Here, we obtained six formulated diets with various pre-and probiotic concentrations, as indicated by Abidin et al. (2022), all of which led to differential growth performances (Supplementary Table 1). The following criteria were maintained while culturing the above samples: 18 75-L aquariums were randomly filled with 180 healthy juvenile white shrimp (2.04 ± .05 g), and the shrimps were divided into the six diet groups, with three replications per group. The shrimps were raised in a recirculating system with 50% water exchange daily and were fed with the experimental diets three times a day at 5% of their body weight for the course of the growth trial, which lasted 60 days. Every 2 weeks, the diet amount was changed based on the shrimp’s weight. The following water quality parameter were used: temperature = 28°C–29°C, oxygen content > 6 mg/L, salinity = 30–33, and pH = 6.8–7.6 (to preserve the water’s purity).

After disinfecting the shrimp’s surface with 70% ethanol, three shrimps were randomly selected from each group. Next, by using sterile tools, we aseptically dissect the intestines and collected gut content. Finally, by using the Quick-DNA Fecal/Soil Microbe Miniprep kit (Zymo Research, USA), microbial DNA was extracted from gut content.

By using universal primers 341F and 805R (which have been used for investigating the microbiota of white shrimp), we performed amplicon sequencing of the V3–V4 region of the bacterial 16S rRNA gene. Illumina NovaSeq 6000 (Illumina, USA) was used for sequencing, and the Illumina TruSeq DNA Library kit was used to create a barcoded library. The sequencing data were uploaded to the NCBI database; the SRA accession numbers are SRR22923665–SRR22923684.

Analysis was performed in accordance with QIIME2’s (Quantitative Insight into Microbial Ecology) guidelines (v2019.10) (Bolyen et al., 2019). By using in-house scripts, the paired-end reads were overlapped to build the sequences. The sequence pre-processing was done using FastQC. The data were then demultiplexed by an internal script and trimmed by Trimmomatics (v0.39) for quality assurance. Cutadapt (v3.3) was then used with these guidelines: read length ≥ 150 bp and default error rate of 0.1. Preprocessing was then carried out using DADA2 (v1.12), which included eliminating chimera to filter amplicon sequence variants (ASVs), combining paired-end reads, and filtering out noisy sequences (denoising). By using the Usearch tool, the candidate sequences were grouped into operational taxonomic units (OTUs) based on a 97% sequence similarity and analysis using principal coordinates. By using the vegan function in R (version 3.4.4; https://www.r-project.org/), principal coordinate analysis (PCoA) plots based on unweighted UniFrac metrics were used to assess beta diversity. Through dimension reduction based on Euclidean and other distances, the probable principal components that could influence the variation in sample community composition were identified. A number of statistical analysis indicators are included in the alpha diversity package to quantify species diversity and abundance in the surrounding ecosystem. The indices, Chao and Ace (http://www.mothur.org/wiki/Chao and http://www.mothur.org/wiki/Ace), were used to determine community richness. Shannon (http://www.mothur.org/wiki/Shannon) and Simpson (http://www.mothur.org/wiki/Simpson) indices were employed to calculate community diversity. Predictions of the gut microbiota’s functional potentials were made using the PICRUSt2 (phylogenetic investigation of communities by reconstruction of observed states, v2.1.0-b) pipeline. Using the picrust2_pipeline.py software, functional profiles were predicted, creating a table of KEGG orthologs (KOs). By using the KEGG Mapper, we reconstructed KEGG reference categories (KEGG level 1) and modules (KEGG level 2) according to the KO annotations.

We prepared 20 libraries from the 6 groups (n = 3–4 group) after sequencing on Illumina NovaSeq 6000 platform and obtained 2,355,035 raw sequences with an average of 117,751 sequences per sample ( Table 1 ). Next, 2,136,094 high-quality sequences were selected after quality filtering and were used to create OTUs. Table 1 provides a summary of the raw and filtered sequences acquired at each data processing stage.

The highest and lowest numbers of sequences were generated from group A and F, respectively. Good’s coverage revealed that >90% of all OTUs were obtained for various treatments, indicating that the sequencing effort accurately reflected most bacterial communities in the group F. Consequently, at the obtained sequencing depth, the rarefaction curves also indicated outstanding bacterial community resolution ( Supplementary Figure 1 ). After applying frequency filters, we obtained 490 OTUs, of which 97.51% were shared by all six groups, and all OTUs shared 97% similarity. Of these 490 OTUs, only 374 could be assigned to putative bacteria under various taxonomical categories.

Taxonomic annotation corresponding to the OTUs was created to illustrate bacterial community structures and diversities. Groups A, B, C, D, E, and F demonstrated 135, 121, 191, 186, 171, and 318 OTUs, respectively; group F demonstrated the highest OTU number and diversity. The 374 OTUs were classified into the domain bacteria under 23 phyla (Acidobacteria, Actinobacteria, Armatimonadetes, Bacteroidetes, Chlamydiae, Chloroflexi, Cloacimonetes, Cyanobacteria, Dadabacteria, Deferribacteres, Deinococcus-Thermus, Dependentiae, Epsilonbacteraeota, Firmicutes, Fusobacteria, Nitrospirae, Patescibacteria, Planctomycetes, Proteobacteria, Rokubacteria, Spirochaetes, Tenericutes, and Verrucomicrobia), 41 classes, 94 orders, 151 families, and 250 genera ( Supplementary Table 2 ).

Of the aforementioned phyla, five were the prevalent core phyla because they represented >1% of all sequences. Proteobacteria represented 38% of the total sequences in white shrimp, followed by Bacteroidetes (30%), Tenericutes (17%), Firmicutes (13%), Patescibacteria (1%), and others (1%) ( Figure 1A ). The most frequent bacterial genus was also revealed by the association between the 30 most abundant genera ( Figure 1B ).

Through treatment-wise sampling, which indicates the microbial community richness and variety ( Figure 2 ), we determined the diversity of the bacterial community and specific taxa within the groups. The results indicated that group F had much higher OTU richness and more species than the other groups. Moreover, group F demonstrated much higher species richness and evenness than did the other groups, as measured using the Shannon index. A strong correlation was also noted between the gut bacteria and shrimp’s body weight. The beta diversity of the bacterial communities in the white shrimp gut from the six groups was examined through PCoA ( Figure 3 ). A total of 39% of the variation was explained by the first two components (PC1, 21%; PC2, 18%). Moreover, groups D and F had a broader dispersion than the other groups, which indicates the significant variations among the shrimp groups. Our analysis of similarities (ANOSIM) revealed no significant difference among the six groups ( Table 2 ). The results of the Multiple Response Permutation Procedure (MRPP) results were comparable to those obtained using ANOSIM ( Table 2 ).

Bacterial phylum compositions in the white shrimp gut in all six groups are shown in the Figure 4A . The dominant phyla in group A were Bacteroidetes (34.79%) and Proteobacteria (29.52%) representing approximately 65% of the total gut microbiota. Moreover, groups B, C, and E demonstrated an abundance of Proteobacteria (37.15%, 63.32%, and 43.83%, respectively) and Bacteroidetes (29.62%, 21.83%, and 28.32%, respectively). Compared with the higher growth groups, groups D and F shows an abundance of slightly different bacterial phyla. Group F shows an abundance of Bacteroidetes (44.17%) followed by that of Proteobacteria (38.05%). Moreover, group D shows an abundance of Tenericutes (29.22%), followed by that of Firmicutes (26.24%; Figure 4A ).

In total, 250 genera were identified from all six groups. Group A demonstrated the highest abundance of Spongiimonas (Flavobacteriaceae bacterium UJ101), followed by Vibrio, Candidatus bacilloplasma, Firmicutes bacterium ZOR0006, and Fusibacter. Group B demonstrated the highest abundance of C. bacilloplasma, followed by Vibrio, Spongiimonas, Photobacterium, and Firmicutes bacterium ZOR0006. Group C had an abundance of Vibrio followed by Spongiimonas, C. bacilloplasma, Firmicutes bacterium ZOR006, and Ruegeria. Group E had the highest abundance of Vibrio followed by C. bacilloplasma, Spongiimonas, Photobacterium, and Carboxylicivirga. Group F exhibited the highest abundance of uncultured Flavobacteriaceae bacterium, followed by Photobacterium, Vibrio, C. bacilloplasma, and Spongiimonas. Finally, group D demonstrated the highest abundance of C. bacilloplasma, followed by Peddiococcus, Vibrio, Spongiimonas, and Carboxylicivirga ( Figure 4B ). Significant between-group differences were found in the abundance of the pathogenic Vibrionaceae bacteria; the occurrence was higher in groups A, B, C, and E (slow-growth groups) but lower in groups D and F (the high growth groups). Three novel Vibrionaceae genera Vibrio fortis, Enterovibrio, and uncultured bacterium were found only in the slow-growth group and group A ( Figure 5A ). In addition, a decline was recorded in the abundance of other pathogenic bacteria in groups D and F (the high-growth groups; Figure 5B ).

To identify the unique gut microbiota among five groups and one control diet group, species abundance of the individual treatment group was first characterized at the taxonomic level of the genus so as to reflect the white shrimp’s adaptation to the treatments ( Figure 2 ). In general, the larger the treatment plot, the higher is the shrimp gut capacity to accommodate different types of microbes. Venn diagrams demonstrating the unique phyla and genera and those shared between the groups are shown in Supplementary Figures 1A, B ). Venn diagram of OTUs for phylum shows at least 8 out of 23 phyla shared between all the groups. Moreover, Fusobacteria was unique to group A; Dadabacteria and Dependentiae were unique to group C; and Armatimonadetes, Acidobacteria, Deferribacteres, Nitrospirae, Chloroflexi, Cloacimonetes, and Deinococcus-Thermus were unique to group F ( Supplementary Figure 2A and Supplementary Table 3A ).

Similarly, as shown in the Venn diagram, 51 genera were shared by all groups but at different abundance levels ( Supplementary Figure 2B and Supplementary Table 3B ). Moreover, groups A, B, C, D, E, and F had 26, 10, 33, 35, 18, and 133 unique genera, respectively.

Our KEGG analysis give rise to six functional categories at level 1 across all the groups: cellular processes, metabolism, organismal system, human diseases, genetic information processing, and environmental information processing. These six categories could be subdivided into 44 functional categories at level 2. The top 10 level 2 functional categories were the biosynthesis of other secondary metabolites, signal transduction, endocrine system, metabolism of terpenoids and polyketides, lipid metabolism, carbohydrate metabolism, amino acid metabolism, glycan biosynthesis and metabolism, and metabolism of cofactors and vitamins. Then, by using online SRPlots (http://www.bioinformatics.com.cn/en?p=1), statistically distinct aspects of functional categories based on KEGG (level 2) were analyzed between the six diet groups. Group A had a higher abundance of infectious diseases, namely, bacterial and parasitic; endocrine system; and folding sorting and degradation. These groups were demonstrated to have a significantly lower abundance in groups F and D. In particular, group D showed an increase in abundance of carbohydrate metabolism, lipid metabolism, transcription, nucleotide metabolism, membrane transport, etc. ( Figure 6A ).

The 44 level 2 functional categories could be subdivided into 346 different level 3 functional groups by using the KO terms. The most commonly occurring functional category, namely, biosynthesis of other secondary metabolites, could be divided into 22 pathways under 22 KO terms: KO00525, KO009,65, KO00232, KO00332, KO00331, KO00944, KO00941, KO00966, KO00901, KO00943, KO00950, KO00261, KO00524, KO00401, KO00311, KO00405, KO00940, KO00333, KO00404, KO00945, KO00521, and KO00960. The functional category of signal transduction could be divided into 20 pathways showing considerable differences among groups A, D, and F ( Figure 6B ). Level-3 KEGG pathway assessment with reference to signal transduction revealed the upregulation of the AMPK, MAPK, and P13k-Akt signaling pathways; peroxisome; lysozyme; and ferroptosis in diet group D, whereas, these pathways were downregulated in group A (control).

We observed that the structure and composition of gut bacteria were related to white shrimp growth, emphasizing the essential role of gut bacteria in shrimp growth. Across all gut content samples, we identified at least 23 phyla, out of which Proteobacteria dominates the bacterial phyla with occurrence of 38%, followed by Bacteriodetes 30%, Tenericutes 17%, Firmicutes 13%, Patescibacteria 1%, and others. Several authors have noted that Proteobacteria predominated in the shrimp gut. On the other hand, He et al. (2020); Zhang et al. (2019), and Zoqratt et al. (2018) have reported that Proteobacteria represent 44.12%, 38.94%, and 50%–80% of all shrimp gut bacteria. Amin et al. (2022) reported that in a white shrimp culture in Indonesia, 53.99% of the gut bacteria were Proteobacteria. Similarly, Gao S. et al. (2019) and Hasyimi et al. (2020) reported the predominance of Proteobacteria in the gut of white shrimp raised in Indonesia with Bacteroides and Firmicutes being the second and third most prevalent.

Fan J. et al. (2019) reported a relatively low prevalence of Proteobacteria (40.83%) in the gut of white shrimp raised in China; nevertheless, the bacterial species remained the predominant gut microbe, followed by Bacteroidetes (19.96%), Verrucomicrobia (8.26%), Firmicutes (6.17%), and Actinobacteria (1.59%). Finally, Huynh et al. (2019) also noted that Proteobacteria were predominant in the white shrimp gut in both the indoor (86.6s%) and outdoor (51.8%) pond culture systems.

The shrimp gut bacterial population appears to vary genus-wise and be influenced by a variety of factors, such as shrimp age, life stage, feeds, culture system, and geographical location (Fan J. et al., 2019; Fan L. et al., 2019; Gao S. et al., 2019; Huynh et al., 2019; Hasyimi et al., 2020). The current findings showed 250 bacterial genera in the total shrimp gut content. Of these genera, only 30 were considered to be included in the core white shrimp gut microbiome because of their frequency in all samples, out of which Vibrio, Candidatus bacilloplasma, Pseudoalteromonas, Donghicola, Agarivorans, Sulfitobacter, Rugeria, Enterobacter, Shimia, etc. show a higher abundance throughout all the white shrimp gut content. These results are remarkably comparable to those of Gao S. et al. (2019) [Vibrio (31%), Photobacterium (23.9%) and, C. Bacillophora (7.6%)]. Hasyimi et al. (2020) earlier reported several dominating genera from shrimp grown in Indonesia, with the five most prevalent genera being Nocardiodes, Neptunomonas, Spongiimonas, Desulfopila, and Bryobacter. These data strongly imply the presence of a wide range for the proportions of different genera of gut bacteria.

In particular, some genera including Photobacterium, Spongiimonas, and Neptunomonas were also frequently reported as the predominant genera in white shrimp gut.

Here, we noted that groups D and F demonstrated a high bacterial diversity based on the Shannon index and species richness index. In comparing with the high growth group, group D shows a fewer diversity and yet shows a high growth rate. These findings might imply that particular community of microbes with high diversity and species richness are preferable, particularly in terms of growth rate. This is in contrast with the results of Lee et al. (2022): less diversity was associated with high growth performance. Daniels et al. (2013) indicated that the gut of shrimp with a high growth rate demonstrated a larger variety of bacterial species and a higher Shannon index. The current findings are in contrast with those of Fan J et al. (2019): white shrimp with higher growth performance demonstrated a higher Shannon index. Huynh et al. (2019), on the other hand, came to the conclusion that no significant relationship between the Shannon index and the growth of white shrimps was observed, since there was no discernible difference in the average values of the Shannon index between shrimp with rapid and sluggish growth. These findings could suggest that other variables, such as the dominating genus or species level, influence the growth of white shrimps in addition to species richness and variety of gut bacteria. In the present study, the most dominant bacterial phylum in the high growth group, i.e., group D, was Tenericutes, followed by Firmicutes, Proteobacteria, and Bacteroidetes. In group F, the Bacteroidetes and Proteobacteria were reported to have high percentage in occurrence. In the present study, we also found a significantly high occurrence of firmicutes in group D in comparison to the other five groups. Candidatus bacilloplasma dominates the intestines of healthy shrimp, but changes in its abundance may cause a change in the disease-causing bacterial population in the intestinal microbial communities in shrimp infecting with many diseases (Chen W. Y. et al., 2017; Chen Y et al., 2017; Huang et al., 2020; Wang et al., 2020b). In this study, the increase in abundance of C. bacilloplasma might influence the growth status of shrimp. However, to date, the functions of C. bacilloplasma in the shrimp intestine are still unknown. In addition, in group D, another genus commonly found was Pedicoccus and was found in significantly higher occurrence when compared to other groups. Given that Firmicutes make up a large portion of aquaculture probiotics (Wang Y. C. et al., 2019), it is possible that these organisms may significantly influence the growth of white shrimps, albeit more research is still needed. Another bacterial genus with high occurrence is Vibrio, being an opportunistic bacterium; it has been found in almost every group irrespective of the diets. According to reports (Huang et al., 2018; Fan L. et al., 2019; Shao et al., 2019), Vibrio was one of the most common genera found in shrimp intestines, and its abundance in shrimp ponds was far higher than that of the surrounding water or soil. Our result shows that the percentage of Vibrio has been decreased in groups D and F where probiotics and M. oleifera have been used significantly. Additionally, it has been discovered that an excess of Vibrio in the gut could alter the shrimp’s health status by increasing the possibility of disease outbreaks (Xiong et al., 2017; Hou et al., 2018). To lessen the risk of disease outbreaks, it is crucial to regularly monitor the prevalence of Vibrio in certain areas because shrimps are susceptible to infections from the water. High abundances of Vibrio in the water can reach the shrimp’s intestinal tract.

To explore the mechanisms underlying the biochemical and physiological changes in shrimp, we performed functional annotation. The KEGG level 2 annotation results demonstrated five dominant functional category groups in group D: carbohydrate metabolism, nucleotide metabolism, lipid metabolism, transcription, and membrane transport (Schweinitzer and Josenhans, 2010; LeBlanc et al., 2017). This is in line with numerous earlier studies on environmental and intestinal microbiota. Moreover, our KEGG level 2 analysis showed that gut microbiota were enriched in carbohydrate metabolism. Studies in shrimp, panda, and turbot (Zhu et al., 2011; Xing et al., 2013; Gao M. et al., 2019) have indicated that carbohydrates may be a primary nutrient supplied to microbiota. Our KEGG level 3 annotation with respect to signal transduction revealed the upregulation of MAPK signaling pathway in group D. Members of the MAPK family’s cascade are crucial in controlling immune response and cell viability. In animals ranging from yeast to humans, MAPKs are evolutionary-conserved signaling pathways that act in response to stimulating signal transduction (Zhang and Dong, 2007; Huang et al., 2017). A variety of phosphorylation processes work together to activate MAPKs. When phosphorylated, MAPK affects a variety of transcription factors, enzymes, and other proteins that regulate cellular activity by phosphorylating a variety of substrates in the cytosol and nucleus (Fujiwara and Denlinger, 2007; Morrison, 2012; Huang et al., 2017). Another metabolic pathway that shows an upregulation is the AMP-activated protein kinase (AMPK) signaling pathway in group D. AMPK controls energy metabolism and serves as a sensor for intracellular energy in terrestrial vertebrates (Hardie, 2003). The role of AMPK in the energy metabolism of aquatic species, particularly shrimp, is, on the other hand, less well understood. Magnoni et al. (2012) indicated that AMPK is critical for promoting glucose uptake in fish skeletal muscles; thus, it might have a role in controlling energy responses to hypoxia in fish.

Ferroptosis is a unique reactive oxygen species (ROS)–dependent cell death process characterized by an iron overload and lipid peroxide formation. The physical signs of ferroptosis include shrinking of the mitochondria, a decrease in or absence of the mitochondrial cristae, and an increase in mitochondrial membrane density. Similar to other cell death processes, glutathione (GSH) synthesis, lipid peroxidation, cysteine transport, iron homeostasis, and NADPH all play vital roles in the regulation of ferroptosis (Zheng and Conrad, 2020). In the present investigation, we have seen comparatively high upregulation of ferroptosis in group D than in group A (control group).

In addition, noticeable upregulation of ferroptosis in group D could be related to the high number of beneficial bacteria present in this groups. In contrast to apoptosis, necrosis, and autophagy, ferroptosis is a type of intracellular iron-dependent cell death. Numerous studies have indicated that ferroptosis is crucial for the inhibition of tumors, opening up new therapeutic possibilities for cancer (Zhang et al., 2022). The molecular mechanisms that distinguish between these two distinct types of cell death remain obscure; mitochondrial ROS are crucial for both the production of apoptosis and ferroptosis (Wang et al., 2020a; Lee et al., 2020; Li et al., 2021). Depending on its substrate, AMP-activated protein kinase (AMPK), a key regulator of ATP homeostasis, has two distinct roles in ferroptosis. While AMPK-mediated ACACA phosphorylation inhibits ferroptosis by inhibiting fatty acid biosynthesis, AMPK-mediated BECN1 phosphorylation promotes ferroptosis by inhibiting SLC7A11 activity or inducing autophagy. This suggests that energy status may affect lipid biosynthesis and peroxidation during ferroptosis (Song et al., 2018; Zhang et al., 2018). Additionally, group D contained more peroxisomes, which are capable of performing various metabolic tasks, such as ROS production and catabolism, the biosynthesis of ether lipids and bile acids (Lodhi and Semenkovich, 2014), and the catabolism of very long chain fatty acids, branched chain fatty acids, D-amino acids, and polyamines. Lysozymes were also identified at larger concentrations; this may be related to peroxisome interactions with lysosomes for free cholesterol transfer (Chu et al., 2015). In group D, Vibrio was less prevalent than that in other groups. Moreover, Vibrio possessed large amounts of “bacterial chemotaxis,” “bacterial secretion system,” and “quorum sensing,” all of which were associated with bacterial virulence (Krukonis and DiRita, 2003; Green and Mecsas, 2016; Kovacikova and Skorupski, 2022). A crucial mechanism that might control the physiological and metabolic activities of the microbial community was quorum sensing. Furthermore, quorum sensing is frequently used to virulence factor secretion (Suntharalingam and Cvitkovitch, 2005; Li and Tian, 2012).

Finally, we noted different results for the beneficiary pathways when the prebiotics concentration was >5 g/kg body weight. Therefore, the ideal feed for balanced gut microbiota that may facilitate growth and enhance the immune system was group D feed, comprising 1 × 10⁷ CFU/g body weight L. acidophillus as the probiotics and 2.5 g/kg of body weight M. oleifera as the prebiotics.