Effect of Feed Protein and Carbohydrate Levels on Feed Intake, Growth, and Gonad Production of the Sea Urchin, Lytechinus variegatus

Collection, Culture, and Initial Measurements

Lytechinus variegatus were collected in October 2004, from Port St. Joseph Peninsula State Park, Florida (30°N, 85.5°W), and transported to the Texas Agrilife Research Mariculture Laboratory in Port Aransas, Texas. At the time of collection, gonads of individuals in wild populations are minimal in size (Cunningham 2008). Sea urchins were held for 1 mo in 750 L tanks at approximately 32 ± 2 ppt salinity and 23 ± 1 C. Natural seawater was filtered to 0.5 µm using stratified sand filtration, a Diamond water filter (Diamond Water Conditioning, Horton, WI, USA), and then piped to the culture tanks under flow-through conditions. During this period sea urchins were fed a maintenance ration (ca. once every 3 d) of a formulated diet (31:33% protein : carbohydrate diet; Table 1).

To stock the experimental systems, L. variegatus (12.6 ± 0.12 g wet weight, 29.5 ± 0.11 mm test diameter, n = 64) were haphazardly selected, blotted dry with a paper towel to remove excess water, and weighed to the nearest mg with a Mettler balance (Mettler Toledo Scales, Dublin, OH, USA). Test diameters were measured at two perpendicular points across the ambitus using calipers and the mean test diameter was calculated. Sixteen individuals were weighed as described above, and dissected into the body compartments (test and spines, Aristotle's lantern, gut, gonads) to characterize the state of the sea urchins at the beginning of the experiment. For dissection, sea urchins were cut through the peristomial membrane on the oral surface. The gut (esophagus, stomach, and intestine) was rinsed in sea water to remove gut contents. Each organ was blotted with a paper towel to remove excess water and weighed to the nearest mg. Organs were placed in aluminum pans, dried to constant weight at 60 C, and reweighed. At the end of the study sea urchins from each of the diet treatments were weighed and dissected as described above.

For the diet trials, sea urchins were placed individually into a cylindrical enclosure. The cylindrical enclosures were made from a plastic mesh (ca. 12 cm diameter, 30 cm high, and a 4-mm open mesh) secured by plastic cable ties. The mesh enclosures were fitted into 11.5 cm internal diameter PVC couplings so that the floors of the cylindrical enclosures were approximately 5.5 cm above the bottom of the tank. Small plastic spacers (ca. 0.5 cm thick) were then placed under the bottom of each coupling to allow water circulation underneath the enclosures. Four cylindrical enclosures were placed into a 0.07-m² bottom surface fiberglass tank with 20 L water volume. Water volume was held constant by a central standpipe below the top of the enclosures to prevent escape. Seawater was supplied to each enclosure approximately at a rate of 25 L/h. Each diet treatment was comprised of four fiberglass tanks containing four cages per tank (n = 16 individuals per diet treatment, assigned haphazardly). The fiberglass tanks were connected to a temperature-controlled semi-recirculating aquaculture system with mechanical and biological filtration, foam fractionation, and UV sterilization. Seawater was exchanged in the semi-recirculating systems at an approximate rate of 10% volume/d.

Culture conditions were checked daily to maintain constant conditions (32 ± 2 ppt salinity, 22 ± 2 C, and 7 ± 2 ppm dissolved oxygen). Photoperiod was 12 h : 12 h light : dark. Ammonia, nitrite, nitrate, and pH levels were checked weekly and were maintained at or below the following concentrations: ammonia 0.1 ± 0.05 ppm, nitrite 0.1 ± 0.05 ppm, nitrate 5 ± 2 ppm, and pH 8 ± 0.3.

Feeds and Feed Preparation

Four semi-purified diets (Table 1) that varied only in levels of protein and carbohydrate (31:33%, 25:39%, 21:44%, and 17:47% protein : carbohydrate diets; to be referred to as 31:33, 25:39, 21:44, and 17:47) were prepared from semi-purified and purified ingredients (ca. 28% marine source ingredients, 34.6% plant source ingredients, 7.5% crude fat, 1.1% carotenoids, 0.7% vitamin premix, 17.9% mineral premix, 10.2% binder-antifungal-antioxidant; Table 1). Protein and carbohydrate levels were adjusted by adding different levels of soy protein isolate (Zeigler Bros. Inc., Gardeners, PA, USA) and pure wheat starch (MP Biomedical, Solon, OH, USA). All other ingredient levels remained constant (Table 1). Dry ingredients were blended with a twin shell dry blender (Patterson-Kelley Co., East Stroudsburg, PA, USA) for 10 min, and mixed in a Hobart mixer (Model A-200, Hobart Corporation, Troy, OH, USA) for 40 min. Deionized water (500 mL/kg) was then added to the dry ingredients and mixed an additional 10 min to achieve a mash consistency appropriate for extrusion. The mash was extruded through a meat chopper attachment (Model A-200, Hobart Corporation) fitted with a 4.8-mm die. Moist feed strands were dried on wire racks in a forced air oven at 35 C to a moisture content of 8–10%, placed into plastic bags and stored in a refrigerator at 4 C until used.

Proximate analysis of the diets (Table 1) was performed by Eurofins Scientific Inc., Des Moines, IA, USA. Energy content was determined by microbomb calorimetry (Parr Instrument Company, Moline, IL, USA).

Feed Intake

A preweighed amount of feed pellets (weighed to the nearest mg) was placed into individually labeled plastic bags and assigned to each individual urchin in each diet treatment (n = 16 plastic bags per treatment). Each sea urchin was fed a uniformly sized pellet from its respective plastic bag once daily at an amount higher than could be consumed in a 24-h period (ad libitum feeding). Daily feed intake was estimated by visual quantification of the amount of uneaten pellet in each individual enclosure as suggested by Jones et al. (2010) and Hammer et al. (2010). Prior to the next feeding, uneaten food was manually removed by siphon. For each 4-wk period, the weight of the feed pellets fed was determined, and feed intake was estimated for each individual. Although this method estimates feed intake, technical assistants were trained in feed intake evaluation prior to the experiments, and data are comparable among treatments. In addition, previous studies indicate that feed leaching/disintegration was consistent among all diet treatments, consequently, these values represent overestimates of actual feed intake. The following variables were calculated:

Growth

The test diameter and total wet weight were measured as previously described. The following variables were calculated:

Dry Matter Production and Production Efficiency

The following variables were calculated:

Statistics

Statistical comparisons for feed intake, protein intake, test diameter, and wet weight at 4, 8, and 12 wk were completed using SAS (version 9.1). These parameters were analyzed as the dependent variable in a repeated measures model with feed protein level as the predictor. The PROC MIXED procedure (SAS) was a general linear model used to analyze the relationship between each parameter and time accounting for group differences. Group means in 4-wk increments were compared using the Tukey's adjustment. Treatment tanks were arranged in a block design; however, tank effects among replicate block tanks were not found to be significant and each individual held within an enclosure was considered to be an independent observation. Statistical comparisons for feed consumption and protein consumption were performed on the Systat 11 software package (Systat Software Inc., Point Richmond, CA, USA). All other statistical comparisons were performed on the Systat 11 software package. If organ growth, production, and production efficiency data were normally distributed and were homoscedastic, parametric tests including ANOVA and ANCOVA were completed. ANCOVA was used when necessary to eliminate confounding effects of animal size for organ analyses. When significant differences were determined, a Tukey's test for pairwise group comparisons was used. A P value of <0.05 was determined statistically significant for all parametric tests. If these data were non-normally distributed or heteroscedastic, data transformations were attempted. If data transformations were not successful, nonparametric tests (Kruskal–Wallis, Mann–Whitney U, or Kolmogorov–Smirnov two-way tests) were employed. To maintain an overall acceptance criterion of α = 0.05 during multiple comparisons, a Bonferroni's adjustment was adopted for all nonparametric analyses.

Survival

Survival was 100% in all feed treatments for the entire 12-wk study.

Feed Intake

The predictive model from The Mixed Procedure indicated there was an overall significant time effect, group (diet) × time effect, but not a significant group effect for total feed intake (Table 2A). Total feed intake was not different among the diet treatments at Week 4 (Table 2B and Fig. 1A). At Week 8, sea urchins fed the 31:33% diet consumed significantly less food than those fed the 17:47% or 21:44% but not the 25:39% diet. At Week 12, sea urchins fed the 31:33% diet consumed significantly less food (14.9 g) than urchins fed the 17:47%, 21:44%, or 25:39% diets (18.3, 18.2, and 16.9 g, respectively).

The predictive model from The Mixed Procedure indicated an overall significant time effect, group × time effect, but not a significant group effect for the total dry protein intake (Table 2A). Sea urchins fed the 31:33% diet consumed significantly more dry protein than sea urchins fed the 17:47% and 21:44% diets but not the 25:39% diet at Week 4 (Table 2B and Fig. 1B). At Week 8 the dry protein intake was significantly different among all the treatments and remained different among treatments for the duration of the study.

The predictive model from The Mixed Procedure indicated an overall significant group effect, time effect, and group × time effect for energy intake (Table 2A). There was no significant difference in energy intake among the diet treatments at Week 4 or 8 (Table 2B and Fig. 1C). At Week 12, energy intake was significantly less in urchins fed the 31:33% diet but was not different among the other diet treatments.

Growth

The predictive model from The Mixed Procedure indicated an overall significant group effect, time effect, and group × time effect for test diameter (Table 3A). Sea urchins fed the 17:47% diet had a significantly smaller test diameter (31.9 mm) than urchins fed the 31:33% (33.2 mm) or 25:39% diets (33.1 mm) but not the 21:44% diet (32.8 mm) at Week 4 (Table 3B and Fig. 2A). At Weeks 8 and 12, sea urchins fed the 31:33% diet had a significantly larger test diameter than urchins fed the 17:47% or 21:44% diets but not the 25:39% diet (Table 3B and Fig. 2A).

The predictive model from The Mixed Procedure indicated an overall significant time effect, group × time effect, but not a group effect for wet weight (Table 3A). At Week 4, the total wet weight of sea urchins fed the 31:33% diet was significantly higher (20.2 g) than sea urchins fed the 17:47% (17.3 g), but not the 21:44% (19.0 g) or 25:39% (19.6 g) diets (Table 3B and Fig. 2B). At Weeks 8 and 12, the total wet weight was different among all the diet treatments and varied directly with protein level.

There was no significant linear relationship (R = 0.088, P = 0.488) between the energy consumption and wet weight gain at Week 12; however, there was a significant linear relationship (R = 0.585, P < 0.001) between protein consumption and wet weight gain (Fig. 3A, B). There were also significant curvilinear relationships between wet weight gain and both protein : carbohydrate ratio (R = 0.692, P < 0.001) and protein : energy ratio (R = 0.694, P < 0.001) at Week 12 (Fig. 4A, B).

Organ Wet and Dry Weight

Test wet and dry weight varied directly with dietary protein and inversely with dietary carbohydrate content at 12 wk (Table 4). Sea urchins fed the 31:33% diet had a significantly higher test dry weight than urchins fed the 17:47% and 21:44% diets but not the 25:39% diet at Week 12. Wet and dry weight of Aristotle's lantern did not vary significantly among the diet treatments (Table 4). Gut wet weight varied directly with dietary protein and inversely with dietary carbohydrate content at 12 wk; however, the increase in gut wet weight was isometric with test wet weight. Gut dry weight did not vary with diet treatment. Gonad wet and dry weight varied directly with dietary protein content and inversely with carbohydrate content at Week 12; however, the increase in gonad wet and dry weight was isometric with test wet and dry weight, respectively.

Production and Efficiency

The feed conversion ratio (FCR) of sea urchins varied indirectly with dietary protein content and directly with the carbohydrate level (Table 5). Sea urchins fed the 31:33% diet had a significantly lower FCR (0.6) than sea urchins fed the 17:47% (1.2) or 21:44% (0.9) diets but not the 25:39% (0.7) diet at Week 12 (ANOVA, F = 39.054, df = 3, P < 0.001). Sea urchins fed the 31:33% diet had significantly higher production (7.1 g) and production efficiency (53%) than sea urchins fed the 17:47% (4.2 g, 27.3%) or 21:44% (5.8 g, 35.3%) diets (ANOVA, F = 14.084, df = 3, P≤ 0.032, production; F = 19.309, df = 3, P≤ 0.003, production efficiency; Table 5). Although production did not vary between urchins fed the 31:33% (7.1 g) and 25:39% (6.0 g) diets, those fed the 31:33% diet had significantly higher production efficiency (53%, ANOVA, F = 19.309, df = 3, P≤ 0.003) than those fed the 25:39% (39.3%) diet (Table 5).

Gonad production was significantly less in sea urchins fed the 17:47% (1.2 g) diet than in urchins fed the other three diets (1.7, 1.8, 1.9 g, respectively, ANOVA, F = 8.112, df = 3, P≤ 0.016; Table 5). Gonad production efficiency was significantly higher in sea urchins fed the 31:33% (12.6%) diet than in urchins fed the 17:47% (6.9%) and 21:44% (9.4%) diets but not the 25:39% (10.9%) diet (ANOVA, F = 17, 47.299, df = 3, P≤ 0.020). There was no difference in the protein efficiency ratio among the diet treatments at Week 12 (Table 5).

Feed Intake

Direct measures of feed intake could not be accomplished in this study due to time and labor available; thus, feed intake was estimated as described previously (Jones et al. 2010). These feed intake estimates are reasonable, as personnel were trained in feed intake evaluation and all evaluations were consistent. It is acknowledged that these estimates may result in increased variation and increase the potential for Type II error. Individual sea urchins feed intermittently and without consistency, that is, some individuals will feed, whole or in part, immediately, whereas others may not feed for several hours, allowing nutrient leaching (Hammer 2006). Consequently, feed estimates are always confounded by leaching/disintegration, and feed intake values will be overestimates of the actual amount of feed consumed. However, these estimated feed intake values are comparable among the diets in this study.

Reduced feed intake of the high protein (high energy) diet suggests that L. variegatus consumed feed to satisfy an energy requirement. However, energy alone cannot easily explain all of the differences observed in growth and production rates. Protein (or amino acid) levels, protein : energy ratios or, more specifically, protein : carbohydrate ratios may affect growth and production rates. Inherent differences in specific nutrient digestibilities may also affect production. Interestingly, the increased feed intake by urchins fed the 17:47%, 21:44%, and 25:39% diets did not become significant until the final period of the study (Weeks 8–12) suggesting that dietary requirements were being satisfied early but were later insufficient to support maximal weight gain. Hammer et al. (2006) reported that adult L. variegatus fed a 9% protein diet consumed more food than urchins fed 20 or 31% protein diet. Similar results were observed with small L. variegatus (Hammer et al. 2004). Fernandez and Boudouresque (1998) and McBride et al. (1998) indicated an inverse relationship between feed intake and protein content for small Paracentrotus lividus and Strongylocentrotus franciscanus, respectively. Otero-Villanueva et al. (2004) indicated that feed intake rates were lower for small Psammechinus miliaris fed a commercial salmon feed (37% protein) than for urchins fed a diet of mussels (31% protein), formulated feed (27% protein), or algae (2% protein). Miller and Mann (1973) suggested that sea urchins in the field may consume large amounts of carbohydrate rich (protein poor) material to process and obtain necessary protein. We do, however, recognize the difficulties in comparing different studies due to differences in source protein quantity and quality as well as differences in other nutrients that may change in response to changing a practical or natural feed ingredient. Additionally, although all urchins were “feed trained” using the 31:33% diet, we do not believe this diet provided an advantage to this treatment, as urchins will readily consume almost any source of organic nutrients.

Organ Weight

Evaluating growth in sea urchin organs (test, Aristotle's lantern, gut, gonad) represent a unique challenge in evaluating treatment effects. Organ allometry can change with organismal size and presumably those factors that induce changes in the size of the organism. The test, gut, and gonad wet weights increased with dietary protein content; however, these organs increased in proportion to the size of the sea urchin. That is, organ allometry did not occur with protein and carbohydrate feed intake, but was isometric. Dry weight values for the gut and gonad showed similar, but less significant, trends with feed protein and carbohydrate levels. Aristotle's lantern did not vary with dietary protein and carbohydrate, but varied directly based on the size of the individual.

Production

The lower food conversion ratio, higher production, and higher production efficiency of urchins fed the 31:33% diet suggests that this diet promotes efficient somatic growth. Total protein and/or indispensable amino acids may be limiting in diets with lower protein levels. Production efficiencies in this study were higher (25–48%) than those reported for large adult L. variegatus (9–26%) (Hammer et al. 2006; Watts et al. 2011) and small L. variegatus (12–33%) (Hammer et al. 2004). The high production efficiencies in this study resulted from a combination of lower consumption and higher production, strongly suggesting that the feeds were of high nutritional quality. Additionally, the feeds used in this study (extruded pellets) had much lower water content (higher nutrient density) than those reported in previous studies (wet pellets, low nutrient density) with L. variegatus. These observed increases in efficiencies are important, as they suggest extruded dry feeds are used effectively by sea urchins for somatic growth.

Fernandez and Boudouresque (1998) reported that production efficiency was directly related to protein content of the diet in small P. lividus cultured for 9 mo. McBride et al. (1998) reported that reduced feed intake in small S. franciscanus fed 40 and 50% protein diets led to higher production efficiencies than for urchins fed a 30% protein diet. Otero-Villanueva (2004) reported highest production efficiencies in small and large P. miliaris fed high protein versus low protein diets. As indicated previously, comparisons among studies are constrained due to differences in source protein quality and quantity.

Higher dietary protein and/or carbohydrate support efficient gonad production. The gonad production efficiencies in this study (6.9–12.6%) are higher than those reported for adult L. variegatus (1.9–9.3%) (Hammer et al. 2006; Watts et al. 2011) and for small L. variegatus (0.59–4.1%) (Hammer et al. 2004), further supporting the value of an extruded dry feed in improving roe production in captured adult urchins.

Protein efficiency ratios (PER) were similar among the feeds examined, suggesting that adequate nonprotein energy was not limiting for the growth rate obtained and that protein was probably not being utilized as a significant energy source. In contrast, Hammer et al. (2006) reported PER values that were significantly lower in a 31% protein feed (compared to 9 and 20% protein feeds) and suggested that the protein was used as an energy source, as carbohydrate was limited in the feed (12 vs. 33% in the current feed). The use of carbohydrate as the primary energy source in sea urchins is suggested by Taylor (2006) and Marsh and Watts (2007). Akiyama et al. (1997) reported PER values ranging from 0.77 to 6.6 (based on wet weight gain) for small P. depressus fed 4 diets of 26 to 30% crude protein. Schlosser et al. (2005) reported that protein efficiency (based on digestible protein) was much higher in adult P. lividus fed a formulated diet (23% protein) than that of urchins fed two different algae feeds (37 and 15% protein) because the formulated diet contained more digestible energy.

The protein : energy ratio of the 31:33% diet was approximately 80.8 mg protein/Kcal and is slightly lower than the ratio used in the diets of many commercial shrimp species (ranging from 90 to 160 mg protein/Kcal reviewed in Cuzon and Guillaume 1997). A better defined protein : energy ratio for this species might allow for a reduction in the amount of protein in the feed without limiting growth. Such a reduction in feed protein would reduce feed cost and pollution potential of the feed in aquaculture systems.

From a commercial standpoint, determination of protein requirements in sea urchins is more complex than in finfish. The ultimate goal of commercial sea urchin culture is the production of consumer-preferred roe. Although certain protein levels may promote maximum somatic weight gain and roe production, these same levels may negatively impact other important qualities (taste, texture, color) for the successful marketability of roe. Pearce et al. (2002) found that high levels of dietary protein increased the degree of bitterness in the roe of cultured gonads in Strongylocentrotus droebachiensis. Thus, determination of protein, carbohydrate, and neutral fat requirements in urchin culture will require an “optimization” of energy levels in formulated feeds, to balance the goal of efficient somatic growth with the production of quality roe.

In summary, sea urchins grew at a rate higher than that reported for any other study using adult L. variegatus, suggesting that the dry diets used were of high quality. Somatic and gonad production were significant, indicating that somatic weight gain and gonad mass of wild-caught L. variegatus could be greatly increased within a 12-wk period using extruded dry diets. These data indicate that we are approaching the upper requirement for feed protein levels, although it may be possible to improve upon the 31:33% diet through evaluation of amino acid requirements and nutrient digestibilities. This study suggests that sea urchins are very efficient at converting protein to tissue (regardless of the protein level of the feed) as long as sufficient digestible energy is available (protein sparing). We suggest that future nutrition research focus on protein : energy ratios, indispensable amino acid levels and requirements, and highly digestible protein and carbohydrate sources. These areas of research are essential to the development of cost-effective, commercially available feeds for the future aquaculture of sea urchins.