Hybrid striped bass (Morone saxatilis × Morone chrysops) were purchased from Delmarva Aquatics (Smyrna, DE, USA). They were held and reared in 17,740 L of fresh water in a closed recirculating aquaculture system (RAS) with an installed sump tank, 1.5 hp water pump, PBF-10 prop washed bead filter, and inline UV sterilizer. The water temperature and the room humidity were maintained at 26 ± 1°C and 60–70%, respectively. Aeration was provided by a central regenerative air blower and individual tank air diffusers and light/dark cycles were set at 12 h. The fish were fed twice daily based on total biomass in the system. Feed-levels began at 0.03% total biomass when the fish were smaller, and reduced to 0.01% as they grew. Diet was Cargill Hybrid Striped Bass Food (Cargill Inc., Minneapolis, MN, USA) including 45% crude protein and 12% crude fat.

For the challenge with Vibrio spp. to the aquaculture tanks, 3 Vibrio spp.: V. anguillarum HB155721 was purchased from Carolina Biological Supply Co. (Burlington, NC, USA) and V. parahaemolyticus O1:Kuk, and V. vulnificus MLT1009 were provided from U.S. Department of Agriculture–Agricultural Research Service, Dover, DE (USDA-ARS). V. parahaemolyticus O1:Kuk strain may potentially cause human illness, because this strain is closely related with strain O3:K6 which was associated with quite a number of Asian outbreaks (Depaola et al., 2003). V. vulnificus MLT1009 has VvhA (virulent factor of V. vulnificus) gene that induces autophagy-related cell death (Song et al., 2016). Each was cultivated overnight in Tryptic Soy Broth, (TSB, Carolina Biological Supply Co.) supplemented with 2% additional NaCl at 37°C, and washed twice with normal saline (0.85% NaCl) buffer. The washed Vibrio spp. were suspended in the saline buffer. To enumerate the bacteria, the suspended Vibrio spp. were serial decimal diluted with the saline buffer and plated on Tryptic Soy Agar (TSA, Carolina Biological Supply Co.) supplemented with 2% NaCl. After incubating for 24 h at 37°C, the colonies were counted. The Vibrio spp. were inactivated by UV radiation just prior to the challenge. V. anguillarum and V. vulnificus were irradiated by UV light for 8 h and V. parahaemolyticus was irradiated overnight. These inactivated Vibrio spp. were administered into the experimental aquaculture tanks.

Sixteen hybrid striped bass were transferred from a rearing tank to each experimental tank and adapted to the experimental tank environments with increasing salinity gradually for 2 weeks. The average weight of the fish was ~450 g (ranging from 340 to 567 g). They were held in separate but identical closed RASs each containing 1,005 L of seawater at 32 ppt salinity. Each system consisted of 1 culture tank, seawater pump, solids filter, and covered biofiltration sump tank filled with bio-ball media (Figure 1). The water temperature and room humidity were controlled at 26 ± 1°C and 60–70%, respectively. Aeration was provided by a diaphragm aerator with individual air diffusers placed in the biofilter sump and light/dark cycles were maintained as 12/12 h. The fish were fed a minimal maintenance ration of two pellets per fish every other day during the trials.

Four experimental groups (control, group2, group3, and group4) were allocated to four experimental tanks. After the adaptation of the fish in the experimental tanks, 2 fish and 500 ml of water were collected from each experimental tank to determine the baseline data (day 0 sampling). On the day following collection of day 0 samples, heat-inactivated cultures of the three Vibrio spp. were inoculated into experimental tanks except for control group at a final concentration of 2 Log cells ml⁻¹ in group 2, 3 Log cells ml⁻¹ in group 3, and 4 Log cells ml⁻¹ in group 4. After 1 h circulation of the inoculated bacteria, two fish and water samples were collected from each experimental tank as a day 1 sampling. The sampling was repeated exactly at 2, 4, and 6 days. Immediately after removing fish, they were euthanized by a manually applied blunt force trauma (cranial concussion) followed by pithing and were fileted for bacterial analyses. The fish fileting was hygienically conducted at bio safety level 2 clean laboratory. This trial was duplicated by exactly the same procedures. The overall experimental design, including fish sampling, was reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at Delaware State University.

To clarify the correlation of bacterial populations between fish and the quality of their surrounding water, randomized aquaculture fish and water samples were collected from the Aquaculture Research Facility at Delaware State University in Dover, DE, USA and Horn Point Research Facility at University of Maryland in Cambridge, MD, USA. The samples were collected from eight independent aquaculture systems, including five water tanks and three ponds. The fish were mummychog (Fundulus heteroclitus), tilapia (Oreochromis niloticus), hybrid striped bass (Morone Chrysops × Morone saxatilis), blue catfish (Ictalurus furcatus), and striped bass (Morone saxatilus) cultured inside the facility. as Additionally, baitfish, large size channel catfish (Ictalurus punctatus), and small size channel catfish (Ictalurus punctatus) were cultured in outdoor ponds. The size and age of the collected fish were random and widely variable. Mummychog, the smallest fish, was less than 10 cm and blue catfish, the largest one, was about 50 cm. The age of each fish ranged anywhere from 1 month to 2 years old. Two fish and water samples were collected from each tank. The samples were processed similarly to the aquarium experiments described previously and skinless filets were used for quantifying Vibrio spp. and total bacteria.

Detection of the three Vibrio spp. and total bacteria in fish filets and aquaculture water samples was carried out as described in a previous study (Kim and Lee, 2014). Briefly, the filets were cut into small pieces and 25 g of fish tissue was homogenized with 75 ml of 0.85% saline buffer by a Bag Mixer (Interscience, St. Nom, France). The homogenized fish filets were centrifuged at 150 × g for 5 min to remove large fish tissue and the supernatants were re-centrifuged at 10,000 × g for 10 min. The pellets were re-suspended and plated on Thiosulfate Citrate Bile Salts Sucrose agar (TCBS; Difco, Sparks, MD, USA) and TSA using Eddy Jet 2 spiral plater (Neutec Group Inc., Farmingdale, NY, USA). Colonies were enumerated on TCBS and TSA at 1 and 3 days incubation, respectively. The pellets were also used for the extraction of bacterial genomic DNA. The DNA was extracted by a boiling method with TZ buffer (20 mg ml⁻¹ Triton X-100 and 2.5 mg ml⁻¹ sodium azide in 0.1 M Tris-HCl, pH 8.0) and 6% chelex solution (Bio-rad Laboratories, Inc., Hercules, CA, USA) from the pellets and precipitated using a Quick-Precip Plus Solution (Edgebio, Gaithersburg, MD, USA) and absolute ethanol (Fisher Scientific, Fair Lawn, NJ, USA). The precipitated DNA was dissolved in nuclease-free water.

The collected aquaculture water samples were also analyzed by culture-dependent enumeration of the bacterial counts using the Eddy Jet 2 spiral plater. In order to retrieve bacteria, the water samples were vacuum filtered through polyethersulfone membrane filters (0.22 μm pore size, 25 mm diameter; Whatman, Buckinghamshire, UK). The filters were soaked in TZ buffer and 6% Chelex solution and then vortexed for 2 min. The bacterial DNA from the filters was extracted by boiling and precipitation methods as described above.

Using the genomic DNA extracted from fish filets and water samples, quantification of Vibrio spp. and total bacteria in the samples were determined using the multiplex real-time PCR assay reported by Kim and Lee (2014). Species specific target genes, tlh for V. parahaemolyticus, toxR for V. anguillarum, vvhA for V. vulnificus were used for detection of each Vibrio spp. and 16S rDNA was used for detection of total bacteria. The 25 μL reaction mixture consisted of 12.5 μL of 2X master mix, 800 nM each of the toxR forward and reverse primers and probe; 200 nM each of the 16S rDNA forward and reverse primers and probe; and 50 nM each of the tlh and vvhA forward and reverse primers and probe. The cycling parameters consisted of a 95°C initial denaturation hold for 2 min followed by 35 cycles of amplification, with each cycle consisting of denaturation at 95°C for 15 s and a combined annealing/extension step at 60°C for 50 s. The detection limits of the multiplex assay were 1 cell ml⁻¹ in seawater, 10 cells g⁻¹ in fish for Vibrio spp. and 4 cells ml⁻¹ in seawater and 40 cells g⁻¹ in fish for total bacteria.

Analysis of variation (ANOVA) was used to evaluate a significant difference between bacterial populations through experimental groups and periods. Tukey post hoc test was applied to determine which groups and days were statistically different. Also, Pearson's correlation was used to assess whether there is a correlation between inoculated bacterial counts in water and the final counts in fish filets. All bacterial concentrations were transformed to log₁₀ scale, and the IBM SPSS Statistics for Windows, Version 22.0 (IBM Corp., Armonk, NY, USA) was used to explore the statistical significance. A confidence interval at the 95% level (P < 0.05) was considered in all cases.

To detect culturable bacteria in the fish and water inoculated with UV-irradiated Vibrio spp. in the aquaculture system, the samples were plated on TSA and TCBS agar. The general medium TSA was used to detect total bacteria and selective medium TCBS was for detection of Vibrionaceae bacteria. Through all experimental periods including before and after Vibrio challenging, total bacteria were counted on TSA from 2.3 log to 3.5 log cells ml⁻¹ in the water samples and from 2.2 log to 4.0 log cells g⁻¹ in the fish samples. In the case of Vibrionaceae, between 1.2 and 2.6 log cells ml⁻¹ in the water samples and below 1.7 log cells g⁻¹ in the fish samples were counted on TCBS agar. Some TCBS agar plated fish samples did not detect any colonies. There were no significant differences between control and Vibrio challenged groups determined by culture-dependent method. Also, there was no change of bacterial number in fish and water samples during all experimental periods. Only minor differences among individual water and fish samples were found.

The inoculated Vibrio spp. and total bacteria counts in the aquaculture water and fish were detected by multiplex real-time PCR assay. In the water samples, total bacteria counts in control tank had no significant (p = 0.717) change through the experimental periods around 3.2 log cells ml⁻¹ (Figure 2A), whereas, the counts of total bacteria in Vibrio challenged tanks peaked on day 1 and then decreased (Figures 2B–D). The bacterial numbers in the water at day 1 were directly proportional to the amount of challenged Vibrio spp. (Control, 3.13 ± 0.39; Group 2, 3.48 ± 0.67; Group 3, 3.98 ± 0.65, Group 4, 4.81 ± 0.56 log cells ml⁻¹). In the fish filet samples, the number of total bacteria in the control group and challenge group 2 were relatively constant from days 0 through 6 of the experiment remaining around 3.7 log cells g⁻¹ (Figures 2A,B). However, the total bacteria numbers in challenge groups 3 and 4 eventually accumulated from 3.87 ± 0.08 log cells g⁻¹ to 4.76 ± 0.26 log cells g⁻¹ in group 3 and from 3.95 ± 0.12 log cells ml⁻¹ to 5.28 ± 0.20 log cells g⁻¹ in group 4 (Figures 2C,D) over the experiment.

V. anguillarum was not detected in our aquaculture systems before challenges (Figure 3). After the challenge, the V. anguillarum showed a sharp peak of bacterial counts at day 1 and the levels in the water were directly proportional with the challenged amounts (Group 2, 1.97 ± 0.12; Group 3, 2.85 ± 0.16; Group 4, 3.85 ± 0.01 log cells ml⁻¹). After day 1, the counts of V. anguillarum quickly decreased in the water and fish (Figure 3). At day 6, V. anguillarum was not detected in the 2 log cells ml⁻¹ challenged fish filets, whereas, 0.78 ± 0.43 log cells g⁻¹ and 1.90 ± 0.21 log cells g⁻¹ of V. anguillarum remained in each in the 3 log and 4 log cells ml⁻¹challenged fish (Figures 3C,D).

Low levels (under 1 log cells ml⁻¹) of V. parahaemolyticus already inhabited in the control aquaculture water, but were not detected in fish (Figure 4A). After the challenge, V. parahaemolyticus counts in the water showed similar patterns with V. anguillarum (Figures 4B–D), which were the highest number at day 1 and then decreasing. However, the V. parahaemolyticus counts in the fish did not decrease but maintained through day 1 to day 6, (Figures 4B–D). On day 6, the residues of V. parahaemolyticus were 2.19 ± 0.23, 2.36 ± 0.53, and 3.21 ± 0.34 log cells g⁻¹ in Group 2, 3, and 4 fish filets, respectively.

Before Vibrio challenge, V. vulnificus existed in the experimental aquaculture systems at relatively high levels, around 2.1 log cells ml⁻¹, and fish, about 2.0 log cells g⁻¹ (Figure 5A). Similar to the other challenges, the levels of V. vulnificus in the water increased on day 1 and then dramatically decreased in groups 3 and 4 (Figures 5C,D). Uniquely, in the group 2 water, the level of V. vulnificus was highest at day 2 and then decreased (Figure 5B). The V. vulnificus counts finally settled back to original levels (day 0, 2.29 ± 0.09; day 1, 4.13 ± 0.77; day 6, 2.42 ± 0.18 log cells ml⁻¹ in group 4) in the water. However, the residues of V. vulnificus in the fish were significantly higher on day 6 than those on day 0 in all challenge groups (group 2, day 0, 2.47 ± 0.16, day 6, 3.08 ± 0.34; group 3, day 0, 2.43 ± 0.25, day 6, 3.39 ± 0.20; group 4, day 0, 2.56 ± 0.09, day 6, 4.14 ± 0.26 log cells g⁻¹; p < 0.001; Figures 5B–D).

On the final day 6, strong correlations were observed between challenged Vibrio amounts in water and their residual population in fish filets (total bacteria, r = 0.794, p < 0.001; V. anguillarum, r = 0.754, p < 0.001; V. parahaemolyticus, r = 0.943, p < 0.001; V. vulnificus, r = 0.942, p < 0.001; Figure 6).

To further clarify the correlation of bacterial populations between fish and its surrounding aquaculture water, eight separate aquaculture systems were assessed. In randomly sampled aquaculture systems, total bacteria were detected from about 3.1 log to 4.1 log cells ml⁻¹ in the water and 2.6 log to 5.1 log cells g⁻¹ in the fish (Table 1). While V. parahaemolyticus and V. vulnificus were detected in water and fish held in aquaculture tanks 1 and 6 by multiplex real-time PCR, Vibrionaceae bacteria in tank 6 was not detected on TCBS agar. In the aquaculture tank 3, only V. vulnificus was detected. Also, in the tank 4 and 5, V. anguillarum and V. vulnificus were detected from aquaculture water and fish by culture independent multiplex PCR but Vibrionaceae was detected from only water by culture dependent assay. There was no detection of Vibrio spp. in aquaculture tanks 2, 7, and 8 (Table 1).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.