Bioprocessed soybean meal replacement of fish meal in rainbow trout (Oncorhynchus mykiss) diets

This 125-day experiment evaluated the growth of adult rainbow trout (Oncorhynchus mykiss) fed one of three isonitrogenous and isocaloric diets (46% protein, 16% lipid). Fish meal was the primary protein source for the reference diet, which was compared to two other diets where bioprocessed soybean meal replaced 60% or 80% of the dietary fish meal. At the end of the experiment, there were no significant differences in gain, percent gain, feed conversion ratio, or specific growth rate among the dietary treatments. There were also no significant differences in intestinal morphology, splenosomatic index, hepatosomatic index, and viscerosomatic index among the diets. Based on these results, bioprocessed soybean meal can replace at least 80% of the fish meal in adult rainbow trout diets. Subjects: Environment & Agriculture; Bioscience; Food Science & Technology


Introduction
Plant-based proteins, such as soybeans (Glycine max), have been extensively researched as alternatives to dietary fish meal (Gatlin III et al., 2007;Li & Robinson, 2015;Nordrum, Bakke-McKellep, Krogdahl, & Buddington, 2000). Alternative protein sources are needed due to the exponential growth of aquaculture without a corresponding increase in sources of fish meal, which is primarily made from small ABOUT THE AUTHORS Jill M. Voorhees is a research biologist with the South Dakota Department of Game, Fish and Parks. She conducts aquaculture and fisheries research on numerous trout and salmon species.
Michael E. Barnes

PUBLIC INTEREST STATEMENT
Aquaculture is rapidly growing to feed the increasing number of people in the world. Many farmed fish diets have contained large amounts of fish meal, which is primarily made from small ocean fish like anchovies or menhaden. Because fish meal supplies are limited, feed for some farmed fish can become expensive and limiting. Soybean meal is much more abundant and sustainable than fish meal, but soybean meal cannot be digested by carnivorous fish like trout and salmon. By-processing, such as fermentation, can make soybeans a better food source for fish. This study showed that bioprocessed soybean meal can replace almost all of the fish meal in the diet of rainbow trout.
pelagic fish (Food and Agriculture Organization of the United Nations (FAO), 2016). Thus, there needs to be other suitable and cost-effective protein sources to replace dietary fish meal.

Methods
This feed trial was conducted indoors at McNenny State Fish Hatchery, Spearfish, South Dakota, using degassed and aerated well water at a constant temperature of 11°C (total hardness as CaCO 3 , 360 mg/L; alkalinity as CaCO 3 , 210 mg/L; pH, 7.6; total dissolved solids, 390 mg/L) in a single-pass, flow-through system. Seventy-two Erwin x Arlee strain rainbow trout (initial weight 130.7 ± 4.2 g, length 213.2 ± 2.0 mm, mean ± SE) were randomly selected and stocked into one of 12 semi-circular fiberglass tanks (190-L) on 15 June 2016, at six fish per tank (n = 4). Flow rates were kept constant throughout the 125-day study.
Three different diets were used (Table 1), with a modified soybean meal replacing 0%, 60%, or 80% of the fish meal as the primary protein source. The modified soybean meal was produced using a proprietary microbial conversion process (SDSU, Brookings, SD, USA). All of the diets were isocaloric and isonitrogenous, and were all manufactured by cooking extrusion (ExtruTech model 325, Sabetha, KS) at Prairie Aquatech (Brookings, SD). The individual fish weights were combined to obtain total tank weight. Fish were individually weighed and measured approximately every four weeks. Weight gain, percent gain, feed conversion ratio (FCR), and specific growth rate (SGR) were calculated for each individual tank. Individual fish weights and total lengths were used to calculate Fulton's condition factor (K).
Fish were fed by hand daily, except on the days they were weighed and measured (days 30, 61, 92, and 125). Feeding amounts were initially determined by the hatchery constant method (Butterbaugh & Willoughby, 1967), with planned feed conversion rates of 1.1 and maximum growth rate of 0.08 cm/day, which was based on historical maximum growth rate of Erwin x Arlee strain rainbow trout at McNenny State Fish Hatchery (Barnes et al., 2011), and then adjusted daily to be at or near satiation. Total feed fed and mortalities were recorded daily. To collect weight and length data on 30-day intervals, the fish were anesthetized using 60 mg/L MS-222 (Tricaine-S, tricaine methanesulfonate, Syndel USA, Ferndale, Washington). On day 125, fish were euthanized using a lethal dose of 250 mg/L MS-222 (American Veterinary Medical Association (AVMA), 2013). In addition to weight and length measurements, fin lengths to the nearest 1.0 mm, and organ (spleen, liver, and visceral) weights to the nearest 1.0 mg, were recorded from three randomly selected trout per tank. Fin indices, hepatosomatic index (HSI) (Strange, 1996), splenosomatic index (SSI) (Goede & Barton, 1990), and viscerosomatic index (VSI) (Goede & Barton, 1990) were calculated for individual fish.
Data were analyzed using the SPSS (9.0) statistical analysis program (SPSS, Chicago Illinois), with significance predetermined at P < 0.05. One-way analysis of variance (ANOVA) was conducted, and if treatments were significantly different, post hoc mean separation tests were performed using Tukey's HSD test.

Results
At the end of this experiment, there were no significant differences in gain, percent gain, feed fed, feed conversion ratios, specific growth rates, or percent mortality among the tanks of fish being fed the three different diets (Table 3). Overall mean (± SE) feed conversion ratios were not significantly different at 1.30 (± 0.04), 1.14 (± 0.03), and 1.25 (± 0.07) for the 0%, 60%, and 80% BSM diets, respectively. There was no mortality observed in any treatments.
There were no significant differences among the diets in gain, percent gain, or SGR overall and after the first rearing period (days 31-125). However, during the first rearing period, the fish in the tanks that were fed the reference (fish meal) diet had significantly higher gain, percent gain, and SGR than the fish in the tanks receiving the 80% BSM diet, but were not significantly different than the fish receiving the 60% BSM diet. Mean (± SE) percent gain at the end of the first rearing period was 26.3 (± 1.7) %, 18.8 (± 3.3) %, and 9.9 (± 2.6) % for the fish being fed the 0%, 60%, and 80% diets, respectively.
Similarly, there were no significant differences overall in individual fish weight, length, or condition factor (Table 4). However, during rearing period 3, the mean (± SE) condition factor of the fish in tanks fed the fish meal reference diet was 1.38 (± 0.01), which was significantly different from the fish in tanks being fed the 80% bioprocessed soybean meal diet at 1.28 (± 0.02). The condition factor of the fish in the tanks that were fed the 60% BSM was 1.31 (± 0.02), which was not significantly different from the other two diets.
Fish receiving the 80% bioprocessed diets had significantly longer dorsal fins than those receiving the 60% diet, but were not significantly different from those fed the reference diet. No significant differences were observed among the dietary treatments for the pectoral and pelvic fin indices. There were also no significant differences in any of the organosomatic indices (HSI, SSI, and VSI), nor any of the histological scores (lamina propria, connective tissue, and vacuoles). Representative images of the distal intestines from fish fed each diet used for the scoring are shown in Figures 1-3. Table 2. Histological scoring system used on rainbow trout fed fish meal or incremental amounts of bioprocessed soybean meal in diets (Barnes et al., 2014, modified from;Goede & Barton, 1990;Barton et al., 2002) Score Appearance Lamina propria of simple folds 1 Thin and delicate core of connective tissue in all simple folds.
2 Lamina propria slightly more distinct and robust in some of the folds. 3 Clear increase in lamina propria in most of simple folds. 4 Thick lamina propria in many folds.

5
Very thick lamina propria in many folds.
Connective tissue between base of folds and stratum compactum 1 Very thin layer of connective tissue between base of folds and stratum compactum.
2 Slightly increased amount of connective tissue beneath some of mucosal folds. 3 Clear increase of connective tissue beneath most of the mucosal folds.

4
Thick layer of connective tissue beneath many folds.

5
Extremely thick layer of connective tissue beneath some of the folds.
At 125 days, this experiment should have lasted long enough to determine any differences in fish rearing performance among the diets (Weathercup & McCraken, 1999). NRC (2011) recommends feed trial durations of 56-84 days, with larger fish attaining at least a 200-300% gain. This experiment provided fish weight gains of approximately 250%, thereby meeting both requirements.
The FCR observed in this experiment was slightly higher than that reported in some other experiments involving rainbow trout (Barnes et al., 2012(Barnes et al., , 2013, but were also similar to other studies (Barnes et al., , 2015aBruce et al., 2017b;Yamamoto et al., 2010Yamamoto et al., , 2012. The SGR was slightly lower in this experiment (0.9-1.0) compared to the 1.0 to 1.3 reported by Bruce et al. (2017b) in a similar study, but were extremely low compared to 1.8 to 3.0 reported by Yamamoto et al. (2010Yamamoto et al. ( , 2012 and Bruce et al. (2017a). The slower growth rate could possibly be due to the size of the fish or water temperatures differences. Yamamoto et al. (2010Yamamoto et al. ( , 2012 and Bruce et al. (2017a, 2017b) used juvenile fish, while this study used adult rainbow trout, which have slower growth rates (Stickney, 1994).
The condition factors observed in this experiment were higher than many other rainbow trout experiments (Barnes et al., 2012(Barnes et al., , 2013(Barnes et al., , 2015a(Barnes et al., , 2015bBruce et al., 2017b). This could possibly be because the fish in this experiment were older and larger, and closer to sexual maturity (Barton, Morgan, & Vijayan, 2002).
The lack of any differences in HSI between the dietary treatments indicates similar energy partitioning. HSI is an indirect measure of glycogen and carbohydrate levels, and can be used to indicate the nutritional state of the fish (Barton et al., 2002;Daniels & Robinson, 1986;Kim & Kaushik, 1992). The HSI levels observed in this study were similar to those reported by Barnes et al. (2013Barnes et al. ( , 2014Barnes et al. ( , 2015b, and slightly higher than those reported by Yamamoto et al. (2010Yamamoto et al. ( , 2012, Barnes et al. (2012), and Bruce et al. (2017a). Differences in HSI among the studies could be related to fish age. Barton et al. (2002) noted that organosomatic indices can vary depending on life stage, and the rainbow trout used in this study were much larger and older than those used in other experiments.
Enteritis was not observed in this study, despite the well-documented and potentially negative effects of soybean products to the distal intestine of rainbow trout (Romarheim et al., 2008;Merrifield, Dimitroglou, Bradley, Baker, & Davies, 2009;Sealey, Barrows, Smith, Overturf, & LaPatra, 2009). The BSM used in this study likely decreased or eliminated the saponins (Krogdahl et al., 2015) and other antinutritional factors responsible for such enteritis (Barnes et al., 2012(Barnes et al., , 2013Yamamoto et al., 2010Yamamoto et al., , 2012. The absolute intestinal scores observed in this study tended to be lower than those reported by Barnes et al. (2014Barnes et al. ( , 2015aBarnes et al. ( , 2015b for rainbow trout fed different fermented soybean meal diets. This could be due to the dietary differences among the studies or scoring difference between readers.
In conclusion, this study indicates that at least 80% of the dietary fish meal can be directly replaced by BSM in diets of adult rainbow trout. It is unknown if the suitability of dietary BSM extends further during the trout life cycle, prior to spawning. Additional research is needed to determine if this BSM can replace all of the fish meal in adult rainbow trout diets.