Effects of Shifts in Bacterial Community on Improving Water Quality and Growth Performance of Pacific Whiteleg Shrimp (Litopenaeus vannamei) in Biofloc Systems

Abstract

This study aimed to induce a shift in a bacterial community by adding substrate into a biofloc system to characterize this shift and estimate its benefits in improving water quality and aquatic animal growth. We compared the bacterial communities between two biofloc systems, either with (sB treatment) or without (nB treatment) the addition of substrate (elastic solid packing filler), and we also analyzed the effects of the shift on the water quality and growth performance of shrimp (Litopenaeus vannamei). Beta diversity analysis indicated that the bacterial communities in the two treatments were significantly different (Jaccard index 0.94 ± 0.01, pseudo-F = 3.96, p = 0.001). The addition of substrate showed significant positive effects on bacterial alpha diversity indices (Shannon, Heip, Pielou, and Simpson; p < 0.05) and the abundances of beneficial genera (e.g., Arenimonas, Arthrobacter, Exiguobacterium, Leadbetterella, Luteolibacter, Marinobacter, Nitratireductor, Novosphingobium, Thermomonas, Plesiocystis, and Rubrivivax; p < 0.05). In addition, the substrate also showed significant positive effects on water quality parameters (TAN, TSS, turbidity, biofloc volume, pH, and carbonate alkalinity; p < 0.05), and it also significantly improved shrimp zootechnical performance indices (survival rate, feed conversion ratio, and productivity; p < 0.05). Redundancy analysis revealed that 94.25–98.58% of the variation in the water quality and the shrimp growth performance between the two treatments could be attributed to the shift in bacterial composition and diversity induced by the addition of substrate. These findings characterize the shift in the microbial community in the biofloc system induced by the substrate, and demonstrate how this shift could be beneficial to the water quality and the growth performance of shrimp.

1. Introduction

Litopenaeus vannamei (Boone, 1931), the most important and valuable crustacean species worldwide [1], is a euryhaline species that can be cultured at salinities as low as 0.5‰ [2]. For this reason, along with its commercial value, this species is cultured in many inland areas of China [3]. Typically, such culturing involves the addition of artificial sea salt into freshwater ponds [4,5,6], which results in the salinization of receiving water bodies from the shrimp culture process [7].

Biofloc technology (BFT) is an innovative approach for raising Litopenaeus vannamei with minimal or zero water exchange [8]. This approach can reduce discharges of saline water from the shrimp culture process into inland zones [4]. In a system utilizing BFT (biofloc system), microorganisms are importance for maintaining good water quality and to promote the growth performance of cultured aquatic animals (e.g., tilapia, L. vannamei, Fenneropenaeus indicus, F. paulensis, Macrobrachium rosenbergii, common carp, African catfish, Goldern crucian carp, and Indian major carp) [9,10]. However, the bacterial community in a biofloc system is highly dynamic and vulnerable to culture operations. For example, manipulations of the carbon source [11,12,13,14,15,16,17,18,19] and the C:N ratio [20,21,22,23] change the bacterial communities in biofloc systems. Such a shift in a bacterial community has significant impacts on the water quality and growth performance of aquatic animals [21,22]. Therefore, identifying and characterizing the shift in bacterial communities beneficial to water quality and the growth of aquatic animals are helpful for the optimization of BFT [15,24,25,26]. Although bacterial roles in biofloc systems are well known [9,10,27], characterizations of shifts in bacterial communities towards beneficial outcomes in biofloc systems have not been well documented yet [28,29].

We hypothesized that an operation which improves water quality and/or animal growth performance does so by inducing a beneficial shift in the microbial community, given the community’s critical role in both aspects of biofloc systems. Therefore, in light of this hypothesis, it was assumed that a beneficial shift in the microbial community in a biofloc system could be induced by adding artificial substrate to the biofloc system [28,29]. This is because an artificial substrate can provide additional surface area and refuge, reducing cannibalism and competition for space and thereby mitigating water quality degradation and potentially improving animal growth [30,31,32,33,34,35,36]. Recently, studies have corroborated that artificial substrates change the structure of microbial composition and improve water quality in biofloc systems [28,29]. However, whether these effects are induced by the beneficial shift in bacterial community has not yet been documented. Therefore, further insights into the effects of such bacterial community shifts and their benefits are needed to characterize these beneficial shifts and establish a basis for community management in biofloc systems.

In this study, aiming to characterize the beneficial shift in the bacterial community of the biofloc system, we induced a shift in bacterial community by adding a substrate (elastic solid packing filler), and then characterized this shift. Furthermore, we analyzed, for the first time, the effects of a substrate-induced shift in bacterial community on water quality and shrimp growth during a nursery procedure (28 days) for L. vannamei postlarvae (PL) under a salinity condition of 5‰. The findings supplied deep insight on the beneficial shift in bacterial community which is helpful for the optimization of BFT.

2. Materials and Methods

2.1. Shrimp Acclimation, Experimental Design, Operation, and Management

The nursery experiment was conducted using the facilities of Hunan Bifuteng Eco-Agriculture Development Co., Ltd. (BEAD, Changde, China). L. vannamei (~PL10) stocked under a salinity condition of 5.0‰ was supplied by BEAD and acclimated for one week before the experiment.

Two biofloc systems were set up in the present study: one system adding substrate (sB treatment), and the other system without addition of substrate as the control (cB treatment). Each system was established in triplicated concrete tanks (width × length × depth = 2 × 2.5 × 1.3 m) aerated using a 3 kW Roots blower (Changsheng Mechanical and Electrical Equipment Co., Ltd., Zhangqiu, Shandong, China). In the sB treatment, the bottom and walls of each tank were fixed with elastic solid packing filler (nylon, 18 cm in diameter) as substrate, covering approximately 50.1% of the tank’s internal surface area below the water line (Figure 1). The elastic solid packing filler (Figure 1a), composed of nylon yarns radially fixed to a central rope forming an open space, has a low floc-capturing capacity similar to a large mesh (2.5 cm) fiber net [34]. This allows biofloc particles formed on the substrate to be released back into the water column, influencing the system’s bacterial community.

After substrate installation, each tank was filled with culture water (salinity 5.0‰, 5.0 m³ per tank, 1.0 m in depth, Figure 1) and prepared and sterilized according to a previous study [37], and then randomly assigned shrimp PL (0.0025 g) at a density of 4000 individuals m⁻³ for a 28-day nursery experiment. During the experimental period, PL were fed a commercial diet with a protein content of 40.0% (Alpha Group Co., Ltd, Shenzhen, China), at a frequency of four times a day (6:00, 12:00, 18:00, 24:00), based on the feeding table for shrimp raised under low-salinity conditions recommended by Van Wyk, Davishodgkins, Laramore, Main, Mountain, and Scarpa [2]. Glucose (food-grade, carbohydrate content 90.0%, Fufeng Biotechnology Co., Ltd., Hohhot, China) was applied as the exogenous carbon source to achieve an inputted carbon to nitrogen ratio (C:N) of 15:1, which was determined according to the procedure used in the previous study [37].

Throughout the whole experimental period, no water exchange was conducted, and only evaporative water loss was compensated with dechlorinated tap water (about 1% per week on average). Additionally, during the stage of 14–28 d of the experiment, partial biofloc was removed using a side-stream settling chamber equipped with a submersible pump (JGP-2500L, SENSEN Group Co., Ltd., Zhoushan, China), which operated for 6 h daily at a flow rate of 2.25 m³ h⁻¹ in both treatments to maintain a biofloc volume < 15 mL L⁻¹.

2.2. High-Throughput Sequencing and Data Processing

Throughout the 28-day nursery experiment, water (50 mL) from each tank was collected on a weekly basis, and pooled by treatment (4 samples per treatment). Then, the bacterial DNA genome was extracted and sequenced following the procedure described in our previous study [37]. Briefly, the V3-V4 region of the 16S rRNA gene of the bacterial DNA genome was amplified. Then, the product was purified, quantified, and normalized to construct a library for high-throughput sequencing on the Miseq platform (Sangon Biotech Co., Ltd., Shanghai, China). The data (58451-93572 reads per sample) were deposited to NCBI (PRJNA646765).

Sequencing reads were processed using the QIIME 2 platform (version 2019.10) as described previously [37]. Briefly, ambiguous, adapter, and primer nucleotides were discarded; bases with a quality score < 25 were trimmed; chimeras were filtered; and paired-end reads were merged and dereplicated. Thereafter, operational taxonomic units (OTUs) were constructed at the 0.97 level.

2.3. Determination of 16S rRNA Gene Copies

The 16S rRNA gene was counted using the quantitative real-time PCR method (qRT-PCR). In brief, the qRT-PCR was performed in a 10 μL reaction mixture with a LightCycler480 II Thermal Cycler (Roche, Rotkreuz, Switzerland), using extracted bacterial DNA from water samples as the template, and 341F: 5′-CCTACGGGNGGCWGCAG-3′ and 805R: 5′-GACTACHVGGGTATCTAATCC-3′ as the pair of primers (corresponding to the V3-V4 variable region of the 16S rRNA gene). The 10 μL qRT-PCR reaction mixture contained 5.0 μL 2× SybrGreen qPCR Master Mix (Roche, Rotkreuz, Switzerland), 0.2 μL of each of the 10 μM forward and reverse primers, and 1.0 μL template (extracted bacterial DNA). PCR was performed under the following thermocycler conditions: 95 °C for 3 min; then, 45 cycles of melting at 95 °C for 5 s; and finally, annealing/extending at 60 °C for 30 s, with a dissociation procedure according to instrument guidelines. All reactions were conducted in triplicate. The 16S rRNA gene concentration (copies per milliliter) of bacterial community in a sample was interpolatively calculated from the standard curves (R² = 0.998) generated by 10-fold serial dilution of a plasmid containing the V3-V4 region of the 16S rRNA gene with the ∆C_t (cycle threshold) method.

2.4. Analysis of Bacterial Composition Profile

OTUs were annotated and collapsed at the phylum, class, order, family, and genus levels with the taxa plugin of QIIME 2 (version 2019.10), with reference to the GreenGene database (13.8). The preliminary abundance of a bacterium in a sample was determined based on its proportion obtained from the annotation process and the total 16S rRNA gene copies of the sample. Thereafter, the preliminary abundance of each bacterium was normalized using the rrnDB database (version 5.6), and the absolute abundance was expressed as bacterial genome copies (normalized 16S rRNA gene copies) per milliliter, on which the proportions of bacteria were recalculated.

2.5. Analysis of Bacterial Diversity

The Jaccard index between bacterial communities of both treatments was used to analyze the beta diversity, and was tested with PERMANOVA [38]. Measures of bacterial alpha diversity (Shannon, Margalef, Berger–Parker, Simpson, Heip, and Pielou) were computed with the diversity plugin prepackaged in QIIME 2 (version 2019.10).

2.6. Water Quality Parameters

Dissolved oxygen and temperature were measured daily with a YSI analyzer (YSI Inc., Yellow Springs, OH, USA). The pH was measured daily using a pH-100 m (LICHEN Sci-Tech, Co., Ltd., Hangzhou, China). Other water parameters (e.g., the three inorganic nitrogen compounds) were determined weekly. In brief, water was sampled and divided into two parts. One aliquot was passed through a pre-weighted 0.45 µm filter membrane (Whatman). The filtrate was used to determine dissolved total nitrogen in water (DTN), total ammonia nitrogen (TAN), nitrite, nitrate, and carbonate alkalinity [39]. At the same time, the membrane was dried to determine the total suspended solid (TSS) content [39]. The remaining aliquot was used to determine whole total nitrogen in the water body (WTN) [39]. Based on the difference between WTN and DTN, we determined total nitrogen contained in biofloc (BTN), which represents the total nitrogen content contained in biofloc per liter of water. Biofloc volume (BFV), representing the volume of settleable solids, was measured weekly using an Imhoff cone where the water sample (1 L) was settled for 15 min [40]. Then, the supernatant in the Imhoff cone was used to spectrophotometrically measure the turbidity [39], which represents the content of unsettleable or small-size particles in the water column.

2.7. Zootechnical Indices

PL weight was monitored on a weekly basis. At harvest, survival rate (SR), weekly increment in body weight (WIW), specific growth rate (SGR), feed conversion ratio (FCR), and productivity were analyzed [40,41].

2.8. Statistical Analyses

Statistical analysis was performed using SPSS 22.0 (IBM Corp, Armonk, NY, USA). Student’s t-test was performed for comparisons of growth parameters between both treatments, after checking data normality and homogeneity with the Shapiro–Wilk test and Levene’s test, respectively. Otherwise, the non-parametric Mann–Whitney U test was used. Percentage data were arcsine-transformed, and bacterial genome copies were log(x + 1)-transformed prior to statistical analysis. For water parameters, alpha diversity indices, and bacterial genome copies (absolute abundances), repeated-measures one-way ANOVA was used to test the significance of main effects and interactions of substrate and time. Differences with a p value of <0.05 were considered to be significant. Biomarkers of bacterial community were detected with LEfSe (linear discriminant analysis effect size) [42]. Redundancy analysis (RDA) was performed using CANOCO 5.0 to estimate the effects of explanatory variables (e.g., counts of bacterial genera) on response variables (e.g., water parameters and zootechnical indices).

3. Results

3.1. Bacterial Composition Profile

RDA showed that the addition of substrate changed bacterial composition at the genus level (Figure 2), as well as at the levels of phylum, class, order, and family (Figure S1). Genera such as Exiguobacterium, Leadbetterella, and Novosphingobium were more abundant in the sB treatment, but Arthrobacter and Rubrivivax were more abundant in the cB treatment (Figure 2). LEfSe analysis revealed that the most significant biomarkers at the genus level were Fluviicola, Leadbetterella, Lutibacterium, and Marinobacter for the sB treatment, and Arthrobacter for the cB treatment (Figure 3).

Regarding the bacterial composition profile, OTUs were assigned to 30 phyla in the present study. The most dominant phyla in both treatments were Actinobacteria (4.0–22.7%), Bacteroidetes (10.4–33.5%), Firmicutes (0.2–11.2%), Planctomycetes (4.0–14.9%), and Proteobacteria (29.4–59.0%) (Figure 4a). In addition, Chloroflexi and Verrucomicrobia were the dominant phyla in the cB treatment (0.1–11.4%) and the sB treatment (0.3–16.2%), respectively (Figure 4a). Significant effects of the substrate were observed on proportions of these seven phyla (p < 0.05), except Firmicutes and Proteobacteria (p > 0.05, Figure 4b). At the genus level, the most dominant ones, such as Paracoccus, Pseudomonas, Rhodobacter, Flavobacterium, and Planococcus, are species belonging to the phyla of Proteobacteria, Bacteroidetes, and Firmicutes (Figure 5).

3.2. Absolute Abundances of Bacterial Taxa

Substrate had a significant effect on the absolute abundances of the dominant phyla of Actinobacteria, Chloroflexi, Firmicutes, and Verrucomicrobia (p < 0.05,

Table 1. Absolute abundances of dominant phyla and genera in the biofloc systems with (sB treatment) and without (cB treatment) addition of an elastic solid packing filler (nylon) as substrate during the 28-day experiment nursing Litopenaeus vannamei postlarvae at a salinity of 5.0‰.

Phylum	Genus	Absolute Abundance (105 Copies of Genome mL−1)		p-Value (Repeated-Measures ANOVA)
		cB	sB	Substrate	Time	interaction
Actinobacteria		23.81 ± 7.15	10.91 ± 3.74	<0.001	0.027	0.008
	Candidatus aquiluna	0.19 ± 0.05	0.49 ± 0.22	<0.001	<0.001	<0.001
	Microbacterium	0.85 ± 0.23	2.13 ± 0.69	0.001	0.019	0.038
	Mycobacterium	3.22 ± 1.25	1.85 ± 0.49	0.021	0.003	0.001
Bacteroidetes		40.98 ± 12.81	41.71 ± 34.45	0.948	0.029	0.183
	Adhaeribacter	0.10 ± 0.05	2.89 ± 1.68	0.041	0.041	0.038
	Aequorivita	2.14 ± 0.65	0.32 ± 0.16	<0.001	0.012	0.016
	Flavobacterium	1.36 ± 0.63	0.59 ± 0.12	0.002	<0.001	<0.001
	Haliscomenobacter	0.07 ± 0.03 a	0.65 ± 0.15 b	0.005	0.046	0.085
	Muricauda	3.00 ± 1.43	0.71 ± 0.22	0.027	0.016	0.039
Chloroflexi		6.88 ± 5.11	1.15 ± 0.53	<0.001	0.006	0.006
Firmicutes		7.18 ± 4.21	10.93 ± 6.39	0.012	0.009	<0.001
	Exiguobacterium	0.08 ± 0.02	0.84 ± 0.3	0.004	0.003	0.002
	Planococcus	1.08 ± 0.36	7.73 ± 2.68	<0.001	0.019	0.037
	Planomicrobium	0.11 ± 0.03	1.15 ± 0.45	0.001	0.013	0.021
Planctomycetes		14.98 ± 4.35	21.86 ± 11.38	0.096	0.071	0.100
	Planctomyces	0.35 ± 0.07	3.82 ± 2.11	0.043	0.043	0.040
Proteobacteria		73.82 ± 13.09	82.51 ± 39.86	0.389	0.064	0.209
	Amaricoccus	9.33 ± 4.17	2.88 ± 1.27	0.031	0.013	0.080
	Anaerospora	0.42 ± 0.19 a	1.07 ± 0.31 b	0.033	0.026	0.554
	Bdellovibrio	0.80 ± 0.17	1.01 ± 0.21	0.079	<0.001	0.566
	Brevundimonas	13.45 ± 4.41	3.02 ± 1.47	<0.001	0.007	0.007
	Halomonas	1.85 ± 0.54	0.37 ± 0.15	<0.001	<0.001	<0.001
	Hyphomonas	3.92 ± 1.06	2.01 ± 0.65	0.034	0.018	0.348
	Lutibacterium	0.44 ± 0.12 a	1.66 ± 0.38 b	0.008	0.038	0.249
	Lysobacter	0.27 ± 0.07	5.53 ± 2.65	0.020	0.051	0.049
	Paracoccus	3.98 ± 1.08	8.66 ± 2.77	<0.001	0.038	0.031
	Pseudomonas	2.99 ± 0.51	3.16 ± 1.52	0.834	0.070	0.029
	Psychrobacter	0.40 ± 0.06	6.31 ± 3.07	0.005	0.007	0.009
	Rhodobacter	5.49 ± 1.26	5.39 ± 1.92	0.931	0.034	0.068
Verrucomicrobia		1.15 ± 0.45 a	7.61 ± 2.86 b	0.001	0.007	0.072
	Luteolibacter	0.16 ± 0.05	1.59 ± 0.57	<0.001	0.019	0.030

Note: different letters for each genus between treatments mean significant difference.

). Substrate also showed significant effects on absolute abundances of the 25 most dominant genera (p < 0.05), except for Bdellovibrio, Pseudomonas, and Rhodobacter (p > 0.05, Table 1).

3.3. Bacterial Diversity

The Jaccard index for bacterial beta diversity between treatments was 0.94 ± 0.01 (pseudo-F = 3.96, p = 0.001, PERMANOVA,

Table 2. Beta diversity (Jaccard index) between the bacterial communities of biofloc systems with (sB treatment) and without (cB treatment) addition of an elastic solid packing filler (nylon) as substrate during the 28-day experiment nursing Litopenaeus vannamei postlarvae at a salinity of 5.0‰.

Group 1	Group 2	Jaccard Index	Permutations	Pseudo-F	p-Value
cB	cB	0.761 ± 0.208	-	-	-
sB	sB	0.768 ± 0.211	-	-	-
cB	sB	0.941 ± 0.007	999	3.96	0.001

), suggesting significant difference between the bacterial communities of the two treatments.

Regarding alpha diversity, substrate addition showed significant effects on the Shannon index, Heip index, Pielou index, and Simpson index (p < 0.05), all of which showed higher values in the sB treatment than the cB treatment (

Table 3. Bacterial alpha diversity indices in the biofloc systems with (sB treatment) and without (cB treatment) addition of an elastic solid packing filler (nylon) as substrate during the 28-day experiment nursing Litopenaeus vannamei postlarvae at a salinity of 5.0‰.

Index	Treatments		p-Value (Repeated-Measures ANOVA)
	cB	sB	Substrate	Time	Interaction
Shannon	6.14 ± 0.18	6.77 ± 0.07	0.002	<0.001	<0.001
Margalef	343.9 ± 22.0	387.7 ± 28.0	0.367	0.007	0.471
Heip	0.013 ± 0.001	0.018 ± 0.001	0.026	0.012	0.544
Pielou	0.493 ± 0.012	0.538 ± 0.003	0.024	<0.001	0.016
Berger–Parker	0.193 ± 0.020	0.152 ± 0.014	0.098	0.002	0.676
Simpson	0.927 ± 0.010	0.955 ± 0.003	0.043	<0.001	0.127

). Overall, the addition of substrate increased the diversity (Shannon index), evenness (Heip index and Pielou index) and richness (Margalef index) of the bacterial community, but decreased bacterial dominance (lower value of Berger–Parker index) in the sB treatment (Figure 6). The interaction of substrate and time was significant on the Shannon index and the Pielou index (p < 0.05, Table 3), indicating different temporal patterns for these indices between treatments (Figure 7). For instance, the Shannon index was highest on day 7 and then slightly decreased in the sB treatment, but it peaked on day 21 in the cB treatment (Figure 7a). The changing trend of the Pielou index was similar with that of the Shannon index in each treatment (Figure 7c).

3.4. Water Quality

Substrate showed significant effects on carbonate alkalinity, pH, TAN, BFV, TSS, and turbidity (p < 0.05,

Table 4. Water parameters in the biofloc systems with (sB treatment) and without (cB treatment) addition of an elastic solid packing filler (nylon) as substrate during the 28-day experiment nursing Litopenaeus vannamei postlarvae at a salinity of 5.0‰.

Parameters	Treatments		p-Value (Repeated-Measures ANOVA)
	cB	sB	Substrate	Time	Interaction
TSS (mg L−1)	148.5 ± 31.3 a	491.3 ± 150.2 b	0.009	0.116	0.594
WTN (mg L−1)	76.8 ± 11.2	76.3 ± 9.6	0.057	<0.001	0.013
DTN (mg L−1)	13.0 ± 2.6	14.7 ± 3.6	0.890	<0.001	<0.001
BTN (mg L−1)	63.8 ± 9.3	61.6 ± 12.5	0.052	<0.001	0.085
TAN (mg L−1)	0.55 ± 0.25	0.43 ± 0.06	0.010	0.227	0.002
Nitrite (mg L−1)	0.37 ± 0.28	0.10 ± 0.03	0.352	0.348	0.342
Nitrate (mg L−1)	0.51 ± 0.14	0.56 ± 0.08	0.924	0.274	0.252
BFV (mL L−1)	8.8 ± 2.3	8.3 ± 2.3	0.044	<0.001	<0.001
Turbidity (NTU)	111.9 ± 56.3	299.5 ± 92.0	0.002	0.005	0.059
CAK (mg L−1 CaCO3)	373.7 ± 12.4 b	306.5 ± 51.7 a	<0.001	<0.001	<0.001
Temperature (°C)	26.9 ± 0.7	27.1 ± 0.6	0.656	<0.001	0.012
pH	7.20 ± 0.14	7.11 ± 0.03	0.001	0.001	<0.001
Dissolved oxygen (mg L−1)	5.85 ± 0.32	5.63 ± 0.59	0.215	<0.001	0.764

Note: BFV, biofloc volume; BTN, the total nitrogen contained in biofloc; CAK, carbonate alkalinity; DTN, the dissolved total nitrogen in water; NTU, nephelometric turbidity units; TAN, total ammonia nitrogen; TSS, total suspended solids; WTN, the whole total nitrogen in water. Different letters for each parameter between treatments mean significant difference.

). In addition, the interaction of substrate with time was significant for WTN and DTN (p < 0.05, Table 4). Overall, substrate explained 33.6% of variation in water quality between the sB and cB treatments (Figure S2). Furthermore, shifts in bacterial composition and diversity following substrate addition accounted for 98.58% (Figure 8a) and 94.33% (Figure 8b) of the variation in water quality, respectively.

The level of dissolved oxygen was above 5.0 mg L⁻¹ and the recorded temperature was above 26.0 °C in both treatments without significant difference (p > 0.05, Table 4). Moreover, the concentrations of the three inorganic nitrogen compounds were at low levels in both treatments (<1.0 mg L⁻¹, p > 0.05, Table 4). Carbonate alkalinity was significantly lower in the sB treatment compared to the cB treatment (p < 0.05, Table 4). In contrast, TSS was significantly higher in the sB treatment than in the cB treatment (p < 0.05, Table 4). Turbidity was also higher in the sB treatment, although this difference was not significant (p > 0.05, Table 4).

3.5. Growth Performance

Shrimp survival rate and productivity were both significantly higher in the sB treatment compared to the cB treatment (p < 0.05,

Table 5. Zootechnical indices of Litopenaeus vannamei postlarvae nursed in the biofloc systems with (sB treatment) and without (cB treatment) addition of an elastic solid packing filler (nylon) as substrate after the 28-day experiment at a salinity of 5.0‰.

Parameters	cB	sB	p-Value
Final ABW (g)	0.36 ± 0.04	0.40 ± 0.03	0.596
SR (%)	81.0 ± 7.1	96.3 ± 3.6	0.011
FCR	0.98 ± 0.08	0.76 ± 0.06	0.044
WIW (g week−1)	0.090 ± 0.014	0.099 ± 0.008	0.652
SGR (% d−1)	17.7 ± 0.5	18.1 ± 0.4	0.581
Productivity (kg m−3)	1.14 ± 0.09	1.54 ± 0.12	0.029

Note: ABW, average body weight; FCR, feed conversion rate; SGR, specific growth rate; SR, survival rate; WIW, weekly increment in body weight. SR (%) = [(survival shrimp counts)/stocked shrimp counts] × 100. WIW (g week⁻¹) = (fbw − ibw)/culture weeks. SGR (%d⁻¹) = [(ln fbw − ln ibw)/culture days] × 100. FCR = total mass of feed given/gained shrimp biomass. Productivity (kgm⁻³) = (harvested shrimp biomass)/(culture water volume). Here, fbw and ibw represent the final and initial mean body weight of PL, respectively.

). FCR in the sB treatment was significantly lower than that in the cB treatment (p = 0.044, Table 5). Throughout the 28-day nursery experiment, the average body weight of shrimp in the sB treatment was higher than that of the cB treatment (Figure 9). Overall, substrate accounted for 41.9% of the variation in shrimp growth performance between the two treatments (Figure S3). Shifts in bacterial composition and diversity induced by substrate addition accounted for 96.65% (Figure 10a) and 94.25% (Figure 10b) of the total variation in shrimp growth performance between the two treatments, respectively.

4. Discussion

The results of the present study indicated that the addition of substrate successfully induced shifts in the bacterial composition and diversity in the biofloc system, which in turn was shown to be beneficial to the water quality and the growth performance of shrimp.

4.1. Bacterial Composition and Diversity of the Biofloc System Shifted After Addition of Substrate

A recent study found that the addition of substrates and carbon sources changes the bacterial community in a pond polyculture system [29]. Similarly, we also found that the bacterial composition profile in the biofloc system obviously changed after the addition of the substrate. In the sB treatment, the relative abundances of Planctomycetes and Bacteroidetes increased, while that of Actinobacteria decreased, when compared with the results in the cB treatment (Table 1). Chloroflexi, Bacteroidetes, and Planctomycetes contain attached-living species [43,44,45], whereas Actinobacteria and Firmicutes contain free-living or free-swimming species [45,46]. Substrates can provide surface area for attached-living species, potentially increasing their abundances [47], in agreement with the observations of the present study. However, we neither observed an increase in the abundance of the attached-living Chloroflexi nor found a decrease in the abundance of the free-living Firmicutes in the sB treatment (Table 1). One explanation for these discrepancies is that lifestyle is not the sole determinant of bacterial abundance, as substrate only explained 38.6% of the variation in bacterial composition (Figure 2). Other factors, such as water quality, could influence the bacterial composition [48,49,50], in line with the findings of the present study (Figure S4). It should be noted that the variation in water quality between both treatment is induced by adding substrate to the sB treatment in the present study (Figure 8). Overall, we can state that there was a shift in the bacterial community in the sB treatment regardless of direct or indirect effects of the addition of substrate, although we did not separate the effects of water quality on the shift in bacterial community from the effects of substrate.

At the genus level, substrate positively impacted abundances of Arenimonas, Bdellovibrio, Exiguobacterium, Fluviicola, Leadbetterella, Luteolibacter, Lutibacterium, Marinobacter, Nitratireductor, Novosphingobium, Plesiocystis, and Thermomonas (Figure 2 and Figure 3). These genera have potential positive implications for water quality, biofloc formation, and shrimp growth, due to their beneficial traits. For example, Arenimonas is involved in denitrification [10]. Bdellovibrio, a probiotic used in aquaculture, exhibits hydrolytic activity against pathogenic Aeromonas and Vibrio [51,52], leading to positive impact upon the survival and growth of aquatic animals [15,53]. Exiguobacterium, a dominant genus in biofloc systems [54,55,56], can potentially be used as a probiotic [26,57,58] due to its functions in nutrient metabolism and digestive enzyme secretion [59,60,61]. Novosphingobium, a common bacterium in biofloc systems [62], could be bioaugmented as a probiotic beneficial for the growth, survival, and immunological status of shrimp [63]. Thermomonas, a denitrifier [64,65] found in biofloc systems [11,66,67,68], has also been associated with trypsin, aminopeptidase, and alkaline protease activities [69]. Nitratireductor is capable of producing microbial flocculants in biofloc systems [9,70,71,72,73]. Similarly, Marinobacter is a biofloc-formation-related genus which is able to produce extracellular polymeric substances [74]. Leadbetterella, Plesiocystis, and Luteolibacter are all dominant bacterial genera in biofloc systems [75,76,77]. Little functional information is available for Fluviicola and Lutibacterium, which are reported here for the first time in a biofloc system. However, substrate showed negative effects on the abundances of the other two potential beneficial genera, Arthrobacter and Rubrivivax (Figure 2 and Figure 3). Arthrobacter is considered a probiotic in shrimp aquaculture [52,78,79,80], and Rubrivivax is a facultative anaerobic denitrifying bacterium commonly observed in biofloc systems [10,14,81].

Substrate addition also increased the bacterial alpha diversity (e.g., Shannon index, Margalef index, Heip index, Pielou index, and Simpson index) in the biofloc system (Table 3 and Figure 8b), leading to increases in bacterial diversity, evenness, and richness. The interactions of substrate and time on the Shannon index and Pielou index were also significant (p < 0.05, Table 3), indicating differences between the treatments in terms of changing patterns. The Shannon index peaked at day 7 in the sB treatment and day 21 in the cB treatment (Figure 7), suggesting that substrate accelerates the formation of a highly diverse bacterial community in the biofloc system. Furthermore, all alpha diversity indices were significantly affected by time (p < 0.05, Table 3), indicating a time-dependent succession pattern of bacterial diversity regardless of substrate addition, in line with the results of previous studies [49,50,82,83].

In terms of dominant bacteria, Actinobacteria, Bacteroidetes, Chloroflexi, Firmicutes, Planctomycetes, Proteobacteria, and Verrucomicrobia were the most abundant phyla in both treatments of the present study, in line with observations by researchers in other biofloc systems stocking whiteleg shrimp [12,37,49,84,85,86,87,88,89]. At the genus level, Amaricoccus, Paracoccus, Planococcus, Pseudomonas, and Rhodobacter were dominant in both treatments (Table 1), in line with the findings of previous studies [10,55,73,76,90,91]. Regarding bacterial diversity, the Shannon index in the present study (6.14-6.77) was within the range of 5.31–7.23 reported in previous studies [37,48,49,50,92,93,94].

4.2. Shift in Bacterial Community Induced by Adding Substrate Positively Affected Water Quality

The addition of substrate showed significant effects on water parameters (Table 4), and it was found to be responsible for 33.6% of the variation in water quality between the treatments in the present study (Figure S2). The effects may have contributed to the shift in bacterial community [21,95]. Shifts in bacterial community are known to significantly impact water quality in biofloc systems [20,21,22]. Previous studies have investigated the influence of water quality on bacterial composition. For example, salinity and electrical conductivity were the most significant environmental factors that influenced the microbial community, and nitrite, nitrate, and total organic carbon were positively correlated with the abundance of Paracoccus, but TAN was negatively correlated with the abundance of Leucothrix [48,49,50]. In the present study, we investigated the effects of the shift in bacterial composition on water quality, and found that the representative genera were Rhodobacter, Bdellovibrio, Devosia, Turneriella, and Moraxella (Figure 8a). We observed that the first three genera correlated positively with BFV, TSS, and nitrate but negatively with TAN and nitrite. In contrast, the remaining two genera correlated positively with pH and turbidity, but negatively with BFV and CAK, in line with the findings of previous studies. For example, Rhodobacter, a common genus in biofloc systems [20,69,96,97], has the ability to assimilate and dissimilate nitrate [65,98]. Devosia, a genus typically found in biofloc systems [69,73,75], is a nitrogen fixer [99]. As for Bdellovibrio, Turneriella, and Moraxella, though their effects on water quality are currently not clear, they are also dominant genera in biofloc systems [47,48,73,90,100,101].

We also found that the shift in the bacterial diversity in the biofloc system after the addition of substrate affected the water quality. The diversity (Shannon index) and the evenness (Pielou index and Heip index) of the bacterial community showed positive effects on BFV, TSS, and nitrate, but the richness index (Margalef) showed negative effects on those parameters (Figure 8b). The dominance of the bacterial community (Berger–Parker index) was positive to CAK, but negative for pH and turbidity (Figure 8b). To our knowledge, this is the first study to directly investigate the effects of bacterial diversity on water quality, although a diverse bacterial community has been qualitatively considered beneficial for maintaining good water quality in biofloc systems [24,25].

4.3. Shift in Bacterial Community Induced by Adding Substrate Improved the Growth Performance of Shrimp

The addition of substrate showed significant effects on SR, productivity, and FCR (Table 5), in line with the findings of the RDA that substrate contributed to 41.9% of the variation in zootechnical indices between the sB treatment and the cB treatment (Figure S3). Furthermore, shifts in bacterial genera following substrate addition accounted for 96.65% of the variation in shrimp growth performance between treatments (Figure 10a). Previous studies have recognized the important roles of microbial community on the survival and growth of cultured aquatic animals in biofloc systems [11,50,83,102]. Xu, Morris, and Samocha [21] and Panigrahi, Saranya, Sundaram, Vinoth Kannan, Das, Sathish, Rajesh, and Otta [22] observed that shifts in bacterial community impact shrimp growth in biofloc systems. Here, we further demonstrate a correlation between shifts in bacterial genus abundances and shrimp growth performance. We found that the most representative shifted bacterial genera were Haliscomenobacter, Truepera, Vogesella, Gillisia and Bdellovibrio. Haliscomenobacter, and Truepera, which showed positive effects on SR, productivity, and FCR (Figure 10a). Previous studies also found that Haliscomenobacter, found in biofloc systems [100], plays a central role in sludge bulking and grows on substrates [10]. Truepera is considered to be related to fish health [103]. However, Vogesella and Gillisia were found to negatively impact SR, productivity, and FCR in the present study (Figure 10a). Corroboratively, Vogesella has been reported to be pathogenic to fish [104,105], indicating harmful effects on growth. Gillisia is considered a producer of antimicrobial compounds and is involved in organic matter mineralization [106]. However, this is the first report of a negative effect of Gillisia on aquatic animal growth. Bdellovibrio was strongly correlated with ABW and WIW (Figure 10a), in line with previous findings that Bdellovibrio has a positive effect on the growth performance of aquatic animals [15,53].

It is generally considered that a diverse microbial community is important to the survival and growth of shrimp in a biofloc system [20,90]. In the present study, we further assessed the contributions of the shift in bacterial diversity to shrimp growth, and found that it accounted for 94.25% of variations in the zootechnical indices of shrimp between the sB treatment and the cB treatment (Figure 10b). We also observed that bacterial diversity (Shannon index), richness (Margalef index), and evenness (Heip index and Pielou index) showed positive effects on SR, productivity, and FCR, but the bacterial dominance was negative for those indices (Figure 10b).

In general, the growth performance of shrimp is usually considered to be most closely associated with water quality [50,102,107]. Corroboratively, in the present study, we found that changes in water quality affected the growth performance of shrimp (Figure S5), in line with the findings of a previous study [108]. Given the major contribution of bacterial community shifts to water quality variation (Figure 8), we propose that the effects on shrimp growth performance are likely indirect, mediated through changes in water quality.

4.4. Limitations and Future Directions

It should be noted that the 28-day trial period may be insufficient to fully validate the findings and hypothesis. Future studies should include a long-term grow-out period (e.g., at least 12 weeks) to obtain more robust insights into beneficial bacterial community shifts. Furthermore, biofloc removal may wash out bacteria, potentially biasing community analysis. Although the same removal procedure was applied to both treatments to mitigate this, future experiments should comprehensively integrate substrate and biofloc removal factors to fully account for such bias. Finally, the linkages between bacterial community shifts, water quality, and growth performance identified here are indirect. Future work should investigate gene expression related to nitrogen cycling (e.g., nirS/nirK) and host–microbe interactions (e.g., immune gene expression) to provide direct molecular insights into the mechanisms underlying beneficial bacterial community shifts in biofloc systems.

5. Conclusions

In the present study, we induced and characterized a shift in bacterial community by adding substrate, and verified the beneficial effects of this shift on the water quality and shrimp growth performance for the first time. This study provides significant insight into beneficial bacterial community shifts in biofloc systems, which will aid in the future management and optimization of BFT. However, long-term experiments and more precise analyses (e.g., variation partitioning, source tracking) are needed to directly link explanatory and response variables and reinforce these findings.