Bacillus subtilis Effects on Growth Performance and Health Status of Totoaba macdonaldi Fed with High Levels of Soy Protein Concentrate

Simple Summary In this study, we investigated Bacillus subtilis 9b effects in Totoaba macdonaldi fed with 30% and 60% of soy protein concentrate substitution. We found that B. subtilis 9b supplementation improved feed intake, weight gain, and internal organs condition of T. macdonaldi fed with 30% substitution. Animals fed with 60% of SPC substitution and B. subtilis doubled their weight and presented 20% more survival than its control diet without B. subtilis 9b probiotic strain. B. subtilis 9b was able to modulate T. macdonaldi intestinal microbiota and increase its resistance to Vibrio harveyi pathogenic strain. Abstract T. macdonaldi is a carnivorous species endemic to the Gulf of California. Indiscriminate exploitation has put totoaba at risk, inducing the development of aquaculture procedures to grow it without affecting the wild population. However, aquafeeds increasing cost and low yields obtained with commercial feeds have motivated researchers to look for more nutritious and cheaper alternatives. Soybean (SB) is the most popular alternative to fishmeal (FM); however, antinutritional factors limit its use in carnivorous species. In this study, we analyzed B. subtilis 9b probiotic capacity to improve growth performance and health status of T. macdonaldi fed with formulations containing 30% and 60% substitution of fish meal with soy protein concentrate (SPC). In addition, we investigated its effect on internal organs condition, their capacity to modulate the intestinal microbiota, and to boost the immunological response of T. macdonaldi against V. harveyi infections. In this sense, we found that T. macdonaldi fed with SPC30Pro diet supplemented with B. subtilis 9b strain and 30% SPC produced better results than SPC30C control diet without B. subtilis and DCML commercial diet. Additionally, animals fed with SPC60Pro diet supplemented with B. subtilis 9b strain and 60% SPC doubled their weight and produced 20% more survival than SPC60C control diet without B. subtilis. Thus, B. subtilis 9b improved T. macdonaldi growth performance, health status, modulated intestinal microbiota, and increased animal’s resistance to V. harveyi infections, placing this bacterium as an excellent candidate to produce functional feeds with high levels of SPC.


Introduction
Aquaculture is one of the fastest-growing industries due to the high demand for aquatic food products. However, this industry requires a fast adaptation to increasing problems such as aquafeed (feed for farmed aquatic animals) ingredients price, diseases, and low yields [1]. Fish products are the main source of proteins and lipids in aquafeed formulations; the balanced amount of essential amino acids and fatty acids required for optimum development, growth, and reproduction of aquacultured animals is of great importance [2]. However, overexploitation of small pelagic fishes to produce aquafeeds

Organisms and Growth Conditions
Juveniles of T. macdonaldi were provided by the marine finfish hatchery of Facultad de Ciencias Marinas, FCM, Universidad Autónoma de Baja California (Ensenada, Baja California, México). Fish were fed with a feed containing 46% protein and 12% lipids and acclimated to experimental facility for two weeks at FCM. A total of 180 fish with an average weight of 153 ± 0.84 g were randomly selected and stocked in 15 tanks of 1100 L. Tanks with 12 fish were connected to a water recirculation system and the flow rate was adjusted to allow 10% of water exchange per day per tank. Water parameters were monitored daily to maintain stable experimental conditions. Temperature was kept at 25 ± 1 • C with thermo-controlled chillers. Salinity was measured with a refractometer and maintained at 35.0 ± 0.5% o . Photoperiod was kept at 12:12 h light:dark. Oxygen concentration was measured with a YSI Pro 20l oximeter and kept above 6 mg/L. Total ammonia-nitrogen (NH 4 + -N) and total nitrite-nitrogen (NO 2 − -N) were measured daily before the first meal of the day with colorimetric test kits (Aquarium Pharmaceutical, Mars, PA, USA). The level of pH was measured using an Oakton pHTestr10 pH meter with an accuracy of 0.01 pH units.
Experiments were performed in triplicate, and fish were fed to apparent satiety twice a day (08:00 and 16:00 h), seven days a week for three months. Feed leftovers were reweighted to measure feed intake. Experimental procedures related to fish husbandry were approved by the Secretariat of Agriculture, Livestock, Rural Development, Fisheries, and Food (Mexican Official Standard NOM-062-ZOO-1999).

Growth Performance
Fish were anesthetized using clove oil at 0.01% (v/v) and weighed individually at the beginning and end of the experiment (24 h fasting before body weight measurement). The following equations were used to evaluate fish growth parameters [22,30]: • Specific growth rate (SGR) • Daily feed intake (DFI) where: FW = Average final weight (g); IW = Average initial weight (g), FIW: Fish initial weight (g); LnFW = Ln final weight; LnIW = Ln initial weight; t = time (days).

Intestine Recollection and Organs Observation
At the end of the bioassay, two fish per treatment were overdosed with clove oil and sacrificed to observe visceral fat content, liver, spleen, distal intestine condition, fillet firmness by tactile and visual assessment and collect samples of the distal intestine to perform a fluorescent in situ hybridization procedure (FISH).

Fluorescent In Situ Hybridization
To analyze bacterial communities in the totoaba distal intestine, fluorescent in situ hybridization was performed following the methodology described by Hernandez & Olmos [34] with slight modifications in control and sample preparation.

Control Preparation
Bacterial samples of B. subtilis 9b and V. harveyi (ATCC 14126) were used as low G + C and γ-Proteobacteria controls, respectively. The strains were grown in specialized media until they reached the mid-exponential growth phase at 600 nm. Samples were fixed by adding 37% (v/v) filtered formaldehyde to a final concentration of 6% (v/v) and incubated at 4 • C for 1 h. A total of 1 mL of cells were collected by centrifugation at 12,000× g for 3 min. Pellets were washed twice with ice-cold phosphate-buffered saline (PBS; 120 mM NaCl, 2.7 mM KCl, 10 mM phosphate, pH 7,4), suspended in 1 mL of PBS and stored at −20 • C for future use; no longer than 2 months.

Samples Preparation
Intestine tissue samples were taken at the beginning and end of the experiment; 1 g of distal intestine was collected from five fish at the beginning of the study and from two per treatment at the end of the experiment, suspended in 5 mL of formaldehyde (10%) and mechanically homogenized. Aliquots of 500 µL were centrifuged at 19,800× g for 8 min; tissue pellets were washed twice with ice-cold PBS and suspended in 1 mL of PBS.
Samples and controls were serially diluted and 15 µL aliquots were spread onto 12-wells Teflon slides. Autofluorescence control was carried out by inoculating wells with reference strains without probes. Slides were allowed to dry at 37 • C, treated with ethanol/formaldehyde solution (90:10 v/v) for 5 min, rinsed twice in distilled water and dried again at 37 • C.

Sample Fixation
In situ hybridization of the controls and tissue samples was performed as described previously [34]. Briefly, 40µL of hybridization mixture [10X SET buffer (1× SET: 150 mM NaCl, 20 mM Tris-HCl (pH 7,8), 1 mM EDTA), 0.2% bovine serum albumin (w/v), 0.01% polyadenylic acid (w/v), and 11% of dextran sulfate (w/v)] containing 1 ng/l of probes were added to samples immobilized on slides, avoiding bubble formation. The slides were placed into a humidified chamber with pieces of 1X SET-saturated paper and incubated overnight in the dark at 37 • C. After hybridization, the slides were washed twice with 1X SET buffer for 20 min each. Slides were dried at 37 • C in the dark, and each well was overlaid with 5 µL of mounting fluid (10× SET buffer, 50% glycerol, 0.1% p-phenylenediamine HCl). Finally, a cover slip was placed, and samples were analyzed.

Microscopy
Fluorescence signals were analyzed using an Olympus microscope with MF filters (400-700 nm: 5FITC or 5(6)-ROX) and Plan Fluorite Universal objectives. The excitation source was a 50-W high-pressure mercury bulb. Digital images were processed with image Fiji/Image J software (version 1.52g, Java 1.80_172, U.S. National Institutes of Health, Bethesda, Maryland, USA).
Bacterial density from the distal intestine was estimated by digital image count, checked for normality, and when appropriate, transformed. For cell quantification, means were calculated from 10 randomly chosen fields for each sample [35]. Fields per well = well area picture area (7) where: well area = 19,635,000 µm; picture area = 203,970 µm.

Pathogens Resistance Assay
After completing the feeding trials, six fish from DCML, SPC30C, and SPC30Pro treatments were placed in tanks isolated from the system and half-filled with water. V. harveyi CAIM1508 pathogenic strain was grown in specialized media until reaching stationary growth phase at 600 nm. An amount of 10 mL were centrifuged at 12,000× g for 10 min, washed twice with cold PBS, and suspended in PBS to obtain a final concentration of 10 7 cells/mL. Clove oil was used as an anesthetic; fish were inoculated intraperitoneally with V. harveyi solution. Fish were placed in their tanks, and recovery was recorded by visualizing the swimming fish activity and feed intake post-infection.

Statistical Analysis
Collected data from growth performance were analyzed through a normality and homoscedasticity test. One way-analysis of variance (ANOVA) was performed to observe differences between treatments and their respective controls without probiotics. p-values < 0.05 were considered significant. Tukey's post hoc test was applied for data presenting normality, and for those data that did not present normality, analysis of multiple comparisons by Kruskal-Wallis was applied.  Table 2 describes the growth performance of fish fed with DCML, SPC30C, and SPC30Pro treatments, as well as with SPC60C and SPC60Pro. Results show 100% survival rate for diets with 30% substitution and 61.11% and 83.33% for diets with 60% substitution, respectively. Although fish fed with SPC30C and SPC30Pro do not show significant differences in their growth performance, SPC30Pro treatment supplemented with B. subtilis 9b probiotic strain produced higher FW, WG, SGR, TGC, and DFI values than SPC30C formulation (Table 2). Fish fed with 60% substitution diets presented lower growth performance than DCML and 30% substitution diets (Table 2); however, significant differences in most parameters were obtained when B. subtilis 9b probiotic strain was included in SPC60Pro formulation, compared to SPC60C without probiotic addition.

Treatment Effect on Organ Conditions
At the end of the feeding trial, three animals of each treatment were sacrificed, and their muscle condition, visceral fat content, as well as their liver, spleen, and distal intestine were measured and analyzed (Table 3). Fish fed with SPC30C and SPC30Pro diets produced firmer fillets than fish fed with DCLM, SPC60C, and SPC60Pro. In this sense, fish fed with SPC60C diet showed worse muscular conditions in the study; however, fish fed with SPC60Pro produced muscular conditions similar to animals fed with DCML commercial diet ( Table 3).
The visceral fat content was lower in diets with 30 and 60% substitution than in DCML commercial diets. In addition, the liver, spleen, and intestine were larger in fish fed with 30% substitution diets than in 60% substitution diets and DCML commercial diets (Table 3). Moreover, bigger spleen and intestine can be observed in fish fed with SPC60Pro than in fish fed with SPC60C treatment without B. subtilis 9b strain (Table 3).

Fluorescent In Situ Hybridization Evaluation
Intestinal samples were taken from fish at the beginning and end of the feeding trial to evaluate the capacity of B. subtilis 9b strain to modulate the intestinal microbiota; increasing low G + C Gram-positive bacteria and reducing γ-proteobacteria that include Vibrio species.  (Table 3). Moreover, bigger spleen and intestine can be observed in fish fed with SPC60Pro than in fish fed with SPC60C treatment without B. subtilis 9b strain (Table 3).

Fluorescent in Situ Hybridization Evaluation
Intestinal samples were taken from fish at the beginning and end of the feeding trial to evaluate the capacity of B. subtilis 9b strain to modulate the intestinal microbiota; increasing low G + C Gram-positive bacteria and reducing γ-proteobacteria that include Vibrio species. Figure 1 shows bacilli-like structures attached to an intestine fragment taken at the end of the assay from a fish fed with SPC30Pro diet. To know the modulation effect of B. subtilis 9b in T. macdonaldi intestinal microbiota throughout the trial, fluorescent in situ hybridization was carried out in fish fed with SPC30Pro, SPC30C, and DCML ( Figure 2). A low G + C Gram-positive bacteria increment To know the modulation effect of B. subtilis 9b in T. macdonaldi intestinal microbiota throughout the trial, fluorescent in situ hybridization was carried out in fish fed with SPC30Pro, SPC30C, and DCML ( Figure 2). A low G + C Gram-positive bacteria increment is observed in fish fed SPC30Pro compared to DCML and SPC30C. However, in SPC30Pro treatment γ-Proteobacteria decreased in number at the end of the trial. On the other hand, treatment with SPC30C without B. subtilis could not reduce the number of γ-Proteobacteria at the end of the experiment as SPC30Pro did. DCML commercial diets showed the lowest bacterial numbers at the end of the feeding trial ( Figure 2).  Intestinal microbiota modulation by B. subtilis 9b in T. macdonaldi fed with SPC30Pro diet can be observed in Figure 3. An increase in low G + C Gram-positive bacterial number can be observed at the end of the feeding trial (Figure 3b,c). However, γ-Proteobacteria number was reduced in the same treatment (Figure 3e,f). Thus, B. subtilis 9b can modulate intestinal microbiota of T. macdonaldi, which could induce beneficial effects observed in totoaba fed SPC30Pro treatment (Tables 2 and 3). Intestinal microbiota modulation by B. subtilis 9b in T. macdonaldi fed with SPC30Pro diet can be observed in Figure 3. An increase in low G + C Gram-positive bacterial number can be observed at the end of the feeding trial (Figure 3b,c). However, γ-Proteobacteria number was reduced in the same treatment (Figure 3e,f) intestinal microbiota of T. macdonaldi, which could induce beneficial effects observed in totoaba fed SPC30Pro treatment (Tables 2 and 3). Intestinal microbiota modulation by B. subtilis 9b in T. macdonaldi fed with SPC30Pro diet can be observed in Figure 3. An increase in low G + C Gram-positive bacterial number can be observed at the end of the feeding trial (Figure 3b,c). However, γ-Proteobacteria number was reduced in the same treatment (Figure 3e,f). Thus, B. subtilis 9b can modulate intestinal microbiota of T. macdonaldi, which could induce beneficial effects observed in totoaba fed SPC30Pro treatment (Tables 2 and 3).

T. macdonaldi Behavior after V. harveyi Infection
A total of six fish from DCML, SPC30C, and SPC30Pro treatments were infected with V. harveyi to evaluate T. macdonaldi behavior after pathogen inoculation. In the first two days, all fish showed reduced and erratic swimming, however, after the third day, fish fed with SPC30Pro showed a recovery, presenting more motility than the SPC30C control diet and DCML commercial diet (Table 4). Additionally, DCML and SPC30Pro treatments similarly improved feed intake after the third day; however, SPC30C showed delayed feed intake ( Figure 4).
SBP addition in feed formulations of carnivorous species such as T. macdonaldi have already been studied. In this sense, Fuentes-Quesada and collaborators [30] found that substitutions of up to 23% of SBM induce adverse effects in fish. In addition, Trejo-
SBP addition in feed formulations of carnivorous species such as T. macdonaldi have already been studied. In this sense, Fuentes-Quesada and collaborators [30] found that substitutions of up to 23% of SBM induce adverse effects in fish. In addition, Trejo-Escamilla and collaborators [31] found that T. macdonaldi can tolerate up to 34.17% of SPC before developing adverse effects. Therefore, authors reported that T. macdonaldi's low tolerance to SBP was mainly due to non-digestible carbohydrates and enzyme inhibitors present in SBM and SPC, respectively [30,31].
In this work, T. macdonaldi fed with SPC30Pro diet presented better growth performance and less visceral fat content than DCML commercial diet (Tables 2 and 3). Thus, considering the health status of animals fed with SPC30Pro, this formulation seems to be an alternative to the Skretting© formulation used in this work (Tables 3 and 4). Taking into account the larger visceral fat content found in animals fed with DCML diet, we speculate that the lipid concentration (12%) used to formulate this aquafeed could be responsible. In this sense, some reports indicate that high lipid concentration in aquafeeds could affect feed intake in carnivorous species [43]. High lipid concentration is principally used to feed cold water fish. Nevertheless, experiments with totoaba were developed at 25 • C; therefore, non-metabolized lipids were accumulated around the viscera. In addition, a 7% lipids concentration used in experimental diets was based on lipid content found in T. macdonaldi [44].
Fish fed with SPC60C and SPC60Pro presented lower growth and a decreased survival rate than SPC30C and SPC30Pro, respectively (Table 2). However, fish fed with SPC60Pro formulation showed 20% more survival and doubled weight gain than SPC60C without B. subtilis 9b strain; results that were not observed with SPC30C and SPC30Pro (Table 2). Therefore, we suggest that FM levels in SPC30Pro could be affecting B. subtilis behavior; maybe high levels of free nitrogen could be inhibiting Bacillus enzymes activity in SPC30Pro treatment [45,46]. In experiments developed with 100% of FM and B. subtilis 9b, T. macdonaldi presented less growth than its control without B. subtilis [32]. On the other hand, the increment of enzyme inhibitors in SPC60 formulations due to an increase in SPC could be affecting T. macdonaldi growth and survival (Table 2) [31]. These results suggest that B. subtilis 9b can degrade carbohydrates, lipids, and partially the proteins from SPC60Pro diet, however, it cannot eliminate all the enzyme inhibitors present in SPC60Pro formulation (Table 2). Thus, high levels of FM and SPC could be affecting B. subtilis 9b nitrogen homeostasis and T. macdonaldi proteases activity, respectively. Therefore, a new or complementary Bacillus strain with higher protease inhibitors degradation capacity must be found to obtain betters yields with 60% SPC substitutions. However, it is important to mention that SPC30Pro treatment produced the best results in this work.
Probiotics can affect muscle quality and the internal organs' conditions. Some authors found that Pengze crucian carps (Carassius auratus var. Pengze) fed with formulations supplemented with B. subtilis or B. cereus produced a firmer, chewy, cohesive, and adhesive meat compared to fish fed with basal feed without Bacillus inclusion [47,48]. In this work, fish fed with SPC30Pro showed firmer and chewy fillets than most diets tested. In addition, fish fed with SPC30Pro showed bigger and larger organs compared with other treatments. These results suggest that B. subtilis 9b could be inducing benefits in T. macdonaldi muscle consistency and in organ health status (Table 3).
Probiotics also can modulate the host microbiota and, in this way, improve their health status. Totoaba autochthonous microbiota mainly comprises Proteobacteria from the Vibrionaceae family [49]. In this work, using FISH methodology, metabolically active γ-Proteobacteria and low G + C Gram-positive bacteria were evaluated in T. macdonaldi treated with SPC30Pro, SPC30C, and DCML diets ( Figure 2). In this sense, B. subtilis 9b modulated T. macdonaldi intestinal microbiota by increasing low G + C Gram-positive bacteria amount and at the same time, it reduced the γ-Proteobacteria population ( Figure 2). These results agree with results obtained by Tachibana and collaborators, who reported that B. subtilis and B. licheniformis added to Nile tilapia diets modulated intestinal microbiota; reducing Proteobacteria and increasing Firmicutes population [50]. However, DCML treatment reduced the bacterial number at the end of the experiment, indicating that DCML formulations contain compounds that inhibit the development of both groups (Figure 2).
Bacillus species produce antimicrobial peptides against many gram-positive and gramnegative pathogen bacteria [51]. These AMP can directly kill pathogens or induce the host immune system to respond against microbial infections. In this work, fish fed with SPC30Pro diet produced a faster recovery against V. harveyi infection. However, fish fed with DCML and SPC30C presented a slower recovery (Table 4 and Figure 4). Challenges with B. subtilis 9b and V. harveyi using agar plates with marine medium, showed that B. subtilis has the capacity to limit V. harveyi growth (data not shown). Therefore, higher resistance observed in fish fed with SPC30Pro could be related to in vitro antimicrobial capacity showed by B. subtilis 9b strain. B. subtilis 9b fish immune system stimulation requires more research before any assumption. However, Bacillus has been shown to boost different aquatic organisms' immune system [22,38,[52][53][54].

Conclusions
T. macdonaldi fed with SPC30CPro and SPC60CPro feeds containing high levels of SPC and B. subtilis 9b for ninety days improved growth performance and enhanced health status compared with SPC30C and SPC60C treatments without B. subtilis 9b. Therefore, B. subtilis 9b probiotic strain had the capacity to increase tolerance to ingredients such as corn starch and SPC present in these formulations. Furthermore, B. subtilis 9b showed the capacity to modulate intestinal microbiota increasing low G + C Gram positive bacteria and lowering γ-Proteobacteria, which seems to increase V. harveyi resistance. Moreover, muscle and organ characteristics were also improved in treatments containing B. subtilis 9b strain. Therefore, these results open the possibility to formulate functional feeds with alternative and economical vegetable ingredients and B. subtilis 9b, without generating animal health problems.