Towards the development of a protocol for the selection of probiotics in marine fish larviculture

Manipulation or control of the microbiological aquaculture environment has been identified as an important focus area for future hatchery development. Subsequently, alternatives to obtain control of the microbiological environment are being sought of which the field of probiotics appears highly promising. Probiotics are usually selected based on various in vitro characteristics, however, the methods used differ and are sometimes unsuccessful due to poor experimentation. The aim of this work is to contribute towards the development of a protocol for the in vitro screening of bacterial candidate probiotics for marine fish larviculture. To reduce the number of candidate probiotics to be tested in vivo, various in vitro experiments need to be conducted, each screening for a particular mode of action – antagonism towards pathogen through production of antimicrobial compounds, growth and attachment to fish intestinal mucus, and the production of other beneficial compounds such as vitamins, fatty-acids and digestive enzymes. A total of 108 bacteria species were isolated from the digestive tract of the adult common clownfish, Amphiprion percula to screen for potential probiotics to be used in clownfish larval rearing. The antagonistic compounds assay identified twelve isolates which showed antagonism towards two or more aquatic pathogens. This was followed by an in vitro test that involved growing the organisms in fish intestinal mucus and modeling their growth parameters. A ranking index (RI) was developed using the lag period (λ) and doubling time (td) of the organism, where (1over λ x td) x 100. Five candidate probionts (AP1-AP5) with varied growth parameters were used for further in vitro experiments. The attachment to mucus assay introduced a novel tool for quantifying competition for attachment sites between candidate probionts and pathogens on mucus. Candidate probiont Pseudoalteromonas AP5 reduced the attachment ability of Vibrio alginolyticus when added before the pathogen and partially out-competed the pathogen for attachment sites when added second. In vitro screening for the production of beneficial compounds tested the candidate probiotics’ ability to produce digestive enzymes - trypsin, lipase and alkaline phosphatase as well as carotenoids and vitamin C. Candidate probiont Pseudoalteromonas AP5 produced high levels of the enzymes (98.2, 34.1 and 91.3 mU product liberated.ml⁻¹, respectively) and contained carotenoids while Kocuria AP4 contained carotenoids but produced low quantities of enzymes (7.8, 0 and 59 mU product liberated.ml⁻¹, respectively). None of the candidate probiotics produced vitamin C. To eliminate potential pathogenic or toxic candidate probionts, Artemia nauplii were exposed to each candidate probiont and the percentage Artemia mortality after 24-hours was determined. Candidate probiont AP2 caused high mortality of Artemia nauplii (98.4%) and was excluded from further studies. Identification of candidate probionts AP3-AP5 was performed using 16S-rDNA molecular techniques and the bacteria were assigned the names Bacillus AP3, Kocuria AP4 and Pseudoalteromonas AP5, respectively. Two methods of larval probiont delivery were tested – attachment to Artemia, and in-water delivery. Attachment to Artemia was high for both Kocuria AP4 and Pseudoalteromonas AP5 (7.2 x10³ and 2.7x10⁴ bacteria.nauplius⁻¹, respectively) while the in-water viability experiment showed that Kocuria AP4 comprised 23.9% of the total culturable water microflora after 24 hours while Pseudoalteromonas AP5 contributed 100%. To validate the findings from the in vitro experiments, in vivo trials using clownfish larvae were performed. Of the four candidate probiotics tested, only Kocuria AP4 showed potential to increase larval survival. In vitro tests produced a better understanding of the possible mode of action and strategies of competition between bacteria, however, the number of criteria in which a candidate probiont is successful in vitro may not be the best predictor for its effectiveness in vivo. Commercial studies that reduce between-treatment variation are required to test predictions about the most suitable probiont or combinations thereof