and the AutoHatch project (reference E114944 co-funded by the EU Framework Programme Horizon 2020 (EUROSTARS-2) and the Innovation Fund Denmark) with support from the PhD school at DTU Aqua. The collaboration with CCMAR was supported by a travel grant awarded to Renata Goncalves by the Fondation Idella (Denmark) and partially funded by FCT – Foundation for Science and Technology (Portugal) through project UIDB/04326/2020.
R. G., I. L. and M. G. conceived the research question and designed the study. R. G. carried out the study trial. Samples and data were analysed by R. G. and V. B. Findings were interpreted by R. G., I. L., M. G., C. N. and M. A. T. All authors contributed to the preparation of the manuscript.
There are no conflicts of interest.
Open access
