Probiotics useful in aquaculture

Abstract

The present invention relates to bacterial probiotics useful in aquaculture. Specifically, it discloses a selection of non-hemolytic, proteolytic, stress-tolerant bacterial strains with enhanced quorum- quenching capabilities. Moreover, the method allows detection of novel species of bacteria whose probiotic attributes could be massive for aquaculture. From the isolated bacteria, selected species were tested for their probiotic efficacy in Nile tilapia (Oreochromis niloticus), a key fish species in global finfish aquaculture. The species demonstrated significant efficacy in improving the health and growth performance of Nile tilapia juveniles. Specific growth rates and weight gain, and improved feed conversion ratios without adverse effects on fish health were observed. The gut microbiome was positively influenced by promoting beneficial bacteria while reducing harmful (opportunistic) species, leading to improved immune modulation. Survival rates were significantly boosted during a pathogenic challenge, and key immune genes were upregulated, indicating stronger local immune defenses. These findings show that the selected probiotic species offer substantial benefits for growth, feed efficiency, and disease resistance, emphasizing their potential for widespread application as probiotics in aquaculture.

Description

[0001] Probiotics useful in aquaculture

[0002] Technical field of invention

[0003] The present invention relates to bacterial probiotics useful in aquaculture. Specifically, it discloses, selected species which were tested for their probiotic efficacy in Nile tilapia (Oreochromis niloticus), a key fish species in global finfish aquaculture. The species demonstrated significant efficacy in improving the health and growth performance of Nile tilapia juveniles. Specific growth rates and weight gain, and improved feed conversion ratios without adverse effects on fish health were observed. The gut microbiome was positively influenced by promoting beneficial bacteria while reducing harmful (opportunistic) species, leading to improved immune modulation. Survival rates were significantly boosted during a pathogenic challenge, and key immune genes were upregulated, indicating stronger local immune defenses. These findings show that the selected probiotic species offer substantial benefits for growth, feed efficiency, and disease resistance, emphasizing their potential for widespread application as probiotics in aquaculture.

[0004] Background art

[0005] Pathogenic bacteria use quorum sensing (QS) to regulate virulence gene expression and enhance antimicrobial resistance (AMR) via cell-cell communication with autoinducers (AIs) such as N-acyl homoserine lactones (AHLs) (Defoirdt, 2018). The disruption of this mechanism of QS, also known as quorum quenching (QQ), can inhibit virulence gene activation and alter bacterial infectious cycles, thereby controlling diseases caused by such pathogens (LaSarre & Federle, 2013). Unlike bactericidal antibiotics, the mechanism of QQ does not significantly affect bacterial growth, minimizing selection pressure for AMR phenotypes, thus making it a promising antivirulence approach (Ghanei-Motlagh et al., 2021). As highlighted by Kuebutornye et al. (2019), certain groups of bacteria such as Bacillus spp., that have demonstrated a high probiotic potential in aquaculture, exhibit these QQ properties. Probiotics are live microorganisms that offer various benefits such as enhanced immunity and resistance to pathogens (El-Saadony et al., 2021). With increasing concerns on the use of antimicrobial agents in aquaculture, probiotic bacteria emerged as a safer alternative (Lu et al., 2022; Serwecinska, 2020). In finfish aquaculture, most used probiotic bacteria belong to genus Lactobacillus, Lactococcus, Pediococcus, Bacillus, Arthrobacter , Burkholderia, Enterobacter , Enterococcus, Vibrio, Pseudomonas, Rhodoseudomonas, Roseobacter, Aeromonas, Shewanella, and Bifidobacterium (Kuebutornye et al., 2019; Yousuf et al., 2023). Probiotics used in aquaculture are allochthonous in nature, isolated from terrestrial animal sources, disregarding their inadequate efficacy for aquatic animals. Consensually, probiotic bacteria isolated from aquatic organisms, or their environment will colonize and establish themselves relatively faster, are more stable, resilient, persistent, and robust. Thus, indigenous QQ bacteria, isolated from aquatic hosts and their environments, already adapted to local conditions (e.g., tolerance to salt and high temperatures) offer inherent resistance to environmental stressors and exhibit improved efficacy against local emerging aquaculture pathogens.

[0006] The present invention discloses new indigenous probiotics for use in aquaculture and the use of these probiotics for improving the production parameters and gut health, and, for offering protection against aquaculture pathogens. Brief description of figures

[0007] Figure 1. AHL degradation by the strains under coculture with an AHL producer (Aeromonas hydrophila LVS3) after 48 hours in LBio at 28°C. Bars represent the mean AHL concentration after 48 hours of coculture. Bars with different superscripts from LVS3 only (negative control) represent treatments that had a significantly lower AHL concentration than the negative control. Error bars represent the standard deviation from the mean AHL concentration for the different treatments.

[0008] Figure 2. Salt tolerance and growth (OD550) of the bacteria at different salt concentrations after 24 hours. Panels A, B, C and D represent data for salt concentrations 1%, 2%, 3% and 4% respectively.

[0009] Figure 3: Tolerance and growth of the strains at different pH levels (3, 4, 5, 6, 7, 8 and 9) after 24 hours of incubation at 28°C. Bars represent mean absorbances at 550nm, and the error bars are for the standard deviations from the mean. No statistical comparisons were made between treatments. Figure 4. Relative substrate utilization by the strains in the Biolog Gen III microplate.

[0010] Figure 5: The Specific Growth Rate (SGR) (Fig. 5A) and Feed Conversion Ratio (Fig. 5B) of Nile tilapia juveniles fed on the different feed treatments. Data is presented as means ± standard deviations for the different treatments after 60 days of feeding. Bars with different superscripts depict significant differences between treatments.

[0011] Figure 6: Immune gene expression in the midgut (Fig. 6A) and spleen (Fig. 6B) of Nile tilapia juveniles after 60 days of feeding on a commercial diet supplemented with two local putative probiotic strains Lysinibacillus fusiformis LFUG and Priestia megaterium PMUG01 at two concentrations T1 - IxlO6CFU.g1and T2 - IxlO8CFU.g1. The three genes included Interleukin 8 (IL8), Interleukin 6 (IL6), Complement component C3 (C3), and Interleukin 1-beta (IL1B). Samples were collected from 3 fish per treatment and gene expression was compared between probiotic treatments and the control treatment. Data is presented as means ± standard deviation and treatments with different superscripts (Fig. 6C) in the same row signify significant differences.

[0012] Figure 7. The mean survival rate of juvenile Nile tilapia 14 days post infection with a clinical and virulent strain of Providencia sp. Data is presented as means+standard deviations and bars with different superscripts denote treatments with significant differences between each other. PC and NC represent the Positive and Negative control treatments. Figure 8. Type Strain Genome Server (TYGS) analysis of a putative probiotic isolate from sediment: Kocuria crassamentum species nova KSNUG showing the Genome BLAST Distance Phylogeny (GBDP) tree based on genome sequences of available reference and type strains. They were supplemented with our query strain Kocuria crassamentum species nova KSNUG (in bold) and representative genomes of species within the Kocuria genus including its closest hit Kocuria rhizophila strain TA68 GCA_003667225.1 (in bold; dDDH: 61% and wgANI: 94.97%).

[0013] Figure 9. Type Strain Genome Server (TYGS) analysis of a putative probiotic isolate from sediment: Heyndrickxia crassamentum species nova HSNUG (in bold) showing the Genome BLAST Distance Phylogeny (GBDP) tree based on genome sequences of available reference and type strains. They were supplemented with our query strain Heyndrickxia crassamentum species nova KSNUG (in bold) and representative genomes of species within the Heyndrickxia, Mesobacillus, and Weizmannia genera. Description of invention

[0014] The present invention discloses selective bacterial isolation from aquaculture environments (i.e., here applied on sediments), phenotypic screening, molecular identification, and in vivo biosafety assessment. The present invention thus relates to isolating indigenous bacterial strains exhibiting enhanced QQ capabilities and enhanced survival under elevated temperature (i.e., 85°C) and high salinity (i.e., 35 g.L-1NaCl) conditions. The disclosed method allows to obtain candidate probiotic strains that could be further exploited for their QQ-based anti-virulence potential. It includes a selective pressure for temperature and salt tolerance to exclude virulent Bacillus cereus strains, known for their involvement in food poisoning incidents and thus of importance for public health concerns. Furthermore, to reassure the probiotic nature of these indigenous bacteria, a phenotypic screening with a focus on key traits relevant for their probiotic application such as QQ rates, hemolysis, and proteolytic activity was conducted to select the most promising candidates for probiotic use in aquaculture. The present invention also relates to an in-depth molecular analyses using Oxford Nanopore Technologies’ long-read whole genome sequencing (WGS) for bacterial identification. The present invention finally assesses the biosafety of the strains in axenic brine shrimp (Artemia franciscana), following a previously established toxicity testing model.

[0015] Strains were selected and their probiotic efficacy was evaluated in Nile tilapia (Oreochromis niloticus), a non-limiting example of fish species. The feeding trial was conducted using two concentrations (T1 and T2: IxlO6and IxlO8CFU.g1of feed). Nile tilapia juveniles were fed a commercial feed supplemented with these strains for 60 days under laboratory conditions. By the end of the trial, fish fed with both strains, regardless of concentration, showed significant weight gain (6.429±1.590 - 8.034±2.316 grams) compared to the control group (5.417±1.988 grams). Importantly, no adverse effects were observed in the fish's condition, visceral somatic index (VSI), hepatosomatic index (HSI), or survival rates across all treatments. The specific growth rate (SGR) and feed conversion ratio (FCR) also improved in fish fed with the probiotic-supplemented diets. SGR values ranged from 2.616±0.318 to 3.027±0.490 %.day compared to the control group's 2.428±0.285 %.day while FCR values improved to 1.050±0.172 - 1.167±0.158, compared to the control group’s 1 ,282±0.192. These results indicate a faster growth rate and better feed utilization in fish fed the probiotic diets. No significant changes were found in the chemical composition of the feed or the fish, suggesting the strains did not alter the nutritional composition of the diet or the fish. However, the gut microbiome was positively influenced, with an increase in symbiotic bacteria and a reduction in harmful (opportunistic) species, which in turn promoted growth and immune modulation. The survival rate of the fish also significantly increased (83.3±11.5% -90.0±10.0%) following a pathogenic challenge with a clinical strain of Providencia sp., a non- limiting example of a pathogenic infection, compared to the positive control (34.8±21.7%). Local immune responses were enhanced, as evidenced by the upregulation of key immune genes (IL8, IL6, C3, and IL1B), indicating a strong local immune defense. Also, the systemic immune responses were modulated to prevent excessive inflammation, thus reducing the risk of tissue damage in probiotic treated groups.

[0016] The results from this Nile tilapia trial demonstrate the potential of probiotic strains isolated and characterized using this method to significantly improve growth, boost immunity, enhance feed utilization, and increase overall health and resistance to infectious diseases in aquaculture species like Nile tilapia.

[0017] Hence, the present invention relates to new and / or effective probiotics, more in particular to isolate and select quorum quenching, heat and salt tolerant, non-haemolytic and proteolytic probiotics useful in aquaculture with reduced likelihood of co-isolation of virulent Bacillus cercus.

[0018] A method to isolate said probiotics comprises:

[0019] a. Sample (sediment) collection, airdrying and storage.

[0020] b. Sediment treatment with physiological saline (PS) prior to bacterial isolation. c. Bacterial strain isolation using buffered Minimal Medium (MM), supplemented with N-hexanoyl homoserine lactone (HHL) as source of carbon and nitrogen. d. Selection of heat tolerant isolates by heat inactivation (85°C, 20 min) while minimizing co-isolation of virulent B. cereus.

[0021] e. Subculturing (3 cycles) with nutrient enrichment introduced in the second cycle by supplementing the MM with LB20 broth (1%), while the third cycle was conducted without additional enrichment.

[0022] f. Selecting salt tolerant (NaCl - 35g.L-1) isolates by salt treatment on LB35 agar. g. Isolate selection based on colony morphology and general appearance.

[0023] h. Gram-staining to confirm isolation of Gram-positive, heat and salt tolerant, and potentially QS signal molecule utilizers.

[0024] i. Isolate selection by:

[0025] i. Confirmation of QQ by degrading HHL and selecting significant HHL degraders.

[0026] ii. Testing haemolytic activity to select non-haemolytic (safe) isolates. iii. Testing proteolytic activity to select for exogenous digestive enzyme producers with potential to enhance feed utilization. iv. Molecular identification of the strains by whole genome sequencing (WGS) to confirm the isolated taxa and guarantee absence of virulent B. cereus.

[0027] The ‘sample (sediment) selection in step a)’ can relate to any sample, but preferably autochthonous (from aquatic organisms and their environment) such as water, and farmed species themselves, but specifically relates to wet sediment samples from the surface layer of Nile tilapia (Oreochromis niloticus) earthen ponds in Uganda (National Aquaculture Research and Development Centre, part of the National Agricultural Research Organization), air-dried in open air, and used for the isolation.

[0028] With ‘sediment treatment with physiological saline’ in step b) is meant soaking the ‘sample’ in sterile physiological saline (PS) for an appropriate time (30 minutes for sediment) to obtain the inoculum for the first isolation cycle in minimal medium (MM).

[0029] The terms ‘bacterial strain isolation using sterile buffered Minimal Medium (MM), supplemented with N-hexanoyl homoserine lactone (HHL)’ in step c) is meant dispensing the suspensions from soaked ‘samples’ in PS, into MM supplemented with buffered HHL (can be any Quorum Sensing Signal Molecule) as a sole source of nitrogen and carbon. Incubation is done at conditions suitable for targeted probiotic strains.

[0030] The terms ‘selection of heat tolerant isolates by heat inactivation (85°C, 20 min) while minimizing co-isolation of virulent B. cereus’ of step d) is meant heat-treating aliquots from cultures from the previous step for 20 minutes at 85°C to minimize the likelihood of (co-)isolating toxin producing Bacillus cereus.

[0031] The terms ‘ Subculturing (3 cycles) with nutrient enrichment introduced in the second cycle by supplementing the MM with LB20 broth (1%), while the third cycle was conducted without additional enrichment’ in step e) is meant inoculating heat-treated aliquots of cultures from the first cycle into new sterile MM, supplemented with buffered HHL (and or any other QS signal molecule) and sterile Luria Bertani (LB20) broth (and or any other suitable media for targeted probiotics) for the second incubation cycle (apply similar incubation conditions as in cycle 1). For the final cycle (cycle 3), inoculate aliquots of cultures from cycle two for a third cycle, similar to cycle 1 (without LB20 broth / media addition) and incubate under similar conditions as in cycle 1 and 2.

[0032] With the terms ‘Selecting salt tolerant (NaCl - 35g.L-1) isolates by salt treatment on LB35 agar’ in step f) is meant plating (culturing) the heat-treated cultures that have undergone all the 3 cycles of subculturing above on sterile Luria Bertani agar (or any other suitable media) plates enriched with 35 g.L'1NaCl (LB35) and incubating for 4 days at 28°C while monitoring colony formation and cell density (CFU.rnL'1). Plates that exhibit growth imply that salt tolerant strains are the ones that grew on such plates and are used for isolate selection.

[0033] With terms ‘Isolate selection based on colony morphology and general appearance’ in step g) is meant selecting the salt tolerant isolates that grew in the previous step, based on colony morphological differences for further screening.

[0034] With the terms ‘Gram-staining to confirm isolation of Gram-positive, heat and salt tolerant, and potentially QS signal molecule utilizers’ of step h) is meant subjecting the selected strains to a Gram-staining protocol to confirm that they are Gram-positive, aligning with the quorum quenching abilities and stress (heat & salinity) tolerance (targeted in the previous steps), used in the selection process.

[0035] With the step ‘Isolate selection by’ of step i) is meant further screening of isolates with better quorum quenching capabilities, thus, higher anti-virulence potential (sub-step i). This can apply to any QS quantification method and for this case, a plate diffusion method to quantitatively detect exogenous HHL degradation is used and is described by Defoirdt et al., 2011). Also, selecting of non-haemolytic strains (safe to use without concerns of profiting from host wounds and lesions (sub-step ii). Any method can be used though the one described by Kuebutomye, Lu, et al., 2020 is used for this case). Still, the capability for producing exogenous digestive enzymes such as proteases to enhance substrate utilization in feeds and the environment (water quality improvement) (sub-step iii) is meant as described by C. H. Liu et al. (2009) even though any other method can be used. Finally, sub-step iv) is meant identifying the remaining strains which fulfill the above criteria (applies for several molecular techniques although whole genome sequencing using the ONT long-read GridlON sequencing platform (Vereecke and Van Hoorde et al., 2023) applies here, to confirm the absence of virulent B. cereus, fulfilling a legislative requirement of using known strains for probiotic applications, and guiding further phenotypic characterization initiatives.

[0036] Overall, this method is superior, concise, efficient, and non-invasive, focusing on the isolation of indigenous bacterial strains with significant probiotic and antivirulence potential. The isolated strains demonstrated notable tolerance to heat and salt, indicating their adaptability to climate change and variability, as well as their suitability for use in fresh, brackish, and marine aquaculture. The protocol is designed to reduce the risk of co-isolating virulent Bacillus cereus, thereby safeguarding human biosafety and minimizing the potential for fish and water-borne zoonoses. Unlike many commercial probiotics for aquaculture, which are often isolated from terrestrial environments and exhibit limited efficacy, this method yields autochthonous strains with superior resilience and adaptability to dynamic aquatic ecosystems. Remarkably, two novel strains were also isolated using this protocol, and thus, more can be discovered if scaled up. The protocol integrates both phenotypic and genomic techniques to evaluate the probiotic potential and safety of bacterial strains, with additional biosafety verification in crustaceans, a crucial group in aquaculture. Furthermore, this approach contributes to the global challenge of developing nonantibiotic therapeutic and prophylactic alternatives, helping to mitigate the negative impacts of antibiotics on hosts and the environment.

[0037] The present invention further discloses a method or protocol for phenotyping and safety assurance of the isolated and selected probiotics. The defined protocol is as follows:

[0038] a. Testing the growth rate of the isolated probiotics as strains with higher growth rates show a competitive advantage for gut colonization.

[0039] b. Testing stress tolerance (pH and salt concentrations) of the isolated probiotic strains.

[0040] v. If needed, pH can be adapted to the relevant fish species.

[0041] vi. If needed, salt concentration can be adapted to the relevant indigenous environment

[0042] c. Testing the strains potential to bind through their glycans to host cell lectins. d. Metabolic fingerprinting of isolated probiotic strains using diverse substrate compositions, depicting the complexity of (local) aquaculture feeds.

[0043] e. Testing the isolated probiotics for sporulation as spore-forming probiotics show extended viability in high stressful environments.

[0044] f. Testing the isolated probiotics for antibiotic susceptibility.

[0045] g. Validating non-toxicity of the isolated probiotics in zebrafish embryos.

[0046] h. Validating gut colonization by the isolated probiotics in zebrafish embryos With ‘Testing the growth rate of the isolated probiotics as strains with higher growth rates show a competitive advantage for gut colonization’ in step a) is meant assessing the growth rate of probiotic bacteria using a suitable medium and method. For this case, LBw broth in sterile nontreated transparent 96 well bacterial culture plates is used, and the OD at 550 nm (hourly for 36 h) for all isolates is measured with a Tecan Infinite 200 microplate reader set at 28°C. A modified Gompertz model is used to determine the growth characteristics of the isolates. With ‘Testing stress tolerance (pH and salt concentrations) of the isolated probiotic strains’ in step b) is meant examining the tolerance and growth of our probiotic strains under varying salt (NaCl) concentrations and pH levels in LB broth. Can be applied to all bacteria using suitable media with pH and salinity adjusted accordingly. In this case, salt (NaCl) concentrations (1%, 2%, 3% and 4%) and pH levels ranging from 3 to 9 are used. Incubation conditions are accordingly adjusted for targeted strains.

[0047] With ‘Testing the strains potential to bind through their glycans to host cell lectins’ in step c) is meant determining the glycan architecture of the probiotics, which plays an important role in bacteria-host binding specificity, regulating their colonization, adhesion, and pathogen exclusion. Can be applied with any suitable protocol although we used a protocol by Bossier and Wongprasert (2019) with some modifications and a CytoFLEX flow cytometer system (Beckman Coulter’s Life Sciences, France).

[0048] With ‘Metabolic fingerprinting of isolated probiotic strains using diverse substrate compositions, depicting the complexity of (local) aquaculture feeds’ in step d) is based on a positive correlation between gut microbial diversity & functional relationships, and metabolic profiles of bacteria. It is thus meant studying the metabolic fingerprints of our isolates using any suitable method. (Biolog Gen 111 MicroPlates (BiOLOGInc. Hayward, CA, USA) featuring diverse substrate compositions according to the manufacturer's guidelines is used in this study).

[0049] With ‘Testing the isolated probiotics for sporulation as spore-forming probiotics show extended viability in high stressful environments’ in step e) is meant for examining the stress tolerance potential of the probiotics, which is associated to production of spores, which tolerate numerous stressors in the environment and within the host and allow for extended shelf life without necessitating sophisticated preservation techniques. Examining bacterial isolates for sporulation applies to all suitable media and methods though sterile sporulation media (SM) as described by Nicholson and Setlow (1990) is used herein.

[0050] With ‘Testing the isolated probiotics for antibiotic susceptibility’ in step f) is meant preventing further development of bacteria that could potentially transfer AMRGs to other bacteria of veterinary and public health significance since the presence of transmissible antimicrobial AMRGs, is paramount when evaluating candidate probiotic strains especially those intended for food-grade applications. This applies to all phenotypic and genetic techniques of assessment. For this method, we followed a disk diffusion method described by Kirby-Bauer (2009) with Mueller Hinton Agar and the comprehensive antibiotic resistance database (CARD) to test the antibiotic susceptibility of the strains. With ‘Validating non-toxicity of the isolated probiotics in zebrafish embryos’ in step g) is meant for screening for in vivo biosafety of the probiotics and applies to all applicable in vivo models. This is a pivotal determinant of safety before applying probiotics to hosts. Evan though it applies to all suitable methods, a bath-treatment to axenic Zebrafish embryos (Danio rerio AB line, wildtype, 3 dpf) with the candidate probiotic strains at a cell density of IxlO7CFU.mL1(28°C, 3 D) (As described by Russo et al., 2015).

[0051] With ‘Validating gut colonization by the isolated probiotics in zebrafish embryos’ in step h) is based on one of the primary goal of probiotics, which is to establish colonization of the gut, to deliver intended health benefits, including enhanced immunity and growth performance. Applies to all applicable methods and aquatic hosts. For our case, Zebra fish (Danio rerio AB line wild type) embryos are used to visualize and verify probiotic bacteria colonization as described by Nayak etal. (2023).

[0052] Overall, this comprehensive approach, combining phenotypic and biosafety evaluations, represents a cost-effective pre-screening strategy for candidate probiotic strains. This screening strategy is further linked to further phenotypic and safety characterization of the QQ, non-haemolytic, and proteolytic strains, after their identification by molecular tools and techniques. Moreover, it gives virgin phenotypic profiles of the novel strains isolated and identified in the latter method to select and screen. Furthermore, it provides an expedited method for selecting promising probiotics before proceeding to more complex and riskier in vivo studies. Importantly, inclusion of biosafety evaluation in another representative aquaculture group (Finfish - Zebrafish) gives further biosafety assurance to both representatives of all categories of fed aquaculture (Anemia - invertebrates in the latter method to select and screen & vertebrates - Zebrafish) which ensures the safety of the host organisms, human consumers of aquatic products, and the environment. Demonstration of gut colonization by the strains is paramount for self-sustenance, a limitation by most commercially available probiotics for aquaculture.

[0053] The present invention further describes a combinatory use of the probiotic isolation and selection method (as described above) with the defined protocol for probiotic phenotyping and safety assurance (as described above). The present invention relates to two novel bacterial species, deposited under the Budapest Treaty in the BCCM / LMG Bacterial Collection, located at K.L. Ledeganckstraat 35, 9000 Ghent, Belgium, under accession numbers LMGP-33748 and LMGP-33747^

[0054] Moreover, the present invention relates to a novel micro-organism showing at least 93 percent homology to the 16S rRNA sequence of said micro-organism having accession number LMGP-33748 or LMGP-33747. The term ‘at least 93 percent homology to the 16 rRNA’ refers to 93, 94, 95, 96, 97, 98 or 99 percent homology to the 16S rRNA sequence of said micro-organism having accession number LMGP-33748 or LMGP-33747.

[0055] The first isolate, designated Kocuria crassamentum sp. nov. (KSNUG) and deposited under accession number LMGP-33748, is characterized by a 2.7 megabase pair (Mbp) genome with a GC content of 70.6%, exhibiting 100% 16S rRNA gene sequence identity with Kocuria rhizophila based on 24 exact rRNA gene matches. Whole-genome comparison with the reference strain Kocuria rhizophila TA68 (GCA_003667225.1) revealed a digital DNA-DNA hybridization (dDDH) value of 61% and a whole-genome average nucleotide identity (wgANI) of 94.97%, confirming its taxonomic distinctiveness below the recognized species delineation thresholds of (>95% wgANI and >70% dDDH (d4%)).

[0056] The second isolate, designated Heyndrickxia crassamentum sp. nov. (HSNUG) and deposited under accession number LMGP-33747, possesses a 4.8 Mbp genome with a GC content of 41.5%, showing 40% 16S rRNA gene sequence identity with Heyndrickxia acidicola and 5 exact rRNA gene matches. Comparative genomic analysis with Heyndrickxia acidicola DSM 14745 (GCA_001636425.1) yielded a dDDH value of 28% and a wgANI of 70.12%, demonstrating clear genomic differentiation consistent with species-level novelty under the same delineation criteria (>95% wgANI and >70% dDDH (d4%)).

[0057] Accordingly, the two isolates represent novel bacterial species, herein designated as Kocuria crassamentum species nova (sp. nov. ) and Heyndrickxia crassamentum species nova (sp. nov. ), respectively, both distinct from previously described members of their respective genera as determined by the internationally accepted method described hereafter.

[0058] Determining the percent homology among 16S rRNA sequences of different strains, and novel species delineation using dDDH and wgANI are techniques well-known in the art.

[0059] The present invention further relates to the usage of a microorganism as described above as a probiotic in aquaculture to improve growth performance, feed utilization, water quality, immune response, disease resistance, and overall health of hosts. Extensive in vitro and in silico characterizations were conducted to prove (augment) their probiotic and biosafety profiles.

[0060] Kocuria crassamentum sp. nov. (KSNUG; LMGP-33748) and Heyndrickxia crassamentum sp. nov. (HSNUG; LMGP-33747) exhibited a combination of desirable probiotic and biotechnological attributes that support their suitability as functional microbial candidates for aquaculture and related applications.

[0061] Specifically, the strains were found to be non-hemolytic, confirming their biosafety and absence of cytolytic activity even under conditions simulating host tissue damage. They demonstrated strong proteolytic activity, indicating potential for enhancing protein digestion in aquafeeds, an essential function given the high cost and nutritional importance of dietary proteins.

[0062] Both strains also exhibited pronounced heat tolerance, suggesting stability during high-temperature feed production processes, and broad salt and pH tolerance, reflecting adaptability and persistence potential to diverse environmental and host conditions.

[0063] Importantly, both isolates displayed quorum-quenching activity, evidenced by their ability to degrade N-hexanoyl homoserine lactone (HHL, 5 mg-L'1), a concentration significantly higher than typically detected in biofilms or aquaculture systems. This indicates strong potential for antivirulence applications, where pathogenic bacteria can be rendered harmless without affecting their proliferation, thereby avoiding the selection pressure for antibiotic resistance.

[0064] Further analyses revealed high lectin-glycan binding indices, suggesting a robust host adhesion capacity and potential for competitive exclusion of pathogens. They also exhibited broad substrate utilization profiles (in Biolog plates), encompassing common and complex compounds present in aquafeeds and aquatic environments, implying roles in nutrient assimilation and water quality improvement.

[0065] Additionally, Heyndrickxia crassamentum sp. nov. (HSNUG) was found to be spore-forming, conferring enhanced stress tolerance, ease of handling, and extended shelf life in formulation. Fluorescent tagging experiments further demonstrated its ability to colonize the gut of axenic zebrafish embryos following a 48-hour bath exposure, confirming host interaction potential. At the genomic level, both strains possess genes associated with amino acid and carbohydrate transport and metabolism, secondary metabolite biosynthesis, defense mechanisms, and carbohydrate-active enzymes (CAZymes) responsible for generating growth-promoting and antimicrobial compounds. These genomic traits collectively underpin their probiotic potential and functional safety. Moreover, toxicity assays confirmed the non-toxic nature of both isolates in Artemia nauplii (instar II stage) and zebrafish embryos (3 days post-fertilization), supporting their environmental and host safety.

[0066] Taken together, these in vitro and in silico findings demonstrate that the deposited strains possess multiple probiotic, anti-virulence, and biosafety attributes, justifying their classification as novel, beneficial bacterial species suitable for safe use in aquaculture and related biological applications. Examples

[0067] Example 1: non-invasive method to isolate and select quorum quenching, heat and salt tolerant, non-haemolytic and proteolytic probiotics useful in aquaculture with reduced likelihood of coisolation of virulent Bacillus cereus

[0068] Materials and methods

[0069] Sample collection and isolation of candidate probiotic bacteria

[0070] Wet sediment samples were collected with a plastic shovel from the surface layer of Nile tilapia (Oreochromis niloticus) earthen ponds at Uganda’s National Aquaculture Research and Development Centre (ARDC), part of the National Agricultural Research Organization (NARO). Sediment samples were airdried in open air on sterile plastic petri dishes until they were completely dried (appr. 7 days). They were then packed in sterile air-tight Ziploc bags (Biohazard, China) and stored at 4°C for 30 days until they were used to isolate native bacteria. Samples were subjected to our bacterial isolation protocol to select for quorum quenching (QQ) and stress tolerant isolates (i.e., heat at 85°C and NaCl at 35 g.L'1). This protocol is thought to eliminate virulent Bacillus cereus.

[0071] Firstly, minimal medium [MM; 0.080 g.L^NaHCCh, 0.250 g.L^KCl, 0.040 g.L^KBr, 1.840 g.L'1MgCl26.H2O, 0.410 g.L’1CaCL2.2H2O, 0.008 g.L’1SrCL2.6H2O, and 0.008 g.L’1H3BO3 in phosphate buffered saline (pH 6.5)] was prepared and autoclaved. Also, aN-hexanoyl homoserine lactone (HHL; fluka, Germany) solution was prepared as follows. A 50 mg.mL'1solution was prepared by dissolving 10 mg of HHL in 200 pL of ethanol (95%), followed by filter-sterilization using 0.22 pm PVDF syringe filters (30 mm diameter; Whatman). The final stock solution (1 g.L'1) was obtained by dilution with sterile distilled water and stored at 4°C until use.

[0072] Next, 2 g of each sediment sample was soaked for 30 minutes in 10 mL of a sterile physiological solution (PS; 8.0 g.L'1NaCl, 0.2 g.L'1KC1, 1.44 g.L'1Na2HPO4, and 0.24 g.L'1KH2PO4). From this 10 mL suspension, 100 pL was dispensed into 20 mL of MM supplemented with buffered [pH adjusted to 6.5 with sterile MOPS (200 mg.L'1) before storage] HHL to a final concentration of 5 mg.L'1as a sole source of nitrogen and carbon. Incubation of the samples was done in sterile glass tubes at 28°C for 48 hours on a shaker (BIOSAN ES-20) at 120 rotations per minute (rpm). After initial incubation, 1 mL of each sample was aliquoted into sterile 1.5 mL microcentrifuge Eppendorf tubes and heat-treated for 20 minutes at 85°C in a water bath to further minimize the likelihood of (co-)isolating toxin producing Bacillus cereus. After heat-treatment, 100 pL of the cultures were spread on sterile Luria Bertani (LB35) agar plates and incubated for 4 days at 28°C with continuous monitoring of colony formation and concentration (CFU.mL'1). Thereafter, 100 |iL from each of the remaining heat-treatments was inoculated into new sterile MM, supplemented with buffered HHL and 1% sterile Luria Bertani 20 (LB20) broth for another incubation cycle. Whereas plating was done on LB35 agar plates to screen for strains with a high salt tolerance capability, alternating addition of LB20 broth (1%) to MM was intended to increase nutrient availability for the already selected strains to enhance their concentration through the subsequent cycles of isolation and purification. Heat-treatment and growth on LB35 agar plates, each time followed by inoculation of colonies in minimal media supplemented with buffered HHL and or 1% LB20 broth was repeated for three cycles. Thus, in summary, three sub-cultures were made in MM supplemented with 5 mg.L'1HHL and or 1% LB20 broth to purify and select distinct colonies on LB35 agar for further characterization and screening. After the final cycle in MM, single colonies on LB35 agar were selected based on morphological differences (i.e., shape, color, size, and general appearance). All isolates were Gram-stained according to Paray et al., (2023) and all Gram-positive candidate probiotic isolates were cryopreserved in 50% sterile glycerol stocks, and stored at -80°C.

[0073] In vitro screening of candidate probiotics

[0074] Assessing the ability of Quorum Quenching (QQ) by measuring HHL degradation

[0075] Due to the importance of QQ in controlling bacterial virulence in aquaculture, the ability for QQ was assessed for our candidate probiotic isolates. We applied a protocol involving a modified plate diffusion method to quantitatively detect exogenous HHL degradation by QQ isolates to aid in the selection of QQ strains as described by Defoirdt et al., (2011). This method relies on the detection of exogenous AHLs being QS autoinducers, by Chromobacterium violaceum strain CV026, a strain devoid of AHL production, but still responsive to the presence of exogeneous AHLs. This is visible upon the production of violacein, a purple pigment produced during QS (Defoirdt et al., 2011a; Tinh et al., 2007).

[0076] In short, candidate probiotic isolates were inoculated on LB 10 agar and after overnight incubation at 28°C, a single colony was inoculated into 5 mL of buffered LB10 broth (pH 6.5 with MOPS, 200mgl'1), supplemented with 5 mg.L'1HHL. Incubation was done at 28°C and 120 rpm for 24 hours. This procedure was repeated for three independent colonies to obtain triplicate read-outs. Simultaneously, the C. violaceum CV026 reporter strain was grown overnight (28°C, shaking at 120 rpm) in LB10 broth buffered to pH 6.5 with sterile MOPS and supplemented with kanamycin (20 mg.L'1). Besides our isolates, negative and positive control strains, Pseudomonas flourescens P3 / pME6000 and P3 / pME6863, being AHL degrading and AHL non-degrading strains, respectively, were included to guarantee valid assay read-outs (Defoirdt et al., 2011; Gopu & Shetty, 2016). To detect HHL degradation by our candidate probiotic isolates, the optical density at 550 nm (OD550) of C. violaceum CV026 was adjusted to 0.1 before making lawns on LB 10 agar plates with 100 pL of the culture. For the standard curve, 10 pL from each HHL dilution series was spotted in the center on the C. violaceum CV026 lawns in triplicate and incubated at 28°C for 24 hours. For the test isolates and controls, 1 mL was aliquoted after 24 hours of incubation, filter-sterilized (0.22 pm PVDF syringe filter, diameter 30 mm, Whatman) to obtain bacteria-free supernatants, and spotted (lOpL) in the center onto C. violaceum CV026 lawns. LB10 plates with C. violaceum CV026 lawns containing spots were incubated for 48 hours at 28°C. The diameters of the purple halos of the violacein pigment on CV026 lawns were measured with a ruler (300 mm) and correlated with the standard curve. A linear regression model was then used to estimate HHL concentrations during the HHL degradation experiments for the test isolates and controls. Isolates capable of significantly reducing HHL concentration below the negative control level were classified as HHL degrading and thus QQ isolates and were retained for further analysis.

[0077] Visualization of in vitro QQ

[0078] The selected isolates were cross streaked with an AHL producer (i.e., Aeromonas hydrophila strain LVS3), and an AHL reporter (C. violaceum strain CV026) parallel to each other at separation not exceeding 15 mm. All the test strains and the negative control (sterile LB10 broth) were streaked at the center between the parallel cross streaks of A. hydrophila LVS3 and C. violaceum CV026. The QQ capability of our test isolates was confirmed when purple pigment formation occurred only where the C. violaceum CV026 streak was unobstructed by the QQ positive isolates, compared to the negative control, which did not inhibit purple pigmentation along the C. violaceum CV026 streak.

[0079] Haemolytic activity of candidate probiotics

[0080] Assessing hemolytic activity is crucial for probiotic biosafety as hemolytic bacteria can profit from small lesions and wounds on the skin of hosts and cause infections (Jinendiran et al., 2019). Therefore, selected QQ confirmed isolates were evaluated further for their hemolytic activity as described by Kuebutornye, Lu, et al. (2020) . Briefly, all isolates were tested for hemolysis on LB 10 agar plates supplemented with 5% (v / v) defibrinated sheep blood. The OD550 of the overnight grown cultures of the isolates was adjusted to 0.1 before spotting 5 pL of each isolate onto the center of the plates in triplicate, followed by incubation at 28°C for 48 hours. Isolates which induced complete hemolysis (P-hemolytic) were identified by clearance around and below the colony spots while isolates which induced partial hemolysis (a-hemolytic) exhibited greenish-brownish zones around and below the colony spots. Non-hemolytic isolates (y-hemolytic) were identified by the absence of clearance around the colony spots. Isolates that displayed P-hemolysis were eliminated, while those that exhibited y or a hemolysis were considered safe (Kuebutornye et al., 2020) and were retained for further screening and molecular characterization.

[0081] Proteolytic activity of candidate probiotics

[0082] Proteolytic activity is an essential indicator for protein utilization, the most expensive nutrient in fed aquaculture production (C. H. Liu et al., 2009). Hence, all QQ confirmed and y or a hemolytic isolates were subjected to an assay to assess their proteolytic activity by streaking them on skim milk agar plates (10% w / v) followed by incubation at 28°C for 72 hours in triplicate (Hossain et al., 2021). Isolates positive for proteolysis as indicated by the presence of clearance zones on the skim milk agar plates were selected for further molecular characterization using Oxford Nanopore Technologies’ (ONT) long-read whole genome sequencing (WGS) and subsequent in vivo biosafety testing.

[0083] Long-read whole genome sequencing, species identification, and in silico evaluation of probiotic activity and potential risks

[0084] All Gram-positive, in vitro confirmed QQ, y or a hemolytic and proteolytic isolates were revived from a freezer (-80°C) and cultured overnight at 28°C on LBw agar plates. These plates were subsequently transported to the PathoSense laboratory (Merelbeke, Belgium), where WGS was conducted using the ONT long-read GridlON sequencing platform, according to Vereecke and Van Hoorde et al. (2023). In short, all bacterial biomass of the putative probiotic isolates was collected into 250 pL dPBS for the isolation of High-Molecular Weight (HMW) DNA using the DNA MiniPrep Kit (Zymo Research), following manufacturer’s instructions. This included two 5-minute cycles of bead bashing (30 oscillations per minute in a TissueLyzer; Qiagen) and the inclusion of a 30-minute Proteinase K (20 pg.pL1; Promega) step post bead-bashing. Resulting HMW DNA was subjected to QC using a NanoDrop device, with the inclusion of an extra DNA clean-up with CleanNGS (CleanNA) beads (1:1 ratio) if A260 / A280 and / or A260 / A230 measures did not reach 1.7 or 1.5, respectively. A total of 400 ng HMW DNA per sample was used in a rapid long-read ONT library preparation (RBK-004) with barcoding and sequencing of the strains for 48h on an R9.4.1 MinlON flow cell. Data was acquired and real-time basecalled / demultiplexed on a GridlON device with super accurate base calling using guppy (v6.1.5; ONT). Resulting reads were used in an in-house established bacterial whole genome sequencing (WGS) pipeline, including read filtering (filtlong vO.2.1; — min length 1,000 — keep_percent 95; (https: / / github.com / rrwick / Filtlong), Trycycler subsampling (v0.5.3; — min read depth 50 —count 10 — genome size 5 M; (Wick et al., 2021), independent genome assemblies using Flye (v2.9; (Kolmogorov et al., 2019), raven (vl.8.1; Ruan & Li, 2020,Vaser & Sikic, 2021), wtdbg2 (vl.12; (Ruan & Li, 2020), and miniasm_miniplish.sh (v0.3; https: / / github.com / rrwick / Minipolish). Further steps include individual try cycler commands as described using default settings, followed by read polishing with minimap2 (v.2.20; (Li, 2018) and medaka (v.1.5.0; ONT). Resulting genome assemblies were identified and classified at the species level using rMLST (pubMLST; (Jolley et al., 2012) and the Type Strain Genome Server (TYGS;(Meier-Kolthoff & Gbker, 2019a). For the latter, all available genome assemblies within the identified genus were included in the TYGS analyses. Genome completeness was determined using CheckM (v.l.l.0;(Parks et al., 2015)) in which family level reference genomes were used for the Bacillaceae (418 marker genes from 162 Bacillaceae reference genomes) and Micrococcaceae (459 marker genes from 38 Micrococcaceae reference genomes). To assess putative new bacterial species, TYGS dDDH (d4) was used along with average nucleotide identities (ANI) as obtained from the ANI calculator on EZBioCloud (Yoon et al., 2017). A new species was determined based on dDDH (d4) and ANI below 70% and 95%, respectively ((Goris et al., 2007; Meier-Kolthoff & Gbker, 2019b). Final genome assemblies were submitted to NCBI under BioProject PRJNA1094437 with ascension numbers presented in Table 2. Next, we used our genomes to perform an in silica prediction and assessment of their probiotic nature and safety using two new tools. A first approach, iProbiotics (v.2023.6.5; (Sun et al., 2022) uses machine learning for the rapid identification of probiotic properties. Here, the “model 1 : Probiotic Predictor” was used for probiotic prediction. Our analysis was supplemented with well-known and relevant bacterial probiotics and pathogens as controls based on Kuebutornye et al, 2019. Secondly, our genomes were submitted to the ProBioMinServer (Y. Y. Liu et al., 2023), an integrated platform to assess the safety and functional properties of putative prokaryotic strains. This approach delivers a Probiotic Potential Risk Score (PPRS) as the overall sum of the presence of ARGs, virulence factors, and MGEs, which can be interpreted as low-risk (<4), medium-risk (4-6), and high-risk (>6). Finally, to reassure ourselves from the rigour of these tools, independent screenings against the CARD (Alcock et al., 2023) and complete VFDB (B. Liu et al., 2019) databases were conducted using Abricate (v.0.9.9; https: / / github.com / tseemann / abricate). This allowed us to analyse the data in “conserved” and “loose” mode, representing 80% / 80% and 60% / 60% for nucleotide identity and query coverage, respectively.

[0085] Data analysis and statistics

[0086] The QQ potential between the test isolates and the controls was compared by an Independent Two sample t-test in which significance was determined if / ? < 0.05. For the HHL degradation assay, a correlation between the diameters of the purple halos and the HHL concentrations in the HHL dilution series was used to compute a standard curve. A linear regression equation from the curve was used to calculate the estimated HHL / AHL concentration for the different isolates. Statistical analysis was conducted in R studio (v. 2023.03.0+386). Graphs were produced by Microsoft excel, GraphPad Prism (version 9, GraphPad Software, San Diego, CA, USA), and Sigma plot (v. 15).

[0087] Results for example 1

[0088] Isolation and selection of potential probiotic strains

[0089] From a total of nine sediment samples, 24 Gram-positive isolates were selected based on their diverse morphological characteristics (i.e., size, color, shape, and overall appearance).

[0090] Assessing the ability of quorum quenching (QQ)

[0091] All 24 isolates were tested in the plate diffusion assay for their HHL degrading capacity. The concentration of HHL (in mg. L'1) was determined using the equation derived from the HHL standard curve obtained from the linear regression on the standard curve The negative control strain (P. flourescens pME6000) gave no HHL degradation, while the positive control strain (P. flourescens pME6863) did. Eleven of 24 isolates (QQB2, QQA2, QQC1, QQA4, QQB8, QQB9, QQA7, QQC4, QQC5, QQC6, QQC7) showed the ability to degrade HHL to concentrations significantly lower (p < 0.05,) than the one observed for the negative control strain P. flourescens pME6000.

[0092] Visualization of in vitro QQ

[0093] The negative control plate (LB streak) showed no AHL degradation. Importantly, all eleven test isolates were indeed able to significantly degrade AHLs which were secreted by the AHL producer strain LVS3. AHL degradation resulted in undetectable HHL concentrations as showed by cross streaking the isolates in parallel with the AHL reporter strain CV026. The eleven isolates originated from different sediment samples and appeared morphologically different. Hence, the isolates were designated as QQ positive and were subjected to further analysis.

[0094] Hemolytic and proteolytic activity of candidate probiotics

[0095] Since all 11 test strains were confirmed to be QQ, they all underwent further evaluation for their hemolytic activity, to align with biosafety considerations to determine their suitability as potential probiotic strains for application in aquaculture. While three isolates demonstrated hemolytic activity (P-hemolysis - QQB8, QQB9, and QQC4), three (isolates QQA2, QQB2, and QQA7) were y- hemolytic and five (isolates QQC1, QQC5, QQC6, QQC7, and QQA4) were a-hemolytic. Hence the former (isolates QQB8, QQB9, and QQC4) were excluded from further consideration and the remaining eight strains were subjected to further testing.

[0096] The eight selected strains (i.e., QQ confirmed and y or a hemolytic - QQA2, QQB2, QQA7, QQC1, QQC5, QQC6, QQC7 and QQA4) were subjected to an assay to test for their proteolytic activity to further confirm and support their potential as probiotic candidates as this is in line with enhanced protein utilization. A total of six (QQA2, QQB2, QQC1, QQA4, QQC5 and QQC6) out of the remaining eight isolates exhibited proteolytic activity and were subjected to subsequent molecular characterization by WGS (Table 1).

[0097] Identification and classification of selected isolates

[0098] Following ONT long-read sequencing and genome assembly, complete circular genome assemblies for the six remaining isolates were obtained for the six isolates that were found suitable for probiotics in aquaculture (i.e., QQ confirmed, non-hemolytic (y and a hemolysis), and proteolytic positive; Table 1). An initial genome identification, using rMLST, resulted in the identification of five different bacterial species; Kocuria rhizophila KSNUG (100% identity with 24 exact rRNA gene matches for its 2.7 MBp genome with 70.6%GC), Lysinibacillus fusiformis LFUG (100% identity with 40 exact rRNA gene matches for its 4.6 Mbp genome with 37.6%GC), Mesobacillus sp. MSNUG (40% identity with 5 exact rRNA gene matches for its 4.8 MBp genome with 41.5%GC), Micrococcus luteus MYUG (100% identity with 42 exact rRNA gene matches for its 2.7 MBp genome with 72.8%GC), and Priestia megaterium PMUG01 andPMUG02 (100% identity with 56 exact rRNA gene matches for its 5.8 MBp genome with 37.8%GC).

[0099] To determine genome completeness, CheckM results were obtained against either the Bacillaceae or Micrococcaceae databases, which showed overall genome completeness > 99%, with the exception of 96.7% and 98.7% for the L. fusiformis LFUG and M. luteus MYUG strains, respectively. Also, no significant contamination was reported in the sequencing data. Due to the observation of some lowered rMLST identities or gene matches, a more elaborated TYGS and whole genome ANI analyses were performed for all strains. This suggested the identification of two new bacterial species (dDDH (d4 in %) < 70% and ANI < 95%) for the Kocuria rhizophila KSNUG (61% dDDH (d4) and 94.97% ANI as compared to the

[0100]

[0101] (GCA_003667225)) and Mesobacillus sp. MSNUG (27.9% dDDH (d4) and 70.12% ANI as compared to the Heyndrickxia acidicola DSM14745 (GCA 001636425.1)) strains. The latter was also shown to be closer related to Heyndrickxia acidicola as evidenced in the Genome BLAST Distance Phylogeny (GBDP) tree based on genome sequences of our new strain and a representative of each species within the genera of Heyndrickxia, Mesobacillus, and Weizmannia. Hence, these new bacterial species will be further referred to as Kocuria species nova KSNUG and Heyndrickxia species nova HSNUG and are also described as novel micro-organisms deposited with BCCM / LMG, Bacterial collection located at KL Ledeganckstraat 35, 9000 Gent, Belgium and having accession number LMGP-33748 or LMGP-33747, respectively. Interestingly, the TYGS analysis also highlighted the wrong speciation of the M. luteus MYUG strain, which was identified as Micrococcus yunnanensis rather than M. luteus in rMLST, which seems to lack data on M. yunnanensis at the moment of our analysis.

[0102] Next, our genomes were used to assess their putative role as probiotics using the iProbiotics tool, which relies on machine learning to rapidly identify probiotic properties. All strains were classified as probiotic with a probability ranging between 92-99%, this is in correspondence with (putative) probiotics (96.3 ± 3.7% probiotic prediction probability; n=9) and in contrast with (putative) pathogens (67.4 ± 37.3% non-probiotic prediction probability; n=23) in aquaculture. Of note, the lowered non-probiotic prediction probabilities were a result of higher probiotic prediction probabilities for Lactococcus piscium. Yersinia enter olilica. Yersinia ruckeri. Streptococcus iniae. and Pseudomonas fluorescens. Removing these from the analysis resulted in a non-probiotic predicition probability of 83.6 ± 22.3% for the remaining (putative) pathogens (n=. Second, a more extended approach was applied to assess the safety of our potential probiotics using the ProBioMinServer platform, which delivers a PPRS. Based on their scoring system, only two of the six putative probiotic strains (Kocuria species nova KSNUG and M. yunnanensis MYUG) were classified as low-risk (PPRS <4). The L. fusiformis LFUG, Heyndrickxia species nova (HSNUG) and P. megaterium PMUG01 / PMUG02 strains were all classified as high(er) risk with PPRSs of 6.08, 9, and 11, respectively. These elevated scores were majorly attributed to pathogenic genes as identified using the PHI-base database. To reassure ourselves to not miss any important aspects on the safety of our new putative probiotics, we independently mined our genomes for the presence of known ARGs and putative virulence factors using the CARD and complete VFDB databases, respectively. When running in “conserved” mode (i.e., 80% nucleotide identity and query coverage), only the IsaB gene, which encodes for an ABC-F subfamily protein conferring resistance to lincosamides (AJ579365.1), was identified in the P. megaterium genomes. Performing this same analysis for virulence factors, showed the identification of three putative virulence factors; The sphaericolysin gene (Bsph_4094 from Lysinibacillus sphaericus C3-41 (VFG043991) was found in the L. fusiformis LFUG strain. Whereas in both the Kocuria species nova KSNUG and M. yunnanensis MYUG, a gene encoding for an isocitrate lyase (id from Mycobacterium avium H37Rv (VFG009263)) and a chaperonin GroEL (Rv0440 from Mycobacterium tuberculosis H37Rv (VFG043550) were identified. When extending our search to “loose” mode (i.e., 60% nucleotide identity and query coverage), the same genes were identified as reported above, along with the identification of additional ARGs and virulence factors. Tables for example 1

[0103] Table 1: Overview of QQ confirmed putative probiotic isolates (n=ll) from aquaculture sediments, comprising in vitro QQ rate (in %), hemolytic activity and proteolytic activity of the isolates. The three f-hemolytic isolates were considered unsafe and were not tested for proteolytic activity (NA: Not Assessed). Isolates with significantly higher QQ rate than the negative control, non-hemolytic and proteolytic in nature were considered suitable (Yes) for probiotic application in aquaculture whereas those that did not satisfy the above criteria were not (No).

[0104] Probiotic

[0105] Isolate In vitro QQ rate (%) Hemolysis Proteolysis suitability

[0106] QQB2 68.4 y + Yes

[0107] QQA2 87.6 y + Yes

[0108] QQC1 86 a + Yes

[0109] QQA4 100 a + Yes

[0110] QQB8 76 p NA No

[0111] QQB9 100 p NA No

[0112] QQA7 80.7 y - No

[0113] QQC4 100 p NA No

[0114] QQC5 100 a + Yes

[0115] QQC6 87.6 a + Yes

[0116] QQC7 99.4 a - No Table 2: Final genome assemblies were submitted to NCBI under BioProject PRJNA1094437 with accession numbers presented below. They were subjected to species classification using both rMLST and TYGS classification and the final species classification is also shown. Two new bacterial species (Kocuria species nova and Heyndrickxia species nova) were identified.

[0117] Strain ID Length (bp) Accession Final classification KSNUG 2,760,674 NCBI | GCA_003667225.1 Kocuria species nova LFUG 4,581,148 NCBI GCA_003049525.1 Lysinibacillus fusiformis HSNUG 4,848,796 NCBI GCA_001636425.1 Heyndrickxia species nova MYUG 2,686,177 NCBI GCA_014138885.1 Micrococcus yunnanensis PMUG01 5,810,842 NCBI | GCA_017086525.1 Priestia megaterium PMUG02 5,815,724 NCBI | GCA 017086525.1 Priestia megaterium Example 2: protocol for phenotyping and safety assurance of the isolated and selected probiotics

[0118] Following whole genome sequencing (WGS) and identification of the isolated strains, molecular characterization and annotation for coding DNA sequences (CDS) and probiotic attributes, probiotic strains can be phenotypically characterized for in vitro probiotic traits. Furthermore, probiotic strains can be assessed for their biosafety in crustaceans (shrimp - Artemia franciscana) and vertebrates (zebrafish -Danio rerio).

[0119] Materials and methods

[0120] N-hexanoyl homoserine lactone (HHL) degradation

[0121] Quorum sensing (QS) regulates several survival attributes in bacteria, including biofilm formation, secretion systems, motility, iron sequestration, hemolysis, substrate utilization, and antimicrobial resistance (Castillo-Juarez et al., 2015; Deng et al., 2021). Interfering with QS can involve degradation of AHLs. Consequently, we evaluated our strains’ capability to degrade AHLs, particularly HHL, a predominantly produced QS signal molecule by Gram-negative aquaculture pathogens including Vibrio sp., Aeromonas sp., and Edwardsiella sp. (Defoirdt et al., 2011a), using a modified plate diffusion assay based on Defoirdt et al. (2011), with slight modifications. Briefly, a single colony from overnight grown cultures of the test isolates, being the six candidate probiotic strains, the negative control (Pseudomonas flourescens P3 / pME6000) as a non-HHL degrader and the positive control (Pseudomonas flourescens P3 / pME6863) as HHL degrader (Molina et al., 2003). All strains were grown on LBio agar at 28°C was inoculated into 5 mL of buffered LBio broth (pH adjusted to 6.5 with sterile 3-(N-morpholino) propanesulfonic acid (MOPS) at 200 mg.L'1) and supplemented with 5 mg.L'1of HHL. All inoculations were performed in triplicate and incubated at 28°cand 120 rpm for 24 hours. Samples of 1 mL were taken at 3, 6, 24, and 48 hours, then filter-sterilized through a 0.22 pm polyvinylidene fluoride (PVDF) syringe filter (diameter -30 mm, Whatman, USA) into sterile 1.5 mL microcentrifuge Eppendorf tubes. From the aliquoted supernatants, 10 pL was spotted in the center of LBio agar plates with lawns Chromobacterium violaceum CV026, a strain that cannot produce AHLs but detects the presence of exogenous AHLs by producing a violacein purple pigment (Ghanei-Motlagh et al., 2021). The plates were incubated at 28°cfor 48 hours, after which the diameter of the purple halos formed was measured. The concentration of HHL was quantified by fitting the diameters of the purple halos into a regression equation derived from a standard curve of known HHL concentrations and their corresponding halo diameters. The smaller the diameter of the purple halos, the lower the HHL concentration. Quorum quenching capacities under co-culture conditions

[0122] Subsequently, the test strains were evaluated for their ability to degrade acyl-homoserine lactones (AHLs) produced by Aeromonas hydrophila LVS3 under co-culture conditions. This was done to test the quorum quenching capability of our strains against other AHLs produced by A. hydrophila such as N-butanoyl-L-homoserine lactone (C4-HSL) in addition to HHL (Rama Devi et al., 2016). Prior to the co-culture experiment, the antagonistic potential of the test strains against LVS3 was verified using the agar well diffusion plate method described by Vaseeharan and Ramasamy (2022). Non-antagonistic strains were selected for the co-culture experiment where the OD550 of all involved bacterial strains was adjusted to 0.1. Co-culturing was done in buffered LB 10 broth (pH adjusted with sterile MOPS at 200 mg.L'1). A negative control, consisting of LVS3 cultured alone without any test strains was included in the experiment and incubated under the same conditions. Following 48 hours of incubation at 28°C and 120 rpm, the concentration of AHLs was determined using a previously described modified procedure from Defoirdt et al. (2011). The AHL concentrations in the co-cultures were compared with the negative control to determine the differences in AHL degradation.

[0123] Growth rates

[0124] Probiotic bacteria pre-conditioning and administration at high densities are strategies used to enhance their rapid colonization, establishment, and dominance within hosts and their environments (Vine et al., 2004). Bacteria with higher specific growth rates gain a competitive advantage by colonizing and establishing themselves in the GIT before being flushed out, thereby excluding pathogens more effectively and conferring other beneficial attributes to hosts (Vine et al., 2004). Hence, the growth rates of our strains were tested using LB10 broth (200 pL) in wells of sterile nontreated transparent 96 well bacterial culture plates (VWR, Belgium). Initially, the absorbance at 550 nm of all the strains was adjusted to 0.1 and 7 replicates per strain were used. Sterile LB 10 broth was used as a blank and the plate was covered and inserted into a Tecan Infinite 200 microplate reader (Tecan, Mechelen, Belgium) at 28°C for 36 hours. Hourly measurements of the optical density (OD) at 550 nm after shaking (3s) were measured and recorded. Absorbance data for all the isolates were fed into GraphPad Prism software (version 9, GraphPad Software, San Diego, CA, USA) to generate growth curves and determine the growth characteristics of the isolates based on a modified Gompertz model (Equation 1) (Zwietering et al., 1990; Vandeputte et al., 2023).

[0125] Equation 1

[0126]

[0127] Parameters Yo and YM are the starting and maximum populations, respectively, expressed as a value of optical density at 550 nm (OD550); lag is expressed in hours and indicates the duration of the lag phase; and K is expressed as a reciprocal time unit (h-1).

[0128] Stress tolerance (pH and salt-NaCl)

[0129] In addition to selecting for robust, resilient, and environmentally stable probiotic strains, stress-tolerant bacterial strains foster enhanced applicability in diverse environmental and host conditions (fresh, brackish, and marine water and fish species) as well as hardiness during handling, processing, and transportation (Lee et al., 2010; Tang et al., 2020). Our candidate probiotic strains were then examined for their tolerance and growth under varying salt (NaCl) concentrations and pH levels in LB broth.

[0130] Different salt (NaCl) concentrations (1%, 2%, 3% and 4%) were tested in sterile nontreated transparent 96 well bacterial culture plates (VWR). Negative controls containing un-inoculated broth were included. The plate was covered and incubated at 28°cfor 24 hours. Absorbance was measured at 550 nm using a Tecan Infinite 200 microplate reader after shaking (3s) (Tecan, Mechelen, Belgium). Absorbance data for all the strains were analyzed using GraphPad Prism software (version 9, GraphPad Software, San Diego, CA, USA) to generate respective growth curves.

[0131] For testing pH tolerance, the pH of LB 10 broth was adjusted to levels ranging from 3 to 9 using 1 M HC1 and 1 M NaOH, respectively and then autoclaved. Strains were inoculated and incubated for 24 hours at 28°cin a shaking incubator (120 rpm), followed by OD550 measurements. Graphs of OD550 for the different strains at different pH levels were plotted to determine the optimal and sub-optimal pH ranges for growth.

[0132] Glycan diversity

[0133] Bacterial glycans are carbohydrate modifications typically found on their proteins or lipids and can act as ligands for glycan-binding proteins, called lectins in the host organism. Host lectins distinguish bacterial cell surface glycans and can achieve species- and strain-level binding specificity. Using a protocol by Bossier and Wongprasert (2019) with some modifications, we utilized a CytoFLEX flow cytometer system (Beckman Coulter’s Life Sciences, France) to assess the glycan diversity of the test strains. The glycan diversity translates thus into the strains’ potential to interact with host cells through specific binding with lectins, which regulate colonization, adhesion, and pathogen exclusion (Poole et al., 2018). Briefly, a single colony was inoculated into LB 10 broth and incubated with constant agitation (120 rpm) at 28°C for 24 hours. The cultures were harvested and diluted to a density of 108CFUrnl'1with a sterile physiological solution (0.9% NaCl). To 1.5-mL sterile microcentrifuge tubes, 250 pL of the diluted bacterial suspensions was transferred for each strain, followed by addition of 1 pL of the 7 different lectins, each coupled with fluorescein isothiocyanate (FITC) (Lectin Kit, Vector Laboratories, USA). For each test strain, one microcentrifuge tube wasn’t mixed with lectins and acted as a negative control. The contents were gently mixed by pipetting, and 100 pL of the mixture was transferred to each well of a sterile transparent 96-well flat-bottom plate. Samples were incubated for 30 minutes in the dark prior to CytoFLEX measurements. Analyses of the samples was done with a 37 pm poresized filter, a 75 pm pore-sized orifice, and a 0.3 neutral density filter. The photomultiplier tube was set at 500 V, following manufacturer’s instructions for standardization. Diluted fluorescent microbeads were used as the standards for fluorescence. Fluorescent signals at the wavelength of 525 nm for FITC were recorded for the different test strains and lectins. The machine settings were adjusted to FITC-113 gain, with an event rate controlled under 1000 events per second. Histograms of FITC intensity per cell on the X-axis against cell count on the Y-axis were generated. Median fluorescent signal values for each lectin were recorded and used to estimate the glycan index using the Shannon diversity (H’) concept (Equation 2). From the equation, / ? / is the proportion of lectin binding of the same test strain for a given lectin, obtained by dividing the median fluorescence intensity (Med-FITC-A) of a given test strain for a specific lectin by the total median fluorescence intensity for all lectins for a that strain.

[0134]

[0135] Metabolic fingerprinting

[0136] A positive correlation between gut microbial diversity and metabolic profiles of bacteria has been reported by Amenyogbe et al. (2022). Additionally, probiotic functional relationships in the GIT are associated with their metabolic fingerprints (Amenyogbe et al., 2022). To study the metabolic fingerprints of our isolates, Biolog Gen 111 MicroPlates (BiOLOG Inc. Hayward, CA, USA) featuring diverse substrate compositions were used according to the manufacturer's guidelines with some modifications. These substrates depict the complexity of aquafeeds that need to be digested and utilized by farmed animals. Briefly, strains were streaked onto Tryptic Soy Agar (TSA) and incubated at 28°C for 16 hours. A single colony was picked using a sterile disposable inoculator swab and transferred into inoculating fluid B (IF-B), provided by the manufacturer. The transmittance was adjusted to 94% using a turbidimeter, with uninoculated IF-B used as the standard. Each well of the plate was inoculated with 100 pL of the bacterial cell suspensions from the different test strains. The plates were covered with aluminum foil and incubated at 28°C for 24 hours. Using a Tecan Infinite M200 machine (Tecan, Mechelen, Belgium), the OD590 was measured across wells after 24 hours of incubation. Absorbance was standardized by subtracting the absorbance of the negative control from that of each substrate. The glycan diversity index (Equation 2) for total substrate and amino acid utilization by the different strains were computed. Sporulation

[0137] Spore-forming probiotics have a competitive edge over non-spore-forming counterparts due to their extended viability under highly stressful environments (Hong et al., 2005; Kavitha et al., 2018). Therefore, all test strains were examined for sporulation in sterile sporulation media (SM) as described by Nicholson and Setlow (1990), with some modifications. The SM medium (pH 7.6) comprised 5 g peptone from meat, 3 g yeast extract, 5 g NaCl, 10 ml of 10% (w / v) KC1, and 10 ml of 1.2% (w / v) MgSO4. 7H2O, diluted with distilled water to 1 liter. After autoclaving, SM medium was supplemented with filter-sterilized (0.2 pm syringe filter, Whatman, UK) metal solutions, including 1 ml of 1 M Ca(NOs)2, 1 ml of 0.01 M MnCh, and 1 ml of 1 mM FeSCE.

[0138] For spore staining, a Schaeffer and Fulton spore staining Kit (Sigma Aldrich, Germany) was used according to manufacturer’s instructions. Visualization of spores, endospores, and vegetative cells was achieved using an upright Eclipse E200 microscope (Nikon, Japan) with 1000X magnification (Nikon 50 Type A, Japan). Images were captured using NIS Elements D.4.51.00 (Nikon, Japan) software.

[0139] Antimicrobial resistance

[0140] In addition to evaluating hemolytic activity of the strains during their isolation and screening, all spore-forming strains were screened for antibiotic susceptibility to establish their potential for transferring antimicrobial resistance to other strains of veterinary and public health importance. A panel of commonly used antibiotics including those deemed critical by the EU for both human medicine and agriculture (Lulijwa et al., 2020; Serwecinska, 2020) was used. The disk diffusion method described by Kirby-Bauer (2009) was followed. Briefly, cultures were spread on Mueller Hinton Agar (Hi-Media), and the antibiotic disks were applied. Zones of inhibition were measured with a ruler (300 mm) in millimeters, and the mean diameters were used to classify the bacteria as resistant (R), intermediate (I) or susceptible (S) according to Abraha et al., (2017).

[0141] In vivo biosafety / non-toxicity verification in crustaceans (shrimp) and fmfish (zebrafish)

[0142] To validate non-toxicity in vivo, axenic Zebrafish embryos (Danio rerio AB line, wild-type, 3 dpf) were bath-treated with the candidate probiotic strains at a cell density of IxlO7CFU.mL1and incubated at 28°C for 3 days according to Russo et al. (2015) with minor modifications. Overnight grown cultures were centrifuged (12°C, 7500xg, 10 min) and washed 3 times with sterile (PBS, pH 7.2) before resuspension in sterile PBS followed by inoculation into the wells containing the embryos. Six-well microtiter plates with each well containing ten zebrafish embryos and 5 ml of E3 media were used in triplicate for each tested strain as well as the negative control (sterile PBS). Axenic brine shrimp (Arlemia Fransiscana) nauplii (instar II stage) was used for the in vivo biosafety verification in Shrimp. They were bath-treated with the bacterial strains at a cell density of 107CFU.mL1. The nauplii were incubated in filtered artificial autoclaved seawater (FAASW -35 g.L'1instant Ocean synthetic sea salt, Aquarium Systems, Sarrebourg, France) on a rotor at 28°C and 6 rpm for 48 hours. They were fed once on autoclaved hydrophilalN^ (108CFU.mL'X).

[0143] Survival rates were calculated for the test treatments and compared with that of the negative controls. Toxicity was inferred when the survival rates of the test treatments were significantly lower than that of the negative control treatments, and vice versa.

[0144] Gut colonization

[0145] Zebra fish (Danio rerio AB line wild-type) embryos were used to visualize and verify probiotic bacteria colonization as described by Nayak etal. (2023) with minor modifications. Briefly, the candidate probiotic strains were transformed and tagged with a red fluorescent plasmid (RFP) tag (pDsRed-monomer, Clonetech laboratories, Inc. CA, US). For this, cells were made competent, and the plasmid was introduced into the competent cells following the protocol detailed in the Gene Pulser X-cell System user manual electroporation system (Bio-Rad, USA). The electrophoresed colonies were plated onto Luria Bertani agar (Merck KGaA, Darmstadt, Germany) with ampicillin (50 mcg.mL'1) and incubated at 30°C for 18 hours. The potential colonies were randomly picked and checked for plasmid insertion by visualizing them under the fluorescent microscope (ZOE fluorescent cell imager, Bio-Rad). The colonies tagged with RFP were stored in 30% glycerol supplemented with ampicillin (50 mcg / ml) at -80°C until further use. Overnight grown cultures of the RFP-tagged isolates were centrifuged at 7500xg and 12°C for 10 minutes to collect the pellet, which was washed thrice with sterile phosphate buffer (PBS, pH 7.2). After the final wash, the pellet was re-suspended in PBS to form an end concentration corresponding to an absorbance of 0.01 at 600 nm (~lxl07CFU.mL'1). Six-well microtiter plates with each well containing ten zebrafish embryos (3 dpf) were used in triplicate for each tested strain as well as the negative control (sterile PBS). Embryos were exposed to the bacteria and incubated for 48 hours at 28°c. After every 24 hours, the zebrafish were visualized under a fluorescent microscope (DM2500, Leica Microsystems, Germany), and the gut region was checked for possible colonization by the tagged isolates. The presence of an RFP signal inside the gut lining would indicate successful colonization by the tagged bacteria. The in vivo presence and attachment of the bacteria were checked by visualizing the tagged, probiotic exposed zebrafish embryos with a fluorescent microscope (DM2500, Leica Microsystems, Germany), and images were taken. Images were analyzed by Image-J software (Schroeder et al., 2021), LAS X software (https: / / www.leica-microsystems.com), and the corrected total cell fluorescence (CTCF) values obtained were plotted on a graph.

[0146] Statistical analyses

[0147] An independent Two sample t-test in which significance was inferred when p < 0.05 was used to compare the HHL degradation rate of the negative control and the other strains at the different time points under independent and co-culture treatments. In vivo survival rate data for the biosafety assays were first transformed (i.e., arcsine) to fulfil the requirements of normality and homoscedasticity prior to running a One-way ANOVA in which significance was determined if p < 0.05. Statistical analysis was conducted in R studio (v. 2023.03.0+386).

[0148] Results for example 2

[0149] N-hexanoyl homoserine lactone (HHL) degradation rates

[0150] By the 3 -hour time point, all strains except PMUG02 and MYUG had significantly reduced HHL concentrations compared to the negative control (p < 0.05). At 6 hours, all strains, excluding PMUG02 (p=0.169), continued to show significantly lower HHL levels. By the 24-hour point, all strains including PMUG02, had degraded HHL to significantly lower concentrations than the negative control, with most strains degrading HHL to undetectable concentrations. PMUG02 achieved undetectable HHL concentrations by 48 hours. Overall, by the end of the 48-hour incubation, all strains had significantly reduced HHL concentrations compared to the negative control (P3 / pME6000) (p < 0.05).

[0151] Quorum quenching under co-culture conditions

[0152] Under co-culture conditions, all strains exhibited no impact on the growth of LVS3 and effectively degraded the AHLs it produced after 48 hours of incubation at 28°C to significantly lower concentrations compared to the negative control, with the exception of MYUG and PMUG02 (Figure 1) (p < 0.05).

[0153] Growth rates

[0154] The maximum specific growth rate was highest for MYUG, followed by KSNUG, PMUG01, LFUG, PMUG02, and HSNUG (Table 3). The maximum specific growth rate of the strains ranged between 0.012 (HSNUG) and 0.023 (MYUG) (log cell / (mL.h)), while the maximum population ranged between an OD550of 0.380 (HSNUG) and 0.728 (PMUG01) (Table 3).

[0155] Stress tolerance (pH and salt-NaCl) All isolates could maintain and / or grow at salt concentrations ranging from 1% to 4 % after 24 hours of incubation at 28°C to a cell density exceeding 106CFU.mL'1(Figure 2). However, there were discrepancies in the timing at which the log phase commenced and ended. Generally, with increasing salt concentration, the growth curves were shifted to the right of the X-axis, and the cell density at the stationary phase decreased.

[0156] The optimum pH range for all bacterial strains ranged between pH 6 and 8. However, all strains demonstrated the capability to either maintain or grow within a broader pH range of 3 to 9, attaining a minimum cell density of 106CFU.mL'1after 24 hours of incubation at 28°C (Figure 3). Glycan diversity

[0157] The glycan diversity index (H’) was highest for HSNUG, followed by KSNUG, PMUG02, PMUG01, MYUG, and LFUG (Table 4). All strains demonstrated preferential lectin binding to lectin WGA except LFUG and PMUG01, whose lectin binding preferences were for lectins ConA and UEAI, respectively. Notably, although the lectin binding preference for KSNUG was for WGA, its preference for lectin RCAI was almost comparably high.

[0158] Metabolic fingerprinting

[0159] The capacity of the strains to metabolize various substrates in the Biolog Gen III microplate was assessed. As depicted in (Figure 4), LFUG and HSNUG utilized substrates from all the six categories, followed by PMUG01 and KSNUG, PMUG02, and MYUG, which utilized substrates from five, four, and three categories, respectively. Carbohydrate utilization, encompassing carboxylic acids and esters, polyols, sugar acids and their compounds, sugar derivatives, and sugars, was most pronounced for strain HSNUG (51), followed by LFUG (32), KSNUG (22), PMUG01(18), and PMUG02 and MYUG (17) (Figure 4). Whereas HSNUG (7) and LFUG (5) presented superior amino acid utilization, the other strains utilized only one amino acid in the microplate, except PMUG02 and MYUG that were unable to utilize any of the amino acids in the plate.

[0160] The metabolic fingerprints of the strains were estimated by the Shannon diversity concept and are presented as Shannon diversity indices in Table 5. A higher index indicates a greater capacity to metabolize a diverse range of substrates by a given strain. For all substrates, the index was highest for HSNUG (1.966), followed by LFUG (1.703), KSNUG (1.605), PMUG01(1.476), PMUG02 (1.119), and MYUG (1.027). Furthermore, the Shannon index for amino acid utilization was highest for strains LFUG (0.365) and HSNUG (0.314), followed by PMUG01 and KSNUG (0.207), in contrast to PMUG02 (0.000) and MYUG (0.000) that did not utilize any amino acid in the plate. Sporulation

[0161] As anticipated, four members of the Bacillaceae family (PMUG01, PMUG02, HSNUG, and LFUG) out of the six (06) sequenced and identified strains were observed to produce both endospores and free spores. This was evident from the staining with malachite green against a pink background of safranin on the vegetative cells after 48 hours of incubation at 37°cin sporulation medium (SM).

[0162] Antimicrobial resistance

[0163] While susceptibility to all antibiotics was observed for all the spore-forming strains, only two strains, HSNUG and LFUG, exhibited recorded resistance to Ampicillin and Streptomycin, respectively. Additionally, HSNUG was the only strain that showed intermediate susceptibility for Azithromycin and Chloramphenicol, while the rest were fully susceptible.

[0164] Toxicity of candidate probiotics in zebrafish embryos andArtemia nauplii

[0165] Following a 48-hour bath-treatment of axenic Zebrafish embryos with all spore-forming putative probiotic strains at a concentration of ICCCFU.mL'1, no significant difference in survival rates was observed between the treatments and the negative control (p > 0.05). Except for PMUG02 and HSNUG, with survival rates of 82.5% and 86.7%, respectively, all the other treatments registered survival rates greater than 95% (Figure 7). For the Artemia nauplii, all the strains were nontoxic after 48 hours of incubation at 28°C. The survival rates for all treatments with different isolates were not significantly different from each other and from the negative control (p = 0.0624). Gut colonization

[0166] After 24 hours of treatment with the RFP-tagged bacteria, the Zebrafish embryos showed the highest mean corrected total cell fluorescence (CTCF) for LFUG treatments, followed by HSNUG, PMUG01, and PMUG02. However, by 48 hours, the mean CTCF of LFUG-treated embryos decreased, while it increased for all the other treatments, eventually reaching a level equivalent to that of PMUG02 treatments, followed by HSNUG and PMUG01. The CTCF in untreated (negative control) treatments remained the lowest and showed minimal change over time. Tables for example 2

[0167] Table 3: Maximum specific growth rates measured at OD550 of all the strains obtained after 36 hours.

[0168] Strains Maximum specific growth rate Maximum Coefficient of

[0169] / n 11 / / Population (Ym) determination (R2) (p [log cell / (mL*h)])

[0170] KSNUG 0.021±0.003 0.492 0.993 HSNUG 0.012±0.001 0.380 0.994 MYUG 0.023±0.003 0.474 0.985 PMUG02 0.017±0.002 0.401 0.997 PMUG01 0.020±0.002 0.728 0.998 LFUG 0.019±0.002 0.631 0.992

[0171] Table 4: The glycan diversity index (H) and the relative lectin binding for the different strains. The median lectin fluorescence intensity (Med-FITC-A) was used to estimate the index from a Shannon Index concept.

[0172] Strain PNA CoNA SBA UEAI DBA WGA RCAI H' MYUG 0.024 0.015 0.102 0.070 0.028 0.714 0.047 1.055 KSNUG 0.098 0.130 0.050 0.161 0.083 0.245 0.233 1.827 PMUG02 0.137 0.092 0.139 0.092 0.037 0.440 0.063 1.642 HSNUG 0.147 0.153 0.159 0.166 0.083 0.247 0.045 1.851 PMUG01 0.134 0.053 0.098 0.443 0.039 0.205 0.028 1.564 LFUG 0.020 0.850 0.018 0.045 0.017 0.037 0.013 0.675 Table 5: The Shannon diversity indices for all the strains based on their capability to metabolize all substrates and amino acids, peptides, and polypeptides in the Biolog Gen III microplate.

[0173] Shannon Diversity Index

[0174] Isolate All substrates Amino acids and peptides LFUG 1.703 0.365

[0175] PMUG01 1.476 0.207

[0176] PMUG02 1.119 0.000

[0177] MYUG 1.027 0.000

[0178] HSNUG 1.966 0.314

[0179] KSNUG 1.605 0.207 Example 3 : use of the bacterial species Lysinibacillus fusiformis LFUG, Micrococcus yunnanensis MYUG, Priestia megaterium PMUGT01 and PMUG02, Heyndrickxia species nova HSNUG and Kocuria species nova KSNUG as probiotics in aquaculture.

[0180] Selected strains (PMUG01 and LFUG) were tested for their probiotic (growth promotion, feed conversion efficiency, and pathogen resistance) efficacy in Nile tilapia juveniles under laboratory conditions.

[0181] Materials and methods

[0182] Ethical approval

[0183] The experiment was approved by the Research Ethics Committee (REC) of Busitema University Maritime Institute (BUMI) (BUMI / Res. Ethics / RP / doc.l) in accordance with the Busitema University Research Ethics policy. It also complied with the stipulations of Section 12 of the Uganda Animals (Prevention of Cruelty) Act (ULII, 2000).

[0184] Isolation and selection of the probiotic strains

[0185] Indigenous strains of probiotic bacteria were isolated from sediment samples collected from earthen Nile tilapia ponds according to our previously described isolation protocol.

[0186] Briefly, the protocol aimed at selecting for non-haemolytic, proteolytic positive, and stress-tolerant probiotic strains (NaCl, heat and pH) with enhanced quorum quenching (QQ) capability. From the six (06) isolated, characterized, and identified strains, we selected two strains (i.e., Priestia megaterium PMUG01 and Lysinibacillus fusiformis LFUG) which belong to species that have already showed high probiotic potential for aquaculture, especially in Nile tilapia culture (Balasubramanian et al., 2023; Beryl Chean et al., 2021; Melo-Bolivar et al., 2021; Rios Castro et al., 2022; Widyastuti et al., 2023; Zabidi et al., 2021).

[0187] In vivo verification of probiotic action

[0188] Experimental fish and culture facility

[0189] Nile tilapia juveniles (1.7+0.03g) were acclimatized for 2 weeks in 60-liter glass tanks containing 50 litres of water. At the end of the acclimatization period, they were examined for quality and overall health (physical and sensory) and fish without any visible signs of disease, deformity and injury were selected for the experiment. They were randomly divided into groups of 50 fish per tank and stocked at a rate of 1 fish.L'1. They were randomly assigned treatments corresponding to the different bacterial strains and concentrations including the control in triplicate. The tanks were cleaned twice daily by siphoning out the dirt before feeding, replenishing the water with 70% aerated freshwater. Physiochemical water quality parameters (pH, dissolved oxygen and temperature) were monitored daily throughout the experiment. Bacteria preparation and processing

[0190] Two strains, Priestia megaterium PMUG01 and Lysinibacillus fusiformis LFUG were selected for the experiment at two different concentrations, T1 (106) and T2 (108) CFU.g1of feed. They were retrieved from a -80°C freezer and streaked on tryptic soya agar (TSA) (Hi-media, India) and incubated for 24 hours at 30°C. Single colonies were inoculated into 100 pL of tryptic soya broth (TSB) and incubated for 48 hours at 30°C. Cultures were then centrifuged at 7000 xg at room temperature for 10 minutes. The supernatants were discarded, and the pellet was washed 3 times with sterile saline water (0.85g.L-1- NaCl) before resuspending it in sterile saline water to make stocks. The absorbance at 550 nm of the stocks was measured, and the cell density estimated according to a McFarland’s standard (BioMerieux, Marcy L’Etoile, France) at a wavelength of 550 nm, assuming the optical density at 550nm of 1, corresponds to 1.2xl09CFU.mL1. This was used to make appropriate dilutions in normal saline corresponding to the required concentrations during feed fortification with the bacteria.

[0191] Feed preparation and feeding regime

[0192] A total of 500 g of feed was measured for each treatment and placed on dry plastic trays. It was evenly mixed with 100 mL of fish oil (MPUTA, Uganda) followed by spraying with 100 mL of the respective bacteria suspensions. For the control, the feed was instead sprayed with 100 mL of sterile normal saline. The treatments were designated the following names: The Control (0 CFU.g'x), PMUGT1 (IxlO6CFU.g'1), PMUGT2 (IxlO8CFU.g'1), LFUGT1 (IxlO6CFU.g'1) andLFUGT2 (IxlO8CFU.g'1). The feeds were allowed to air dry in open air in the lab whose temperature was maintained at 30°C for 5 days with daily mixing to avoid feed cacking. The experimental diets were stored in sterile airtight glass bottles and stored in a refrigerator at 4°C from where they would be picked when needed. Feeds were freshly prepared every three weeks until the end of the experiment (60 days). The commercial feed that was used in the experiment (crumbles of 0.5-0.8 mm in size) had proximate composition of: 49% crude protein, 4% crude fiber, 6.5% crude fat, 11% moisture, 14% ash, 1.3% phosphorus, and 1.5% calcium according to the manufacturer (Koudijs, Uganda). Fish in all treatments were fed at 4% of their average body weight (ABW) twice a day for the 60 days of the experiment while monitoring the feeding response.

[0193] Sampling and evaluation of fish growth performance

[0194] Sampling for weight (g), total length (cm) and standard length (cm) was done after every 10 days on 30 randomly selected fish from each treatment. The growth performance indicators including weight gain (g), specific growth rate (%.day ''), feed conversion ratio (FCR), condition factor (K) and survival rate (%) were calculated and compared between treatments. At the end of the feeding trial, the viscerosomatic and hepatosomatic indices (VSI and HSI) were determined and compared between treatments. The growth parameters were calculated according to the following formulae:

[0195] 1. Weight gain (g) = Final weight (g) — Initial weight (g)

[0196] 2. Specific growth rate (%.Day-1) =

[0197] (In Final average body weight (g)-In Initial average body weight (g)) - QQ

[0198] (Time of sampling (Days))

[0199] > > , > > . Quantity of feeds fed (s)

[0200] 3. Feed Conversion Ratio (FCR) = , , - Fish biomass gained (g)

[0201] . , .

[0202] 4. Survival rate 100

[0203]

[0204] 5 Condition factor (K) body w.tght (g)x 1Q0

[0205] v

[0206]

[0207] (Total Length (cm))3 7

[0208] - -r r- •T1 z-r r^rx Weight of the viscera (g

[0209] 6. Vi scerosomatic Index (V SI) = - - - —

[0210] Body weight (g)

[0211] >TT•Ti Weight of the liver (g

[0212] 7. Hep

[0213] 1atosomatic Index (

[0214] vHSI) = - - - — — —

[0215] Body weight (g)

[0216] Proximate composition

[0217] Proximate analysis of the feeds and fish at the end of the experiment was conducted to determine whether feed supplementation with the bacteria influenced the proximate composition of the feed and fish, according to AO AC (1990).

[0218] The effect on the gut microbiome

[0219] At the end of the feeding trial, 6 juveniles per treatment were randomly selected, euthanized, and weighed. They were dissected to obtain hindgut tissues which were also weighed and immediately transferred into sterile 1 ,5-mL Eppendorf tubes on ice. Additionally, the sampling was repeated 10 days after cessation of probiotic feeding, and fish were fed on a bacteria-free commercial diet during this period. This was done to find out if the putative probiotic strains could still be present in the GIT of fish as well as their influence on the microbiome after cessation of probiotic feeding. Genomic DNA was extracted from the samples with a QIAamp Fast DNA stool mini kit (Qiagen, Germany) according to manufacturer’s instructions. The DNA was stored at -20°C until metagenomic analysis was done. All samples were subjected to targeted full-length 16S rRNA sequencing on a PromethlON device to identify the bacterial composition and diversity within the samples. Samples with a sequencing output >100,000 reads were randomly subsampled to 100,000 reads. All reads were then filtered based on quality and length and samples that had no reads left after filtering were left out of further analyses. Rarefaction plots, representing the number of observed taxa against the number of sequenced reads per sample were generated by performing an iterative random subsampling of the total reads. Rarefaction curves that reached a plateau phase, indicated that sufficient sequencing depth was obtained for those samples, and were the only ones selected for further statistical analyses. Each read was mapped against sequences in the SILVA database to obtain taxonomic classification of the reads using the EMU algorithm. The relative abundance of all bacterial phyla, families, genera and species in the samples were determined. To find out whether there were changes in microbial diversity and composition between treatments at the two sampling points, alpha and beta diversity analyses were performed to compare the bacterial diversity and composition. Additionally, we evaluated whether there were any differentially abundant taxa between treatments at the different time points.

[0220] The challenge test

[0221] Bacterial growth conditions and preparation for challenge

[0222] A local clinical strain of Providencia spp. was obtained from the Department of Biomolecular Resources and Biolab Sciences, College of Veterinary Medicine, Animal Resources and Biosecurity at Makerere University. The strain was cultured overnight at 30°C on brain heart infusion agar (BHIA) (Hi-media, India) to isolate single colonies. These colonies were then inoculated into brain heart infusion broth (BHIB) and incubated on shaker at 30°C for 24h. The cultures were centrifuged at room temperature at 7000 xg for lOmin. The supernatant was discarded, and the pellet was washed three times with sterile normal saline (0.85g.L-1NaCl). The pellet was then suspended in sterile normal saline, vortexed to mix, and used as the stock for infection. A relationship between optical density and cell density (CFU.mL'1) was established by plating dilutions of known absorbance at 550nm on BHIA (30°C, 24h), which was used to estimate the infection doses.

[0223] The experimental infection

[0224] After the feeding trial, 30 fish were randomly picked from each treatment and stocked in 15-L troughs containing 10-liters of water in triplicate (10 fish per replicate). The control treatments were obtained from the fish fed on bacteria-free diets and were divided into the negative control (NC) and positive control (PC). Except the NC, all the other treatments were challenged with a local virulent strain of Providencia sp. at a concentration of 1.72xl08CFUmL1by an intraperitoneal (IP) injection (0.2 mL). The negative control was unchallenged, but rather injected with sterile saline water of equal volume (0.2 mL). Prior to the injection, the fish were anaesthetized with clove oil (naissance, UK) at a concentration of 0.1 mL.L'1(v / v) for 5 minutes and then placed back into well aerated challenge containers for recovery. The experiment was run for 14 days while monitoring and scoring progression of morbidity and mortality. Abnormal behaviour, clinical signs, and daily mortality were recorded. Freshly dead and moribund fish were examined for clinical manifestation of disease (necropsy). The pathogen was re-isolated from liver and spleen tissues of challenged fish by plating on selective media (violet-red bile agar - VRBA). Liver, spleen, midgut tissues, and blood samples were collected to evaluate the histopathological, immunological, and haematological effects of the pathogen on the host under the different treatments.

[0225] Total and differential leucocyte counts

[0226] Peripheral blood was collected from the caudal vein of the fish after euthanasia before the challenge (TO), day 7 (T7), and day 14 (T 14) of the challenge. A drop of blood was used to make blood smears on fresh and clean glass slides whereas the other portion was aliquoted into red-top non-heparinized tubes for serum collection. Blood smears were left to air-dry at room temperature and then stacked into glass slide holders. They were stained with Hemacolor (Sigma-Aldrich, Germany) and analysed for total leucocyte and differential leucocyte counts under a light microscope according to ANTACHE et al. (2014); Correa et al. (2017); De et al. (2003); Khunrang et al. (2023) & Wang et al. (2021).

[0227] RNA extraction and immune gene expression assay

[0228] Total RNA was extracted from spleen and midgut tissues collected on days 0, 7, and 14 of the challenge. Tissues were stored in RNA later tissue protect tubes (1.5 mL) (Qiagen, Benelux) at 4°C for 24 hours before storing them at -20°C until they RNA extraction. Total RNA was extracted using a RNeasy Mini Kit (Qiagen, Germany) according to manufacturer’s guidelines with minor modifications. Briefly, frozen tissues were thawed on ice, transferred into sterile nuclease free 1.5 mL Eppendorf tubes, and homogenized (Kimble Pellet Pestle Motor, USA) in RLT buffer mixed with P -Mercaptoethanol (10 pL.mL'1) (Sigma-Aldrich, Germany) for 2 minutes. The mixture was centrifuged (awel-MF-20R) at 15,000 xg for 3 minutes. The supernatant was transferred into fresh sterile 1.5 mL Eppendorf tubes, mixed with an equal volume of 70% ethanol (Chem-lab nv, Belgium), and transferred into a RNeasy spin column placed in a supplied 2 mL collection tube. The spin column lids were closed, and contents centrifuged at 15,000 xg for 30 seconds followed by discarding the flowthrough. To minimize DNA contamination, the extracted RNA was treated with RNase-Free DNase (Qiagen, Germany) following the guidelines of the manufacturer. To enhance the purity of the extracted RNA, a cleanup step involving addition of 500 pL RPE buffer mixed with an equal volume of 100% ethanol (HPLC grade, Chem-lab nv, Belgium) to the spin columns, followed by centrifugation at 15,000 xg for 30 seconds. The flow-through was discarded and the RPE treatment step was repeated, followed by centrifugation under similar conditions for 2 minutes this time. The flow through was discarded, spin columns transferred into fresh 2 mL collection tubes and centrifuged at 15,000 xg for 30 seconds to dry. Finally, the spin columns were placed into supplied sterile nuclease free 1.5 mL Eppendorf tubes, and 30 pL of RNase free water was added directly to the columns and centrifuged at 15,000 xg for 1 minute to elute the RNA. The concentration and purity of the eluted RNA were measured using a Nanodrop 2000 Spectrophotometer (Thermo Scientific). The cDNA of each sample was synthesized following the manufacturer’s protocol (thermoscientific RevertAid H Minus First Strand cDNA Synthesis Kit, Lithuania). The reaction mixture was carried out in a total volume of 20 pL, consisting of total RNA up to 1 pg, 4 pL 5x Reaction buffer, 1 pL of Ribolock RNase Inhibitor (20 U / pL), 2 pL of 10 mM dNTP mix, 1 pL of RevertAid H Minus M-MuLV Reverse Transcriptase (200 U / pL) and nuclease free water up to 20 pL. The final reaction mixture was placed in a thermal cycler, and the following programme was carried out: primer annealing at 25°C for 05 minutes, reverse transcription at 42°C for 1 hour and followed by inactivation at 70°C for 5 minutes. The samples were held at 4°C. After cDNA synthesis, it was diluted with 180 pL of nuclease free water to make a total volume of cDNA of 200 pL.

[0229] Quantitative Realtime PCR

[0230] Relative quantification of mRNA levels of immune markers (C3, IL-ip, IL8, IL10, TNF-a, and 1L6) in the midgut and spleen of Nile tilapia juveniles was performed by Realtime qPCR using a Maxima SYBR Green / ROX qPCR Master Mix (2X). (Thermo Scientific, Lot. 2974659, Lithuania). The housekeeping gene EFl was used as an internal control to normalize gene expression data. The reaction mixture was carried out in a total volume of 20 pL consisting of 12.5 pL of Maxima SYBR Green / ROX qPCR Master Mix (2X), 2 pl cDNA, 3.5 pL of nuclease free H2O and 1 pL of each primer. The PCR cycling conditions were as follows: Pre-heating at 50.0°C for 2 minutes, an initial denaturation at 95°C for 10 minutes, followed by 40 cycles of denaturation at 95°C each for 15 s and 40 cycles of primer annealing and extension at 60°C each for 60 s. At the end of the amplification phase, a melting curve analysis was performed to confirm the specificity of the PCR product. The gene's relative expression in each sample versus the control in comparison to the P-actin / EFl genes were calculated according to the 2AAClmethod (Pfaffl, 2001). Data analysis

[0231] Growth performance parameters were statistically compared between treatments over the different sampling times using analysis of variance (ANOVA) after examining normality and homoscedasticity. Survival analysis was done with the Kaplan-Meier method with logrank pairwise comparisons between treatments. A Cox Proportional -Hazards Model was used to analyse the effect of bacterial supplementation on the survival of infected juveniles with the positive control as a reference treatment. Proximate composition for all nutrients between the different treatments was done by a Kruskal-Wallis test. An independent sample T-test was also conducted to compare proximate nutrient composition of probiotic treatments with the control. A Wilcoxon rank sum test was used to assess significant differences in microbiome Shannon diversity between treatments at the two time points. For beta P-diversity, analysis was performed to evaluate the dissimilarity in bacterial communities between samples. Here, the Bray-Curtis dissimilarity index was used to determine the degree of community differentiation between all samples within a data set. These were visualized using Principal Coordinates Analysis (PCoA) plots, which allowed us to assess dissimilarity between all samples and visually evaluate which samples cluster closer to each other. The community dissimilarity between the 2 timepoints for each treatment group was statistically tested for significant differences using a Permutational ANOVA (PERMANOVA) test with 9,999 permutations. Linear discriminant analysis (LDA) effect size (LEFSe) was used to identify the bacterial species with a significantly different relative abundance between the different evaluated groups at the different time points. Only significant taxa with a LDA log score >2.0 were considered in the results. For total and differential leucocyte counts, comparisons between treatments at the different time points was done by a One-way analysis of means (not assuming equal variances), followed by pairwise comparisons using T tests with pooled SD.

[0232] Results for example 3

[0233] Growth performance

[0234] Throughout the 60-days of feeding Nile tilapia juveniles with a local commercial feed supplemented with the two putative probiotic strains Lysinibacillus fusiformis LFUG and Priestia megaterium PMUG01 at two different concentrations (T1 - IxlO6and T2 - IxlO8CFU.g1of feed), the physiochemical water quality parameters were recorded (DO - 6.1±0.3 mg.L'1, Temperature - 24.7±1.3 °C and pH - 8.2±0.1), and the dynamics in growth performance parameters are summarized in (Table 6). Furthermore, there was no significant difference in the condition factor, visceral somatic index & hepatosomatic indices (HSI), and survival rates across all treatments.

[0235] For the SGR and FCR, fish fed on probiotic treated diets exhibited significantly higher specific growth rates and feed conversion ratios than the control except the LFUGT1 treatment (Figure 5 Panel A & B respectively).

[0236] The weight gain changes over the respective 10-day sampling intervals were compared between treatments and treatment PMUGT2 was the only treatment whose fish gained significantly more weight than the control (p = 0.008). After 20 days, all the other probiotic treatments significantly enhanced the weight gain except LFUGT1. A similar trend was maintained until day 50, when LFUGT1 also significantly enhanced the weight gain compared to the control (p = 0.001). Additionally, the cumulative survival rates for the different treatments at the different sampling intervals were computed. All treatments including the control presented a similar trend in mortalities which corresponded to statistically similar final survival rates between treatments. However, PMUGT2 exhibited the lowest survival rate (86.0±8.0%), followed by for PMUGT1 (88.0±4.0%) and LFUG treatments which recorded the same survival rate of 92.0% (Table 6). Proximate composition

[0237] Compared to the negative control, there were no significant differences in the proximate nutrient composition of the different feed treatments (data not shown). Similarly, there were not significant differences in nutrient composition for all the nutrients between treatments (X2= 7.875, df = 4, p = 0.096). However, when we performed an independent sample T-test to compare probiotic treatments with the control, no significant differences were detected for all the nutrients except the crude protein content, which was significantly higher in fish fed on PMUGT2 and LFUGT1 treatments compared to the control (p = 0.017 and 0.007 respectively).

[0238] The effect on the gut microbiome

[0239] The alpha diversity was estimated for all treatments as the Shannon species diversity at the end of the 60-days feeding trial, and 10 days after termination of the feeding trail while feeding all treatments on a basal diet, devoid of probiotic supplementation. Significant differences in Shannon species diversity were only detected after 10 days without probiotic feeding (p = 0.002) but not at the end of the 60-days of probiotic feeding (p = 0.052), and only between the control and PMUGT1 (p = 0.043). Also, there were no significant differences in Shannon species diversity within treatments at both sampling times except LFUGT2 (p = 0.004) and PMUGT1 (p = 0.041).

[0240] Furthermore, the within treatment divergences in the bacterial communities at the two time points are presented as PCoA plots of the P-diversity Bray-Curtis dissimilarity indices. Significant differences in the bacterial communities between treatments at the two time points were detected in LFUGT2, PMUGT1 and LFUGT1 but not the control and PMUGT2. Whereas both indices showed no significant differences in bacterial community structure between the control and the other treatments at dayO, the PERMANOVA analysis of the P-diversity Bray-Curtis dissimilarity indices identified significant dissimilarities in bacterial communities between the control and all probiotic treatments at day 10, contrary to the alpha-Shannon diversity analysis, which portrayed significant differences between the control and only PMUGT1 at day 10. To test the changes and differences in dominant taxa between the treatments at the two time points, linear discriminant analysis (LDA) effect size (LEFSe) was used to identify the bacterial species with a significantly different relative abundance within the different groups using only significant taxa with an LDA log score > 2.0. Probiotic treatments especially on dayO (during probiotic feeding) were significantly more dominated with probiotic taxa than the control, which was dominated by opportunistic pathogenic bacteria. These included: Rhodobacter sp, Allorhizobium- Neorhizobium- Pararhizobium Rhizobium sp, Phreatobacter sp, Tepidiphilus succinalimandens. Romboutsia sp, Schlegelella sp, Bosea sp, Vulcaniibacterium sp, etc, as opposed to the pathogenic taxa, that were more dominant in the control (i.e., Aeromonas sp, Plesiomonas sp, and Shewanella sp ). After cessation of feeding with probiotic-supplemented diets, opportunistic pathogenic species, which were dominant in the control, also became notably abundant in probiotic treatments, contrary to when fish were still being fed on the probiotics. Immune gene expression post feeding trial

[0241] After 60 days of feeding Nile tilapia juveniles on a commercial feed supplemented with the two local putative probiotic strains PMUG and LFUG at concentrations (T1 - IxlO6CFU.g1and T2 - IxlO8CFU.g'1), the expression of four immune genes (IL1B, IL8, C3, and IL6) was monitored in the spleen and midgut, and the results are presented in Figure 6. In the spleen, IL8 was significantly underregulated in all probiotic treatments compared to the control. However, in the midgut, IL8 was significantly upregulated in PMUGT1 and LFUGT2 treatments, while the other probiotic treatments exhibited significant under-regulation compared to the control. For IL6, in the spleen, a similar trend was observed as with IL8, where IL6 was underregulated in most probiotic treatments. However, in the midgut, there were no significant differences in IL6 expression across the treatments compared to the control, except for PMUGT1 and LFUGT2, where IL6 was significantly upregulated. C3 expression in the spleen was significantly underregulated in all probiotic treatments compared to the control. However, in the midgut, PMUGT1 and LFUGT1 treatments showed significant upregulation of C3 compared to the control, unlike PMUGT2, where C3 was significantly underregulated. The expression of C3 in LFUGT2 was not significant different from the controls. For IL1B, in the spleen, it was significantly underregulated in all probiotic treatments compared to the control. In contrast, in the midgut, IL1B was significantly upregulated in all probiotic treatments compared to the control except for PMUGT2, where its expression was like that in the control.

[0242] The experimental infection

[0243] After feeding Nile tilapia juveniles with a diet supplemented with our strains LFUG and PMUG01 at concentrations T1 and T2 for 60 days, they were challenged with a local virulent strain of Providencia sp. for 14 days. The survival rates and survival probabilities, as represented in Figure 7, show that there were significant differences in the survival rates between the NC (96.7±5.8%) and PC (34.8±21.7%) treatments, but not the other probiotic treatments. Even though the negative control treatment exhibited the highest survival rate, there was no significant difference compared to LFUGT2 and PMUGT2 (90.0±10.0%), PMUGT1 (86.7±5.8%), and LFUGT1 (83.3±11.5%). There were no significant differences in survival rates between the probiotic treatments. Moreover, the Cox proportional hazards regression analysis revealed significant differences in hazard ratios among the treatment groups, with the positive control treatment as the reference group. As expected, the negative control (uninfected) treatment exhibited the most substantial risk reduction (95.24%), with the lowest significant hazard ratio (0.049), followed by LFUGT2 and PMUGT2 (0.148), PMUGT1 (0.197), and LFUGT1 (0.24821) (Table 7). These results suggest a substantial reduction in Providencia sp. -induced mortality risk with PMUG01 and LFUG feed supplementation irrespective of concentration (T1 or T2).

[0244] Total and differential leucocyte counts

[0245] Analysis of immune cell counts including total leucocytes, neutrophils, lymphocytes, monocytes, and basophils of juvenile Nile tilapia subjected to various probiotic treatments for 60 days and then challenged with virulent clinical Providencia sp. Samples were taken on day 0 before the challenge, and then on days 7 and 14 after the challenge to monitor the immune responses across the different treatments. Results are described as follows:

[0246] On day 0, before the challenge, all probiotic treatments resulted in significantly higher neutrophil, lymphocyte, monocyte, basophil counts, and total leucocyte counts (TLC) compared to the negative and positive controls.

[0247] By day 7, after the pathogen challenge, all probiotic treatments having significantly lower TLCs than the negative control except PMUGT2. The positive control had significantly lower TLCs than the negative control and PMUGT2 but higher than LFUGT1. Neutrophil counts in all probiotic treatments, except PMUGT2, were significantly lower than the negative control. Lymphocyte counts were significantly lower for all probiotic treatments compared to the negative control, except for PMUGT2. The positive control had significantly higher lymphocyte counts than LFUGT1 but lower than PMUGT2 and the negative control. Monocyte counts were lower in all probiotic treatments and the positive control compared to the negative control, except for PMUGT2. Basophil counts were also lower in all probiotic treatments compared to the negative control, except for PMUGT2, with LFUGT1 showing the lowest counts.

[0248] On Day 14, all probiotic treatments and the positive control had significantly higher TLCs than the negative control. However, the negative control had higher TLCs than all the probiotic treatments. Whereas LFUGT1 had the lowest TLCs among the other probiotic treatments, LFUGT2 had higher TLCs than LFUGT1 but lower than PMUGT1, as opposed to PMUGT2, which had higher TLCs than PMUGT1. Furthermore, all probiotic treatments had significantly lower neutrophil counts than the negative control, with the positive control having higher counts than all probiotic treatments. LFUGT1 had the lowest neutrophil counts, and LFUGT2 had lower counts than PMUGT1 but not PMUGT2. For PMUGT2, the neutrophil count was lower than PMUGT1. Lymphocyte counts were lower in probiotic-treated fish compared to the negative control, while the positive control had higher counts than all probiotic treatments. LFUGT1 had the lowest counts, and for LFUGT2, it was lower than PMUGT1 but not PMUGT2. PMUGT1-treated fish had higher counts than PMUGT2. Monocyte counts were higher in all treatments, including the positive control, compared to the negative control, with the positive control having higher counts than probiotic treatments. LFUGT1 consistently had the lowest monocyte counts, and LFUGT2 had lower counts than PMUGT1 but not PMUGT2. PMUGT1 expressed significantly higher monocytes than PMUGT2. Basophil counts were higher in all probiotic treatments compared to the negative control but lower than the positive control. LFUGT1 had the lowest counts, and for LFUGT2, it was lower than PMUGT1 but not PMUGT2, while PMUGT2 exhibited higher counts than PMUGT1.

[0249] Immune gene expression during infection

[0250] After the feeding trial with probiotic fortified diets, the effect of Providencia sp. on the expression of the four immune genes (IL8, C3, IL6 and IL1B) in the spleen and midgut was monitored at days 7 and 14 post challenge, and the results are summarized in (Table 8). Seven days after the challenge, IL8 expression in the spleen was significantly upregulated in all probiotic treatments and the positive control compared to the negative control, except for the PMUGT1 treatment, where IL8 was instead significantly under-expressed. In the midgut, IL8 was significantly downregulated in the positive control and LFUG treatments compared to the negative control, whereas the PMUGT1 treatment showed significant upregulation, and PMUGT2 showed no significant difference from the negative control. Compared to the positive control, IL8 was significantly upregulated in both PMUG treatments in the midgut, while in LFUGT1 treatment, it was significantly downregulated, and LFUGT2 showed no significant difference. For IL6 in the spleen, significant upregulation was observed in the positive control and LFUGT2 compared to the negative control. IL6 was significantly downregulated in the PMUG treatments, while in the LFUGT1 treatment, it showed no difference compared to the negative control. Compared to the negative control, IL6 was significantly down regulated in all other treatments except LFUGT2, where it was significantly upregulated. In the midgut, IL6 was significantly upregulated in the positive control and other probiotic treatments compared to the negative control, with a similar trend observed when compared to the positive control, except for LFUGT2, which was not different from the positive control. C3 expression in the spleen showed no significant difference between the positive control, PMUGT1, and LFUGT2 compared to both controls. However, C3 was significantly downregulated in LFUGT1 and PMUGT2 compared to both the negative and positive controls. In the midgut, C3 was significantly upregulated in both PMUG treatments, while it was significantly downregulated in the LFUG treatments and the positive control compared to the negative control. Significant upregulation of C3 was observed in all probiotic treatments compared to the positive control. For IL1B expression in the spleen, significant upregulation was recorded in the positive control and LFUGT1 compared to the negative control, while the other probiotic treatments showed no difference. Compared to the positive control, IL IB levels were significantly higher in LFUGT1, not different in PMUGT2, and significantly lower in LFUGT2 and PMUGT1. In the midgut, IL1B was significantly upregulated in PMUGT treatments compared to the negative and positive controls, while LFUG treatments showed no significant difference from either control.

[0251] On day 14, IL8, expression in the spleen was significantly upregulated in the positive control and all probiotic treatments compared to the negative control. IL8 was significantly upregulated in all probiotic treatments except PMUGT2 compared to the positive control. In the midgut, IL8 was significantly downregulated in all treatments compared to the negative control, with all probiotic treatments showing significantly lower expression compared to the positive control, except LFUGT2, which showed significantly higher levels. IL6 expression was significantly downregulated in all treatments compared to the negative control. Compared to the positive control, there was no difference in IL6 expression in LFUGT1, while it was significantly downregulated in the other probiotic treatments in the spleen. A similar pattern was observed in the midgut compared to the negative control, with LFUGT2 and PMUGT 1 significantly upregulating IL6, while LFUGT1 and PMUGT2 significantly downregulated IL6 compared to the positive control. C3 expression in the spleen was significantly upregulated in the positive control and all other probiotic treatments except PMUGT 1, where it was significantly down regulated compared to both controls. In the midgut, significant downregulation of C3 was recorded in the positive control and all probiotic treatments compared to the negative control. Compared to the positive control, significant upregulation of C3 was observed in PMUGT1 and LFUGT2, while LFUGT1 and PMUGT2 showed significant downregulation. For IL1B expression in the spleen, there was no significant difference across all treatments compared to the controls, except in LFUGT1, where it was significantly upregulated. In the midgut, IL1B was significantly downregulated across all treatments compared to the negative control, except in PMUGT 1, where the levels were significantly higher than in the positive control. The other probiotic treatments showed similar IL1B expression profiles compared to the positive control. Tables for example 3

[0252] Table 6: Growth performance indicators of Nile tilapia juveniles after treatment with a commercial feed supplemented with the two putative probiotic strains LFUG and PMUG01 at two different concentrations T1 and T2 for 60 days. The concentrations T1 and T2 were IxlO6and IxlO8CFU.g1of feed. Data is presented as means standard deviation and values in the same row with different superscripts depict significant differences between the respective treatments. IABW - Initial average body weight; FA W Final average body weight; A WG - Average weight gain; K - Condition factor; FCR - Feed conversion ratio; S - Survival rate, HIS - Hepato somatic index; and VSI- Visceral somatic index

[0253] Parameter Control PMUGT1 PMUGT2 LFUGT1 LFUGT2 IABW (g) 1.763±0.032 1.786±0.072 1.770±0.096 1.688±0.024 1.714±0.058 FABW (g) 7.179±0.584 8.686±0.626 9.797±1.508 8.115±0.720 9.025±0.411 AWG(g) 5.417±1.988a6.901±1.715bc8.034±2.316e6.429±1.590b7.305±1.276cdSGR (%.Day-1) 2.428±0.285a2.728±0.288b3.027±0.490b2.616±0.318ab2.867±0.289aK 3.127±0.654a3.000±0.540a3.034±0.660a3.034±0.567a3.009±0.581aFCR 1.282±0.192a1.112±0.151b1.050±0.172b1.167±0.158a1.056±0.100bS (%) 90.0±2.0a88.0±4.0a86.0±8.0a92.0±5.3a92.0±7.2aHSI 9.063±1.460a9.179±1.726a8.589±1.131a9.079±1.246a8.199±1.372aVSI 1.568±0.507a1.448±0.515a1.518±0.254a1.446±0.402a1.341±0.497a Table 7: The Cox regression analysis showing the % reduction in mortality by Juvenile Nile tilapia from a virulent clinical strain of Providencia spp. following a 60-days feeding trial with a commercial diet supplemented with our putative probiotic strains LFUG and PMUG01 at concentrations T1 and T2 using the positive control as our reference treatment for of risk progression. The concordance index (0.779), likelihood ratio test (fr = 30.61, df= 5, p < 0.00001), Wald test (fr = 28.93, df = 5, p < 0.00002), and Score (logrank) test (fr = 40.53, df = 5, p < 0.0000001).

[0254] Regression Hazard % Risk

[0255] Treatments coefficients Ratios Reduction z-score P (> z ) sig.

[0256] LFUGT1 -1.393 0.248 75.179 -2.771 0.006 ** LFUGT2 -1.913 0.148 85.238 -3.079 0.002 ** NC -3.021 0.049 95.124 -2.944 0.003 ** PMUGT1 -1.622 0.197 80.256 -2.948 0.003 ** PMUGT2 -1.910 0.148 85.185 -3.073 0.002 ** Table 8: Relative changes in juvenile Nile tilapia immune gene expression in the spleen and midgut at days 7 and 14 post an intraperitoneal infection with a clinical strain of Providencia sp. The targeted immune genes included: IL8; Interleukin 8, IL6; Interleukin 6, C3; Compliment Component C3, and Interleukin 1 beta. Data is represented as the mean slandard deviation and treatments in the same row for a given immune gene with different superscripts are significantly different compared to the negative and positive controls

[0257] Time Tissue Target NC PC PMUGT1 LFUGT1 PMUGT2 LFUGT2

[0258] 7 Spleen IL8 1.0±0.2a4.9±0.2b1.4±0.1ac4.7±1.4b3.1±0.3b5.9±0.6bIL6 1.0±0.2a2.7±0.2bO.3±O.OCl.l±0.1acO.1±O.OC6.8±l.lcC3 l.O±O.Oal.l±0.1a1.5±0.0aO.O±O.ObO.O±O.Ob1.5±0.3aILip 1.0±0.2a2.8±0.0bO.5±O.Oac12.2±0.8c2.3±0.6ab0.4±0.0ac14 IL8 1.0±0.1a1.3±0.0b1.7±0.1c5.6±0.5c1.4±0.1b2.1±0.0cIL6 1.0±0.1aO.3±O.ObO.1±O.OC0.4±0.0b0.2±0.0cO.1±O.OCC3 l.O±O.OaO.3±O.ObO.O±O.OC26.4±l.lc3.5±0.3c7.9±0.9cILip 1.0±0.1a2.9±1.3a2.7±0.4a5.7±0.7b2.1±0.1a2.4±0.3a

[0259] 7 Mdigut IL8 1.0±0.1aO.5±O.Ob3.0±0.1cO.3±O.OC1.2±0.1acO.5±O.ObIL6 1.0±0.1a2.9±0.8b21.1±4.4C7.9±0.2c12.4±0.4c4.5±2.1bC3 l.O±O.OaO.O±O.Ob23.3±0.6cO.1±O.OC6.1±0.1cO.1±O.OCILip l.O±O.Oa1.9±0.1a12.4±1.8b2.0±0.3a32.1±1.0b2.0±0.2a

[0260] Day IL8 1.0±0.1a0.2±0.0bO.1±O.OCO.1±O.OCO.O±O.OC0.4±0.0c14

[0261] IL6 1.0±0.1aO.O±O.ObO.O±O.OCO.O±O.OCO.O±O.OCO.3±O.OCC3 1.0±0.1aO.O±O.Ob0.0±0.0bcO.O±O.OCO.O±O.OCO.5±O.OCILip 1.0±0.1aO.l±O.Ob2.5±0.4cO.O±O.ObO.O±O.Ob0.4±0.1b References

[0262] 1. Alcock, B. P., Huynh, W., Chalil, R., Smith, K. W., Raphenya, A. R., Wlodarski, M.

[0263] A., Edalatmand, A., Petkau, A., Syed, S. A., Tsang, K. K., Baker, S. J. C., Dave, M., McCarthy, M. C., Mukiri, K. M., Nasir, J. A., Golbon, B., Imtiaz, H., Jiang, X., Kaur, K., ... Mcarthur, A. G. (2023). CARD 2023: expanded curation, support for machine learning, and resistome prediction at the Comprehensive Antibiotic Resistance Database. Nucleic Acids Research, 57(1 D), D690-D699. https: / / doi.org / 10.1093 / nar / gkac920

[0264] 2. ANTACHE, A., CRISTEA, V, GRECU, I., DEDIU, L., CRETU, M., BOCIOC, E., & PETREA, S. M. (2014). Effects of Dietary Supplementation at Nile tilapia with Thymus vulgaris, Trigonela foenum graecum and Azadirachta indica on Welfare Status. Bulletin of University of Agricultural Sciences and Veterinary Medicine Cluj- Napoca. Animal Science and Biotechnologies, 77(2). https: / / doi.Org / 10.15835 / buasvmcn-asb:10386

[0265] 3. Balasubramanian, S., Bettin Thomas, T., Mathavan, D., Sathish Kumar, R., Uma, G., Stevens Jones, R. D., & Citarasu, T. (2023). Isolation and Screening of Probiotic Bacteria from the Gut of Polychaetes as a Probiotic Potential for Fish Aquaculture. Nature Environment and Pollution Technology, 22(2), 861-868. https: / / doi.org / 10.46488 / NEPT.2023.v22i02.028

[0266] 4. Beryl Chean, M. Y, Puvanasundram, P., Yaminudin, J., & Karim, M. (2021).

[0267] Evaluation of antagonism activity and control of vibrio alginolyticus in artemia culture using mixed probiotic. Pertanika Journal of Tropical Agricultural Science, 44(1), 117-137. https: / / doi.org / 10.47836 / PJTAS.44.E07

[0268] 5. Correa, S. A. da S., Abessa, D. M. de S., Santos, L. G. dos, da Silva, E. B., & Seriani, R. (2017). Differential blood counting in fish as a non-destructive biomarker of water contamination exposure. Toxicological and Environmental Chemistry, 99(3), 482- 491. https: / / doi.org / 10.1080 / 02772248.2016.1189554

[0269] 6. De, N., Rodrigues Bittencourt, L., Molinari, L. M., De Oliveira Scoaris, D., Pedroso, R. B., Nakamura, C. V, Ueda-Nakamura, T., Alves De Abreu Filho, B., Prado, B., & Filho, D. (2003). Haematological and biochemical values for Nile tilapia Oreochromis niloticus cultured in semi-intensive system. In Acta Scientiarum. Biological Sciences Maringd (Vol. 25, Issue 2).

[0270] 7. Defoirdt, T. (2018). Quorum-Sensing Systems as Targets for Antivirulence Therapy.

[0271] In Trends in Microbiology (Vol. 26, Issue 4, pp. 313-328). Elsevier Ltd. https: / / doi.Org / 10.1016 / j.tim.2017.10.005

[0272] 8. Defoirdt, T., Thanh, L. D., Van Delsen, B., De Schryver, P., Sorgeloos, P., Boon, N., & Bossier, P. (2011a). N-acylhomoserine lactone-degrading Bacillus strains isolated from aquaculture animals. Aquaculture, 377(1-4), 258-260. http s : / / doi . org / 10.1016 / j . aquaculture .2010.11.046

[0273] 9. Defoirdt, T., Thanh, L. D., Van Delsen, B., De Schryver, P., Sorgeloos, P., Boon, N., & Bossier, P. (2011b). N-acylhomoserine lactone-degrading Bacillus strains isolated from aquaculture animals. Aquaculture, 377(1-4), 258-260. http s : / / doi . org / 10.1016 / j . aquaculture .2010.11.046

[0274] 10. El-Saadony, M. T., Alagawany, M., Patra, A. K., Kar, I., Tiwari, R., Dawood, M. A.

[0275] O., Dhama, K., & Abdel-Latif, H. M. R. (2021). The functionality of probiotics in aquaculture: An overview. In Fish and Shellfish Immunology (Vol. 117, pp. 36-52). Academic Press. https: / / doi.Org / 10.1016 / j.fsi.2021.07.007

[0276] 11. Ghanei-Motlagh, R., Mohammadian, T., Gharibi, D., Khosravi, M., Mahmoudi, E., Zarea, M., ELMatbouli, M., & Menanteau-Ledouble, S. (2021). Quorum quenching probiotics modulated digestive enzymes activity, growth performance, gut microflora, haemato-biochemical parameters and resistance against Vibrio harveyi in Asian seabass (Lates calcarifer). Aquaculture, 531. https: / / doi.Org / 10.1016 / j.aquaculture.2020.735874

[0277] 12. Gopu, V, & Shetty, P. H. (2016). Regulation of acylated homoserine lactones (AHLs) in beef by spice marination. Journal of Food Science and Technology, 53(6), 2686- 2694. https : / / doi .org / 10.1007 / s 13197-016-2240-x

[0278] 13. Goris, J., Konstantinidis, K. T., Klappenbach, J. A., Coenye, T., Vandamme, P., & Tiedje, J. M. (2007). DNA-DNA hybridization values and their relationship to wholegenome sequence similarities. International Journal of Systematic and Evolutionary Microbiology, 57(1), 81-91. https: / / doi.Org / 10.1099 / ijs.0.64483-0

[0279] 14. Hossain, T. J., Das, M., Ali, F., Chowdhury, S. I., & Zedny, S. A. (2021). Substrate preferences, phylogenetic and biochemical properties of proteolytic bacteria present in the digestive tract of Nile tilapia (Oreochromis niloticus). AIMS Microbiology, 7(4), 528-545. https: / / doi.org / 10.3934 / microbiol.2021032 15. International, A. (n.d.). By Authority Of THE UNITED STATES OF AMERICA Legally Binding Document.

[0280] 16. Ismail, R. F., Saleh, N. E., El-Din, A., & Sayed, H. (n.d.). Impacts of microplastics on reproductive performance of male tilapia (Oreochromis niloticus) pre-fed on Amphora coffeaeformis . https: / / doi.org / 10.1007 / sll356-021-14984-2 / Published 17. Jinendiran, S., Boopathi, S., Sivakumar, N., & Selvakumar, G. (2019). Functional Characterization of Probiotic Potential of Novel Pigmented Bacterial Strains for Aquaculture Applications. Probiotics and Antimicrobial Proteins, 77(1), 186-197. https: / / doi.org / 10.1007 / sl2602-017-9353-z

[0281] 18. Jolley, K. A., Bliss, C. M., Bennett, J. S., Bratcher, H. B., Brehony, C., Colles, F. M., Wimalarathna, H., Harrison, O. B., Sheppard, S. K., Cody, A. J., & Maiden, M. C. J. (2012). Ribosomal multilocus sequence typing: Universal characterization of bacteria from domain to strain. Microbiology, 158(A), 1005-1015. https: / / doi.Org / 10.1099 / mic.0.055459-0

[0282] 19. Khunrang, T., Pooljun, C., & Wuthisuthimethavee, S. (2023). Correlation of Streptococcus agalactiae concentration on immune system and effective dose of inactivated vaccine for Chitralada 3 strain Nile tilapia (Oreochromis niloticus) in Thailand. BMC Veterinary Research, 19(\). https: / / doi.org / 10.1186 / sl2917-023- 03835-6

[0283] 20. Kolmogorov, M., Yuan, J., Lin, Y, & Pevzner, P. A. (2019). Assembly of long, error- prone reads using repeat graphs. Nature Biotechnology, 37(5), 540-546. https : / / doi .org / 10.1038 / s41587-019-0072-8

[0284] 21. Kuebutomye, F. K. A., Abarike, E. D., & Lu, Y. (2019). A review on the application of Bacillus as probiotics in aquaculture. In Fish and Shellfish Immunology (Vol. 87, pp. 820-828). Academic Press. https: / / doi.Org / 10.1016 / j.fsi.2019.02.010

[0285] 22. Kuebutomye, F. K. A., Lu, Y, Abarike, E. D., Wang, Z., Li, Y, & Sakyi, M. E. (2020).

[0286] In vitro Assessment of the Probiotic Characteristics of Three Bacillus Species from the Gut of Nile Tilapia, Oreochromis niloticus. Probiotics and Antimicrobial Proteins, 12(2), 412-424. https: / / doi.org / 10.1007 / sl2602-019-09562-5

[0287] 23. LaSarre, B., & Federle, M. J. (2013). Exploiting Quorum Sensing To Confuse Bacterial Pathogens. Microbiology and Molecular Biology Reviews, 77(1), 73-111. https: / / doi.org / 10.1128 / mmbr.00046-12

[0288] 24. Li, H. (2018). Minimap2: Pairwise alignment for nucleotide sequences.

[0289] Bioinformatics, 34(\A), 3094-3100. https: / / doi.org / 10.1093 / bioinformatics / btyl91 Liu, B., Zheng, D., Jin, Q., Chen, L., & Yang, J. (2019). VFDB 2019: A comparative pathogenomic platform with an interactive web interface. Nucleic Acids Research, 47(f)V), D687-D692. https: / / doi.org / 10.1093 / nar / gkyl080

[0290] Liu, C. H., Chiu, C. S., Ho, P. L., & Wang, S. W. (2009). Improvement in the growth performance of white shrimp, Litopenaeus vannamei, by a protease-producing probiotic, Bacillus subtilis E20, from natto. Journal of Applied Microbiology , 707(3), 1031-1041. https: / / doi.Org / 10.llll / j.1365-2672.2009.04284.x

[0291] Liu, Y. Y, Hsu, C. Y, Yang, Y. C., Huang, C. H., & Chen, C. C. (2023). ProbioMinServer: an integrated platform for assessing the safety and functional properties of potential probiotic strains. Bioinformatics Advances, 3(1). https: / / doi.org / 10.1093 / bioadv / vbadl53

[0292] Lu, S., Na, K., Li, Y, Zhang, L., Fang, Y, & Guo, X. (2022). Bacillus-derived probiotics: metabolites and mechanisms involved in bacteria-host interactions. In Critical Reviews in Food Science and Nutrition. Taylor and Francis Ltd. https: / / doi.org / 10.1080 / 10408398.2022.2118659

[0293] Meier-Kolthoff, J. P, & Goker, M. (2019a). TYGS is an automated high-throughput platform for state-of-the-art genome-based taxonomy. Nature Communications, 10( ). https: / / doi.org / 10.1038 / s41467-019-10210-3

[0294] Meier-Kolthoff, J. P, & Goker, M. (2019b). TYGS is an automated high-throughput platform for state-of-the-art genome-based taxonomy. Nature Communications, 10(\). https: / / doi.org / 10.1038 / s41467-019-10210-3

[0295] Melo-Bolivar, J. F., Ruiz Pardo, R. Y, Hume, M. E., & Villamil Diaz, L. M. (2021). Multistrain probiotics use in main commercially cultured freshwater fish: a systematic review of evidence. \n Reviews in Aquaculture (Vol. 13, Issue 4, pp. 1758-1780). John Wiley and Sons Inc. https: / / doi.org / 10.llll / raq.12543

[0296] Paray, A. A., Singh, M., & Amin Mir, M. (2023). Gram Staining: A Brief Review. International Journal of Research and Review, 70(9), 336-341. https: / / doi.org / 10.52403 / ijrr.20230934

[0297] Parks, D. H., Imelfort, M., Skennerton, C. T., Hugenholtz, P, & Tyson, G. W. (2015). CheckM: Assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome Research, 25(f), 1043-1055. https: / / doi.org / 10.1101 / gr.186072.114

[0298] Rama Devi, K., Srinivasan, R., Kannappan, A., Santhakumari, S., Bhuvaneswari, M., Rajasekar, P, Prabhu, N. M., & Veera Ravi, A. (2016). In vitro and in vivo efficacy of rosmarinic acid on quorum sensing mediated biofilm formation and virulence factor production in Aeromonas hydrophila. Biofouling, 32(10), 1171-1183. https: / / doi.org / 10.1080 / 08927014.2016.1237220

[0299] 35. Rios Castro, N., Zavala, J., Gil-Kodaka, P., & Diringer, B. (2022). Evaluation of bacterial strains to improve the productivity of microalgae used in bivalve hatcheries in Peru. Journal of the World Aquaculture Society, 53(1), 95-105. https: / / doi.org / 10.llll / jwas.12858

[0300] 36. Ruan, J., & Li, H. (2020). Fast and accurate long-read assembly with wtdbg2. Nature Methods, 17(2), 155-158. https: / / doi.org / 10.1038 / s41592-019-0669-3

[0301] 37. Schroeder, A. B., Dobson, E. T. A., Rueden, C. T., Tomancak, P., Jug, F., & Eliceiri, K. W. (2021). The ImageJ ecosystem: Open-source software for image visualization, processing, and analysis. Protein Science, 30(1), 234-249. https: / / doi.org / 10.1002 / pro.3993

[0302] 38. Serwecinska, L. (2020). Antimicrobials and antibiotic-resistant bacteria: A risk to the environment and to public health. Water (Switzerland), 72(12). https: / / doi.org / 10.3390 / wl2123313

[0303] 39. Sun, Y, Li, H., Zheng, L., Li, J., Hong, Y, Liang, P, Kwok, L. Y, Zuo, Y, Zhang, W., & Zhang, H. (2022). iProbiotics: a machine learning platform for rapid identification of probiotic properties from whole-genome primary sequences. Briefings in Bioinformatics, 23(1). https: / / doi.org / 10.1093 / bib / bbab477

[0304] 40. Tinh, N. T. N., Asanka Gunasekara, R. A. Y. S., Boon, N., Dierckens, K., Sorgeloos, P, & Bossier, P. (2007). N-acyl homoserine lactone-degrading microbial enrichment cultures isolated from Penaeus vannamei shrimp gut and their probiotic properties in Brachionus plicatilis cultures. FEMS Microbiology Ecology, 62(1), 45-53. https: / / doi.org / 10.1111 / j .1574-6941 ,2007.00378.x

[0305] 41. Vaser, R., & Sikic, M. (2021). Time- and memory-efficient genome assembly with Raven. Nature Computational Science, 7(5), 332-336. https: / / doi.org / 10.1038 / s43588-021-00073-4

[0306] 42. Vereecke, N., Van Hoorde, S., Sperling, D., Theuns, S., Devriendt, B., & Cox, E.

[0307] (2023). Virotyping and genetic antimicrobial susceptibility testing of porcine ETEC / STEC strains and associated plasmid types. Frontiers in Microbiology, 14. https: / / doi.org / 10.3389 / fmicb.2023.1139312 Wang, X., Wu, Z., Wu, S., Chen, X., Hanif, M., & Zhang, S. (2021). Hematological and cytochemical characteristics of peripheral blood cells in the argus snakehead (Ophiocephalus argus Cantor). PeerJ, 9. https: / / doi.org / 10.7717 / peerj.11234 Wick, R. R., Judd, L. M., Wyres, K. L., & Holt, K. E. (2021). Recovery of small plasmid sequences via oxford nanopore sequencing. Microbial Genomics, 7(8). https: / / doi.Org / 10.1099 / MGEN.0.000631

[0308] Widyastuti, Y. R., Mundayana, Y, Kadarini, T., Mumiasih, S., & Saputra, A. (2023). Multi-species biofloc bacteria and application test on snakehead fish (Channa striata bloch, 1793) juvenile. E3S Web of Conferences, 442. https: / / doi.org / 10.1051 / e3sconf / 202344202041

[0309] Yoon, S. H., Ha, S. min, Lim, J., Kwon, S., & Chun, J. (2017). Alarge-scale evaluation of algorithms to calculate average nucleotide identity. Antonie van Leeuwenhoek, International Journal of General and Molecular Microbiology, 110(10), 1281-1286. https : / / doi .org / 10.1007 / s 10482-017-0844-4

[0310] Yousuf, S., Tyagi, A., & Singh, R. (2023). Probiotic Supplementation as an Emerging Alternative to Chemical Therapeutics in Finfish Aquaculture: a Review. Probiotics and Antimicrobial Proteins, 75(5), 1151-1168. https: / / doi.org / 10.1007 / sl2602-022-09971-z

[0311] Zabidi, A., Yusoff, F. M., Nurul Amin, S. M., Yaminudin, N. J. M., Puvanasundram, P, & Karim, M. M. A. (2021). Effects of probiotics on growth, survival, water quality and disease resistance of red hybrid tilapia (Oreochromis spp.) fingerlings in a biofloc system. Animals, 77(12). https: / / doi.org / 10.3390 / anilll23514

[0312]

[0313] BUDAPEST TREATY ON THE INTERNATIONAL RECOGNITION OF THE DEPOSIT OF MICRO-ORGANISMS FOR THE PURPOSES OF PATENT PROCEDURE

[0314]

[0315]

[0316]

[0317]

[0318]

[0319] BUDAPEST TREATY ON THE INTERNATIONAL RECOGNITION OF THE DEPOSIT OF MICRO-ORGANISMS FOR THE PURPOSES OF PATENT PROCEDURE

[0320]

[0321]

[0322]

[0323]

Claims

Claims1. A micro-organism deposited with BCCM / LMG, Bacterial collection located at KL Ledeganckstraat 35, 9000 Gent, Belgium and having accession number LMGP-33748 or LMGP-33747, respectively, or, a micro-organism showing at least 93 percent homology to the 16SrRNA sequence of the micro-organism deposited with BCMM / LMG, Bacterial collection located at KL Ledeganckstraat 35, 9000 Gent, Belgium and having accession number LMGP-33748 or LMGP-33747, respectively.

2. A micro-organism deposited with BCCM / LMG, Bacterial collection located at KL Ledeganckstraat 35, 9000 Gent, Belgium and having accession number LMGP-33748 or LMGP-33747, respectively.

3. Use of a micro-organism according to claims 1 or 2 as probiotic in aquaculture.

Images