INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Several alternative strategies for the use of antimicrobials in disease control have been proposed and have already been applied successfully in aquaculture, such as the use of vaccines (11), the use of immunostimulants for the enhancement of the nonspecific defense mechanisms of the host, and the use of probiotic bacteria (5). Considering their recent successes, these alternative approaches have been defined by the Food and Agriculture Organization of the United Nations (18) as major areas for further research in disease control in aquaculture.

Already in 1980 Yasuda and Taga (26) anticipated that bacteria would be found to be useful not only as food for cultured aquatic species but also as biological controllers of disease. Several well-documented studies on the use of probiotics as biological control agents in the farming of bivalve mollusks, crustaceans, and fish were recently published (6, 8; S. Rengpipat and S. Rukpratanporn, Book Abstr. Fifth Asian Fish. Forum, 1998).

The probiotic application of Aeromonas media A199 was found to prevent death of the oyster Crassostrea gigas larvae when they were challenged in vivo with the pathogen Vibrio tubiashii, although A. media A199 was not able to persist more than 4 days on the host (6). The administration of the probiotic strain to the larvae caused a spectacular decrease of the pathogen densities in the larvae compared to those in the larvae treated with V. tubiashii only.

Rengpipat and Rukpratanporn (Book Abstr. Fifth Asian Fish. Forum) reported the use of a Bacillus strain, S11, as a probiotic administered to larvae of the black tiger shrimp Penaeus monodon via enriched brine shrimp, Artemia spp. The P. monodon larvae fed with the Bacillus-fortified Artemia had significantly shorter development times and fewer disease problems than larvae reared without the Bacillus. After feeding for 100 days, P. monodon postlarvae were challenged with the pathogenic Vibrio harveyi D331 by immersion. Ten days later all the groups treated with Bacillus S11 had 100% survival, whereas the control group had only 26% survival.

Siderophore-producing Pseudomonas fluorescens has been successfully applied as a biological control agent; it limited the mortality of 40-g rainbow trout (Oncorynchus mykiss) experimentally infected with Vibrio anguillarum (8). Siderophores are low-molecular-mass compounds with a very high affinity for ferric iron, whose biosynthesis is iron-regulated (26). A correlation was found between the production of siderophores and the protective action of P. fluorescens, suggesting that competition for free iron is involved in the mode of action (8).

Juvenile and adult brine shrimp are used increasingly as suitable live diets for different aquaculture species (17). The intensive culture of the brine shrimp Artemia has always suffered from unpredictable results due to incidental crashes in individual production tanks (24). In previous research, manipulation of the microbiota by preemptive colonization of the culture water with selected bacterial strains has been shown to improve the culture performance of Artemia (23). It was demonstrated under monoxenic conditions that the selected bacterial strains improved the nutritional quality of the dry food. The aim of this study was to investigate whether these selected strains can also be active as biological control agents against bacterial infections. Experimental infections of Artemia were done with Vibrio proteolyticus CW8T2, which has previously been shown to cause mortality in monoxenic Artemia cultures (23). The infection route was determined by means of transmission electron microscope observations. In vivo antagonism tests were performed to see whether the selected bacterial strains are able to protect Artemia from the pathogenic actions of V. proteolyticus CW8T2. In addition, filtrate experiments were done to verify whether extracellular compounds were involved in the protective action.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains. The nine probiotic strains originate from well-performing Artemia cultures and were selected based on their positive effect on the monoxenic and xenic culture of Artemia juveniles, as described by Verschuere et al. (23). V. proteolyticus CW8T2 was isolated from artificial feed from a sea bass hatchery in Spain and was kindly provided by L. Verdonck of the Department of Microbiology, University of Ghent.

Axenic hatching and culture of Artemia. In order to examine the action of added microbiota and avoid interference by other microorganisms, Artemia were disinfected with Merthiolate and axenically hatched as described by Verschuere et al. (23). Subsequently, 20 Artemia were transferred to sterile culture tubes containing 30 ml of autoclaved artificial seawater (Instant Ocean [33 g/liter]; Aquarium Systems, Sarrebourg, France). The culture tubes were kept at 28°C on a rotor. Gamma-irradiated (10 kGy) dry food was administered at a rate of 5 mg/day on days 0 and 1 after the transfer of the nauplii and 5.7 mg/day from day 2 onwards. Depending on the experiments, cultures were grown up to 3 or 5 days (see in vivo antagonism tests). The axenic condition of the control cultures was assessed regularly by inoculating 100 µl of the culture water on marine agar 2216 (MA) (Difco Laboratories, Detroit, Mich.). If any contamination occurred, the results of the experimental run were not accepted. Unless otherwise stated, this protocol was used whenever Artemia had to be cultured.

Pathogenicity of V. proteolyticus CW8T2 under conditions of feeding and starvation. The first experiment was aimed at detecting the effect of several concentrations of V. proteolyticus CW8T2 on Artemia survival, in order to determine an appropriate range for further experimental infections. V. proteolyticus CW8T2 was grown overnight in marine broth, centrifuged at approximately 10,000 × g and resuspended in Nine Salts Solution (NSS) (12). Beforehand, the relationship between the optical density at 550 nm and the plate count was established by plating dilutions of a suspension with known values of optical density at 550 nm on marine agar plates. Twenty-four hours after the transfer of the nauplii to axenic culture tubes, V. proteolyticus CW8T2 was added at different concentrations (1 × 0, 1 × 10², 1 × 10³, 1 × 10⁴, 1 × 10⁵, and 5 × 10⁶ CFU/ml). The highest concentration used in these experiments corresponded to the highest concentrations of Vibrionaceae present in Artemia culture water as determined on TCBS Cholera medium in previous experiments (24). The culture conditions of the Artemia nauplii were similar to the description above. Forty-eight hours after the addition of V. proteolyticus CW8T2, the surviving Artemia nauplii were counted. The experiment was performed twice, with three replicates per treatment.

Histological observation of infection process. Artemia nauplii experimentally infected with V. proteolyticus CW8T2 (10³ CFU/ml) 1 day after the transfer of the nauplii to the culture tubes were sampled after 1 day of infection for histological analysis. Axenically reared Artemia nauplii of the same age were also sampled. The Artemia nauplii were fixed in Karnovsky's fixative containing 2% (vol/vol) paraformaldehyde and 2.5% (vol/vol) glutaraldehyde in 0.2 M sodium cacodylate buffer, which was further diluted three times before use. After being rinsed in the same buffer the samples were postfixed in 2% (wt/vol) osmic tetroxide diluted in the same buffer. The samples were dehydrated in a graded concentration of ethanol and embedded in LX resin.

In vivo antagonism test. In vivo antagonism tests were done to examine the pathogenic action of V. proteolyticus CW8T2 in Artemia cultures preemptively colonized by each of the nine selected bacterial strains. Artemia nauplii were axenically hatched, transferred to the culture tubes, and maintained as described above. Immediately after the transfer of the nauplii, the culture water was inoculated with one of the selected bacterial strains at a calculated concentration of 5 × 10⁶ cells/ml. Twenty-four hours later, the Artemia cultures were experimentally infected with V. proteolyticus CW8T2 administrated in the culture water at a concentration of either 10² or 10³ CFU/ml. As control treatments Artemia nauplii were either (i) maintained under axenic conditions, (ii) infected only with V. proteolyticus CW8T2 (10² or 10³ CFU/ml), or (iii) inoculated only with the selected bacterial strain (5 × 10⁶ cells/ml). The survival of the Artemia nauplii was recorded 2 days after the experimental infection (3-day-old Artemia nauplii) by counting the number of living animals. For each strain at least two experimental runs were performed, with three or four replicates.

In vitro antagonism test. The production of inhibitory compounds against V. proteolyticus CW8T2 was assessed in vitro using the double-layer method (4). Marine agar plates were spot-inoculated with 10 µl of an overnight marine broth culture of one of the nine strains and were incubated at 28°C for 48 h. After killing the macrocolonies with chloroform vapors, a soft marine agar overlay (marine broth plus 7.5 g of agar-agar/liter) just inoculated with an overnight broth culture of V. proteolyticus CW8T2 (1/100 dilution) was poured on top of it. The plates were then incubated for 24 h at 28°C, and the presence of a clear inhibition zone around the macrocolony indicated that inhibitory compounds had been produced by the macrocolony. Each strain was tested in triplicate, and autoinhibition of V. proteolyticus CW8T2 was examined.

Filtrate experiments. In order to determine whether extracellular components are involved in the protective action of LVS2 and LVS8, experimental infections were done in Artemia cultures grown in sterile filtrates of these strains. The filtrates were made by inoculating the selected strains at a concentration of 5 × 10⁶ cells/ml in culture tubes containing 30 ml of autoclaved artificial seawater, but without Artemia. Dry food was added once at a dose of 7.14 mg/tube. The tubes were incubated for 2 days at 28°C. Subsequently the filtrates were prepared through filter sterilization (0.22 µm) of the bacterial suspensions. Tubes receiving only dry food were used to obtain a control filtrate. Another control consisted of untreated autoclaved artificial seawater. Furthermore, some of the culture tubes containing the filtrates of LVS2 and LVS8 were supplemented with 5 × 10⁶ CFU/ml of the corresponding strain to check whether results similar to those of the in vivo antagonism tests would be obtained.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Identification of bacteria. The nine bacterial strains were first identified based on several biochemical tests. These results are shown in Table 1. The strains that were tentatively identified as Vibrionaceae were further examined by FAME analysis and/or with API20NE. LVS3 and LVS9 were identified by FAME analysis as Aeromonas sp. and Vibrio alginolyticus, respectively. API20NE identified LVS3, LVS8, and LVS9 as Aeromonas hydrophila or Aeromonas caviae, V. alginolyticus, and V. alginolyticus, respectively, and confirmed the results obtained by FAME analysis.

Pathogenicity of V. proteolyticus CW8T2 under conditions of feeding and starvation. Under conditions of feeding, all Artemia died within 48 h after the experimental infection, regardless of the applied concentration (from 1 × 10² to 5 × 10⁶ CFU/ml), while less than 10% mortality was observed in the axenic control treatments.

View larger version (20K): [in this window] [in a new window]   FIG. 1.   Cumulative mortality of starved Artemia nauplii 24 and 48 h after administration of V. proteolyticus CW8T2 at a concentration of 103 CFU/ml and LVS8 at a concentration of 5 × 106 cells/ml. Different letters (above each bar) indicate significant differences in mortality after 48 h.

Histological observation of infection process. The process of infection of Artemia by V. proteolyticus CW8T2 on a histological level is illustrated in Fig. 2 to 5. In Fig. 2 epithelial cells from the anterior part of the gut of axenic Artemia are shown. The apical surface bordering the gut lumen shows short irregular microvilli and zonulae adherents between neighboring cells. Figure 3 shows a similar part of the gut of Artemia nauplii infected with V. proteolyticus CW8T2. The microvilli have disappeared, and the border of the epithelial cells seems to be liquefied. The cell junctions are affected, and the bacterial cells penetrate between the epithelial cells and force their way through the gut epithelium. Figures 4 and 5 show the body cavity in which the bacterial cells have penetrated and continue their devastating activity. One can see host cells that are not affected yet, such as muscle cells, fat-storing cells, and unidentified embryonic cells. Eventually these cells and whole tissues are also affected. In Fig. 5 one can see the bacterial cells surrounded by debris of destroyed host cells.

View larger version (137K): [in this window] [in a new window]   FIG. 2.   Epithelial cells of the anterior part of the gut of axenic Artemia. The apical cell surface is coated with short irregular microvilli (MV). Apical zonulae adherents link adjacent epithelial cells (arrow). N, nucleus. Bar = 1 µm. Magnification, ×9,000.

View larger version (143K): [in this window] [in a new window]   FIG. 3.   Electron micrograph demonstrating how V. proteolyticus CW8T2 causes lysis of the apical part of the epithelial cells of the anterior part of the gut. The apical zonulae adherents are destroyed, allowing the bacteria to penetrate between the epithelial cells. V, bacterial cells. Bar = 1 µm. Magnification, ×10,000.

View larger version (139K): [in this window] [in a new window]   FIG. 4.   V. proteolyticus CW8T2 in the body cavity of Artemia. PH, phagocytic storage cell; EP, epidermal cell; EM, unidentified embryonic cell; V, bacterial cell. Bar = 1 µm. Magnification, ×9,000.

View larger version (141K): [in this window] [in a new window]   FIG. 5.   V. proteolyticus CW8T2 among cell debris in the body cavity of Artemia. PH, phagocytic storage cell; EM, unidentified embryonic cell; MC, muscle cell; V, bacterial cell. Bar = 1 µm. Magnification, ×9,000.

In vivo antagonism test. The results of the in vivo antagonism tests are shown in Fig. 6. The survival of the Artemia nauplii was recorded 2 days after the infection. The treatments shown in Fig. 6 are the axenic control treatment, the cultures inoculated with only the selected bacterial strains (5 × 10⁶ cells/ml), and the cultures both inoculated with the selected strains (5 × 10⁶ cells/ml) and infected with V. proteolyticus CW8T2 at a dose of 10² or 10³ CFU/ml. In all Artemia cultures infected only with V. proteolyticus CW8T2 total mortality occurred within 2 days, regardless of the applied initial concentration of the pathogen (data not shown). The in vivo antagonism tests showed that all the selected strains were able to protect the Artemia from the pathogenic action of V. proteolyticus CW8T2, as the survival rate in the presence of each of these strains was always higher than that when only V. proteolyticus CW8T2 was added to the culture.

View larger version (53K): [in this window] [in a new window]   FIG. 6.   Survival of Artemia nauplii 48 h after infection with V. proteolyticus CW8T2 in culture water preemptively colonized with the nine selected bacterial strains (LVS1 to LVS9). Culture types are indicated by bar fill patterns, as follows: black, axenic Artemia culture (control); cross-hatched, Artemia culture preemptively colonized by the selected bacterial strain; white, Artemia culture preemptively colonized by the selected bacterial strain and infected with V. proteolyticus CW8T2 at a concentration of 102 CFU/ml; and striped, Artemia culture preemptively colonized by the selected bacterial strain and infected with V. proteolyticus CW8T2 at a concentration of 103 CFU/ml. Data are averages plus standard deviations (error bars).

View larger version (19K): [in this window] [in a new window]   FIG. 7.   Viable counts of V. proteolyticus CW8T2 in the culture water (A) and survival of Artemia up to 4 days after infection with V. proteolyticus CW8T2 (103 CFU/ml) (B) in the cultures preemptively colonized with LVS2 () and LVS8 () and in axenic culture () Data are averages ± standard deviations (error bars). d, days.

Colonization capacity of selected bacterial strains and the pathogen. The colonization of Artemia by the selected bacterial strains was quantified 24 h after the inoculation of the culture water. The results of the MA counts are given in Table 2. It is striking that the strains that gave only partial protection against the pathogen in the in vivo antagonism test (Fig. 6) also showed a restricted capability to colonize Artemia (from 2.23 to 2.954 log CFU/Artemia nauplius for LVS1, LVS2, LVS5, and LVS7). In comparison, the strains totally protecting the Artemia nauplii showed a higher colonization capacity (from 3.37 to 4.18 log CFU/Artemia nauplus for LVS3, LVS4, LVS6, LVS8 and LVS9).

In vitro antagonism test. The results of the double-layer in vitro antagonism tests were all negative, as no inhibition zone could be observed around the macrocolonies of the selected strains. Also, no autoinhibition of V. proteolyticus CW8T2 was observed. The same observations were made with the modified procedure.

Filtrate experiments. Artemia was cultured in sterile filtrates of LVS2 and LVS8 to see whether extracellular compounds are involved in the protection of the animals against V. proteolyticus CW8T2. Two days after the experimental infection (10³ CFU/ml) the survival in the different treatments was determined (Table 3). The filtrates of both LVS2 and LVS8 did not significantly increase the survival compared to the infected control filtrate and the infected control seawater, as virtually total mortality occurred in all those treatments. When living cells of LVS2 and LVS8 were added to their corresponding filtrates, results were similar to those obtained in the in vivo antagonism tests (Fig. 6): the living cells of LVS2 and LVS8 partially and totally protected Artemia against the pathogenic action of V. proteolyticus CW8T2, respectively.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this study V. proteolyticus CW8T2 has been identified as a virulent pathogen for Artemia (Fig. 1) and the infection process of V. proteolyticus CW8T2 has been visualized by transmission electron microscopy (Fig. 2 to 5). It was shown that preemptive colonization of the culture water with selected bacterial strains did protect, at least partially, the Artemia nauplii against the pathogenic effects of the strains (Fig. 6 and 7). The in vitro antagonism tests and the filtrate experiments showed that probably no extracellular compounds were involved in the protective action (Fig. 7 and Table 3).

Although V. proteolyticus CW8T2 demonstrated a high virulence under both feeding and starvation conditions, mortality occurred faster when Artemia nauplii were fed (Fig. 1). It is possible that under fed conditions V. proteolyticus CW8T2 developed faster due to the nutrient enrichment of the culture medium or that ingestion of the pathogen was increased due to adhesion to the food particles. Infection of fed Artemia nauplii has also been accomplished in previous experiments 4 days after hatching instead of 1 day after hatching. Despite the later development stages, all Artemia nauplii in those experiments also died within 48 h (data not shown), indicating that a better resistance towards infection with V. proteolyticus CW8T2 in the in vivo antagonism tests could not simply be assigned to a later developmental stage of Artemia due to the nutritional contribution of the selected strains.

The infection route of V. proteolyticus CW8T2 was monitored by electron microscopy (Fig. 2 to 5). It was observed that the bacteria penetrated through the gut epithelium and invaded the body cavity. The gut epithelium and the underlying cells and tissues were clearly affected by the devastating action of the pathogen, which probably eventually causes death of the organism. The observed decrease in blood cell numbers in infected Artemia compared to axenic ones could not be explained. No phagocytizing blood cells were found in the infected Artemia, contrary to the observations of G. G. Martin et al. (submitted for publication). It was also striking that the colonization capacity of V. proteolyticus CW8T2 was much higher than that of the selected strains, despite the much lower initial density (10³ CFU/ml versus 5 × 10⁶ cells/ml) (Table 2). This explosive growth capacity probably contributed to the virulence of the pathogen, causing a rapid death of infected Artemia.

Several pathogens for Artemia have been described so far (2, 10, 13, 14, 16). Overton and Bland (13) described extensively the infection process of Artemia by the fungus Haliphthoros milfordensis, from the attachment to the exoskeleton to the utilization of the host tissues, with final invasion of the gut by the fungus. Gunther and Catena (10) exposed Artemia nauplii to three species of Vibrio at a concentration of 10⁸ cells/ml. Two of the three strains coated the shrimp's body in 2 to 8 h, as was shown on scanning electron micrographs of the nauplii, and completely inhibited swimming. Solangi et al. (16) observed also infestation of brine shrimp by a filamentous bacterium tentatively identified as Leucothrix mucor at the shrimp's exterior, causing a slow death of the Artemia. Tyson (19, 20) observed spirochetes inside the Artemia's body, but the infection route was not clarified and it was not clear whether this infection caused mortality or not. Similarly to this study, Grisez et al. (9) examined the infection route of V. anguillarum in turbot Scophthalmus maximus larvae and showed with immunohistochemistry that V. anguillarum was transported stepwise through the gut wall and subsequently by the blood to the different organs, eventually leading to septicemia and mortality.

The in vivo antagonism tests showed that all the selected strains were able to protect Artemia, at least partially, against the pathogenic action of V. proteolyticus CW8T2 (Fig. 6 and 7). LVS8 suppressed the growth of V. proteolyticus CW8T2 dramatically (Fig. 7A). The growth suppression of V. proteolyticus CW8T2 shown in Fig. 7A was clearly correlated with Artemia survival depicted in Fig. 7B, as survival after 2 days was the highest in the culture waters carrying the lowest density of V. proteolyticus CW8T2 and vice versa. Although the viable count of the pathogen still increased during the culture period, the survival of the Artemia cultured with LVS8 was apparently not affected, probably as a consequence of a lower exposure to the pathogen. Contrary to the observations of Gibson et al. (6), the densities of the pathogen did not decrease in the culture, but the proliferation of V. proteolyticus CW8T2 was slowed down. It is likely that eventually the concentration of V. proteolyticus CW8T2 will increase to a lethal concentration. Delay in mortality following a probiotic treatment was also observed by Gildberg and Mikkelsen (7). They supplemented a commercial dry feed with two strains of Carnobacterium divergens isolated from fish intestines and administered it during 3 weeks to Atlantic cod (Gadus morhua) fry. Twelve days after the infection with a virulent V. anguillarum a lower cumulative mortality was recorded, but 4 weeks after the infection the same cumulative mortality as in the control group was reached. Thus, the main effect of the probiotic treatment was a delay in mortality of infected cod fry. The authors argued, however, that this observation does not exclude the considerable importance of such methods under normal rearing conditions in the presence of moderate levels of opportunistic bacteria. Furthermore, total protection by LVS8 was demonstrated at least up to 4 days after the experimental infection (Fig. 7B). This time period can be qualified as considerable, as the culture period for Artemia juveniles in stagnant culture system is usually limited to 7 days (3).

No inhibition of V. proteolyticus CW8T2 was observed in the in vitro antagonism test (double-layer methods). Also the experimental infections of Artemia grown in the filtrates of LVS2 and LVS8 led to total mortality within 48 h (Table 3). One can conclude that no extracellular compounds such as antibiotics or siderophores are involved in the suppression of V. proteolyticus CW8T2 but that living cells are required. The observation that all the selected strains at least partially protected the Artemia after the experimental infection (Fig. 6) suggests a general mode of action, such as competition for chemicals, available energy, or adhesion sites. The correlation between the colonization potential and the protective ability of the selected strains is striking (Table 2 and Fig. 6, respectively). The lowest level of protection was observed for the strains with the lowest colonization capacity. This observation could support the hypothesis of competition for adhesion sites on or in the shrimps, but a higher colonization may also be a consequence of a more efficient use of resources like chemicals and available energy present in the ambient environment. The observed growth suppression in the culture medium (Fig. 7) and the fact that extracellular compounds do not seem to be involved in the protective action (Table 3) indicate that preemptive colonization allows the selected bacterial strains to compete efficiently for chemicals or available energy with the pathogen and to suppress its development.

It has been demonstrated in this study that preemptive colonization by the selected bacterial strains could prevent the proliferation not only of V. proteolyticus CW8T2 but probably also of other pathogens or opportunistic pathogens in the culture of Artemia juveniles. This shows that, apart from their nutritional contribution demonstrated by Verschuere et al. (23), the selected bacterial strains could also act as biological control agents of infections.