Comparative microbial community occurrence pattern, growth attributes, and digestive enzyme indices of Puntius gonionotus (Bleeker, 1850), Pangasianodon hypophthalmus (Sauvage, 1878) and Heteropneustus fossilis (Bloch, 1794) under freshwater biofloc based polyculture system

Abstract

Background

The biofloc system (BFS) provides a sustainable aquaculture system through its efficient in situ water quality maintenance by the microbial biomass, besides continuous availability of these protein-rich microbes as feed to enhance growth and immunity of the reared organism. This study explores the gill architecture, growth performance, digestive enzyme activity, intestinal microbial composition, and histology of three freshwater fish species, Puntius gonionotus, Pangasianodon hypophthalmus, and Heteropneustus fossilis reared in biofloc based polyculture system.

Results

The three species in T2 showed significantly higher WG and SGR, followed by T1 and T3. The wet mount of gill architecture showed smaller inter-filament gaps in gill arches of silver barb followed by stinging catfish and stripped catfish, but showed no correlation with the weight gain. However, silver barb being an omnivore and filter-feeder, accumulated a more diverse microbial community, both in T1 and BFS (T2 and T3), while the bottom feeder H. fossilis exhibited unique gut bacterial adaptability. The presence of floc in T2 and T3 enhanced bacterial abundance in water and fish gut, but their microbial diversities significantly reduced compared to T1 receiving only feed. Next-generation sequencing revealed that the Pseudomonas dominated in gut of P. gonionotus and P. hypophthalmus in T1, Enterobacterales and Fusobacterium prevailed in those of T2 and T3, respectively. In contrast, gut of H. fossilis had the highest proportion of Clostridium in T1, while Rhizobiaceae dominated in T3. Similarly in floc samples, Enterococcus dominated in T1 while Micrococcales and Rhizobiaceae dominated in T2 and T3, respectively. A positive correlation of enterobacteria, with the digestive enzyme activities and growth patterns was observed in all treatments.

Conclusion

The present study revealed feeding behaviour to play crucial role in distinguishing the gut microbial composition patterns in fishes reared in Biofloc System. Further it revealed the requirement of supplementary feed along with floc in these three species for higher growth in the biofloc system.

Introduction

Biofloc aquaculture is a scientifically established method that addresses fundamental challenges in intensive fish farming, such as maintaining water quality, lowering feed costs, controlling pathogenic bacteria, improving biosecurity, increasing disease resistance, and eventually achieving high production. In such system, the resuspended organic waste, such as faces, unused feed along with the added carbon sources to balance the C: N ratio supports growth of heterotrophic bacteria, that not only convert organic waste into protein rich microbial food for fish, but also maintains water quality by enhancing nitrification [1, 2] in the system. Previous research into the biofloc culture system for finfish has predominantly emphasized aspects such as growth, survival, immunity, replacement of dietary protein, and the use of suitable carbon sources [3], as well as the system’s impact on water quality specific to certain species. Many of these have revealed that biofloc based aquaculture provides better growth, higher protein digestibility, reduces dietary protein requirement, and it enhances antioxidant activity, and cellular and humoral defence mechanism [4]. The success of BFT depends on management of bacterial community of culture system and further environmental microbes have a strong correlation with animal gut microbiome [5]. To explore on biofloc based diversity of bacterial communities and investigate the interaction of this microbial environment with animal gut microbiome, a couple of experiment have applied next generation sequencing (NGS). According to Cardona et al. [6], shrimp reared in BFT displayed four out of five dominant gut bacterial communities different than shrimp reared in clear water. According to Zhao et al. [7], addition of a consortium of probiotic bacteria such as Bacillus, Pseudomonas, Nitrobacter reduces Vibrio count in BFS. Deng et al. [8] has reported dominance of bacterium involved in carbohydrate fermentation and short-chain fatty acids production in gut microbiota of BFT reared nile tilapia. In a recent study, Vala et al. [9] reported that gut microbiota of pangasius reared in BFT system is significantly different than one reared in cage and pond system. However, comparing the similarity of gut microbiota of fish with floc microbiome is one of the criteria to to reveal the floc consumption pattern of a particular fish.

Zhao et al. [10] selectively studied upon performance of filter feeder (bighead carp) and bottom feeder (mirror carp) in BFT based polyculture. Few of the filter feeding freshwater finfish such as tilapia, common carp and Indian major carps perform better in BFT in terms of growth, immunity and digestive enzyme activities [11, 12]. Filtration from the column or ingestion of the deposited floc from the bottom are the two modes of entry of biofloc particles into the fish digestive system. Therefore, feeding habit and the filtration efficiency of gill architecture are considered important factors to determine suitability of a fish for BFT. However, study on correlation of the gill architecture, particularly the inter-filamental gaps, with the floc intake and subsequent growth indices is necessary to examine suitability of few omnivorous and carnivorous fish species for their possible inclusion as candidate for BFT. Additionally, correlating the gut microbiome with feeding habits, gill anatomy, growth characteristics, and digestive enzyme profiles would aid in identifying the most suitable fish species for such systems. Within the limited information on the fish species’ suitability for freshwater BFT, the current experiment was designed to culture three species of varied feeding habit: a herbivore (silver barb, Puntius gonionotus), a carnivore (stinging catfish, Heteropneustus fossilis) and an omnivore (striped catfish, Pangasianodon hypophthalmus) together under polyculture in BFT. The suitability of these species in BFT were assessed based on their growth, gut microbial populations, biofloc-associated microorganisms digestive enzyme indices, and tissue microarchitecture.

Materials and methods

Experimental setup

The study was performed for 90 days in the Biofloc Research Facilities of ICAR- Central Institute of Freshwater Aquaculture (ICAR-CIFA) (20° 11’ 39.012’’ N, 85° 51’ 0.0108’’ E), Bhubaneswar, India. Nine indoor 1000 L capacity circular fibre-reinforced plastic (FRP) tanks, filled with 800 L with dechlorinated freshwater (10 mg l^-1 Cl^-) and equipped with continuous aeration supply through air-oxy tubes, were used for the study. Evaporation loss in the tanks was replenished fortnightly with dechlorinated water. Single batch of fingerlings each of striped catfish (19.2 ± 0.84 g), silver barb (20.5 ± 1.02 g) and stinging catfish (5.6 ± 0.28 g) were procured from ICAR-CIFA farm and acclimatized to the experimental condition for seven days prior to release in the tanks. The three different species were randomly distributed at a density of 50 juveniles of each species in every tank (187.5 juveniles m^-3). The nine tanks were grouped into three treatments namely T1 (Control; only feed), T2 (feed + floc) and T3 (only floc). Fishes of T1 and T2 were fed with commercial floating pellet feed (28% CP, 4% fat) at daily ration of 3% of biomass, split in two meals (09: 00 h and 15: 00 h). The daily ration was further adjusted based on the fish biomass estimated during fortnight sampling. No chemicals, antibiotics and other medicines were used in the tanks during the study period. A graphical picture of the experiment is provided as Fig. 1.

Fig. 1

Graphical abstract representing experimental design for biofloc based polyculture system (T1-only feed, T2-feed + floc, T3-only floc) with three different fish species namely Puntius gonionotus, Pangasianodon hypophthalmus, and Heteropneustus fossilis and their respective results of growth, digestive enzymes and gut microbiome

Biofloc preparation

Preparation for biofloc development in tanks was commenced three days before fish stocking. The required quantity of molasses (200 g; 34% organic carbon), ammonium sulphate (10.

g m^− 3) and ready-to-use probiotic (1 g m^− 3; CIBAFLOC™) were mixed in a container with dechlorinated water (10 mg l^− 1 of Cl⁻) and kept for 24 h under vigorously aerated condition before adding to the treatment tanks. Subsequently, molasses and probiotic (1 g m^− 3) were added in the tanks every day till the desired floc volume (12–15 ml l^− 1) is achieved, followed by discontinuation. Daily requirement of molasses was determined based on its carbon content and percentage of carbon and nitrogen in the supplied feed so as to maintain C: N ratio at 10:1. Floc volume was monitored every day and addition of molasses and probiotic resumed when there was a decline in the floc below the desired level. Bottom siphoning was carried out for sediment removal from tanks only once after two months and the water loss was compensated with dechlorinated water.

Water quality analysis

Water quality parameters in the tanks were analysed fortnightly. Temperature, dissolved oxygen and pH were measured using battery operated probe (HANNA, USA). The total alkalinity was measured following APHA [13] and was monitored intermittently for continued maintenance in the range of 80–100 mg as CaCO₃ l^-1 with lime (CaO) application. Total ammonia nitrogen (TAN), nitrite (NO₂-N) and nitrate (NO₃-N) concentrations were measured by using Photometric Ammonium test method (Sigma- Aldrich, 1.14752.0001, Batch. No: HC316909), Photometric Nitrate test method (Sigma- Aldrich, 1.14773.0001, Batch. No: HC153335) and Photometric Nitrite test method (Sigma- Aldrich, 1.14776.0001, Batch. No: HC299570) respectively, with the spectrophotometer (Thermo Scientific, USA). Besides the scheduled fortnight sampling, frequent interim measurement of all these parameters were also done to monitor the tank environment.

Fish sampling for growth assessment

Fish sampling was carried out in the tanks at month intervals. Fifteen fishes of each species were collected from each tank to record their weight. Further, the body surface was checked for presence of any parasites or lesion. At the end of the study, final body weight (FBW), % weight gain (WG), specific growth rate (SGR), and survival rate (SR) were calculated using following formulae:

% WG = (final body weight - initial body weight) x 100 x (Initial body weight)⁻¹.

Specific growth rate (SGR) (% day^− 1)) = {Ln (Final length) – Ln (Initial length)}x 100 x (days of culture)⁻¹.

Survival Rate (%) = (number of fish harvested) ×100 x (number of fish stocked) ^-1.

Feed conversion ratio (FCR)= (feed intake) x (weight gain of fish) ^-1.

Proximate analysis

The floc samples were randomly collected from single replicate of each treaments T2 and T3 at the end of experiment and sun dried. The proximate composition. was examined as by Swain et al. 2022 [14]. The Kjeldahl method (nitrogen × 6.25) was used to determine crude protein using the Kjeltec System (Tecater 1002 Distilling Unit), while crude fat content was analyzed using ether extraction using the Soxtech System (Tecater 1043 Extraction Unit). The ash content was determined by incineration in a muffle furnace at 550 ± 10 °C for 12 h. The crude protein and crude fat content of floc varied from 22.27 to 23.05% and 2.65 to 2.84% in T3 and T2. However, the ash content varied from 22.30 to 24.44% in T2 and T3 respectively.

Microscopy of Gill architecture

The fishes were anaesthetized by using bath treatment with 120 mg l^− 1 tricaine methane sulphonate (MS-222; Sigma-Aldrich, USA) before sacrifice. Fresh gill samples were collected by dissecting out from the opercular region for microscopic observation and measurement. The gap between the gill rackers and gill filaments were measured from freshly prepared wet mount of gill clip using a stereo microscope (Wild Heerbrugg, Switzerland) and camera (BestScope; BUC2E-530 C) to correlate the gill architecture with their floc filtration pattern.

Histology of gut

At the end of the study, intestinal villus size was measured in fishes of different treatments to correlate the growth and digestive enzyme of fish with gut histology. For histological analysis of gut, samples of the three species were randomly collected from one replicate of each treatment at the end of the study after anaesthetizing the fish as mentioned earlier. Briefly, gut samples were dissected, and midgut portion were preserved in neutral buffered formalin. After 72 h of fixation, the tissue was washed under tap water, dehydrated and cleaned with xylene prior to paraffin embedding. Sections of 5 µ thick were prepared and stained in H & E and mounted with DPX solution. The stained slide was observed under light microscope (Olympus; RXLr-5).

Digestive enzyme activities study

To estimate the different enzyme activities, the intestine tissue was aseptically dissected from fishes of each replicated tank, pooled together, and stored at -20⁰C until determination of enzyme activities. Frozen pooled samples of intestine were pulverized hygienically in liquid nitrogen using mortar and pestle. 5% crude enzyme extract with chilled 0.25 M sucrose (w/v) in ice cold conditions was prepared. The homogenate was centrifuged at 10,000 g for 15 min at 4 °C; the supernatant was collected and stored at − 80 °C for estimation of enzymatic activities. Total protein concentration in enzyme extracts was estimated [15], using bovine serum albumin as a standard. The intestinal amylase activity was assayed using soluble starch (1% w/v) as a substrate [16]. The protease activity was quantified using casein digestion method [17] while intestinal lipase activity determined with slight modification of the method given by German and Bittong [18].

Fish and floc sample collection for DNA isolation

One individual of each species was randomly selected from each replicate tank in every treatment at the end of culture period for gut microbiome analysis. Briefly, sampled fishes were anaesthetized as mentioned earlier and midguts were dissected. Replicate samples of each species were pooled for every treatment and were processed for DNA isolation using Xploregen tissue gDNA extraction kit (Xploregen, India) as per the instructions provided. Similarly, the concentrated floc samples were collected from replication tanks and were pooled in treatments for DNA isolation using the same protocol as explained for the gut tissue. The Extracted DNA from the samples were quantified using Nano Drop (Thermo Fisher Scientific, USA) reading at 260 nm and quality was checked at 260/280nm with a range of 1.8 to 2. The DNA quality was also check by using gel electrophoresis on 8% agarose.

Polymerase chain reaction (PCR) and next-generation sequencing (NGS)

For PCR reaction, 40 ng of extracted DNA was used for amplification of V3-V4 region of small ribosomal subunit, along with 10pM of each primer 16sF:- 5’ AGAGTTTGATGMTGGCTCAG3’ and 16sR:- 5’ TTACCGCGGCMGCSGGCAC3’ in taq master mix (high-fidelity DNA Polymerase, 0.5mM dNTPs, 3.2mM MgCl₂, PCR Enzyme Buffer). The amplification condition was set as initial denaturation − 95 ⁰C 25 Cycles of the following condition, denaturation @ 95⁰C for 15 s, annealing @ 60⁰C for 15 s, elongation @ 72⁰C for 2 min, final extension at 72⁰C for 10 min and hold at 4⁰C. The amplified 16s PCR Product was subjected to gel check and around 600 bp size amplicons were purified with Ampure beads to remove unused primers. Additional 8 cycles of PCR was performed using Illumina barcoded adapters to prepare the sequencing libraries. Libraries were purified using Ampure beads and quantitated using Qubit dsDNA High sensitivity assay kit. The sequencing work was performed by Biokart (India) company using Illumina Miseq with 2 × 300PE v3-v4 sequencing kit.

Bioinformatic analysis

The data received from the sequencer was de-multiplexed into fastq raw data and quality checked using Fastqc (Version 0.11.9) and Multiqc (Version 1.10.1) tools. The QC passed samples were qualified for further analysis using meta genomic pipeline (chimera detection-OTU clustering-pick representative e-sequences-assign taxonomy-taxonomic table). The sequence reads shorter than length 20 bp and merged paired end reads shorter than 400 bp were discarded during data trimming and the sequences with 97% similarity were grouped as OTU. Before the downstream statistical analysis the data were filtered for low count at a 20% prevalence with the minimum count of 4 and the low variance was filtered based on inter-quantile range of 10%. The final raw OTU table were used for data analysis. The NCBI refseq database is used for the OTU selection. The abundance feature tables in each sample were analysed using Microsoft excel (2016). For richness and alpha diversity analysis, the filtered data were measured in 4 different methods i.e. Chao1, Shannon, Simpson and Fisher and statistical significance was measured using ANOVA. The beta-diversity of the bacterial communities was investigated through a principal coordinate analysis (PCoA) performed with Bray-Curtis index distance method based on Permutational MANOVA (PERMANOVA). The other analyses like core microbiome, alpha diversity, beta diversity was built using Microbiome analyst (https://www.microbiomeanalyst.ca/).

Data analysis

To meet the requirements for homoscedasticity and normality, the data were subjected to arcsine transformation and analysed by one-way analysis of variances (ANOVA) followed by Duncan’s multiple range test (DMRT) in Statistical Package for the Social Sciences (SPSS) version 24.0. Differences were deemed significant at p ≤ 0.05. The correlation between the core bacterial groups and environmental factors was estimated using the Pearson’s rank correlation matrix presenting the correlation efficiency using the “corrplot” packages in R studio.

Results

Growth attributes

The Table 1 depicts the various growth attributes of the three species in the treatments. No mortality was observed in any species in treatments (100% survival) during the study period. Growth performances of all the three species were significantly higher in T2 followed by those in T1 and T3 (Fig. 2c). The weight gain of striped catfish, silver barb and stinging catfish in T2 were 26.01 ± 0.29, 28.19 ± 0.31 and 8.8 ± 0.1, respectively (P < 0.05). The lowest growth was recorded in all three species when reared only with floc in T3. A similar trend was also observed in the specific growth rate of the three species among the treatments. The FCR varied (p < 0.05) from 1.66 ± 0.09 in T2 to 2.25 ± 0.13 in T1 (Fig. 2b). The harvested fish biomass differed significantly among the treatments and depicted in Fig. 2a. The maximum biomass was harvested in T2 (5415 ± 139.63 g) followed by T1 (4634 ± 180.64 g) and T3 (3469 ± 144.86 g).

Table 1 Growth attributes of the three freshwater species reared over a period of 90 days in treatments (T1-only feed, T2-feed + floc, T3-only floc) in biofloc system

Fig. 2

(a) Harvested biomass (g) (b) FCR and (c) Growth curves of three freshwater fish species reared in different treatments over a period of 90 days (T1-only feed, T2-feed + floc, T3-only floc). Bar with different letters differ significantly at p < 0.05

Water quality indices

The mean water quality indices like temperature, dissolved oxygen (DO), pH, total alkalinity (TA), total ammonia nitrogen (TAN), nitrite (NO₂-N) and nitrate (NO₃-N) concentrations are depicted in Table 2. The temperature varied between 28.5 and 31.5 °C in the treatments during the study period. No significant difference in the DO, pH and TA levels were observed among treatments as desirable levels of these parameters were maintained in tanks through continuous aeration and periodic liming intervention. The TAN and nitrite level were significantly higher in T1 followed by T2 and T3. However, nitrate concentrations did not vary significantly among treatments.

Table 2 Ranges of water quality parameters in tanks during rearing of three freshwater species for 90 days in treatments (T1-only feed, T2-feed + floc, T3-only floc) in biofloc system

Measurements of gill rackers, filaments and histology of gut

The gaps between the gill rackers and gill filaments were measured to identify the species’ capability for floc filtration in the biofloc system. The highest gap between the gill rackers and filaments were observed in stripped catfish (1245.81 ± 19.8 and 216.24 ± 30.56 μm, respectively) followed by stringing catfish and silver barb. However, the difference is not significant among stringing catfish and silver barb (Table 3).

Table 3 Gaps between gill rakers, gill filaments and size of intestinal villous of the three freshwater species reared over a period of 90 days in different treatments (T1-only feed, T2-feed + floc, T3-only floc) in biofloc system

Histology of the gut revealed the villus size of the species is comparatively higher in T2 (186.3 × 38.3, 270.4 × 151.8 and 82.25 × 45.21 μm in striped catfish, silver barb and stinging catfish, respectively) followed by T1. However, the villus in T3 are slightly atrophied and smaller in size (Table 3; Plate 1).

Plate 1

Gut histological sections of the three freshwater species after 90 days of rearing in different treatments (T1-only feed, T2-feed+ floc, T3-only floc) with measured villous length and width at 4X magnification. A- Striped catfish in T1; B- Striped catfish in T2; C- Striped catfish in T3; D-Silver barb in T1; E- Silver barb in T2; F- Silver barb in T3; G- Stinging catfish in T1; H- Stinging catfish in T2; I- Stinging catfish in T3

Digestive enzymes analysis

The amylase activities were significantly higher in all the three species in treatment T2 when co-fed with feed and floc. Similarly, protease and lipase activities were higher in these three species in T2, but the differences were significant (p < 0.05) only in striped catfish and silver barb (Table 4; Fig. 3). However, a comparison between T1 and T3 revealed similar activities of the three enzymes in all species except the significantly higher amylase activities (p < 0.05) in striped catfish in T1 and significantly lower amylase and lipase activities (p < 0.05) in silver barb in T3.

Table 4 Digestive enzyme activities of the three freshwater species after 90 days of rearing in different treatments (T1-only feed, T2-feed + floc, T3-only floc) in biofloc system

Fig. 3

Activities of different digestive enzymes (amylase, lipase and protease) of the three freshwater species after 90 days of rearing in different treatments (T1-only feed, T2-feed + floc, T3-only floc). (Bar with different letters differ significantly at p < 0.05)

Bacterial diversity and composition in floc and fish gut

Bacterial diversities and abundance in the floc and the fish gut are presented in Table 5; Fig. 4, respectively. The generated sequence raw data in the form of fastq were submitted to NCBI SRA database with reference ID PRJNA1106429. The library size overview (Fig. 5) indicates total 1,613,401 sequences to be obtained from floc and gut samples, with an average of 1,344,50 reads per sample (n = 12). The floc in T3 (with only probiotic and carbon source addition) have the maximum reads (6,606,18) whereas floc in T1 (only feed) represented the lowest number of reads (4,629,4). Total 1,620,333 operational taxonomic units (OTUs) were obtained at genus level. The Floc in T3(667368) and stinging catfish gut (131747) in T2 represented the highest OTUs compared to floc and fish gut samples of other treatments. The fish gut (irrespective of species) in T1 represented the highest alpha diversity values among the treatments (Table 5). Overall, there is a significant difference between the treatment wise diversity index like Chao 1, Shannon and Fisher (p < 0.05) (Table 6). The beta-diversity of the bacterial communities associated with treatment-wise floc and fish gut samples (Fig. 6) has revealed that major clusters were formed based on habitat (water and fish gut) and fish species. The first axis separates floc samples from fish gut samples with 28.1% total variation, except T1 and T3. The two PC axes explained 46.79% of total variation among all samples.

Table 5 OTUs value, chao1, Shannon, Simpson and fisher diversity index of each sample according to different treatments (T1, T2 and T3)

Fig. 4

Relative abundance of core bacterial groups at phylum (a) and genus (b) level of floc and gut samples the three freshwater species after 90 days of rearing in different treatments (T1-only feed, T2-feed + floc, T3-only floc)

Fig. 5

Overview of nucleotide library size in floc and gut samples of three freshwater fish species reared in BFS with varied treatments (T1-only feed, T2-feed + floc, T3-only floc)

Table 6 Statistical analysis of diversity index of each sample according to different treatments (T1, T2 and T3) expressed as (mean ± SD). P value < 0.05 (ANOVA test) is marked with different superscripts next to the values in same row

Fig. 6

Beta diversity of microbiota at genus level OTUs represented by principal coordinates analysis (PCoA) plot with Bray-Curtis index distance method based on Permutational MANOVA (PERMANOVA, F = 1.4355, R square = 0.34933, P = 0.07) statistical method between four sample (floc, silver barb: puntius, striped catfish: pangas and stinging catfish: singhi) in T1, T2 and T3.Each colour spot represents a sample as indicated in right side of figure based. An R-value [range (-1, 1)] greater than zero indicates that intergroup dissimilarities are greater than intragroup dissimilarities, while an R-value less than zero means that there are no intergroup differences

Total 25 different phyla and 532 genus level OTUs were observed out of which five phyla and 20 genera of bacteria were identified as core bacterial group, across the nine fish gut samples and three floc samples. The core phyla group across all samples with mean relative abundance were i.e. Proteobacteria (51.93 ± 26.35%), Firmicutes (19.36 ± 21.64%), Bacteroidetes (11.37 ± 7.36%), Fusobacteria (9.99 ± 17.22%) and Actinobacteria (6.12 ± 8.90%). The core phyla vary significantly with respect to different treatments. For example, the Proteobacteria abundance was higher in floc (76.26%) and stinging catfish gut (96.89%) samples of T3, whereas the Firmicutes dominated in floc (77.26%) and stinging catfish gut in T2 (10.73%). According to top 10 genus level grouping, bacteria in the fish gut and floc samples varied substantially in the treatments as well as fish species (Fig. 3). In case of silver barb and striped catfish, Pseudomonas dominated in the intestinal flora of T1 group, whereas Enterobacterales and Fusobacterium dominated in that of T2 and T3, respectively. In contrast, gut microflora of stinging catfish in T1 had the maximum proportion of Clostridium which is replaced by Rhizobiaceae in T3. Similarly, floc samples of T3 had higher proportion of Rhizobiaceae than floc in T2 whereas floc in T1 had Micrococcales and Enterococcus as major groups.

Discussion

Correlation of fish growth with gill architecture, gut microbiome and gut histology

Biofloc technology is thriving in the global inland aquaculture market, particularly in areas with limited water and high land expenses. The key influence on successful aquaculture system diversification is the selection of appropriate candidate species. Currently, the lack of information on suitable species in freshwater systems is the primary concern for the overall adaptability of the BFT technology. The current investigation envisaged evaluating three freshwater species: stripped catfish, stinging catfish, and silver barb which were tested under three different treatments (only feed, floc + feed and only floc). The 100% survival rate of the three species after three months of indoor tank culture irrespective of treatments (Table 1) suggested the suitability of these species not only for the feed-based system (T1), but also the suspended floc laden systems like BFS for their survival (T2 and T3). Further, significantly higher (p < 0.05) growth attributes such as FBW, WG and SGR of all the three species in T2 than those of T1 clearly delineated the added advantage of the biofloc in fish growth. The crude protein and crude fat content of the biofloc may be influenced the growth in T2 and T3. The supplementation of additional feed may be an added advantages for significantly higher growth in T2 than T1 and T3. Significantly lower FCR in T2 than T1 also indicated contribution of floc as additional nutrient source for growth. The harvested fish biomass varied significantly between treatments due to growth differences. Consumption of biofloc also might have imputed to higher growth due to the presence of immunostimulant compounds (peptidoglycan, lipopolysaccharides and glucan) and antioxidants in the floc. This apart high levels of crude protein in biofloc might have contributed for such higher growth of the fishes fed with biofloc along supplementary diet [19,20,21]. The current findings are in confirmation with several previous reports indicating that biofloc supplementation ultimately contributes to improving the growth, survival, feed utilization, and production of various cultured species like Nile tilapia [22, 23], shrimp [24,25,26], carp [27], sea cucumber [28], and catfish [29]. However, the lower growth indices of the three species in T3 with only biofloc compared to those in T2 might be attributed to either the suboptimal nutrient supply from floc, a possible lower digestibility of the floc or the antinutritional factors that might be present in it which needs investigation in the future studies. Nevertheless, such reduced growth in T3 suggested the essentiality of supplementary feed along with the floc to ensure higher yield in the BFS.

All the water quality parameters presented in Table 2 were within the suitable ranges for the three species cultured, as also reported earlier for temperature, DO, pH, total alkalinity, ammonia, nitrite and nitrate [30]. Further, continuous aeration helped in maintaining the temperature, pH and DO in the tanks stable and within the suitable range of the carp fry [31, 32]. Though there was a higher build up of the TAN, nitrite and nitrate towards end of second month, it was corrected through water exchange. Prevalence of a suitable growing environment for the three species is also indicated in all the three treatments from their 100% survival in all tanks and the growth increment during the study, attributed to management of the water quality and floc volume in the system through periodic liming and the water exchange.

One of the aims of the present study was to find correlation between the gill filament architecture with the weight gain as flocs are ingested mostly by filtration besides the foraging. The floc particles might get caught on gill filaments by sieve filtration and then carried to the esophagus for digestion [33]. The authors also speculate that the fish reared in BFS are fed floc by sieve filtration via gill. Microscopic study revealed smaller inter-raker and inter-filament gaps in the gill arches of silver barb followed by stinging catfish and stripped catfish (Table 3). Floc intake, based on the gill architecture, is therefore expected to be the highest in silver barb followed by the other two, while stinging catfish, being a bottom feeder, can also feed on the settled floc at the bottom. In contrast, gut microbiome of stinging catfish had the highest number of OTUs followed by silver barb and striped catfish in all the treatments including T3 (Table 5). Such result in stinging catfish, despite its wider gill inter-filament gaps, indicated additional floc entry through its feeding on the settled floc from the tank bottom. Further, availability of floc contributed 49 and 45% additional weight gain in striped catfish and stinging catfish, respectively, in T2 than T1 (Table 1), while the gain was only 18% higher in silver barb. Similarly, only floc in T3 contributed the lowest % weight gain in silver barb (48.8%) as compared to 71% in the other two species. As highlighted in most studies, bottom foraging is a common feeding strategy for catfish, although their feeding behavior can vary based on food availability [34]. The wider mouth, small esophagus and vertically displaceable buccal cavity enables stinging catfish as well as the striped catfish to efficiently suck the bottom detritus [35]. In addition, their muscular stomach could better digest the protein rich flocs compared to herbivore fish. While the higher additional weight gain and microbial abundance in stinging catfish can be reasoned to its efficient bottom collection of floc, occasional bottom as well as prompt catching of floating feeds might have enabled the striped catfish to gain higher weight% despite larger inter-filament gaps. Whereas the contrasting growth results in silver barb implicates two possible reasons: (i) its sieve filtration by gill filaments assisted in consumption of limited amount of suspended flocs, resulting in less bacterial abundance; (ii) being a herbivore with absence of stomach, ineffective digestion of the protein and fat rich biofloc could have led to less contribution towards weight gain.

Previous reports have suggested biofloc to help maintaining the integrity of intestinal tract through improving the structure of intestinal mucosa and villus height and hence, increasing the gut absorptive areas; and it also protect the intestinal epithelial cells covering the villi through the immunostimulatory effect [36]. The biofloc reared Oreochromis niloticus and Cyprinus carpio L. with suitable carbon sources have shown significantly higher villus height compared to control [37, 38]. In the present study also, histological observation revealed significant increase (p < 0.05) in villus height in T2 group (Plate1) which might have facilitated better nutrient absorption and immunostimulatory effect resulting in the significantly higher weight gain (Table 1).

Correlation of fish growth with gut enzyme

Many researchers have investigated the digestive enzymes of aquatic animals in biofloc systems [39, 40]. It is recognized that the production and activity of digestive enzyme are endogenously controlled by digestive system and physiology rhythms of the aquatic animals [41]. Few other studies have also showed that the production of digestive enzymes was continuous during the 24 h cycle of a day, and their production and activity were not strongly influenced by the feed or feeding frequency [41, 42]. Positive effect of biofloc on digestive enzymes activities have been reported in cultured shrimp [4]. It has been reported in tilapia (O. niloticus), trout (Oncorhynchus mykiss), and stinging catfish (H. fossilis) where probiotic supplementation has been shown to improve intestinal morphology, digestive enzyme activities, nutrient absorption and growth performance [43, 44]. Similarly in the present study, significantly higher amylase, protease and lipase activities in all three species in T2 (p < 0.05; Table 4) over T1 (only feed) indicated additional availability of digestive enzymes for digestion of feed and floc in the former. Such higher digestive enzyme activities in the fish reared in BFS might be attributed to the endogenous digestive enzymes production as stimulated by the biofloc and the exogenous digestive enzymes contribution from the floc microbes themselves [4]. Further, fish fed only feed (T1) and only floc (T3) showed almost similar activities of the three enzymes in striped catfish and stinging catfish. Whereas, silver barb showed significantly higher amylase and lipase activities in T3, revealing its higher dietary biofloc consumption that contributed for the higher digestive enzymes activities.

Bacterial community and diversity in floc and fish gut

The present study showed higher bacterial richness index (OTU) of the floc samples than those of intestine samples of all the species in every treatment. Similarly, the Shannon diversity index of floc sample in both T1 (feed only) and T2 (BFS) (Table 5) were higher compared to the fish intestinal sample in the respective treatments. Such results signify that bacterial density and diversity in rearing medium is higher compared to the fish intestine. Similar results were also revealed from amplicon sequence data of floc sample and intestine of the Litopenaeus stylirostris which explained that floc samples had a large pool of OTUs those were not found in the intestines [6]. Another report from outdoor pond culture system of pacific white shrimp also reported higher proportion of bacteria in rearing water compared to shrimp gut [45]. It indicates that host microbiome allows selective bacteria to colonise and establish in the gut. However, Kim et al. [46] reported contrasting results where the diversity and richness of intestinal sample in L. vannamei was higher compared to the rearing water which is mostly attributed to the role of carbon source. The contrast results are mostly due to differences in culture environment and applied technique in BFT.

The bacterial abundance in floc sample and intestines of all three species, as observed from the nucleotide read and OTU counts, significantly increased in T2 and T3 groups (Fig. 5) whereas their diversity decreased in comparison to the T1 (feed only) sample (Table 5). However, shrimp raised with in situ biofloc, supplemented with various carbon and nitrogen sources, displayed increased gut microbial richness and diversity compared to the control group [6, 46]. The differences in microbial diversity can be attributed to the addition of commercial probiotics, which likely led to the colonization of specific bacterial groups. According to Hooper et al. [47], the gut microbiota varied significantly due to the variation in age, diet and health status of animal. Sturgeon fed with different level fishmeal diet had significant difference in gut microbial diversity and composition [48]. Addition of wheat flour as a carbon source in BFT strongly correlated with Flavobacterial dominance in rearing water and weakly correlated with dominant gut bacterium Planctomycetes [46]. Based on the above evidence, it can be suggested that, addition of carbon source and commercial probiotic in BFS allows selective bacteria to colonize and proliferate in the rearing water medium and gut samples.

Gut microbial analysis of species revealed higher sequence reads and OTUs value (201768) in stinging catfish to have across all the treatments in comparison to silver barb and striped catfish. Yang et al. [49]. had explained that gut microbes are mostly transient and introduced from their environment based on their feeding habits. The richer microbial communities in stinging catfish in this study might be attributed to its omnivorous and bottom feeding habit that helped to accumulate more microbes in the control (T1) as well as biofloc tanks (T2 and T3). However, microbial diversity is higher in the silver barb, followed by the striped catfish. This difference in diversity may be attributed to the filter-feeding and swift feeding behaviors of the silver barb and striped catfish, which enable them to ingest both feed and suspended particles more frequently compared to the bottom-dwelling stinging catfish.

The beta diversity analysis in the present study indicated clear clustering according to the habitat i.e. floc and fish intestine (Fig. 6). In silver barb and striped catfish, the beta diversity was clustered distinctly than the stinging catfish which is strictly a bottom feeder, having distinct gut bacterial adaptability. Similar results were also shown by Kim et al. [50] where clustering in beta diversity plot was clearly based on habitats like water, stomach and intestine.

The major bacterial group analysis in gut revealed that in silver barb and striped catfish, Pseudomonas and Prevotella dominated in T1 (feed only) which were replaced by Enterobacterales and Fusobacterium in T2 and T3, respectively. Similar groups were also reported earlier in the gut microbiome studies of freshwater fish like grass carp, freshwater pufferfish and Nile tilapia which included Gamma-proteobacteria, Fusobacteria, Firmicutes, Betaproteobacteria, Clostridia and Deltaproteobacteria [51,52,53,54]. Food digestion, absorption and disease resistance are important functions of gut microbes. Being facultative anaerobes, Pseudomonas and Enterobacterales are having broad metabolic range to grow in the presence of different organic compounds like carbohydrate, protein and fat [55]. The higher proportion of Pseudomonas and Enterobcaterales in T1 and T2 in the present study may be attributed to the functional variabilities of these bacteria in nutrient metabolism. In contrast to the Proteobacteria, the Bacteroids and Fusobacteria have specific groups of metabolic enzymes and are mostly associated with carbohydrate metabolism. Yang et al. [53] reported that fish guts are rich in genes associated with carbohydrate metabolism and are positively correlated with a higher proportion of fusobacteria and bacteroids. These polysaccharide degrading bacteria were also dominated in T3 in the present study attributed to the regular enrichment of its water medium with carbohydrate sources like molasses. The present study with such varied abundance of the bacterial community in treatments thus corroborates the earlier reports of changes in the fish gut microbiota in accordance with rearing water medium [56, 57].

Further in the present study, gut bacterial colonization pattern varied in the stinging catfish than the other two species. In T3 supplied with only probiotic, the gut microflora of stinging catfish had major proportion of Rhizobiales which also had similar dominance in its floc sample, whereas in T1 receiving only feed, Clostridium were the major group which is associated with cellulose and fiber metabolism [58, 59]. Few earlier studies have reported the gut microbes to be clustered mostly based on the diet, regardless of species [60, 61]. Zhang et al. [62] and Zhou et al. [63] have reported benthic organisms to be the major source of microbes to colonise in the gut of bottom detritus feeding fish. Gao et al. [5] reported predominant gut bacteria of sea cucumber to be like the microbes rich in marine sediments i.e. gamma-proteobacteria. The varied gut microbial colonization of stinging catfish across treatments in the present study thus indicated feeding behaviour of the species (surface or bottom feeding) to be an attribute to determine the pattern of gut microbial composition.

Pearson’s rank correlation matrix (Fig. 7a and b) showing a positive and negative correlation between each factor (growth, enzyme activity, and core bacterial group) in each treatment in different species of fish was generated. The correlation matrix highlights that due to the functional diversity of enterobacteria in nutrient metabolism it shows a positive correlation between protease and amylase digestive enzymes and growth across all treatments. The T1 and T2 groups are mostly dominated by phylum Proteobacteria which can grow on a range of organic compounds including protein, carbohydrates, and lipids and hence positively correlated with growth [55]. The T2 and T3 group dominated with nitrogen-fixing floc bacteria Rhizobiales which enhances fatty acid oxidation [64], and hence positively correlated with lipase. Whereas the high protein content in fish meal-based diet in the control group (T1) was positively correlated with protease which is also supported by Zu and Pan [4].

Fig. 7

Pearson’s rank correlation matrix presenting the correlation efficiency between mean weight gain of the three fish species (a), digestive enzymes such as amylase, lipase and protease (b) and core bacterial genera in each treatment. Colour bar (below) indicates effect size, with red denoting positive correlation and blue indicating negative correlation

Conclusion

The experiment’s findings suggest that relying solely on floc availability cannot meet the nutritional needs of fish in BFS, and feeding habit along with gill architecture play a major role in floc accumulation into the intestine. However, despite differences in feeding habits and gill architecture, all three species showed increased growth when fed with a combination of floc + feed. As per our hypothesis, gaps between the gill rackers have a major role in floc accumulation in the gut. However, contradiction result in silver barb with closer racker gaps and lower growth indicates possibilities of underutilization of the ingested biofloc of single cell protein as it is an omnivore with affinity towards plant material. Due to the bottom foraging behaviour, stinging catfish have resulted in higher biofloc intake, richer gut microbiome and improved growth, The closer inter filamentous gap in striped catfish, favorable environmental conditions in the presence of the floc and the effective utilization of both feed and floc helped in higher biofloc consumption and growth. Such results indicate better suitability of Stinging catfish and striped catfish in biofloc system as compare to silver barb.