For the present investigation, L. catla fish (n = 10; 5 males and 5 females) were collected from the two rearing ponds of ICAR-CIFA, Bhubaneswar, India during June 2022. The fish were not subjected to vaccination and appeared to be in good health, showing no signs of infection. All fish were anesthetized using MS-222 also known as (a.k.a.) (tricaine methanesulfonate, Sigma Aldrich, USA) at a dose of 0.1 g L⁻¹ and dissected immediately after catching. For microbial analysis, only foregut tissue samples were carefully excised, and separated from each fish under aseptic conditions, providing 5 biological replicates per sex. Tissue samples were collected using sterile scissors and scalpels, and stored individually at −20°C until further analyses. The blood samples were collected from each pre-weighed fish via the caudal peduncle and transferred to a microfuge tube at room temperature for 4 h to allow coagulation. The tubes were centrifuged at 1500 × g for 20 min and supernatant was collected to obtain serum. All experiments were approved by the Institutional Animal Ethics Committee of ICAR-CIFA, Bhubaneswar, Odisha, India. Animal studies were conducted in compliance with ARRIVE 2.0 guidelines, which include considerations for study design, sample size determination, randomization, and statistical analysis.

Surface water samples were collected once at the time of fish sampling from two rearing ponds at ICAR-CIFA to avoid any temporal bias. Using a sterile plastic bottle, three independent water samples were collected from each pond to provide biological replicates for physico-chemical analysis. Environmental parameters, e.g., water temperature, pH, conductivity, alkalinity, hardness, dissolved oxygen (DO), phosphate (PO43--P), ammonia (NH4+-N), and nitrite (NO2--N) in each rearing pond were assessed following the method outlined by Leoni et al. (2018). Body weight (W_t), standard length (SL), gonad weight (GW), liver weight (LW), and gut weight (GtW) were measured for both male and female fish. Eviscerated weight (EW) (i.e., whole fish excluding the gonad, liver, and gut tissue) was calculated. Moreover, additional bioenergetic indices were computed to evaluate the fish's condition. Gonadosomatic index (GSI) expressed as 100 × GW/EW served as a marker for fish reproductive cycles. Hepatosomatic index (HSI) was calculated as 100 × LW/EW which provided insights into liver function and metabolic health, and fatness (K) was determined as 100 × W/L³ offering an indicator of adiposity and overall fish condition. These indices together offer valuable insights into the reproductive dynamics as well as the physiological state of the fish species under investigation.

Gonadal tissues were fixed in 10% neutral buffered formalin for over 24 h before undergoing a dehydration process through a graded ethanol series. The samples were then cleared with xylene, embedded in paraffin, and sectioned into 4 μm slices using a microtome. If necessary, the section ribbons were floated on water, then transferred onto pre-cleaned albumenized glass slides and placed on a slide warmer at 60 °C for fixation. The slides were incubated for 12 h to ensure complete stretching. Subsequently, the paraffin sections were stained using the standard hematoxylin and eosin (H and E) protocol. This involved dewaxing in xylene, rehydrating through a graded ethanol series, staining the nuclei with hematoxylin, and counterstaining the cytoplasm with eosin. Finally, the slides were sealed with resin. Images were captured at 10 × and 40 × magnifications using a Nexscope RXLr-5 microscope equipped with a camera (Radical Procam). ImageJ software was also used for further analyses.

The hormone levels of 11-KT, estradiol, FSH, and LH in the serum of L. catla were quantified using enzyme-linked immunosorbent assay (ELISA) kits specific for fish (GENLISA; Krishgen Biosystems, USA; Product Code- LH: KLF0040; 11-KT: KLF0000; Estradiol: KLF0245; FSH: KLF0039). The ELISA assays were conducted following the manufacturer's protocol, and absorbance measurements were recorded using an iMark™ Microplate Absorbance Reader (Bio-Rad, India). The assay procedure involved the simultaneous incubation of the target analyte with a biotin-labeled antibody, followed by washing and the addition of an affinity-labeled horseradish peroxidase (HRP) conjugate. After further incubation and washing steps to remove unbound enzyme conjugates, substrates A and B were introduced, leading to a colorimetric reaction. The intensity of the color developed was directly proportional to the concentration of the respective hormone in the sample. All standards and samples were analyzed in triplicate to ensure accuracy and reproducibility.

DNA from foregut tissues (100 mg) and surface water (600 μl) samples was extracted using the DNeasy PowerSoil Pro Kit (Qiagen) following manufacturer's instructions. Extracted DNA samples were quantified with Thermo Scientific™ NanoDrop™ 2000/2000 c Spectrophotometers, USA. DNA quality and reliability were checked using 0.8% agarose gel electrophoresis, and subsequently kept at −20 °C for further processing. The samples underwent quality assessment, demonstrating satisfactory yield and concentration levels, which render them suitable for Illumina library preparation (Supplementary Table 1).

The V3-V4 hyper-variable region of the 16S rRNA gene was amplified via PCR utilizing the primer pair 341F-805R (Liu et al., 2020). The PCR protocol comprised an initial denaturing step at 95 °C for 5 min, followed by 30 cycles of denaturation at 94 °C for 30 sec, annealing at 56 °C for 30 sec, and extension at 72 °C for 30 sec. A final extension step was conducted at 72 °C for 10 min to complete the reaction. The PCR products were then migrated on an agarose gel and purified using the Qiagen^TM QIAquick PCR Purification Kit. Subsequently, library indexing and preparation were performed using the Nextera method. Finally, Illumina MiSeq sequencing was carried out using a 2 × 300 bp paired-end sequencing protocol. Briefly, adapter sequences were added to the primer pairs to align with Illumina index and sequencing adapters. Master mixes contained total DNA (12.5 ng) and KAPA HiFi HotStart ReadyMix (2 ×; KAPA Biosystems, Wilmington, MA). Negative controls (no template) were also included and amplified. Amplicons were purified using AMPure XP reagent. Samples were then amplified with Illumina sequencing adapters and dual-index barcodes i.e., index 1(i7) and index 2(i5) followed by purification, quantification, and normalization before pooling.

Data de-multiplexing was performed initially based on the Illumina 3^rd index reads through bcl2fastq v.2.20 software, which facilitated the conversion of raw files to fastq files. Subsequently, sequenced data quality was assessed using FastQC v.0.11.8 (Andrews, 2017). The low-quality bases (< q30) were filtered off during read pre-processing using Trimgalore v.0.4.01 (Krueger, 2015) and used for downstream analysis. The raw sequences obtained from this study have been deposited in SRA—Sequence Read Archive (SUB14182840) under the BioProject ID: PRJNA1070889.

The initial FASTQ files were subjected to quality filtering and subsequent merging, employing the QIIME 2 (Quantitative Insights into Microbial Ecology) software v.2018.11 (Caporaso et al., 2010). Following the quality filtering and merging procedures, which incorporated the use of DADA2 software for error correction in Illumina-sequenced amplicons, the outcomes consisted of (i) a feature table showing ASVs (i.e., Operational Taxonomic Units OTUs) and (ii) a feature data table with sequences linked with these ASVs (Callahan et al., 2016). Any potential contaminants originating from fish gut, like chloroplasts and mitochondria, were eliminated from the final files. Taxonomic assignment was performed and aligned with the Greengenes reference OTU database (Langille et al., 2013). Venn diagram was generated to detect the unique and shared OTUs by using a web-based tool i.e., Bioinformatics Tools for genomics and transcriptomics analysis (https://www.biotools.fr/misc/venny).

Diversity analysis was performed on nine rarefaction depths (ASVs rarefied at 1,667, 3,334, 5,000, 6,667, 8,333, 10,000, 11,666, 13,333, and 15,000 sequences per sample) by analyzing alpha rarefaction curves. Sequence coverage of ≥97% was supported by QIIME2 diversity plugin-generated rarefaction curves. α and β-diversity estimates were calculated using PAST software v.4.03. Several α-diversity metrics were estimated, including richness (observed species i.e., Sobs, Chao1) as well as the diversity indices (Shannon, Simpson). Spearman's correlations (ρ) were computed and identified abiotic factors that are strongly associated with richness, revealing the key environmental drivers of bacterial α-diversity. Distances of β-diversity i.e., Bray-Curtis (taxonomic) as well as UniFrac (phylogenetic) [weighted (abundance matrix), and un-weighted (presence and/or absence matrix)] distances were calculated to explore the differences among the structure of gut bacterial communities between the sample groups (Lozupone and Knight, 2005). Both Bray-Curtis and weighted UniFrac dissimilarity matrices of these bacterial groups were graphically visualized by two-dimensional NMDS (non-metric multidimensional scaling) plots. Pairwise dissimilarity assessment of β-diversity between sex-specific groups (i.e, male and female) was performed via Multivariate Analysis of Variance (MANOVA), a non-parametric test. One-way ANOVA, followed by Games-Howell post-hoc test (p-value < 0.05) was applied to compare significant differences in α-and β-diversity indices between L. catla gut sample groups.

The variation within the bacterial communities was evaluated by using SIMPER (Similarity Percentage Analysis) based on Bray-Curtis distance matrix. This analysis was performed in PAST v.4.03, a software application designed for statistical analysis and data visualization. To identify variations in taxa composition, LEfSe analysis, i.e., Linear Discriminant Analysis (LDA) Effect Size, was applied (Segata et al., 2011). Taxa meeting the criteria where their logarithmic base 10 of LDA score >2.0 and FDR (false discovery rate) < 0.05 determined by the Kruskal-Wallis test, were regarded as differential taxa. One-way ANOVA were performed to assess the sex stages on serum hormone levels. The Mantels' pearson correlation test was utilized to assess relationship between the bacterial community and serum hormones, conducted using the PAST v.4.03 software.

We used PICRUSt2 analysis i.e., Phylogenetic Investigation of Communities by Reconstruction of Unobserved States to anticipate bacterial functionality (Douglas et al., 2019). Metagenome prediction accuracy was estimated by calculating the NSTI value i.e., Nearest Sequenced Taxon Index. Normalization of OTU table was carried out, taking into account the expected 16S copy number for each organism, and these OTU tables were further employed to predict the functional profile using KOs (Kyoto Encyclopedia of Genes and Genomes i.e., KEGG Orthologs). To aggregate all KEGG Orthologs (KOs) into their respective KEGG pathways, the KOs were grouped by their functions. The role of individual OTU in influencing a specific gene function was determined by calculating the product of OTU abundances and the corresponding gene abundances for each taxon. We also examined functional categories within the gut microbiomes of male and female L. catla, comparing aspects like molecular communication, disease suppression, nutrition, and other factors crucial for growth promotion and fitness. The outcomes from PICRUSt2 were examined using a software, STAMP v.2.1.3 (Statistical Analysis of Metagenomic Profiles), employing Welch's two-sided t-test (i.e., for unequal variance) to detect significant variation in bacterial functional abundances between male and female sex types (Parks et al., 2014). Extended error bar plots were utilized across three KEGG pathway levels for visualization.

Canonical correspondence analysis (CCA) was carried out to explore the link between the gut bacterial community compositions of distinct sex-specific L. catla groups and their essential biological and environmental factors. Co-occurrence networks were created to investigate the inter-taxa relationship in the gut microbiomes specific to each sex of the L. catla fish. The algorithm for Network Analysis for Metagenomic Abundance Profiles calculated Pearson's correlation coefficients “r” using “MetagenoNets” (web-based tool available at https://web.rniapps.net/metagenonets) (Nagpal et al., 2020). Taxa were filtered using a maximum prevalence threshold of ≥0.0001, and they were included if they occurred in at least 10% of the samples. For further analysis, we exclusively considered relatively strong correlations (|r| > 0.7, p-value < 0.01) through 100 bootstrapping iterations, and the visualization of the networks was conducted using the Gephi platform (v.0.10.1) accessible at https://gephi.org. Moreover, several descriptive measures of co-occurrence networks, including network-level features (e.g., nodes, edges, average degree, clustering coefficient, graph density, modularity, and statistical inference) and node-level features (e.g., degree, closeness centrality, betweenness centrality, and eigenvector centrality) were calculated using Gephi software v.0.10.1 for each network. Average degree represents the mean number of connected edges per node. Average clustering coefficient denotes the mean ratio of node connectivity to the maximum connectivity among the connected nodes. This analysis aimed to discern variations in sex-specific L. catla gut microbiome co-occurrences and identify keystone species. The top 4 nodes were considered as keystone taxa, which followed high values of closeness centrality and average degree, including low scores of betweenness centrality (Berry and Widder, 2014).

Detailed information on environmental variables, e.g., water temperature, pH, conductivity, alkalinity, hardness, DO, PO43-P, NH4+-N, and NO2--N are listed in Table 1. The data on the essential biological characteristics of the fish demonstrated a consistent progression in the male fish groups, with SL, Wt, LW, EW, and HSI all showing an increasing trend. During the same period, noteworthy variations in GW and GSI were observed between the two distinct groups, with the female fish displaying higher values (Table 1).

Histological analysis of gonadal tissues revealed clear morphological features indicative of active gametogenesis in both male and female specimens (Figures 1A, B). In the testicular sections, the presence of spermatogonial cells along the basal lamina and primary spermatocytes beneath it signifies the initiation of spermatogenesis. The testes exhibited a well-developed tubular structure, with intertubular connective tissue sparsely populated by sustentacular cells (Figure 1A). Although a few spermatozoa were observed in certain areas of the testes, they appeared to be in an immature stage, suggesting ongoing development.

Ovarian sections displayed oocytes at various stages of development, including early perinucleolar, cortical alveolar, and pre-vitellogenic stages (Figure 1B). The ovarian tissue was characterized by Stage I differentiation, comprising primary follicles and cells in the cortical alveolus stage. Follicular cells were observed surrounding the ovarian follicles, indicating their role in supporting oocyte maturation. The presence of asynchronously developing oocytes suggests a batch-spawning reproductive strategy. These observations confirm active gametogenic processes in both sexes and provide valuable insights into the reproductive status and maturation stages of the analyzed specimens.

Serum hormone levels (11-KT, estradiol, FSH, and LH) of L. catla were measured during pre-spawning stages (Figures 1C, D). No significant differences in 11-KT, FSH, and LH levels were observed between male and female (p-value > 0.05). However, one-way ANOVA revealed significant differences in estradiol levels between the sexes (p-value < 0.05).

For bacterial communities, data from the 16S rRNA gene sequencing run produced 2,640,687 reads for 10 samples from both female (n = 5) and male (n = 5) fish. After quality filtering and removal of the chloroplast, mitochondria, and other non-target ASVs, 2,384,007 reads were retained (Supplementary Table 1) representing 1,194 OTUs from gut samples. Individual samples had a wide range of reads (minimum = 75,349; maximum = 643,407; mean = 238,401 ± 192,955). Full sequencing data are publicly available at NCBI BioProject ID PRJNA1070889.

MiSeq Illumina sequencing of female L. catla gut tissue samples (n = 5) resulted in a total of 647,044 reads. After quality filtering, each sample had a mean of 129,408.8 ± 56,437.5 sequences. A total of 35 bacterial phyla were identified that were mainly belonged to Proteobacteria (59.97%), Fusobacteria (18.93%), Bacteriodetes (14.20%), and Firmicutes (5.02%), together accounting for over 97% of total bacterial communities (Figure 2A). In the female gut samples, the dominant classes were γ-Proteobacteria (57.32%), Fusobacteriia (18.97%), Bacteroidia (13.15%), and Clostridia (3.97%) (Figure 2B). At the order level, female L. catla gut bacterial communities were greatly signified by Enterobacteriales (26.57%), Fusobacteriales (20.50%), Bacteroidales (14.04%), Alteromonadales (11.18%), Aeromonadale (10.90%), Clostridiales (4.29%), Vibrionales (3.27%), and Flavobacteriales (1.03%) (Figure 2C). Within the 218 different families, the most predominant taxa collectively represent over 85% of the total number of sequences were Enterobacteriaceae (27.37%), Fusobacteriaceae (21.54%), Bacteroidaceae (13.71%), Shewanellaceae (10.98%), Aeromonadaceae (10.96%), Peptostreptococcaceae (1.47%), Acidaminobacteraceae (1.30%), and Campylobacteraceae (1.00%) (Figure 2D). Of the 400 definite genera, Cetobacterium (33.76%), Shewanella (18.91%), Edwardsiella (9.92%), Serratia (6.56%), Bacteroides (6.16%), Plesiomonas (2.22%), Fusibacter (2.04%), Sulfurospirillum (1.57%), Klebsiella (1.33%), and Akkermansia (1.19%) were the most dominant (Figure 2E).

Sequencing of male L. catla gut tissue samples (n = 5) produced a total of 1,736,963 sequences, with a mean of 347,392.6 ± 225,583.2 sequences per sample. Among 50 different phyla that were detected in the male L. catla gut, the dominant ones were Proteobacteria (45.85%), Fusobacteria (27.86%), Bacteroidetes (14.17%), and Firmicutes (10.34%) (Figure 2A). In the male gut samples, the dominant classes were γ-Proteobacteria (42.10%), Fusobacteriia (27.88%), Bacteroidia (12.85%), Clostridia (8.97%), β-Proteobacteria (2.26%), α-Proteobacteria (1.19%) and Flavobacteriia (1.09%) (Figure 2B). At the order level, out of the 197 definite orders, Fusobacteriales (30.27%), Enterobacteriales (15.02%), Bacteroidales (14.29), Clostridiales (9.84%), Alteromonadales (6.70%), Pseudomonadales (1.69%), Burkholderiales (1.7%), Flavobacteriales (1.27%), Neisseriales (1.03%), collectively signified 92.55% of total bacterial communities in the male L. catla gut tissue (Figure 2C). Within the 289 different families, the most dominant taxa accounting for >85% of the total number of sequences were Fusobacteriaceae (31.36%), Enterobacteriaceae (15.45%), Bacteroidaceae (11.94%), Aeromonadaceae (11.17%), Shewanellaceae (5.60%), Peptostreptococcaceae (4.54%), Porphyromonadaceae (2.50%), Clostridiaceae (2.28%), and Neisseriaceae (1.06%) (Figure 2D). Major genera detected in male L. catla gut were Cetobacterium (44.92%), Shewanella (9.80%), Bacteroides (5.05%), Edwardsiella (4.80%), Dysgonomonas (3.53%), Serratia (3.21%), Clostridium (3.20%) Plesiomonas (2.79%), Rheinheimera (1.77%), Chryseobacterium (1.69%), Pseudomonas (1.29%), Acinetobacter (1.06%), Epulopiscium (1.05%), and Fusobacterium (1.02%) (Figure 2E).

Venn diagram is applied to demonstrate the shared and unique OTUs in two sex-specific L. catla gut bacterial groups. A total of 798 and 1,103 OTUs were observed in the female and male gut microbial groups, respectively (Supplementary Figure 1). A total of 707 OTUs were shared between the male and female guts, accounting for 59.21% in the gut microbial community. The female gut microbial group possessed the fewest unique OTUs (91; 7.62%), whereas the male gut microbial group displayed the highest number of unique OTUs (396; 33.17%).

LEfSe analysis was carried out to exemplify and highlight significant variations in the gut bacterial abundances among distinct sex-specific L. catla groups (Supplementary Figure 2). The analysis revealed that two families viz. Clostridiaceae, Exiguobacteraceae, and one genus (Aquitalea) were significantly different and higher (LDA score > 2.0; p-value < 0.05) in male gut samples compared with the female groups. In contrast, the female gut bacterial group demonstrated markedly greater abundances of ε-Proteobacteria (and their sister lineages, namely, Campylobacterales, and Campylobacteraceae) compared with the male gut microbiome groups (Supplementary Figure 2). Clostridiaceae and Exiguobacteraceae exhibited a significant increase of 9.97 and 8.54 folds, respectively, in male gut samples as compared to female gut samples. Meanwhile, Campylobacteraceae showed a significant 10.54-fold increase in the female gut group compared to the male gut.

The functional profiling of bacterial communities harboring the L. catla intestine in male and female groups was predicted using PICRUSt assignment through 16S rRNA gene sequences (Figure 5). NSTI value i.e., 0.14 ± 0.04, determined from PICRUSt analysis, signifies high prediction accuracy. The study involved analyzing bacterial 16S rRNA gene sequencing dataset from gut samples of L. catla to predict genes and their functional pathways using KEGG databases. In total, 7,305 KO groups were identified, encompassing 247 pathways in L. catla gut samples (n = 10). The analysis showed that the majority (63.31%) of the detected genes at KEGG level-1 were enriched in metabolic pathways, followed by genetic information processing (23.74%), environmental information processing (9.78%), cellular processes (1.50%), human diseases (1.07%), and organismal systems (0.57%) (Figure 5A).

In addition, amino acid metabolism (9.69%), metabolism of terpenoids and polyketides (8.75%), carbohydrate metabolism (8.28%), replication and repair (5.79%), metabolism of cofactors and vitamins (5.78%), xenobiotics biodegradation and metabolism (5.11%), lipid metabolism (5.02%), enzyme families (4.92%), metabolism of other amino acids (4.77%), folding, sorting and degradation (4.38%), biosynthesis of other secondary metabolites (3.69%), translation (3.45%), transport and catabolism (3.10%), energy metabolism (3.02%), and others were detected as the major predicted functional pathways in the KEGG level-2 (Figure 5B).

The study found that at KEGG level-3, the relative abundances of genes linked with various metabolic pathways were particularly prominent (Figure 5C). These pathways included valine, leucine, and isoleucine degradation, peptidases, phenylalanine metabolism, glycolysis/gluconeogenesis, polyketide sugar unit biosynthesis, limonene and pinene degradation, carbon fixation in photosynthetic organisms, pentose phosphate pathway, N-glycan biosynthesis, synthesis and degradation of ketone bodies, vitamin B6 metabolism, glutathione metabolism, aminobenzoate degradation, D-glutamine and D-glutamate metabolism, novobiocin biosynthesis, tropane, piperidine, and pyridine alkaloid biosynthesis, geraniol degradation, glycerophospholipid metabolism, glyoxylate and dicarboxylate metabolism. Moreover, this study revealed a dominance of relative gene abundances in various pathways, including those related to genetic information processing (such as proteasome, ribosome biogenesis in eukaryotes, homologous recombination, chromosome, and DNA replication proteins), environmental information processing (secretion system), cellular processes (peroxisome, cell cycle in Caulobacter, bacterial chemotaxis), as well as genes associated with human diseases (Huntington's disease and tuberculosis) (Figure 5D).

The heatmap analysis revealed significant correlations (p-value < 0.05) between specific bacterial phyla and genera with serum hormone levels. At the phylum level, Bacteroidetes showed a strong negative correlation with 11-KT (r = −0.756, p-value < 0.05) (Supplementary Figure 3A). At the genus level, Dysgonomonas (r = −0.799) and Sulfurospirillum (r = −0.729) displayed significant negative correlations (p-value < 0.05) with LH, and Bacteroides showed negative correlation with 11-KT (r = −0.843, p-value < 0.05) (Supplementary Figure 3B). In contrast, estradiol showed significant (p-value < 0.05) positive correlation with Shewanella (r = 0.830) and Serratia (r = 0.837). These findings suggest that specific microbial taxa may modulate reproductive hormone dynamics in L. catla.

Male and female L. catla gut microbiomes were studied to unveil the abiotic and biotic factors shaping their bacterial community structure (Figure 7A). At the order level, the cumulative variation explained by the first two CCA axes was 79.59%. The CCA axis 2 showed a strong positive association with three environmental factors such as conductivity (ρ = 0.871), PO43–P (ρ = 0.741), NH4+-N (ρ = 0.647), whereas negatively associated with water temperature (ρ = −0.870), pH (ρ = −0.870), alkalinity (ρ = −0.869), hardness (ρ = −0.846), DO (ρ = −0.838), NO₂–N (ρ = −0.771).

Significant correlations have been observed between gut bacterial communities and environmental factors (Supplementary Figure 4A). For instance, bacterial families namely, Porphyromonadaceae (r = 0.784 with NO₂, r = 0.690 with DO, r = 0.672 with pH, r = 0.661 with WT and alkalinity, r = 0.649 with hardness, r = −0.637 with conductivity), Bacteroidaceae (r = −0.679 with NH₄), Campylobacteraceae (r = 0.685 with NO₂, r = 0.672 with DO), Aeromonadaceae (r = −0.663 with DO) and Acidaminobacteraceae (r = 0.663 with DO, r = 0.643 with NO₂) were markedly (p-value < 0.05) associated with environmental factors of L. catla fish gut. Significant correlations have been observed between gut bacterial communities and the biological factors (Supplementary Figure 4B). For instance, Enterobacteriaceae (r = −0.766 with SL, r = 0.754 with GtW), Porphyromonadaceae (r = −0.721 with Wt, r = 0.681 with GtW, r = −0.721 with EW), Clostridiaceae (r = −0.778 with GW, r = −0.745 with GSI), Acidaminobacteraceae (r = −0.770 with EW, r = 0.685 with GSI, r = 0.650 with GtW), Campylobacteraceae (r = −0.779 with EW, r = 0.656 with GSI, r = 0.652 with GtW) were markedly (p-value 0.05) linked with biological factors of L. catla fish gut.

CCA based VPA i.e., variance partitioning analysis revealed that differences in the bacterial communities were mainly explained by biological (63.97%) than environmental (20.99%) factors (Figure 7B). VPA also showed that weight parameters (Wt, GW, LW, GtW, and EW) accounted for 37.34%, GSI, HIS, and K accounted for 29.04%, and body length (SL) accounted for 4.22% variance in the gut microbiome composition. The analysis also revealed that Wt accounted for the maximum variance (16.27%), followed by GtW (14.10%), K (12.70%), GSI (10.02%), EW (7.35%), GW (6.21%), LW (6.04%), and HSI (5.20%).

Co-occurrence network analysis in each L. catla gut bacteria group unveiled notable differences in their interactive patterns. These variances were evident in established positive and negative correlations depicted in the network plots (Figure 8). For instance, co-occurrence network indicating potential interactions among the bacterial taxa of the male gut microbiome comprised 17 nodes and 32 edges (Figure 8A). Female gut bacterial communities demonstrated a greater inter-connectedness (average clustering coefficient 0.805 vs 0.630), average degree (5.176 vs 3.765), edges (44 vs 32), graph density (0.324 vs 0.235), modularity (0.279 vs 0.271), and statistical inference (124.182 vs 83.921) as compared to the male gut bacterial communities (Figures 8C–E). Both male and female gut bacterial communities showed a maximum prevalence of negative association and the numbers of negatively correlated nodes were lower in female L. catla gut (52.27%) than in male fish (53.12%).

Co-occurrence network analyses emphasized the noteworthy inter-taxa interactions in L. catla gut bacterial communities (Figure 8). Species of the class γ-Proteobacteria were found to have strong positive associations within the gut bacterial communities of males and females. For example, Pseudomonas stutzeri showed strong positive correlations with Salmonella enterica and Serratia marcescens. Salmonella enterica was also positively correlated with Serratia marcescens. Furthermore, various nodes on the networks exhibited negative correlations with other nodes. For instance, Cetobacterium somerae (Fusobacteriia) showed a negative correlation with Pseudomonas stutzeri, Salmonella enterica, and Serratia marcescens.

The co-occurrence network revealed specific bacterial relationships unique to female L. catla gut communities, absent in males. In the female network, Clostridium perfringens and Propionibacterium acnes had strong positive relationships with Pseudomonas stutzeri, Salmonella enterica, and Serratia marcescens (Figure 8B). Plesiomonas shigelloides displayed significant positive correlations with Pseudomonas stutzeri, but these associations were absent in males (Figure 8A). Similarly, specific bacterial relationships unique to male L. catla gut communities were absent in females. For instance, Akkermansia muciniphila (Verrucomicrobiae) displayed a negative association with Cetobacterium somerae (Fusobacteriia) and a positive association with Pseudomonas stutzeri, Salmonella enterica, and Serratia marcescens belonging to the class γ-Proteobacteria (Figure 8B). The key taxa were represented by Clostridium perfringens, Pseudomonas stutzeri, Plesiomonas shigelloides, and Serratia marcescens (Supplementary Table 3).

Histological observations confirmed active gametogenesis in both male and female L. catla, highlighting distinct reproductive maturity during the pre-spawning phase. Testes displayed multiple spermatogenic stages, indicating continuous sperm production, while ovaries contained vitellogenic oocytes with yolk deposition, reflecting advanced reproductive readiness and batch-spawning behavior (Uribe et al., 2014; Udit et al., 2024). Ovarian histology further demonstrates that optimal dietary conditions and specific supplements, such as balanced protein-to-carbohydrate ratios and choline chloride, enhance oocyte maturation and reproductive performance, while imbalances or environmental contaminants can disrupt normal gonadal development and hormone profiles (Das et al., 2023; Udit et al., 2024). Elevated estradiol levels in females are consistently linked to active vitellogenesis and oocyte maturation, supporting findings in other cyprinids, while 11-KT, FSH, and LH levels do not show significant sex differences, suggesting shared endocrine mechanisms for reproductive initiation but a unique estradiol-driven pathway in females (Ganguly et al., 2023; Udit et al., 2024). Such sex-specific hormonal differences provide an important physiological backdrop to interpret variations in microbial composition and functions. These findings provide valuable insights into the species' reproductive status, informing effective aquaculture breeding strategies.

The gut microbiome was dominated by members of Proteobacteria, Fusobacteria, Bacteriodetes, and Firmicutes, which are core taxa commonly observed in freshwater fish such as common carps, grass carp, as well as in the snakehead murrel (Han et al., 2010; Wu et al., 2012; Ye et al., 2014; Li H. et al., 2015; Li T. et al., 2015; Rasal et al., 2022). Notably, Proteobacteria are involved in the fish metabolism and cycling of biogeochemicals such as carbon, nitrogen, and sulfur (Fjellheim et al., 2010; Han et al., 2010). However, striking sex-based differences were detected. Within the phylum Proteobacteria, γ-Proteobacteria dominated in both male and female L. catla gut samples, with genera such as Plesiomonas (Enterobacteriaceae) and Acinetobacter (Moraxellaceae) commonly observed in several carp species (Namba et al., 2007; van Kessel et al., 2011). Female gut samples showed higher abundances of Serratia marcescens, Plesiomonas shigelloides, and Salmonella enterica in some samples such as CGF1 (15.26%), CGF2 (8.39%), and CGF3 (7.92%), respectively. These bacterial species can thrive as facultative pathogens in both fish and humans without causing overt disease (Walczak et al., 2017). Serratia generates serrawettin (a wetting agent) reducing environmental surface tension (Chan et al., 2013), while P. shigelloides has been linked to pathogenicity in freshwater carp (Behera et al., 2018). The natural intestinal flora exhibited stress-induced symptoms such as body darkening, hemorrhaging, fin rot, ascitic fluid accumulation, and internal organ lesions (Liu et al., 2015). P. stutzeri strains improved water quality, growth, gut microbiota, immune response, and hold probiotic potential for aquaculture through water supplementation (Fu et al., 2021).

Fusobacteria, particularly Cetobacterium, dominated male gut samples (97.51% in CGM2) and are known for butyrate and acetate production, supporting gut health, anti-inflammatory effects, and vitamin B12 synthesis (Tsuchiya et al., 2008; van Kessel et al., 2011; Wang et al., 2021). Shewanella was one of the abundant genera observed in both sexes of gut samples of L. catla in this study, produces omega-3 PUFAs i.e., polyunsaturated fatty acids (EPA: eicosapentaenoic acid and DHA: docosahexaenoic acid) and is being explored as a probiotic to enhance growth and immunity (Monroig et al., 2013; Egerton et al., 2018; Ghosh et al., 2021). Within the Firmicutes, Clostridium (gram-negative and obligate anaerobes) contributed to polysaccharide degradation, pathogen inhibition, SCFA production, and overall gut health, including fostering anti-inflammatory and antibacterial properties (Flint et al., 2008; Ikeda-Ohtsubo et al., 2018; Piazzon et al., 2020). In addition, our study detected the presence of Klebsiella aligns with previous reports of fish gut as a reservoir for opportunistic pathogens (Austin et al., 2007; Wu et al., 2012; Li H. et al., 2015; Mukherjee et al., 2020). Bacteroidetes, particularly Bacteroides were notably abundant in female L. catla gut samples, aiding in polysaccharide degradation, SCFA production, fatty acid metabolism, and overall gut function (Hu et al., 2018; Mohapatra et al., 2020; Foysal et al., 2020; Mukherjee et al., 2020). The genus also facilitates vitamin synthesis, oxygen depletion to maintain anaerobiosis, and utilization of dietary polysaccharides (Xia et al., 2014; Schwalm and Groisman, 2017), strengthening their candidacy as next-generation probiotics (Wexler and Goodman, 2017).

Sex-specific differences were evident for bacterial community composition, with female gut samples showing higher proportions of ε-Proteobacteria with a higher proportion of ε-Proteobacteria (and their sister lineages, namely, Campylobacteraceae). In contrast, male gut samples exhibited greater abundance of Clostridiaceae, Exiguobacteraceae, and Aquitalea. Members of Campylobacteraceae, often linked with reduced growth and disease symptoms in fish (Lastovica et al., 2014; Legrand et al., 2018), were more abundant in females, suggesting a correlation with lower microbial diversity (Legrand et al., 2018; Mehta et al., 2021). The genus Exiguobacterium, enriched in male gut samples, is recognized for its adaptability and roles in lipid droplet formation, stress tolerance, enzyme production, and amino acid release (Semova et al., 2012; Lei et al., 2014; Kasana and Pandey, 2018). The strain contributes to environmental pollutants degradation, pesticides removal, and bioremediation of arsenic, thereby mitigating oxidative stress and enhancing fish health (Shanthakumar et al., 2015; Pandey and Bhatt, 2015).

Functional predictions revealed distinct metabolic signatures between sexes. In females, microbial pathways related to fatty acid biosynthesis and carbohydrate metabolism were enriched, potentially reflecting the heightened energy requirements for vitellogenesis and gonadal maturation. In contrast, males displayed enrichment of amino acid biosynthesis pathways, such as lysine metabolism, along with antioxidant pathways including glutathione metabolism. These functions may enhance protein turnover and stress tolerance, traits that are beneficial for growth and spermatogenesis. The alignment of microbial functions with sex-specific reproductive and metabolic demands underscores the adaptive significance of gut microbiota in shaping host physiology.

Lysine, an essential amino acid synthesized via the DAP, i.e., diaminopimelic acid, and AAA, i.e., aminoadipic acid, pathways (Liu et al., 2010), was significantly enriched in male microbiota. Bacterially derived lysine has been reported to contribute to host amino acid pools (Metges, 2000), suggesting that enhanced microbial lysine biosynthesis may support male growth and spermatogenesis. Likewise, glutathione-related pathways, including gst, gpx, and gsr, were more abundant in males, potentially enhancing oxidative stress tolerance, cellular detoxification, and immune regulation (Wu et al., 2004; Vallejos-Vidal et al., 2020). These sex-specific functional enrichments highlight the adaptive alignment of microbial metabolism with host reproductive and physiological needs. In both sexes, enrichment of amino acid degradation pathways, particularly for Val, Leu, and lle, was associated with fatty acid metabolism (Zhang et al., 2019), indicating a broader microbial role in nutrient turnover and metabolic health. Such microbial contributions to amino acid and lipid metabolism align with observations in trout hepatocytes (Chang et al., 2011; Dai et al., 2015), underscoring their importance for aquaculture productivity. Moreover, cellulolytic and hydrolytic enzymes (e.g., cellulase, chitinase, amylase) from gut bacteria remain vital for digesting plant-based feed ingredients, supporting host nutrient acquisition (Jiang et al., 2011; Ray et al., 2012; Ni et al., 2014).

The α and β-diversity indices are pivotal for evaluating microbiota structure (Luo et al., 2022), and in this study, diversity analysis revealed lower gut bacterial diversity in female L. catla compared to males, likely due to a significant water shift experienced by the female L. catla community during cultivation, leading to a decline in pond diversity, consistent with earlier observations in rainbow trout and Atlantic cod (Bakke et al., 2013; Wong et al., 2013). NMDS analysis further highlighted clear sex-specific differences in microbial composition, suggesting that distinct physiological demands between the sexes may potentially influence nutrient absorption, immune function, and overall health.

Gut microbiome showed significant correlations with serum hormone levels, suggesting potential microbial influences on reproductive endocrinology. Negative correlation between Bacteroidetes and 11-KT, as well as between Dysgonomonas and Sulfurospirillum with LH, implies that these bacterial taxa may play inhibitory roles in androgenic and luteinizing hormone regulation (Thackray, 2019). In contrast, positive correlations of Shewanella and Serratia with estradiol point toward a role in supporting vitellogenesis in females. This observation consistent with other vertebrates showing gut microbial modulation of steroidogenesis by influencing hormone metabolism and signaling pathways (Neuman et al., 2015; Colldén et al., 2019), highlight the complexity of host–microbiome–hormone interactions and underscore the need for further research to explore microbiome-targeted strategies for improving fish reproductive performance in aquaculture.

Abiotic factors strongly influenced gut microbiota composition in L. catla, as indicated by CCA, which explained 79.6% of the variation. Nutrient-rich conditions with higher conductivity, phosphate, and ammonium favored bacteria involved in nitrogen and phosphorus cycling, while stable, oxygenated environments appeared less conducive for certain taxa. These results suggest that gut bacteria adapt to fluctuating water quality, with parameters such as temperature, pH, and DO shaping microbial metabolism and immune-related functions (Li et al., 2017; Zhao et al., 2018). Such insights highlight the critical role of environmental conditions in modulating fish gut microbiota and emphasize the importance of maintaining optimal water quality in aquaculture to enhance nutrient absorption, immunity, and reproductive performance.

Co-occurrence network analysis highlighted sex-specific variation in the L. catla gut microbiota interactions. Female networks exhibited higher clustering and modularity, suggesting stronger and more structured microbial associations compared to males. Network topology reflects gut bacteria interactions, and the average degree, betweenness centrality, and closeness centrality indicate the individual bacteria species' importance in influencing interactions within the network plot (Ma et al., 2016). Bacterial taxa with elevated values of betweenness centrality, positioned near the network core, likely exert significant influence on community interactions compared to other species (Li et al., 2021). Taxa within γ-Proteobacteria showed robust positive relationships and central roles in both sexes, while cross-class interactions were largely negative, reflecting competition or niche exclusion. Overall, these patterns reveal that sex influences microbial interrelationships, with positive correlations likely reflecting mutualistic adaptations to resource use and stress, while negative ones arise from competition and co-exclusion mechanisms. Notably, Verrucomicrobiae exhibited strong negative correlations with Fusobacteriia, while positive correlations were observed with γ-Proteobacteria, indicating commensalism, mutualism, symbiosis, and cooperative interactions (Coyte and Rakoff-Nahoum, 2019; Wang et al., 2024). Akkermansia muciniphila (Verrucomicrobiae), a mucin-degrading anaerobe, produces butyric acid that fuels gut epithelial cells and supports barrier integrity. Beyond this, it contributes to intestinal immunity and regulates glucose and lipid metabolism (Dao et al., 2015; Naito et al., 2018). Overall, these patterns reveal that sex influences microbial interrelationships, with positive correlations likely reflecting mutualistic adaptations to resource use and stress, while negative ones arise from competition and co-exclusion mechanisms (e.g., niche exclusion, direct competition for limited resources, edaphic/environmental modification, and toxin production) (Faust and Raes, 2012).