Nguyen Thi Ngoc Tinh, Nguyen Ngoc Phuoc, Kristof Dierckens, Patrick Sorgeloos, Peter Bossier

Transcription

1 Aquaculture 253 (2006) Gnotobiotically grown rotifer Brachionus plicatilis sensu strictu as a tool for evaluation of microbial functions and nutritional value of different food types Nguyen Thi Ngoc Tinh, Nguyen Ngoc Phuoc, Kristof Dierckens, Patrick Sorgeloos, Peter Bossier Laboratory of Aquaculture and Artemia Reference Center, Ghent University, Rozier 44, 9000 Gent, Belgium Received 24 November 2004; received in revised form 6 September 2005; accepted 9 September 2005 Abstract Axenic rotifers (Brachionus plicatilis sensu strictu, clone 10) were obtained by treating amictic eggs with glutaraldehyde. Depending of the batch of rotifers, total disinfection could be obtained by exposure to ppm from 1 to 2 h at 28 C. The hatched axenic neonates were used to test the effect of microbial communities (MCs) which were isolated from either normalperforming or crashed rotifer cultures. These MCs were either used directly or were first regrown on Marine Agar. MCs were introduced to gnotobiotic Brachionus cultures in combination with three different food types, i.e. Chlorella, wild-type yeast and the mnn9 yeast mutant, which is deficient in cell wall-bound mannoprotein. In the absence of MCs or when heat-killed MCs were added, Chlorella was always the best food, while lower growth rates were observed with wild-type yeast and the mnn9 mutant as food. In the presence of live MCs and when rotifers were fed with Chlorella, the added MCs had no effect on rotifer performance. When yeasts were used as major food, all the tested MCs were able to increase the rotifer growth rate. The experiments with heatkilled MCs yielded no increase in rotifer growth rate, suggesting that the observed enhancement in rotifer growth rate was truly a probiotic effect rather than a nutritional effect. The results of this study demonstrate that gnotobiotic rotifer cultures obtained from axenic amictic eggs can be used as a test system for studying microbial-attributed as well as nutritional functions in the aquatic food chain. In addition, since the MCs originating from the crashed rotifer cultures did not decrease the growth rate in the tested rotifer cultures, it is likely that the observed crashes were not due to the presence of a standing deleterious MC Elsevier B.V. All rights reserved. Keywords: Brachionus plicatilis; Gnotobiotic rotifers; Amictic eggs disinfection; Microbial communities; Food types 1. Introduction Rotifers (Brachionus spp.) have been found to be valuable and indispensable food organisms in the industrial larviculture of fish and crustaceans throughout the world (Lubzens et al., 1997; Lee and Ostrowski, Corresponding author. Tel.: ; fax: address: Thingoctinh.Nguyen@Ugent.be (N.T.N. Tinh). 2001; Liao et al., 2001; Shields, 2001; Marte, 2003). Rotifers possess several characteristics that make them suitable live prey for the newly hatched fish larvae, e.g. relative small size, slow swimming behaviour, rapid reproduction rate, possibility to be cultured at high densities (Lubzens, 1987; Lubzens et al., 1989, 2001), possibility of bioencapsulation with highly unsaturated fatty acids, vitamins or antibiotics, which are required for the growth and survival of fish larvae (Gatesoupe, /$ - see front matter 2005 Elsevier B.V. All rights reserved. doi: /j.aquaculture

2 422 N.T.N. Tinh et al. / Aquaculture 253 (2006) ). Several studies have shown that rotifers can be used for transferring probiotic bacteria to fish larvae (Makridis et al., 2000; Martinez-Diaz et al., 2003; Rombaut et al., 1999). One of the bottlenecks in using rotifers as food, is the large diversity of microbiota associated with this filterfeeding organism. Although most of the associated bacteria are not pathogenic to rotifers, they can be easily transferred via the food chain to their larval predators and can cause detrimental effects (Dhert, 1996). The dominant bacterial groups in rotifer cultures were classified as Pseudomonas, Vibrio and Aeromonas (Nicolas et al., 1989). Verdonck et al. (1997) revealed that Vibrio species were dominantly present in the microbiota of rotifer cultures. Vibrio spp. account for 45 73% of culturable bacteria in the startcultures and 15 67% in mass production. Nicolas et al. (1989) observed 10 7 CFU ml 1 of total bacteria in the culture water by direct counts, and from 10 4 to 10 5 CFU rotifer 1 were observed in the rotifers. Skjermo and Vadstein (1993) reported large variations in the number of rotifer-associated ( CFU rotifer 1 ) and free-living bacteria ( CFU ml 1 ). Hayashi et al. (1975) and Miyakawa and Muroga (1988) came up with total aerobic rotifer-associated bacterial numbers ranging from 10 7 to CFU g 1 dry weight of rotifers, while the numbers in the culture water ranged from 10 4 to 10 7 CFU ml 1. Treatment of rotifers, as well as other live food organisms, with antibiotics or disinfectants prior to feeding to the fish larvae has become a routine practice in many hatcheries, as it can improve the larval survival rates. Various chemical and physical methods for reducing the rotifer external bacterial load have been developed by Munro et al. (1993). The bacteriolytic enzyme, lysozyme, was considered as a potential decontaminating agent. Tanasomwang and Muroga (1992) have found that the Vibrio numbers rapidly declined by four log units of CFU g 1 wet weight of rotifers after 5-h exposure to sodium nyfurstyrenate. Physical methods for surface disinfection of rotifers have been widely studied, although complete decontamination can not be achieved. Exposure of rotifers to ultraviolet radiation can reduce the bacterial load by 90% within 2 min (Munro et al., 1999). Utilization of ozone-treated seawater is also suitable for rotifer disinfection, as long as the total residual oxidants (TRO) value does not exceed the NOEC (non-observable effect concentration) value (Davis and Arnold, 1997). On the other hand, the approaches of disinfection of live food as well as fish eggs or culture water may disturb the balance of microbial communities in the larval rearing environment and result in unfavourable conditions (Olafsen, 2001). In recent years, research has focused on the bioencapsulation of rotifers and other live food organisms with selected bacteria, which can favour the growth and survival of the predating fish larvae. Axenic rotifers were used as a tool for studying the role of specific bacterial strains or microbial communities, in both nutritional and probiotic aspects. The first attempts to obtain axenic cultures of rotifers were reported a few decades ago (Dougherty et al., 1960; Plasota et al., 1980). Recently, several studies have been directed toward obtaining axenic rotifer cultures either from disinfected resting eggs (Douillet, 1998; Rombaut et al., 1999) or from disinfected parthenogenetic eggs (Martinez-Diaz et al., 2003), by using antibiotic mixtures or different kinds of disinfectants. The aim of this study is to develop a technique to obtain axenic rotifer cultures from amictic eggs using disinfectants of non-antibiotic nature. Gnotobiotic rotifer cultures obtained this way are used for evaluating the effect of different microbial communities as well as the effect of food types on the rotifer growth performance. 2. Materials and methods 2.1. Rotifer strains Clones L1 and L3 of Brachionus plicatilis were obtained from the University of Valencia (Instituto Cavanilles de Biodiversidad y Biología Evolutiva) in Spain. Clone 10 was obtained from CIAD (Centro de Investigación en Alimentación y Desarrollo, Mazatlan Unit for Aquaculture) in Mexico. All the clones were identified as B. plicatilis sensu strictu using the methodology published by Gomez et al. (2002) (unpublished results). The rotifer stocks were maintained at controlled culture conditions: 28 C, light intensity 2000 lx, 25 g l 1 seawater, and fed with Chlorella at cells ml Preparation of food An axenic inoculum of Chlorella sp., strain CCAP 211/76, was obtained from the Culture Collection of Algae and Protozoa (Dunstaffnage Marine Laboratory, Dunberg) in Scotland. Axenic Chlorella was grown in closed 500 ml bottles provided with 0.22 μm-filtered aeration. The culture was maintained at 19 C, light intensity lx, using a standard Walne medium supplemented with vitamins and 0.22 μm-filtered and autoclaved regular seawater (FASW), which was diluted with tap water to have a salinity of 25 g l 1. The wild-type strain of baker's yeast (Saccharomyces cerevisiae) and its isogenic mutant strain mnn9 were

3 N.T.N. Tinh et al. / Aquaculture 253 (2006) obtained from EUROSCARF (Institute of Microbiology, University of Frankfurt) in Germany. The mnn9 strain is deficient in mannose chain elongation of mannoproteins resulting in a low concentration of mannoproteins in the outer layer of the cell wall (Magnelli et al., 2002; Marques et al., 2004). Axenic yeast cultures were grown in sterile Erlenmeyer's on a shaker at 150 rpm and 30 C. The culture medium used was YEPD (Yeast Extract Peptone Dextrose) medium, containing yeast extract (Sigma, 1% w/v), peptone bacteriological grade (Sigma, 1% w/v) and D-glucose (Sigma, 2% w/v). This medium was prepared in 25 g l 1 FASW. Cultures of axenic Chlorella and two yeast strains were harvested by centrifugation (1600 g for 5 min) at exponential growth phase (after seven days for Chlorella, 24 h for yeast strains). Cells were resuspended twice in sterile falcon tubes (TRP, γ- irradiated) with 10 ml of 25 g l 1 FASW. All the manipulations were performed in a laminar flow hood to maintain axenicity. Chlorella and yeast densities were determined by measuring twice the cell concentration, using a Bürker haemocytometer Sources of microbial communities Two types of MCs were used in the experiments, which were isolated either from normal-performing or from crashed rotifer cultures. The latter were isolated from 2-L cultures which were maintained under controlled conditions (25.5 C, aeration, light regime 12 h day 1, Chlorella was used as food). The rotifer densities in these cultures dropped drastically from 30 to 40 rotifers ml 1 to 0 rotifers ml 1 within three days. In addition, the crash conditions were preceded by very low egg ratio (<0.1 eggs rotifer 1 ) and reduced food consumption. Resting egg production was observed in L3 culture following the crash Experimental protocol Disinfection of rotifer amictic eggs Glutaraldehyde (Fluka) and benzyldimethyldodecylammonium bromide (BAB) (Sigma) were used as disinfectants for the rotifer culture (clone 10). Different concentrations and exposure times were tested at a temperature of 28 C, each treatment was done in triplicate (Table 1). The rotifer culture was first filtered through a 250 μm mesh to remove all the algal flocci which may affect the disinfectant's activity, then washed three times with an equal volume of 25 g l 1 FASW using a 60 μm mesh, before being transferred to sterile falcon tubes which contained different concentrations of disinfectants in 40 ml of 25 g l 1 FASW. The initial rotifer number was animals per tube. The swimming behaviour of the rotifers was checked every 30 min under a binocular microscope. As soon as all the rotifers were dead (i.e. no more cilia movement) in any treatment, the content of each falcon tube was filtered through an autoclave-sterilized 30 μm mesh mounted to a Büchner filter (Nalgene ). The rotifer adults were not separated from the amictic eggs. The mesh with all the retained dead rotifers and amictic eggs was subsequently placed in another falcon tube containing 40 ml of fresh 25 g l 1 FASW. The falcon tube was placed on a rotor for hatching. The culture was checked after 30 min, to verify that all the rotifers were killed. After 3 h of incubation, the presence of newly hatched neonates was observed under a microscope. After disinfection the amictic eggs were tested for axenicity. Ten milliliters of the culture from each replicate was filtered through an autoclave-sterilized 30 μm mesh mounted to a Büchner filter. The mesh was then placed aseptically in a sterile plastic bag containing 10 ml of autoclaved Nine Salt Solution (NSS), which was composed of 17.6 g l 1 NaCl, 1.47 g l 1 Na 2 SO 4, 0.08 g l 1 NaHCO 3, 0.25 g l 1 KCl, 0.04 g l 1 KBr, Table 1 Concentrations and exposure times of two disinfectants (glutaraldehyde and benzyldimethyldodecyl-ammonium bromide) used for rotifer amictic eggs disinfection Disinfectant Experiment 1 Experiment 2 Experiment 3 Concentration (ppm) Exposure time Concentration (ppm) Exposure time Concentration (ppm) Exposure time BAB 5 2 h 50 1, 2 h 10 2, 3, 4, 5 h , 30 min 25 1, 2 h Glutaraldehyde 10 1 h 10 6, 9, 12, 24 h 60 2 h 25 1 h 25 6, 9, 12, 24 h 80 2 h 50 1, 2, 3 h h h

4 424 N.T.N. Tinh et al. / Aquaculture 253 (2006) g l 1 MgCl 2, 0.41 g l 1 CaCl 2, g l 1 SrCl 2 and g l 1 H 3 BO 3. The content of the plastic bag was homogenized for 6 min by means of a stomacher blender (Seward). Three sub-samples of 50 μl from the suspension were spread plated on Marine Agar (MA, Difco). After 24 h of incubation at 28 C, the presence of bacterial colonies was checked. Axenicity was also verified by staining vital bacterial cells. Each sample was treated with MTT ( 3-(4,5-dimethylthazol-2-yl)- 2,5-diphenyl tetrazolium bromide) (Sigma, 0.5% w/v) in a sterile eppendorf (1 part of MTT to 9 parts of sample) and incubated at 30 C for 30 min. Under a light microscope (1000 magnification), the samples were checked for the presence of blue-stained viable cells (Sladowski et al., 1993) Experiments on the effect of microbial communities and effect of food types Rotifers (clone 10), hatched from axenic amictic eggs, were used in all experiments. Disinfection of amictic eggs was done using 50 ppm of glutaraldehyde with 1 h exposure time at 28 C. After 3 h of incubation in fresh FASW, the cultures were allowed to stand for 5 min during which most of the dead rotifers were settled down. The density of newly hatched animals was counted afterwards, and the neonates were distributed to sterile falcon tubes containing 20 ml of 25 g l 1 FASW, to have a density of 8 rotifers ml 1 at the start of each experiment. Axenic cultures of Chlorella and two yeast strains were used as food for rotifers in all the experiments. The optimum feeding level for each food type was determined in previous experiments (unpubl. data). The optimum feeding levels for Chlorella, wild-type yeast and mnn9 was 150,000 cells rotifer 1, 240,000 cells rotifer 1 and 106,000 cells rotifer 1, respectively. The different feeding levels in cell number were due to the difference in the cell ash-free dry weight (Marques et al., 2004). Feeding was done once a day under a laminar flow hood, using MultiGuard Barrier pipette tips (Sorenson BioScience). Two series of experiments were conducted (Table 2). Experiments presented in the same row were carried out at the same time. Live MCs (freshly isolated or preserved at 80 C and regrown on MA) were added in the first series. For isolation of MCs, the culture water collected from respective rotifer culture was filtered through 250 and 60 μm meshes to remove big food particles and all the rotifers, subsequently was centrifuged at 1600 g for 5 min to remove the algal cells, thus only the MCs were retained in the supernatant. The MCs from normal-performing cultures were collected the Table 2 Outline of the experiments on the effect of microbial communities Bacterial treatment Control Treatment 1 Treatment 2 (Exp. 1.1) (Exp. 1.2) (Exp. 1.3) (Exp. 1.4) (Exp. 2.1) (Exp. 2.2) (Exp. 2.3) MC from clone (Exp. 1.5) MC from clone (Exp. 1.6) MC from clone (Exp. 1.7) MC from clone (Exp. 1.8) Autoclaved MC from clone 10 normal culture (Exp. 2.4) Autoclaved MC from clone 10 normal culture (Exp. 2.5) Autoclaved MC from clone 10 normal culture (Exp. 2.6) MC from L1 crashed culture (Exp. 1.9) MCR from L1 (Exp. 1.10) MCR from L3 (Exp. 1.11) MCR from clone 10 (Exp. 1.12) Autoclaved MCR from L1 crashed culture (Exp. 2.7) Autoclaved MCR from L3 crashed culture (Exp. 2.8) Autoclaved MCR from clone 10 (Exp. 2.9) Experiments in the same row were performed at the same time. MC: live and freshly collected microbial community. MCR: live microbial community, which was preserved at 80 C and regrown on MA. same day when experiment started. There was only one experiment where the MC from was used fresh (exp. 1.9). Part of the supernatants collected from L1, L3 and clone 10 s was centrifuged at 4450 g for 15 min, the yielded bacterial pellets were re-suspended in autoclaved Nine Salt Solution (NSS). These MCs were preserved for further experiments in 1 ml eppendorfs containing 20% glycerol and 80% bacterial suspension and kept at 80 C. Before starting each experiment, the eppendorfs were defrosted, subsequently 50 μl of the corresponding MC suspension was spread plated on Marine Agar (MA). After 24 h of incubation at 28 C, the bacteria were harvested by swabbing the MA plate and resuspending in autoclaved NSS. The optical density measurement (OD 550 ) was taken for each bacterial suspension, appropriate volume to be added to each treatment was calculated in order to have a density of 10 6 CFU ml 1 at the start of each experiment. It is acknowledged that the microbial composition of the culture obtained in this way cannot be the same as the one originally available in the rotifer cultures. The second experimental series was run in order to investigate whether the effects observed in the first series of experiments are nutritional or probiotic/

5 N.T.N. Tinh et al. / Aquaculture 253 (2006) pathogenic in nature. The MCs used in these experiments, either freshly isolated or preserved and regrown on MA, were killed by autoclaving at 121 C for 20 min before addition to the rotifer cultures. All the treatments took place in 50 ml sterile falcon tubes (TRP, γ- irradiated) containing 20 ml of 25 g l 1 FASW, with four replicates per treatment. The falcon tubes were put on a rotor (4 rpm) which was placed inside a temperaturecontrolled room (28 C, light intensity 2000 lx). Axenicity was tested on the starting day (day 0) and on the last day (day 5) for the experiments where no bacteria were added. Fifty microliters of the culture water was spread plated on MA plate, bacterial growth was checked after incubation at 28 C for 24 h. Axenicity was also checked by bacterial staining using MTT. All the results were discarded if contamination was found in any replicate Data analysis Two sub-samples of 500 μl were withdrawn daily from each replicate for estimation of rotifer densities. Population growth rate (GR) was calculated as: GR ¼ðlnN t lnn o Þ=t where N o is the initial rotifer density, N t is the rotifer density on day t of culture, t is the duration in days. Parametric assumptions were evaluated using Levene's test for homogeneity of variances and Shapiro Wilk's test for normality. As data were normaldistributed and homoscedastic, the growth rates on day 5 were compared by food type between synchronous experiments using one-way ANOVA, followed by Tukey test. All the tests were performed using the computer program SPSS release Results 3.1. Disinfection of rotifer amictic eggs Different concentrations and different exposure times were tested for two kinds of disinfectants (glutaraldehyde and BAB). The effectiveness of each treatment was evaluated based on the observation of the following parameters: mobility of adult rotifers after exposure time, hatchability after incubation in new seawater, bacterial count of amictic eggs after disinfection. The results of disinfection treatments with BAB are presented in Table 3. The rotifers were found dead in all the treatments after a certain exposure time. However, hatchability of amictic eggs was zero in the treatments with high concentrations of disinfectant or with prolonged exposure times. In addition, axenicity could not be obtained in any treatment. Hence, BAB is not suitable for disinfection of rotifer amictic eggs. Table 4 shows the results of glutaraldehyde treatment with different combinations of concentration and exposure time. The concentration of 10 ppm at 1, 6 and 9 h exposure time and the concentration of 25 ppm at 1 h exposure time were not sufficient to kill all the rotifer adults. In contrast to BAB, all of the treatments did not affect the hatchability of amictic eggs. However, only one treatment (100 ppm at 2 h exposure time) showed a complete elimination of the microbiota associated with amictic eggs. The four treatments 50, 60, 80 and 100 ppm were later repeated four times for different rotifer cultures (results not shown). In all the cases, the hatched rotifers showed normal swimming behaviour and ability to produce eggs afterwards. However, the disinfecting efficacy varied between rotifer batches, probably depending on the residual organic matter load as well as the bacterial Table 3 Effectiveness of BAB as a disinfectant for rotifer amictic eggs Concentration Exposure time Mobility after exposure to disinfectant Hatchability of treated amictic eggs Axenicity of treated amictic eggs 5 ppm 2 h ± ppm 2 h ±2.39 3h 3.38±2.16 4h 3.02±1.91 5h 2.76± ppm 1 h ±1.65 2h 2.04± ppm 1 h 2.34±1.45 2h 2.33± ppm 15 min 1.94± min 1.90±1.11 +: positive result, : negative result. MA plate count (log CFU ml 1 ) (mean±sd)

6 426 N.T.N. Tinh et al. / Aquaculture 253 (2006) Table 4 Effectiveness of glutaraldehyde as a disinfectant for rotifer amictic eggs Concentration Exposure time Mobility after exposure to disinfectant Hatchability of treated amictic eggs Axenicity of treated amictic eggs 10 ppm 1 h + 6h + 9h + 12 h ± h ± ppm 1 h + 6h ±3.25 9h ± h ± h ± ppm 1 h ±3.73 2h ±3.12 3h ± ppm 2 h ± ppm 2 h ± ppm 1 h ±2.86 2h + + +: positive result, : negative result. MA plate count (log CFU ml 1 ) (mean±sd) load in the culture water and in the rotifer's gut, which cannot be eliminated during the washing process. For one rotifer batch 50 ppm glutaraldehyde during 1 h was sufficient to eliminate the egg-associated microbiota, for another batch only 100 ppm during 2 h was effective Evaluation of nutritional effect of different food types in axenic rotifer cultures The effect of three different food types (Chlorella, wild-type yeast (BY) and mnn9 yeast mutant) on rotifer performance was evaluated based on the results of the second experimental series, in the absence of bacteria or when autoclave-killed MCs were added (Table 5). Rotifers fed on algae had the highest growth rate compared to those fed on yeast strains (p<0.001). These results were highly reproducible. On the other hand, no significant difference was observed between the treatments with yeast strains (p > 0.05) Evaluation of the probiotic/nutritional effect of microbial communities In the first series of experiments, the effect of different live MCs isolated either from normal-performing or from crashed rotifer cultures was evaluated. Table 6 represents the population growth rates over five days for each food type. Comparison is made between synchronous experiments with and without addition of MCs (Table 7). No significant stimulation of the growth rate by both types of MCs was found (p > 0.05) in the treatments where Chlorella was used as food source, except in experiments 1.1, 1.5 and 1.9. In contrast, differences were seen when the two yeast strains were used as food. Growth rates were significantly improved (p < 0.001) when MCs from normal-performing cultures were Table 5 Growth rate over 5 days (mean±sd, n=4) of Brachionus plicatilis sensu strictu (clone 10) hatched from disinfected amictic eggs and fed three types of food: effect of food type (see Table 2 for experimental outline) Food type Experiment Autoclaved MC Autoclaved MCR Exp. 2.1 Exp. 2.4 Exp. 2.7 Chlorella 0.60±0.02 b 0.59±0.02 b 0.57±0.04 b Wild-type yeast 0.25±0.008 a 0.26±0.01 a 0.26±0.03 a mnn9 yeast 0.28±0.03 a 0.30±0.04 a 0.29±0.02 a mutant Exp. 2.2 Exp. 2.5 Exp. 2.8 Chlorella 0.69±0.03 b 0.68±0.05 b 0.69±0.03 b Wild-type yeast 0.22±0.02 a 0.24±0.04 a 0.24±0.02 a mnn9 yeast 0.26±0.02 a 0.25±0.04 a 0.28±0.04 a mutant Exp. 2.3 Exp. 2.6 Exp. 2.9 Chlorella 0.66±0.04 b 0.66±0.02 b 0.66±0.05 b Wild-type yeast 0.35±0.07 a 0.36±0.02 a 0.34±0.03 a mnn9 yeast mutant 0.33±0.09 a 0.33±0.10 a 0.30±0.08 a Treatments with different superscripts in each experiment are significantly different from each other (Tukey test, p<0.05).

7 N.T.N. Tinh et al. / Aquaculture 253 (2006) Table 6 Growth rate over 5 days (mean±sd, n=4)of Brachionus plicatilis sensu strictu (clone 10) hatched from disinfected amictic eggs and fed three types of food: effect of the addition of live microbial communities Exp. Bacterial treatment Food type Growth rate (mean±sd) 1.1 Chlorella 0.64±0.07 Wild-type yeast 0.18±0.06 mnn9 yeast mutant 0.36± Chlorella 0.55±0.05 Wild-type yeast 0.36±0.07 mnn9 yeast mutant 0.32± Chlorella 0.58±0.02 Wild-type yeast 0.34±0.03 mnn9 yeast mutant 0.31± Chlorella 0.60±0.03 Wild-type yeast 0.35±0.07 mnn9 yeast mutant 0.33± MC from clone 1.6 MC from clone 1.7 MC from clone 1.8 MC from clone 1.9 MC from L MCR from L MCR from L MCR from clone 10 crashed culture Chlorella 0.83±0.04 Wild-type yeast 0.63±0.01 mnn9 yeast mutant 0.64±0.07 Chlorella 0.62±0.02 Wild-type yeast 0.54±0.01 mnn9 yeast mutant 0.58±0.01 Chlorella 0.59±0.05 Wild-type yeast 0.55±0.01 mnn9 yeast mutant 0.58±0.02 Chlorella 0.67±0.05 Wild-type yeast 0.55±0.02 mnn9 yeast mutant 0.58±0.01 Chlorella 0.79±0.02 Wild-type yeast 0.56±0.03 mnn9 yeast mutant 0.71±0.07 Chlorella 0.60±0.05 Wild-type yeast 0.45±0.08 mnn9 yeast mutant 0.40±0.03 Chlorella 0.62±0.02 Wild-type yeast 0.44±0.04 mnn9 yeast mutant 0.40±0.04 Chlorella 0.59±0.02 Wild-type yeast 0.44±0.05 mnn9 yeast mutant 0.42±0.04 MC: live and freshly collected microbial community. MCR: live microbial community, which was preserved at 80 C and regrown on MA. added. The behaviour of MCs from s was more variable, since they were collected from the cultures of different rotifer strains, and were utilized under two different forms: freshly isolated or preserved at 80 C and regrown afterwards on MA. When comparing by origin of preserved MCs, only the addition of MCR from L3 showed a significant improvement (p<0.05) in growth rate (exp. 1.3 and 1.11). MC from L1 was used in two forms. The fresh-isolated MC could stimulate the growth performance significantly (p < 0.001, exp. 1.1 and 1.9), while no stimulation was found (p > 0.05) when that MC was preserved and regrown before use (exp. 1.2 and 1.10). The experimental setup of the second series of experiments was similar to that in the first series. The only difference was that autoclave-killed instead of live MCs were added, allowing to investigate whether the change in growth rate observed in the first experimental series was solely the consequence of the extra nutrients added through the MCs. As shown in Tables 8 and 9,no significant difference in growth rate was observed (p > 0.05) in the presence or absence of heat-killed MCs. This observation was reproducible for all three types of food and in all experiments. 4. Discussion Two disinfectants with strong biocidal property were used in this study. Glutaraldehyde is a disinfectant which displays a broad spectrum of activity and a rapid killing rate against the majority of microorganisms. It is capable of destroying all forms of microbial life including bacterial and fungal spores, tubercle bacilli and viruses. The mode of action of glutaraldehyde is based on the formation of intercellular bonds with the outer layers of Table 7 Comparison in growth rate over 5 days of Brachionus plicatilis sensu strictu (clone 10) in the presence of live microbial communities (Tukey test) Pair of experiments Food type Significance level (p-value) Exp. 1.1 and exp. 1.5 Chlorella Wild-type yeast mnn9 yeast mutant Exp. 1.2 and exp. 1.6 Chlorella Wild-type yeast mnn9 yeast mutant Exp. 1.3 and exp. 1.7 Chlorella Wild-type yeast mnn9 yeast mutant Exp. 1.4 and exp. 1.8 Chlorella Wild-type yeast mnn9 yeast mutant Exp. 1.1 and exp. 1.9 Chlorella Wild-type yeast mnn9 yeast mutant Exp. 1.2 and exp Chlorella Wild-type yeast mnn9 yeast mutant Exp. 1.3 and exp Chlorella Wild-type yeast mnn9 yeast mutant Exp. 1.4 and exp Chlorella Wild-type yeast mnn9 yeast mutant See Table 2 for more information on the treatments.

8 428 N.T.N. Tinh et al. / Aquaculture 253 (2006) Table 8 Growth rate over 5 days (mean±sd, n=4) of Brachionus plicatilis sensu strictu (clone 10) hatched from disinfected amictic eggs and fed three types of food: effect of the addition of autoclave-killed microbial communities Exp. Bacterial treatment Food type Growth rate (mean±sd) 2.1 Chlorella 0.60±0.02 Wild-type yeast 0.25±0.008 mnn9 yeast mutant 0.35± Chlorella 0.69±0.03 Wild-type yeast 0.22±0.02 mnn9 yeast mutant 0.26± Chlorella 0.66±0.04 Wild-type yeast 0.35±0.07 mnn9 yeast mutant 0.33± Autoclaved MC from clone 10 normal culture 2.5 Autoclaved MC from clone 10 normal culture 2.6 Autoclaved MC from clone 10 normal culture 2.7 Autoclaved MCR from L1 crashed culture 2.8 Autoclaved MCR from L3 crashed culture 2.9 Autoclaved MCR from clone 10 Chlorella 0.59±0.02 Wild-type yeast 0.26±0.01 mnn9 yeast mutant 0.31±0.04 Chlorella 0.68±0.05 Wild-type yeast 0.24±0.04 mnn9 yeast mutant 0.25±0.04 Chlorella 0.66±0.02 Wild-type yeast 0.36±0.02 mnn9 yeast mutant 0.33±0.10 Chlorella 0.57±0.04 Wild-type yeast 0.26±0.03 mnn9 yeast mutant 0.36±0.02 Chlorella 0.69±0.03 Wild-type yeast 0.24±0.02 mnn9 yeast mutant 0.28±0.04 Chlorella 0.66±0.05 Wild-type yeast 0.34±0.03 mnn9 yeast mutant 0.30±0.08 bacterial cells or bacterial spores and in this way on the interference with the functionality of the cell wall (Scott and Gorman, 1991). Glutaraldehyde is used widely in disinfection of marine fish eggs before incubation. Salvesen and Vadstein (1995) have found that glutaraldehyde was the most promising candidate of four disinfectants (Buffodine, glutaraldehyde, chloramine-t and sodium hypochlorite) tested on plaice eggs. Benzyldimethyldodecyl-ammonium bromide (BAB) belongs to the group of quaternary ammonium compounds (QACs). By binding to the phospholipid and protein layers, QACs impair the permeability of the cell membrane (Maris, 1995). While a good disinfectant should sterilize amictic eggs completely, it should not affect the viability of amictic eggs neither the mobility of hatched rotifers. In our study, none of the BAB treatments met both axenicity and zootechnical criteria. The amictic eggs ceased to hatch at high concentrations of BAB, probably due to toxicity. In addition, such high concentrations were still not sufficient to kill all the bacteria in the culture. For glutaraldehyde, appropriate conditions could be identified, allowing obtaining axenic amictic eggs that were still able to hatch after 3 h incubation and develop further normally. Depending on the batch, rotifers should be exposed to ppm glutaraldehyde for 1 2 h at 28 C. Bacteria-free organisms are being used as a tool for studying microbiota-attributed functions. It allows studying the probiotic potential or pathogenicity of selected strains in a gnotobiotic environment (Martinez- Diaz et al., 2003). Germ-free organisms are also useful for the studies of nutritional requirements (Scott, 1983). Attempts to make axenic rotifer cultures from resting eggs were conducted by several authors using different kinds of disinfectants. Hagiwara et al. (1994) suggested soaking the resting eggs in 0.5 and 0.25 ppm sodium hypochlorite solution for 60 and 30 min, respectively. Douillet (1998) found that the disinfection of resting eggs was most effective at much higher concentration of sodium hypochlorite (0.5%) and shorter exposure time (3 min). Rombaut et al. (1999) tested the effect of merthiolate and glutaraldehyde. The latter was found to be effective at low concentration (0.05 ppm) and prolonged incubation time (6 h). Some authors went for another approach, using amictic eggs as material for starting an axenic culture. Most of them MC: freshly collected microbial community. MCR: microbial community which was preserved at 80 C and regrown on MA. Table 9 Comparison in growth rate over 5 days of Brachionus plicatilis sensu strictu (clone 10) in the presence of autoclave-killed microbial communities (Tukey test) Pair of experiments Food type Significance level (p-value) Exp. 2.1 and exp. 2.4 Chlorella Wild-type yeast mnn9 yeast mutant Exp. 2.2 and exp. 2.5 Chlorella Wild-type yeast mnn9 yeast mutant Exp. 2.3 and exp. 2.6 Chlorella Wild-type yeast mnn9 yeast mutant Exp. 2.1 and exp. 2.7 Chlorella Wild-type yeast mnn9 yeast mutant Exp. 2.2 and exp. 2.8 Chlorella Wild-type yeast mnn9 yeast mutant Exp. 2.3 and exp. 2.9 Chlorella Wild-type yeast mnn9 yeast mutant See Table 2 for more information on the treatments.

9 N.T.N. Tinh et al. / Aquaculture 253 (2006) used antibiotic mixtures as disinfecting medium (Plasota et al., 1980; Hirayama and Funamoto, 1983; Dhert, 1996), which is not desirable in the light of antibiotic resistance among bacterial populations in rotifer cultures. Martinez-Diaz et al. (2003) have tried out two disinfectants (PVP-Iodine, Hydrogen peroxide) and two antibiotic mixtures, starting either with adult females or amictic eggs. However, only the antibiotic treatment applied to amictic eggs was effective in eliminating the associated bacteria, while sustaining the viability of amictic eggs. Hence, the method applied in this study has several advantages. Firstly, axenic rotifer cultures can be obtained from amictic eggs of one single clonal strain, allowing for instance to obtain an axenic culture from a strain that does not produce resting eggs under lab conditions. It also avoids using resting eggs which are normally costly and may contain (if collected in the field) an uncharacterized rotifer mixture. Secondly, a non-antibiotic disinfectant (glutaraldehyde) is used, which eliminates the risk of antibiotic resistance build up. Several studies are investigating the effect of different diets on the growth as well as on the dietary value of rotifers to the fish larvae. Those studies indicate that green microalgae such as Chlorella sp. and Nannochloropsis sp. are the most suitable diets for rotifers. Moreover rotifers fed on those algae can satisfy the nutritional demand of fish larval predators. Yeast can only be used at low concentrations to supplement algal requirement due to its deficiency in HUFA content (Caric et al., 1993; Tamaru et al., 1993; Sarma et al., 2001), or rotifers can be grown on yeast and then enriched with DHA or EPA (Dhert, 1996). In the present study, the mnn9 mutant which is defective in mannoprotein synthesis was chosen for evaluation of its nutritional effect on rotifer growth in comparison to wild-type yeast. In a similar study done on Artemia nauplii, Marques et al. (2004) found a strong reverse correlation between yeast cell wall mannoprotein content and Artemia performance. According to Telford (1970) and Coutteau et al. (1990), β-glucanase activity is detected in the digestive tract of Artemia, but no mannase activity. This finding coincides with the poor performance of Artemia fed on wild-type yeast, suggesting that the external mannoprotein layer of the yeast cell wall presents the major barrier to digestion. However, it was not the case for rotifers as shown in the present study. Rotifer performance was statistically equal when a wild-type yeast strain or a mnn9 mutant was given as food. Biochemical studies on hydrolytic enzymes from the rotifer B. plicatilis revealed the presence of β-1,3-glucanase and chitinase (Kleinow, 1993; Hara et al., 1997). There is no evidence of mannase activity in rotifers. However, as opposed to Artemia in which a mandibular grinding of food particles is unknown (Coutteau et al., 1990), the mastax organ present in rotifers may help to break the yeast cell wall, facilitating the access of digestive enzymes to the internal layers. Axenically grown rotifers were used in this study as a test model to reveal the role of endogenous microbiota which were isolated from rotifer cultures. When yeast strains were given as food, rotifer growth performance was dependent on the MCs origin. Yet, the nature of MC had a strong influence. The effect on growth rate was significant and was reproducible when MCs were fresh-isolated, irrespective of their origin (exp. 1.1 vs. 1.5, 1.2 vs. 1.6, 1.3 vs. 1.7, 1.4 vs. 1.8, 1.1 vs. 1.9; in the latter experiment, MC was collected from a crashed rotifer culture) (Table 7). In the other experiments with live MCs, all MCRs (regrown microbial communities) were not able to stimulate the growth of Brachionus (except in one case, when the MCR originated from L3 crashed culture), suggesting that regrown MCs have a less favourable effect or no effect on Brachionus even when low quality food (such as yeast cells) is used. In one particular case it was possible to directly compare the effect of a MC (exp. 1.1 vs. 1.9) and the effect of the corresponding regrown MC (exp. 1.2 vs. 1.10) (although these experiments were performed after each other, which is for logistical reasons unavoidable; the MC from exp. 1.9 would always need to be preserved in some way or another in attendance of the MCR of exp. 1.10). In this single case the MCR effect was not significant, while the effect of the corresponding MC was significant. Overall, it can be concluded that the presence of MCs is essential for good rotifer growth when suboptimal food is used. Yet, the source of MC attenuates the extent of the beneficial effect (direct use of MCs vs. sub-cultured MCs). The results of the second experimental series seem to exclude the role of microbiota as an initial and additional source of nutrients for rotifers, although it has to be acknowledged that through autoclavation nutrients might be deprived from the cytoplasmic content of the microbial cells. It is possible that bacterial enzymes may help to improve the yeast digestibility, thus improving the rotifer growth, or that live bacteria have a direct effect on the rotifer metabolism. Bacterial enzymes, such as N-acetyl-βglucosaminidase (chitinolytic activity) and β-glucosidase (cellulolytic activity), were detected in relatively

10 430 N.T.N. Tinh et al. / Aquaculture 253 (2006) high concentrations in the putative probiotic isolates (Rombaut, 2001), indicating a possible role in the digestion of yeast, which synthesizes chitin as a cell wall constituent (Smits et al., 1999). An additional reason for the beneficial effect of live bacteria might lay in their eventual capacity to grow on the yeast or Chlorella degradation product, generating new microbial biomass that can serve as food for rotifers. This nutrient recycling might also have a beneficial effect on the water quality. The third hypothesis is that bacterial cell wall material might induce a change in digestion competence in the rotifers, resulting in better growth rate. Associated microbiota have been shown to affect a wide range of biological processes in the host organism, as revealed in the assay with gnotobiotic zebrafish (Rawls et al., 2004). Although the present experimental setup does not allow to clearly distinguish between the nutritional and probiotic effect of bacteria on rotifers, it does strongly suggest that killed bacterial cells (some of them might have lost the cytoplasmic content) alone do not promote the growth of rotifers. In the present study, MCs isolated from crashed rotifer cultures did not show any negative effect in the rotifer growth test, suggesting that the MCs associated with those particular crashed rotifer cultures were not responsible for the crash. This is in contradiction with the postulation of many authors, that the microbiota associated with rotifer production systems play a major role in the instability and variability of the rotifer cultures (Hirayama, 1987; Gatesoupe et al., 1989; Gatesoupe, 1991; Skjermo and Vadstein, 1993; Harzevilli et al., 1997). Some bacterial strains such as Plavobacterium, Aeromonas spp. and Vibrio anguillarum were isolated from collapsing rotifer cultures and showed toxicity for the rotifer population (Harzevilli et al., 1997; Balompapueng et al., 1997). Hino (1993) formulated a more subtle hypothesis, suggesting that changes in the composition of the MC, and not the standing microbiota itself, are the cause of the collapse of rotifer cultures. Rapid succession in the MC was observed during batch culture of rotifers (Maeda and Hino, 1991). Shifts in the fingerprinting of the MC in the rotifer culture water, as revealed by means of denaturing gradient gel electrophoresis (DGGE), were found to be often associated with technical problems resulting in a reduced water quality in the rotifer production systems (Rombaut et al., 2001). The present results seem to support the view that the standing MCs, isolated either from a normal-performing rotifer culture or from a crashing rotifer culture, have no negative influence on rotifer performance. The experimental setup however does not allow to either support or negate the hypothesis that a rapid changing MC can be a facilitator of rotifer culture crashes. The results however do suggest that regrowing a MC in a rich medium, such as Marine Agar (most probably resulting in a shift of its composition), is reducing the beneficial effect of that MC, both when Chlorella or yeast is used as the main food source. Future research, with other MCs isolated from especially freshly crashed rotifer cultures, might further substantiate this hypothesis. 5. Conclusions The glutaraldehyde treatment of amictic eggs represents a novel approach for obtaining axenic rotifers, as a disinfectant of non-antibiotic nature is used. This technique can be used in strains that do not produce resting eggs. Axenic rotifers obtained this way have shown to be an excellent tool for studying probiotic and nutritional function of microorganisms, as well as nutritional value of different food sources. It was proven that the presence of an endogenous microbiota is essential for the growth of rotifers, especially when low quality food (yeast) is offered. Such an effect could be obtained with a microbial community originating from a normal or a, but not with a MCR, namely MC regrown in the lab on rich medium. The results further suggest that the MC isolated from one particular crashed rotifer culture was probably not responsible for the crash, as no negative effects were observed by adding this MC to axenic rotifers. This of course does not exclude the possibility that crashes in other rotifer cultures are due to the presence of certain microorganisms. Acknowledgements This study was supported by a doctoral grant for candidates from developing countries (Bijzonder Onderzoeksfonds, grant number B/ DS502) given to the first author by Ghent University, Belgium, and in part by the EU ROTIGEN project QLRT ( The authors thank Serra M. and Roque A. (Spain) for providing rotifer strains, the Department of Genetics, Development and Molecular Biology, Aristotle University of Thessaloniki, Greece (a partner of Rotigen project), for identifying rotifer clones. Special thanks go to Jean Dhont and Mathieu Wille for critical reading the manuscript.

11 N.T.N. Tinh et al. / Aquaculture 253 (2006) References Balompapueng, M.D., Hagiwara, A., Nishi, A., Imaizumi, K., Hirayama, K., Resting egg formation of the rotifer Brachionus plicatilis using a semi-continuous culture method. Fish. Sci. 63, Caric, M., Sankonjire, J., Skaramuca, B., Dietary effects of different feeds on the biochemical composition of the rotifer (Brachionus plicatilis Müller). Aquaculture 110, Coutteau, P., Lavens, P., Sorgeloos, P., Baker's yeast as a potential substitute for live algae in aquaculture diets: Artemia as a case study. J. World Aquacult. Soc. 21, 1 8. Davis, D.A., Arnold, C.R., Tolerance of the rotifer Brachionus plicatilis to ozone and total oxidative residuals. Ozone: Sci. Eng. 19, Dhert, P., Rotifers. In: Lavens, P., Sorgeloos, P. (Eds.), Manual on the Production and Use of Live Food for Aquaculture. FAO Technical Paper, pp Dougherty, E.C., Solberg, B., Harris, L.G., Synxenic and attempted axenic cultivation of rotifers. Science 132, Douillet, P.A., Disinfection of rotifer cysts leading to bacteriafree populations. J. Exp. Mar. Biol. Ecol. 224, Gatesoupe, F.J., Nutritional and antibacterial treatments of live food organisms: the influence on survival, growth rate and weaning success of turbot (Scophthalmus maximus). Ann. Zootech. 31, Gatesoupe, F.J., The effect of three strains of lactic bacteria on the production rate of rotifers, Brachionus plicatilis, and their dietary value for larval turbot, Scophthalmus maximus. Aquaculture 96, Gatesoupe, F.J., Arakawa, T., Watanabe, T., The effect of bacterial additives on the production rate and dietary value of rotifers as food for Japanese flounder, Paralichthys olivaceus. Aquaculture 83, Gomez, A., Serra, M., Carvalho, G.R., Lunt, D.H., Speciation in ancient cryptic species complexes: evidence from the molecular phylogeny of Brachionus plicatilis (Rotifera). Evolution 56 (7), Hagiwara, A., Hamada, K., Hori, S., Hirayama, K., Increased sexual reproduction in Brachionus plicatilis (Rotifera) with the addition of bacteria and rotifer extracts. J. Exp. Mar. Biol. Ecol. 181, 1 8. Hara, K., Pangkey, H., Osatomi, K., Yatsuda, K., Hagiwara, A., Tachibana, K., Ishihara, T., Some properties of β-1,3-glucan hydrolyzing enzymes from the rotifer Brachionus plicatilis. Hydrobiologia 358, Harzevilli, A.R.S., Van Duffel, H., Defoort, T., Dhert, P., Sorgeloos, P., Swings, J., The influence of a selected bacterial strain Vibrio anguillarum TR27 on the growth rate of rotifers in different culture conditions. Aquacult. Int. 5, Hayashi, K., Kimura, T., Sugahara, I., Microbiological studies on the artificial seedling production of ayu fish, Plecoglossus altevelis (Temminck et Schlegel): III. Bacterial contamination of Brachionus plicatilis and Moina macrocopa. Bulletin of the Faculty of Fisheries. Mie University, pp Hino, A., Present culture systems of the rotifer (Brachionus plicatilis) and the function of micro-organisms. In: Lee, C.S., Su, M.S., Liao, I.C. (Eds.), Finfish hatchery in Asia: Proceedings of finfish hatchery in Asia'91. TLM Conference Proceedings 3, Tungkang Marine Laboratory. Taiwan Fisheries Research Institute, Tungkang, Pingung, Taiwan, p. 14. Hirayama, K., A consideration of why mass culture of the rotifer Brachionus plicatilis with baker's yeast is unstable. Hydrobiologia 147, Hirayama, K., Funamoto, H., Supplementary effect of several nutritive deficiency of baker's yeast for population growth of the rotifer Brachionus plicatilis. Nippon Suisan Gakkaishi 49, Kleinow, W., Biochemical studies on Brachionus plicatilis hydrolytic enzymes, integument proteins and composition of trophi. Hydrobiologia 255, Lee, C.-S., Ostrowski, T., Current status of marine finfish larviculture in the United States. Aquaculture 200, Liao, C., Su, H.M., Chang, E.Y., Techniques in finfish larviculture in Taiwan. Aquaculture 200, Lubzens, E., Raising rotifers for use in aquaculture. Hydrobiologia 147, Lubzens, E., Tandler, A., Minkoff, G., Rotifers as food in aquaculture. Hydrobiologia 186/187, Lubzens, E., Minkoff, G., Zmora, O., Barr, Y., Mariculture in Israel-past achievements and future directions in raising rotifers as food for marine fish larvae. Hydrobiologia 358, Lubzens, E., Zmora, O., Barr, Y., Biotechnology and aquaculture of rotifers. Hydrobiologia 446/447, Maeda, M., Hino, A., Environmental management for mass culture of rotifer, Brachionus plicatilis. Rotifer and Microalgae culture systems. Proc. US-Asia Workshop. The Oceanic Institute, Honolulu, pp Magnelli, P., Cipollo, J.F., Abeijon, C., A refined method for the determination of Saccharomyces cerevisiae cell wall composition and β-1,6-glucan fine structure. Anal. Biochem. 301, Makridis, P., Fjellheim, A.J., Skjermo, J., Vadstein, Ø., Control of the bacterial flora of Brachionus plicatilis and Artemia franciscana by incubation in bacterial suspensions. Aquaculture 185, Maris, P., Modes of action of disinfectants. Rev. Sci. Tech. Off. Int. Epiz. 14, Marques, A., François, J.-M., Dhont, J., Bossier, P., Sorgeloos, P., Influence of yeast quality on performance of gnotobiotically grown Artemia. J. Exp. Mar. Biol. Ecol. 310, Marte, C.L., Larviculture of marine species in Southeast Asia: current research and industry prospects. Aquaculture 227, Martinez-Diaz, S.F., Moreno-Legorreta, M., Vazquez-Juarez, R., Barrios-Gonzalez, J., Elimination of the associated microbial community and bioencapsulation of bacteria in the rotifer Brachionus plicatilis. Aquacult. Int. 11, Miyakawa, M., Muroga, K., Bacterial flora of cultured rotifer Brachionus plicatilis. Suisanzoshoku 35, Munro, P.D., Birkbeck, T.H., Barbour, A., Bacterial flora of rotifers (Brachionus plicatilis): evidence for a major location on the external surface and methods for reducing the rotifer bacterial load. In: Reinertsen, Jorgensen, Tvinnereim (Eds.), Fish Farming Technology. Munro, P.D., Henderson, R.J., Barbour, A., Birkbeck, T.H., Partial decontamination of rotifers with ultraviolet radiation: the effect of changes in the bacterial load and flora of rotifers on mortalities in start-feeding larval turbot. Aquaculture 170, Nicolas, J.L., Robic, E., Ansquer, D., Bacterial flora associated with a trophic chain consisting of microalgae, rotifers and turbot larvae: Influence of bacteria on larval survival. Aquaculture 83,