Bacillus velezensis (named as BV007) was previously isolated from healthy L. vannamei intestines. The intestines were immersed in sterile phosphate buffered saline (PBS) (0.05 M, pH 7.4), shook, and eluted at 30°C for 30 min. Spread the eluate on LB nutrient agar (Beijing Land Bridge Tech. Co. Ltd.) plates with gradient dilution and incubated at 30°C for 24 h. After picking a single colony and streaking for purification, the strain 007 with a wide range of enzyme-producing abilities was obtained. The genetic characterization of B. velezensis was accomplished by using 16S rDNA sequence analysis, and the strain was deposited by the China General Microbiological Culture Center (CGMCC no. 20039).

In the diet preparation, the bacterial strains were inoculated in nutrient broth, collected in sterilized water, and adjusted to 10⁹ CFU/ml before feed production. The ingredients of the basal diet were measured to contain 41.35% crude protein and 7.67% crude lipid with fish meal, corn gluten meal, and soybean meal as the fundamental protein source whereas the main lipid sources were soybean oil and soy lecithin oil. Three experimental diets were supplemented with different levels of BV007: 10⁵ CFU/g (BV1), 10⁷ CFU/g (BV2), and 10⁹ CFU/g (BV3). The group without BV007 was served as a control (C). The bacterial amounts of experimental diets were also confirmed by the plate count method, and the colonies that emerged on the plates were counted. The four formulated experimental diets were extruded through a 1.2-mm die. The resulting pellets were dried at 25°C with the aid of an air conditioner. After drying, all diets were stored at 4°C until usage.

Healthy L. vannamei (body length: 3.0 ± 0.3 cm, body weight: 0.32 ± 0.8 g) were purchased from a local farm (Yantai, Shandong Province, China) and acclimated at 25–27°C and 30% of salinity for 2 weeks before processing. The shrimp were randomly dispersed into 12 fiberglass tanks (60 L), each containing 100 shrimp. The shrimp were fed four times daily (06:00, 12:00, 17:00, and 22:00) at 6% of body weight. The cumulative mortality of the shrimp was recorded daily. After 42 days of feeding, the total number of shrimp and the weight in each tank was quantified to calculate the survival rate (SR, in percentage), final body weight (FBW), final body length (FBL), plumpness index (PI), specific growth rate (SGR, %/d) according to the methods described previously (Velmurugan et al., 2015; Amoah et al., 2019b).

Hemolymph samples (eight replicates were tested in each treatment, and three shrimps were tested in each replicate [n = 24]) were individually withdrawn from the pericardial cavity of the shrimp using a 1-ml sterile syringe containing an anticoagulant (30 mM trisodium citrate, 0.34 M sodium chloride, and 10 mM EDTA at pH 7.55 with an osmolality adjusted to 780 mOsm/kg with glucose). The volume ratio of anticoagulant to hemolymph was 1:1. The harvested hemocytes were centrifuged at 4°C, 4,000 rpm/min for 10 min, collected, and adjusted to 10⁶ cells/ml with PBS (0.05 M, pH 7.4) for the phagocytic activity and respiratory burst assays. In addition, the hepatopancreas was collected for the assay of digestive, antioxidant, and immune-related enzyme activities, eight replicates were tested in each treatment, and three shrimps were tested in each replicate (n = 24). The middle intestines and hepatopancreas of each group were collected sterile into 4% paraformaldehyde and stored at 4°C for the analysis of microorganisms (three parallel assays).

The phagocytosis assay was adapted from the method of Delaporte (Delaporte et al., 2003) with minor modifications. Briefly, the hemolymph was mixed with 2.3% yellow-green FluoSpheres (diameter 2.0 μm, Polyscience, Eppelheilm, Germany), and incubated in the dark at 18 °C for 1 h with rotation. After incubation, the hemocytes were analyzed by flow cytometry (BD Accuri C6 Plus, BD Biosciences, America) using the FL-1 tunnel to detect hemocytes containing fluorescent beads. The phagocytosis rates were expressed as the percentage of hemocytes that engulfed three or more beads.

Respiratory burst activity was analyzed according to the protocols described by Kalgraff et al., 2011. Briefly, hemocytes (1 × 10⁶ cells/ml) were challenged by phorbol 12-myristate 13-acetate (PMA; final concentration of 0.1 μg/ml) for 10 min. Subsequently, dihydrorhodamine 123 (DHR 123) was added to a final concentration of 2 μg/ml and then incubated for 30 min prior to flow cytometry analysis. The flow cytometry analyses were done on a BD FACSCalibur flow cytometer equipped with a 15 mW 488 nm argon ion laser. Further data analyses were done using FCS Express 3 software (De Novo Software, CA, USA).

The digestive enzymes and immune enzymes were measured as described by Cao et al. (2012). Briefly, the digestive enzymes included α-amylase (AMS), trypsin (TRP), lipase (LPS), gltamic-pyruvic transaminase (GPT), glutamic oxaloacetic transaminase (GOT), and the immune enzymes included catalase (CAT), superoxide dismutase (SOD), acid phosphatase (ACP), and alkaline phosphatase (AKP). The enzymatic activities were quantified with corresponding commercial detection kits (C016-1-1, A080-2-2, A054-1-1, C009-2-1, C010-2-1, A007-1-1, A001-3-2, A001-3-1, and A060-2-1, Nanjing Jiancheng Bioengineering Institute, P.R. China) according to the instructions of the manufacturer.

Total RNA was isolated from hepatopancreas of different treatments using RNAiso (TaKaRa, Japan) reagent. cDNA synthesis was performed according to Promega M-MLV RT Usage information. The temporal expression profiles were carried out with an Applied Biosystem 7500 fast real-time PCR system (Applied Biosystems, MA, USA) using SYBR Green I (Biotek, VT, USA). The PCR amplification program was 95°C for 10 min, then 40 cycles of 95°C for 15 s, 60°C for 1 min, followed by 95°C for 15 s, 60°C for 1 min, and 95°C for 15 s. At the end of each PCR, a dissociation curve analysis of amplification products was performed to confirm the purity of the PCR product. The 2^−ΔΔCT method was used to analyze the expression levels (Livak and Schmittgen, 2001). All data were given in terms of relative mRNA expressed as mean ± SD (N = 6). The primers for qRT-PCR of genes for antioxidant-related genes (cat and sod), antibacterial peptides genes (crustin and penaiedin 3a), pattern recognition receptors (lgbp and lec), along with the housekeeping gene β-actin, are shown in Table 1.

Three similar-sized shrimp from each tank were sampled and dissected to obtain the midgut and the hepatopancreas (5 mm³). The fixed intestinal and hepatopancreas samples were dehydrated in a graded series of ethyl alcohol and embedded in paraffin. Then, the paraffin blocks were sliced transversely into 5-μm thick sections and stained with H&E. The villus length of intestinal slice and the cell type of hepatopancreas slice were measured by a light microscope with a computerized image system (Olympus, DP73, Tokyo, Japan).

The intestinal community of the micro-organisms was performed according to Suo (Suo et al., 2017) with slight changes. The total DNA of microbes in the intestine was extracted directly with the E.Z.N.A. Stool DNA Kit (Omega Bio-tek, Inc., GA, USA) (Xin et al., 2015). The amplification and sequencing of the hypervariable region (V4 + V5) of the bacterial 16S DNA gene was performed for the sequencing analysis and species identification. High-throughput sequencing was performed on the IonS5TMXL platform (Novogene, China). The sequencing reads were assigned to each sample according to the individual unique barcode. The sequences were analyzed with the QIIME software package (Quantitative Insights Into Microbial Ecology) and the UPARSE pipeline (Caporaso et al., 2011). The reads were first filtered and clustered into operational taxonomic units (OTUs) at an identity threshold of 97%. The alpha and beta diversity analyses were then performed as described by Amoah et al. (2019b).

Vibrio parahaemolyticus was used for the challenge experiment (Amoah et al., 2019b). At the end of the culture experiment, 30 shrimps from each tank were injected with V. parahaemolyticus (final concentration 2 × 10⁷ CFU). In each tank, 20 shrimps were used for cumulative mortality assay, which was recorded daily for 14 days. The other 10 shrimps were cultured for immunoassays and 1 week after bacterial challenge, both the hemocytes and hepatopancreas were sampled and analyzed as described above. In addition, each treatment group was continuously fed with the experimental diets. The cumulative mortality rate was calculated following the formula of Liu et al. (2017).

The experimental data are presented as the mean ± SE. The data were statistically analyzed for significance using a one-way ANOVA, and then the differences among the means at P < 0.05 were tested with Duncan's multiple range test using SPSS 23.0 (SPSS Inc., 2005, IL, USA).

The growth performance and survival of L. vannamei fed experimental diets are shown in Table 2. The results showed that FBL, FBW, PI, and SGR were significantly improved by the different levels of probiotics supplementation, whereas no significant differences were observed in SR among the control and probiotic diets.

After 42 days of feeding, there was a significant increase in the phagocytic activity of shrimp fed diets supplemented with B. velezensis as compared to the control group (Figure 1A). However, no significant increase in the rates of phagocytosis was found in either the probiotics-supplemented or non-supplemented shrimps after the V. parahaemolyticus challenge (Figure 1B). Compared with the control group, significantly reduced reactive oxygen species (ROS) production was observed in the BV3 group, but not in the BV1 or BV2 groups (Figure 1C). In contrast, the production of ROS by hemocytes decreased statistically in the BV1, BV2, and BV3 groups when shrimp were challenged with V. parahaemolyticus (P < 0.01; Figure 1D).

As shown in Table 3, the specific activity of the α-amylase enzyme was enhanced in B. velezensis-supplemented groups. As compared with the control group, the AMS activity was enhanced significantly in shrimp fed BV1, BV2, and BV3 diets with the highest activity observed in the BV2 group (P < 0.01). However, no significant differences were observed in TRP or LPS activities among any of the groups. In addition, the GOT and GPT activities were significantly reduced in the BV3 and BV1 groups as compared with the control group, respectively (P < 0.05).

The activities of ACP showed no differences among the tested groups after being challenged with V. parahaemolyticus (Table 4). Similarly, no significant difference in AKP activity was detected among the BV1, BV2, and BV3 groups. However, the AKP activity was enhanced significantly after V. parahaemolyticus challenge among the tested groups (P < 0.01; Table 4). Both the CAT and SOD activities were enhanced in B. velezensis-supplemented groups and in the groups challenged with pathogenic bacteria (Table 4). The highest SOD and CAT activities were observed in the BV2 (P < 0.01; Table 4) and BV3 (P < 0.01; Table 4) groups.

As revealed in Figure 2, the feeding dosage of B. velezensis had a significant effect on the expression levels of immune-related genes. The expression levels of cat were significantly induced in BVs groups (P < 0.05, Figure 2A). After V. parahaemolyticus challenge, the cat transcripts in BV1 group were significantly higher than other groups (P < 0.05, Figure 2A). As concerned to sod expression, the higher expression levels were observed in BV2 and BV3 groups compared with C and BV1 groups, while the levels were significantly upregulated in BV1 and BV2 groups after the bacterial challenge (P < 0.05, Figure 2B). No significant changes in the expression levels of antibacterial peptides (crustin and penaeidin 3α) were observed among C, BV1, BV2, and BV3 groups. After the bacterial challenge, the expression levels of crustin and penaeidin 3α in BV2 were significantly higher than the other groups (P < 0.05, Figures 2C,D). In addition, the expression levels of lgbp and lec were highly enhanced regardless of whether the shrimp were bacterial challenged, especially in BV1 and BV2 groups (P < 0.05, Figures 2E,F).

The intestinal and hepatopancreas morphometric parameters were presented in Figure 3 and Table 5. The liver tubule of hepatopancreas in the BV addition groups are arranged relatively tightly and significantly generally star-shaped or polygonal (P < 0.05), while the lumen of the control group is mostly circular. Besides, the control group distributed many differentiated E cells (embryonic cells), while the BV addition group distributed more B cells (secreting cells), F cells (fibrocyte), and R cells (absorbing cells). The shrimps from all dietary treatments had an intact epithelial barrier with extensive mucosal folds and abundant Microvilli. The mucosal structure of the control group was slightly loose, the length of intestinal villi was significantly shorter (P < 0.05), and the number of mucosal epithelial cells was less. The intestinal tissue structures of the BV addition groups were relatively complete, and the length of the intestinal villi was increased and closely connected with the intestinal wall.