Bacterial Cytochrome C Orchestrates Larval Metamorphosis Through Bioelectrical Control of Extracellular Vesicles Biogenesis

ABSTRACT

Beyond the well‐established roles in biomedical processes, extracellular vesicles (EVs) from Gram‐negative bacteria are now identified as potent inducers of larval metamorphosis in marine invertebrates, a critical process for sustaining the health and function of global marine ecosystems. However, the specific inductive molecules within EVs remain largely unknown. Our findings reveal that cytochrome c (CytC), a previously unrecognized chemical signal packaged within EVs, induces larval metamorphosis. CytC maturation is essential for this process, as it enhances the efficiency of electron transfer, thereby supplying the energy required for EV biogenesis and loading with CytC cargo. Molecular docking analysis suggests that bacterial CytC induces the metamorphosis of mussels by binding to Toll‐like receptor 4 (TLR4). Notably, a marked rise in CytC gene abundance was observed in the associated bacterial community during larval metamorphosis, which may indicate a programmed enhancement of electrogenic activity within the biofilm microenvironment to support larval developmental transition. Overall, a new paradigm for interkingdom signaling is proposed in this study, wherein CytC serves as a bioelectrical driver of EV biogenesis and a potentially broadly acting inductive signal within EVs for larval metamorphosis in invertebrates.

The ccmH‐encoded CytC serves as a bioelectrical driver, facilitating EV production and orchestrating biofilm‐mediated larval metamorphosis.

1. Introduction

Invertebrates comprise 1.25 million documented species, accounting for almost 95% of known species on earth (Eisenhauer and Hines 2021). A significant feature of the early life history of most marine invertebrates is the existence of a free‐living, dispersive larval stage which exhibits morphological differences from the adult (Pechenik 1999; Wray and Raff 1991). Larvae must subsequently identify appropriate settlement sites and undergo metamorphosis to develop into adults and establish populations (Yang et al. 2013). Abundant literature demonstrates that larvae of numerous species are induced by specific chemical cues to settle and undergo metamorphosis (Hadfield and Paul 2001). Over recent decades, bacterial biofilms have been recognized as effective cues for inducing larval settlement and metamorphosis across all major invertebrate phyla, including sponges, cnidarians, bryozoans, molluscs, annelids, echinoderms, and crustaceans (Hadfield and Paul 2001). Biofilms consist of bacteria and their self‐produced extracellular polymeric substances (EPS), which act as a vehicle for diverse signaling molecules, including the frequently observed extracellular vesicles (EVs) (Flemming and Wingender 2010; Saad et al. 2024). The broad phylogenetic distribution of bacteria‐induced larval settlement and metamorphosis points to the involvement of a conserved bacterial signal molecule. However, the identity of this universal signal remains a crucial gap in our understanding.

Recently, the recognition of bacterial EVs as a conserved signal is growing, with evidence from their efficacy in inducing larval settlement and metamorphosis in both the serpulid polychaete Hydroides elegans and the mussel Mytilus coruscus (Freckelton et al. 2022; Hu et al. 2024a). By transporting diverse bioactive molecules (proteins, nucleic acids, lipids, and metabolites), bacterial EVs act as a common language in bacteria‐host communication (Schwechheimer and Kuehn 2015). The composition of this complex molecular cargo is modulated by the environment. The functional conservation of EVs in inducing larval metamorphosis across different species suggests the presence of conserved signal molecules within them. As a ubiquitous component of Gram‐negative bacterial EVs, lipopolysaccharide (LPS) is the only constituent identified so far as a key inducer of larval metamorphosis in EVs (Freckelton et al. 2022; Hu et al. 2024a). A key unanswered question is whether other conserved molecules, beyond LPS, are responsible for regulating larval settlement and metamorphosis.

Our previous study proved that an electrochemically active bacterium (EAB), Shewanella has a remarkable ability to facilitate larval settlement and metamorphosis in mussels (Hu et al. 2024b; Yang et al. 2013). Interestingly, studies have revealed that CytC, a ubiquitous bacterial protein essential for electron transfer and ATP synthesis, is abundantly localized on the surface and within the lumen of Shewanella EVs (Bergman 1999; Zhang et al. 2024). CytC is a highly conserved globular protein widely present across diverse organisms, from bacteria and plants to animals and humans (Zhou et al. 2024). Under physiological conditions, it primarily localizes within mitochondria, where it functions as an electron carrier in the respiratory chain (Zhou et al. 2024). Beyond this classical role, the translocation of CytC into the cytosol, nucleus, and extracellular space highlights its multifaceted involvement in critical processes including cell death, cellular signaling, and immune regulation (Zhou et al. 2024). CytC also could act as a signaling molecule. Released CytC from damaged cells caused immune activation of microglia via toll‐like receptor (TLR) 4, a highly conserved pattern recognition receptor in the innate immune system that is widely distributed across the animal kingdom (Wenzel et al. 2019). However, the specific mechanisms underlying bacterial production of CytC‐containing EVs, along with the potential role of vesicular CytC as a metamorphosis‐inducing signal, remain to be elucidated.

In natural environments, biofilms consist of complex bacterial communities (Flemming and Wingender 2010), and it is well established that their capacity to induce larval settlement and metamorphosis varies with community composition (Wieczorek and Todd 1997). Different bacterial species produce distinct sets of signaling molecules; nevertheless, certain conserved signals, such as CytC, may be widely utilized across diverse bacteria in natural biofilm to induce larval metamorphosis. Thus, clarifying the role of Shewanella and CytC in natural biofilm‐induced larval metamorphosis is crucial for understanding host‐microbe interactions in marine ecosystems and further highlights the ecological significance of this work.

In the present study, we first demonstrated that CytC is a novel potential universal chemical cue in EVs that induces larval metamorphosis. Genetic mutation of the CytC synthesis gene ccmH revealed its dual role in regulating the biogenesis of outer membrane vesicles and in the precise packaging of CytC into EVs. Mechanistically, the ccmH mutant disrupted extracellular electron transfer (EET) and consequently reduced extracellular ATP levels, which uncovered a key energy metabolic pathway for the regulation of larval development by CytC‐containing vesicles. Molecular docking analysis suggests that bacterial CytC may induce the metamorphosis of mussels by interacting with larval Toll‐like receptor 4 (TLR4). During larval metamorphosis, we discovered a spatial redistribution of Shewanella from the periphery to the core of the natural biofilm community, coincident with an enrichment of CytC‐related gene abundance. This coordinated shift suggested a programmed enhancement of electrogenic activity within the biofilm micro‐environment to support the developmental transition. Together, we proposed that the electron carrier CytC is a bioelectrical driver for the biogenesis of EVs, potentially serving as a universal cue within EVs for invertebrate developmental transition.

2. Materials and Methods

2.1. ∆ccmH and ∆ccmH::ccmH Construction

The gene ccmH was knocked out by homologous recombination. The bacterial strains, plasmids and primers used for gene knockout were listed in Table S1. In accordance with the protocol established by Zeng et al. (2015), the upstream and downstream of the ccmH gene were PCR‐amplified, incorporating restriction enzyme sites. The amplified fragments were then cleaved using the appropriate enzymes and inserted into the pK18mobsacB‐Ery vector. The recombinant plasmid was introduced into Escherichia coli WM3064 cells. The E. coli WM3064 containing the pK18mobsacB‐Ery plasmid was mixed with wild‐type strains and co‐cultured on mating agar for conjugation transfer. Following culture at 25°C for 12–48 h, the mixture was spread onto Zobell 2216E agar medium supplemented with erythromycin (25 µg/mL). Positive colonies were screened and confirmed via PCR using LF/SR and SF/LR primer pairs. Verified transformants were then plated onto 20% sucrose‐containing Zobell 2216E medium and incubated for 2 d. A single colony was picked up and the knockout strains were verified by PCR using SF/SR, LF/SR, SF/LR, and LF/LR primers. The positions of the PCR primers for generating the mutation via recombination and for confirming the mutation are provided in Figure S1A.

To restore ccmH function, genetic complementation was carried out using the pBBR1MCS‐Cm broad‐host vector. The full‐length ccmH coding sequence was PCR‐amplified using primers pBBR‐F/pBBR‐R. The purified PCR product was digested and inserted into the corresponding sites of pBBR1MCS‐Cm. Following sequence confirmation, the recombinant plasmid was introduced into E. coli WM3064 donor cells via chemical transformation.

The E. coli WM3064 containing the pBBR1MCS‐Cm plasmid was mixed with ∆ccmH and co‐cultured on mating agar for conjugation transfer. Positive clones bacteria were selected by plates containing chloramphenicol (30 µg/mL) and verified by PCR. The validated strains were cultured on the 20% sucrose‐containing Zobell 2216E medium for 2 d. The ∆ccmH::ccmH was verified by PCR using SF/SR, LF/SR, SF/LR, and LF/LR primers.

2.2. EVs Isolation

To evaluate EV production capacity, EVs were isolated in parallel from two sources: bacterial cell lysates and culture supernatants. Bacterial cultures of wild‐type, ΔccmH, and ∆ccmH::ccmH strains were grown in Zobell 2216E medium at 25°C with shaking (200 rpm) for 18 h. The cultures were centrifuged at 8000 × g for 10 min at 4°C to separate cells from supernatant. For supernatant‑based isolation, the culture supernatant was sequentially filtered through 0.45‑µm and 0.22‑µm pore‑size filters to remove residual bacteria and large debris. The filtrate was ultracentrifuged at 130 000 × g for 1 h at 4°C to pellet crude EVs. The pellet was washed twice by resuspension in 0.05 M Tris‐HCl (pH 8.0) and ultracentrifugation at 130 000 × g for 1 h. For lysis‑based isolation, the cell pellet was resuspended in 30 mL of 0.05 M Tris‑HCl (pH 8.0), heated at 56°C for 45 min, and then incubated on ice for 30 min with lysozyme (final concentration 1 mg/mL). The cells were ultrasonically disrupted, and the lysate was centrifuged at 5000 × g for 30 min to remove debris. The resulting supernatant was ultracentrifuged at 130 000 × g for 75 min to pellet crude EVs. The pellet was washed twice by resuspension in 0.05 M Tris‑‐HCl (pH 8.0) and ultracentrifugation at 130 000 × g for 60 min. To further purify EVs from the crude preparations, we used density gradient centrifugation at 4°C, 130 000 × g for 12 h. The density gradient was prepared using different concentrations of OptiPrep (Serumwerk Bernburg AG, Germany) dissolved in 50 mM HEPES‐150 mM NaCl. A density gradient was formed by sequentially layering 5 mL each of 50%, 45%, 40%, 35%, 30%, and 25% OptiPrep solutions from the bottom to the top of an ultracentrifuge tube. The 5 mL EVs sample was carefully loaded onto the top of the gradient. Following ultracentrifugation at 130 000 × g for 12 h at 4°C, fractions were collected from top to bottom and analyzed by SDS‐PAGE with silver staining to identify vesicle‐enriched fractions. The fraction from the 50% OptiPrep layer was collected (Figure S2). The collected fraction was diluted with 50 mM HEPES‐150 mM NaCl and centrifuged at 130 000 × g for 1 h at 4°C to pellet the purified EVs. After repeating this washing step twice, the supernatant was discarded, and the pellet was resuspended in sterile 1× PBS. To verify sterility, 5 µL of the purified EV suspension was incubated overnight in Zobell 2216E medium to ensure no colony growth, after which the EVs were stored at −80°C for subsequent experiments.

2.3. Immunomagnetic Bead‐based Validation of Purified EVs

To further validate that the particles obtained by the lysis‐based method were genuine vesicular structures rather than intracellular debris, the purified lysate‐derived EVs were subjected to additional immunomagnetic bead‐based enrichment using the Magnetic Exosomes Enrichment Kit (TIANGEN, Cat. No. ES201, product link: https://www.tiangen.com/content/details_40_28427.html; China) according to the manufacturer's instructions with minor modifications. This kit employs proprietary TiO₂‐coated magnetic beads that specifically bind to phosphate groups on the EV surface, enabling enrichment independent of species‐ or strain‐specific antibodies. Briefly, the purified EVs were first diluted 200‐fold with sterile 1× PBS, and then filtered through a 0.22‐µm filter to remove any residual aggregates. Two hundred microliters of MagAttract Suspension XBE (magnetic beads) were added to the filtered EV sample, and the mixture was incubated at room temperature for 5–10 min with rotation at 10 rpm to allow binding. The tube was then placed on a magnetic stand for 30 s to 1 min until the solution became clear, and the supernatant was completely removed. The magnetic bead–EV complexes were washed three times with 1 mL of Wash Buffer XW, with thorough resuspension and magnetic separation each time. Finally, the EVs were eluted by adding 200 µL of Elution Buffer XE, followed by incubation at room temperature for 5 min with rotation at 10 rpm. After a brief centrifugation and magnetic separation, the eluate containing the enriched EVs was collected. Protein concentrations were measured before (Pre‐enrichment) and after (Post‐enrichment) bead capture to assess the capture efficiency.

2.4. Monospecific Biofilm Development

The single colony of wild‐type, ΔccmH, ∆ccmH::ccmH of S. marisflavi was picked and cultured in Zobell 2216E medium for 16–18 h at 25°C with shaking at 200 rpm. The bacterial culture was centrifuged at 1300 × g for 15 min to separate cells from the medium. Following centrifugation, the liquid supernatant was discarded, leaving the cell pellet. This pellet was then thoroughly rinsed three times using 50 mL of autoclaved and filtered seawater (AFSW) to remove residual medium components. Seawater was prepared from natural seawater collected from the breeding farm. For filtered seawater (FSW), the seawater was filtered through 0.45 µm filters. For AFSW, the seawater was sequentially filtered through 0.45 µm filters and then autoclaved at 121°C for 20 min. No additional chemical supplements were added to either seawater preparation. The bacteria were resuspended into 50 mL of AFSW. A 100 µL of bacterial suspension was stained by 0.1% acridine orange for 5 min, filtered by 0.22 µm filter membrane and the bacterial concentration was calculated under fluorescence microscope at 1000 × magnification. The bacterial suspension of wild‐type, ∆ccmH, ∆ccmH::ccmH was adjust to 5 × 10⁸ cells/mL according to the concentration. The biofilm formed in a Petri dish (64 mm × 19 mm) contained glass slides with ITO film (35 mm × 25 mm) and bacterial suspension at 18°C for 48 h.

To determine whether CytC and EVs could restore the biofilm‐inducing activity of ∆ccmH, biofilms were formed by co‐culturing ∆ccmH bacteria with CytC and EVs. Standard CytC (Sigma, USA) was added to the Petri dish containing ∆ccmH bacterial suspension at final concentrations of 0.1, 0.5, 1.0, 5.0, 10.0, and 20.0 mg/L to form ∆ccmH + CytC biofilms. For each tested concentration of CytC, six independent biological replicates were prepared. Isolated EVs were added to the Petri dish containing ∆ccmH bacterial suspension at final concentrations of 1.0, 5.0, 10.0, 50.0, and 100.0 mg/L to form ∆ccmH + EV biofilms. For each tested concentration of EVs, nine independent biological replicates were prepared. The compositions of all bacterial media and supplement preparation methods are detailed in Supplementary Table S2.

2.5. Larval Culture of M. coruscus and Crassostrea angulata

After collection from a breeding farm, the pediveliger larvae were separated from seawater and transported to the laboratory in a dry, ice‐cooled environment. For experiments conducted during periods when M. coruscus larvae were seasonally unavailable, C. angulata larvae were used as a complementary model. Larvae of both species were cultured in 60 L plastic buckets (upper diameter 440 mm, lower diameter 370 mm, height 520 mm) containing 50 L of FSW, at a density of 5 individuals per milliliter, and maintained in darkness. They received daily feedings of Isochrysis zhanjiangensis at a rate of 1 × 10⁴ cells per individual. I. zhanjiangensis was cultured in‐house. Briefly, stock cultures were first expanded in 300 mL of AFSW supplemented with 0.1% nutrient salts in conical flasks for 7 days. The cultures were then transferred to 500 mL of AFSW with 0.1% nutrient salts and incubated for another 7 days, followed by final scale‐up in 5 L of AFSW with 0.1% nutrient salts and cultured for an additional 4 days prior to use. The detailed composition of the nutrient salts is provided in Supplementary Table S3. The culture water of larvae was renewed with FSW every 48 h. The culture temperature was maintained at 18°C for M. coruscus and at 25°C for C. angulata. Both species were acclimated for a minimum of 2 days prior to bioassays to restore their planktonic activity. Following this period, they were utilized in settlement and metamorphosis assays, as detailed in the subsequent sections.

2.6. Larval Settlement and Metamorphosis Bioassays in M. coruscus

The established protocol was followed to conduct the larval settlement and metamorphosis bioassays in mussels (Wang et al. 2012). To test the inducing activity of biofilms, each Petri dish contained biofilm, ca. 20 pediveliger larvae, and 20 mL of AFSW containing either biofilm alone, biofilm supplemented with CytC, or biofilm supplemented with EVs. Separately, to test the inducing activity of purified EVs, each well of a 6‐well plate contained EVs (10 mg/L), 10 pediveliger larvae, and 10 mL of AFSW. An Olympus stereomicroscope (100×; Tokyo, Japan) was used to observe larval behavior and morphology. The settlement and metamorphosis sequence, completed within 96 h, was characterized by progressive behavioral modifications (swimming, substrate contact, crawling, permanent settlement) and concurrent anatomical remodeling (velum disappearance, shell growth, gill emergence, juvenile form attainment). The percentages of larvae displaying distinct behaviors and physical structures were meticulously recorded. Negative controls consisted of Petri dishes containing a cleaned and sterilized glass slide, ca. 20 larvae, and 20 mL of AFSW. As positive controls, ca. 20 larvae were placed in Petri dishes with 20 mL of 10⁻⁴ M epinephrine (EPI) solution (Yang et al. 2011). The clean glass without biofilms was used as a blank. Petri dishes and 6‐well plates were incubated at 18°C without light.

2.7. Larval Settlement and Metamorphosis Bioassays in C. angulata

To test the inducing activity of EVs, each well of a 6‐well plate contained EVs (10 mg/L), 10 larvae, and 10 mL of AFSW. To test the inducing activity of biofilms, each well of a 6‐well plate contained biofilm, 10 larvae, and 10 mL of AFSW. Negative controls consisted of each well of a 6‐well plate containing a cleaned and sterilized glass slide, 10 C. angulata pediveliger larvae, and 10 mL of AFSW. As positive controls, 10 C. angulata pediveliger larvae were placed in each well of a 6‐well plate with 10 mL of 10⁻⁴ M epinephrine solution. An Olympus stereomicroscope (100×; Tokyo, Japan) was used to observe larval behavior and morphology. The settlement and metamorphosis sequence, completed within 96 h, was characterized by progressive behavioral modifications (swimming, substrate contact, crawling, permanent settlement) and concurrent anatomical remodeling (shell growth, gill emergence, and juvenile form attainment).

2.8. Identification of Protein Concentration, Color, and Morphology of EVs

The 500 µL of EVs were resuspended in PBS and tested the protein concentration by a Modified BCA Protein Assay Kit (Sangon Biotech, China). Then, the EVs were diluted to the same concentration using PBS and the color of EVs was tested by measuring the absorbance values at 475 nm. The number and particle size of EVs was analyzed by using the Nanoparticle Tracking Analysis (NTA) system. To test the morphology of EVs, the EVs were negatively stained with osmium tetroxide and observed under a transmission electron microscope (TEM).

2.9. CytC Enzyme‐linked Immunosorbent Assay

To test CytC production in bacteria, the bacteria were cultured for 48 h using the streak plate method. Single colonies of uniform size were selected and suspended in 1 mL of AFSW, then ultrasonicated at 99% power for 2 min with a 5‐s on/10‐s off cycle. The suspension was centrifuged at 5000 rpm for 15 min, and the supernatant was collected. To test CytC content in biofilms, bacterial biofilms were scraped from 12 glass slides, suspended in 1 mL of AFSW, centrifuged at 5000 rpm for 15 min, and the supernatant was collected. To test CytC content in EVs, the EV suspension was adjusted to a protein concentration of 1000 mg/L, ultrasonicated at 99% power for 2 min with a 5‐s on/10‐s off cycle.

The content of CytC of each solution was determined using a microorganism CytC enzyme‐linked immunosorbent assay kit (MREDA, Cat. No. M132497, product link: https://www.mreda.com.cn/product/664793.html; China). The purified CytC antibodies were immobilized on a microplate to create solid‐phase antibodies. Samples containing CytC were then incubated with HRP‐labeled antibodies, forming an antibody‐antigen complex. Following a washing step, the substrate TMB (3,3′,5,5′‐Tetramethylbenzidine) was introduced. Under HRP catalysis, TMB turned blue and subsequently yellow upon acid addition. CytC concentrations in the samples were positively associated with the color intensity. Absorbance was measured at 450 nm, and CytC concentrations were quantified using a standard calibration curve. The schematic diagram of the standard curve is presented in Figure S3.

2.10. Biofilms Stained With Fluorescent Dyes and Biovolume Analysis

Biofilms were stained following the method of González‐Machado et al. (2018). The concanavalin A, tetramethylrhodamine conjugate (ConA‐TMR) (Invitrogen, Germany) was used for staining α‐polysaccharides, the calcofluor white M2R (CFW) (Sigma, United States) was for β‐polysaccharides, the fluorescein isothiocyanate isomer I (FITC) (Sigma, United States) was for proteins and DiIC18(5) oil, 1,1′‐dioctadecyl‐3,3,3′,3′‐tetramethylindodicarbocyanine perchlorate (DiD'oil) (Invitrogen, Germany) was for lipids. Three biological replicates were set for each dye. The Leica TCS SP8 CLSM and the LAS X Version (Leica, Germany) software were used for acquisition of CLSM images. Images were captured using a 63 × (1.4 NA) objective, then the biovolume of stained polysaccharides, proteins, and lipids was quantified by threshold analysis using ImageJ (National Institutes of Health, United States).

2.11. Assessment of eATP of Biofilms

For the assessment of extracellular ATP (eATP) in biofilms, six biofilm slides were grouped to form one replicate group. Biofilms were scraped off using sterilized glass slides and transferred into 1.5 mL centrifuge tubes containing sterilized double‐distilled water, and the final volume was adjusted to 1 mL. The samples underwent centrifugation at 860 × g for 10 min, after which the supernatant was retained as the ATP measurement sample from the biofilm. To evaluate eATP levels in the surrounding water of biofilms, 1 mL of water was collected and subjected to centrifugation at 860 × g for an additional 10 min. The supernatant then served as the ATP measurement sample from the surrounding water where the biofilm resided. For assessing eATP in EVs, EVs isolated from wild‐type strains, ΔccmH mutants, and ΔccmH::ccmH complemented strains were concentrated to a final concentration of 2000 mg/mL. The eATP from EVs was extracted by adding trichloroacetic acid (TCA; Sigma) to achieve a final concentration of 1%. Following incubation on ice for 10 min, pH adjustment of these samples was performed using Tris–HCl buffer solution until reaching pH 7.75; subsequently, the supernatant was collected as the ATP measurement sample from EVs.

Detection of ATP was carried out utilizing the Enliten ATP detection kit (Promega), with bioluminescence monitored using a non‐transparent format on a 96‐well plate. Each well contained an aliquot of 100 µL of sample along with an immediate injection of another 100 µL comprising rLuciferase/Luciferin (rL/L) reagent; thereafter, bioluminescence intensity readings were recorded promptly. Each experimental group set six replicates.

2.12. Natural Biofilm Development

Sterile and cleaned glass slides were fixed on PVC holders and submerged in the coastal seawater of Gouqi Island (30.73°N, 122.77°E) in Zhejiang Province, China. After remaining submerged for a total of 21 d (March 16th to April 6th, 2018), the slides had natural biofilms formed on them. Biofilms from Gouqi Island were harvested, placed into insulated containers filled with seawater, transported to the laboratory on ice, and utilized in bioassays on the same day.

2.13. Biofilm Collection and 16S rRNA Gene Sequencing

One thousand pediveliger larvae were transferred into a tank containing 1 L of filtered (0.22 µm) seawater, in which 50 glass slides with biofilms were placed. Each biofilm sample was prepared by scraping seven biofilms from glass slides and pooling them together. Biofilm samples were collected every 3 h during the settlement and metamorphosis process. Biofilm samples were cryopreserved in liquid nitrogen and maintained at −80°C. Biofilm sample DNA was extracted, PCR‐amplified using a barcoded primer set 338F (5′‐ACTCCTACGGGAGGCAGCAG‐3′) and 806R (5′‐GGACTACHVGGGTWTCTAAT‐3′), and sequenced on an Illumina HiSeq2500 platform at Guangdong Magigene Biotechnology Co., Ltd. (Guangzhou, Guangdong Province, China).

2.14. Electrochemical Testing

The biofilm was washed with AFSW before being taken out for electrochemical experiments. The AFSW was used as the electrolyte solution and the electrochemical testing was conducted on an electrochemical workstation (Gamryreference 600 plus, Gamry Instruments) at room temperature. The electrochemical cell was configured with biofilm as the working electrode, saturated calomel as the reference, and thin platinum electrode as the auxiliary electrode. The electrochemical impedance spectroscopy, cyclic voltammetry, and potential polarization tests was conducted on three bacterial biofilms in AFSW. Before testing, an open circuit potential test was run for 1 h. The frequency range of electrochemical impedance spectroscopy test was 100 000–0.1 Hz, with an amplitude of 5 mV. The scanning speed of the cyclic voltammetry test was 5 mV/s, and the scanning range was −0.6 to 0.2 V. The cyclic voltammetry test was performed three times and data from the second cycle were used as the final result for analysis. The polarization curve was measured within the range of corrosion potential (Ecvs. SCE) −0.5∼0.5 V, with a scanning rate of 1 mV/s.

2.15. GeoChip Hybridization and Data Processing

GeoChip 5.0, which contains more than 60 000 probes targeting diverse biogeochemical processes, was utilized to unveil the functional succession of biofilm communities during metamorphosis. The GeoChip hybridization experiments were conducted by Guangdong Magigene Biotechnology Co., Ltd. (Guangzhou, Guangdong Province, China). Briefly, 1 µg of biofilm DNA was labeled with the fluorescent dye Cy‐3 and hybridized with the GeoChip microarray overnight, as previously described (Shi et al. 2019). After scanning the arrays, the signals were converted into signal intensities using ImaGene 6.0. After quality control and data normalization, the signal intensities of specific genes were calculated as relative abundances.

2.16. Network Construction and Analysis

Seventeen networks were constructed for biofilm communities during larval settlement and metamorphosis, based on the relative abundances of OTUs at each time point. Covariations were measured across 12 biological replicates to create each network. Only OTUs detected in at least 8 of the 12 samples were included for network construction. RMT was used to identify the appropriate similarity threshold (St) prior to construction, St defines the minimum strength of the connections between node pairs (Zhou et al. 2010). Global network properties were calculated using the Molecular Ecological Network Analyses (MENA) Pipeline (http://ieg2.ou.edu/MENA/) (Deng et al. 2016). Cytoscape 2.8.2 was used for graphing networks (Shannon et al. 2003). Characters of each network were analyzed in the MENA region. The greedy modularity optimization method was used for detecting modules, which are groups of highly connected nodes (i.e., OTUs) (Newman 2006). Modularity indicates the extent to which a network is divided into modules (Newman 2006).

2.17. Molecular Docking

The three‐dimensional structure of the mussel TLR4 protein was predicted using AlphaFold3 (https://alphafoldserver.com/) (Abramson et al. 2024, Jumper et al. 2021), relying on the amino acid sequence. The predicted structure of TLR4 was processed with AutoDock Tools 1.5.6 (Morris et al. 2009) to prepare for molecular docking. This preprocessing included adding hydrogen atoms, assigning partial charges, defining atom types, and exporting the structure in PDBQT format. The ligand CytC was constructed in ChemDraw. The structure was then subjected to energy minimization with BIOVIA Discovery Studio Visualizer 2019, followed by preprocessing using AutoDock Tools 1.5.6 for hydrogen addition, charge calculation, and atom type assignment. Molecular docking was performed using AutoDock Vina (Eberhardt et al. 2021), generating 100 output poses. The pose with the lowest binding energy and a reasonable conformation was selected as the final docking result. Visualization of the docked ligand‐receptor complex was carried out with BIOVIA Discovery Studio Visualizer 2019.

2.18. Statistical Analyses

Normality was assessed using the Shapiro–Wilk test when n > 3; when n ≤ 3, normality was assessed by direct inspection of the distribution for symmetry and absence of extreme outliers. Homogeneity of variances was assessed using Levene's test. If the data met the assumptions of normality and homogeneity of variance (p > 0.05), comparisons between two groups were performed using independent samples t‐test, and comparisons among multiple groups were performed using one‐way ANOVA followed by Tukey's HSD post hoc test. If the data met normality but not homogeneity of variance (p < 0.05), Welch's t‐test was used for two‐group comparisons, and Welch's ANOVA followed by Games‐Howell post hoc test was used for multiple‐group comparisons. If the data did not meet normality, the Wilcoxon rank‐sum test was used for two‐group comparisons, and the Kruskal–Wallis test was used for multiple‐group comparisons; post hoc pairwise comparisons were performed using the Steel‐Dwass test for ≤ 5 groups, or the Dunn's test for > 5 groups. p‐value < 0.05 was considered statistically significant. All percentage data were arcsine transformed prior to analysis. The analysis of community diversity was performed using R version 3.6.3 software (R: vegan). Microbial successional patterns were revealed through detrended correspondence analysis (DCA) using the Bray‐Curtis dissimilarity. Three non‐parametric multivariate methods‐Adonis, ANOSIM, and MRPP were employed to evaluate biofilm community differences across developmental stages. The relative importance of stochastic processes was quantified by the modified stochasticity ratio (MST), as previously described (Ning et al. 2019). MST analysis can be performed using the package “NST” written with the R language. The MST index ranges from 0 to 1, where a value of 1 indicates purely stochastic assembly and a value of 0 signifies fully deterministic processes. A threshold of 0.5 serves to distinguish predominantly stochastic assembly (>0.5) from predominantly deterministic assembly (<0.5).

3. Results

3.1. CytC Is Required for the Biogenesis of CytC‐packaged EVs

S. marisflavi ECSMB14101, isolated from natural biofilms, formed a monospecific biofilm that exhibited potent inducing activity for larval settlement and metamorphosis. To elucidate the role of CytC in larval metamorphosis, an in‐frame deletion mutant of CytC biogenesis gene ccmH was constructed (Figure 1A and S1). The wild‐type colony exhibited a red‐brown phenotype, while the ∆ccmH colony displayed a light pink coloration, indicating a reduction in CytC content following the deletion of ccmH (Figure 1B). For lysis‐based isolation, the production of EVs exhibited a decline, with the size of the isolated EVs measuring approximately 50–200 nm as determined by transmission electron microscopy (Figure 1C). The color of isolated EVs shifted from red‐brown in wild‐type strains to light pink in the ∆ccmH mutants, indicating a reduction in the CytC content of EVs in ∆ccmH mutant (Figure 1D). The ∆ccmH showed a 38% reduction in EV production compared to the wild‐type strain (p < 0.0001), as measured by BCA protein assay (Figure 1E). For the supernatant‐based isolation, EVs derived from the ∆ccmH mutant also exhibited significantly lower protein yields compared to those from the wild‐type and complemented strains (Figure S4, p < 0.0001). The reddish color of wild‐type EVs containing enriched CytC correlated with a higher optical density at 475 nm, which was significantly higher than that of ∆ccmH mutant EVs (Figure 1F, p < 0.0001).

FIGURE 1.

CytC is required for the biogenesis of CytC‐packaged EVs and its inducing activity on larval metamorphosis of M. coruscus. (A) The gene cluster for CytC maturation and the location of ccmH within it. (B) The phenotypic characteristics of S. marisflavi bacterial colonies. (C) Transmission electron microscopy images of EVs, with black arrows indicating the presence of EVs. (D) Isolated EVs from bacterial cultures of different strains. (E) Protein quantification of bacterial EVs produced by various bacterial strains (n = 3). Statistical significance was determined by one‐way ANOVA followed by Tukey's HSD post hoc test. (F) Absorbance measurements of EVs at 475 nm (n = 3). Statistical significance was determined by one‐way ANOVA followed by Tukey's HSD post hoc test. (G) Nanoparticle tracking analysis of EVs. (H) CytC concentration in EVs isolated from different S. marisflavi strains (n = 12). Statistical significance was determined by one‐way ANOVA followed by Tukey's HSD post hoc test. (I) Inducing activity of isolated S. marisflavi EVs on larval metamorphosis (n = 6). Statistical significance was determined by the Kruskal–Wallis test followed by Steel‐Dwass post hoc comparisons. Data are presented as means ± SEM. *p < 0.05, **p < 0.01, ****p < 0.0001, and ns, not significant (p > 0.05).

To further verify that the particles obtained by the lysis‐based method were EVs rather than debris, purified EVs were subjected to immunomagnetic bead‐based enrichment using the Magnetic Exosomes Enrichment Kit. Over 90% of the protein content was retained by the beads after capture (p > 0.05), confirming that the majority of the isolated material consisted of vesicular structures (Figure S5). Meanwhile, EVs isolated by both lysis‐based and supernatant‐based methods exhibited comparable TEM morphologies, featuring intact membrane‐enclosed structures and an absence of noticeable contaminants (Figure S6).

The EV profiles of the wild‐type, ∆ccmH, and complemented strains showed comparable peaks, and the deletion of ccmH resulted in a marked reduction in peak area (Figure 1G). EV particle numbers were quantified per milligram of protein. The wild‐type (8.50 × 10¹¹ particles/mg), ∆ccmH mutant (8.43 × 10¹¹ particles/mg), and ∆ccmH::ccmH strains (8.51 × 10¹¹ particles/mg) exhibited no statistically significant differences (Figure S7, p > 0.05). Analysis further revealed a significantly decreased CytC content (p < 0.01) in ∆ccmH EVs relative to the wild‐type strains (Figure 1H). Moreover, quantitative analysis showed EVs derived from the ∆ccmH mutant consistently exhibited significantly lower CytC content levels than those from the wild‐type and ∆ccmH::ccmH strains, with reductions of 16% and 15% under lysis‐based and supernatant‐based isolation methods, respectively (Figure S8A, p < 0.05).

3.2. The CytC Cargo of EVs Is Essential for Larval Metamorphosis

To determine whether CytC‐packaged EVs influence larval metamorphosis of M. coruscus, the inducing activity of EVs isolated from wild‐type, ∆ccmH, and complemented strains was compared (Figure 1I). The inducing activities for the EPI and blank control groups were 51.2% and 0%, respectively. The inducing activity of wild‐type EVs was 35.0% ± 5.5%, which was 1.6‐fold higher than that of ∆ccmH EVs (p < 0.05). Furthermore, this activity was restored mostly to wild‐type levels upon genetic complementation (Figure 1I, p > 0.05). Notably, for wild‐type strain, EVs from the two isolation methods exhibited comparable metamorphosis‐inducing activities toward C. angulata larvae (Figure S8B, p > 0.05). The ΔccmH mutant‐derived EVs showed significantly decreased activity in inducing C. angulata larval metamorphosis compared with those from the wild‐type strain, with reductions of 47% and 43% under lysis‐based and supernatant‐based isolation, respectively (Figure S8B, p < 0.01).

3.3. The CytC Maturation Regulates Biofilm Formation and Its Inducing Activity

To prepare bacterial biofilms for larval metamorphosis assays, different strains were washed and co‐cultured with ITO‐coated glass slides for 48 h (Figure 2A). The biofilms were then transferred to Petri dishes containing AFSW for larval metamorphosis induction (Figure 2A). The inducing activity showed no significant difference between glass and ITO (indium tin oxide)‐coated surfaces (Figure S9A, p > 0.05). Bacterial density showed a significant difference between the two surfaces (Figure S9B, p < 0.0001). In M. coruscus larvae, the positive control, EPI, induced 45.8% of larvae to settle and undergo metamorphosis; the blank control group induced no larval metamorphosis. The metamorphosis‐inducing activity of the wild‐type biofilm on M. coruscus was approximately 44%. The ∆ccmH monospecific biofilm displayed a drastically reduced activity of only 14% in inducing M. coruscus metamorphosis, representing a 67% decrease compared to the wild‐type (Figure 2B, p < 0.05). The inducing activity on M. coruscus was rescued upon genetic complementation (∆ccmH::ccmH), confirming a functional role for ccmH in the inductive effects of the bacterial biofilm (Figure 2B, p > 0.05). In C. angulata larvae, the induction activity was 52.2% for the EPI control, whereas no larval metamorphosis occurred in the blank control. The ∆ccmH mutant biofilm exhibited significantly lower metamorphosis‐inducing activity on C. angulata, with a 35% reduction relative to the wild‐type (Figure S10, p < 0.01). The complemented strain restored the activity on C. angulata to the wild‐type level (Figure S10, p > 0.05).

FIGURE 2.

The CytC maturation regulates S. marisflavi biofilm formation and its inducing activity on larval metamorphosis of M. coruscus. (A) Bacterial biofilms were formed on the glass with ITO film, which were subsequently transferred to a Petri dish to induce larval metamorphosis. (B) The inducing activity of S. marisflavi biofilms on larval metamorphosis (n = 6). Statistical significance was determined by the Kruskal‐Wallis test followed by Steel‐Dwass post hoc comparisons. (C) The bacterial density of biofilm formed by different bacterial strains (n = 30). Statistical significance was determined by one‐way ANOVA followed by Tukey's HSD post hoc test. (D) Confocal images of biofilms, showing protein distribution. (E) The biovolume of proteins in bacterial biofilms (n = 9). Statistical significance was determined by Welch's ANOVA followed by Games‐Howell post hoc test. (F) CytC contents of biofilms (n = 4). Statistical significance was determined by one‐way ANOVA followed by Tukey's HSD post hoc test. (G) CytC contents of bacteria (n = 6). Statistical significance was determined by one‐way ANOVA followed by Tukey's HSD post hoc test. Data are presented as means ± SEM. *p < 0.05, ****p < 0.0001, and ns, not significant (p > 0.05).

The biofilm‐forming capacity of the ∆ccmH strain, as indicated by bacterial density, exhibited a 24% reduction compared to the wild‐type strain (Figure 2C, p < 0.0001). This capacity was restored to wild‐type levels following complementation with the ccmH gene (Figure 2C, p > 0.05). Confocal microscopy revealed a visible reduction in protein content within the ∆ccmH mutant biofilms (Figure 2D), a finding confirmed by quantitative analysis showing that the protein biovolume of the wild‐type strain was 3.4‐fold greater than that of the mutant (Figure 2E, p < 0.0001). There were no significant reductions observed in polysaccharides and lipids within the biofilms following the deletion of the ccmH gene (Figure S11, p > 0.05). The CytC production in bacterial colonies and biofilms was quantified. Compared to the wild‐type strain, the ∆ccmH mutant exhibited a 40% reduction in CytC production in bacterial cells (Figure 2G, p < 0.0001). Furthermore, the ∆ccmH biofilm showed a 21% reduction in CytC production compared to the wild‐type (38 nmol/L, p < 0.05), while complementation of the ccmH gene restored the level to that comparable to the wild‐type (Figure 2F and G, p > 0.05).

3.4. The Genetic Disruption of CytC Maturation Creates a Functional Deficiency in Biofilm EET and eATP

Redox‐active EVs play a crucial role as mediators in facilitating EET in natural ecosystems (Liu et al. 2020). Given the significant reduction in EV production observed in the ∆ccmH mutant, a further experiment was conducted to assess the regulatory role of CytC in biofilm EET (Figure 3). The biofilm served as the working electrode, while a saturated calomel electrode was employed as the reference electrode, and a thin platinum electrode functioned as the auxiliary electrode (Figure 3A). The impedance arc radius of ∆ccmH biofilms exhibited a 50% increase in resistance compared to wild‐type and ∆ccmH::ccmH biofilms, indicating that strains containing CytC facilitate electron transfer more effectively (Figure 3B). Potentiodynamic polarization studies demonstrated that the biofilms of both the wild‐type and ∆ccmH::ccmH strains exhibited lower current density and higher current efficiency in comparison to the ∆ccmH biofilms (Figure 3C). Furthermore, the distinct redox peak of biofilm decreased by 20% compared to the wild‐type biofilm, indicating that the deletion of the ccmH gene may impede direct extracellular electron transfer processes mediated by CytC (Figure 3D).

FIGURE 3.

The genetic disruption of CytC maturation creates a functional deficiency in biofilm EET and eATP. (A) Schematic representation of the electrochemical experimental process. (B) Nyquist diagram of S. marisflavi biofilm in electrochemical impedance spectroscopy test. (C) Polarization curves of S. marisflavi biofilms. (D) The cyclic voltammetry curve of biofilms. The extracellular ATP of (E) the isolated EVs, (F) biofilms, and (G) the surrounding seawater of biofilm of wild‐type strains, ∆ccmH, and ∆ccmH::ccmH (n = 6). For (E), statistical significance was determined by the Kruskal–Wallis test followed by Steel‐Dwass post hoc comparisons. For (F), statistical significance was determined by Welch's ANOVA followed by Games‐Howell post hoc test. For (G), statistical significance was determined by one‐way ANOVA followed by Tukey's HSD post hoc test. Data are presented as means ± SEM. *p < 0.05, **p < 0.01, and ns, not significant (p > 0.05).

The assessment of extracellular ATP (eATP) revealed that ccmH knockout reduced eATP levels, indicating that CytC is essential for maintaining intracellular electron transfer‐driven ATP synthesis. The results demonstrated that EVs isolated from wild‐type strains exhibited an eATP level that was 4.03 times higher compared to those derived from ∆ccmH strains (Figure 3E, p < 0.05). The eATP level in the wild‐type biofilms was 3.77 times higher than that in ∆ccmH biofilms (Figure 3F, p < 0.05), while eATP level in seawater from wild‐type was 4.91 times greater than that observed in ∆ccmH biofilms (Figure 3G, p < 0.01).

3.5. The Supplementation With CytC Restored EET and the Inductive Capacity of the Mutant Biofilm

To test the hypothesis that the impaired function of the ∆ccmH biofilm is due to a lack of CytC, we added purified CytC during biofilm formation to assess the recovery of inducing activity and EET (Figure 4A). With positive and negative controls yielding metamorphosis rates of 55.7% and 6.5%, respectively, the inducing activity of ∆ccmH biofilms supplemented with CytC (5 mg/L) on larval metamorphosis of M. coruscus was comparable to that of the wild‐type biofilms (Figure 4B, p > 0.05). Moreover, at concentrations of 5 mg/L, CytC restored the biofilm density to over 90% of the levels observed in both wild‐type and complemented (∆ccmH::ccmH) biofilms (Figure 4C, p > 0.05).

FIGURE 4.

The supplementation with CytC restored EET and the inductive capacity of the mutant biofilm on larval metamorphosis of M. coruscus. (A) Monospecific biofilms of S. marisflavi were established using wild‐type and ∆ccmH strains. The ∆ccmH + CytC biofilm was formed through co‐culturing CytC with bacterial strains for 48 h. Bioassays were conducted in Petri dishes containing the biofilm and ca. 20 larvae. (B) Inducing activity of the ∆ccmH + CytC biofilm, generated by adding CytC at varying concentrations (n = 6). Statistical significance was determined by the Kruskal–Wallis test followed by Dunn's post hoc comparisons. (C) Bacterial density of ∆ccmH + CytC biofilm (n = 30). Statistical significance was determined by the Kruskal‐Wallis test followed by Dunn's post hoc comparisons. (D) Nyquist diagram obtained from electrochemical impedance spectroscopy testing of the biofilms. (E) Polarization curves of biofilms. (F) Cyclic voltammetry curve of biofilms. Data are presented as means ± SEM. ***p < 0.001, ****p < 0.0001 and ns means not significant (p > 0.05).

The impaired electron transfer in the ∆ccmH biofilm, was evidenced by its high impedance arc radius. The ∆ccmH mutant exhibited a significantly larger radius of 1731 Ω, approximately 2.8‐fold and 2.2‐fold higher than that of the wild‐type and the ∆ccmH::ccmH biofilm (794 Ω), respectively (Figure 4D). CytC supplementation conversely reduced this impedance in a dose‐dependent manner, with a notable 43% decrease at 10 mg/L (Figure 4D). Moreover, the addition of CytC increased the current density in ∆ccmH biofilms in potentiodynamic polarization tests, with the 5 mg/L treatment restoring it to a level comparable to the ∆ccmH::ccmH biofilm (Figure 4E). The functional recovery of electron transfer in the ∆ccmH biofilm, as further confirmed by cyclic voltammetry, was marked by a heightened oxidation peak following CytC supplementation, demonstrating restored EET capability (Figure 4F).

3.6. The Supplementation With CytC‐packaged EVs Restored the Inductive Capacity of the Mutant Biofilm

To directly assess the functional capacity of CytC‐packaged EVs in biofilm‐mediated larval development of M. coruscus, purified EVs were exogenously added during biofilm formation (Figure 5A). The positive control induced a metamorphosis rate of 37.2%, whereas the negative control group showed a rate of 0%. Upon addition of 50 or 100 mg/L EVs, the inducing activity of ∆ccmH biofilms on M. coruscus increased approximately 1.2‐fold compared to untreated biofilms (Figure 5B). Although pairwise comparisons between the untreated ∆ccmH biofilms and each of the EV‐treated groups revealed no significant differences (p > 0.05), a general trend of a gradual, non‐linear increase in larval metamorphosis rate was observed with increasing EV concentrations. This phenotypic rescue supports the involvement of CytC‐packaged EVs in the developmental induction mechanism of M. coruscus.

FIGURE 5.

The supplementation with CytC‐packaged EVs restored the inductive capacity of the ∆ccmH mutant biofilm on larval metamorphosis of M. coruscus. (A) Monospecific biofilms of S. marisflavi were established using wild‐type and ∆ccmH strains. The ∆ccmH + EVs biofilm was generated by co‐culturing ∆ccmH strains and EVs for 48 h. Bioassays were performed in Petri dishes containing the biofilm and 20 larvae. (B) The inducing activity of ∆ccmH + EVs biofilm, formed by adding EVs at different concentrations (n = 9). Statistical significance was determined by the Kruskal–Wallis test followed by Dunn's post hoc comparisons. Particle numbers of the wild‐type, ∆ccmH, and ∆ccmH::ccmH EVs showed no significant differences (Figure S4, p > 0.05). Data are presented as means ± SEM. ***p < 0.001, ****p < 0.0001 and ns, not significant (p > 0.05).

3.7. TLR4 Protein Binds to CytC

CytC bonded with TLR4 of M. coruscus through multiple interactions and the results of docking are shown in Figure 6. CytC formed hydrogen bonds with the TLR4 residues SER65, LYS113, and ASN45. Electrostatic interactions between CytC and TLR4 were observed with ASP67 and GLU89, and hydrophobic interactions were observed with LEU7, ILE10, TYR11, VAL14, and ALA22. The calculated binding free energy was −8.625 kcal/mol, corresponding to a spontaneous binding process and a stable complex.

FIGURE 6.

Structure of the TLR4‐cytochrome C complex. (A) The TLR4 protein of M. coruscus is depicted as a green cartoon model. The CytC molecule is shown in a ball‐and‐stick representation, colored by atom type: oxygen atoms in red, nitrogen atoms in blue‐purple, polar hydrogens in white, and carbon atoms in blue. (B) The 3D interactions between the molecule and protein using a ball‐and‐stick model. The interacting protein residues are shown as green sticks. (C) The 2D interactions between the protein and the molecule. The molecule is displayed in a ball‐and‐stick model.

3.8. Redistribution of Shewanella and Enrichment of CytC‐related Genes During Larval Metamorphosis in Natural Biofilms

To understand the dynamic roles of Shewanella and CytC in the bacterial community during larval metamorphosis, we examined the biofilm profiles and larval development every 3 h during the 48 h metamorphosis of M. coruscus (Figure 7A). Natural biofilms induced almost 60% larval metamorphosis of M. coruscus (p < 0.0001); during this period, larvae underwent a series of behavioral and morphological changes (Figure S12). We obtained an average of 64 300 ± 3800 sequence reads per sample for bacterial biofilm communities. The complementary non‐parametric multivariate statistical tests, including the permutational multivariate analysis of variance (Adonis, F = 11.03, p = 0.001), analysis of similarity (ANOSIM, R = 0.54, p = 0.001), and multiple response permutation procedure (MRPP, δ = 0.27, p = 0.001), showed that the community compositions were different among the four stages (0–9 h, pre‐metamorphosis; 12–24 h, metamorphic onset to juvenile appearance; 27–36 h, sustained increase in juvenile percentage; 39–48 h, plateau in juvenile percentage). As visualized by detrended correspondence analysis (DCA), the microbial communities significantly diverged after 36 h (Figure 7B). Deterministic processes dominated community assembly during most time points from 0–36 h (Figure 7C, p < 0.0001), while the stochastic processes were dominant during the later stage (39–48 h) of larval metamorphosis (Figure 7C, p > 0.05).

FIGURE 7.

Dynamic profile of microbial community structure and CytC related genes in natural biofilm during larval metamorphosis of M. coruscus. (A) Experimental design. Nature biofilms collected from the sea were co‐cultured with larvae to assess their inducing activity. Biofilm was taken every 3 h and sequenced by 16S rDNA and Geochip 5.0 (n = 12). The rate of larval metamorphosis induced by biofilms was recorded (n = 9). (B) Detrended Correspondence Analysis (DCA) was performed on bacterial communities at the OTU level. (C) MST indices of the bacterial community within biofilms were presented (n= 12). Statistical significance was determined by the Kruskal–Wallis test followed by Dunn's post hoc comparisons. (D) Succession of Shewanella species in biofilm networks during larval settlement and metamorphosis, with Shewanella highlighted by a red square. Networks represented random matrix theory co‐occurrence models derived from 12 biological replicates across 17 time points. Nodes represented OTUs, and the links between these nodes indicated significant correlations. Modules were randomly assigned colors at each time point, while nodes within modules containing fewer than five nodes were colored black. (E) The heat map illustrating the gene encoding CytC in relation to larval settlement and metamorphosis. Data are presented as means ± SEM. ****p < 0.0001 and ns, not significant (p > 0.05).

The biofilm networks at 48 h displayed a higher number of connections (links) between nodes compared to those at 0 h (Figure 7D). Prior to inducing larval metamorphosis, only one node from the Shewanella genus was present in a small module of the biofilm. With larval metamorphosis, the number of nodes belonging to the Shewanella genus expanded and relocated to a larger module with more connections to other species (Figure 7D). This suggests that the inductive bacterial species from the Shewanella genus increase along with larval metamorphosis, and ecological interactions within bacterial communities may influence larval metamorphosis. The correlation networks were constructed to uncover potential species interactions within biofilms. The majority of correlations were positive (Figure S13), implying that the bacterial species within the biofilms share a similar ecological niche or engage in positive interactions during larval metamorphosis. From the result of Geochip 5.0, the gene abundance of CytC exhibited a significant increase during the later stage of larval metamorphosis (i.e., 27, 36, 48 h; Figure 7E). Concurrently, there was also a notable significant increase in the abundance of hydrogenase, an additional enzyme involved in electron transfer (Figure 7E).

4. Discussion

EVs play significant roles across various domains, including horizontal gene transfer, biofilm formation, cell communication, and interactions between bacteria and their hosts (Roier et al. 2016). Recently, EVs have served as universal chemical signals to regulate the settlement and metamorphosis of mussel larvae. This study established that CytC‐packaged EVs of S. marisflavi are novel key chemical cues for larval metamorphosis in the mussel M. coruscus. We demonstrated that deletion of ccmH gene in the S. marisflavi strain could directly decrease the level of EVs production, affecting CytC content of EVs. We established that the diminished inducing activity of the ccmH mutant biofilms was primarily due to the loss of CytC‐packaged EVs, confirming CytC as the key bioactive component essential for biofilm‐induced larval metamorphosis. We further revealed that CytC modulates key energy production systems including EET and eATP generation. Molecular docking analysis suggested that bacterial CytC can bind to TLR4, a receptor known for its conserved function in sensing microbial patterns in marine invertebrates. Moreover, the importance of CytC extends from single bacteria to complex biofilm communities, as evidenced by increased CytC gene abundance in natural biofilms during larval metamorphosis. Given the broad phylogenetic distribution of CytC across bacterial species, it potentially acts as a universal regulator, influencing larval metamorphosis through its role in EV production in natural environments.

It has long been recognized that larvae of a broad spectrum of marine invertebrate taxa do not settle randomly across the bottom of the sea and survive only if they settle a site where they can successfully metamorphose (Hadfield and Paul 2001). Upon detecting inductive cues from the “right spots” via chemoreceptors, larvae transduce these signals through neural or hormonal pathways, ultimately triggering metamorphosis (Hadfield and Paul 2001). Biofilms act as a vehicle containing bacteria and their self‐produced EPS, which include proteins, polysaccharides, lipids, and other signaling molecules (Flemming and Wingender 2010). EVs are also commonly found in biofilm EPS (Saad et al. 2024). Biofilms are known to induce larval metamorphosis in many invertebrate species (Hadfield and Paul 2001). Our previous studies demonstrated that EPS components such as flagellin, colanic acid, and lipopolysaccharides possess inducing activity on mussel larval metamorphosis (Hu et al. 2024a; Liang et al. 2020; Peng et al. 2020). However, the signals within EPS responsible for this inductive effect have not been fully elucidated. Previous studies have found that the induction activities of bacterial biofilms and EVs display a similar trend (Hu et al. 2024b). Furthermore, the EPS containing EVs of P. marina biofilms showed similar inducing activity with that of biofilms; however, after removing the EVs in the EPS by ultracentrifugation, EPS without EV was similar to the AFSW control group without significant differences (Hu et al. 2024a). Thus, it becomes clear that the EVs are important for EPS biofilm in inducing larval settlement and metamorphosis. However, the main chemical cues remain poorly understood.

EVs contain a diverse array of chemical components including DNA, proteins, and small signaling molecules. EVs exhibit high environmental stability at higher ambient temperatures and a wide range of pH values. These capacity enable them to deliver a broad spectrum of cargo over long distances while shielding the contents from degradation in the external environment (Bomberger et al. 2009). For example, bacterial DNA of potential EV origin has been detected in human blood and urine (Park et al. 2018; Yoo et al. 2016) and even in normally sterile sites like the heart (Svennerholm et al. 2017), implying that EVs can migrate far from where they are produced. The remarkable stability of EVs makes them a highly efficient delivery system in marine biofilms, enabling the precise transmission of inductive cues for larval metamorphosis. Beyond the well‐established role of LPS as an inductive cue in EVs, our data demonstrate that CytC acts as a novel essential signaling molecule within EVs that induces larval metamorphosis. The inducing activity of ∆ccmH mutant EVs on M. coruscus larvae was drastically lower than wild‐type EVs, demonstrating that CytC is the key inductive cue within EVs responsible for metamorphosis (Figure 1I). Meanwhile, this phenomenon was also observed on larval metamorphosis of C. angulata (Figures S8B and S10). By extending the assay to C. angulata, we have thereby demonstrated that the metamorphosis‐inducing capacity of CytC in EVs is not confined to mussel, but reflects a general principle shared among invertebrate species.

The CytC functions as an immune mediator in maintaining homeostasis, regulating cell apoptosis and survival pathways, and participating in immune responses (Zhou et al. 2024). In vertebrates, CytC is released to the extracellular space and enhances apoptosis induced by caspase‐9 and caspase‐3 activation (Jiang and Wang 2004). During larval metamorphosis in insects, larval tissues undergo autophagy, apoptosis, and cell proliferation, among which apoptosis is an essential process (Lee and Park 2021). The release of CytC is necessary for the activation of caspase‐3 during apoptosis in lepidopteran insects (Xie et al. 2016). In M. coruscus, caspase genes showed high expression at the pediveliger larvae stage, and the apoptosis‐mediated larval metamorphosis was confirmed through TUNEL apoptosis detection in the velum (Cao et al. 2024). A key unresolved question is whether CytC‐packaged EVs influence apoptotic signaling during larval metamorphosis. Molecular docking analysis indicated that CytC can bind to mussel TLR4, pointing to a potential receptor‐mediated mechanism for metamorphosis induction (Figure 6). Based on the broad phylogenetic distribution of CytC among bacteria and the wide conservation of the TLR4 receptor in animals, we hypothesize that EVs containing CytC function as universal metamorphosis‐inducing signals, thereby connecting environmental signals to the transition between life stages in mussels and mediating metamorphosis development.

Despite extensive research on EVs, the mechanisms underlying their production and regulatory processes remain largely unknown. This study revealed that disrupting CytC maturation reduced both the production and CytC content of EVs (Figure 1). Most studies indicate that EV production originates from the outer membrane (OM) through three proposed models. One model suggests that EV formation is driven by differential OM growth rates caused by the loss or relocation of its covalent bonds to the peptidoglycan layer (Schwechheimer and Kuehn 2015; Roier et al. 2016). The second model suggests that the accumulation of peptidoglycan fragments or misfolded proteins within the periplasmic space generates turgor pressure on the OM, leading to its bulging and eventual pinching off (Schwechheimer and Kuehn 2015; Roier et al. 2016). The third model proposes that specific molecules induce curvature in the membrane (Schwechheimer and Kuehn 2015; Roier et al. 2016). The production of EVs is an expensive process that requires significant expenditure of energy for prokaryotic cells (Kulp and Kuehn 2010). The energy (ΔG) required to form a spherical vesicle from a flat membrane is estimated to be ΔG∼250–600 kBT (where kBT represents thermal energy), indicating that this process is non‐spontaneous (Bloom et al. 1991; Gui et al. 2016). Although EV production is highly energy‐consuming, it is notable that the associated energy costs have long been overlooked. Our results show that the ΔccmH mutant produced fewer EVs and exhibited a concurrent decrease in eATP and EET levels (Figure 3). A previous study demonstrated that the CytC‐deficient mutant also showed reduced electron transfer efficiency (Yang et al. 2017). These results suggest that a disruption in intracellular energy metabolism is the cause of the reduced EV production. There are two possible routes for CytC regulating EVs production.  The first is that, when ccmH is knocked out, the CytC can not be completely synthesized, and there are many misfolded proteins which accumulated in the periplasmic space (Schwechheimer and Kuehn 2015). Secondly, the energy production related to the CytC was blocked, the lack of energy may cause a reduction in EVs production. Similarly, it was proved that under iron restriction conditions, the levels of lipopolysaccharides (LPS) in Helicobacter pylori and EV decrease, possibly due to reduced energy supply (Keenan et al. 2008). This study therefore provides the first evidence that CytC, which is involved in energy production, is critical for the EV biogenesis.

The biosynthesis of CytC is facilitated by the CytC maturation (ccmABCDEFGH) system. Deletion of the ccmH gene in the classical Escherichia coli model has been shown to nearly abolish the synthesis of CytC (Fabianek et al. 1999). The CytC of ccmH mutant in this study decreased, but the mutant did not lose all of the CytC production (Figure 2G). It is consistent with previous reports that ccmH function is not conserved and varies across bacterial species. In Desulfovibrio desulfuricans, its Ccm system naturally lacks CcmH, yet it can still effectively catalyze the maturation of CytC (Goddard et al. 2010). In Shewanella oneidensis, it was demonstrated that although CcmH is required for efficient CytC synthesis, its deletion does not result in a complete loss; residual CytC could still be detected in the ΔccmH mutant strain (Jin et al. 2013). Multiple subsequent studies have also provided evidence for compensatory mechanisms or alternative pathways in CytC synthesis (Guo et al. 2022; Han et al. 2023). Our study indicates the functional diversity of CytC maturation systems among different bacterial species.

The purine adenosine 5′‐triphosphate (ATP) is a universal intracellular energy carrier, synthesized through CytC‐mediated electron transfer in Gram‐negative α‐ and γ‐proteobacteria. CytC mediates electron transfer to terminal electron acceptors under both aerobic and anaerobic conditions to drive ATP synthesis. The present results support that reduced CytC maturation effectively suppresses EET in biofilms, consequently lowering energy production. Multi‐heme CytC‐based electron transfer enables the bacterium to utilize nutrients for ATP synthesis (Breuer et al. 2015; Kaila and Wikström 2021). During biological oxidation, high‐energy electrons derived from metabolites must subsequently be transferred through a network of redox proteins to reach a terminal electron acceptor (Breuer et al. 2015; Kaila and Wikström 2021). Under aerobic conditions, electrons are ultimately transferred to oxygen, reducing it to water (Breuer et al. 2015; Kaila and Wikström 2021). Under anaerobic conditions, electrons can be transported across the cytoplasmic membrane via CytC to the terminal electron acceptors like sulfite, thiosulfate, elemental sulfur, and oxidized metals (Breuer et al. 2015; Kaila and Wikström 2021). The eATP can be produced by bacteria in culture fluid at concentrations ranging from micro‐ to millimolar levels (Ivanova et al. 2006). As electrochemically active bacteria (EAB), Shewanella species are distinguished by their capacity to utilize a diverse array of electron acceptors for energy generation (Hau and Gralnick 2007). We proved that the ccmH deletion in S. marisflavi disrupts intracellular electron transport and energy production, which is further supported by the significantly reduced eATP production in both biofilms and EVs of the ΔccmH mutant (Figure 3). Moreover, bacterial eATP acts as an inter‐kingdom signaling molecule in various organ systems, regulating inflammation, cell growth, survival, and drug resistance via multiple signaling pathways (Evers et al. 2022; Spari and Beldi 2020). Bacteria also can actively uptake and metabolize ATP in the periplasmic space and cytoplasm, depending on their physiological requirements (Alvarez et al. 2017; Spari and Beldi 2020). As larvae undergo metamorphosis, their metabolic rate increases (Lucas et al. 1979), and the successful completion of metamorphosis partly depends on the larvae's ability to mobilize substrates to meet their metabolic needs (Sánchez‐Lazo and Martínez‐Pita 2012). It needs to be clarified in further studies whether the eATP production regulated by CytC plays a role in larval metamorphosis by providing energy or signaling.

Larval settlement and recruitment in natural ecosystems are induced by chemical cues from biofilms with multispecies bacteria. Although CytC‐containing EVs induced larval metamorphosis in Shewanella monospecies biofilms, the role of Shewanella bacteria and CytC in natural biofilm‐induced metamorphosis remains to be elucidated. Clarifying this is essential for understanding host‐microbe interactions in marine ecosystems. Shewanella sp. isolated from the multispecies natural biofilms not only induces larval metamorphosis in both mussels and sea urchins (Huggett et al. 2006; Yang et al. 2013), but also may affect the settlement of planktonic oyster larvae by producing exopolysaccharides and melanin precursors (Mitra et al. 2015). Certain Shewanella sp. strains are utilized as probiotics in aquaculture to protect fish, molluscs, and echinoderms from infections (Chi et al. 2014; Nayak 2010). Following the initiation of larval metamorphosis, Shewanella shifted from the periphery to the core of the natural biofilm community (Figure 7B). This shift suggests that Shewanella tended to interact with a wider range of species (Guimerà and Nunes Amaral 2005) and participated in a greater number of functional modules (Guimerà and Nunes Amaral 2005), performing analogous functions such as ecological niche overlap (Faust et al. 2015), electron transfer (He et al. 2023), or co‐evolution (Toju et al. 2017). The predominantly positive correlations in the co‐occurrence network (Figure S13) provide further support for this proposition, as they reflect cooperative ecological processes such as bacterial co‐colonization and niche overlap (Faust et al. 2012). By testing the gene abundance of biofilm community, we found a significant upregulation of CytC related genes (Figure 7D). This study proposes that as Shewanella transitions into a keystone taxon, it synergizes with CytC‐upregulated energy metabolism to orchestrate a functional shift in the microbiome, thereby driving the critical life‐history transition of marine invertebrates.

Collectively, for the first time, we propose a novel paradigm wherein CytC acts as a bioelectrical driver for EV production, thereby orchestrating biofilm‐dependent larval metamorphosis. This discovery offers a novel bacteria‐larvae interaction mechanism mediated by the conserved electron carrier CytC within EVs. Future studies should focus on elucidating the activation of host cell signaling pathways by the CytC of EVs and their underlying molecular mechanisms.

Author Contributions

X.L., J.L.Y., and J.X.X. designed the study. Y.T., L.H.P., J.Q.H., and Y.Y.W. conducted experiments. L.H.P. analyzed the data. L.H.P., X.L., and J.L.Y. were major contributors in writing the manuscript. All authors read and approved the final manuscript.

Funding

This study was supported by the Natural Science Foundation of Shanghai (25ZR1401161), the National Key Research and Development Program of China (No. 2023YFE0115500), and the National Key Research and Development Program of China (No. 2023YFD2401902).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting Information: jev270373‐sup‐0001‐SuppMat.docx

Data Availability Statement

All sequences are under the accession number: PRJNA886047 in the SRA database.