Importance of quorum sensing crosstalk in the brown alga Saccharina latissima epimicrobiome

Summary Brown macroalgae are colonized by diverse microorganisms influencing the physiology of their host. However, cell-cell interactions within the surface microbiome (epimicrobiome) are largely unexplored, despite the significance of specific chemical mediators in maintaining host-microbiome homeostasis. In this study, by combining liquid chromatography coupled to mass spectrometry (LC-MS) analysis and bioassays, we demonstrated that the widely diverse fungal epimicrobiota of the brown alga Saccharina latissima can affect quorum sensing (QS), a type of cell-cell interaction, as well as bacterial biofilm formation. We also showed the ability of the bacterial epimicrobiota to form and inhibit biofilm growth, as well as to activate or inhibit QS pathways. Overall, we demonstrate that QS and anti-QS compounds produced by the epimicrobiota are key metabolites in these brown algal epimicrobiota communities and highlight the importance of exploring this epimicrobiome for the discovery of new bioactive compounds, including potentially anti-QS molecules with antifouling properties.


Production of QS chemical inhibitors by cultivable fungal strains of the algal epimicrobiota
The extracts used in the bioassays demonstrated no impact on the growth of the biosensors.However, their efficacy in modulating QSmediated signaling was assessed.The results revealed that, among the 42 fungal strains (Figure 3; Tables S5 and S6), 61 and 76% of fungal extracts were able to inhibit short-and long-chain homoserine lactone (HSL)-based QS, respectively, in MEA and PDA culture media.AI-2-based QS was inhibited by 83% of fungal extracts in MEA culture medium and 26% in PDA.Among the extracts, Acremonium sp.strains stood out in the MEA culture medium by inhibiting all the signals involved in the QS tested.Penicillium sp. were also able to interact with QS compounds, with 80 and 92% of extracts inhibiting short-and long-chain HSL-based QS, respectively, and 84% inhibiting AI-2-based QS in MEA medium.In PDA medium, 84-92% of Penicillium sp.extracts were able to inhibit short-and long chain-HSL and 38% of AI-2-based QS.For Aspergillus sp.20-70% inhibited long and short chain-HSL based QS and 80% inhibited AI-2-based QS in MEA, whereas in PDA, 10-40% inhibited long and short chain-HSL-based QS and 0% inhibited AI-2-based QS.G. intricans sp.strain inhibited all the compounds involved in QS tested (AHLs and AI-2) in MEA medium, while P. salina was able to inhibit them in PDA.

Chemical mediators involved in QS produced by cultivable bacterial epimicrobiota
The 272 bacterial strains isolated were screened for their ability to produce and inhibit QS mechanisms (short-HSL, long-HSL, and AI-2-based QS) (Figure 4; Table S7) in marine broth (MB).The results showed that 40% of the strains were able to produce AHL.Among the Micrococcaceae and Microbacteriaceae, 100% of the strains isolated were able to produce AHLs, as were 85% of the Colwelliaceae.Surprisingly, under our cultural conditions, Moraxellaceae, Alteromonadaceae, Planococcaceae, and Granulosicoccaceae were not able to produce AHLs.However, Granulosicoccaceae produced AI-2, as did Staphylococcaceae and Halomonadaceae.Overall, 44% of the strains in the collection produced AI-2 under our culture conditions.
Regarding QS inhibitors, 100% of strains of Granulosicoccaceae, Paracoccaceae, and Planococcaceae were able to inhibit short and long-HSL-based QS.Of the 272 strains, 83.5% produced short-HSL-based QS inhibitors, compared with 65% for long-HSL-based QS.Among short-HSL based-QS-inhibiting strains, Moraxellaceae (100%), Microbacteriaceae (100%), and 98% of Pseudomonadaceae were particularly productive of this type of inhibitor, while for long-HSL inhibition Vibrionaceae were the most active (87%).Most strains in the Pseudoaltermonadaceae family produced both short and long-HSL-based QS inhibitors (98% and 87%, respectively).Overall, for each bacterial genus, at least one strain is capable of inhibiting AHLs.However, only 3% of the strains isolated were able to inhibit AI-2-based QS, notably 20% of Microbacteriaceae and 12% of Flavobacteriaceae.

Ability to form and inhibit bacterial biofilm formation
Extracts of 42 fungal strains grown in 2 culture media (MEA 75% NSW and PDA 75% NSW) were also tested for their ability to inhibit bacterial biofilms (Figure 3; Tables S5 and S6).Vibrio harveyi biofilm formation was inhibited in MEA by 20% of extracts produced by Acremonium sp. and Aspergillus sp.strains, and by 25% of those of Penicillium sp.Such an antibiofilm effect was also shown for the extract of Geomyces pannorum cultivated in MEA and for Gibberella intricans cultured in PDA.In this culture medium, 80% of Aspergillus sp., 40% of Acremonium sp., and 25% of Penicillium sp.strains were able to inhibit biofilm formation.The biofilm of the Labrenzia sp.BBCC2184 strain was inhibited in MEA by 67% of the Penicilliums metabolome and 40% of the Acremoniums metabolome, but not by Aspergillus sp.In PDA, only 20% of Aspergillus sp. was able to inhibit biofilm formation.The biofilm-forming capacities of 272 bacterial strains were also tested and 59% were able to form biofilms under our culture conditions, notably 100% of Staphylococcaceae and Halomonadaceae, and 90% of Pseudoalteromonadaceae. Bacteria from the Microbacteriaceae and Colwelliaceae families were not able to form biofilms under our culture conditions (Figure 4; Table S7).
The biofilm growth inhibiting capacities of bacterial and fungal strains were also tested.Two biofilm inhibition tests were carried out, one on the biofilm-forming capacity of Vibrio harveyi and the second on Labrenzia sp.BBCC2184.Eight percent of bacterial supernatants were shown to inhibit V. harveyi biofilm without affecting its viability, notably 33% of Rhodobacteraceae and Planococcaceae strains, but also 28% of strains from the Colwelliaceae family.Regarding Labrenzia sp.BBCC2184 inhibition, 38% of bacterial strains inhibited its biofilm formation, 72% of Pseudoalteromonas, and 62% of Shewanellaceae inhibited its biofilm formation.Among the Flavobacteriaceae family, the genus Dokdonia was shown capable of inhibiting this type of biofilm.

Cultivable diversity of the epimicrobiota of Saccharina latissima
The macroalgal surface epimicrobiome forms a complex biofilm with intricate microbial interactions. 12,15Few studies have been dedicated to the investigation of the epiphytic fungi associated with brown macroalgae, and the cultivable fungal diversity of the S. latissima epimicrobiome remains poorly explored.Our findings primarily noted the presence of epiphytic fungi belonging to the Dothideomycetes class (95%), with a predominant occurrence of Penicillium sp.(57%), Aspergillus sp.(24%), and a few Acremonium sp.(12%).Interestingly, our Culture media MEA PDA results reveal a significant disparity between the fungal epimicrobiota and the endomicrobiota, composed of Sordaryomycetes, 30 indicating a potential specificity of the fungal community based on colonized algal tissues.Other studies have focused on the epiphytic fungi of brown macroalgae, with the authors isolating Penicillium sp., Acremonium sp., Aspergillus sp., as well as Geomyces sp. and Paradendryphiella salina from Fucus sp. 32,38Penicillium sp. is a commonly occurring genus within the algal epimicrobiota, and its presence is therefore not surprising. 100egarding the cultivable bacterial diversity within the S. latissima epimicrobiota, our study revealed the presence of Pseudomonadota (Gammaproteobacteria and Alphaproteobacteria), Bacteroidota, Bacillota, and Actinomycetota.2][103] Notably, Lu et al. (2023) highlighted the significant presence of Flavobacteriaceae and Rhodobacteraceae within the core microbiota of brown macroalgae, constituting 23% and 8% of the cultivable bacterial community in our study.Furthermore, Dong and collaborators showed the predominance of Pseudomonas sp. and Psychromonas sp.within S. latissima from the Arctic, whereas, in our study, Pseudomonas sp. and Pseudoalteromonas sp.prevailed, with Psychromonas sp.being infrequently isolated. 101However, culturebased and metabarcoding approaches could lead to different but complementary biodiversity patterns.Variation in the composition and proportions of bacterial epimicrobiota among different studies can be attributed to diverse sampling locations, environmental factors (Arctic, Baltic Sea, China), algae habitat (natural or cultivable), and algal maturity. 104These variations may also reflect the capacity of macroalgae to modulate their surface epibionts based on biotic and abiotic parameters, similar to the ''gardening'' phenomenon observed in plants, where they chemically recruit microorganisms that contribute to the enhancement of seaweed fitness. 105

Diversity of fungal chemical mediators
7][108] However, the investigation into metabolites produced by macroalgal-associated fungi remains limited.As highlighted in a review by Papagianni et al., 109 each fungus exhibits unique metabolite production under distinct morphological and physiological conditions.Previous studies highlighted the production of secondary metabolites with antioxidant, antimicrobial and cytotoxic activities [33][34][35] as well as anticancer activity. 32LC-MS/MS analysis of the epiphytic fungal metabolomes yielded 573 metabolites.We focused on annotating metabolites potentially involved in microbial interactions, revealing over 99 metabolites from the fungal microbiota of S. latissima engaged in a molecular dialogue.These include antibacterial (18), antifungal (9), and antimicrobial compounds (26).Interestingly, some of the identified putative compounds are related to one species, whereas others seem to be common in different fungal strains, suggesting both specific and shared biological roles.
1][42] Our group has also previously demonstrated the key role of QS in the interspecies interactions between fungal and bacterial endophytes associated with brown algae. 615][116] Similarly, furanones and butenolides have been linked to anti-QS activities. 59,117,118Although less studied, beta-carbolines have also demonstrated potential in this regard. 87Consequently, many of the other annotated compounds in our study may possess QS or anti-QS activity, but they have never been tested for such activity. 119These results highlight the complex interactions that fungal strains can have with other microorganisms within the algal epimicrobiota through specific secondary metabolites.

Role of the epimicrobiota in regulating bacterial biofilms
Fungal strains isolated from algal epimicrobiota have a strong potential to inhibit QS mechanisms, including AHLs and AI-2 based QS.Remarkably, 95% of strains cultivated in MEA culture media possess anti-QS activity, a phenomenon previously demonstrated in the marine environment [120][121][122][123][124][125][126] and in endophytic fungi associated with brown algae. 30,61However, to our knowledge, the presence of such activities in epiphytic fungal strains of macroalgae has not been documented.Notably, our research findings also highlight the presence of bacterial biofilm-inhibiting strains in the MEA culture media, accounting for approximately 57% of the total strains.This indicates that fungi possess significant potential in regulating bacterial biofouling.To highlight the role of fungi in regulating the algal epimicrobiome, it would be particularly relevant to test the ability of fungal strains to inhibit biofilm formation by bacteria belonging to the core microbiota identified by Lu et al. 102 While the production of antifouling compounds by marine fungi has been the subject of several studies, with an overview provided by Dobretsov et al., 127 only a limited number of studies have focused on the ecological role of fungi in the epiphytic biofilm of macroalgae.Our findings suggest that epiphytic fungi may play a regulatory role in the epimicrobiota of S. latissima, particularly in shaping bacterial communities acting on QS mechanisms.In particular, the production of secondary metabolites or anti-QS enzymes could regulate bacterial QS and biofilm formation.However, the diversity of quorum quenching (QQ) enzymes produced by fungal strains has not been studied here.
We investigated bacterial strains for their communication abilities through QS mechanisms (AHLs and AI-2 based QS), and biofilm formation capacity.Results revealed that 40% of bacterial strains were able to produce AHLs (short-, medium-, and long-chain HSL), 44% of the strains communicated via AI-2 and 59% demonstrated biofilm-forming capabilities under our specific culture conditions.This substantial communication capacity suggests a crucial role of QS in maintaining the surface biofilm of brown macroalgae.7]128 Specifically, 87.5% of the bacterial strains inhibited short-chain HSL, and 65% inhibited long-chain HSL base QS.However, only 3% of the strains were capable of inhibiting AI-2 communication, and 8% exhibited biofilm formation inhibition.Moreover, the spectrum of QS compounds, which includes both activating and inhibiting agents sometimes produced by the same bacterial strains, shows remarkable diversity, with no exclusive association with a single bacterial genus.Therefore, the bacterial strains present in the S. latissima epimicrobiota exhibit the ability to influence the formation of algal biofilms.However, additional research is necessary to elucidate the precise mechanisms by which bacteria regulate the communities of epiphytic fungi.While Tourneroche et al. 61 demonstrated the involvement of AI-2 in crosskingdom signaling between bacteria and endophytic fungi in Saccharina latissima, it is plausible that other mechanisms may also contribute to these interactions.

Brown algae epimicrobiota, a source of new bioactive and antifouling compounds
The microbiota of macroalgae is a source of new biologically active compounds such as antioxidant, antibacterial, and anticancer agents. 5,6,1291][132][133][134] One of the most well-known example are the halogenated furanones extracted from the red macroalga D. pulchra, 59,135 although some others have since been identified such as 2-dodecanoyloxyethanesulfonate from Asparagopsis taxiformis. 25However, most of the bioactive anti-QS compounds previously identified, such as halogenated furanone, were also noted for their potential toxicity, reinforcing the need to pursue this exploration.Limited attention has been given to studying the anti-fouling properties of fungal strains of the brown algal epimicrobiota.Collectively, our For the 272 bacterial isolates, DNA was extracted and sequenced by the company Genoscreen.Their two pairs of primers were used: 1040R-1040F and P8-PC535.The results were received as .ab1and .seqfiles containing the unassigned sequences.Subsequently, phylogenetic affiliations were manually annotated on EZBioCloud (http://ezbiocloud.net/using BLAST. All sequences were submitted to NCBI (ITS sequences, accession numbers Database: OR461711-OR461752, 16S rRNA sequences, accession numbers Database: OR491302-OR491573) in BIOPROJECT: PRJNA1010225.All sequences (fungi and bacteria) were aligned using Muscle implemented in MEGA 11. 155,156 The alignments were manually reviewed for any mismatches, and a phylogenetic tree was constructed using a maximum composite likelihood tree with the Kimura 2-parameter Model (K2), G+I model, and nearest-neighbor-interchange and neighbor-joining methods. 156,157The reliability of each node in the tree was assessed by bootstrapping over 500 replicates.The phylogenetic tree was then exported in Newick format and processed on the iTOL online platform, where the biotest data were incorporated.

Culture of strains
Bacterial strains were inoculated with 50 mL (from cryotube) in glass tubes containing 5 mL of MB (Marine Broth, BD Difco) for 24 h at 18 C.For the AHL inhibition tests (short chain-homoserine lactones and long chain-homoserine lactones also called short chain-HSL and long chain-HSL), the bacterial strains were cultured with 1 mM (C6-HSL or 3-oxo-C10-HSL, Cayman Chemical, Ann Arbor, MI, USA) corresponding to the specific biotest being conducted.Fungal strains were inoculated with 3 pieces of agar (1 cm 2 ) from Petri dish cultures.The fungal strains were grown at 18 C for 7 days in 50 mL of PDB 75 % NSW (Potato Dextrose Broth, BD Difco) and liquid MEA 75 % NSW without agar (see previous).

Fungal and bacterial supernatants
Bacterial supernatants were obtained after sampling 2 mL of culture from glass tubes and centrifuged at 10,000 x g for 10 min.The resulting supernatant was stored at -80 C until further use for the biotests.Fungal supernatants were obtained after 7 days of growth, and 25 mL of culture was sampled and centrifuged at 10,000 x g for 30 minutes.The supernatant was then filtered through 4 mm and 0.2 mm filters.The filtered fungal supernatants were stored at -80 C for subsequent analysis.
Fungal metabolomes were extracted from the supernatants by the addition of 25 mL of ethyl acetate to 25 mL of supernatant and were then shaken overnight at 25 C. Subsequently, the ethyl acetate phase was carefully collected using a Pasteur pipette and transferred to preweighed hemolysis tubes.The collected ethyl acetate phase was evaporated to dryness and the resulting dry mass of each sample was recorded.Sterile culture media (liquid MEA and PDB) was also extracted using the same method.The samples were then stored at -20 C until further analysis.

LC-MS-based metabolomic analysis of fungal strains
The dried extracts of fungal supernatant were solubilized in methanol at a concentration of 10 mg.mL -1 and injected in High Performance Liquide Chromatography coupled to tandem Mass Spectrometry (HPLC-MS/MS).The analysis was performed in one batch and in a random sequence, with five samples followed by three quality controls (QCs), according to established protocols. 30,61The separation was carried out using a C18 Acclaimä RSLC PolarAdvantage II column (2.1 3 100 mm, 2.2 mm pore size; Thermo Fisher Scientific, United States) connected to a Dionex Ultimate 3000 HPLC system.The column was coupled to a Maxis IITM QTOF mass spectrometer (Bruker, United States) equipped with an electrospray ionization source.The flow rate was set at 300 mL.min -1 .The MS parameters were 3.5 kV of electrospray voltage, 35 psi of nebulizing gas (N 2 ) pressure, drying gas (N 2 ) flow rate of 8 L.min -1 , and 200 C of drying temperature.The mobile phases were water (0.1 % formic acid) and acetonitrile (0.1 % formic acid, solvent B) following a gradient of Solvent B at 5, 50, 90, and 5 % for 0, 9, 15, and 21 min, respectively.
LC-MS/MS data were analyzed using DataAnalysis (version 4.4 Bruker Daltonik GmbH) and converted to ''.mzXML''.The LC-MS/MS data were preprocessed on MzMine 3.2.8(Pluskal et al., 2010; Schmid et al., 2023) with a noise level of 1E3 for MS1 and 1E2 for MS2, and the ''ADAP chromatogram'' was built with a minimum group size of scan of 2, a group intensity threshold and minimum highest intensity at 3E3. 144 The chromatograms were resolved with ''ADAP Module Disclaimer'' with S/N threshold at 10, minimum feature height at 1, coefficient/area threshold at 10, 0.02-1.00for peak duration range and 0.10-0.50for RT wavelet range.The data were then deisotoped and filtered to only keep peaks with MS2 scans and a ''RANSAC'' alignment was performed.After alignment, the compounds in the culture media (MEA and PDB) were subtracted from the matrix using the "blank subtraction" tool.This step eliminates the compounds present in the culture media from the dataset.The data were then exported to Feature-Based Molecular Networking of the platform Global Natural Products Social Molecular Networking (FBMN-GNPS) to build a molecular network with a cosine of 0.7 and 4 minimum matched fragment ions, precursor ion mass tolerance (PIMT) and fragment ion mass tolerance (FIMT) at 0.05 Da. 147,158 Metabolites were annotated with various GNPS tools, including Library Search, Dereplicator, Dereplicator +, 159 and Sirius 5.6.3 1608][169] The annotations were thus added to the network using Cytoscape 3.10.0software. 148
V. harveyi MM32 reporter strain was grown at 30 C in marine broth medium supplemented with kanamycin (30 mg.mL -1 , Sigma-Aldrich) and chloramphenicol (10 mg.mL-1, Sigma-Aldrich).In a 96-well plate, 180 mL of biosensor (diluted at 1/100 in growth media) was added to wells containing 20 mL of bacterial strain supernatants (in triplicate) for AI-2 production testing.All tests were performed with two negative controls: (i) sterile MB alone and (ii) biosensor supplemented with supernatant culture medium.The positive controls were a culture supernatant from strain BBCC292 (MOLA292) and BBCC707 (MOLA707), known to be a strong AI-2 producer. 136After 24 hours at 30 C, cell growth (OD 630nm ) and luminescence (OD 540nm ) of MM32 were measured using a Victor 3 spectrofluorometer (Pelkin Elmerâ).

Quorum quenching bioassays: Detection of AI-1 and AI-2 inhibition
To measure AI-1 inhibition by culture supernatants or extracts, the biosensor cultures were supplemented with 3 mM AHLs, 3-oxo-C10-HSL was added for F117 and MT102 and C6-HSL for CV026 and MT102, which can detect the inhibition of both types of AHLs.In the microplate, 150 mL (for bacterial test) or 180 mL (for fungal test) of biosensors (diluted at 1/50 in growth media) supplemented with AHLs were added, followed by 50 mL of bacterial culture supernatants or 20 mL of 1 mg.mL -1 fungal extract (dissolved in 10% DMSO).The microplates were incubated, and after 24 h, the same measurements as described before were performed.However, the inhibition of the biosensor signal was calculated to determine the percentage of AI-1 inhibition.Two negative controls were used for the tests: (i) biosensors without AHLs with the addition of culture medium or culture medium extract dissolved in dimethylsulfoxyde (DMSO, 10 %) and (ii) biosensors with AHLs and the addition of sterile culture medium or culture medium extract diluted in 10% DMSO.After 24h hour, the same measurements were made as above.
For the detection of AI-2 inhibition, Vibrio harveyi MM32 biosensor was used in the presence of 1 mM of 4,5-dihydroxy-2,3-pentanedione (DPD, a precursor of AI-2 from Rita Ventura's research Group at ITQB, Oeiras, Portugal). 137In a microplate, 20 mL of culture supernatant or fungal extract was added to 180 mL of biosensor (diluted at 1/100 in growth media) and DPD.The negative controls were (i) biosensor without DPD and (ii) biosensor with DPD with the addition of the culture medium of the tested supernatants or the extracted culture medium of the extracts.The positive control used was the active extract of the fungal strain Microsphaeropsis olivacea AN329T dissolved in DMSO 10%, as highlighted in. 61After 24h of incubation, the measurements as in the section Detection of AI-1 and AI-2 production, were performed.

Evaluation of bacterial biofilm-forming capacity
The biofilm-forming capacity of bacterial strains in the collection was assessed following previously described protocols. 136,172The strains were grown in microplates at 18 C for 5 days.After measuring the optic density OD 630nm with a Victor 3 spectrofluorometer (Perkin Elmerâ), the microplates were rinsed with Phosphate Buffered Saline (PBS), dried, and then treated with crystal violet (0.2 % Sigma, diluted in 20 % ethanol) for 15 min in the dark at room temperature.Subsequently, the microplates were rinsed with water (three times) and the wells were decolorized with a decolorizing solution (33 % glacial acetic acid) to reveal the presence of biofilm.The OD 540nm was measured with a spectrofluorometer.

Biofilm inhibition
To evaluate the capacity of microbial strains to inhibit biofilm formation, two reference strains were used: Vibrio harveyi and Labrenzia sp.BBCC2184, which were obtained from our laboratory collections.These strains were grown in 48-well microplates (900 mL) at 25 C in MB, in the presence of 100 mL of bacterial supernatants for 48h before biofilms revelation (as described above).For inhibition test of biofilm formation by fungal metabolomes, 950 mL of bacterial strains was incubated with 50 mL of fungal extract at a concentration of 1 mg.mL -1 in DMSO (10 %) to limit the potential toxicity of fungal strains.The biofilm inhibition test was performed following the same procedure as that for biofilm formation (as described above).

Figure 1 .
Figure 1.Fungal and bacterial cultivable diversity from the epimicrobiota of S. latissima (A) fungal diversity at the genus level of the 42 strains isolated (B) bacterial diversity at the genus level of the 272 strains isolated.

Figure 2 .
Figure 2. Metabolites putatively involved in microbial interactions produced by fungal strains isolated from the S. latissima epimicrobiota Molecular subnetworks obtained after extraction of 100 metabolites potentially involved in interactions between microorganisms.The complete network containing 573 metabolites is shown in Figure S1 and annotations are shown in TablesS3 and S4.

Figure 3 .
Figure 3. Bioactivities of fungal strains in MEA and PDB media Phylogenetic tree of fungal strains and biotest results for each strain in MEA and PDA culture media.The numbers on the branches correspond to the genetic distance between strains.

Figure 4 .
Figure 4. Bioactivities of bacterial strains in MB media Phylogenetic tree of bacterial strains and biotest results for each strain in MB culture medium.The numbers on the branches correspond to the genetic distance between strains.