Antagonistic Mechanisms of Probiotic Aliivibrio sp. Strain Vl2 Against Moritella viscosa: Evidence from Co-cultivation and Targeted Transcriptomic Analysis

TL;DR

A probiotic Aliivibrio strain reduces the growth and harmful effects of Moritella viscosa, a bacterium causing winter ulcers in salmon.

Contribution

This study identifies in vitro antagonistic mechanisms of a probiotic Aliivibrio strain against Moritella viscosa using co-cultivation and transcriptomic analysis.

Findings

Aliivibrio Vl2 inhibits the growth and pathogenicity of Moritella viscosa in vitro.

Transcriptomic analysis reveals potential mechanisms used by the probiotic to antagonize the pathogen.

The findings support the role of the probiotic in reducing winter ulcers in salmon.

Abstract

Winter ulcers, primarily caused by Moritella viscosa, represent a significant challenge for the Norwegian aquaculture industry. Effective control measures are hampered by the lack of effective vaccines and limited use of antibiotics, driven by the global effort to combat antibiotic resistance. Recent studies have shown that probiotic Aliivibrio spp. colonize the skin and ulcers of Atlantic salmon and are linked to a reduced prevalence of winter ulcers. These observations suggest that M. viscosa and Aliivibrio spp. may interact within ulcers in vivo. In this study, we investigated how the probiotic Aliivibrio sp. strain Vl2 (hereafter Aliivibrio Vl2) modulates M. viscosa in vitro, using both co-cultures and cultures within the salmonid cell line CHSE-214. We found that this probiotic strain antagonizes M. viscosa, reducing its growth and its pathogenicity toward the salmonid cells. Targeted transcriptome analysis of Aliivibrio Vl2 during co-culture revealed potential antagonistic mechanisms that impede the growth of this competing pathogen. Together, our findings demonstrate that this probiotic bacterium inhibits M. viscosa growth in vitro. Furthermore, the identified transcriptomic changes suggest potential mechanisms underlying the reduced prevalence of winter ulcers observed in field studies after probiotic administration.

The online version contains supplementary material available at 10.1007/s12602-025-10669-1.

Figures

Mean CFU/mL (log) with standard deviations (SD) of (A) *M. viscosa* (strain 7) and *Aliivibrio* Vl2 co-cultured in equal concentrations at 8 °C (*n* = 12) and (B) three different *M. viscosa* strains and *Aliivibrio* Vl2 cultivated in a 10:1 ratio as co-cultures and as monocultures at 4 °C (*n* = 30). Asterisks indicate statistical significance: p* < 0.05; *p* < 0.01; p* < 0.001. Figure created in Rstudio

Interactions on blood agar plates. A Cross streak of *Aliivibrio* Vl2 (horizontal streak) with zone lines (arrow) in the contact area against *M. viscosa* (vertical streak). B Growth reduction of *M. viscosa* by *Aliivibrio* Vl2 supernatant

Mean OD_600_ with SD of *M. viscosa* (NVI-3632) and *Aliivibrio* Vl2 cultivated in semipermeable bags, either as monocultures in separate vials or as co-cultures in shared vials. Duplicate bags also contained *M. viscosa* (NVI-3632) with *Aliivibrio* Vl2 BEVs (*n* = 10). Figure created in Rstudio

Cell viability (A) and LDH concentration (B) measured in vitro after 24-h incubation of CHSE-214 cell cultures with *Aliivibrio* Vl2, *M. viscosa* (strain 7), or control (*n* = 6). Cell viability (% reduction relative to control) after exposure to *M. viscosa* (strain 7) and probiotic supernatant or BEVs from *Aliivibrio* Vl2 (*n* = 20) (C). Asterisks indicate statistical significance: p* < 0.05; *p* < 0.01; p* < 0.001. Figure created in Rstudio

Representative images of CHSE-214 cells after exposure to (A) *M. viscosa* only, B *M. viscosa* and *Aliivibrio* Vl2 supernatant 1, C *M. viscosa* and *Aliivibrio* Vl2 supernatant 2, D *M. viscosa* and *Aliivibrio* Vl2 BEV’s, and (E) control. Arrows indicate elongated cells, while arrowheads indicate rounded cells

Heatmap with dendrograms of the 100 most significant DEG’s with co-cultures and monocultures in separate vertical clusters (*n* = 8). Blue indicates higher relative expression; red indicates lower relative expression (row-wise *z*-score). Figure created in Rstudio

Functional enrichment analysis (GSEA) comparing the expression profiles of *Aliivibrio* Vl2 when co-cultured with *M. viscosa* to *Aliivibrio* Vl2 in monoculture. This analysis is based on the sequences recognized by BLAST (*n* = 8). GO names are sorted based on Normalized Enrichment Score (NES) in descension for up-regulated genes (A) and ascension for down-regulated genes (B). Figure created in Omicsbox

Full text

Introduction

Bacterial ulcers are a major challenge for Norwegian open pen aquaculture of Atlantic salmon (Salmo salar) and the prevalence has increased substantially over the last decade [1]. This is primarily due to winter ulcers, a disease occurring most frequently during the winter months, typically causing round ulcers on the lateral sides of the fish. In 2023, outbreaks were documented in 320 different Norwegian farms (32% of all farms) [1] and the actual prevalence is likely even higher, as it is relatively easy to diagnose macroscopically and not a notifiable disease.

Moritella viscosa is the main aetiological agent causing winter-ulcer disease [2]. This Gram-negative bacterium is psychrophilic and has been shown to be more adherent when cultured at 4 °C compared to 15 °C, which has been proposed as a reason for its increased pathogenicity in the winter, when water temperatures are lower [3, 4]. The pathogen does not invade host cells, as demonstrated by immunohistochemical staining of infected tissues and cell-culture infection models [3, 5]. Tunsjø et al. proposed that cell damage is caused by extracellular proteins secreted by the bacteria, which leads to cytoskeleton disruption, pore formation, and ultimately cell lysis [3].

M. viscosa is often isolated from winter ulcers together with other bacteria [6, 7]. One of these agents, Aliivibrio wodanis, has been shown to strongly reduce the growth and acute virulence of *M. viscosa *[8]. However, A. wodanis is a pathogen by itself, and a co-infection generally leads to a prolonged course of illness [9].

Despite bacterial infections constituting a major challenge in the Norwegian salmon farming industry, there is an overall low use of antibiotics [1]. This is partly due to effective vaccines, along with strict regulatory oversight and biosecurity measures [10, 11]. Vaccines for M. viscosa, however*,* have failed to adequately protect the Atlantic salmon from winter ulcers, leaving the salmon farming industry vulnerable. This situation opens the door for alternative health management options such as probiotic bacteria, bacteriocins, bacteriophages, and other novel interventions [12, 13].

The relationship between vertebrates and resident bacteria dates back hundreds of millions of years and serves a wide variety of functions in the host [14]. In fish, microbial communities have been studied in skin, gills, and gut with several links made to disease resistance [15]. Beneficial bacteria introduced to the host are often referred to as probiotics, defined as “live microorganisms which when administered in adequate amounts confer a health benefit on the host” [16, 17]. Probiotic bacteria have been shown to serve many different purposes in fish, such as control of infectious diseases [18–20], water quality improvement [21, 22], and functional ingredients [23, 24]. These benefits can be conferred through competitive exclusion of pathogens [25–28], quorum quenching [29], immunomodulation [30], and more [31, 32]. The Aliivibrio genus consists of six established species [33]. Although two of these species are pathogenic, the remaining four species are considered to be commensal to marine host animals [34–38]. Some Aliivibrio isolates have also been reported to have probiotic effects, such as reduced prevalence of winter ulcers in Atlantic salmon [39]. We have recently recovered these ulcer-reducing probiotic isolates from skin and co-cultures in ulcerated tissues following bath administration [40]. The mechanisms by which these probiotics exert their effects remain unknown.

To characterize these effects and underlying mechanisms, we first conducted co-cultivation experiments to study interactions between M. viscosa and Aliivibrio sp. strain Vl2. We then evaluated the cytotoxic potential of the Aliivibrio strain toward a CHSE-214 cell line and examined whether its culture supernatant could mitigate the harmful effects of M. viscosa on these cells. Finally, we characterized the transcriptomic profile of Aliivibrio Vl2 during this interaction using RNA sequencing (RNA-seq).

Materials and Methods

Strain Collection

Three different isolates of M. viscosa (strain 7 (from a strain collection isolated at Stokkasjøen in 2019), NVI-3632 (RefSeq genome accession no. GCF_900120025.1) and NVI-5427 (RefSeq genome accession no. GCF_900120285.1)), all originally isolated from ulcerated fish were collected from the freeze stock (− 80 °C) at the Norwegian University of Life Sciences (NMBU) Department of Paraclinical Sciences, Bacteriology and Mycology Unit. These isolates represent the three clonal complexes (CC) of M. viscosa that have been isolated from farmed fish in Norway, where CC1 and CC3 are currently reported to dominate [41]. Strain NVI-3632 is the classic strain belonging to the CC1 category, while strain NVI-5427 belongs to CC2 and strain 7 belongs to the CC3 category.

Aliivibrio Vl2, NCIMB 42592 was originally isolated, cultivated, and stored as detailed by Klakegg et al. [39] and collected from freeze stock (− 80 °C) at Previwo AS (Oslo, Norway).

All bacteria were cultured as given here unless otherwise specified: bacteria were taken from stock at − 80 °C and first cultured on blood agar plates (OXOID, CM0271, blood agar base No. 2, with 5% bovine blood and 0.9% NaCl) for 5 days at 10 °C. A starter culture was prepared by transferring single colonies from a blood agar plate to Luria–Bertani (LB) broth (Merck, Germany) with 0.9% NaCl added. The cultures were incubated at 8 °C with 120 rpm for 3 days before expansion 1:10, followed by incubation for three more days.

Co-cultures in LB Broth

Starter cultures of Aliivibrio Vl2,* M. viscosa* (strain 7, NVI-5427 and NVI-3632), were prepared as described above. In two different experiments, the growth of M. viscosa and Aliivibrio Vl2 was assessed in liquid co-cultures and compared to monocultures. Each liquid co-culture experiment was conducted in triplicate.

First, we investigated if the growth of M. viscosa (strain 7) and Aliivibrio Vl2 was altered in liquid co-cultures at equal concentrations, compared to their growth in monocultures. Ten milliliters of M. viscosa (strain 7) at 0.5 OD_600_ (1.23 × 10^10^ CFU/mL) was added to 50-mL Erlenmeyer vials together with 10 mL of Aliivibrio Vl2 at 0.5 OD_600_ (1.51 × 10^10^ CFU/mL) (co-culture). The same volumes of the same bacterial cultures were also added to 10 mL of 0.9% LB broth without bacteria (monoculture). All cultures were incubated at 8 °C and 120 rpm for 5 days, with daily samplings for serial dilution and transfer to blood agar plates to determine colony-forming units (CFU) per millilitre (mL) of liquid culture.

In the second experiment, we assessed the inhibitory effects of Aliivibrio Vl2 on all three strains of M. viscosa. To show the probiotic’s inhibitory activity, the temperature and the relative proportion of the probiotic strain in the co-culture were reduced. To set up the co-cultures, 10 mL of M. viscosa culture (OD_600_ = 0.5, 1.8 ± 7 × 10^9^ CFU/mL) and 10 mL of Vl2 culture (OD_600_ = 0.05, 2 × 10^8^ CFU/mL) were combined in 50-mL Erlenmeyer vials. Monocultures were prepared using the same concentrations of each bacterial culture, added to 10 mL of 0.9% LB broth without other bacteria. The respective cultures were kept at 120 rpm at 4 °C for 5 days and sampled daily for serial dilution and transfer to blood agar plates. The growth was evaluated based on the number of colony-forming units (CFU) per mL of liquid culture.

Co-cultures on Blood Agar Plates

M. viscosa (strain 7) and Aliivibrio Vl2 were cultured as previously described. Both strains were then crossed with each other on a fresh blood agar plate by streaking one strain horizontally and the other vertically, creating an intersection between the two bacteria (shown in Fig. 3). The plate was incubated at 10 °C for 6 days, after which the interactions between the two species were evaluated. This experiment was conducted in triplicate.

Next, a starter culture of Aliivibrio Vl2 was prepared as previously described before transfer to 50-mL Falcon tubes. To isolate supernatants, the tubes were centrifuged at 4600 × g for 10 min, and the supernatants were double sterile filtered with a 0.2-µm pore size Minisart Syringe Filter (Sartorius, Germany), before storage in aliquots at − 80 °C until use.

M. viscosa (strain 7) was transferred from a blood agar plate and uniformly spread onto fresh blood agar plates using a Retro C80 plate carousel (Montebello diagnostics, Norway). Twenty microliters (µL) of Aliivibrio Vl2 supernatant thawed from − 80 °C storage was spotted onto the spread bacteria as a single droplet. The plates were then incubated at 4 °C for 7 days before evaluation. This experiment was conducted in triplicate.

Scanning Electron Microscopy (SEM)

Five-day-old colonies of Aliivibrio Vl2 were transferred to 10-mL fixative consisting of 5 mL 4% paraformaldehyde, 2.5 mL 0.5 M PIPES, 0.5 mL 25% glutaraldehyde, and 2 mL dH_2_O. Fixated bacteria were incubated on a glass slide coated with 1 mg/mL poly-L-lysin, dehydrated in an alcohol gradient, critical point dried, and sputter-coated with platinum particles. Bacteria on prepared slides were inspected with a Zeiss EVO 50 scanning electron microscope (20,000 × magnification, 10 kV EHT, 10–15 pA probe). The same was carried out for bacterial extracellular vesicles (BEVs), prepared and isolated as described below (see “Cell Culture Infection Assays” section).

Spatially Separated Co-cultures

For spatially separated co-cultures, semi-permeable 25-mm diameter regenerated cellulose bags with MW cutoff 12–14 kDa (Spectra/Por™, Los Angeles, CA) were prepared after the method of Colquhoun et al. [42]. Three knots were used to seal the bottom end of each bag and a sterile Venofix IV-tube (with cap and without needle, B. Braun, Germany) was inserted. The top of the bag was sealed and secured with knots using a separate sterile IV-tube and the sealed bags were disinfected using 70% ethanol. Five milliliters 0.5 OD_600_ M. viscosa (NVI-3632) or Aliivibrio Vl2 was added to the respective bags through the IV-tube. In one bag of M. viscosa (NVI-3632), 50 µL of Aliivibrio Vl2 BEVs was also added (see “Cell Culture Infection Assays” section for preparation and isolation). Each bag was submerged in 100 mL 0.9% NaCl LB broth in 500-mL Erlenmeyer vials. In addition, one vial contained both a bag of M. viscosa NVI-3632 and Aliivibrio Vl2. In total, this yielded four cultures: M. viscosa monoculture, Aliivibrio Vl2 monoculture, Vl2 and M. viscosa co-culture, and M. viscosa with Aliivibrio Vl2 BEVs. Each bag was incubated at 8 °C with 120 rpm and sampled for OD_600_ measurement at day 0, 1, 2, 3, 7, and 11. The LB broth outside the bags in each vial was inspected visually and cultured on blood agar plates to assess contamination at each timepoint. The experiment was performed in duplicate.

Cytotoxicity Assay

The supernatant from Aliivibrio Vl2 and M. viscosa (strain 7) was harvested as previously described for supernatant growth inhibition, but not sterile filtered, resulting in a mixed bacterial culture containing supernatant and approximately 10^3^ CFU/mL. The cultures were then aliquoted and stored at − 80 °C until use.

CHSE-214 salmon cell line (Merck, Germany) was grown in 75-cm^2^ culture flasks with Leibovitz’s L-15 Medium, GlutaMAX™ supplement (Gibco, USA), and 10% FBS media (Gibco, USA) at 20°C^3^. Cells were detached with 0.05% trypsin (Gibco, USA) and split weekly at ≈80% confluence in a 1:4 ratio. Cells were counted with a TC20 Automated Cell Counter (Bio-Rad Laboratories, USA) and viability was assessed with erythrosine B (Logos biosystems, South Korea). One hundred microliters of fresh media containing approximately 20,000 cells at passage 20–30 with > 97% viability was transferred to each well in 96-well plates and incubated at 20 °C. When the wells reached ≈80% confluence, the cytotoxicity assays were conducted by adding 20 µL of the mixed bacterial culture to duplicate wells and incubating at 8 °C for 24 h.

To evaluate toxic effects on the cells, all wells were inspected visually using an EVOS M5000 Imaging System (Thermofisher, USA). The supernatant was transferred to new wells (without cells) and the LDH concentration was measured using the CyQUANT™ LDH cytotoxicity assay kit (Thermofisher, USA). The wells with supernatant removed, still containing CHSE-214 cells, were then carefully washed five times with 1 × PBS (Gibco, USA) and incubated with fresh media containing 200 µg/mL gentamycin for 4 h, before measuring cell viability with Alamar Blue (Thermofisher, USA) as per the manufacturer’s instructions using a Cytation 3 Cell Imaging Multi-Mode Reader (BioTek, USA). Wells without CHSE-214 cells were used to subtract the background signal. Two biological replicates were used per group in this experiment [8, 42, 43].

Cell Culture Infection Assays

CHSE-214 cells were prepared as described above. Bacterial cultures of M. viscosa (strain 7) and Aliivibrio Vl2 were prepared as described above and adjusted to an OD_600_ of 0.5. Ten milliliters of each of these cultures was combined and co-cultured for 3 days. The culture was then expanded 1:10 and incubated for an additional 3 days, followed by a second 1:3 expansion and 3 more days of incubation, before harvest of the supernatant (Supernatant 1). A monoculture of Aliivibrio Vl2 without M. viscosa was also prepared in the same manner (Supernatant 2). Both supernatants were harvested by centrifugation at 15,000 × g at 4 °C for 10 min, double sterile filtered with a 0.2-µm pore size Minisart Syringe Filter, and aliquoted to Eppendorf tubes. In addition, bacterial extracellular vesicles (BEVs) were isolated from 300 mL Supernatant 1 by ultracentrifugation as described by Brudal et al. [44]. The BEVs were suspended in 600 µL PBS and aliquoted into Eppendorf tubes. This procedure was repeated for another 300 mL of supernatant 1, but the BEVs were suspended in 6 mL PBS, introducing a 1:10 dilution. All supernatants and BEVs were stored at − 80 °C until further use. To verify the absence of culturable bacteria, the supernatant and BEV samples were cultured on 0.9% NaCl blood agar plates. The presence of BEVs was also controlled by SEM as described in the “Scanning Electron Microscopy (SEM)” section.

The cell culture infection assay was conducted on CHSE-214 cells with ≈80% confluency by adding 20 µL of M. viscosa (strain 7) at 0.8(± 0.1) OD_600_ (≈100 MOI) and 20 µL of either PBS, Aliivibrio Vl2 supernatant, or Aliivibrio Vl2 BEVs. Additional groups were included to control for potential unintended pathogenic effects by adding 20 µL PBS and 20 µL Aliivibrio Vl2 supernatant or 20 µL PBS and 20 µL Aliivibrio Vl2 BEVs. For negative controls, 40 µL PBS was added to the CHSE-214 cells. Probiotic safety to CHSE-214 cells was investigated in the “Cytotoxicity Assay” section and therefore not included in this assay. After incubation at 8 °C for 18 h, each well was carefully washed five times with PBS and incubated with fresh media containing 200 µg/mL gentamycin for 4 h, before measuring cell viability with Alamar Blue (Thermofisher) as per the manufacturer’s instructions. Each group was replicated in four wells (technical replicates) per plate, and the experiment was repeated four times (biological replicates). Each plate contained four wells per group. Wells exceeding the detection limit were excluded, resulting in the omission of one well. Cell viability was calculated by subtracting the background signal from wells without CHSE-214 cells, and the mean value for each group was compared to the controls. Visual assessment of cell damage was performed using the EVOS M5000 Imaging System (Thermofisher, USA), with images captured from a fixed location using the Cytation 3 Cell Imaging Multi-Mode Reader (BioTek, USA).

RNA-seq Analysis

Monocultures and co-cultures of Aliivibrio Vl2 and M. viscosa (NVI-3632) in LB broth (0.9% NaCl) were prepared as described for liquid co-cultures. Four replicate co-cultures consisting of 10 mL Aliivibrio Vl2 at 0.5 OD_600_ and 10 mL M*. viscosa* (NVI-3632) at 0.5 OD_600_ were set up, in addition to four monocultures of Aliivibrio Vl2 at 0.5 OD_600_ in 10 mL 0.9% NaCl LB broth. All eight cultures were incubated at 8 °C and 120 rpm for 72 h, with samples collected at 0, 24, 48, and 72 h for serial dilution and determination of CFU/mL (Table S1 and Figure S1). The same experiment was then repeated for the preparation of samples for RNA sequencing, with all cultures sampled after 24 h. The cultures were centrifuged at 4600 × g, and the supernatant was discarded. The bacterial pellets were resuspended, and RNA was extracted using the RNeasy Protect Bacteria Mini Kit (Qiagen, Germany) following the manufacturer’s instructions. RNA concentration and purity were measured using a Multiskan Sky Spectrophotometer (Thermofisher, USA), and RNA integrity (RIN) was evaluated with a 4200 TapeStation (Agilent, USA). Samples were sent to Novogene (UK) for library preparation, including rRNA removal and prokaryotic mRNA paired-end sequencing with Illumina PE150, generating 5–10 million reads per sample for biological quadruplicates of both groups.

The Raw sequencing reads were quality filtered using “bbduk,” version 37.48 (BBMap – Bushnell B., https://sourceforge.net/projects/bbmap/) and Filtlong v0.2.0 (https://github.com/rrwick/Filtlong), respectively. Hybrid (short and long reads) whole-genome assemblies were obtained with Unicycler v0.4.8 [45]. The quality of the assembled genomes was assessed by mapping the sequencing reads to the de novo assembled genomes using Minimap2 and samtools ver.1.3.1 as well as by the assessment of the assembly graphs. The assembled genomes were annotated using Prokka [46].

The transcriptome reference file was obtained using an AGAT conversion Perl script, “agat_convert_sp_gff2gtf.pl,” to obtain the GTF file from the Prokka-generated GFF file (REF: Dainat J. AGAT: Another Gff Analysis Toolkit to handle annotations in any GTF/GFF format. (version v0.7.0). Zenodo. https://www.doi.org/10.5281/zenodo.3552717). The reads were aligned using STAR to the annotated genome using the genome sequence (from the de novo assembly) and the GTF file as input [47].

The obtained gene counts were processed using the DESeq2 package [48] in RStudio (version 2024.04.2 + 764). Normalization was performed using DESeq2’s built-in median ratio method, which accounts for sequencing depth and library size differences, and normalized counts were extracted. For visualization, log_2_ transformation and variance-stabilizing transformation (VST) were applied to stabilize variance across gene expression levels. The normalized and transformed data were visualized using boxplots to ensure consistency across samples. The DESeq2 pipeline was subsequently applied to prepare for differential expression analysis, including size factor estimation and dispersion modeling.

The OmicsBox bioinformatics suite (formerly Blast2Go [49]) was used for functional annotation with BLAST [50] and the Gene set enrichment analysis (GSEA), filtered for 0.05 P-value and 1000 permutations.

The sequence data is available at Fairdomhub (https://doi.org/10.15490/fairdomhub.1.datafile.7605.1).

Statistics

Bacterial growth kinetics were analyzed by Welch’s t-test. The cytotoxicity assay was analyzed by a nonparametric Mann–Whitney test by comparing the cell viability per well for each group. Damage reduction was analyzed by the unpaired two-tailed t-test by comparing the average cell viability per plate for each group. All assumptions have been fulfilled. All data reported are untransformed values.

Results

To assess the probiotic potential of Aliivibrio Vl2 against the fish pathogen M. viscosa, a series of co-culture experiments were performed under varying in vitro conditions. The results presented below detail the effects of the probiotic on M. viscosa viability, its ability to protect host cells, and associated changes in bacterial gene expression during co-culture.

Co-cultures in LB Broth

When co-cultured with Aliivibrio Vl2 in equal concentrations, the number of detected M. viscosa (strain 7) colonies was reduced by 2 logs after 24 h compared to the M. viscosa monoculture (Fig. 1A). No colonies of M. viscosa were detected after 32 h. At all timepoints from 24 h and onwards, the M. viscosa co-cultured with Aliivibrio Vl2 showed significantly lower CFU/mL compared to the M. viscosa monoculture (P < 0.05 by Welch’s t-test). The growth of Aliivibrio Vl2 was not significantly affected by M. viscosa at any timepoint in co-culture (P > 0.05 by Welch’s t-test), but the CFU/mL was reduced by 1 log after 96 h and 0.6 log after 120 h, when compared to the Aliivibrio Vl2 monoculture (n = 12).Fig. 1. Mean CFU/mL (log) with standard deviations (SD) of (A) M. viscosa (strain 7) and Aliivibrio Vl2 co-cultured in equal concentrations at 8 °C (n = 12) and (B) three different M. viscosa strains and Aliivibrio Vl2 cultivated in a 10:1 ratio as co-cultures and as monocultures at 4 °C (n = 30). Asterisks indicate statistical significance: p < 0.05; ** p < 0.01; **** p < 0.001. Figure created in Rstudio

When evaluating the growth of three strains of M. viscosa co-cultured with Aliivibrio Vl2 in a 10:1 ratio, all strains were outcompeted by Vl2 within 120 h (Fig. 1B). Notably, no colonies from M. viscosa (strain 7) were detected after 72 h. Strain NVI-3632 was strongly reduced after 72 h and was not detectable after 120 h. Strain NVI-5427 appeared less impacted by the probiotic species during earlier sampling points but was undetectable at the final sampling. All three M. viscosa strains exhibited significantly lower CFU/mL in at least the final timepoint compared to the M. viscosa monoculture (P < 0.05 by Welch’s t-test). Aliivibrio Vl2 appeared to be less affected by the presence of any M. viscosa strain as the growth in monoculture was similar to the co-cultures.

Co-cultures on Blood Agar Plates

Six days after crossing M. viscosa (strain 7) and Aliivibrio Vl2, the streaked colonies were assessed. Aliivibrio Vl2 exhibited growth extending over M. viscosa on both sides (Fig. 2A). When exposing M. viscosa to supernatant from Aliivibrio Vl2, reduced growth was observed after 7 days in the area where the supernatant had been placed (Fig. 2B). Similar results were observed in triplicates of both experiments.Fig. 2. Interactions on blood agar plates. A Cross streak of Aliivibrio Vl2 (horizontal streak) with zone lines (arrow) in the contact area against M. viscosa (vertical streak). B Growth reduction of M. viscosa by Aliivibrio Vl2 supernatant

Spatially Separated Co-cultures

Spatially separated co-cultures of M. viscosa and Aliivibrio Vl2 showed clear differences in growth between co-cultures and monocultures for both bacteria (Fig. 3). The growth of M. viscosa was lower at every sampling point past day 0 when co-cultured with Aliivibrio Vl2 compared to when monocultured. By day 11, there was an average reduction in growth of 53.1%. There was also a reduction of 19% in Aliivibrio Vl2 growth when co-cultured with M. viscosa as compared to a monoculture*.* Scanning electron images visualize the vegetative Aliivibrio Vl2 with surrounding vesicles (Figure S2A) and the vesicles isolated by ultracentrifugation from Aliivibrio Vl2 (Figure S2B). However, there was no apparent growth reduction of M. viscosa when exposed to Aliivibrio Vl2 vesicles.Fig. 3. Mean OD_600_ with SD of M. viscosa (NVI-3632) and Aliivibrio Vl2 cultivated in semipermeable bags, either as monocultures in separate vials or as co-cultures in shared vials. Duplicate bags also contained M. viscosa (NVI-3632) with Aliivibrio Vl2 BEVs (n = 10). Figure created in Rstudio

Probiotic Cytotoxicity Assay

Aliivibrio Vl2 showed no visible signs of damaging CHSE-214 cells, and cell viability, as measured by Alamar blue, was comparable to the control (Fig. 4A) (P > 0.05 by nonparametric Mann–Whitney test). In contrast, M. viscosa reduced the viability of the CHSE-214 cell culture to 46% in 24 h, a significant difference compared to the control group (P = 0.022 by nonparametric Mann–Whitney test). Additionally, LDH release measured by relative fluorescent units (Fig. 4B) was significantly higher in wells with M. viscosa than in the controls (P = 0.028 by nonparametric Mann–Whitney test).Fig. 4. Cell viability (A) and LDH concentration (B) measured in vitro after 24-h incubation of CHSE-214 cell cultures with Aliivibrio Vl2, M. viscosa (strain 7), or control (n = 6). Cell viability (% reduction relative to control) after exposure to M. viscosa (strain 7) and probiotic supernatant or BEVs from Aliivibrio Vl2 (n = 20) (C). Asterisks indicate statistical significance: p < 0.05; ** p < 0.01; **** p < 0.001. Figure created in Rstudio

Cell Culture Damage Reduction Assay

Cell viability in wells containing only M. viscosa was significantly lower than in wells with M. viscosa and Aliivibrio Vl2 supernatant 1 (from co-culture of Aliivibrio Vl2 and M. viscosa, P < 0.0001) or supernatant 2 (from monoculture of Aliivibrio Vl2, P = 0.002) (Fig. 4C). Visually, more live cells were observed in wells containing probiotic supernatant compared to those with only M. viscosa (Fig. 5). However, signs of cell damage typically seen prior to cell death were present in wells with supernatant, such as elongation and rounding (as indicated by arrows and arrowheads in Fig. 5B and C). Addition of Aliivibrio Vl2 BEVs did not significantly reduce the damage caused by M. viscosa as measured by cell viability (P > 0.05), and no damage reduction was observed visually. The wells with added supernatant 1 had a higher level of cell viability on average compared to supernatant 2, but they were not significantly different from each other (P > 0.05). The wells containing only supernatants or Aliivibrio Vl2 BEVs were not significantly different from the control, which yielded 100% viability (data not shown).Fig. 5. Representative images of CHSE-214 cells after exposure to (A) M. viscosa only, B M. viscosa and Aliivibrio Vl2 supernatant 1, C M. viscosa and Aliivibrio Vl2 supernatant 2, D M. viscosa and Aliivibrio Vl2 BEV’s, and (E) control. Arrows indicate elongated cells, while arrowheads indicate rounded cells

Transcriptomic Response of Aliivibrio Vl2 to Co-culture with M. viscosa

To investigate adaptive transcriptomic changes in Aliivibrio Vl2 during interaction with M. viscosa, we compared gene expression profiles of Aliivibrio Vl2 in co-cultures with M. viscosa to those of Aliivibrio Vl2 in monoculture. The growth in monocultures and co-cultures was recorded daily for 72 h, with a strong reduction in CFU/mL of M. viscosa in the co-culture starting at 48 h (Table S1 and Figure S1).

A total of 140,355,970 transcripts annotated to 3870 different genes were recognized for Aliivibrio Vl2 overall. Of these, 708 genes were significantly differentially expressed when comparing the gene expression profiles of Aliivibrio Vl2 monoculture to Aliivibrio Vl2 co-cultured with M. viscosa (Padj < 0.05). Of these significantly differentially expressed genes (DEGs), 326 were upregulated and 382 were downregulated in the co-cultured Aliivibrio Vl2. The 100 most significant DEGs were used to compare the samples, presented in Fig. 6, clearly separating co-cultures and monocultures into distinct clusters based on the Euclidean distance.Fig. 6. Heatmap with dendrograms of the 100 most significant DEG’s with co-cultures and monocultures in separate vertical clusters (n = 8). Blue indicates higher relative expression; red indicates lower relative expression (row-wise z-score). Figure created in Rstudio

DEGs were functionally annotated and assigned to 712 Gene Ontology (GO) terms. In the co-culture, downregulated GO terms (651) significantly outnumbered the upregulated ones (63). The 32 most significantly upregulated and 32 most significantly downregulated GO terms from the GSEA are presented in Fig. 7.Fig. 7. Functional enrichment analysis (GSEA) comparing the expression profiles of Aliivibrio Vl2 when co-cultured with M. viscosa to Aliivibrio Vl2 in monoculture. This analysis is based on the sequences recognized by BLAST (n = 8). GO names are sorted based on Normalized Enrichment Score (NES) in descension for up-regulated genes (A) and ascension for down-regulated genes (B). Figure created in Omicsbox

Most downregulated GO terms were associated with metabolic and biosynthetic processes for nucleotides and amino acids, while the upregulated GO terms were more varied in their functional classifications (Fig. 7). Several significantly upregulated genes in the co-cultured Aliivibrio Vl2 were associated with production, secretion, and uptake of siderophores (Table 1). The gene encoding N(2)-citryl-N(6)-acetyl-N(6)-hydroxylysine synthase, an enzyme that synthesises the direct precursor to aerobactin (KEGG pathway map00997), was the most significantly differentially expressed gene in the entire transcriptomics dataset (Padj = 6.86E − 26; fold-change = 6.9). Aerobactin is a siderophore previously shown to be produced by A. fischeri in co-cultures with other bacteria [51]. Aerobactin synthase, which catalyses the terminal step of the aerobactin biosynthetic pathway, was detected but not significantly differentially expressed in our dataset. Several genes related to ABC transporters are significantly upregulated in the co-culture (fold-change = 1.4–3.7), reported to facilitate export and uptake of iron and iron-siderophore complexes across the inner bacterial membrane [52]. Genes encoding TonB-dependent receptors, required for ferric-siderophore uptake across the outer bacterial membrane [53], are also upregulated (fold-change = 1.9). Genes encoding systems for uptake of ferric coprogen, rhodotulic acid, enterochelin, and ferrioxamine were also significantly upregulated, while genes related to production of these siderophores were not detected. Table 1. Upregulated genes (Padj < 0.05) for siderophoresGeneAverage expression (monoculture)Sd (monoculture)Average expression (co-culture)Sd (co-culture)Fold-changeP-value (adjusted)N(2)-citryl-N(6)-acetyl-N(6)-hydroxylysine synthase2415516634356.96.86E − 26Iron ABC transporter substrate-binding protein122204481143.71.51E − 12Iron-siderophore ABC transporter substrate-binding protein10829272722.50.000052TonB-dependent receptor53514510152051.90.0021Outer membrane receptor for ferric coprogen and ferric-rhodotorulic acid237594401021.90.0041ABC-type enterochelin transport system, periplasmic component43836616741.40.0055ABC-type Fe3 + transport system, periplasmic component23571548626317312.70.007Imelysin-like iron-regulated protein IrpA, duplicated M75 peptidase-like domain7818063137102440.0073TonB-dependent receptor plug domain-containing protein3152240081.30.015High-affinity Fe2 +/Pb2 + permease2411024691161.90.024Ferrioxamine B receptor31414562215220.02

Alterations in motility-related gene expression were also observed, with flagellar genes downregulated and genes negatively regulating motility upregulated.

Finally, several genes involved in antimicrobial activity were also detected in Aliivibrio Vl2. These include genes encoding antibiotic biosynthesis monooxygenases, isopenicillin N synthase, and toxin-antitoxin systems.

Discussion

In the present study, we have investigated how two naturally occurring bacteria in Atlantic salmon, the ulcer pathogen M. viscosa and the probiotic strain Aliivibrio Vl2, interact under different conditions in vitro. Our findings show that the growth of M. viscosa is significantly inhibited by Aliivibrio Vl2, suggesting a competitive relationship driven by the probiotics mechanisms to outcompete the pathogen. These findings are similar to what is reported for A. wodanis; however, A. wodanis is a pathogen. These mechanisms, which appear to be contact-independent, also reduce the level of damage inflicted on CHSE-214 cell cultures by M. viscosa.

Growth Suppression of M. viscosa in Co-cultures

The controlled in vitro setting enabled a clear assessment of the antagonistic effects of Aliivibrio Vl2 against M. viscosa, with consistent growth inhibition observed across varying concentrations, temperatures, and bacterial strains, demonstrating stable antimicrobial performance that persists under diverse conditions. There are differences in how long the different strains of M. viscosa were able to resist the presence of Aliivibrio Vl2. These variations in susceptibility may be related to their respective CC groups [41]; however, as only one strain per CC group was tested, the results may reflect strain-specific characteristics. Despite these differences, all three strains were ultimately outcompeted in every replicate, indicating broad-spectrum efficacy of the probiotic towards M. viscosa. A substantial amount of evidence documents similar antagonistic effect [54–56], underscoring the extensive potential for desirable characteristics among selected probiotic strains.

Bacteria can develop specific responses when encountering other bacteria on solid medium. Plate crossing experiments have previously been carried out for M. viscosa and A. wodanis, with documented contact-independent inhibition of M. viscosa. Hjerde et al. attributed the inhibition to increased gene expression level within a bacteriocin locus and siderophore systems [8]. Here, we observed a different phenomenon, with overgrowth of Aliivibrio Vl2 and contact-dependent inhibition of M. viscosa. The absence of contact-independent inhibition on blood agar plates, unlike in LB medium, has been reported for A. wodanis [8] and may suggest that the same inhibitory compounds are either not produced under these conditions or are unable to diffuse effectively through the solid medium. The overgrowth indicates that the probiotic bacterium is unaffected by the presence of M. viscosa and may even be attracted to it, possibly stimulated by substances produced by the pathogen—a phenomenon previously documented in other bacterial interactions [57, 58]. However, the growth of Aliivibrio Vl2 was reduced when co-cultured with M. viscosa in semi-permeable tubings, which may also be attributed to the altered culture conditions. Another observation was the zonal lines at the contact point, which suggests that the probiotic bacteria produce distinct substances in this location, not found elsewhere in the colony [59].

Finally, we investigated if the growth reduction was contact-dependent through semi-permeable bags. The growth of M. viscosa was still strongly reduced (Fig. 3). Hence, the inhibition is non-contact dependent, which is also supported by the growth inhibition seen on blood agar plates by the addition of probiotic supernatant (Fig. 2B). This means that the probiotic bacterium secretes one or more diffusible compounds capable of impeding M. viscosa growth at a distance, consistent with observations previously reported for *A. wodanis * [8]. The inhibitory effects of Aliivibrio Vl2 were more pronounced during direct co-culture (Fig. 1), likely due to higher concentrations of these compounds in proximity to M. viscosa. However, the possibility of additional contact-dependent mechanisms contributing to this antagonism cannot be excluded [60].

Reduced Cytotoxicity

Mixed cultures of Aliivibrio Vl2 and M. viscosa were screened for cytotoxic effects (Fig. 4A and B). While the probiotic caused no damage to the CHSE-214 cells, M. viscosa reduced the cell viability to 46% compared to that of the controls. The presence of M. viscosa elevated LDH level in the culture medium indicates cell lysis, resulting from cytoskeleton disruption and pore formation [3]. This was not observed in any of the wells with Aliivibrio Vl2.

When CHSE-214 cells were exposed to M. viscosa in a damage reduction assay, the cell survival increased from 3.8 to 49.4% on average if probiotic supernatant 1 (from co-culture) or 2 (from monoculture) was also added to the wells (Fig. 4C). This indicates that Aliivibrio Vl2 can effectively reduce the damage to CHSE-214 cells during M. viscosa infections through the secretion of inhibitory or protective compounds into its supernatant. These findings are consistent with other in vitro studies, demonstrating that several probiotic-derived supernatants can improve cell viability during pathogen exposure [61]. Although the cell survival increased drastically by the presence of both probiotic supernatants, most cells were still negatively affected by the presence of M. viscosa, evident by widespread cell elongation and some cell rounding (Fig. 5B and C), which are signs of stress typically seen prior to detachment and cell death [39]. This suggests that the probiotic supernatants may mitigate the stress caused by M. viscosa, potentially prolonging cell survival despite the observed signs of cellular distress. These effects may be due to competitive exclusion of the pathogen [62], protective effects directly on the cell line [63], or both [64].

When comparing the effects of the two supernatants on the cell survival, no apparent difference was detected. This means that the production of inhibitory compounds by Aliivibrio Vl2 was not dependent on the presence of M. viscosa. However, supernatant 1 resulted in more cells surviving in most experiments, implying that Aliivibrio Vl2 may have adjusted its extracellular production repertoire to specifically target M. viscosa during co-culture, a trait related to ecological competition documented in other species within the Vibrionaceae family [65].

Lack of Antagonism by BEVs

Many bacterial species continuously release vesicles into the surrounding environment, and some vesicles from probiotic bacteria have been shown to harbour anti-bacterial properties in fish [66]. While we observed the production of BEVs in Aliivibrio Vl2, these BEVs did not affect the viability of M. viscosa when Aliivibrio Vl2 BEVs were added directly to the semi-permeable bags used for cultivation (Fig. 3) or when introduced to cell cultures (Fig. 4C and Fig. 5D).

Transcriptomic Responses in Aliivibrio Vl2

Bacterial transcriptomic responses to co-cultivation have been shown to include varied changes in protein synthesis, including upregulation, downregulation, or no change at all [67–69]. Several factors may contribute to the different responses [70, 71]. The transcriptome analysis of Aliivibrio Vl2 revealed a significant downregulation of genes associated with various metabolic and biosynthetic pathways, particularly those involved in protein synthesis. The downregulated genes were associated with multiple GO terms with major overlaps between them (bubble sizes, Fig. 7B). This is likely due to these transcripts being applicable to several similar metabolic processes, while the upregulated genes presented in Fig. 7A appear to be associated with effects in which the specific transcripts are exclusively linked to individual pathways. Despite the diversity of transcriptomic responses observed across different studies, the findings presented here are consistent with omics data reported for A. wodanis and *A. salmonicida * [8, 72]. Bjelland et al. hypothesized that this may represent a survival strategy, enabling the bacteria to evade the host immune system and thereby facilitate successful colonization [73]. However, a similar trend was also observed in A. wodanis and in the present study in response to the presence of M. viscosa, absent a host. We therefore propose that these changes in global gene expression may also reflect a reprioritisation of core metabolic functions, favoring processes more relevant to adaptation and survival within a co-culture environment, including host-associated niches.

Motility-related processes appeared to be restricted in the co-cultured Aliivibrio Vl2, as genes associated with flagellar assembly were downregulated, while those involved in the negative regulation of motility showed increased expression. Instead of prioritizing motility-related gene expression, the co-cultured Aliivibrio exhibited increased expression of genes associated with cell adhesion, aggregation, and biofilm formation. Other bacterial transcriptomic studies have reported increased biofilm production triggered by co-culture conditions [68]. When experiencing environmental stressors, bacteria can gather on a surface and encase themselves in protective biofilm [74–76]. The population must communicate through cell signals to coordinate the production of biofilm matrix [77], which is represented by several upregulated GO-terms (Fig. 7A) and standalone DEGs, such as acyl-homoserine-lactone (AHL) synthase. These changes suggest that Aliivibrio Vl2 may be initiating survival-associated pathways in response to the presence of M. viscosa.

Iron is essential for the growth and survival of nearly all bacteria. However, due to its limited bioavailability, many bacteria produce siderophores to acquire iron. Siderophores are molecules with a high affinity for iron, released by bacteria into the surrounding environment to scavenge iron and transport it back. Many different siderophore systems have been documented over the years [53]. In this study, we observed the inhibitory effects of Aliivibrio Vl2 on the viability of M. viscosa in co-cultures and documented significant upregulation of siderophore pathways in Aliivibrio cells during this interaction. Although a direct causal link cannot be confirmed, these findings imply that Aliivibrio Vl2 competes for available iron by producing siderophores, thereby depriving M. viscosa of this essential growth factor. As many siderophores are small (< 1000 Da) [78, 79], they are able to pass through the semi-permeable tubings and have contact-independent antagonistic effects, as observed in these studies.

Aliivibrio species have previously been shown to produce siderophores to drive competitive exclusion in co-cultures with other species [51, 80]. Eickhoff et al. showed that the siderophore known as aerobactin allows Aliivibrio fischeri to establish a niche by denying growth of its competitor. To capture transcripts relevant to the observed interactions, the samples taken at 24 h were selected for transcriptomic analysis. As our most significant DEG at this timepoint encodes the final intermediary biosynthetic step, the subsequent strong reduction of M. viscosa abundance 24–48 h later likely coincides with the completion of aerobactin biosynthesis and the secretion of substantial quantities into the environment. In addition to competitive exclusion via resource competition, aerobactin can enhance bacterial defences in biofilm formation and oxidative stress protection [81].

Some bacteria can steal iron sequestered by siderophores produced by other bacteria, a competitive strategy known as siderophore piracy [82]. Even unculturable bacteria can be stimulated to grow when co-cultured with a siderophore-producing bacterium through piracy [57]. Although there is a spike in aerobactin-precursor transcripts, the final step is not (yet) upregulated. Despite this, several TonB-receptors and ABC-transporters important for siderophore-complex import are significantly upregulated. Upregulated genes in four additional siderophore systems were recognized but without the genes related to the production of these siderophores. Although this can be due to uncharacterized annotations, we speculate that the probiotic bacteria may be a siderophore pirate, capable of scavenging siderophore complexes produced by other species. Notably, M. viscosa is known to produce siderophores [83] which could be pirated by Aliivibrio Vl2. This phenomenon may also explain the observed attraction towards M. viscosa on blood agar plates (Fig. 2A).

Competitive exclusion can also be exerted through the secretion of antimicrobial substances [84, 85]. Aliivibrio Vl2 expressed genes encoding antibiotic biosynthesis monooxygenase and isopenicillin N synthase with related dioxygenases, which are involved in the production of potent antibiotics [86, 87]. Several genes linked to toxin and anti-toxin systems were also upregulated in the co-cultured Aliivibrio Vl2, indicating an activation of multiple offensive and defensive mechanisms that can benefit the probiotic in the competition with M. viscosa.

The differential gene expression patterns observed in Aliivibrio Vl2 suggest the activation of several competitive strategies in response to M. viscosa. It remains unclear whether the observed reduction in M. viscosa abundance in co-cultures results from iron starvation, bactericidal activity, or a combination of both. Future studies are required to elucidate the precise mechanisms involved.

Winter-Ulcer Disease

In Norwegian salmon aquaculture, it is not uncommon for M. viscosa-induced winter ulcers to be co-infected with *A. wodanis * [9]. This pathogenic Aliivibrio species exhibits a robust non-contact-dependent inhibitory effect on M. viscosa, likely mediated by bacteriocins and siderophore production [8]. Recent work has shown that probiotic Aliivibrio spp. also colonize salmon ulcers, potentially reducing the ulcer prevalence through competitive exclusion [40]. Our findings parallel those reported for A. wodanis, with one notable distinction; Aliivibrio Vl2 does not exhibit cytotoxicity. This suggests that the attenuation of acute winter-ulcer symptoms observed in the presence of A. wodanis may also occur with Aliivibrio Vl2, but without the associated progression to chronic ulceration that is linked to A. wodanis. This interpretation is further supported by the overall reduction in ulcer prevalence in both Atlantic salmon and lumpsuckers following administration of probiotic Aliivibrio species [39, 88].

Conclusion

The interaction between pathogenic and probiotic bacteria is a dynamic process that can significantly influence disease outcomes. In this study, we have shown that Aliivibrio Vl2 exhibits antagonistic traits against M. viscosa that resemble those of A. wodanis but with potential mechanisms that resemble those of A. fischeri. The observed effects are likely linked to the expression of genes involved in siderophore production, synthesis of antibacterial compounds, and biofilm production. These findings provide a foundation for further investigation into the specific mechanisms underlying these interactions, guided by the targeted transcriptome analysis presented here.

Supplementary Information

Below is the link to the electronic supplementary material.Supplementary Material 1 (DOCX 271 KB)