Yeast Production and Processing.
The yeasts C. jadinii and B. adeninivorans were cultivated in a company demonstration plant at 200 L scale (Biorefinery Demo, Borregaard® AS, Sarpsborg, Norway), using a medium composed of enzymatic hydrolyzates of pre-treated spruce wood (Picea abies)54 and hydrolyzates of chicken by-products (Norilia®, Oslo, Norway), as described in Lapeña, et al.22B. adeninivorans was cultivated for 18.5 h in batch mode, while C. jadinii was cultivated for 42 h in fed-batch fermentation mode with the addition of wood sugars, urea, KH2PO4, CaCl2.2H2O, MgSO4.7H2O and NaCl (See supplementary Fig. S1). W. anomalus yeast was cultivated in 20 L scale according to the protocols described in Lapeña, et al.33 For washing, the yeasts were separated by centrifugation and re-suspended in the same volume of 7 oC deionized water in a 30 L EINAR bioreactor system (Belach Bioteknik, Sweden), equipped with a helical impeller. The washed yeasts were then again centrifuged to obtain yeast creams with 12.5%, 5.5% and 15% dry matter contents for C. jadinii (CJ), B. adeninivorans (BA) and W. anomalus (WA), respectively. Half of these microbial biomasses were dried and heat-inactivated by spray-drying using a SPX 150 MS (SPX Flow Technology, Denmark AS) spray-dryer with inlet and outlet temperatures of 180 oC and 80 oC, respectively. The spray-dryer was fitted with a co-current nozzle and the pump speed was set to auto and stabilized at around 35% of maximum speed of the pump. The other half of the yeast creams underwent autolysis by incubating the creams at 50 oC for 16 h in a 30 L EINAR bioreactor system, with constant stirring at 50 rpm using a helical impeller, followed by spray-drying using the same conditions as for the untreated yeast. Dried yeasts were kept at 4 oC until use.
Formulation and Production of Fish Feeds.
Nine experimental diets were produced in this experiment. The diets were as follows: a fishmeal (FM) control; a diet with 40% SBM as a positive control; 6 treatment diets containing 40% SBM and 5% yeast ingredients [inactivated CJ (ICJ), autolyzed CJ (ACJ), inactivated BA (IBA), autolyzed BA (ABA), inactivated WA (IWA) and autolyzed WA (AWA)], respectively. An extra control diet containing 40% SBM and 5% of a reference preparation of C. jadinii (ICU) already described for its ability to counteract enteritis15 was also used in this trial. The feed formulation is as presented in supplementary Table S1. The diets were formulated to have a similar ratio of digestible protein to digestible energy, and to meet the nutrient requirements of Atlantic salmon as recommended by NRC55. To meet fish amino acid requirements, crystalline lysine and methionine were added to the diets due to the high inclusion of plant-based ingredients. All dry ingredients were mixed in a Moretti Forni mixer (Spiry 25, Mondolfo, Italy). Gelatin was mixed in cold water and heated up to 60 °C in a microwave oven before mixing with dry ingredients and fish oil using the same mixer as above. The mash was cooled down to room temperature prior to cold-pelleting using a P35A pasta extruder (Italgi, Carasco, Italy). The pellets were dried in small experimental dryers at approximately 60 °C drying temperature and stored at 4 °C prior to feeding.
Fish Management and Feeding.
The fish experiment was conducted at the Fish Laboratory of Norwegian University of Life Sciences (NMBU, Ås, Norway), which is an experimental unit approved by the National Animal Research Authority, Norway (permit no. 174). The experimental procedures were performed in accordance with the institutional and national guidelines under the applicable laws and regulations controlling experiments with live animals in Norway (regulated by the “Norwegian Animal Welfare Act” and “The Norwegian Regulation on Animal Experimentation” derived from the “Directive 2010/63/EU” on the protection of animals used for scientific purposes). The study was carried out in compliance with the ARRIVE guidelines.
In total, 1215 Atlantic salmon fry with an average start weight of 5.71 ± 0.05 g were sorted, batch weighed and randomly distributed into 27 fiberglass tanks (80 L) equipped with automatic feeders. Each tank was randomly stocked with 45 fish. Each diet was fed to triplicate tanks, 20% in excess based on feed consumption in each tank. Feeding was done twice a day with automatic feeders, and uneaten pellets were collected after each feeding from the outlet water settling on a screen for each tank. Daily feed intake was calculated from the dry weight of the feed given and the dry weight of recovered uneaten pellets, adjusted for feed recovery rate from fish tanks. Feeds were kept under refrigerated conditions (4 °C) throughout the experiment. Fish were exposed to a 24 h light regime and recirculated freshwater with an average temperature of 15.0 °C. The water flow was standardized to about 6 L min− 1, and the oxygen content of the outlet water was kept within 8.2–10.1 mg L− 1. The experiment lasted for 37 days, after which the fish were counted and grouped weighed to estimate the growth performance.
Sampling Procedure for Fish Tissue.
For tissue sampling, six fish per tank were randomly selected, anesthetized with metacaine (MS-222™; 50 mg L− 1 water) and killed with a gentle blow to the head. The individual body weight of each fish was recorded and included in the total tank mean. Distal intestine and pyloric caeca tissues were collected from each fish and further processed, as described below. The distal intestine was opened longitudinally, the content was removed and the tissue was carefully divided into two parts: one part was fixed in 10% phosphate-buffered formalin for 24 h before storage in 70% ethanol until further processing for histological analysis; the second part was immediately snap-frozen in liquid nitrogen and stored at -80 °C for indirect enzyme-linked immunosorbent assays (ELISA). Pyloric caeca were treated in the same way as the distal intestine samples for histological analysis.
Morphometric and Histological Examination of Fish Tissues.
Formalin-fixed distal intestine and pyloric caeca samples were dehydrated in ethanol, equilibrated in xylene and embedded in paraffin using standard histological techniques. Longitudinal sections of approximately 6 µm in thickness were prepared. The sections were stained with hematoxylin, eosin and Alcian blue 8 GX. Changes in villi length were captured using a DMLS light microscope (Leica Microsystems, Wetzlar, Germany) equipped with a Leica E3 digital imaging camera and LAS EZ v4.9 software. Randomly selected villus of 18 distal intestine tissues from each dietary group (at least 80 measurements per group) was measured from the stratum compactum to the tip of the fold by ImageJ software. For histological evaluation, changes associated with intestinal tissues were blindly evaluated with a focus on the characteristic changes known for SBMIE in Atlantic salmon.9 The histological scores were obtained through a semi-quantitative scoring system measuring changes in three morphological parameters: loss of supranuclear vacuoles in absorptive enterocytes; widening of lamina propria in mucosal folds; and increase of connective tissue between the base of folds and stratum compactum. Each parameter was given a score of 1–5, where 1–2 represents normal morphology; 3–4 mild and moderate enteritis; and 5 for severe enteritis (Supplementary Fig. S2a-b). To measure changes associated with pyloric caeca, the longitudinal enterocytes area was selected (image 20x) and measured from the base to the apex. Measurement of enterocyte height was performed using the Easy Scan Software. The total number and average mucous cell size in the caeca mucosal area were measured using the ImageJ software. The number of mucous cells was counted per 1 mm2 of the mucosal area (Supplementary Fig. S2c-e).
Indirect ELISA of Distal intestine Tissues.
Immunological parameters were analyzed using the distal intestine samples by indirect ELISA56. Briefly, distal intestine samples from nine fish per treatment were homogenized using metal beads and lysis buffer (Tris 20 mM, NaCl 100 mM, Triton X-100 0.05%, EDTA 5 mM, and protease inhibitor cocktail 1x, pH = 7.2). Subsequently, the homogenate was centrifuged at 12000 x g for 25 min at 4 °C. The supernatant containing soluble proteins was stored at -20 °C until use. The protein concentration was quantified using the BCA protein assay kit (Thermo Fisher Scientific) following the manufacturer’s instructions. Then, each sample was diluted in carbonate buffer (NaHCO3 60 mM, pH 9.6) and seeded (in duplicate) in a 96-well plate (Maxisorp, Thermo Fisher Scientific) at 50 ng µL− 1 (100 µL) for overnight incubation at 4 °C. After blocking with 5% Block solution (Biorad), diluted in PBS, for 2 h at 37 °C, the plates were incubated for 90 min at 37 °C with the first antibody (Supplementary Table S2). Then, the second antibody-HRP (Thermo Fisher Scientific), at 1:7000 dilution, was added, followed by incubation for 1 h at 37 °C. Finally, the chromogenic substrate 3,3′,5,5′-tetramethylbenzidine (Invitrogen) was added (100 µL) followed by incubation for 30 min at room temperature. The reaction was stopped with 50 µL of 1 N sulfuric acid and absorbance at 450 nm was measured using a Spectramax microplate reader (Molecular Devices).
Chemical Analysis of Yeast and Fish Feeds.
The yeasts and diets were analyzed for dry matter by drying to constant weight at 105 °C, for crude protein using Kjeldahl nitrogen (N × 6.25) (Commission dir. 93/28/EEC), for crude lipid by Accelerated Solvent Extractor (ASE200, Dionex, California, USA)(ISO 6492) and for ash by incineration at 550 °C (Commission dir. 71/250/EEC). Gross energy content was determined using an adiabatic bomb calorimeter (Parr 1281; Parr Instruments, Moline, IL, United States), according to ISO (1998).
Calculations for Fish Growth Parameters.
The average biomass gain, feed conversion ratio (FCR), and specific growth rate (SGR) were calculated according to the equations presented in Agboola, et al.57. Briefly, the biomass gain was calculated as the difference between the average final weight and the average initial body weight of fish per tank. The FCR was calculated as the ratio between average feed consumption per day and average biomass gain per day. The SGR was calculated as logarithm differences between average final and initial weight of fish divided by the experimental duration.
Morphology and Ultrastructure of the Yeast Cells.
The ultrastructure of yeast cells was examined using a scanning electron microscope (SEM) and a transmission electron microscope (TEM). For each yeast, the SEM and TEM samples were taken before and after the autolysis process, i.e. before spray drying. Three yeast samples per treatment were prepared according to the procedure described in Straume, et al.58 for SEM and TEM imaging. Samples for SEM were coated with Pt-Pl and examined in a Zeiss EVO 50 EP (Zeiss International, Germany) scanning electron microscope at an accelerating voltage of 15 kV in the secondary emission mode. The sections for TEM were examined in a FEI MORGAGNI 268 (FEI, USA) transmission electron microscope, and photographs were recorded with a VELETA camera. The imaging was performed at the Imaging Centre, Faculty of Biosciences, Norwegian University of Life Sciences. The cell wall thickness was obtained by measuring the length of five random locations on the cell wall surface of twenty TEM micrographs of each yeast using ImageJ.
Cell Surface Properties of Yeast as Determined by AFM.
Atomic force microscopy (AFM) measurements were done following the protocol described in Schiavone, et al.32 and Schiavone, et al.50. Experiments were carried out with a Nanowizard III atomic force microscope (Bruker-JPK Instruments). The spring constants of each MLCT cantilever (Bruker) were determined using the thermal noise method59 and were found to be in the range of 10–20 pN nm− 1. Yeast sample preparation was done by re-suspending the dry yeast mass in sodium acetate buffer (18 mM CH3COONa, pH 5.2, 1 mM CaCl2 and 1 mM MnCl2) and immobilized on polydimethylsiloxane (PDMS) stamps, as described in Dague, et al.60. 100 µL of yeast suspension was deposited on the PDMS stamps by convective/capillary assembly. Using bare AFM tips, AFM heights (expressed in nm) were recorded in Quantitative Imaging mode61 with a maximal force of 1 nN, at 20 °C in buffer solution. Elasticity of cells was determined from 3072 force curves recorded in force volume mode at an applied force to the surface of 0.5 nN and speed of approach and retraction of 2 µm.s− 1. Elasticity histograms were generated by analyzing the force-distance curves according to the Hertz model described in Schiavone, et al.52, with an indentation of 50 nm and considering a conical tip geometry with half-opening angle α of 0.31 rad and a Poisson ratio ν of 0.5.
To probe cell surface polysaccharides, AFM tips were functionalized with Concanavalin A (ConA) from Canavalia ensiformis (Sigma-Aldrich, L7647) via a silicon nitride dendritip as described in Jauvert, et al.62. To analyze the stretching of polysaccharides at the surface of the cell, elongation forces were stretched using the worm-like chain model introduced in Bustamante, et al.63, which describes the polymer as a curved filament. The contour length from this model represents the length of mannoprotein unfolded. At least three cells were analyzed for each treatment, representing a total of 3072 force curves for each treatment. The force curves were analyzed with the JPK data processing software (JPK BioAFM, Bruker Nano, Germany). All specific adhesion peaks were considered for the histograms, which were generated using the Origin® 2020 software (OriginLab Northampton, MA, USA). A Gaussian distribution curve fitted on the histogram was used to determine the maximal values of Young modulus, length of mannoprotein unfolded and adhesion force for each yeast group.
Quantification of Yeast Cell Wall Polysaccharides.
The total polysaccharide content of the yeast cell wall was estimated without prior cell wall isolation according to the protocol described by François64. Briefly, the yeast samples were hydrolyzed with sulphuric acid and the released sugar monomers (mannose, N-acetylglucosamine and glucose) were quantified by high-performance anion-exchange chromatography with pulsed amperometric detection as described in Dallies, et al.65 and Hansen, et al.23. The content of β-glucan in the yeast samples was determined using a Megazyme® kit (reference K-YBGL) and α-glucan was calculated as the difference between total glucan and β-glucan.
Immunofluorescence Analysis of Yeast for Determining Mannan Specificity for ConA.
Approximately 200 mg of each spray-dried yeast was fixed with 10% formalin for 30 min at room temperature in Eppendorf tubes. Thereafter, the sample was centrifuged at 1000 x g for 5 min at 4 °C and re-suspended in PBS. For fluorescence detection of mannan with ConA lectin, the sample was blocked for 1 h at room temperature with PBS containing 1% bovine serum albumin. Subsequently, the sample was incubated with 5 mg mL− 1 of ConA-conjugated FITC (Sigma-Aldrich) for 1 h at room temperature in the dark. The samples were then gently layered on slides and allowed to dry for 10 min, before mounting in the Vectashield Medium (Vector Lab). Between all the steps of this procedure, the samples were washed in PBS. The slides were analyzed using a Zeiss LSM800 confocal microscope (Zeiss International, Germany).
Statistical Analysis.
Fish performance, morphometric, histological and immune parameters were analyzed using the SPSS statistical software package version 26 (IBM Institute, Armonk, NY, USA). Fish performance, morphometric and immune response data were tested for treatment effects using one-way ANOVA. Significance difference (P < 0.05) between means for fish performance and morphometric data were detected using the Tukey HSD test, whereas, for immune response parameters, Dunnett’s multiple comparison test was used for detecting significant differences. Data from morphometry measurements (villi length) was tested for normality by the Shapiro-Wilk test and homogeneity of variance using Levene’s test. Data from the histological evaluation were analyzed using a non-parametric Kruskal-Wallis test by ranks followed by Dunn’s multiple comparison test. Significance was set at P < 0.05. The tank effect was considered for all parameters and found to have no influence on the statistical analyses. Correlation coefficients between the diets using five immune markers were examined using corrplot package in R. Likewise, correlations between dietary intake of yeast cell wall components and immune markers were determine using the same R package. Also, the correlations between cell wall components and AFM data were examined using the corrplot package in R (CRAN: http://cran.r-project.org/package=corrplot).