Digestibility and gastrointestinal transit of Ulva fasciata seaweed meal in tilapia ( Oreochromis niloticus ) juveniles : basis for the inclusion of a sustainable ingredient in aquafeeds Digestibilidade e trânsito gastrointestinal do farelo de alga Ulva fasciata em juvenis de tilápia ( Oreochromis niloticus ) : bases para a inclusão de um ingrediente sustentável na aquicultura Digestibilidad y tránsito gastrointestinal de harina de alga Ulva fasciata en juveniles de tilapia ( Oreochromis niloticus ) : bases para

The seaweed Ulva fasciata has many features favorable to integrated multi-trophic aquaculture (IMTA). It is efficient at biofiltering, shows high biomass production, and is rich in many nutrients useful in aquatic animal diets. We evaluated the digestibility of the seaweed meal of U. fasciata produced in IMTA and its effects on gastrointestinal transit time in tilapia (Oreochromis niloticus) juveniles. Juveniles (6.30 ± 1.80 g initial weight, and 5.5 ± 0.61 cm initial length) were cultivated in six tanks (50 individuals per tank) in a closed recirculating aquaculture system. The digestibility of Ulva meal was 57.92 ± 5.21% for dry material, 78.59 ± 1.91% for protein, and 69.87 ± 3.72% for energy. The inclusion of 10% seaweed meal did not alter the gastrointestinal transit time in tilapia juveniles as compared to controls. The earliest colored feces were observed four hours after first feeding in both treatments (feed diets with [10%] and without seaweed); all fecal material was colored after ten hours. The digestibility of seaweed meal was satisfactory for dry material, protein, and gross energy, and the inclusion of 10% of that meal did not change gastrointestinal transit time - indicating that the inclusion of 10% seaweed meal in tilapia diet is safe and without any nutritional use losses.


Introduction
Food supply sustainability is integrally related to the development of high-efficiency ecofriendly production systems. Integrated Multi-Trophic Aquaculture (IMTA) systems are based on the efficient utilization of materials as one organism uses the wastes of another in a recycling loop (Troell et al., 2014). Seaweeds are often used as biofilter in IMTA with marine fish and shrimp and can produce green biomass rich in protein and bioactive compounds of great interest for animal nutrition (Øverland et al., 2018).
The possibility of producing plant ingredients within the same fish production system reduce costs, diminish gaseous emissions related to transport, and simplify the overall production chain. As such, the use of seaweed biomass as an ingredient in fish diets would contribute to the economic viability of IMTA and diminish its footprint (Marinho et al., 2013).

Methodology
Three hundred tilapia juveniles (mean weight 6.30±1.80 g; mean length 5.5 ± 0.61 cm) were used [Ethical Committee for Animal Uses of the Fisheries Institute of Rio de Janeiro State (FIPERJ) -number 001/2017]. The fish were randomly distributed in six cylindrical polyethylene tanks (0.1 m³ volume; 50 juveniles per tank) in a closed water recirculation system. Each tank had a continuous water supply stream and drainage system (0.1 m³ h -1 ) with constant, soft, superficial aeration (15 cm deep) to avoid bottom disturbance. The effluents were conducted to a filtering and sterilizing system composed of a sedimentation tank (0.05 m³), a physical filter (1 mm mesh), and six biological filter tanks (0.05 m³ each) containing expanded clay as the substrate for nitrifying bacteria. The water then flowed by gravity to a reservoir (0.2 m³), from which it was pumped (5.80 m³ h -1 ) by a submersible pump (JATO 6000; CUBOS ® ) through a UV filter, and finally redistributed to the fish tanks. The juveniles were acclimated for 15 days before initiating the experiments.
Two experiments were conducted: 1) to evaluate the gastrointestinal transit time (GTT); 2) to quantify the digestibility of the seaweed meal. The basal diet used in the two experiments was formulated from Soybean, Corn, Fish, Poultry by-product meals, in addition to Fish oil, Butyl-Hydroxy-toluene (BHT; antioxidant) and premix Nutrifish-GUABI ® . It presented a proximate composition of 389.12 g kg -1 of crude protein, 92.91 g kg -1 of crude lipid, 13.27 g kg -1 of crude fiber, 114.21 g kg -1 of ash and 18.73 J kg -1 of gross energy.
To prepare the seaweed meal, the biomass of U. fasciata was harvested from an experimental IMTA of seaweed and marine fishes (Eugerres brasilianus and Rachycentron canadum) at the Almirante Paulo Moreira Experimental Aquaculture Station at FIPERJ. The seaweed biomass was dried in an oven (TE-394/3; TECNAL®) at 50 °C for 24 h, and then ground in a rotary mill (TE-651/2; TECNAL®) equipped with a 0.5 mm mesh; 853.9 g -1 of seaweed meal was produced. All ingredients cited in Table 1 were weighed in a centesimal balance, manually mixed, ground in a rotary mill (0.5 mm mesh), moistened with water (50% by weight), homogenized, pelletized (1.0 -2.0 mm), and subsequently dried in a forced air recirculation oven at 50 °C for 24 h.  Development, v. 9, n. 10, e3889108497, 2020 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v9i10.8497 6 In addition to the seaweed meal, two inert markers were used in the experimental feeds, chromium oxide (Cr2O3) and titanium dioxide (TiO2), which lend green and white colors, respectively, to fish feces. During both experiments, the fish were fed until apparent satiety three times a day (at 08:00; 12:00; 16:00 h). The diets, feed administration, feces collections, and tank cleaning schedules varied according to the characteristics of each experiment, as described below.

Experiment 1 -Gastrointestinal transit time (GTT)
The feeds used for GTT evaluations are presented in Table 2. GTT was evaluated using a color index (CI) representing the visual changes in the colors of the fish feces caused by the inclusion of chromium oxide (Cr2O3) or titanium dioxide (TiO2) (which result in green or white tones respectively) (Storebakken, 1985). The CI is expressed as a score (0, 20, 40, 60, 80 and 100%) assigned to the colors of the feceswhere 0% represents completely white feces(without any influence of the diet provided at the beginning of the time count) and 100% represents completely green matter (all feces were from the diet provided after the start of the time count).
The feces for CI attribution were collected by siphoning (using a 4 mm silicone tube) and then transferred to a glass beaker. Tilapia feces are cohesive, which allows their collection by suction without affecting their physical structure. Color evaluations were performed on a gray bench illuminated by white light (LED 5w-6500k), facilitating distinctions of the different tones and colors. When collections coincided with feeding times, tank cleaning was performed immediately after feeding to avoid mixing the collected feces and feed remains.
For those experiments, the fish were fed with a reference diet containing TiO2 for two days, without any seaweed meal (Wref, Table 1) to dye all feces in white (and thus increase color contrasts). After the feces were completely white (CI-0%), feeding was initiated with a reference green test feed containing Cr2O3, without a seaweed meal (Gref). The feces of each tank (n=6) were then collected separately, every two hours for 24 h, and the CI assessed. The next step involved feeding the fish with a white test diet (Wtest , Table 1), containing TiO2 and seaweed meal, for two days, to dye all of the feces white again, and then restart the feeding-cleaning-feces collection cycle with green test feed (Gtest, Table 1) containing Cr2O3 and seaweed meal. Development, v. 9, n. 10, e3889108497, 2020 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v9i10.8497 Similarly, the feces of each tank were collected every two hours, during 24 h, and the CI determined. Total GTT was determined by the time required for 100% CI in all replicates.
The fish in all six tanks were first fed with the reference diet, with feces collections beginning two days later. The feces were siphoned from each tank separately (n=6), every 2 hours, for five days. The collected material was dried at 60 °C for 12 hours and subsequently frozen at -15 °C for posterior analysis (gross energy, crude protein, dry matter, and Cr2O3 content). The same procedure was used for the test diet.
Composition analyses were made of the reference diet, test diet, the respective feces produced, and the seaweed meal (n=3). Gross energy was determined by burning the sample in a calorimetric pump; crude protein content was determined using the Dumas method; dry matter content was determined by drying loss; and Cr2O3 content was determined using atomic absorption spectroscopy, following the methodology recommended by the Brazilian Association of Animal Feeding Compendium (CBAA, 2009).
The normality (Shapiro-Wilk) and homogeneity of the variances (Cochran) of the IC data were tested according to Zar (1996). Analysis of variance (ANOVA) for repeated measures was used to verify differences between the ICs of diets with and without seaweed meal over time. The limit of tolerance for all tests was 95% (p<0.05). The statistical analyses were conducted using Statistica 6.0 StatSoft Inc. Software
Four hours after the beginning of feeding on diets with and without seaweed meal, the feces began to change color from white to green. The CI was 100% after 10 h in both cases for all treatments and replicates, suggesting that 10 h was the total GTT for both tested diets (Figure 1). Differences in ICs between diets were not observed at any time (F = 1.569; p=0.271). The seaweed meal contained 243.80 g kg -1 of crude protein and 12.73 KJ g -1 of gross energy, on a dry matter basis. Digestible protein was 191.60 g kg -1 , and digestible energy 8.89 KJ g -1 . The digestibility coefficients of both diets and the seaweed meal are presented in Table 2.  Research, Society and Development, v. 9, n. 10, e3889108497, 2020 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v9i10.8497

Discussion
Seaweed meal derived from U. fasciata contained more than 20% ulvan, an important component of NSP; it represented 2.26% of the dry matter in the diet used to determine gastrointestinal transit times. The inclusion of a rich NSP seaweed meal (10%) in the tilapia diet did not change its gastrointestinal transit time. That amount of seaweed meal may not alter the chime enough to affect the GTT or the activities of the digestive enzymes and diminish the expected growth rate of the species. The NSPs present in algae are viscous (Zhu et al., 2015) and, at high levels, can affect the viscosity of both the diet and feces, impairing nutrient digestion and production performance. Those effects were not observed at low inclusion levels, however, and our results, therefore, support the safe inclusion of up to 10% Ulva sp. for tilapia. Likewise, several other studies identified no production performance impacts at 10% inclusion levels (Güroy et al., 2007;Azaza et al., 2008;Ergün et al., 2009;El-tawil, 2010;Marinho et al., 2013).
Fiber enhancements in fish diets tend to diminish both the GTT and nutrient digestibly (Lanna et al., 2004;Rodrigues et al., 2010), although Meurer et al. (2003) observed that crude fiber enhancements of 8.5% in tilapia diets increased the GTT without fish productivity losses.
Fish, as well other monogastric animals, did not have gastrointestinal enzymes capable of degrading NSP (Sinha et al., 2011), although part of the energy contained in those carbohydrates could be obtained from the volatile fatty acids produced by fermentation by microorganisms. In addition to serving as energy sources, those fatty acids could promote immune system gains, ion absorption, the growth of beneficial microorganisms, and the inhibition of pathogens (Montagne et al., 2003;Amirkolaie et al., 2006). According to Haidar et al. (2016), the efficiency of volatile fatty acid use by tilapia depends on the polysaccharide source, feed practices, and processing.
The digestibility of the seaweed meal was similar to that of other plant ingredients. The coefficients of apparent digestibility (CAD) of the dry matter, protein, and energy of seaweed meal were within the ranges reported by Pezzato et al. (2002) for many vegetable meals of tilapia, ranging: from 23.44 to 71.04% for dry matter; from 67.83 to 94.86% for protein; and from 51.00 to 91.29% for energy. The CADs of dry matter and protein, however, were lower than those observed for corn, wheat, and soybeans fed to tilapia juveniles (Furuya et al., 2001). Digestibility can be affected by anti-nutritional factors often observed in vegetable ingredients and, at high concentrations, they can be toxic and impact diet performance (Francis et al., 2001). Azaza et al. (2008) attributed low observed performances of tilapia juveniles to high contents of saponins, tannins, and phytic acid in diets containing more than 30% U. rigida. In comparison with other seaweeds, however, Ulva sp. showed one of the highest protein CAD values for tilapia, although Pereira et al. (2012) reported values of 63.4% and 57.1% for protein and energy respectivelyvalues lower than those observed in the present study.
Variations of CAD values in Ulva sp. could be related to specific chemical (genotypic) heterogeneity or too high chemical plasticity in response to environmental (phenotypic) conditions (Fleurence, 1999). In this sense, IMTA has the potential to produce algae strains rich in proteins and with low chemical heterogeneity, as environmental conditions can be more controlled in closed aquaculture systems than in the open sea. However, optimal parameters have yet to be determined in cultivation, such as stocking density, nutrient availability, water flow and light intensity, as these parameters can determine algae growth and chemical composition (Diamahesa et al., 2017;Oca et al., 2019;Shpigel et al., 2019).
In addition to genotypic (specific) and phenotypic (environmental) expressions of chemical composition, differences in CAD values could be explained by diet processing methods. Extruding processes can alter the digestibility of nutrients and diminish protein digestibility through the denaturation that occurs during the high temperatures usually required (Cheng & Hardy, 2003).
Fermentation can also improve seaweed digestibility, and Felix & Brindo (2014) reported that the fermentation of Ulva lactuca meal increased the digestibility of its dry matter, lipids, and proteins in relation to raw meal (used as feed for the giant freshwater prawn Macrobrachium rosenbergii).
That fermentation process allowed the inclusion of up to 30% of Ulva lactuca meal with no detriment to productive performances.
The advance of the use of Ulva in animal nutrition demands improvements in utilization studies of nutrients present in the seaweed, nutraceutical effect of the present compounds and evaluation of productive performance. In addition, the deepening of the processing of the material, such as the extraction of the ulvan could increase the income of the activity, reduce a possible antinutritional factor and concentrate the other nutrients in this meal.

Conclusion
The seaweed meal of U. fasciata therefore proved to be a generally attractive complementary ingredient for tilapia diets, as the inclusion of 10% of that meal did not affect gastrointestinal transit times and its digestibility was found to be acceptable to that omnivorous fish.