Environmental Exosomes/Small Extracellular Vesicles: Evidence of Extracellular RNA Release by Aquatic Organisms

Abstract

Aquatic organisms continuously interact with surrounding water, yet whether they release extracellular vesicles into this vast medium remains unknown. We hypothesized that pearl oysters (Pinctada fucata) release exosomes/small extracellular vesicles (sEVs) into the aquatic environment. To this end, we collected exosome/sEV-sized components by ultrafiltration from tank water and open-sea culture areas. Microscopy revealed abundant vesicles consistent with exosome/sEV size, and small RNA sequencing identified pearl oyster-specific PIWI-interacting RNAs (piRNAs) that matched sequences previously detected in hemolymph exosomes. These findings demonstrate that pearl oysters actively released exosomes containing species-specific nucleic acids into surrounding water, effectively protecting the RNA from rapid degradation. We propose referring to these vesicles as environmental exosomes/environmental sEVs (eExosomes/esEVs). These findings suggest that aquatic exosomes/sEVs serve as carriers of RNA and may contribute to inter-organismal communication networks. Beyond their functional role, eExosomes/esEVs also hold promise as highly stable, novel targets for environmental DNA/RNA (eDNA/eRNA) analysis, offering new opportunities for ecological monitoring and biodiversity research.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10126-026-10618-1.

Introduction

Aquatic organisms continuously exchange biological components with their surrounding waters. These exchanges can profoundly influence ecological interactions. We hypothesized that, in addition to soluble biomolecules, exosomes and small extracellular vesicles (sEVs) may also be secreted into aquatic environments. These vesicles could act as stable carriers of biological information not only within organisms (as they do in terrestrial species, transferring information among tissues and organs), but also between organisms. Exosomes and sEVs are lipid bilayer vesicles typically 50–200 nm in diameter (Pegtel and Gould 2019; Gurung et al. 2021; Nazri et al. 2023; Ważny et al. 2024; Gabaran et al. 2025; Zhou et al. 2025) known to encapsulate nucleic acids, proteins, and other bioactive molecules (Meldolesi 2018; Pegtel and Gould 2019; Kalluri and LeBleu 2020; Negahdaripour et al. 2021) and mediate a wide range of intercellular communication processes (Valadi et al. 2007; Meldolesi 2018; Pegtel and Gould 2019; Gurung et al. 2021). In terrestrial and biomedical research, extensive studies have revealed their critical roles in physiological regulation, disease mechanisms, and cell-to-cell signaling (Kim et al. 2015; Zhao et al. 2020; Dad et al. 2021; Moros et al. 2021; Chen and Li 2025). However, investigations into the roles of exosomes/sEVs in aquatic organisms remain limited. Most prior work has focused on intracellular or organismal functions, while their potential release into the external aquatic environment and possible contributions to inter-organismal communication remain essentially unexplored.

Extracellular vesicles, such as exosomes, possess a lipid bilayer that protects their enclosed nucleic acids and proteins from enzymatic degradation (Koga et al. 2011; Chen et al. 2024), potentially enabling their persistence in aquatic environments. If released into surrounding waters, these vesicles could serve as vehicles for the transfer of functional molecules between individuals, thus contributing to ecological communication networks. Our previous studies on the Akoya pearl oyster (Pinctada fucata) have demonstrated that hemolymph-derived exosomes are enriched in small RNAs, particularly PIWI-interacting RNAs (piRNAs; Huang et al. 2022), which are expressed in various somatic tissues of invertebrates (Huang et al. 2019, 2021). Notably, invertebrate piRNAs appear to serve broader functions, potentially resembling those of microRNAs (miRNAs; Huang et al. 2021). Empirical evidence supports this possibility; for instance, plant-derived miRNAs have been shown to regulate insect gene expression (Shi et al. 2024), and algae-derived exosomes containing miRNAs have been detected in P. fucata (Zheng et al. 2022). Based on these integrated biological foundations, we conceived the idea that aquatic organisms might continuously release exosomes/sEVs into the surrounding environment, functioning as highly stable, transmissive carriers for small RNA-mediated inter-organismal communication.

In recent years, environmental nucleic acids, especially eRNA in aquatic ecosystems, have emerged as powerful tools for assessing biodiversity and conducting ecological monitoring (Minamoto et al. 2021; Veilleux et al. 2021; Yates et al. 2021; Glover et al. 2025). Because RNA reflects the active physiological state of living organisms, eRNA applications have expanded from precise ecological monitoring (Miyata et al. 2021, 2025) to the detection of functional and stress responses via environmental transcriptomics (Hiki et al. 2023; Hechler et al. 2025; Hiki and Jo 2025). However, eRNA analysis faces significant technical challenges and limitations, such as the overwhelming dominance of non-target microbial RNA (Hiki et al. 2023; Hechler et al. 2025; Hiki and Jo 2025) and inherently rapid degradation (Wang et al. 2025). Unlike eDNA, free eRNA is highly susceptible to rapid degradation depending on environmental factors such as temperature and pH (Kagzi et al. 2022; Jo et al. 2023), with the half-lives of some specific transcripts estimated to be as short as a few tens of minutes (Aminaka et al. 2025). Consequently, while controlled tank and laboratory experiments have provided valuable insights into eRNA dynamics, standardized protocols and successful applications in complex natural field environments remain scarce (Bunholi et al. 2023). While some researchers in the eRNA field have recently begun to speculate that eRNA persistence might depend on protection by biological structures like extracellular vesicles (Jo et al. 2022), direct empirical evidence from natural environments has been lacking. This lack of evidence is largely due to the technical difficulty of recovering and concentrating exosomes/sEVs from highly diluted environmental water, as no standardized isolation protocols exist for field samples.

Therefore, driven by our biological hypothesis regarding small RNA dynamics and vesicle-mediated protection, in this study, we aimed to provide the first direct evidence that Akoya pearl oysters release exosomes and sEVs into surrounding seawater as highly stable alternative carriers, by combining microscopy-based characterization with small RNA sequencing. By validating the presence of these carriers in the natural environment, we sought to uncover a previously unrecognized mechanism of molecular exchange in aquatic ecosystems.

Materials and Methods

Biological Materials

In September 2023, a tank containing Akoya pearl oysters (summer-harvested seed bivalves, 300 individuals in a 30-L tank with 20 L of filtered seawater) was maintained for 5 days under static water conditions without feeding. As the initial findings were derived from closed-environment tank water, we extended our investigation to open-water conditions conducted concurrently. In September 2024, 2 L of seawater was collected in the morning and divided into two 1-L replicates (hereafter referred to as AM1 and AM2). Similarly, another 2 L was collected in the evening and divided into two 1-L replicates (PM1 and PM2), from the Akoya pearl oyster aquaculture area (34°17’41"N 136°48’09"E) in Ago Bay, Mie Prefecture, Japan (Fig. 1). After collection, the water samples were stored at 4 °C for 1 day and transported to the University of Tokyo, Tokyo, Japan, via refrigerated delivery at 4 °C. Samples were processed immediately upon arrival. This research was approved by the Animal Experiment Ethics Committee of the Graduate School of Agricultural and Life Science, The University of Tokyo (Accession No. P21-103).

Fig. 1.

Sampling sites and collection procedures in the aquaculture area. (A): The raft in Ago Bay, Shima, Mie prefecture, where Akoya pearl oysters are cultured. (B): Raft used for farming Akoya pearl oysters. (C): Water collection procedure. (D): Suspended net cage for Akoya pearl oyster marine farming

Environmental Exosome/sEV Isolation and Identification

As shown in Fig. 2, exosomes/sEVs were recovered by modifying an exosome purification protocol originally designed for culture supernatants (Chen et al. 2022). To isolate vesicles within the exosome/sEV size range (50–200 nm), seawater was first filtered through a 20-µm mesh to remove large particulates. The filter was then passed through a 0.22 μm Stericup-GP filter (Merck Millipore, MA, USA) to eliminate residual smaller particulates. The final filtrate, containing vesicles smaller than 220 nm, was subjected to ultrafiltration.

Fig. 2.

Experimental design and analysis workflow performed in this study

Ultrafiltration was performed on 1 L of sample using Centricon Plus-70 10 kDa Ultracel-PL units (Merck Millipore) at 3,500 × g for over 25 min (Centrifuge; Model 6200; rotor: SF-2504 S; KUBOTA, Tokyo, Japan). After removing the flow-through, the retained fraction was replenished with the remaining sample and repeatedly concentrated. The final concentrate was replaced with phosphate buffered saline (PBS) and stored at −80 °C until further use. For the aquaculture open-water samples, vesicle concentration and particle size distribution were analyzed using interferometric light microscopy (Videodrop), which measures Brownian motion (Sausset et al. 2023). For this analysis, 100 µL samples from the AM and PM collections were sent to Meiwafosis (Tokyo, Japan). Vesicle counts per liter were estimated based on these measurements.

For transmission electron microscopy (TEM), samples from both the rearing tank and aquaculture area were prepared. Exosome/sEV fractions were further purified via ultracentrifugation at 100,000 × g for 70 min at 4 °C, supernatant was removed, and the pellet was resuspended in PBS. Ultracentrifugation was performed using an Optima MAX-TL ultracentrifuge (Beckman Coulter, CA, USA) with a TLA-100.3 rotor (Beckman Coulter). Samples were applied to self-made carbon coated copper grids (200 mesh), allowed to adsorb for less than 5 min, and negatively stained with 2% uranyl acetate (Bio-Rad, CA, USA) for approximately 30 s. The grids were then visualized using a transmission electron microscope (JEM-1400plus; JEOL, Tokyo, Japan) operated at an accelerating voltage of 120 kV.

RNA Extraction, Library Construction, and Small RNA Sequencing

Following the methodology described by Pan et al. (2025), small RNA was extracted from each exosome/sEV fraction obtained from tank water (250 µL) and open water (300 µL) utilizing the exoRNeasy midi kit (Qiagen), in accordance with the manufacturer’s protocol. The quantity and quality of the extracted small RNA were evaluated using the Qubit microRNA kit with the Qubit 2.0 fluorometer (Thermo Fisher Scientific, MA, USA) and the Agilent Small RNA kit with the Agilent 2100 Bioanalyzer (Agilent Technologies, CA, USA). Library construction was carried out using the NEXTFLEX^® Small RNA-Seq Kit v4 (Revvity, MA, USA), following the manufacturer’s instructions. For the evening samples, the cDNA amplification step was modified to increase the number of PCR cycles to 32 in order to obtain sufficient cDNA yields. Final library quality was evaluated using the High-Sensitivity D1000 ScreenTape Kit on the Agilent 2200 TapeStation. High-quality libraries were sequenced using the DNBSEQ-G400 platform (BGI) with 100 bp paired-end reads by BGI JAPAN (Hyogo, Japan).

Sequencing Data Analysis

Following the method described by Meng et al. (2025), we first trimmed adapter sequences and removed low-quality reads. Reads outside the 18–40 nt range were filtered using TrimGalore (https://github.com/FelixKrueger/TrimGalore). The remaining reads were mapped to the P. fucata reference genome (Takeuchi et al. 2022) using Bowtie (Langmead et al. 2009) with zero mismatches (bowtie -f -v 0 -a --al) to analyze expression and distribution across the genome. To improve alignment efficiency, reference haplotypes, versions 4.1 A (reference) and 4.1B (alternative), were used. To classify and annotate small RNAs, mapped reads were compared against Mollusca sequences in the miRBase v22.1 database (Kozomara et al. 2019) allowing for one mismatch, using Bowtie (bowtie -v 1 -a --best --strata). The reads were also annotated against the Rfam 14.10 database (Kalvari et al. 2020) to identify other small RNA types (e.g., rRNAs, tRNAs), using Bowtie with no mismatches (bowtie -v 0 -a --best --strata). GNU Awk v4.0.2 (https://www.gnu.org/software/gawk/) was used to extract the sequences.

Subsequently, a custom Perl script (casify_rna.pl; Supplementary data 1), developed with the assistance of ChatGPT (GPT-5; OpenAI) based on reference (Zhang et al. 2023), was used to classify RNA and output FASTA sequences (perl classify_rna.pl -fa < fasta> -rfam < rfam_file> -mirna < miRNA.bwt> -ncrna < ncRNA.bwt> -outpre < output_prefix>). The authors reviewed and validated the generated code to ensure its accuracy. During annotation, sequences of unexpected lengths were observed among the classified miRNAs. Canonical miRNA sequences (20–24 nt) were extracted using seqkit (Shen et al. 2016); all others were designated as “unknown”. As Rfam lacks sequences specific to P. fucata, previously published sequences (28 S rRNA: AB214477, 18 S rRNA: AB214462, 5.8 S rRNA: AB205102, ITS1: AB214218, ITS2: AB214265, rRNA intergenic spacer: AB214291 and AB214307) were used to recover unannotated rRNAs from the unknown category via local BLAST (BLAST 2.16.0+). Tandem repeats were then removed from the unknown sequences using MISA (Beier et al. 2017) and TRF (Benson 1999). Sequences 25–32 nt in length were extracted and designated as predicted piRNAs, following the criteria in Huang et al. (2022). In model organisms, piRNAs frequently show a 1U bias, with uridine at the first position (Kawaoka et al. 2011; Izumi et al. 2013; Stein et al. 2019). Accordingly, we analyzed the nucleotide composition at the first base of the predicted piRNAs, evaluating both the diversity and frequency of occurrence. Finally, we used piRNAs derived from P. fucata hemolymph exosomes (Huang et al. 2022) as a reference database to identify perfectly matching sequences via BLASTn (option: -task blastn-short).

Additionally, for the Tank and aquaculture site (AM1) sample, which yielded the highest number of sequencing reads, adapter trimming was performed using fastp (Chen et al. 2018). The resulting paired-end reads were mapped to the P. fucata reference genome using BWA-MEM (Li 2013), and alignment files were processed with SAMtools (Li et al. 2009) to generate high-quality mapped reads. To account for potential splicing, alignments were also performed using HISAT2 (Kim et al. 2019), followed by transcript assembly with StringTie (Pertea et al. 2015). A merged GTF file, generated from both genome versions 4.1 A and 4.1B, was used to detect transcripts from both haplotypes. CDS and peptide sequences were predicted from the assembled transcripts using TransDecoder (Haas et al. 2013). Functional annotation was conducted via BLASTx and BLASTp searches against the Swiss-Prot database (downloaded December, 2023), and comprehensive annotations were generated using Trinotate (Bryant et al. 2017), which incorporated protein domain and functional annotation data.

Results

Identification of Exosomes/sEVs from Environmental Seawater

Exosomes/sEV fractions were isolated from the high-molecular-weight component of filtered environmental seawater using ultrafiltration. One sample from the rearing tank and one each from the AM and PM aquaculture samples were examined via microscopy. Vesicle size and concentration were measured using interferometric light microscopy (Videodrop; Myriade, Paris, France). The median diameters of the AM and PM samples were 176 nm and 174 nm, respectively (Fig. 3). The concentrations were approximately 5.9 × 10⁸ and 8.0 × 10⁸ vesicles/mL, respectively. These values were similar to those reported for exosomes derived from Akoya pearl oyster tissue and hemolymph in previous studies by Meng et al. (2025) and Huang et al. (2022), as well as to preliminary data from rearing tank water (data not shown). When extrapolated to 1 L of environmental seawater, the estimated vesicle concentration ranged from approximately 1.0 × 10¹⁰ to 10¹¹ vesicles/L. Transmission electron microscopy (TEM) revealed numerous vesicles with diameters primarily in the 50–200 nm range. In addition, vesicles smaller than 50 nm and filamentous structures were observed that were not detectable via interferometric light microscopy (Fig. 4A and B).

Fig. 3.

Exosome/sEV fraction identification using interferometric light microscopy of Videodrop. Vesicle size distribution of exosomes/sEVs isolated from open-water samples collected in the morning (AM) and evening (PM)

Fig. 4.

Transmission electron microscopy (TEM) images of exosome/sEV fractions. (A) Exosomes/sEVs isolated from tank water. (B) Exosomes/sEVs isolated from morning (AM) open water sample. Spherical vesicles are indicated; vesicles smaller than 50 nm and filamentous structures may correspond to ribosomes and nucleic acids, respectively. Scale bars: 500 nm

Small RNA was extracted from aquaculture water samples collected at the same site in the morning (AM1 and AM2) and in the evening (PM1 and PM2) as well as once from the rearing tank sample. The resulting exosome/sEV preparations were then quantified using the Qubit microRNA assay. Adequate RNA yields were obtained, with concentrations of 9.0 ng/µL for the rearing tank sample and 7.2 ng/µL (AM1), 15.2 ng/µL (AM2), 11.1 ng/µL (PM1), and 20.6 ng/µL (PM2) for the aquaculture samples. Electrophoretic analysis using an Agilent Bioanalyzer showed that the majority of RNA fragments were shorter than 40 nucleotides. In most samples, more than 70% of sequences fell within the characteristic size range for miRNAs.

Small RNA Profiling of Exosomes/sEVs

Duplicate samples were collected from the aquaculture area (AM and PM), and one sample was collected from the rearing tank. The raw sequencing outputs for the rearing tank and aquaculture samples (AM1, AM2, PM1, and PM2) was 40 million, 30 million, 8 million, 13 million, and 9 million reads (M reads), respectively. After adapter trimming and quality filtering, reads were restricted to a length of 18–40 nucleotides (nt). The remaining read counts were: rearing tank, 15 M reads; AM1, 13 M reads; AM2, 4 M reads; PM1, 6 M reads; PM2, 5 M reads. Mapping these reads to the P. fucata reference genome resulted in alignment rates of 4.6% for the tank water sample and 2.8 ± 0.2% for aquaculture area samples. Even though nearly two days elapsed between seawater collection and vesicle isolation, a considerable proportion of reads still mapped to the P. fucata reference genome.

Reads were annotated by mapping to the Rfam and miRbase databases. Sequences of 20–24 nt were classified as canonical miRNAs, while sequences of 25–32 nt that did not map to known small RNA categories were designated as estimated piRNAs. Annotated reads were further categorized into piRNA, miRNA, rRNA, tRNA, snRNA, other ncRNAs, and unknown RNA types. Corresponding read counts for each category are presented in Table 1. Among these, rRNA was the most abundant RNA type across all samples (Fig. 5). In contrast, miRNAs accounted for less than 0.1% of mapped reads. Notably, despite being extracted from environmental seawater, piRNAs were detected in higher-than-expected proportions, 2.5% in the rearing tank sample and 0.1 ± 0.06% in aquaculture samples.

Table 1.

RNA type distribution based on small RNA-seq mapping and annotation

RNA/Reads	Tank	AM1	AM2	PM1	PM2
piRNA	19,248	372	81	246	145
miRNA	27	30	9	22	7
rRNA	373,789	361,592	103,368	171,123	107,696
tRNA	45,384	4,174	1,606	4,321	2,328
snRNA	1,239	1,638	589	717	1,347
other_ncRNAs	2,293	830	173	342	1,088
unknown	338,564	8,702	3,025	5,376	3,306

Fig. 5.

Classifications of mapped small RNA reads from rearing tank and aquaculture water samples. piRNA, piwi-interactive RNA; rRNA, ribosomal RNA; tRNA, transfer RNA; unknown, non-annotated sequence; Other, other small RNAs

The sequence most frequently identified among the estimated piRNAs was observed 251 times in the rearing tank samples and 33 times in aquaculture samples (Supplementary data 2). The 1U bias (Kawaoka et al. 2011; Izumi et al. 2013; Stein et al. 2019), calculated based on the number of estimated piRNA sequence types, was 31% in the rearing tank and 22.4 ± 2.1% in the aquaculture area. The corresponding frequency rates were 19.1% and 15.4 ± 5.8%, respectively (Table 2).

Table 2.

1U bias based on the number of sequence types and observed frequencies

Sample	Types of sequence			Frequencies of sequence
	U	A, G, C	1U bias (%)	U	A, G, C	1U bias (%)
Tank	863	2,067	31.3	3,679	15,569	19.1
AM1	71	230	23.6	74	298	19.9
AM2	16	51	23.9	16	65	19.8
PM1	27	112	19.4	35	211	14.2
PM2	8	27	22.9	11	134	7.6

Sequence-type-level analysis indicated that a 1U bias was present only in the rearing tank sample. Comparison of the estimated piRNAs with exosome-derived piRNAs reported by Huang et al. (2022) using BLASTn, revealed that 116 sequence types from the rearing tank and 6 types from the aquaculture area matched exactly with previously reported sequences. Several piRNAs frequently detected in earlier studies were also found in this dataset. Among these, the most abundant piRNA was observed 198 times in the rearing tank sample and 13 times in the aquaculture water sample (Tables 3 and 4; Supplementary data 3).

Table 3.

Frequently detected piRNA sequences in the rearing tank samples

Top5	Sequence	nt	Number of occurrences in this study	Number of occurrences in previous study*
1	GGAAGGTCCGGAGAGTCTAGGTTACCCAATT	31	198	247
2	GGAAGGTCCGGAGAGTCTAGGTTACCCAATTC	32	171	405
3	TGGGAATACCGGGTGTTGTAGGCAT	25	110	543
4	GGGGTGAGCCGTGGCGTGCTTGGGTTCAAATC	32	79	14
5	ATGGAAGGTCCGGAGAGTCTAGGTTACCCAAT	32	34	336

*: Sequences from the sixth onward are provided in Supplementary data 2. The previous study referenced is Huang et al. (2022)

Table 4.

Occurrence counts of piRNA sequences matching known sequences in aquaculture water samples

Sample	Sequence	nt	Number of occurrences in this study	Number of occurrences in previous study*
AM1	GATAAATGAAGTCAGCTTATTATCTCTGAGT	31	1	1
	GGCCGTGATCGTATAGTGGTTAGTACTTCGCG	32	1	7,823
	GATTGTCAATGGGTTTTGAGAGAAAACC	28	1	11
	GAGGTCTGCTGCCGATGGACTTTACGATTTCT	32	1	143
	TGTCAATGTACTTTGGATGAGCTGCAGG	28	1	1
PM1	TCAATACCCCTAGCTCAATATTCGACCTTCTC	32	2	3,038
PM2	GGCCGTGATCGTATAGTGGTTAGTACTTCGCG	32	13	7,823

*: The previous study referenced is Huang et al. (2022)

mRNA Detection from Small RNA-seq Data

mRNA analysis was performed using paired-end reads from the rearing tank and AM1 aquaculture samples, both of which yielded sufficient sequencing depth. After filtering, reads were aligned to the P. fucata reference genome. The number of paired-end reads mapped was 56 K reads (Tank) and 120 K reads (AM1). Transcript assembly was performed using the Hisat2–StringTie–Trinotate pipeline. This analysis yielded 164 transcripts from the rearing tank (125 of which were annotated) and 49 transcripts from AM1 (10 annotated), including transcripts lacking gene annotations (Supplementary data 4). Among the detected transcripts were lysosome-associated membrane glycoprotein 1 (LAMP1; Mathieu et al. 2021), moesin (MOES; Li et al. 2024), BRO1 domain-containing protein BROX (BROX; Ichioka et al. 2008), and filamin-A (FLNA; Eguchi et al. 2020), all of which have previously been reported in exosomes/EVs. Additionally, several genes associated with key cellular functions were detected, including those involved in signal transduction (Signal transducer and activator of transcription 5 A; STAT5A), reproduction (Spermatogenesis-associated serine-rich protein 2; SPAS2, Zonadhesin; ZAN), and transport (ATP-binding cassette sub-family A member 2; ABCA2, Sodium- and chloride-dependent glycine transporter 2; SLC6A5).

Discussion

Exosomes are small extracellular vesicles secreted by eukaryotic cells that contain nucleic acids, proteins, lipids, and other bioactive molecules (Meldolesi 2018; Pegtel and Gould 2019; Kalluri and LeBleu 2020). They have been extensively characterized in bodily fluids and tissues, where they participate in intercellular communication (Gurunathan et al. 2019; O’Brien et al. 2020). However, nearly all prior studies have focused on intracellular or intercellular functions; their release into and potential interaction with environmental seawater remain largely unexplored. Our findings provide the first direct evidence that aquatic organisms, specifically P. fucata, release exosomes/sEVs into their surrounding environmental seawater. This observation reveals a previously unrecognized biological phenomenon. We propose that these extra-individual exosomes/sEVs be termed “environmental Exosomes/environmental sEVs (eExosomes/esEVs)”.

By adapting an established exosome/sEV isolation protocol originally developed for cell culture supernatants and liquid biopsy samples (Chen et al. 2022), we successfully obtained vesicle fractions from environmental seawater. These fractions exhibited size distributions consistent with those of tissue- and hemolymph-derived exosomes (Huang et al. 2019, 2022; Meng et al. 2025), as determined by interferometric light microscopy (Videodrop). Transmission electron microscopy further confirmed the presence of abundant vesicles sized 50–200 nm, which aligns with the known size range of exosomes (Huang et al. 2022). These results demonstrate that vesicles morphologically and physically consistent with standard exosomes/sEVs are present in environmental seawater and validate the utility of this modified method for environmental exosome/sEV isolation.

Small RNA-seq of these vesicle fractions revealed that despite a delay of nearly 2 days between sampling and isolation (including storage and transport at 4 °C), a considerable proportion of reads could still be mapped to the P. fucata reference genome. While previous studies have shown that naked or free eRNA is highly susceptible to rapid degradation driven by ubiquitous environmental RNases, as well as fluctuations in pH and temperature under such conditions—often declining below the detection threshold within 72 h (Kagzi et al. 2022; Qian et al. 2022; Jo et al. 2023; Wang et al. 2025)—our mapping rate was comparable to or exceeded that reported in conventional eRNA analyses (Veilleux et al. 2021; Miyata et al. 2025). This successful recovery despite the delayed processing strongly supports the hypothesis that aquatic eRNA can exist in a state protected by vesicles (Jo et al. 2022). Specifically, these findings provide compelling evidence that the lipid bilayer of exosomes/sEVs confers robust protection against enzymatic degradation, microbial digestion, and other environmental stressors, thereby preserving the integrity of encapsulated nucleic acids (Koga et al. 2011; Chen et al. 2024). Most of the identified sequences corresponded to rRNA, which may partially reflect the co-isolation of free nucleic acids (Linney et al. 2021) and ribosomes (Maslov et al. 2006) as supported by the presence of filamentous structures and vesicles smaller than 50 nm observed via TEM (Fig. 2). As shown in Tables 2 and 3, we also detected piRNAs with sequences identical to those previously identified in hemolymph-derived exosomes (Huang et al. 2022), providing strong evidence that species-specific small RNAs are released into the environment via exosomes/sEVs.

In the Akoya pearl oyster, piRNAs are not only abundant in reproductive tissues but also enriched in hemolymph-derived exosomes (Huang et al. 2019, 2022). The detection of these piRNAs in environmental seawater strongly supports the analytical capability of our approach to capture species-specific RNA signatures released from the organisms. mRNA analysis using small RNA-seq data identified transcripts for LAMP1 (Mathieu et al. 2021), MOES (Li et al. 2024), BROX (Ichioka et al. 2008), and FLNA (Eguchi et al. 2020), all of which have previously been shown to be associated with exosomes/sEVs. Among these, LAMP1 is considered an exosome-specific surface protein, FLNA is frequently detected in exosomes, and MOES is regarded as a universal pan-exosome marker (Muriel et al. 2016; Hoshino et al. 2020; Mathieu et al. 2021). These findings provide indirect support for the interpretation that the vesicles detected in the environmental seawater samples correspond to exosomes/sEVs. Although their precise association with exosomes/sEVs remains unclear, the presence of transcripts related to reproduction, transport, and signaling may suggest roles in inter-individual communication, such as spawning and horizontal immune transfer. Collectively, our detection of exosome/sEV-encapsulated piRNAs and mRNAs suggests that biological molecules are continuously disseminated beyond individual organisms. While the precise biological functions of these environmental piRNAs and specific mRNAs currently remain uncharacterized and warrant future experimental validation, their stable presence in the water presents significant methodological advantages. Although the specific biological roles of eExosomes/esEVs remain to be determined, our quantification of more than 10 billion vesicles per liter of seawater highlights their potential as a robust source of molecular information. Given that pheromones and other chemical signals exert biological effects at far lower concentrations (Cummins and Bowie 2012), exosomes/sEVs may represent a novel and previously unrecognized mechanism for inter-individual or organismal communication in aquatic ecosystems.

Despite these promising findings, our study has several limitations. First, the exosome/sEV fractions isolated from environmental seawater may include co-isolated particles such as ribosomes and/or free nucleic acids, which could complicate interpretation of the sequencing data. Second, although piRNAs and mRNAs detected in these vesicles suggest potential biological roles, functional validation of their activity in recipient organisms was beyond the scope of this study. Third, our analyses were limited to a single species and sampling period, and additional studies across multiple organisms and environmental contexts will be required to assess the generality of our observations. Addressing these limitations will be essential to fully elucidate the ecological and physiological significance of eExosomes/esEVs.

In conclusion, this study demonstrates for the first time that aquatic organisms release exosomes/sEVs containing species-specific nucleic acids into their surrounding waters. By stabilizing and transmitting functional molecules, eExosomes/esEVs may represent a previously unrecognized mode of inter-organismal communication and contribute to the formation of complex environmental information networks. Notably, unlike eRNA, which rapidly degrades under natural conditions, eExosomes/esEVs provide a stable source of molecular information, suggesting that they may complement existing eDNA and eRNA approaches in ecological monitoring. Future research should aim to elucidate the mechanisms, dynamics, and ecological significance of eExosomes/esEVs across diverse aquatic organisms.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Ryo Yonezawa: Study conception and design, methodology, formal analysis, software, investigation, visualization, writing—original draft, writing—review and editing; Lingxin Meng: Methodology, formal analysis, software, investigation, visualization; Naoki Hashimoto: Methodology, investigation, resources; Ibuki Igarashi: Investigation, visualization; Satoshi Kimura: Investigation, resources; Nina Yasuda: Validation, visualization, supervision; Susumu Mitsuyama: Resources; Takanori Kobayashi: Validation, supervision; Kazutoshi Yoshitake: Software, resources; Shigeharu Kinoshita: Investigation, resources; Nahoko Bailey-Kobayashi: Validation; Kaoru Maeyama: Supervision; Kiyohito Nagai: Resources, visualization, supervision; Shugo Watabe: Validation, supervision; Tetsuhiko Yoshida: Validation, supervision; Shuichi Asakawa: Study conception and design, resources, validation, writing—review and editing, supervision. All authors read and approved the final manuscript.

Funding

Open Access funding provided by The University of Tokyo. This study was supported in part by Grant-in-Aid for Challenging Research (Pioneering) from the Japan Society for the Promotion of Science (JSPS) (25K21737; SA, RY), Core Research for Evolutional Science and Technology (CREST) from Japan Science and Technology Agency (JST) (JPMJCR23J2; NY).

Data Availability

The NGS data have been deposited with links to BioProject accession number PRJDB37606 in the DNA Data Bank of Japan (DDBJ) BioProject database. Other data generated or analyzed during this study are included in this article and supplementary information.

Declarations

Ethical Approval

The authors declare that this manuscript complies with Springer’s Ethical Guidelines for Journal Publication.

Competing interests

The authors declare the following potential competing interests: Financial support for this study was provided by a research donation from Mikimoto Pharmaceutical Co., Ltd. Additionally, funding was provided through a social cooperation program endowed by Toagosei Co., Ltd. Technical and experimental support was provided by K. Mikimoto & Co., Ltd. Naoki Hashimoto, Nahoko Bailey-Kobayashi, Kaoru Maeyama, Kiyohito Nagai and Tetsuhiko Yoshida are employees of these respective companies. However, the authors declare that this study constitutes fundamental basic research. The affiliated companies and funding sources did not compromise the scientific objectivity or integrity of the study design, data analysis, or data interpretation. The findings do not promote any specific commercial products and yield no direct commercial advantage to the affiliated companies. The authors declare no other competing interests.