Phylogenetically distinct Vibrio mediterranei lineages confer robust protection under thermal stress against oyster pathogens

Abstract

Introduction:

Marine bivalve mortality events cause substantial economic losses in aquaculture and threaten global food security. While pathogenic Vibrio species are frequently implicated, growing evidence suggests that loss of beneficial microbes can increase host susceptibility to disease. We previously observed that Vibrio mediterranei was consistently isolated from healthy oysters but systematically disappeared prior to mortality events, coinciding with proliferation of pathogenic Vibrio species. Here, we test whether this pattern reflects a protective functional role.

Methods:

Screening eleven V. mediterranei strains against eastern oyster (Crassostrea virginica) larvae revealed three distinct virulence phenotypes – avirulent, pathogenic, and intermediate – that correspond to monophyletic phylogenetic clades.

Results:

Pre-colonization with the avirulent strain Vm02 increased larval survival from 10–19% (pathogen-only controls) to 94–97% when challenged with V. harveyi or V. coralliilyticus, representing near-complete protection maintained at both ambient (28 °C) and thermal stress (32 °C) temperatures. Protection was rapid, effective even under simultaneous co-inoculation with the pathogen, and durable through >96 hours post-exposure. Fluorescence microscopy confirmed persistent association of fluorescently tagged Vm02 with larval digestive tissues. Phylogenomic analysis of 33 V. mediterranei genomes placed the three phenotypic groups within well-supported clades separated by 97.1–97.8% average nucleotide identity, approaching species-level thresholds. Pangenome analysis revealed that protective-clade strains harbor 230 unique orthogroups encoding regulatory systems, stress tolerance, metabolic versatility, and a conserved bacteriocin biosynthetic operon. These same strains do not encode the Type I secretion system, Type VI secretion system variants, and TCP pathogenicity island found in pathogenic lineages.

Discussion:

Together, these findings demonstrate that beneficial and pathogenic phenotypes are phylogenetically constrained within distinct V. mediterranei lineages, providing a framework for probiotic development and disease forecasting in shellfish aquaculture.

1 Introduction

Marine bivalve mortality events represent a major obstacle to the economic viability of aquaculture, causing losses in the billions of dollars annually and undermining the consistency and reliability of shellfish markets while threatening food security across coastal regions (Rato et al., 2022; Masanja et al., 2024; Smaal et al., 2019; The state of food security and nutrition in the world 2023, 2023). These events impact all major shellfish species and can result in stock losses exceeding 90%, with frequency and severity increasing due to ocean warming and environmental stress (Stewart-Sinclair et al., 2020; Ben-Horin et al., 2024; Matkarimov et al., 2025). Disease management efforts have largely focused on identifying and controlling pathogenic microorganisms, particularly Vibrio species, which are frequently implicated as causative agents of mortality across marine invertebrates (Richards et al., 2015; Bruto et al., 2017; Xu et al., 2022). Members of the Harveyi clade, including V. harveyi and V. coralliilyticus, are particularly problematic in shellfish hatcheries where they cause acute larval mortality through extracellular toxins and proteases (Zhang et al., 2020; Richards et al., 2015; Ushijima et al., 2022).

However, increasing evidence suggests that a pathogen-centric view overlooks critical features of host-microbe interactions that govern disease susceptibility. Across terrestrial, human, and marine systems, the depletion of beneficial microbes can predispose hosts to pathogenic colonization and contribute to dysbiosis-driven mortality (Zaneveld et al., 2017; Egan and Gardiner, 2016; Stevens et al., 2021). In Pacific oysters infected with Ostreid herpesvirus (OsHV-1), microbiome destabilization precedes mortality and involves depletion of diverse bacterial communities coupled with proliferation of opportunistic Vibrio species (Petton et al., 2021; De Lorgeril et al., 2018; Clerissi et al., 2020). Similar patterns occur in scallops exposed to harmful algal blooms, where paralytic shellfish toxins induce gut dysbiosis characterized by increased Vibrio abundance and reduced microbial diversity (Wei et al., 2024). These observations suggest that community composition, including the presence or absence of specific beneficial taxa, may ultimately determine disease trajectories in marine bivalves.

Our previous analysis of Vibrio communities in eastern oysters (Crassostrea virginica) revealed distinct assemblages associated with healthy versus diseased states (Smith et al., 2025). Most notably, Vibrio mediterranei was consistently isolated only in low-mortality periods and was systematically absent during mortality progression. The consistent loss of V. mediterranei associated with mortality raised the possibility that specific beneficial Vibrios may influence host resilience.

Vibrio mediterranei is ubiquitous across marine habitats, isolated from seawater, sediments, and diverse marine invertebrate hosts worldwide. Yet this species presents an ecological paradox: while some strains cause disease in corals (Kushmaro et al., 2001; Rubio-Portillo et al., 2020), pen shells (Andree et al., 2021), razor clams (Fan et al., 2023), and macroalgae (Yang et al., 2020), other strains are consistently recovered from healthy animals, including bacteriocin-producing isolates with probiotic potential (Serrano et al., 2018; Carraturo et al., 2006; Worden et al., 2022; Smith et al., 2025). V. mediterranei was first identified as a coral pathogen based on temperature-depending bleaching of Oculina Patagonia, with subsequent work identifying superoxide dismutase and a photosynthesis-inhibiting peptide toxin as key virulence factors (Banin et al., 2001, Banin et al., 2003). Transcriptomic analyses of coral co-infections revealed that V. mediterranei upregulates genes encoding the type VI secretion system (T6SS), siderophores, lytic enzymes, and toxins during interspecific competition with V. coralliilyticus (Rubio-Portillo et al., 2020). In bivalve systems, pathogenic strains associated with pen shell mortality carry virulence-associated genes (sod, rtx, and mshA) encoding virulence factors with temperature-dependent expression (Andree et al., 2021), while strains causing razor clam mortality secrete extracellular products including proteases that vary in composition between high- and low-virulence strains (Fan et al., 2023, Fan et al., 2024). Comparative genomics has further identified a toxin-coregulated pilus pathogenicity island (TCP-PAI) variably distributed among pathogenic isolates (Zhang et al., 2025). Many of these characterized pathogenic strains have been genome-sequenced and are included in our comparative analysis (Table 1).

Figure 1

Such dual ecological roles are well-established in other bacterial systems; Staphylococcus aureus, for example, exists as a harmless commensal in many humans while causing severe disease in others. These contradictory observations, together with the systematic disappearance of V. mediterranei prior to oyster mortality events (Smith et al., 2025), suggest that strain-level variation determines whether this species functions as a pathogen or beneficial symbiont.

Environmental temperature strongly influences Vibrio community composition and virulence (Kimes et al., 2012; Green et al., 2019; Vezzulli et al., 2016). Temperature-dependent virulence has been demonstrated in V. mediterranei infections of Pinna nobilis (Andree et al., 2021), and in coral systems where warming drives synergistic co-infection dynamics (Rubio-Portillo et al., 2020). Given that marine heatwaves are intensifying globally (Oliver et al., 2018), it is essential to understand how temperature affects both pathogenic and beneficial interactions between hosts and their associated microbiota.

Based on these observations, we hypothesized that: (1) V. mediterranei isolates consistently associated with healthy oysters function as protective symbionts rather than opportunistic colonizers; (2) protective versus pathogenic phenotypes represent phylogenetically constrained strategies within the V. mediterranei species complex; and (3) loss of protective V. mediterranei strains may create ecological conditions that facilitate pathogenic Vibrio-induced mortality.

Here, we test these hypotheses using controlled small-scale larval challenge experiments in which larval survival, bacterial colonization patterns, and physiological conditions were monitored, combined with comparative genomics across strains isolated from healthy and diseased hosts. We show that a specific V. mediterranei lineage provides rapid and robust protection against established oyster pathogens under thermal stress conditions, and that beneficial and pathogenic phenotypes are constrained within distinct evolutionary clades. These findings provide mechanistic insight into how mutualistic traits evolve within Vibrio species complexes and offer practical applications for disease forecasting and biological control in aquaculture systems.

2 Materials and methods

2.1 Bacterial strains and culture conditions

All bacterial strains used in this study are listed in Table 1. Avirulent Vibrio mediterranei strains Vm02 (PNB23_20_7), PNB23_3_3, PNB22_9_1, PNB22_6_1, PNB22_7_2, and PNB22_2_1 were isolated in 2022–2023 from healthy adult eastern oysters (Crassostrea virginica) in North Carolina as described previously (Smith et al., 2025). Intermediate V. mediterranei strains PPP23_1_1 and PPP23_1_2 were isolated from healthy adult eastern oysters at a separate North Carolina commercial aquaculture farm during the same sampling period. Intermediate strains FYC25 and FYC26 were isolated by the University of Rhode Island from larvae collected from a mortality event at an eastern oyster hatchery in the Northeast US.

Pathogenic V. mediterranei strains McD53–9 and McD43–10 were isolated from Montastraea cavernosa corals displaying tissue loss lesions in the Florida Keys, USA (corals collected from Sand Key under permit issued by the Florida Keys National Marine Sanctuary [FKNMS-2017-128-A2]). Coral samples were collected and processed as previously described (Ushijima et al., 2020). However, initial coral tissue/mucus slurry samples were plated onto seawater agar (SWA) and incubated under anaerobic conditions at 28 °C within a BD GasPack EZ container system (Becton, Dickinson and Company) for five days. Colonies were streaked for purification on SWA before being incubated again anaerobically. Purified strains were grown in seawater broth (SWB) before being cryopreserved with 20% (v/v) glycerol at -80 °C.

Pathogenic controls included V. harveyi Vh11 (DNB23_25_5), isolated from oysters during a 2023 mortality event (Smith et al., 2025), and V. coralliilyticus OCN008, previously characterized as a coral and oyster pathogen (Ushijima et al., 2014, Ushijima et al., 2022).

All strains were stored at -80 °C in Marine Broth 2216 (MB2216; BD Difco) supplemented with 20% (v/v) glycerol. For experiments, frozen stocks were streaked onto MB2216 agar plates and incubated at 28 °C for 24h. Single colonies were inoculated into 3 mL MB2216 broth and grown overnight (16–18 hours) at 28 °C with shaking at 200 rpm. On the day of experimentation, overnight cultures were diluted 1:100 into fresh MB2216 and grown to mid-exponential phase (OD600 = 0.3-0.5, ~3–4 hours) at experimental temperature (28 °C or 32 °C). Cells were harvested by centrifugation (7, 000 × g, 3 min, room temperature), washed twice in sterile artificial seawater (ASW; Instant Ocean, 30 ppt), and resuspended to 1×10⁸ CFU mL^-1 based on optical density.

2.2 Construction of GFP-expressing V. mediterranei Vm02

For colonization imaging, V. mediterranei Vm02 was transformed with plasmid pVSV102 (Dunn et al., 2006), which constitutively expresses green fluorescent protein (GFP) and confers kanamycin resistance. Transformation was performed by tri-parental conjugative mating using Escherichia coli S17-1 λpir as the donor. Donor and recipient cells were grown to mid-exponential phase, mixed at a 1:1 ratio with helper plasmid pEVS104 (Dunn et al., 2006), pelleted by centrifugation, and spotted onto MB2216 agar plates. After 8 hours incubation at 28 °C, cells were resuspended in MB2216 broth and plated on LBS agar supplemented with 50 µg mL^-1 kanamycin. Fluorescent colonies were confirmed by epifluorescence microscopy and maintained on MB2216 agar containing 50 µg mL^-1 kanamycin.

2.3 Larval oyster culture

Eastern oyster (C. virginica) larvae were obtained from the Shellfish Research Hatchery (SRH; University of North Carolina Wilmington, NC, USA) and maintained following standard hatchery procedures. Larvae were reared in filtered, UV-treated water obtained from the hatchery at 28 °C and ~30 ppt with gentle aeration. Algal diets consisted of a mixed diet (Chaetoceros mulleri, C. neogracilis, Isochrysis lutea, and Pavlova lutheri) supplied at rations following Helm et al., 2004 (Helm et al., 2004).

For challenge experiments, veliger stage larvae (5–7 days post-fertilization) were suspended in rearing water to a working density of ~20 larvae mL^-1. Larvae were distributed into sterile 50 mL polystyrene tubes at ~400 larvae per replicate volume (20 mL) and acclimated at the target experimental temperature (28 °C or 32 °C) for 1h prior to bacterial inoculation.

2.4 Larval challenge experiments

2.4.1 Single-strain exposure and temperature dependence

To compare baseline virulence of avirulent, intermediate, and pathogenic V. mediterranei strains and control V. harveyi and V. coralliilyticus pathogens to C. virginica larvae, larvae were exposed to individual strains at a defined cell density. Mid-exponential phase cultures were washed and resuspended in ASW as described above, then added to larval suspensions to achieve final concentrations of 10⁶ CFU mL^-1. Experiments were conducted at 28 °C (ambient hatchery temperature) and 32 °C (elevated temperature representing thermal stress).

Treatments included: (1) non-inoculated controls (larvae and algal feed only), (2) exposure to individual V. mediterranei strains, and (3) exposure to pathogenic controls Vh11 and OCN008. Larvae were incubated for 24h post-exposure before survival was assessed.

2.4.2 Protection assays and pre-colonization experimental design

To test whether Vm02 confers protection against pathogenic Vibrio, larvae were exposed to combinations of Vm02 with Vh11 or OCN008. Each strain was added to a final concentration of 10⁶ CFU mL^-1. Experimental treatments included: (1) non-inoculated control (algae only), (2) Vm02 only, (3) pathogen only, and (4) combined Vm02 + pathogen.

Pre-colonization experiments tested the dependence of protection on colonization timing. Larvae were first inoculated with Vm02 at 10⁶ CFU mL^-1, incubated at 28 °C, and then challenged with Vh11 at 10⁶ CFU mL^-1 after 0, 24, 48, or 96 h. Pre-colonization treatments were maintained at 28 °C with a single Vm02 and Vh11 inoculation per treatment. After pathogen addition, larvae were incubated an additional 24 hours before survival assessment. It should be noted that water changed were not performed after initial Vm02 inoculation. Thus, experimental tubes functionally represent a closed system where both waterborne and host-associated Vm02 may be present at pathogen challenge.

Co-inoculation evaluated the rapidity of protection: Vm02 and pathogen were added simultaneously at 10⁶ CFU mL^-1 each, and survival was measured after 24h.

All larval challenge experiments included at least two independent biological replicates (separate tubes) per treatment, and each experiment was repeated with larvae from two independent spawning cohorts. Experiments were conducted in temperature-controlled incubators set to the target temperature.

2.5 Larval survival assessment and relative larval survival

Larval survival was quantified at 24h after the final bacterial exposure using light microscopy. The full volume of each replicate tube was gently mixed and transferred into a sterile Petri dish. Larvae were scored as live or dead based on established criteria: live larvae exhibited active velar ciliary beating, coordinated swimming, and intact shell and tissue morphology; dead larvae displayed complete absence of movement, tissue disintegration, or empty shells. All counts were performed by the same observer to ensure consistency.

Relative larval survival (RLS) was calculated to normalize for background mortality in controls:

Where % survival = (number live larvae/total larvae counted) x 100. For protection experiments, RLS values were calculated relative to non-inoculated (larvae and algae only) controls. RLS >1.0 indicates survival enhanced above control levels; RLS = 1 indicates no effect; RLS<1.0 indicates mortality exceeding background.

2.6 Statistical analyses

All statistical analyses were performed in R version 4.3.1 using the tidyverse, rstatix, and ggpubr packages. Relative larval survival (RLS) was calculated for each replicate tube and served as the response variable for all statistical comparisons.

For baseline virulence comparisons, temperature effects on RLS were assessed independently for each bacterial strain using two-sample t-tests comparing 28 °C versus 32 °C treatments.

For protection experiments testing multiple conditions, differences between pathogen-only and Vm02-protected treatments were evaluated using two-sample t-tests performed independently within each pathogen x temperature combination. P-values were adjusted for multiple comparisons across the four pathogen-temperature conditions using the Benjamini-Hochberg false discovery rate (FDR) method.

For pre-colonization timing experiment, protection efficacy was compared between Vm02-protected and unprotected larvae using two-sample t-tests performed independently at each pre-colonization interval (0h, 24h, 96h).

For age-dependent susceptibility experiments, differences in larval survival across developmental stages were assessed separately for protection and evaluated whether survival differed across larval ages (7, 8, and 11 days post-fertilization). When significant (p<0.05), pairwise comparisons between age groups were conducted using Wilcoxon rank-sum tests with Benjamini-Hochberg FDR correction.

All experiments included two independent biological replicates (separate spawning cohorts) with two technical replicates (tubes) per treatment per cohort, for a total of four replicate tubes per treatment. Approximately 400 larvae were assessed per tube. Data are presented as mean ± standard error (SE) calculated across replicate tubes. Statistical significance was defined as α = 0.05 for all tests.

2.7 Fluorescence microscopy of larval colonization

To visualize bacterial colonization, larvae were inoculated with GFP-expressing Vm02 using the same inoculation schemes described for challenge experiments. At 24h and 96h post-inoculation, ~1 mL of larval suspension (~20 larvae) per replicate was fixed in 4% paraformaldehyde (PFA) in PBS and stored at 4 °C until imaging.

Immediately before imaging, fixed larvae were transferred onto glass microscope slides and immobilized under 2% (w/v) agarose pads prepared in sterile ASW adjusted to 20 psu. Coverslips were applied and slides were imaged immediately.

Fluorescence microscopy was performed using a Nikon Eclipse E600 upright microscope equipped with a Nikon DS-Fi3 camera and Sola Lumencor light engine. Images were acquired using a Nikon CFI Plan Fluor 40X objective with standardized GFP filter sets. Transmitted light and fluorescence images were captured sequentially for each field of view to visualize both overall morphology and bacterial colonization patterns. Exposure time, gain, and illumination settings were kept constant within each experiment. Image processing was performed using FIJI/ImageJ v1.53 (Schindelin et al., 2012). Brightness and contrast adjustments were applied uniformly across all images; no other manipulations were performed.

2.8 Genome sequencing and assembly

Whole-genome sequencing was performed on six V. mediterranei strains not previously characterized: pathogenic strains McD43–10 and McD53-9, and intermediate strains FYC25, FYC26, PPP23_1_1, and PPP23_1_2. All other V. mediterranei genomes used in comparative analyses were obtained from our previous study (Smith et al., 2025) or obtained from NCBI GenBank (accession numbers in Table 1).

Long-read sequencing was performed at SeqCoast Genomics on a PromethION 2 platform (Oxford Nanopore Technologies) (FYC26, PPP23_1_1, PPP23_1_2, McD53-9, McD43-10), or at the University of Rhode Island on a MinION Mk1B (Oxford Nanopore Technologies) (FYC25). Raw reads were trimmed using PoreChop and assembled using Unicycler with the nf-core/bacass v2.5.0 workflow (Daniel et al., 2025). Assembly quality metrics were assessed with QUAST. Draft genomes were deposited under BioProject accession PRJNA1379678, with individual assembly accession identifiers listed in Table 1.

2.9 Genome annotation, ANI, and phylogenomic analysis

Draft genomes, including reference assemblies retrieved from NCBI GenBank, were annotated using Prokka v1.14.6 (Seemann, 2014) with consistent default parameters to ensure uniform gene prediction across the dataset.

Pangenome analysis and core genome alignment were performed using Panaroo v1.3.4 (Tonkin-Hill et al., 2020) with ‘—clean-mode strict’ to minimize spurious gene calls from assembly errors or potential contamination. The core genome was defined as genes present in ≥99% of all input genomes, yielding 3, 684 core genes. Panaroo generated a multi-FASTA alignment of concatenated core genes that was used for phylogenomic reconstruction.

Maximum likelihood phylogenetic trees were constructed using IQ-TREE v2.2.0 (Minh et al., 2020). The best-fit substitution model was automatically selected using ModelFinder Plus (Kalyaanamoorthy et al., 2017) with Bayesian Information Criterion, identifying GTR+F+I+G4 as the optimal model. Phylogenetic inference was performed with 1, 000 bootstrap replicates and 1, 000 SH-aLRT tests to assess branch support. Trees were visualized with TreeViewer (Bianchini and Sánchez‐Baracaldo, 2024).

Pairwise average nucleotide identity (ANI) was calculated using pyANI v0.2.12 (Pritchard et al., 2016) with the ANIm method, which uses MUMmer (Marçais et al., 2018) for whole-genome alignment. Default parameters were used (minimum alignment length 1, 000 bp, minimum alignment identity 30%, minimum alignment coverage 50%). Pairwise ANIm values were organized into a similarity matrix and visualized as a heatmap using pheatmap v1.0.12 in R with hierarchical clustering based on complete linkage. Within-clade and between-clade ANI distributions were summarized to quantify genomic divergence corresponding to phenotypic classifications.

2.10 Pangenome and functional analyses

Orthogroup classifications from Panaroo were used to characterize the pangenome structure of V. mediterranei. Clade-specific orthogroups were identified using custom Python scripts (Python v3.9.7) that parsed the Panaroo gene presence-absence matrix. Analysis was restricted to V. mediterranei strains specifically ascribed a host phenotype, either experimentally in this study or in the literature. Protective clade-specific orthogroups were defined as present in 90% of protective clade strains (n=6) and absent from 90% of pathogenic (n=7), intermediate high-virulence (n=4), and intermediate low-virulence (n=6) strains. Pathogenic-specific, intermediate high-virulence-specific, and intermediate low-virulence-specific orthogroups were defined analogously. This criterion (90% presence within-clade, 10% presence between-clades) ensured identification of genes consistently associated with each phenotypic group.

Functional annotation of all clade-specific protein sequences was performed using eggNOG-mapper v2.19 (Cantalapiedra et al., 2021) against the eggNOG v5.0 database with default parameters. This analysis assigned Clusters of Orthologous Groups (COG) functional categories, KEGG Orthology (KO identifiers), Gene Ontology (GO) terms, and PFAM domains to each protein. Orthogroups were assigned functional annotations based on their representative sequences.

Phylogenetic clade designations were assigned based on experimental phenotypic validation. The protective clade includes strains that demonstrated avirulence or enhanced survival (RLS≥1.0) when tested experimentally against bivalve larvae in this study or previous reports. The pathogenic clade includes only strains experimentally confirmed as pathogens in marine invertebrates. Strains from the same broader phylogenetic group but isolated from healthy hosts (A203 from healthy fireworm, CyArs1 from healthy sponge, Mu-109 from mucilage) are excluded from the pathogenic clade designation. Strain 00-90-15, which enhanced oyster larval survival experimentally (Kehlet-Delgado et al., 2020) but falls phylogenetically between clades, was not assigned to any specific group.

3 Results

3.1 V. mediterranei strains exhibit source-dependent pathogenicity phenotypes

Given the contradictory reports of V. mediterranei as both pathogen and commensal across marine systems, we first asked pathogenicity to oyster larvae virulence varies among strains. We screened eleven V. mediterranei isolates spanning the known diversity of this species complex for pathogenicity against eastern oyster larvae in single-strain challenge assays (Table 1). All strains were tested clonally at 32 °C to account for potential temperature-mediated virulence previously reported in V. mediterranei interactions with other marine hosts. These assays revealed three distinct phenotypic categories that corresponded closely with strain isolation source (Figure 1).

Six strains isolated from healthy adult oysters during low mortality periods demonstrated complete avirulence in oyster larvae (RLS = 1.03-1.05), including strain Vm02 (PNB23_20_7) and five additional isolates (PNB22_7_2, PNB22_2_1, PNB22_3_3, PNB22_6_1, PNB22_9_1). The consistency of this avirulent phenotype across multiple independent isolates from healthy oysters, collected across different sampling periods and locations, indicated that avirulence to C. virginica larvae is a reproducible characteristic of oyster-associated V. mediterranei populations.

Two strains isolated from diseased coral tissues, McD53–9 and McD43-10, proved highly pathogenic to oyster larvae, causing 42% and 30% mortality respectively (RLS = 0.58 ± 0.03 and 0.70 ± 0.05). These strains were originally isolated from Montastraea cavernosa showing signs of tissue loss disease in the Florida Keys.

Three strains displayed intermediate virulence with slight but statistically significant larval mortality compared to non-inoculated controls. These included PPP23_1_2 (RLS = 0.96 ± 0.02), PPP23_1_1 (RLS = 0.98 ± 0.01), and FYC26 (RLS = 0.95 ± 0.03). PPP23_1_1 and PPP23_1_2 were isolated from healthy adult oysters during the same sampling period as Vm02, although at a different farm, and FYC26 was isolated from larvae from a hatchery experiencing larval mortality. While these strains caused minimal mortality, we categorize them here as “intermediate low-virulence” strains.

The presence of avirulent, intermediate, and pathogenic phenotypes raised the question of whether avirulent strains might actively influence larval health. We selected Vm02 as a representative avirulent strain for detailed characterization because it was isolated from a healthy oyster in the weeks preceding a mortality event where we previously documented systematic V. mediterranei depletion during mortality progression (Smith et al., 2025).

3.2 Vm02 is avirulent to oyster larvae while pathogen virulence is temperature-enhanced

To further examine V. mediterranei Vm02, we next characterized its pathogenicity across temperature conditions alongside known Vibrio pathogens. Larvae were challenged with Vm02 or pathogenic control strains at 10⁶ CFU mL^-1 at both ambient (28 °C) and elevated (32 °C) temperatures (Figure 2). Vm02 demonstrated complete avirulence across both temperature conditions. At ambient hatchery temperature (28 °C), Vm02-exposed larvae maintained relative larval survival (RLS) of 1.03 ± 0.02 compared to non-inoculated controls, indicating not only absence of pathogenicity but a modest enhancement of survival. This avirulent phenotype remained stable at elevated temperature (32 °C), with RLS of 1.03 ± 0.01, demonstrating that elevated temperature does not induce virulence in this protective strain.

Figure 2

In contrast, the pathogenic control strains exhibited severe, temperature-enhanced lethality to larvae. V. harveyi Vh11, isolated from an oyster several weeks after Vm02 during a mass mortality event, caused severe mortality at both temperatures tested (RLS = 0.19 ± 0.04 at 28 °C; 0.10 ± 0.02 at 32 °C; p<0.01), representing 81% mortality above background levels. Elevated temperature (32 °C) exacerbated pathogenicity, with RLS declining to 0.10 ± 0.02 (90% excess mortality). Similarly, V. coralliilyticus strain OCN008, a known coral pathogen in the Pacific and previously shown to cause mortality in Pacific oyster larvae (Ushijima et al., 2018), proved highly pathogenic to eastern oyster larvae. OCN008 caused comparable mortality to Vh11, with RLS of 0.16 ± 0.03 at 28 °C and 0.09 ± 0.02 at 32 °C. The temperature-independent avirulence of Vm02 contrasted sharply with the temperature-enhanced pathogenicity of both species. This divergent response to temperature led us to ask whether Vm02 might not only be commensal but actively protective against pathogen challenge.

3.3 V. mediterranei Vm02 pre-colonization protects against pathogens

To test whether Vm02 could protect larvae against pathogen challenge, we used a sequential inoculation design in which larvae were pre-colonized with a single dose of Vm02–24 hours before challenge with either V. harveyi Vh11 or V. coralliilyticus OCN008 (Figure 3A). V. mediterranei pre-colonization dramatically altered infection outcomes at both temperature conditions (Figure 3B). At 28 °C, larvae pre-colonized with Vm02 and subsequently challenged with Vh11 maintained 0.97 ± 0.02 RLS, compared to only 0.19 ± 0.04 survival in larvae exposed to Vh11 alone (p<0.0001, two-sample t-test, FDR-adjusted). This represents an 82% rescue effect, effectively restoring larval survival to control levels. Protection remained robust at 32 °C, with pre-colonized larvae achieving 0.95 ± 0.02 RLS versus 0.10 ± 0.02 in unprotected controls (p<0.0001, two-sample t-test, FDR-adjusted), demonstrating that protective efficacy is maintained despite thermal stress conditions that enhance pathogen lethality.

Figure 3

Protection against V. coralliilyticus OCN008 proved equally effective. Pre-colonization with Vm02 increased survival from 0.16 ± 0.03 to 0.97 ± 0.02 at 28 °C (p<0.0001, two-sample t-test, FDR-adjusted) and from 0.09 ± 0.02 to 0.94 ± 0.03 at 32 °C (p<0.0001, two-sample t-test, FDR-adjusted). The consistency of protection across two phylogenetically distinct pathogens demonstrates that Vm02-mediated protection is not limited to a single pathogen species.

To test the effect of colonization timing and stability on protection, we varied the time interval between Vm02 pre-colonization dose and Vh11 pathogen challenge. Each interval represents an independent treatment, with a single dose of Vm02 and a single subsequent dose of Vh11. Protection proved remarkably durable across all time intervals tested (Figure 3C). Even with co-inoculation (0h) when larvae were exposed to protector and pathogen simultaneously, Vm02 provided substantial protection, increasing RLS from 0.33 ± 0.05 in Vh11-only controls to 0.93 ± 0.03 in co-inoculated larvae (p<0.0001, two-sample t-test). Standard 24-hour pre-colonization yielded near-complete protection (RLS = 0.97 ± 0.02, p<0.0001), while extended pre-colonization periods of 96 hours maintained similarly high protection levels (RLS = 0.95 ± 0.03, p<0.0001). The rapid establishment and stability of protection suggested persistent colonization rather than transient effects. All time points assessed (up to 96 hours post-inoculation, corresponding to 11 dpf) occurred prior to larval metamorphosis.

Vm02 protection remained effective across early larval developmental stages, even as unprotected larvae became progressively more susceptible to Vh11 pathogen challenge with age (Figure 3D). We tested larvae at different ages pre-competence following a standard pre-colonization with Vm02. At 7 days post-fertilization (dpf), unprotected larvae maintained 0.35 ± 0.05 RLS when challenged with Vh11. This susceptibility increased significantly by 8 dpf (0.28 ± 0.08 RLS) and reached maximum vulnerability at 11 dpf (0.07 ± 0.02 RLS; Kruskal-Wallis χ²=6.8, p=0.033; pairwise Wilcoxon test with FDR correction: 7 vs 11 dpf=0.038). In contrast, Vm02-protected larvae maintained consistently high survival (>0.90 RLS) regardless of developmental stage (Kruskal-Wallis χ²=0.5, p=0.78), showing that protection remained robust across early larval stages.

3.4 V. mediterranei Vm02 persistently associates with larval tissues

The durability of protection across extended pre-colonization intervals suggested that Vm02 establishes a persistent physical association with larval tissues. To test this, we transformed Vm02 with a stable GFP-expressing plasmid (pVSV102) and used fluorescence microscopy to visualize colonization dynamics. GFP-expressing Vm02 cells were consistently observed in association with the larval digestive system, with particularly dense colonization of the stomach and digestive diverticula (Figure 4). Colonization also extended to the velum, the primary feeding and swimming organ of larvae. In contrast, larvae exposed to V. harveyi Vh11 showed rapid tissue degradation, damaged cilia, and mortality within 24 hours.

Figure 4

The colonization pattern established by 24 hours persisted through at least 96 hours post-inoculation, with GFP-expressing Vm02 cells maintaining their association with digestive tissues. This persistent colonization corresponds with the sustained protection observed in our challenge experiments, indicating a durable physical association with larval digestive tissues.

3.5 V. mediterranei comprises three phylogenetically distinct lineages

The consistent association between virulence phenotype and isolation source across our strain panel suggested that these phenotypic differences might reflect underlying evolutionary divergence (Figure 5). To test this, we constructed a maximum likelihood phylogeny using 1, 247 single-copy core genes from 33 V. mediterranei genomes, including our 11 experimental strains and 22 reference genomes from diverse marine sources. This analysis also included seven closely-related V. barjaei strains and two V. thalassae strains for phylogenetic context (Figure 5; Table 1). The phylogeny revealed three well-supported monophyletic clades that corresponded precisely with the experimentally determined phenotypes.

Figure 5

All six avirulent strains formed a tight monophyletic clade with 100% bootstrap support; this clade includes all isolates that were non-pathogenic in single-strain assays as well as Vm02 which conferred robust larval protection in challenge experiments. Given the validated protective capacity of Vm02 and above-control RLS of the additional avirulent strains, we designate this the “protective clade”. This clade includes reference strain V7A from healthy Argopecten purpuratus (scallop) digestive tissue, which was previously reported to produce bacteriocin-like inhibitory substances and proposed as a potential probiotic (Serrano et al., 2018). Additional closely-related reference strains were isolated from seawater and sediment samples. While environmental reservoirs likely exist for all V. mediterranei lineages, the representation of isolates from the protective clade in seawater and sediment samples in particular suggests they persist outside host associations. The protective clade showed unexpectedly low internal genetic diversity, with all 15 strains from various sources sharing >99% average nucleotide identity (ANIm), consistent with recent shared ancestry or strong purifying selection maintaining beneficial traits.

The two pathogenic experimental strains clustered with reference strains isolated from diseased marine invertebrates, forming a “pathogenic clade” with 100% bootstrap support. This clade included V. mediterranei strains AK1 and Vic-OC-097, well-characterized coral pathogens, strain 117-T6 from diseased red algae (Pyropia yezoensis), and strains from moribund pen shells (Pinna nobilis). Within-clade ANIm values ranged from 98.5-99.9%, indicating somewhat greater genetic diversity than the protective clade while maintaining clear genomic cohesion.

An “intermediate clade” bridged the protective and pathogenic lineages, containing strains with variable virulence phenotypes from bivalve aquaculture systems and marine environments. This clade included our three intermediate-virulence experimental strains along with reference genomes from both healthy and diseased bivalves. The phylogenetic position of this clade, sister to the protective clade and divergent from the pathogenic lineage, indicates an intermediate genomic grouping consistent with their variable phenotypes.

Genomic divergence among clades approaches species-level thresholds while within-clade similarity remains high (Supplementary Figure 1). ANIm analysis revealed consistently high within-clade values (Protective: 99-100%; Intermediate: 97.6-100%; Pathogenic: 98.2-100%), indicating recent shared ancestry or strong purifying selection. Between-clade comparisons revealed substantial genomic divergence, with Pathogenic-Intermediate strains showing 96.5-97.6% ANIm, Pathogenic-Protective showing 97.1-97.8% ANIm, and Intermediate-Protective showing 97.5-98.2% ANIm. These between-clade values approach the 95-96% ANI threshold typically used for species delineation (Richter and Rosselló-Móra, 2009), highlighting substantial divergence within the species complex that is supported by their differing host-interaction phenotypes.

3.6 Protective and pathogenic lineages are distinguished by genomic specialization

Having established that virulence phenotypes correspond to monophyletic clades, we next asked whether these lineages differ in gene content. Pangenome analysis revealed striking asymmetry in specialized gene content between clades. Protective strains harbored 230 unique orthogroups (genes present in 100% of protective strains and absent in all others), compared to 80 unique orthogroups in pathogenic strains and 27 in intermediate high-virulence strains (Supplementary Figure 2A; Supplementary Table 1). Intermediate low-virulence strains notably lacked any clade-specific genes.

COG functional analysis of protective-specific genes identified enrichment in several categories: regulatory systems (21 transcriptional regulators including two-component systems EnvZ/OmpR), stress response mechanisms (27 DNA repair genes, nitric oxide dioxygenase), metabolic functions (glyoxylate carboligase, menaquinone biosynthesis pathway genes), and cell surface modification (19 cell wall biogenesis genes including multiple glycosyltransferases) (Supplementary Figure 2B). Additional protective-specific elements included eight phage integrases and resolvases, Type IV secretion system components, and multiple toxin-antitoxin systems.

Among the protective-specific orthogroups, we identified an 88 amino acid predicted bacteriocin located within a putative biosynthetic operon containing genes encoding a peptide biosynthesis protein (YydG-like), acetyltransferase, and secretion machinery. This protein shares 45.6% identity with UniProt A0A240EP04, an annotated class II bacteriocin with double-glycine leader peptide in Vibrio thalassae.

Additional protective-specific orthogroups included genes involved in polysaccharide degradation (ß-1, 3 glucan and glycogen metabolism), peptide transport systems, cell surface modification, biofilm formation, and pilus assembly.

The 80 pathogenic-specific orthogroups showed relatively limited functional diversity, distributed among secondary metabolism, transcriptional regulation, and amino acid/organic acid metabolism. The pathogenic clade uniquely encoded a complete Type I secretion system (T1SS) consisting of an ATP-binding protein (PrsD), membrane fusion protein (PrsE), and a large repetitive protein containing 44 hemolysin-type calcium-binding domains (1, 480 amino acids). Additional pathogenic-specific functions included multidrug efflux pumps and specialized regulatory proteins.

Within the intermediate clade, virulence phenotypes correlated with the presence of specific genetic elements. All V. mediterranei strains regardless of clade were found to encode a type VI secretion system on chromosome I (T6SS1). However, intermediate high-virulence strains uniquely possessed an additional T6SS on chromosome II (T6SS2), which was absent from both protective and pathogenic strains. Comparative analysis revealed that T6SS1 and T6SS2 are indeed structurally distinct systems, differing in both gene organization and amino acid identity of homologous core proteins (Supplementary Figure 3).

Screening for the recently described V. mediterranei TCP pathogenicity island (TCP-PAI; (H. Zhang et al., 2025)) revealed this genomic region was absent from all protective strains. In contrast, TCP-PAI was variably present across intermediate and pathogenic clades, with considerable variation in gene content and completeness among strains where present. In protective strain Vm02, the genomic regions flanking the TCP-PAI insertion site are directly adjacent, confirming absence of the intervening pathogenicity island (Supplementary Figure 4).

4 Discussion

Our findings establish that V. mediterranei comprises phylogenetically distinct lineages with fundamentally different host-interaction strategies, and we organize this discussion around four themes: the protective phenotype demonstrated by Vm02 and its phylogenetic basis, the genomic architecture underlying protective versus pathogenic strategies, implications for aquaculture under climate change, and broader considerations for Vibrio taxonomy and disease management.

4.1 Phylogenetically constrained protective versus pathogenic phenotypes in V. mediterranei

This study addresses the longstanding paradox of V. mediterranei’s contradictory roles across marine systems by demonstrating that beneficial and pathogenic phenotypes are phylogenetically constrained within distinct evolutionary lineages. The protective clade, comprising strains exclusively from healthy bivalves and environmental sources, forms a cohesive monophyletic group that consistently rescued oyster larvae from near-complete pathogen-induced mortality. Pathogenic strains formed a separate lineage with between-clade divergence approaching commonly cited species delineation thresholds (Richter and Rosselló-Móra, 2009) (Supplementary Figure 1).

This phylogenetic structure reconciles the contradictory literature on V. mediterranei ecology: all previously described pathogenic strains cluster within the pathogenic or intermediate clades, while commensal and beneficial isolates – including the bacteriocin-producing V7A isolate from a scallop (Serrano et al., 2018) – cluster within the protective clade.

More broadly, strain-specific virulence is a recurring pattern across Vibrios. In V. parahaemolyticus, clinical strains often harbor pathogenicity islands carrying the tdh and/or trh genes together with an accessory type III secretion system (T3SS2) that mediates enterotoxicity (Matsuda et al., 2020). In contrast, environmental strains lacking these islands are typically avirulent despite sharing >99% ANI with pathogenic lineages (Ceccarelli et al., 2013). The pandemic O3:K6 clone was further distinguished by acquisition of a T6SS and additional genomic islands (Boyd et al., 2008). In V. alginolyticus, population genomics have identified constrained lineages encoding T3SS and other virulence-associated genes, with pathogenic capacity restricted to specific lineage-host combinations (Huang et al., 2024; Darshanee Ruwandeepika et al., 2012). Similarly, in V. fischeri, different T6SS clusters mediate niche partitioning and eukaryotic antagonism among closely related strains (Speare et al., 2018; Gaddy et al., 2025). The parallels to our findings in V. mediterranei are striking: just as pandemic V. parahaemolyticus lineages are defined by specific genomic islands absent from environmental strains, our pathogenic and intermediate V. mediterranei clades are distinguished by T1SS, T6SS2, and the TCP-PAI, while protective-clade strains entirely lack these offensive modules.

4.2 Genomic architecture of protection: specialized defenses

The marked asymmetry in clade-specific gene content (Supplementary Figure 2) reveals fundamentally different evolutionary strategies, with protective lineages maintaining substantially greater genomic complexity than pathogenic strains. This pattern suggests that stable beneficial associations require more specialized molecular machinery than opportunistic pathogenicity, paralleling other mutualistic systems where symbionts maintain expanded metabolic and regulatory repertoires to persist within host environments (Schwartzman and Ruby, 2016; Soto and Nishiguchi, 2021).

The co-inoculation result, in which Vm02 provided substantial protection even without prior host colonization (RLS = 0.93 at the 0h interval), has important mechanistic implications. This demonstrates that waterborne interactions between Vm02 and the pathogen are sufficient to confer strong protection independent of prior host association. The modest additional benefit of Vm02 pre-colonization (RLS = 0.97) suggests that tissue-associated mechanisms may provide a supplementary but not essential contribution under these experimental conditions.

Within this context, we consider four candidate mechanisms with differing levels of support. (i) Direct antimicrobial activity: the protection observed under co-inoculation is consistent with Vm02 producing diffusible compounds that inhibit pathogen growth in the water column, and the conserved bacteriocin biosynthesis operon identified across all protective-clade genomes provides a strong candidate molecular basis. (ii) Competitive exclusion: the protective clade’s genomic enrichment in nutrient acquisition systems is consistent with resource competition that could limit pathogen proliferation in the shared larval rearing environment. (iii) Niche preemption through tissue association: fluorescence microscopy confirmed persistent Vm02 association with larval digestive tissues, and the protective clade’s unique cell surface modification genes are consistent with specialized colonization capacity. However, strong protection under co-inoculation indicates that tissue preemption is not required, though it may enhance durability. (iv) Enhanced stress tolerance: the protective clade’s enrichment in stress-response genes may enable persistence under conditions of host immune activity. The maintenance of protection at 32 °C is consistent with this interpretation, though it may also reflect the simple absence of temperature-inducible virulence-associated genes in this clade.

Our experimental and genomic evidence most strongly supports waterborne mechanisms, where direct antimicrobial activity and/or competitive exclusion act as primary drivers of protection. The genomic architecture is consistent with protection operating through multiple complementary pathways rather than a single mechanism, aligning with emerging models of microbiome-mediated defense (Buffie and Pamer, 2013; Spragge et al., 2023). Resolving the relative contributions of these mechanisms requires targeted experiments that will include water-exchange protocols to isolate tissue-associated from waterborne protection, knockout mutations in putative protective pathways, and quantitative tracking of pathogen loads in both the water column and host tissues during co-challenge.

In contrast, pathogenic lineages encode broad-spectrum offensive systems characteristically associated with anti-eukaryotic activity and tissue invasion. While our experimental validation of pathogenicity focused on C. virginica larvae, the genomic architecture of these offensive systems is consistent with broader anti-eukaryotic activity. The pathogenic clade harbors a complete Type I secretion system (T1SS) with associated hemolysin-type calcium-binding domains. The presence of RTX-family toxin genes in this clade is consistent with the rtx virulence determinants identified in V. mediterranei strains pathogenic to Pinna nobilis (Andree et al., 2021) and aligns with the broader role of T1SS-exported hemolysins and cytotoxins in Vibrio virulence. The T6SS and toxin genes upregulated during V. mediterranei co-infection of corals (Rubio-Portillo et al., 2020) further implicate these secretion systems in pathogenic interactions, and our finding that T6SS2 is variably present across intermediate high-virulence strains is consistent with the role of this system in mediating anti-eukaryotic activity in the closely related V. coralliilyticus (Mass et al., 2024). The variable distribution of the TCP pathogenicity island (TCP-PAI; Zhang et al., 2025) across pathogenic and intermediate clades – but absence in Vm02 (Supplementary Figure 4) – provides an additional genomic marker distinguishing these lineages and suggests that horizontally acquired pathogenicity islands contribute to the virulence repertoire of non-protective V. mediterranei.

The strain-level variation in virulence factor composition we observe across clades parallels recent proteomics work in bivalve-pathogenic V. mediterranei, where high-virulence strains secrete a greater diversity of virulence-associated extracellular products compared to low-virulence strains (Fan et al., 2024). Together, these observations suggest that intraspecific variation in V. mediterranei is driven by the presence or absence of discrete virulence modules (T1SS, T6SS2, TCP-PAI) rather than by gradual genome-wide divergence alone.

The protective clade’s complete absence of T1SS, T6SS2, and TCP-PAI, combined with unique metabolic specialization and stress-tolerance pathways, is consistent with a model in which protective strains favor long-term niche occupation over host damage (Figure 6).

Figure 6

4.3 Protection under thermal stress and climate change implications

The maintenance of protective efficacy at 32 °C, despite increased virulence of both V. harveyi and V. coralliilyticus at elevated temperatures, has critical implications for aquaculture under warming ocean conditions. The durability of this protection across temperatures and early developmental stages supports the practical viability of protective V. mediterranei as a biological control agent in shellfish hatcheries.

The persistent association of Vm02 with larval digestive tissues, maintained even under co-exposure with pathogens, is compatible with practical hatchery probiotic applications, where strain instability and inconsistent host-association frequently limit efficacy (Kesarcodi-Watson et al., 2008).

4.4 Taxonomic implications and evolutionary dynamics

The discovery that V. mediterranei comprises deeply divergent protective and pathogenic lineages with mutually exclusive ecological strategies raises important considerations for Vibrio taxonomy. When V. shiloi was first described as a coral pathogen, it was explicitly distinguished from V. mediterranei on the basis of pathogenicity and cellular characteristics (Kushmaro et al., 2001). Our phylogeny recapitulates these distinctions: pathogenic isolates cluster with the original V. shiloi strains AK1 and Vic-OC-097, whereas protective and environmentally associated isolates group with the classical V. mediterranei lineage. The subsequent synonymization of these taxa (Thompson et al., 2001) may therefore have consolidated biologically heterogeneous lineages under a single name.

Regardless of formal nomenclature, our results underscore that species-level identification alone is insufficient for predicting ecological behavior within V. mediterranei. The clade-specific genomic markers identified here (e.g., bacteriocin operons in protective strains or T1SS/T6SS2 in pathogenic strains) provide practical tools for distinguishing beneficial from harmful populations in aquaculture settings and for interpreting historical reports of “V. mediterranei” in the literature.

The near-clonal structure of the protective clade across years and sources (>99% ANIm) contrasts with greater genomic diversity in the pathogenic lineage (98.5-99.9% ANIm), potentially reflecting broader ecological opportunity or adaptation to diverse hosts. The intermediate clade’s variable virulence, defined by presence or absence of T6SS2, highlights how mobile genomic elements may enable rapid shifts between commensal and pathogenic lifestyles (Bruto et al., 2018; Lemire et al., 2015).

The association between clade membership and isolation source (Table 1) raises the question of whether host identity or environmental context may drive lineage sorting in V. mediterranei. Protective-clade strains were isolated exclusively from bivalve hosts (Crassostrea virginica, Magallana gigas, Argopecten purpuratus) and environmental reservoirs (seawater, sediment), while pathogenic-clade strains derive predominantly from coral hosts (Oculina patagonica, Montastraea cavernosa) or diseased macroalgae. The intermediate clade spans both bivalve and environmental sources, consistent with its genomic and phenotypic variability. This host-associated distribution suggests that the selective pressures encountered in bivalve versus coral tissue environments may favor distinct ecological strategies.

Temperature may represent an additional selective pressure reinforcing this divergence. Our data show that protective-clade strains remain avirulent at both 28 °C and 32 °C, while pathogenic strains exhibit enhanced virulence at elevated temperatures. This pattern is consistent with temperature-dependent virulence reported in V. mediterranei infections of Pinna nobilis, where virulence onset occurs above 24 °C (Andree et al., 2021), and in coral co-infection systems where V. mediterranei and V. coralliilyticus show temperature-dependent transcriptional shifts in virulence gene expression (Rubio-Portillo et al., 2020). Whether avirulence in the protective clade at elevated temperatures reflects active suppression of virulence pathways or simply the absence of temperature-responsive virulence genes (e.g. absent T1SS, T6SS2, and TCP-PAI) remains an open question. The genomic architecture of each clade is consistent with either scenario: pathogenic strains possess virulence machinery that can be temperature-activated, while protective strains lack these systems entirely.

These findings suggest new monitoring approaches that extend beyond traditional pathogen surveillance. Our recent field study showed systematic depletion of protective V. mediterranei preceding mortality events (Smith et al., 2025), indicating that protective taxa may function as early-warning biomarkers of community disruption. Clade-specific dPCR assays targeting the protective lineage’s unique genes could provide high-resolution tracking of beneficial Vibrio populations, while assays detecting T1SS or T6SS2 may allow hatcheries to identify emerging threats before clinical disease becomes apparent.

4.5 Limitations and future directions

Several limitations should be considered. Our larval model provides proof-of-concept but may not fully represent interactions in adult oysters or in complex field environments. Additionally, our experiments assess protection only through the pre-metamorphosis larval stage (up to 11 days post-fertilization, 96 hours post-bacterial exposure). Whether protective V. mediterranei can persist through metamorphosis and continue providing benefits in juvenile or adult oysters remains unknown but is critical for practical probiotic deployment.

Community-level processes, such as polymicrobial interactions and host genotype-by-microbe effects, remain to be explored. While we cannot exclude geographic sampling bias, strains from each clade derive from diverse geographic locations (protective strains from North Carolina, Australia, Chile, and the Mediterranean; pathogenic strains from Florida and the Mediterranean), suggesting that phenotype rather than geography drives phylogenetic clustering. Notably, colonization capacity was directly demonstrated for strain Vm02. The high genomic similarity within the protective clade (>99% ANIm) raises the hypothesis that colonization may be a shared trait of this lineage, but this requires experimental validation across additional strains. While genomic analyses suggest candidate protection mechanisms involving metabolic competition, antimicrobial activity, and enhanced stress tolerance, direct functional validation of these mechanisms is still needed. Finally, the evolutionary stability of beneficial phenotypes requires assessment. The protective clade’s monophyletic structure, extensive genomic investment in specialized functions, and lack of T1SS and T6SS2 suggest stable mutualistic traits. However, the intermediate clade demonstrates that horizontally acquired elements can rapidly alter ecological strategy. Long-term monitoring of protective strain genomic stability, particularly under different selection pressures, will be important for ensuring reliable probiotic performance.

5 Conclusions

This study redefines V. mediterranei as a species complex comprising phylogenetically distinct lineages with fundamentally different ecological strategies. Beneficial and pathogenic phenotypes correspond to monophyletic clades distinguished by the presence or absence of discrete virulence systems rather than by gradual genome-wide divergence. The identification of a protective clade whose representative strain Vm02 confers robust pathogen protection under thermal stress, even without prior tissue colonization, challenges pathogen-centric approaches to aquaculture disease management and underscores the importance of preserving beneficial microbiome members. Protective strains are consistently isolated from bivalves and environmental sources, whereas pathogenic strains are isolated predominantly from corals, suggesting that host-associated selective pressures contribute to maintaining these divergent strategies. As marine ecosystems face increasing thermal and disease stress, recognizing and maintaining protective symbionts like V. mediterranei will be as critical as controlling pathogens. By combining field observations, experimental validation, and comparative genomics, we present a framework for leveraging beneficial host-microbe relationships in sustainable shellfish aquaculture.

References

• 1

  AndreeK. B.CarrascoN.CarellaF.FuronesD.PradoP. (2021). Vibrio mediterranei, a potential emerging pathogen of marine fauna: investigation of pathogenicity using a bacterial challenge in Pinna nobilis and development of a species‐specific PCR. J Appl Microbiol130, 617–631. doi: 10.1111/jam.14756

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  BaninE.KhareS. K.NaiderF.RosenbergE. (2001). Proline-Rich Peptide from the Coral Pathogen Vibrio shiloi That Inhibits Photosynthesis of Zooxanthellae. Appl Environ Microbiol67, 1536–1541. doi: 10.1128/AEM.67.4.1536-1541.2001

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  BaninE.VassilakosD.OrrE.MartinezR. J.RosenbergE. (2003). Superoxide Dismutase Is a Virulence Factor Produced by the Coral Bleaching Pathogen Vibrio shiloi. Current Microbiology46, 418–422. doi: 10.1007/s00284-002-3912-5

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  Ben-HorinT.CiesielskiM.LucasJ.NobleR. T.WilburA. (2024). Pathology associated with summer oyster mortality in North Carolina. Aquaculture Reports34, 101901. doi: 10.1016/j.aqrep.2023.101901

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 5

  BianchiniG.Sánchez‐BaracaldoP. (2024). TreeViewer : Flexible, modular software to visualise and manipulate phylogenetic trees. Ecology and Evolution14, e10873. doi: 10.1002/ece3.10873

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  BoaventuraJ.CaetanoT.MendoS.ClearyD. F. R.GomesN. C. M.OliveiraV. (2022). Draft Genome Sequence of Vibrio mediterranei Strain CyArs1. Microbiol Resour Announc11, e00155-22. doi: 10.1128/mra.00155-22

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  BoydE. F.CohenA.NaughtonL. M.UsseryD. W.BinnewiesT. T.StineO. C.et al. (2008). Molecular analysis of the emergence of pandemic Vibrio parahaemolyticus. BMC Microbiol8, 110. doi: 10.1186/1471-2180-8-110

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  BrutoM.JamesA.PettonB.LabreucheY.ChenivesseS.Alunno-BrusciaM.et al. (2017). Vibrio crassostreae , a benign oyster colonizer turned into a pathogen after plasmid acquisition. The ISME Journal11, 1043–1052. doi: 10.1038/ismej.2016.162

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  BrutoM.LabreucheY.JamesA.PielD.ChenivesseS.PettonB.et al. (2018). Ancestral gene acquisition as the key to virulence potential in environmental Vibrio populations. The ISME Journal12, 2954–2966. doi: 10.1038/s41396-018-0245-3

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  BuffieC. G.PamerE. G. (2013). Microbiota-mediated colonization resistance against intestinal pathogens. Nat Rev Immunol13, 790–801. doi: 10.1038/nri3535

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 11

  CantalapiedraC. P.Hernández-PlazaA.LetunicI.BorkP.Huerta-CepasJ. (2021). eggNOG-mapper v2: Functional Annotation, Orthology Assignments, and Domain Prediction at the Metagenomic Scale. Molecular Biology and Evolution38, 5825–5829. doi: 10.1093/molbev/msab293

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 12

  CarraturoA.RaietaK.OttavianiD.RussoG. L. (2006). Inhibition of Vibrio parahaemolyticus by a bacteriocin-like inhibitory substance (BLIS) produced by Vibrio mediterranei 1. J Appl Microbiol101, 234–241. doi: 10.1111/j.1365-2672.2006.02909.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 13

  CeccarelliD.HasanN. A.HuqA.ColwellR. R. (2013). Distribution and dynamics of epidemic and pandemic Vibrio parahaemolyticus virulence factors. Front. Cell. Infect. Microbiol.3, 97. doi: 10.3389/fcimb.2013.00097

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  ClerissiC.De LorgerilJ.PettonB.LucassonA.EscoubasJ.-M.GueguenY.et al. (2020). Microbiota Composition and Evenness Predict Survival Rate of Oysters Confronted to Pacific Oyster Mortality Syndrome. Front. Microbiol.11, 311. doi: 10.3389/fmicb.2020.00311

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  DanielV. M.PeltzerAStraubDnf-core bot.PantanoLYiHDet al. (2025). nf-core/bacass: nf-core/bacass v2.5.0. Zenodo. doi: 10.5281/ZENODO.2669428

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  Darshanee RuwandeepikaH. A.Sanjeewa Prasad JayaweeraT.Paban BhowmickP.KarunasagarI.BossierP.DefoirdtT. (2012). Pathogenesis, virulence factors and virulence regulation of vibrios belonging to the Harveyi clade. Reviews in Aquaculture4, 59–74. doi: 10.1111/j.1753-5131.2012.01061.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  De LorgerilJ.LucassonA.PettonB.ToulzaE.MontagnaniC.ClerissiC.et al. (2018). Immune-suppression by OsHV-1 viral infection causes fatal bacteraemia in Pacific oysters. Nat Commun9, 4215. doi: 10.1038/s41467-018-06659-3

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 18

  DunnA. K.MillikanD. S.AdinD. M.BoseJ. L.StabbE. V. (2006). New rfp - and pES213-Derived Tools for Analyzing Symbiotic Vibrio fischeri Reveal Patterns of Infection and lux Expression In Situ. Appl Environ Microbiol72, 802–810. doi: 10.1128/AEM.72.1.802-810.2006

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 19

  EganS.GardinerM. (2016). Microbial Dysbiosis: Rethinking Disease in Marine Ecosystems. Front. Microbiol.7, 991. doi: 10.3389/fmicb.2016.00991

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 20

  FanC.DaiW.ZhangH.LiuS.LinZ.XueQ. (2024). Genomic and Proteomic Analyses of Extracellular Products Reveal Major Virulence Factors Likely Accounting for Differences in Pathogenicity to Bivalves between Vibrio mediterranei Strains. Animals14, 692. doi: 10.3390/ani14050692

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  FanC.LiuS.DaiW.HeL.XuH.ZhangH.et al. (2023). Characterization of Vibrio mediterranei Isolates as Causative Agents of Vibriosis in Marine Bivalves. Microbiol Spectr11, e04923-22. doi: 10.1128/spectrum.04923-22

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  GaddyK. E.SepterA. N.MrukK.MiltonM. E. (2025). A mutualistic model bacterium is lethal to non-symbiotic hosts via the type VI secretion system. mBio16, e00157-25. doi: 10.1128/mbio.00157-25

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  GreenT. J.SiboniN.KingW. L.LabbateM.SeymourJ. R.RaftosD. (2019). Simulated Marine Heat Wave Alters Abundance and Structure of Vibrio Populations Associated with the Pacific Oyster Resulting in a Mass Mortality Event. Microb Ecol77, 736–747. doi: 10.1007/s00248-018-1242-9

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  HelmM. M.BourneN.LovatelliA. (2004). Hatchery culture of bivalves: a practical manual. (Rome: Food and Agriculture Organization of the United Nations).

  • Google Scholar
  • View reference in article

• 25

  HuangZ.LiY.YuK.MaL.PangB.QinQ.et al. (2024). Genome-wide expanding of genetic evolution and potential pathogenicity in Vibrio alginolyticus. Emerging Microbes & Infections13, 2350164. doi: 10.1080/22221751.2024.2350164

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  KalyaanamoorthyS.MinhB. Q.WongT. K. F.Von HaeselerA.JermiinL. S. (2017). ModelFinder: fast model selection for accurate phylogenetic estimates. Nat Methods14, 587–589. doi: 10.1038/nmeth.4285

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 27

  KanarekK.KeppelK.CohenH.FridmanC. M.GerlicM.SalomonD. (2025). Assessing toxicity and competitive fitness of Vibrio isolates from coastal waters in Israel. mSphere10, e00025-25. doi: 10.1128/msphere.00025-25

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 28

  Kehlet-DelgadoH.HäseC. C.MuellerR. S. (2020). Comparative genomic analysis of Vibrios yields insights into genes associated with virulence towards C. gigas larvae. BMC Genomics21, 599. doi: 10.1186/s12864-020-06980-6

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 29

  Kesarcodi-WatsonA.KasparH.LateganM. J.GibsonL. (2008). Probiotics in aquaculture: The need, principles and mechanisms of action and screening processes. Aquaculture274, 1–14. doi: 10.1016/j.aquaculture.2007.11.019

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 30

  KimesN. E.GrimC. J.JohnsonW. R.HasanN. A.TallB. D.KotharyM. H.et al. (2012). Temperature regulation of virulence factors in the pathogen Vibrio coralliilyticus. The ISME Journal6, 835–846. doi: 10.1038/ismej.2011.154

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 31

  KushmaroA.BaninE.LoyaY.StackebrandtE.RosenbergE. (2001). Vibrio shiloi sp. nov., the causative agent of bleaching of the coral Oculina patagonica. International Journal of Systematic and Evolutionary Microbiology51, 1383–1388. doi: 10.1099/00207713-51-4-1383

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 32

  LemireA.GoudenègeD.VersignyT.PettonB.CalteauA.LabreucheY.et al. (2015). Populations, not clones, are the unit of vibrio pathogenesis in naturally infected oysters. The ISME Journal9, 1523–1531. doi: 10.1038/ismej.2014.233

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 33

  LinH.YuM.WangX.ZhangX.-H. (2018). Comparative genomic analysis reveals the evolution and environmental adaptation strategies of vibrios. BMC Genomics19, 135. doi: 10.1186/s12864-018-4531-2

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 34

  MarçaisG.DelcherA. L.PhillippyA. M.CostonR.SalzbergS. L.ZiminA. (2018). MUMmer4: A fast and versatile genome alignment system. PLoS Comput Biol14, e1005944. doi: 10.1371/journal.pcbi.1005944

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 35

  MasanjaF.LuoX.JiangX.XuY.MkuyeR.ZhaoL. (2024). Environmental and social framework to protect marine bivalves under extreme weather events. Science of The Total Environment946, 174471. doi: 10.1016/j.scitotenv.2024.174471

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 36

  MassS.CohenH.PodichetiR.RuschD. B.GerlicM.UshijimaB.et al. (2024). The coral pathogen Vibrio coralliilyticus uses a T6SS to secrete a group of novel anti-eukaryotic effectors that contribute to virulence. PLoS Biol22, e3002734. doi: 10.1371/journal.pbio.3002734

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 37

  MatkarimovI.SallaahM. H.SalayevU.KumarS.KhaitovaD.RU. (2025). Climate-induced stress and disease dynamics in aquaculture species. IJARES5, 1–11. doi: 10.70102/IJARES/V5S1/5-S1-01

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 38

  MatsudaS.HiyoshiH.TandhavanantS.KodamaT. (2020). Advances on Vibrio parahaemolyticus research in the postgenomic era. Microbiology and Immunology64, 167–181. doi: 10.1111/1348-0421.12767

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 39

  MinhB. Q.SchmidtH. A.ChernomorO.SchrempfD.WoodhamsM. D.Von HaeselerA.et al. (2020). IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era. Molecular Biology and Evolution37, 1530–1534. doi: 10.1093/molbev/msaa015

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 40

  MoreiraA. P. B.Chimetto TononL. A.Do Valle P. PereiraC.AlvesN.Amado-FilhoG. M.Francini-FilhoR. B.et al. (2014). Culturable Heterotrophic Bacteria Associated with Healthy and Bleached Scleractinian Madracis decactis and the Fireworm Hermodice carunculata from the Remote St. Peter and St. Paul Archipelago, Brazil. Curr Microbiol68, 38–46. doi: 10.1007/s00284-013-0435-1

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 41

  OliverE. C. J.DonatM. G.BurrowsM. T.MooreP. J.SmaleD. A.AlexanderL. V.et al. (2018). Longer and more frequent marine heatwaves over the past century. Nat Commun9, 1324. doi: 10.1038/s41467-018-03732-9

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 42

  PettonB.Destoumieux-GarzónD.PernetF.ToulzaE.De LorgerilJ.DegremontL.et al. (2021). The Pacific Oyster Mortality Syndrome, a Polymicrobial and Multifactorial Disease: State of Knowledge and Future Directions. Front. Immunol.12, 630343. doi: 10.3389/fimmu.2021.630343

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 43

  PritchardL.GloverR. H.HumphrisS.ElphinstoneJ. G.TothI. K. (2016). Genomics and taxonomy in diagnostics for food security: soft-rotting enterobacterial plant pathogens. Anal. Methods8, 12–24. doi: 10.1039/C5AY02550H

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 44

  PujalteM.-J.GarayE. (1986). Proposal of Vibrio mediterranei sp. nov.: A New Marine Member of the Genus Vibrio. International Journal of Systematic Bacteriology36, 278–281. doi: 10.1099/00207713-36-2-278

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 45

  RatoA.JoaquimS.MatiasA. M.RoqueC.MarquesA.MatiasD. (2022). The Impact of Climate Change on Bivalve Farming: Combined Effect of Temperature and Salinity on Survival and Feeding Behavior of Clams Ruditapes decussatus. Front. Mar. Sci.9, 932310. doi: 10.3389/fmars.2022.932310

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 46

  RichardsG. P.WatsonM. A.NeedlemanD. S.ChurchK. M.HäseC. C. (2015). Mortalities of Eastern and Pacific Oyster Larvae Caused by the Pathogens Vibrio coralliilyticus and Vibrio tubiashii. Appl Environ Microbiol81, 292–297. doi: 10.1128/AEM.02930-14

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 47

  RichterM.Rosselló-MóraR. (2009). Shifting the genomic gold standard for the prokaryotic species definition. Proc. Natl. Acad. Sci. U.S.A.106, 19126–19131. doi: 10.1073/pnas.0906412106

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 48

  Rubio-PortilloE.Martin-CuadradoA. B.Caraballo-RodríguezA. M.RohwerF.DorresteinP. C.AntónJ. (2020). Virulence as a Side Effect of Interspecies Interaction in Vibrio Coral Pathogens. mBio11, e00201-20. doi: 10.1128/mBio.00201-20

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 49

  SchindelinJ.Arganda-CarrerasI.FriseE.KaynigV.LongairM.PietzschT.et al. (2012). Fiji: an open-source platform for biological-image analysis. Nat Methods9, 676–682. doi: 10.1038/nmeth.2019

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 50

  SchwartzmanJ. A.RubyE. G. (2016). A conserved chemical dialog of mutualism: lessons from squid and vibrio. Microbes and Infection18, 1–10. doi: 10.1016/j.micinf.2015.08.016

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 51

  SeemannT. (2014). Prokka: rapid prokaryotic genome annotation. Bioinformatics30, 2068–2069. doi: 10.1093/bioinformatics/btu153

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 52

  SerranoW.TarazonaU. I.OlaecheaR. M.FriedrichM. W. (2018). Draft Genome Sequence of a New Vibrio Strain with the Potential To Produce Bacteriocin-Like Inhibitory Substances, Isolated from the Gut Microflora of Scallop (Argopecten purpuratus). Genome Announc6, e00419-18. doi: 10.1128/genomeA.00419-18

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 53

  SmaalA. C.FerreiraJ. G.GrantJ.PetersenJ. K.StrandØ. (2019). Goods and Services of Marine Bivalves. (Cham: Springer International Publishing). doi: 10.1007/978-3-319-96776-9

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 54

  SmithS.CiesielskiM.ClerkinT.Ben-HorinT.NobleR. T. (2025). Farmed oyster mortality follows consistent Vibrio community reorganization. mSystems10, e01078-25. doi: 10.1128/msystems.01078-25

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 55

  SotoW.NishiguchiM. K. (2021). Environmental Stress Selects for Innovations That Drive Vibrio Symbiont Diversity. Front. Ecol. Evol.9, 616973. doi: 10.3389/fevo.2021.616973

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 56

  SpeareL.CecereA. G.GuckesK. R.SmithS.WollenbergM. S.MandelM. J.et al. (2018). Bacterial symbionts use a type VI secretion system to eliminate competitors in their natural host. Proc. Natl. Acad. Sci. U.S.A.115, E10368–E10377. doi: 10.1073/pnas.1808302115

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 57

  SpraggeF.BakkerenE.JahnM. T.B. N. AraujoE.PearsonC. F.WangX.et al (2023). Microbiome diversity protects against pathogens by nutrient blocking. Science382, eadj3502. doi: 10.1126/science.adj3502

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 58

  StevensE. J.BatesK. A.KingK. C. (2021). Host microbiota can facilitate pathogen infection. PLoS Pathog17, e1009514. doi: 10.1371/journal.ppat.1009514

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 59

  Stewart-SinclairP. J.LastK. S.PayneB. L.WildingT. A. (2020). A global assessment of the vulnerability of shellfish aquaculture to climate change and ocean acidification. Ecology and Evolution10, 3518–3534. doi: 10.1002/ece3.6149

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 60

  TarazonaE.LucenaT.ArahalD. R.MaciánM. C.RuviraM. A.PujalteM. J. (2014). Multilocus sequence analysis of putative Vibrio mediterranei strains and description of Vibrio thalassae sp. nov. Systematic and Applied Microbiology37, 320–328. doi: 10.1016/j.syapm.2014.05.005

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 61

  FAO. (2023). The State of Food Security and Nutrition in the World 2023. (Rome: FAO). doi: 10.4060/cc3017en

  • CrossRef
  • Google Scholar

• 62

  ThompsonF. L.HosteB.ThompsonC. C.HuysG.SwingsJ. (2001). The Coral Bleaching Vibrio shiloi Kushmaro et al. 2001 is a Later Synonym of Vibrio mediterranei Pujalte and Garay 1986. Systematic and Applied Microbiology24, 516–519. doi: 10.1078/0723-2020-00065

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 63

  Tonkin-HillG.MacAlasdairN.RuisC.WeimannA.HoreshG.LeesJ. A.et al. (2020). Producing polished prokaryotic pangenomes with the Panaroo pipeline. Genome Biol21, 180. doi: 10.1186/s13059-020-02090-4

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 64

  UshijimaB.MeyerJ. L.ThompsonS.PittsK.MarusichM. F.TittlJ.et al. (2020). Disease Diagnostics and Potential Coinfections by Vibrio coralliilyticus During an Ongoing Coral Disease Outbreak in Florida. Front. Microbiol.11, 569354. doi: 10.3389/fmicb.2020.569354

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 65

  UshijimaB.RichardsG. P.WatsonM. A.SchubigerC. B.HäseC. C. (2018). Factors affecting infection of corals and larval oysters by Vibrio coralliilyticus. PLoS ONE13, e0199475. doi: 10.1371/journal.pone.0199475

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 66

  UshijimaB.SawJ. H.VideauP.HäseC. C. (2022). Comparison of Vibrio coralliilyticus virulence in Pacific oyster larvae and corals. Microbiology168, 001169. doi: 10.1099/mic.0.001169

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 67

  UshijimaB.VideauP.BurgerA. H.Shore-MaggioA.RunyonC. M.SudekM.et al. (2014). Vibrio coralliilyticus Strain OCN008 Is an Etiological Agent of Acute Montipora White Syndrome. Appl Environ Microbiol80, 2102–2109. doi: 10.1128/AEM.03463-13

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 68

  VezzulliL.GrandeC.ReidP. C.HélaouëtP.EdwardsM.HöfleM. G.et al. (2016). Climate influence on Vibrio and associated human diseases during the past half-century in the coastal North Atlantic. Proc. Natl. Acad. Sci. U.S.A.113, E5062. doi: 10.1073/pnas.1609157113

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 69

  WeiZ.ZhaoL.WangS.ChangL.ShiJ.KongX.et al. (2024). Paralytic shellfish toxins producing dinoflagellates cause dysbacteriosis in scallop gut microbial biofilms. Ecotoxicology and Environmental Safety273, 116146. doi: 10.1016/j.ecoenv.2024.116146

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 70

  WordenP. J.BogemaD. R.MicallefM. L.GoJ.DeutscherA. T.LabbateM.et al. (2022). Phylogenomic diversity of Vibrio species and other Gammaproteobacteria isolated from Pacific oysters (Crassostrea gigas) during a summer mortality outbreak. Microbial Genomics8, 000883. doi: 10.1099/mgen.0.000883

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 71

  XuK.WangY.YangW.CaiH.ZhangY.HuangL. (2022). Strategies for Prevention and Control of Vibriosis in Asian Fish Culture. Vaccines11, 98. doi: 10.3390/vaccines11010098

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 72

  YangR.LiuQ.HeY.TaoZ.XuM.LuoQ.et al. (2020). Isolation and identification of Vibrio mediterranei 117-T6 as a pathogen associated with yellow spot disease of Pyropia (Bangiales, Rhodophyta). Aquaculture526, 735372. doi: 10.1016/j.aquaculture.2020.735372

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 73

  ZaneveldJ. R.McMindsR.Vega ThurberR. (2017). Stress and stability: applying the Anna Karenina principle to animal microbiomes. Nat Microbiol2, 17121. doi: 10.1038/nmicrobiol.2017.121

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 74

  ZhangH.ZouX.HuJ.LiuS.FanC.DaiW.et al. (2025). Genomic insights into mechanism underlying virulence variations between Vibrio mediterranei strains different in pathogenicity toward bivalves. Aquaculture605, 742524. doi: 10.1016/j.aquaculture.2025.742524

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 75

  ZhangX.-H.HeX.AustinB. (2020). Vibrio harveyi: a serious pathogen of fish and invertebrates in mariculture. Mar Life Sci Technol2, 231–245. doi: 10.1007/s42995-020-00037-z

  • Pubmed Abstract
  • CrossRef
  • Google Scholar