Introduction

Bivalves are an important part of the marine ecosystem and have been widely used as food, providing a valuable source of protein for humans1,2. Global aquaculture reached a record high of 130.9 million tons in 2022, with 94.4 million tons coming from aquaculture animals, and 16.6% of this being aquaculture bivalves produced in China3. In addition to their food functions, bivalves have important ecological functions. Bivalve shellfish can sequester blue carbon and act as carbon sinks4,5. Global climate change, caused by emissions of greenhouse gases (especially carbon dioxide) from human activities in the past century, is one of the most significant challenges facing humanity in the twenty-first century6. Bivalve shellfish contribute to carbon sequestration in two main ways. Firstly, they convert bicarbonate (HCO3-) in seawater into calcium carbonate (CaCO3) shells through calcification. Secondly, they filter-feed on particulate organic carbon in the water column to synthesize their own organic material and increase their carbon content7. The faeces and pseudofaeces produced by these organisms are deposited on the seafloor, accelerating the transport of organic carbon to the seafloor.

The river contributes significantly to the estuary by depositing a substantial amount of silt, which results in a large expanse of shallow water. Additionally, the river discharges considerable organic matter and nutrients into the sea, enriching the estuary area with nutrients8. Geligang, located adjacent to the estuary of the Liao River and Shuangtaizi River, covers an area of approximately 10,000 hectares. During high tide, the entire region of Geligang is submerged, while certain portions become exposed during low tide9. Estuarine intertidal mudflats are crucial habitats for macrobenthos, serving as growth locations and breeding grounds10. Meretrix meretrix and Mactra veneriformis, belonging to the Mollusca, Lamellibranchia, and Veneroida orders, are common benthic economic shellfish species. They primarily inhabit the intertidal mudflats and shallow marine sediments11. These two species, M. meretrix and M. veneriformis, are widely distributed in both the intertidal and subtidal areas of Geligang and represent the dominant shellfish species in the region’s mudflats12. Consequently, Geligang has emerged as a significant breeding site for northern mudflat shellfish in China. However, in recent years, the quantity of shellfish resources for M. meretrix and M. veneriformis in Geligang has been declining annually due to the increasingly detrimental effects of human activityM. meretrixM. veneriformis13. This decline underscores the need for sustainable management practices to protect and preserve these valuable marine resources. The knowledge of the diet composition and feeding habits of aquatic species is essential for their resource conservation and ecological studies14,15. Therefore, it is crucial from an ecological perspective to understand the dietary characteristics of M. meretrix and M. veneriformis in Geligang and to provide fundamental knowledge for developing their resource enhancement and conservation strategies.

Information on food composition forms the basis for feeding ecology research and is a critical component in constructing high-resolution food webs16,17. M. meretrix and M. veneriformis are filter-feeding shellfish that consume a variety of suspended particles, including inorganic particulate matter, organic detritus, biological waste, bacteria, phytoplankton, and zooplankton in the water column18.Current studies on the diet of filter-feeding shellfish use microscopic examination19, fatty acid analysis20, and stable carbon and nitrogen isotopic analyses21.Microscopic examination is a widely used technique to determine the diet of organisms; however, it faces challenges in identifying smaller individuals such as pico- and nano-phytoplankton, and in distinguishing digested and semi-digested food in stomach contents22,23,24. While fatty acid and stable carbon and nitrogen isotopic analyses can characterize the extent of the food source, they often fall short in determining the precise composition of the food25. In recent years, Illumina Miseq-based high-throughput sequencing technology has been extensively applied to investigate the diet composition and feeding habits of aquatic species, such as Notorynchus cepedianus26, Crassostrea gigas27, Apostichopus japonicus28, Acanthopagrus schlegelii29, and Anguilla Anguilla30. With the ability to analyze food with varying degrees of degradation and species categorization at the species level, high-throughput sequencing offers greater sensitivity and accuracy. This enables more comprehensive and in-depth dietary assessments22,25.

In this study, we employed 23 S rDNA as a marker gene to characterize the food composition of M. meretrix and M. veneriformis by Illumina Miseq high-throughput sequencing. The results obtained from this analysis will provide a theoretical foundation for the conservation of their resources in Geligang. By employing this advanced sequencing technique, we aim to gain valuable insights into the dietary preferences and ecological roles of these shellfish species, ultimately contributing to their sustainable management and protection.

Results

Quantitative traits and delineation of OTUs. The dimensions of the M. meretrix were (5.36 ± 0.14) cm for shell length, (2.35 ± 0.16) cm for shell width, (4.34 ± 0.22) cm for shell height, and (42.29 ± 4.88) g for weight. The dimensions of the M. veneriformis were (3.25 ± 0.32) cm for shell length, (2.21 ± 0.30) cm for shell width, (3.20 ± 0.38) cm for shell height, and (15.86 ± 2.41) g for weight.

A total of 1,423,459 high-quality sequence fragments were obtained from 20 samples by Illumina high-throughput sequencing. The lowest sequence number of the samples was 36,957, the maximum was 173,507, and the average number of sequences was (71,173 ± 28,892). The 20 samples were clustered into OTUs after 97% concordance of the sequences, and a total of 2,485 OTUs were obtained. The results of OTUs identification and classification status are shown in Table 1. The number of OTUs shared by the M. meretrix and M. veneriformis was 956 (Fig. 1), of which 1,117 were unique to the M. meretrix and 412 to the M. veneriformis. The dilution curves of every sample leveled off as the sequence depth increased (Fig. 2), suggesting that there were enough samples in this study to capture the great majority of the information. The results of the alpha indices are shown in Fig. 3, with Chao1 and Observed species indices of M. meretrix significantly higher than those of M. veneriformis. Good’s coverage of M. meretrix was significantly lower than that of M. veneriformis, while the Shannon, Simpson and Pielou’s evenness indices of M. meretrix and M. veneriformis were not significantly different.

Fig. 1
figure 1

Venn diagram showing the number of the overlapping OTUs between M. meretrix and M. veneriformis.

Full size image
Fig. 2
figure 2

Sample rate faction curves.

Full size image
Fig. 3
figure 3

Alpha diversity for M. meretrix and M. veneriformis.

Full size image
Table 1 Statistics of operational taxonomic units (OTUs) identification and classification status for M. meretrix and M. veneriformis.
Full size table

Identification of eukaryotic species in stomach contents. A total of 50 bait organisms from 11 phyla were identified in the total sample, along with a few OTUs that were unrelated to any known taxonomic groups. The primary taxa of the stomach contents in the M. meretrix were Chlorophyta, Cryptophyta, Pyrrophyta, and Bacillariophyta, with relative mean abundances of 44.66%, 20.36%, 12.31%, and 9.22%, respectively. Similarly, the primary taxa of the stomach contents in the M. veneriformis were Cryptophyta, Chlorophyta, Pyrrophyta, and Chrysophyta, with relative mean abundances of 44.66%, 20.36%, 12.31%, and 9.22%, respectively (Fig. 4). Euglenozoa, Haptophyta, Xanthophyta, Streptophyta, Chordata, and Cercozoa make up a smaller proportion. The main groups in the Chlorophyta are Haematococcus, Nephroselmis, Pyramimonas, and Chlorella; the main group in the Cryptophyta is Cryptomonas; the main group in the Pyrrophyta is Dinophysis; Bacillariophyta with Thalassiosira as the major group; Chrysophyta with Ochromonas.

Fig. 4
figure 4

Species composition of eukaryotes for M. meretrix and M. veneriformis.

Full size image

Compositional characteristics of stomach contents. NMDS analyses based on Bray-Curtis distances found that samples from M. meretrix were grouped together, indicating low variability in the stomach contents composition, while samples from M. veneriformis were found to be more distant from one another, indicating greater variability in the stomach contents composition (Fig. 5). Figure 6 displays the results of the clustering heat map analysis. Adonis test results showed significant differences (R2 = 0.263, P < 0.05) in the composition of the stomach contents of M. meretrix and M. veneriformis.

LEfSe analysis allowed the identification of species with significant differences in abundance between groups31. The results of LEfSe analyses showed that the stomach contents of M. meretrix had fifteen genera with LDA values greater than 2 (Fig. 7), namely Thalassiosira, Resultomonas, Pseudochlorella, Navicula, Monomorphina, Rhodomonas, Micractinium, Lobosphaera, Fragilaria, Phaeodactylum, Ulnaria, Chaetoceros, Climaconeis, Gloeochaete, and Lepocinclis; while the stomach contents of M. veneriformis had seven: Ochromonas, Cloniophora, Nannochloropsis, Aphanochaete, Euglenaria, Micromonas, and Nitzschia.

Fig. 5
figure 5

NMDS analysis of stomach contents for M. meretrix and M. veneriformis.

Full size image
Fig. 6
figure 6

Cluster heat map of species composition of stomach contents for M. meretrix and M. veneriformis.

Full size image
Fig. 7
figure 7

LEfSe analysis of species composition of stomach contents for M. meretrix and M. veneriformis.

Full size image

Discussion

Filter-feeding shellfish are ecologically significant species that not only provide immense economic benefits but also play an important ecological role in driving the transport and transformation of various forms of carbon32,33. Filter-feeding shellfish consume a wide range of suspended particles, including organic debris, phytoplankton, and zooplankton, from the water column. The distribution of phytoplankton in the water column will be directly impacted by the filter-feeding behaviors of shellfish, which will, in turn, affects the level of marine primary production and ultimately influences the flow of energy and materials throughout the marine ecosystem34. Filter-feeding shellfish have selective feeding on phytoplankton. Zhang et al.35 discovered that Chlamys farreri preferred Bacillariophyta to Pyrrophyta, Bougrier et al.36 discovered that Crassostrea gigas was able to selectively excrete Skeletonema costatum, Chaetoceros calcirruns, and Nitzschia closterium via pseudofaeces, and Jiang et al.37 discovered that Argopecten irradians favored ingesting micro- and nano- phytoplankton with larger particle sizes. However, it has been proposed that selective feeding characteristics of filter-feeding shellfish may be affected by the geographical environment. Laboratory experiments often fail to accurately capture the way in which shellfish consume aquatic phytoplankton in their natural environment38,39. In this study, we selected in situ M. meretrix and M. veneriformis from Geligang for high-throughput sequencing to investigate their feeding ecological characteristics. This approach aims to provide a more accurate understanding of how these shellfish species interact with their natural environment and contribute to marine ecosystem dynamics.

High-throughput sequencing provides a solution to these problems by identifying not only the major food components but also tiny, highly digestible bait organisms40,41. With advancements in high-throughput sequencing technology, its application has expanded widely in feeding ecology, food web structure, species identification, and classification of aquatic organisms42,43,44,45,46,47. For example, Qiao et al.48 used the 23 S rRNA gene as a molecular marker to explore differences in the feeding selectivity of Mercenaria mercenaria, M. meretrix and Ruditapes philippinarum in natural waters. Their results showed that these three filter-feeding shellfish exhibite different preferences for phytoplankton genera, with some picophytoplankton dominating the hepatopancreas samples. Similarly, Chen et al.49 utilized the mitochondrial cytochrome c oxidase subunit I (CO I) as a molecular marker to identify the food composition of Diaphus splendidus in the South China Sea. They identified Ostracoda, Copepoda, and Amphipoda as the dominant groups in its food composition, and also discovered jellyfish that were not identified through traditional morphological identification. Li et al.50 employed the 18 S rDNA high-throughput sequencing method to study the food composition and feeding selection of Hyporhamphus sajori juveniles at different developmental stages under pond culture in the northern Yellow Sea. Their study revealed differences in the food composition of H.sajori juveniles at different developmental stages and detected Fungi, Intramacronucleata, and Streptophyta, which were not identified by traditional morphological identification. In this study, we used 23 S rDNA as a molecular marker to investigate the diet of M. meretrix and M. veneriformis. Our findings showed that the primary taxa in the stomach contents of M. meretrix were Chlorophyta, Cryptophyta, Pyrrophyta, and Bacillariophyta. For M. veneriformis, the primary taxa were Cryptophyta, Chlorophyta, Methanophyta, and Chrysophyta.

Liu et al.51 employed the 18 S rDNA high-throughput sequencing method to study the feeding selectivity of R.philippinarum and found that R.philippinarum actively avoided Bacillariophyta, Chrysophyta, and Cryptophyta, while it preferentially selected Pyrrophyta and Chlorophyta. Similarly, Zhang et al.52 also employed the 18 S rDNA high-throughput sequencing method and noted that Bellamya aeruginosa preferred feeding on Chlorella and Euglena, showing lower selectivity for Cryptophyta and Oscillatoriales. In this study, there was a significant difference in food composition between M. meretrix and M. veneriformis. M. meretrix preferred Bacillariophyta, Chlorophyta and Cryptophyta, whereas M. veneriformis favored Chrysophyta. The filter-feeding shellfish employ two primary feeding mechanisms: one relies on the coordinated movement of gill filaments and cilia to capture food particles53, hydrodynamic transport of water with different food concentrations for bivalve filtration. While the other utilizes water flow to transport food particles through the gill filaments to the back of the gills and along the dorsal grooves into the labellum54, selective feeding by bivalves. Both species are passive filter feeders, The ecological niches occupied by M. meretrix and M. veneriformis in this study are basically the same, so their feeding selectivity may be related to differences in their gill structures. Filter-feeding shellfish with large anterolateral cilia, such as M.mercenaria, retained microalgae larger than 4 μm with 100% efficiency. In contrast, species with small or no anterolateral cilia, like Crassostrea virginica and A. irradians, have retention efficiencies of only 75-85%55. The gill structure of the M. meretrix belongs to the heterorhabdic gill type. The gill filaments are divided into main gill filaments and common gill filaments. The common gill filaments have dense anterior and lateral cilia on the surfaces, which can effectively retain particles larger than 8 μm in size56. On the other hand, the gills of M. veneriformis are divided into inner and outer gills, which are petal-shaped. The inner gills are larger than the outer gills and have dense cilia, allowing them to feed on small nutrients flowing into the outer coat cavity. Moreover, the concentration of phytoplankton and their nutritional content can also impact the filter-feeding shellfish’s ability to selectively feed57,58.

High-throughput sequencing can detect DNA sequences present in low quantities at a reduced cost. The number of DNA sequences can reflect the relative amount of bait ingestion59,60. However, it cannot provide absolute quantification due to the overestimation of OTUs for species61. Additionally, a portion of the OTUs in this study were not annotated to the species level. There is still much taxonomic information need to be added to the present incomplete 23 S rDNA database62. In addition to phytoplankton, filter-feeding shellfish also consume zooplankton and organic detritus. However, there are few studies on the potential diets of these filter-feeding shellfish63. In high-density aquaculture, the selective feeding of filter-feeding shellfish on different groups of phytoplankton can lead to changes in the structure of the phytoplankton community64. In future study, it is also necessary to improve the coverage of taxonomic information, enhance the 23 S rDNA database, and increase the identification of potential food sources for zooplankton and other filter-feeding shellfish. Additionally, establishing a long-term monitoring mechanism to observe the succession mechanism of phytoplankton communities is crucial. Furthermore, improving the analysis of the feeding habits of M. meretrix and M. veneriformis is essential to provide valuable data for the study of their feeding ecology and aid in their scientific management and conservation.

Conclusion

In this study, we employed high-throughput sequencing of 23 S rDNA as a molecular marker to analyze the diet composition and feeding habits of M. meretrix and M. veneriformis in Geligang. Our findings revealed that the main taxa in the stomach contents of M. meretrix were Chlorophyta, Cryptophyta, Pyrrophyta, and Bacillariophyta. For M. veneriformis, the predonminant taxa were Cryptophyta, Chlorophyta, Pyrrophyta, and Chryptophyta. NMDS analyses based on Bray-Curtis distances found that samples from M. meretrix were grouped together, indicating low variability in the stomach contents composition, while samples from M. veneriformis were found to be more distant from one another, indicating greater variability in the stomach contents composition. The adonis test results indicated significant differences in the composition of the stomach contents between the two species. M. meretrix showed a preference for Bacillariophyta, Chlorophyta, and Cryptophyta, while M. veneriformis favored Chrysophyta. This study provides fundamental insights for research on the feeding ecology and resource conservation of M. meretrix and M. veneriformis in Geligang.

Materials and methods

Sample collection. As seen in Fig. 8, the M. meretrix and M. veneriformis were randomly collected from the Geligang of Liaodong Bay in June 2023. A total of 10 M. meretrix and 10 M. veneriformis were chosen for predation investigation. Following the collection of the shellfish samples, they were immediately cleaned with water and subjected to standard biological measurements (shell length, shell width, shell height and weight). The stomach contents were sampled, immediately preserved in liquid nitrogen and returned to the laboratory for DNA extraction.

Fig. 8
figure 8

Location map of Geligang. Map was drawn by authors, using ArcGIS 10.2 (Environmental Systems Research Institute, USA. https://www.esri.com).

Full size image

DNA extraction. The stomach contents were homogenized after being washed with phosphate buffer (pH 7.2–7.6). The whole DNA was extracted according to the instructions using the Marine DNA Extraction Kit (Tiangen, Beijing, China). The acquired total DNA was measured using a UV spectrophotometer, identified by electrophoresis using 0.8% agarose gel, and the qualified DNA was kept at −20 °C.

High-Throughput Sequencing. The PCR reaction was conducted in a 20 µL system and sample DNA was amplified by PCR using the 23 S rDNA primers p23SrV_f1 (5′-GGA CAG AAA GAC CCT ATG AA-3′) and p23SrV_r1 (5′-TCA GCC TGT TAT CCC TAG AG-3′)65. 10 ng of DNA template, 0.5 µL of Q5 high-fidelity Taq enzyme (NEB, Beijing, China), 1 µL of each of the positive and negative primers (5 µmol/L), 4 µL of 5× buffer, and 2 µL of dNTPs (2.5 mmol/L) were used. With a reaction cycle of 29, the PCR process was as follows: pre-denaturation at 95 °C for 7 min, denaturation at 95 °C for 45 s, annealing at 55 °C for 30 s, extension at 72 °C for 45 s, and final extension at 72 °C for 10 min. Then the PCR products were identified using 2% agarose gel electrophoresis, and purified by, and purified by AxyPrep DNA Gel Extraction Kit (Axygen, Silicon Valley, USA) before being submitted to bipartite sequencing on the Illumina Miseq sequencing platform (Illumina, San Diego, USA).

Data analysis. Firstly, the primer fragments of the sequence were removed using cutadapt (v2.3, http://cutadapt.readthedocs.io/en/stable/)66, and then the raw sequencing data were analyzed by Vsearch (v2.13.4, https://github.com/torognes/vsearch)67 for quality control procedures such as quality filtering, noise reduction, splicing and chimera removal to obtain high-quality sequences. QIIME2 (v2019.4) was used to combine the sequences and classify them into operational taxonomic units (OTUs) based on 97% sequence similarity. Representative sequences within the OTUs were chosen for species annotation analysis using the NCBI locally-accounted sequence database (v20230626, ncbi.nlm.nih.gov/). OTU analyses can provide information on the diversity of bait organisms and the relative abundance of different species.

The alpha diversity index (Chao1 index, Good’s coverage, Simpson index, Pielou’s evenness, Shannon index, Observed species index) was calculated using QIIME2. The Kruskal-Wallis test and Dunn’s test were used to determine whether group differences were statistically significant (P < 0.01 indicates highly significant differences, and P < 0.05 indicates a significant difference). Non-metric multidimensional scaling (NMDS) and hierarchical clustering plots based on Bray-Curtis similarity coefficients were plotted using the vegen package of R 3.6.2 and combined with permutational multivariate analysis of variance (Adonis) to analyze the significance of the differences between the stomach content composition of the M. meretrix and M. veneriformis. Linear discriminant analysis Effect Size (LEfSe) was performed by the microeco package of R 3.6.2 to identify differential bait organisms in M. meretrix and M. veneriformis, with the LDA threshold set at 2.