Intestine bacterial community affects the growth of the Pacific white shrimp (Litopenaeus vannamei)

Abstract

Abstract

Increasing evidence suggests that intestine microorganisms are closely related to shrimp growth, but there is no existing experiment to prove this hypothesis. Here, we compared the intestine bacterial community of fast- and slow-growing shrimp at the same developmental stage with a marked difference in body size. Our results showed that the intestine bacterial communities of slow-growing shrimp exhibited less diversity but were more heterogeneous than those of fast-growing shrimp. Uncultured_bacterium_g_Candidatus Bacilloplasma, Tamlana agarivorans, Donghicola tyrosinivorans, and uncultured_bacterium_f_Flavobacteriaceae were overrepresented in the intestines of fast-growing shrimp, while Shimia marina, Vibrio sp., and Vibrio campbellii showed the opposite trends. We further found that the bacterial community composition was significantly correlated with shrimp length, and some bacterial species abundances were found to be significantly correlated with shrimp weight and length, including T. agarivorans and V. campbellii, which were chosen as indicators for a reverse gavage experiment. Finally, T. agarivorans was found to significantly promote shrimp growth after the experiment. Collectively, these results suggest that intestine bacterial community could be important factors in determining the growth of shrimp, indicating that specific bacteria could be tested in further studies against shrimp growth retardation.

Key points

• A close relationship between intestine bacterial community and shrimp growth was proven by controllable experiments.

• The bacterial signatures of the intestine were markedly different between slow- and fast-growing shrimp, and the relative abundances of some intestine bacterial species were correlated significantly with shrimp body size.

• Reverse gavage by Tamlana agarivorans significantly promoted shrimp growth.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00253-023-12897-3.

Introduction

Intestine microorganisms serve as virtual endocrine organs that can provide key metabolic pathways that play critical roles in maintaining host growth (Sommer and Bäckhed 2013; Tap et al. 2015; Dvergedal 2020; Choi et al. 2021). For instance, intestine microorganisms can affect the growth rates of bovines across different developmental stages, and some specific microorganisms are critical to host growth, such as Oscillospira, which can utilize host-derived glycans and partly explains leanness (Fan et al. 2021). In this regard, it is important to explore the tight link between host growth and intestine microorganisms. Aquatic products, as an important source of high-quality protein, are crucial for maintaining food security and obtaining nutrition in both developed and developing countries (FAO 2018). Recently, accumulating evidence has revealed a tight connection between intestine microorganisms and the host growth of aquatic animals. One study revealed that intestine bacterial diversity is higher in identically strategy farmed abalones with slower growth velocity, and discrepancies in bacterial compositions in the intestine are obvious between large and small abalone cohorts (Choi et al. 2021). Another study demonstrated that several intestine bacterial species are associated with carbon metabolism in adipose tissue and feed efficiency of juvenile Atlantic salmon, while other species are associated with host weight gain (Dvergedal 2020). Thus, the intestine microorganisms are indeed closely related to the growth of aquatic animals.

Shrimp is one of the most popular aquaculture products in the world (Wang et al. 2021a). Unfortunately, growth retardation has hindered the healthy development of shrimp farming industry in recent years. Thus, it is very important for scientific research and commercial farming to explore the causes and solutions of this dilemma. Likewise, the growth performance of shrimp is related to intestine microorganisms (Tacon et al. 2002). Some studies found that the structures and functions of the intestine microbiota of shrimp were different among shrimp with retarded growth, those that were overgrown and normal cohorts from the same cultural pond, and the marked changes in intestine microbiota contribute to marked variations in host digestive activities, which may further affect host growth (Dai et al. 2017, 2022). These findings have improved our understanding of the close association between intestine microorganisms and shrimp growth. However, no controllable experiment has been designed to verify whether intestine microorganisms are closely related to shrimp growth and whether intestine microorganisms actually affect host growth.

This study aimed to explore the relationship between intestine bacterial community and host growth of the Pacific white shrimp Litopenaeus vannamei. We hypothesized that the structure and function of the intestine bacterial community were significantly different between fast- and slow-growing shrimp and that specific bacteria in the intestine affected host body size. To test this idea, we conducted a controllable experiment to compare the intestine bacterial communities between fast- and slow-growing shrimp at the same developmental stage by full-length 16S rRNA genes amplicons sequencing. Furthermore, we selected several culturable bacterial species in the shrimp intestine that were associated with host body size to test whether these species indeed promoted or inhibited growth. This study provides a useful reference for the close link between intestine microorganisms and host growth and could aid in the prevention and treatment of shrimp growth retardation in the future.

Materials and methods

Experimental design and sample collection

We designed a controllable experiment in this study (Fig. 1): (i) First, all fifty L. vannamei groups were incubated into larvae on the same day, and all groups were from the Haida Group Co. Ltd (Guangzhou, China). During the experimental period, the cultural management strategy for all shrimp was consistent: water temperature varied with the indoor natural temperature, and the salinity was set at 30‰; each tank had the same breeding density with 30 shrimp per tank. After beginning to eat feed, these shrimp were given four commercial feeds at 7:00, 11:00, 15:00, and 20:00 every day (~1 g feed for each tank), and the residual feed and feces in the aquarium were cleaned at 8:30, 12:30, 16:30, and 21:30. (ii) Two months later, three fast-growing (numbered 0711, 1011, and 4012 and assigned to the Fast group) and three slow-growing (2612, 2711, and 8203; the Slow group) shrimp groups were selected, and these six different groups were from different tanks. The body length and weight of shrimp in 0711, 1011, and 4012 were much higher than those in 2612, 2711, and 8203. (iii) After 1 month, fifteen shrimp from the Fast or Slow group were randomly selected, their body length and weight were measured, and intestine samples were collected. Shrimp body length and weight in the Fast group were both significantly (P < 0.01) higher than those in the Slow group (Supplemental Table S1 and S2). Some studies have shown that Enterocytozoon hepatopenaei (EHP) affects shrimp growth (Chayaburakul et al. 2004; Patil et al. 2021). For this reason, we adopted some strict careful management in our experiments. Firstly, shrimp seedlings without specific pathogens such as EHP were adopted. All foreign substances such as seawater and feed were disinfected and tested for pathogens. Then, the access of other personnel except laboratory personnel to shrimp breeding laboratory was restricted. We further found that the experimental shrimp was not infected with EHP. The method of intestine collection is detailed in our previous study (Hou et al. 2020). Intestine samples were stored in a −80°C refrigerator prior to nucleic acid extraction.

Fig. 1.

Diagram of experimental design

Full-length 16S rRNA genes amplicons sequencing and bioinformatic analysis

Genomic DNA of all samples was extracted by the PowerFecal DNA Isolation Kit (Mobio, Carlsbad, CA, USA). Then, the 27F (5’-AGRGTTTGATYNTGGCTCAG-3’) and 1492R (5’-TASGGHTACCTTGTTASGACTT-3’) primers were used to amplify the bacterial full-length 16S rRNA gene with the following program: 5 min at 95°C; 30 cycles of 95°C for 30 s, 50°C for 30 s and 72°C for 1 min; and 7 min at 72°C. The subsequent construction of an amplicon sequencing library was employed as described elsewhere (Wang et al. 2021b). The qualified library was sequenced on a PacBio Sequel II platform (Pacific Biosciences, Menlo Park, CA, USA) by Biomarker Technologies Co. Ltd. (Beijing, China), and single-end reads were generated.

The raw sequences were filtered and demultiplexed by SMRT Link (version 8.0) (https://www.pacb.com/support/software-downloads/) with minPasses ≥ 5 and minPredictedAccuracy ≥ 0.9 to obtain the circular consensus sequencing (CCS) reads. Then, CCS assignment of corresponding samples was performed using Lima (version 1.7.0) (https://github.com/PacificBiosciences/barcoding). The Cutadapt quality control process (version 2.7) (http://cutadapt.readthedocs.io/en/stable/) was employed to discard the CCS reads with no primers and whose length was between 1200 and 1650 bp. Chimeric sequences were detected and removed using the UCHIME algorithm (V8.1) (Edgar et al. 2011), and clean reads were then obtained. Then, the sequences with similarity ≥ 97 % were clustered into the same operational taxonomic unit (OTU) by USEARCH (V10.0) (Quast et al. 2012). The SILVA database (release 132) (https://www.arb-silva.de/documentation/release-132) was used to annotate OTUs taxonomy based on the RDP Classifier (version 2.2) (http://sourceforge.net/projects/rdpclassifier/) (Wang et al. 2007). The analysis of diversity (Caporaso et al. 2010) and construction of a phylogenetic tree (Huson et al. 2007) were employed as described in other studies. Finally, the function prediction of bacterial community was gained by Tax4Fun (Aßhauer et al. 2015).

Statistical analysis

Venn analysis was used to compare the intestine bacterial OTU numbers between the Fast and Slow groups (Huang et al. 2021). Student’s t-test was employed to analyze the differences in α-diversity indices between these two groups, as well as the relative abundances of genera and species (Sun et al. 2022). The unweighted pair-group method with arithmetic means, nonmetric multidimensional scaling (NMDS), and analysis of similarity (ANOSIM) analyses were performed to evaluate the community differences of these two groups based on the Bray Curtis distance (Hou et al. 2018b, 2021). Principal component analysis (PCA) was employed to analyze the similarity of predicted function content (level 3) between the two groups, and then the Student’s t-test was used to identify differentially functional genes (Hou et al. 2018a). Spearman correlation analysis (Mo et al. 2018) was employed to evaluate the relationship between bacterial dissimilarity and the Euclidean distance of host body size and between the relative abundances of bacterial species and body size.

Experiments verify the influence of indicator species on shrimp growth

Finally, we selected culturable indicator species (the bacterial strains were purchased from the Marine Culture Collection of China and which isolated from L. vannamei intestine) and used reverse gavage (to ensure that each shrimp ingested the same amount of target bacteria) (Huang et al. 2020) to test whether these bacteria affected shrimp growth. Shrimp with at same growth stage and with similar body weight and body length specifications were randomly selected and divided into the control and experimental groups. The body length and body weight of each shrimp in each group were measured, and the differences in body weight and body length between any two groups were compared by Student’s t-test (Sun et al. 2022). Then, each shrimp in the control or experimental groups was supplemented with 5 μL of PBS or bacterial solution (concentration was 2 × 10⁴ colony-forming units^.μL⁻¹ in PBS) by the reverse gavage. An automatic pipette was used to slowly inject 5μL liquid into the anal cavity of shrimp to avoid traumatic effects. During the experimental period, the cultural management strategy for shrimp was consistent: water temperature varied with the indoor natural temperature, and the salinity was set at 30‰; each tank had the same breeding density with 20 shrimp per tank. These shrimp were given four commercial feeds at 7:00, 11:00, 15:00, and 20:00 every day (~1 g feed for each tank), and the residual feed and feces in the aquarium were cleaned at 8:30, 12:30, 16:30, and 21:30. One week later (shrimp grows rapidly and is generally ready for market after 50 days in the farm, so this experiment was only conducted for one week), the host body length and weight were measured, and the differences of length or weight between any two groups were examined by Student’s t-test. During experiment, the death of shrimp in each group was recorded.

Results

Differences in intestine community diversity of the Fast and Slow groups

Our sequencing results showed that a total of 175,634 valid reads (mean = 5754 ± 896) were identified from all intestines, with an average sequence length of 1454 bp (Supplemental Table S3). There were 668 bacterial OTUs in all intestines, among which were 494 and 472 OTUs of the Fast and Slow groups, respectively (Fig. 2A). Student’s t-test results showed that the OTU numbers in each sample of the Fast group (193 ± 19) were obviously (P < 0.001) higher than those of the Slow group (147 ± 36) (Fig. 2B and Supplemental Table S4). The α-diversity indices of the Fast group (Chao1 = 254 ± 34, Shannon = 0.94 ± 0.03) were also significantly (P < 0.001, P < 0.01) higher than those of the Slow group (185 ± 35, 0.89 ± 0.04) (Fig. 2C and Supplemental Table S4). For β-diversity, the UPGMA (Fig. 2D) and NMDS results (Fig. 2E) together showed marked variation in bacterial community structures in the intestine in the two groups, which was further corroborated by the ANOSIM (r = 0.515, P = 0.001) (Fig. 2E). Especially, the Bray-Curtis distances of the intestine community of the Slow group (0.77 ± 0.28) were significantly (P < 0.001) higher than those of the Fast group (0.56 ± 0.14) (Fig. 2F). Thus, the intestine bacterial communities of slow-growing shrimp had lower diversity indices and were more heterogeneous than those of fast-growing shrimp.

Fig. 2.

A Venn analysis results of bacterial OTU numbers in shrimp intestine in the Fast and Slow groups. B and C Statistical significance of intestine bacterial OTUs and α-diversity indices between two groups. D–F UPGMA, NMDS, and ANOSIM results of bacterial community structure in intestine of the Fast and Slow groups based on the Bray-Curtis distance. **: P < 0.01, ***: P < 0.001

Differences in bacterial taxonomic composition of the Fast and Slow groups

Candidatus Bacilloplasma, Vibrio, Formosa, Motilimonas, and Pseudoalteromonas were the five most abundant genera in the shrimp intestines of the two groups (Supplemental Fig. S1), and uncultured_bacterium_g_Candidatus Bacilloplasma, Vibrio sp., Bacteroidetes bacterium, Vibrio brasiliensis, Motilimonas eburnea, Pseudoalteromonas sp., Halocynthiibacter arcticus, uncultured_bacterium_f_Mycoplasmataceae, Pseudoruegeria aquimaris, and Tamlana agarivorans were the ten most abundant species (Fig. 3A and Supplemental Table S5). Student’s t-test results showed that the relative abundances of four genera and seven species in the intestine were significantly (P < 0.05 in all cases) different between the two groups: Candidatus Bacilloplasma and Tamlana were overrepresented of the Fast group, but Vibrio and Shimia behaved in the opposite way (Supplemental Fig. S2); T. agarivorans, uncultured_bacterium_f_Flavobacteriaceae, Donghicola tyrosinivorans, and uncultured_bacterium_g_Candidatus Bacilloplasma were overrepresented in the Fast group, but Vibrio sp., Shimia marina, and Vibrio campbellii behaved in the opposite way (Fig. 3B).

Fig. 3.

A Relative abundance of bacterial species in shrimp intestine. B Statistical significance of bacterial taxonomic composition in shrimp intestine of the Fast and Slow groups at the species level. *P < 0.05, ***P < 0.001

Differences in potential bacterial functions of the Fast and Slow groups

The PCA results indicated that the predicted functional contents of the bacterial community in the intestine of the Fast group were clearly separated from those of the Slow group (Fig. 4A). Several pathways were significantly enriched in the intestine of the Fast group (Fig. 4B), especially the genes associated with metabolism (e.g., pyrimidine metabolism) and genetic information processing (ribosome and aminoacyl-tRNA biosynthesis). In contrast, the genes associated with cellular processes (e.g., bacterial chemotaxis), environmental information processing (e.g., bacterial secretion system), and microbial metabolism in diverse environments were significantly (P < 0.001 in all cases) more abundant in the Slow group. Thus, the potential functions of the intestine bacterial community were markedly different between the fast- and slow-growing shrimp.

Fig. 4.

A PCA results of predicted functional content of bacterial community in shrimp intestine of the Fast and Slow groups. B Predicted bacterial functional gene comparisons in shrimp intestine of these two groups. *P < 0.05, ***P < 0.001

Relationship between shrimp body size with intestine bacterial taxa

We observed that there was a significant correlation between host body length (r = 0.242, P < 0.001) or weight variables (r = 0.251, P < 0.001) and intestine bacterial community dissimilarity (Supplemental Fig. S3). We further showed significant (P < 0.05 in all cases) relationships between some species and body length and/or weight (Table 1) in that the relative abundances of uncultured_bacterium_g_Winogradskyella, Tenacibaculum mesophilum, Aliiroseovarius sediminilitoris, Tenacibaculum lutimaris, Roseovarius aestuarii, and uncultured_bacterium_g_Candidatus Bacilloplasma were significantly positively correlated with both body length and weight of shrimp, while S. marina, uncultured_bacterium_c_WWE3, uncultured_bacterium_g_Thiothrix, Jiulongibacter sediminis, and uncultured_bacterium_f_Saprospiraceae were significantly negatively correlated. The relative abundance of one species (Flavobacteriaceae_bacterium_UJ101) was significantly negatively correlated with only the body length, while four (e.g., T. agarivorans) and eight species (e.g., Vibrio sp., V. campbellii) were significantly positively and negatively correlated with only body weight. Of interest, culturable T. agarivorans was enriched in the intestines of the Fast group, with relative abundances significantly increased with body weight, and this trend was opposite for S. marina and V. campbellii.

Table 1.

Spearman correlation relationships between relative abundances of bacterial species in the intestine and shrimp body length and/or weight

	Body length (cm)		Body weight (g)
	r	P value	r	P value
uncultured_bacterium_g_Candidatus Bacilloplasma	0.759	<0.001	0.823	<0.001
uncultured_bacterium_g_Winogradskyella	0.648	<0.001	0.611	<0.001
Tenacibaculum mesophilum	0.609	<0.001	0.615	<0.001
Aliiroseovarius sediminilitoris	0.496	0.005	0.545	0.002
Tenacibaculum lutimaris	0.485	0.007	0.432	0.017
Roseovarius aestuarii	0.413	0.023	0.385	0.036
uncultured_bacterium_c_WWE3	−0.598	<0.001	−0.496	0.005
uncultured_bacterium_f_Saprospiraceae	−0.499	0.005	−0.546	0.002
Jiulongibacter sediminis	−0.428	0.018	−0.388	0.034
Shimia marina	−0.383	0.037	−0.449	0.013
uncultured_bacterium_g_Thiothrix	−0.371	0.044	−0.372	0.043
Flavobacteriaceae_bacterium_UJ101	−0.432	0.017	−0.345	0.062
Tamlana agarivorans	0.022	0.352	0.483	0.001
uncultured_bacterium_f_Flavobacteriaceae	0.352	0.056	0.483	0.007
Primorskyibacter aestuariivivens	0.282	0.132	0.367	0.046
Sedimentitalea nanhaiensis	0.346	0.061	0.365	0.047
Vibrio sp.	−0.321	0.084	−0.484	0.007
Vibrio jasicida	−0.324	0.081	−0.449	0.013
uncultured_bacterium_f_Cellvibrionaceae	−0.171	0.367	−0.437	0.016
uncultured_bacterium_g_Halobacteriovorax	−0.170	0.368	−0.437	0.016
Vibrio sinaloensis	−0.178	0.347	−0.436	0.016
Vibrio rotiferianus	−0.173	0.362	−0.433	0.017
Vibrio campbellii	−0.173	0.360	−0.431	0.017
uncultured_bacterium_g_Rhodopirellula	−0.306	0.100	−0.423	0.020

T. agarivorans promote shrimp growth performance

Finally, we selected two culturable bacterial indicator species, T. agarivorans and V. campbellii, which were enriched in the intestines of fast- and slow-growing shrimp, respectively; their abundances significantly increased and decreased with body size, respectively) and used reverse gavage to test whether they affected shrimp growth (Fig. 5A). Before bacterial supplementation, there were no significant (P > 0.05 in all cases) differences in host body length and weight between any two groups (Fig. 5B, Supplemental Table S4). After 1 week, the body length and weight of the T. agarivorans group were significantly (P < 0.05) higher than those of the other two groups (Fig. 5B, Supplemental Table S6). In contrast, there was no significant (P > 0.05 in all cases) difference in shrimp body length/weight between the Control and V. campbellii groups (Fig. 5B, Supplemental Table S6). Thus, T. agarivorans could indeed promote shrimp growth.

Fig. 5.

A Diagram of reverse gavage experiment. B Before the bacterial supplementation, there were no significant differences in the body length and weight of shrimp between any two groups. One week after supplementation, the body length and weight of shrimp in Tamlana agarivorans group were significantly higher than those in other two groups. *P < 0.05

Discussion

Given the insufficient control for discrepancies in farming conditions, management strategies, diet, growth, development, and others that contribute decisive effects on intestine microorganisms of hosts, identifying the causality of the relationship between intestine microorganisms and host growth remains a challenge. This study conducted a controllable experiment, and the findings reinforced the evidence that intestine bacteria influenced shrimp growth.

In accordance with other aquatic animals (Sha et al. 2016; Xiong et al. 2017; Piazzon et al. 2020; Dai et al. 2022), we found that bacterial signatures in the intestine were markedly different between fast- and slow-growing shrimp. Moreover, we identified seven bacteria (e.g., uncultured_bacterium_g_Candidatus Bacilloplasma, T. agarivorans, Vibrio sp., and V. campbellii) of the intestine that can be used as indicators of shrimp growth. Thus, intestine bacteria may be involved in influencing shrimp growth. Some studies have revealed that Candidatus Bacilloplasma is a common genus in the intestines of some aquatic animals and is closely related to host health (Zhang et al. 2021). However, no study has focused on the effects of Candidatus Bacilloplasma on shrimp growth. T. agarivorans is also commonly found in the intestines of aquatic animals and can utilize trehalose, maltose, succinate, and so on in host intestines (Yoon et al. 2008). In addition, the relative abundances of some metabolism pathway genes (e.g., glycolysis/gluconeogenesis) were significantly higher in the intestines of fast-growing shrimp, which indicates that the fast-growing shrimp have more active energy metabolism in their intestines. Furthermore, we found that T. agarivorans could promote shrimp growth, reinforcing the evidence that intestine microorganisms could indeed influence host growth.

We also found that Vibrio sp. and V. campbellii were enriched in slow-growing shrimp intestines. Vibrio species are common opportunistic pathogens of aquatic animals (especially shrimp) (Huang et al. 2016; Dong et al. 2017; Li et al. 2017), and their enrichment in host intestines is generally accompanied by bacterial disease outbreaks (Hou et al. 2018a; Xiong et al. 2018). In our experiment, there was no disease outbreak in any of the experimental stages. One potential reason might explain our phenomenon: the adaptation of host intestine to opportunistic pathogens might come at the cost of impaired growth, and this cost was attributed to the damage of these opportunistic pathogens to the intestine epithelium, as a result affecting the nutrient absorption of host (Amadi et al. 2021). As we have observed, the intestines of slow-growing shrimp contain a relative abundance of opportunistic pathogens, and the host’s tolerance to these opportunistic pathogens may come at the expense of growth. However, our findings from reverse gavage experiment revealed that there was no significant inhibitory effect of V. campbellii on shrimp growth. Possible explanation is that another Vibrio species, namely Vibrio sp., exerted a more potent inhibitory effect on host growth due to its tenfold higher abundance compared to V. campbellii, and its negative correlation coefficient with body weight and length was also higher than that of V. campbellii. Unfortunately, Vibrio sp. is an uncultivable bacterium, and it should be cultured and investigated for its potential effects on shrimp growth.

Generally, some specific microorganisms in the intestines are selected by the hosts, and these microorganisms play crucial roles in host growth and health status (Burns et al. 2016). Thus, we raised an important question: how do hosts select these specific intestine microorganisms? Intestine microorganisms are a complex system that can provide indispensable growth- and health-related functions (Matamoros et al. 2013; Lee and Hase 2014). The coevolution among hosts and intestine microorganisms forms specific trophic symbiosis (LeBlanc et al. 2013; Morrison and Preston 2016). Namely, some intestine microorganisms are beneficial for host nutrition, and these taxa appear to be preferentially selected by the hosts via genetic factors. Recently, some studies have focused on host genetic effects on intestine microorganisms and found that host genetics exert lifelong effects on the intestine microbiota and their association with host growth and immunity (Ramayo-Caldas et al. 2016; Li et al. 2019; Piazzon et al. 2020; Fan et al. 2021). One study even revealed a clear linear relationship between some intestine bacteria (including fiber-digesting and pathogenic bacteria) and bovine breed composition (e.g., as detected by SNPs (single nucleotide polymorphisms)) (Fan et al. 2020). For shrimp, future work should focus on the effects of host genetic on intestine microorganisms and the efficacy of these effects on host phenotypes (e.g., health and growth).

In summary, our work clarified the close relationship between shrimp growth and intestine bacterial community. We mainly found that the intestine community of slow-growing shrimp was less diverse and more heterogeneous than that fast-growing shrimp, and some taxa’s abundances were markedly different between these two cohorts. Some species were enriched in fast-slowing shrimp intestines and caused more active energy metabolism. There was a significant positive correlation between body size variables and intestine bacterial dissimilarity, and some species’ abundances were significantly correlated with body size. We further found that T. agarivorans could actually promote shrimp growth. Our findings provide experimental evidence revealing that intestine bacterial communities are tightly linked to host growth and could help solve the problem of shrimp growth retardation.

Author contributions

DH, SW, HL, XJ, HL, CL, SW, JH, and ZH conceived and designed the research. DH, BY, SW, and HL conducted the experiments. DH analyzed the data. DH, JH, and ZH wrote and revised the manuscript. All authors read and approved the manuscript.

Funding

This work was financially supported by the Key-Area Research and Development Program of Guangdong Province (2021B0202040001 and 2020B0202010009); National Natural Science Foundation of China (31902392); Innovation Group Project of Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai) (SML2021SP203); and earmarked fund for CARS-48.

Data availability

The raw sequencing data acquired in this study were deposited at the NCBI under the accession code PRJNA788291.

Declarations

Ethical approval

This study does not contain any studies with human participants and endangered or protected species. No specific permits were required for the described field studies. No specific permissions were required for access to artificial aquarium in Maoming, Guangdong Province, China. This study was reviewed and approved by the ethics committee of Sun Yat-sen University.

Conflict of Interest

The authors declare no competing interests.