Six strains (accession number from MW24846 to MW24858) of Bacteroidetes were isolated from the larvae and adults of sponge Tedania sp. and were found to produce OMVs, including Cryomorphaceae bacterium SW-4, Formosa sp. SW-1, Salegentibacter sp. SP-1, Sunxiuqinia dokdonensis SP-8, Muricauda aquimarina SP-4, and Tenacibaculum mesophilum SP-7 (Li et al., 2021). They were cultivated aerobically in marine 2216E broth at 28°C. The OMVs were isolated from 500 ml bacterial culture (OD₆₀₀ = ~1.0) in accordance with a modified procedure. After all bacterial cells were removed via centrifugation at 2,627 × g for 40 min, the OMVs were separated from other extracellular products by ultracentrifugation at 100,000 × g for 1 h at 4°C (41 Ti Rotor, Beckman Coulter, California, USA). The resulting pellets were washed and re-suspended in 10 mM phosphate-buffered saline (PBS; Solarbio, Beijing, China.) (Brown et al., 2015). The amount of OMVs was determined corresponding to the protein content by using the Modified Bradford Protein Assay Kit (Sangon Biotech, Shanghai, China.). The bacteria and OMVs were observed by scanning electron microscopy (SEM) and transmission electron microscopy (TEM), respectively.

Free-living larvae of marine sponge Tedania sp. were prepared by a tuck net with a mesh screen (150 μm) (Li et al., 2021). The effects of six Bacteroidetes strains and their OMVs were assessed on larval settlement and metamorphosis in sterile six-well plates (NEST Biotechnology Co., Ltd., Wuxi, China). The larvae were incubated in the control groups of nature seawater (NSW) and filtered seawater (FSW) with 0.22 μm pore-sized polyvinylidene fluoride membrane filters (Millipore Co., Ltd, Darmstadt, Germany). The NSW group was used to demonstrate the health of the collected sponge larvae and their ability to live in a natural setting. To evaluate the effect of bacteria and OMVs addition on larval settlement, FSW was used as the basic experiment seawater without taking other biological components into account. All bacteria were quantified up to 10³ cfu/ml and added into FSW. According to references reported, the method of quantitative protein was an approved way for quantifying OMVs (Klimentová and Stulík, 2015). OMV proteins were extracted using a Bacterial Protein Extraction Kit (Sangon Biotech) and quantified by using the Bradford Protein Quantification Kit (Vazyme Biotech Co., Ltd, Jiangsu, China). Based on previous experimental results, most bacterial species-OMVs showed the highest settlement rates at a medium concentration (~100 ng/ml protein) (Li et al., 2021). Therefore, for each OMV experiment, a protein concentration of approximately 100 ng/ml was chosen, and the corresponding OMVs were incubated in 8 ml FSW. One parallel group contained ≥30 sponge larvae and three parallel plates were set for each trial. After 8–90 h, the amount of settled larvae was counted to evaluate the ability of larval settlement. The effects of bacteria and vesicles on settlement rates were computed independently by SPSS (a one-way ANOVA analysis), and descriptive analyses, homogeneous subset tests, and tests of homogeneity of variances were carried out. Afterward, the OMVs and sponge cell suspensions after incubation were observed by SEM.

The salt was initially removed from OMVs samples using a Millipore volume with a 100-kDa filter membrane. Four volumes of cold methanol were added into the samples, broken for 12 min by ultrasonication using a noise isolating chamber (Scientz Biotechnology Co., Ltd, Ningbo, China), and kept for 1 h at −20°C for albumen precipitation. The samples were subsequently centrifuged at 12,000 rpm for 15 min at 4°C. Using a Termovap Sample Concentrator (N-EVAP), the supernatant was reduced to the smallest possible volume before being dried and re-dissolved in a minimal amount of water. The flow rate for LC-MS was 3 μl/min, and a gradient of acetonitrile (ACN) was used to elute the analytes. The analytes were initially eluted at 1% ACN for 3 min, followed by 60% for 10 min, 95% for 12 min, and then back to the initial 1% ACN for 20 s. Re-equilibration took place for 20 min. Data were manually annotated using Xcalibur Qual Browser software (Thermo Fisher Scientific).

RNA was extracted from a variety of materials using the miRNeasy Mini Kit. The samples used in the NOS gene quantification experiment were an embryo, competent larva, post-larva, and “OMVs+post-larva”. The embryo and competent larvae were selected for collection from NSW due to the particularity of the embryo in the adult sponge and its special time node when floating larvae were just being released from the body. Following a roughly 24 h co-incubation in FSW, the post-larva and “OMVs+post-larva” were collected for qRT-PCR analysis, which were consistent with transcriptome. Among them, the competent larvae act as the blank control. RNA was reverse-transcribed into cDNA by the PrimeScript^TM RT reagent kit with a gDNA Eraser (Takara Co., Ltd.). qPCR was performed using a QuantStudio 6 Flex instrument (Life Technologies) via TB Green^®Premix Ex Taq™II (Tli RNaseH Plus) (Takara Co., Ltd.). Specific primers were designed for the NOS and succinate dehydrogenase complex, subunit A (SDHA) genes in line with the transcriptome of Tedania sp. larvae. The quantity of NOS transcripts throughout normal development was evaluated using qPCR with the following oligonucleotide primer set: forward 5′- AATTTGCTCAAGCCCATTGC-3′ and reverse 5′-GAGGGACAATCCACACCCAG-3′. The transcription level was normalized using the reference gene SDHA (Gardères et al., 2014; Song et al., 2021) to obtain the relative gene expression values for Tedania sp. The oligonucleotide primer set for SDHA was as follows: forward 5′- CTCTTTCCCACCCGATCTCAC-3′ and reverse 5′- GGGCATTCCGTAATTCTCAAGC-3′. The relative level of the NOS gene expression was calculated using the 2 ^−ΔΔCt method (Livak and Schmittgen, 2001), and three replicates were performed for each treatment.

All data were expressed as the means ± standard deviations from at least three sets of independent experiments. Both GraphPad Prism 9.0 software (GraphPad Software, Inc., San Diego, California, USA) and Microsoft Excel (Microsoft Corporation, Seattle, Washington, USA) were used to analyze all data. Using a one-way ANOVA and t-tests, significant variations in the sample characteristics were assessed. The results were displayed using online tools (HiPlot, Figdraw, and Vecteezy), as shown in the Figures 2, 5.

To determine the effect of bacteria and bacteria-OMVs on sponge larval settlement, six Bacteroidetes species and their OMVs were examined. Without the addition of any other Bacteroidetes species and their OMVs, it was found that the sponge larval settling rates in NSW and FSW were approximately 68% and 56%, respectively, after approximately 24 h. All bacterial additions increased the larval settling rates in comparison to the FSW and NSW control groups. When strains SP-8, SP-4, and SP-7 were added to the FSW group, they exhibited considerable activities to promote larval settlement, even more than the NSW group (Figure 1A, Supplementary Tables S1, S2). Among them, OMVs from the strain SP-7 (83.33% settlement rates) showed a nearly equivalent effect to bacteria. OMVs produced by the strain SP-7 had an integral membrane structure and were measured at 50–200 nm (Figures 1B, C). When comparing the surface of the larval cells under SEM observation without OMVs addition (Figure 1D), more vesicular OMVs were enriched in the surface of sponge cells throughout the process of promoting sponge larval settlement (Figure 1E).

Based on the fact that SP-7-OMVs significantly promoted sponge larval settlement, it was implied that OMVs could communicate with larval cells and the packaged signals could motivate larval development. To gain effective information, the non-targeted metabolome of T. mesophilum SP-7-OMVs was analyzed. The map of the total ion flow over time of six samples in negative and positive ion modes, which was used to determine their total ion count (TIC), was shown in Supplementary Figure S1, Supplementary Table S3. All known metabolites in SP-7-OMVs were annotated with the KEGG pathway, and 22 classes were traced to the KEGG B class (Figure 2A, Supplementary Tables S4, S5). Most of the metabolites were annotated to the kind of metabolism of the cluster KEGG A class. Among them, the categories belonging to global and overview maps, amino acid metabolism, and carbohydrate metabolism were the top three with the most amount of metabolites.

Superclass analysis was performed for 185 metabolites from MS2 that could be annotated into the KEGG pathway (Figure 2B, Supplementary Tables S6, S7). These organic compounds identified as organic acids and derivatives, organic nitrogen compounds, organic oxygen compounds, and organoheterocyclic compounds were predominate metabolites. Among them, the most abundant categories of metabolites were “organic acids and derivatives” (61 metabolites) and “organoheterocyclic compounds” (39 metabolites), many of which have been identified as potential effectors connected to animal development (Figures 2B, C). Arginine was reported to be involved in the growth and development of the marine sponges A. queenslandica (Song et al., 2021) and Geodia barretti (Hedner et al., 2008). Gamma-aminobutyric acid and L-alanine were shown to promote the development of the blue mussel Mytilus edulis L. (Dobretsov and Qian, 2003; Rivera-Ingraham et al., 2011). D-glutamic acid, glycine, D-glutamine, L-phenylalanine, L-valine, L-methionine, and creatine were confirmed to be connected with cypris larval growth and the development of the barnacle Balanus amphitrite (Tegtmeyer and Rittschof, 1988; Inoue et al., 2018; Labriere et al., 2021). Betaine was verified to induce the life cycle of the coral Acropora cytherea (Jorissen et al., 2021). L-histidine, as a signal molecule, was related to Strongylocentrotus purpuratus (Sutherby et al., 2012) and Holopneustes purpurascens (Swanson et al., 2004) (Figure 2C).

The multiomics analysis, which included the bacterial genome and the transcriptome of the larval host, validated the true bacterial origin of these metabolites and their possible impact on the larval host. The strain T. mesophilum genome, which had 3,360,630 bp in total length and 31.76% in GC content, was sequenced to confirm the bacterial origin of the metabolites found in OMVs (Figure 3A). A total of 7834 DEGs (Supplementary Table S8) were enriched into 347 KEGG pathways (Supplementary Table S9) from the Tedania sp. larval transcriptome. In total, 100 metabolites from “organic acids and derivatives” and “organoheterocyclic compounds” were enriched into 138 KEGG pathways (Supplementary Table S10) from the SP-7-OMVs metabolome. A total of 715 genes from the T. mesophilum genome were enriched into 127 KEGG pathways (Supplementary Table S11). These findings showed that the connectivity of 47 KEGG pathways (Supplementary Table S12) was shared in bacteria, OMVs, and sponge larvae (Figure 3B). Six metabolites were selected based on the above analysis, and a probable correlation with the larval settlement was previously described, including arginine, hypoxanthine, adenine, glycine, glutamic acid, and gamma-aminobutyric acid (Figure 3C, Supplementary Figures S2–S6).

Depending on the optimal concentration, arginine, glycine, glutamic acid, and gamma-aminobutyric acid could increase the sponge larval settlement rate by 35.84%, 30.83%, 28.33%, 25.83%, respectively (Figure 3D, Supplementary Figure S7, Supplementary Table S13). Among them, arginine exhibited the best performance and was targeted as one of the important metabolites for further verification of the interaction between bacterial OMVs and larvae. Afterward, the argininosuccinate lyases (argH) and nitric oxide synthases (NOS) that associated arginine as a substrate and that were involved in larval development were retrieved from the NCBI database. It was found that the majority of species contained both the NOS genes and argH genes, while sponges lacked the latter (Figure 3E, Supplementary Table S14). In other words, although sponges were unable to biosynthesize arginine, the NOS gene, which used arginine as a substrate, was fully expressed. Therefore, bacterial OMVs were reasonably speculated to transmit exogenous signaling molecules to sponge larvae, such as arginine.

The procedure was arranged for easier understanding, as shown in the illustration below (Figure 4). During this study, it made sense to use the three biologically relevant groups, including bacteria (T. mesophilum), bacterial OMVs, and sponge larvae (Tedania sp. larvae). The KEGG pathway analysis was performed by the T. mesophilum genome, SP-7-OMVs metabolome, and Tedania sp. larval transcriptome, respectively. Finally, a simple Venn diagram was used to locate the shared KEGG pathways.

As a representative of many active substances, arginine was selected to testify to the function of bacterial OMVs by arginine quantification in OMVs, arginine's assay in larval settlement, and the related gene expression of larvae. First, the key enzyme argininosuccinate lyase (ArgH) located in the arginine biosynthesis pathway (ko00220) was detected with the gene size of 1,272 bp (Figure 5A) from the bacterial genome. Second, the arginine content was further measured by LC-MS according to the reference spectrum (m/z = 175.12) (Figure 5B, Supplementary Table S15). Compared with the standard, arginine from purified OMVs samples reached a high base peak intensity (RT = 1.45 min). The concentration of arginine in vesicles was estimated to be approximately 39.49 μg/L by HPLC, which was greater than the concentrations in seawater (approximately 5.58 μg/L) and seawater OMVs (approximately 0.07 μg/L). It implied that the presence of arginine in vesicles influenced sponge larval settlement more than NSW and FSW at the laboratory level. Finally, by qPCR test, the gene NOS expression increased with the transition of developmental stages in the sponge larval life cycle, from embryo to competent larva (swimming larvae) to post-larva (settled larvae) (Figure 5C, Supplementary Table S16). The post-larva incubated with SP-7-OMVs showed the highest relative expression of the NOS gene (p = 0.034).

Marine sponges arose more than 600 million years ago and could become promising models of animal-microbe symbioses (Li et al., 1998; Hentschel et al., 2012). Symbiotic bacteria coevolved with the sponge host, developing various strategies to grant hosts advantages or the “priority effect” in exchange for “refuge” from the sponge host (Björk et al., 2019). These bacteria aggregations could affect the larval growth process, even if it was unclear whether the proliferation of Bacteroidetes species in the late-larvae stage was the result of vertical transmission from parents' seed or horizontal transmission from surrounding seawater (Li et al., 2021). The OMVs from sponge-associated Bacteroidetes were favorably involved in larval settlement and metamorphosis according to our repeated tests, including the effect of different dose gradients of OMVs, the EM observation of larvae with or without OMVs addition, and the internalization of OMVs by larva using a confocal fluorescence microscope (Li et al., 2021). Similar roles of bacterial OMVs in larval development appeared in other invertebrate animals. Freckelton et al. (2017) found that large numbers of bacterial extracellular vesicles triggered the settlement of the larvae of Hydroides elegans. Guo et al. (2021) demonstrated that the enriched OMV fractions derived from Pseudoalteromonas sp. P1-9 induced high transformation rates of Hydractinia echinata larvae. The development of Euprymna scolopes could be stimulated by Vibrio cholera OMVs that contain a homolog of OmpU (Lynch et al., 2019). Prokaryotic organisms and animals have coevolved in the symbiotic connection for millions of years, and this interaction has been engineered in a variety of beneficial forms, such as the release of bacterial OMVs, to establish partners with the host (Hill and Round, 2021).

Despite the fact that free-arginine considerably increased sponge larval settlement in this study was consistent with earlier references (Song et al., 2021), the concentration was far greater than it was in seawater and bacteria OMVs. The availability of free-arginine in the dissolved form and the barrier effect of larval cells may limit the utilization of the bioactive molecule. OMVs are ubiquitous, lipid-bilayer-coated nanoscale membrane vesicles that enable active intercellular communications (Margolis and Sadovsky, 2019). Numerous OMVs from marine bacteria protect and condense biomaterials to create a stable transport channel for host-microbe cell interactions in a complicated marine environment with unlimited dilution and mobility (Zhao et al., 2022). These modularized OMVs from diverse bacteria with individualized components were delivered to eukaryotes and prokaryotes (Margolis and Sadovsky, 2019; Ñahui Palomino et al., 2021). For marine benthic species, ecological niche development and population viability depended on the successful larval settlement (Freckelton et al., 2017; Hamel et al., 2019). Therefore, bacterial OMVs could play an essential role in the flow of information and energy in marine ecosystems (Schatz and Vardi, 2018).

Once the biological role of OMVs in promoting larval settlement has been established, the identification of biomolecules becomes crucial for understanding the developmental and symbiotic mechanism of the host. A wide variety of metabolites, including steroid hormones, lipids, metabolic intermediates of nutrient anabolism, and monomers of major biopolymer classes, were present in SP-7-OMVs (< 1 kDa). These molecules may serve as a common language for host cells to learn (Ñahui Palomino et al., 2021). Several studies reported that the metabolites adenosine, acetylcholine, serotonin, palmitoleic acid, and others could induce larval settlement and metamorphosis of a variety of aquatic invertebrates, such as mussels, barnacle cyprids, and sea urchins (Kitamura, 1993; Faimali et al., 2003; Glebov et al., 2014; He et al., 2021). Numerous chemical cues carried by bacterial OMVs suggested their possible functions in interactions between bacteria and their animal host (Biller et al., 2014; Rudnicka et al., 2022). Furthermore, it was also verified that SP-7-OMVs, which carried signal molecules such as glycine, glutamic acid, and gamma-aminobutyric acid, could overcome cell barriers to aid in the growth and development of the host. The characteristic of bacterial OMVs effectors could give bacteria the inherent capacity to become an important component of sponge holobiont.

When bacterial genomics, the OMVs metabolome and the host transcriptome were combined with the shared KEGG pathway analysis to filter the impact molecules, a clear, complete information chain could be found present in the multiomics profile. There must be numerous bioactive molecules involved in the interaction between SP-7-OMVs and larvae that could be connected to larval settlement. Of them, arginine was selected since it clearly increased larval settlement rates. The arginine showed superior activity in enhancing sponge larval settlement, which was consistent with the impact of the exogenous NO donor (sodium nitroprusside and S-Nitroso-N-acetylpenicillamine) in A. queenslandica (Ueda et al., 2016). One crucial method for regulating larval settlement and metamorphosis in lower invertebrates is the NO pathway, which uses arginine as a substrate (Song et al., 2021). It has also been noted that the NO signal-induced larval settlement and metamorphosis occur via the MAPK/ERK signaling pathway (Ueda et al., 2016) in diverse marine invertebrates (Zhang et al., 2015; Yang et al., 2018; Song et al., 2021; Locascio et al., 2022). The NO signaling pathway involves the synthesis and utilization of arginine and includes the argH gene and the NOS gene. The argH genes, which are expressed in sponge-associated bacteria, are not encoded in the published genomic and transcriptomic data of the sponges (Srivastava et al., 2010). It was noteworthy that most invertebrates could synthesize arginine, whereas marine sponges were found to be defective in critical enzymes involved in arginine synthesis (Payne and Loomis, 2006). However, the associated bacteria with the sponge developed a “compensation mechanism” by which they could supply arginine to the host as a substrate for the NOS pathway (Song et al., 2021). Bacterial-OMVs could deliver arginine to cells and aid the growth and development of animal hosts (Figure 5D). Amino acid auxotrophy may be an evolutionarily optimal option for lowering biosynthetic burden (Mee et al., 2014), while related bacteria and their OMVs appear to provide metabolites as an ancient and alternative approach.