Bacterial Skin Microbiota of Seabass from Aegean Fish Farms and Antibiotic Susceptibility of Psychrotrophic Pseudomonas

Farming seabass (Dicentrarchus labrax) is an essential activity in the Mediterranean basin including the Aegean Sea. The main seabass producer is Turkey accounting for 155,151 tons of production in 2021. In this study, skin swabs of seabass farmed in the Aegean Sea were analysed with regard to the isolation and identification of Pseudomonas. Bacterial microbiota of skin samples (n = 96) from 12 fish farms were investigated using next-generation sequencing (NGS) and metabarcoding analysis. The results demonstrated that Proteobacteria was the dominant bacterial phylum in all samples. At the species level, Pseudomonas lundensis was identified in all samples. Pseudomonas, Shewanella, and Flavobacterium were identified using conventional methods and a total of 46 viable (48% of all NGS+) Pseudomonas were isolated in seabass swab samples. Additionally, antibiotic susceptibility was determined according to standards of the European Committee on Antimicrobial Susceptibility Testing (EUCAST) and Clinical and Laboratory Standards Institute (CLSI) in psychrotrophic Pseudomonas. Pseudomonas strains were tested for susceptibility to 11 antibiotics (piperacillin-tazobactam, gentamicin, tobramycin, amikacin, doripenem, meropenem, imipenem, levofloxacin, ciprofloxacin, norfloxacin, and tetracycline) from five different groups of antibiotics (penicillins, aminoglycosides, carbapenems, fluoroquinolones, and tetracyclines). The antibiotics chosen were not specifically linked to usage by the aquaculture industry. According to the EUCAST and CLSI, three and two Pseudomonas strains were found to be resistant to doripenem and imipenem (E-test), respectively. All strains were susceptible to piperacillin-tazobactam, amikacin, levofloxacin, and tetracycline. Our data provide insight into different bacteria that are prevalent in the skin microbiota of seabass sampled from the Aegean Sea in Turkey, and into the antibiotic resistance of psychrotrophic Pseudomonas spp.


Introduction
Seafood, especially fish, is an increasingly important component of human diets. Thus, aquaculture is an important source of food suitable for human consumption [1], and could provide a sustainable supply of affordable seafood to an increasing global population. Mediterranean marine aquaculture grew exponentially during the last decades of the 20th century, though at a slower pace over the past 20 years or so [2]. European seabass (Dicentrarchus labrax) is the 31st most-reared fish in worldwide aquaculture [3]. Seabass production increased by 2.9% in 2020 and reached 243,900 tons globally [4]. More than 95% of the world's seabass and sea bream (Sparus aurata) production comes from aquaculture, of which, 97% accounts for the production in Mediterranean countries. Turkey and Greece are the primary producers, while Spain, France, Italy, Greece, and Turkey are the primary consumers [5].
During June 2022, 96 seabass with an average weight of 300 g and average length of 220 mm were obtained from fish farms in 12 locations (8 samples per farm) in the Aegean Sea. These fish farms belonged to five different aquaculture companies and were labeled using capital letters with a numerical subscript (i.e., A1, A2, A3, B1, B2, C1, C2, D1, D2, E1, E2, and E3) ( Figure 1). The collected fish were stored in styrofoam boxes containing aseptic ice and transported within 4-6 h in refrigerated vehicles (+4 • C) to the international market chain in Istanbul. The styrofoam boxes were opened immediately on arrival under aseptic conditions. The central temperature in the boxes was ≤+4 • C measured with a thermometer (Testo, Lenzkirch, Germany). Under the same conditions, the samples were taken by rubbing off the skin of the seabass with sterile swabs containing a transport liquid medium (Becton Dickinson, NJ, USA). The swabs were transported under refrigeration temperatures in thermal boxes (≤+4 • C) to the laboratory (Department of Food Science and Technology, Istanbul University-Cerrahpasa) for immediate analyses.

Total DNA Extraction
Total DNA extraction was carried out directly from the swab samples by applying the phenol/chloroform/isoamyl alcohol method [15]. For this purpose, 2 mL swab samples were centrifuged at 14,000 rpm for 5 min at room temperature. The pellet was resuspended in 500 µL 1×TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) containing 5 mg/mL lysozyme (Applichem, Darmstadt, Germany) and the phenol/chloroform/isoamyl alcohol was applied. Finally, the extracted DNA samples were resuspended in 30 µL sterile deionized water and stored at −20 °C for amplicon PCR experiments in NGS studies.

Next-Generation Sequencing
16S rRNA amplicon sequencing and DNA library preparation were carried out according to the 16S metabarcoding sequencing library preparation guide [16]. The primers for the amplicon PCR were F-primer:5′-TCGTCGGCAGCGTCAGATGTGTATAAGAGA-CAGCCTACGGGNGGCWGCAG-3′ and R-primer: 5′-GTCTCGTGGGCTCGGA-GATGTGTATAAGAGACAGGACTACHVGGGTATCTAACC-3′. Bacterial 16S rRNA V3-V4 gene regions were amplified using a KAPA HiFi HS kit (Roche, Mannheim, Germany). PCR products from each seabass sample were indexed with dual indexes using a Nextera ® XT Index Kit v2 Set-A (Illumina, San Diego, USA). All the amplicon PCR products and indexed amplicons were purified using AMPure XP magnetic beads (Beckman Coulter, Indianapolis, USA). The prepared equimolar proportions (10 nM) of the samples were pooled, and diluted to a 35 pM library containing 5% (v/v) PhiX control DNA (Illumina).

Total DNA Extraction
Total DNA extraction was carried out directly from the swab samples by applying the phenol/chloroform/isoamyl alcohol method [15]. For this purpose, 2 mL swab samples were centrifuged at 14,000 rpm for 5 min at room temperature. The pellet was resuspended in 500 µL 1×TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) containing 5 mg/mL lysozyme (Applichem, Darmstadt, Germany) and the phenol/chloroform/isoamyl alcohol was applied. Finally, the extracted DNA samples were resuspended in 30 µL sterile deionized water and stored at −20 • C for amplicon PCR experiments in NGS studies.

Next-Generation Sequencing
16S rRNA amplicon sequencing and DNA library preparation were carried out according to the 16S metabarcoding sequencing library preparation guide [16]. The primers for the amplicon PCR were F-primer: 5 -TCGTCGGCAGCGTCAGATGTGTATAAGAGAC AGCCTACGGGNGGCWGCAG-3 and R-primer: 5 -GTCTCGTGGGCTCGGAGATGTGTATAA GAGACAGGACTACHVGGGTATCTAACC-3 . Bacterial 16S rRNA V3-V4 gene regions were amplified using a KAPA HiFi HS kit (Roche, Mannheim, Germany). PCR products from each seabass sample were indexed with dual indexes using a Nextera ® XT Index Kit v2 Set-A (Illumina, San Diego, CA, USA). All the amplicon PCR products and indexed amplicons were purified using AMPure XP magnetic beads (Beckman Coulter, Indianapolis, IN, USA). The prepared equimolar proportions (10 nM) of the samples were pooled, and diluted to a 35 pM library containing 5% (v/v) PhiX control DNA (Illumina). Subsequently, a 20 µL library was loaded into an iSeq100 v1 cartridge. The sequencing was carried out using the iSeq100 system (Illumina) pair end read type and two reads of 151 bp read length.

Metabarcoding Analysis
The sequencing reads from the 16S rRNA gene were analyzed using Silva NGS software version 138.1, VSEARCH 2.17.0, SINA v1.2.10 for ARB SVN (revision 21008), and BLASTn version 2.11.0+. Trimming of adapter sequences from short NGS read data was performed using Genious Prime software. The amplicons were clustered based on the sequence identity operational taxonomic unit (OTU) approach. Clustering Ward's analysis was applied using the PAleontological STatistics (PAST) Software version 4.11 package (2022) at the genus levels in the seabass samples [17].

Isolation and Identification of Pseudomonas
Pseudomonas spp. isolation and identification were performed using the modified conventional TS EN ISO 13720 standard [18]. First, 250 µL of the swab sample containing each liquid medium was taken and placed in 2 mL of Pseudomonas Broth (Z699101 Merck, Darmstadt, Germany) and incubated at 22 ± 2 • C for 44 ± 4 h (Pre-enrichment). Subsequently, 0.
The E-test (Bioanalyse, Turkey) was applied to determine the Minimum Inhibitory Concentration (MIC) of Pseudomonas strains that were found to be resistant to antibiotics in the disc diffusion test. Results were evaluated according to the EUCAST [13] and CLSI [14] breakpoint tables.

NGS and Metabarcoding Analysis Results
Modern high-throughput methods have substituted conventional culture-based microbiological techniques, increasing our understanding of fish microbial communities throughout the production chain, from harvesting through storage distribution, until the end of shelf life [20]. In this study, the alpha diversity of bacteria was estimated to determine the diversity within samples, and the Shannon species diversity index values were determined using Silva NGS software (Table 1). This diversity index is a quantitative measure for estimating the number of different species in a given environment and their relative abundance [21]. This can be relevant for identifying the bacterial diversity in skin seabass samples because skin mucus harbors a complex bacterial community [22]. Metabarcoding analysis of 189,207 sequences from 96 seabass skin samples led to 123,391 OTUs, 39,737 clustered sequences, and 164,870 classified sequences. The results indicated that the phylum Proteobacteria was dominant in all seabass skin samples. The skin microbiota samples also contained bacteria belonging to the phyla Firmicutes and Bacteroidota (Figure 2). At the genus level, Pseudomonas was the dominant genus among the 96 seabass swab samples. (Figure 3). Additionally, Shewanella, Acinetobacter, and Flavobacterium were also among the most prevalent genera (Figure 3). Similar results were reported from the Bodrum coast in seawater, Mugla [23]. The genus Pseudomonas is considered to be an important fish pathogen as it comprises some (sub) species which are opportunistic pathogens to humans [23]. Another study dedicated to examining the microbiota of whole and filleted seabass [20] presented results similar to those we obtained. Pseudomonas was dominant in seabass samples, based on the 16S rRNA metabarcoding analysis, followed by the presence of Shewanella. Among animal food products, fish are the most vulnerable to bacterial spoilage and Shewanella has previously been reported as a main contributor in the microbiota of spoiled seafood, such as hake fillets [24]. Additionally, Shewanella was the dominant genus in MAP-stored seabass fillets, but its relative abundance declined dramatically towards the end of the products' shelf life [19]. Acinetobacter are abundant in aquatic environments and frequently isolated from the skin and gills of fresh fish [25]. In a previous study, Acinetobacter were the dominant bacteria in seabass fillets [20] and rainbow trout samples [26]. However, Acinetobacter are not recognized as important spoilage bacteria [27] as they cannot hydrolyze fish proteins and are thus, a weak producer of biogenic amines, as well as a weak degrader of ATP-related compounds [28].
the most vulnerable to bacterial spoilage and Shewanella has previously been reported as a main contributor in the microbiota of spoiled seafood, such as hake fillets [24]. Additionally, Shewanella was the dominant genus in MAP-stored seabass fillets, but its relative abundance declined dramatically towards the end of the products' shelf life [19]. Acinetobacter are abundant in aquatic environments and frequently isolated from the skin and gills of fresh fish [25]. In a previous study, Acinetobacter were the dominant bacteria in seabass fillets [20] and rainbow trout samples [26]. However, Acinetobacter are not recognized as important spoilage bacteria [27] as they cannot hydrolyze fish proteins and are thus, a weak producer of biogenic amines, as well as a weak degrader of ATP-related compounds [28].     P. lundensis was identified by NGS analysis of all seabass samples. Similar to our results, Elbehiry et al. [29] reported that, in red meat samples, P. lundensis was the dominant species. Pseudomonads are highly opportunistic and may become a highly threatening fish pathogen causing serious illness including ulcerative syndrome and hemorrhagic septicemia [30]. Enterococcus were found in 15 seabass samples, of which samples S1, S4, and S5 were sampled from the same fish farm. The other Enterococcus-containing samples were S11, S34, S45, S53, S54, S57, S64, S65, S66, S67, S68, and S69 identified from four different fish farms (D2 and C2 located in Mugla, A2 and E2 located in Izmir). Detection of Enterococcus spp. in sea bass skin samples may indicate fecal contamination in seawater.
The highest Shannon diversity index in this study (7.25) was obtained for samples S3 and S13, indicating that these samples had the highest diversity of skin microbiota. The S28 sample contained the lowest species diversity with a value of 6.16 (Table 1). Ward's analysis demonstrated that two main clusters were present at the genus level ( Figure 4). The composition of the microbiota, however, did not cluster at the genus level. This might be attributable to differences in the composition of the fish skin microbiomes between individual fish from the same population and differences between the skin microbiome and the surrounding water [6].
Foodborne pathogens such as Salmonella, Escherichia, and Mycobacterium genera were not found in the samples. On the other hand, Vibrio (V.) ordalii was detected in three seabass swab samples (numbers 65, 66, and 67) originating from E-2 fish farms in Izmir. Similarly, many researchers have reported V. ordalii from seabass in the Aegean Sea [31,32], including Izmir [33]. Bacterial infections most frequently detected in cultured seabass and gilthead sea bream are caused by bacteria belonging to the family Vibrionaceae. Associated losses have been reported with Vibrionaceae in many fish species, including seabass, sea bream, and salmonid species etc. [34].

Temperature Measurement of Seabass Samples in Styrofoam Boxes Containing Ice
The lowest average temperature was 1.7 • C in the samples from the fish farm B2 located in Izmir, and the highest temperature was 3.4 • C in the samples from the fish farms A3 (Izmir) and E2 (Izmir) ( Table 2). The average and standard deviation of the inner temperature of seabass samples were 2.58 ± 0.53 • C. In addition, the internal temperature values measured in all fish samples were below +4 • C. Similarly, a study reported the internal temperature of iced styrofoam-packaged seabass from the Aegean Sea to be 4.15 ± 1.12 • C [35]. The extension of shelf life by chilling is essentially due to the reduction in the growth rate and metabolic activity of spoilage microorganisms such as Pseudomonas spp. [35] and Acinetobacter spp. Acinetobacter species have been found in great abundance in fresh seabass at 12 • C [19] and fish fillets at 10 • C [36], and were the dominant species at the end of the shelf life of rainbow trout stored aerobically at 4 • C [25]. Indeed, upon storage the psychrophilic bacteria proliferated slowly and dominated the mesophilic load, as the low temperature favored their growth [37]. Similar to our study, Syropoulou et al. [38] reported that Pseudomonas spp. were found from the beginning of shelf life, whilst in seabass products from Greece, Shewanella were detected at later storage stages.

Isolation of Psychrotrophic Pseudomonas spp. in Seabass Swab Samples using Conventional Methods
In total, 46 seabass swab samples (48%) were positive for psychrotrophic Pseudomonas strains isolated with the conventional ISO method [18] (Table 2). Pseudomonas strains were isolated from four fish farms in Izmir, i.e., A2 (n = 6), E2 (n = 6), E3 (n = 6), and A3 (n = 5), and farm C1 (n = 5) in Mugla. The cultivation-based method will detect live Pseudomonas strains, which is an important characteristic when compared to NGS and metabarcoding methods that are used in the detection of DNA fragments and DNA structures, as these do not necessarily indicate the presence of living bacteria [39].

Temperature Measurement of Seabass Samples in Styrofoam Boxes Containing Ice
The lowest average temperature was 1.7 °C in the samples from the fish farm B2 located in Izmir, and the highest temperature was 3.4 °C in the samples from the fish farms A3 (Izmir) and E2 (Izmir) ( Table 2). The average and standard deviation of the inner temperature of seabass samples were 2.58 ± 0.53 °C. In addition, the internal temperature val-

Antibiotic Groups
Name of Antibiotics Thirty (65.2%) Pseudomonas strains were susceptible to all antibiotics according to the CLSI [14]. On the other hand, thirty-three (71.7%) Pseudomonas strains were susceptible to all antibiotics according to the EUCAST [13]. All Pseudomonas strains from A1 (Izmir), B1 (Izmir), and E1 (Izmir) fish farms were susceptible to all antibiotics (Table 4).  Sixteen (34.8%) Pseudomonas strains were resistant to more than one antibiotic based on the CLSI [14]. Eight (17.4%) Pseudomonas strains were resistant to one antibiotic only, including carbapenem (doripenem) and aminoglycoside group (tobramycin). Six Pseudomonas strains were resistant to doripenem, and two strains were resistant to tobramycin based on the CLSI [14]. However, only five (10.9%) Pseudomonas strains were resistant to two antibiotics, according to the CLSI [14]. All Pseudomonas strains from fish farms in Izmir [(A2; n = 1) and (A3; n = 1)] and Mugla [(C1; n = 1) and (D1; n = 1)] were resistant to doripenem and imipenem (carbapenem group). In addition, one strain originating from D2 fish farms (Mugla) was found to be MDR to doripenem, imipenem, and meropenem, all included in the carbapenem group, based on the CLSI (Table 4).
Pseudomonas spp. have been identified as primarily invasive or opportunistic pathogens for many organisms and this genus has also grown in importance in terms of antimicrobial resistance [9]. Many researchers have evaluated the antimicrobial sensitivity of Pseudomonas species isolated from fish, and have reported them as MDR, based on their resistance to ampicillin, cefotaxime, aztreonam, trimethoprim-sulfamethoxazole, nitrofurantoin and other groups of antimicrobials [9,40]. Recently, Rezgui et al. [41] showed an abundance of antibiotic-resistant bacteria isolated from the gills and intestinal tract of seabass and sea bream. The antibiotic-resistant bacteria belong to several species of the genera Pseudomonas, Vibrio, Aeromonas, and Enterobacterales. They were resistant to tetracycline and penicillin, which are commonly used in treating infections in animals and humans. In another study, almost all Pseudomonas strains were resistant to penicillins (ampicillin), macrolides (erythromycin, clindamycin), sulfonamides (trimethoprim-sulphamethoxazole-), and chloramphenicol [9]. We report similar results, i.e., that the Pseudomonas strains were susceptible to penicillins (piperacillin-tazobactam), aminoglycosides (amikacin and gentamycin), fluoroquinolones (levofloxacin, norfloxacin), and tetracyclines (tetracycline, ciprofloxacin) based on the CLSI [14]. Likewise, a study reported that enrofloxacin, oxytetracycline, and ciprofloxacin were found to be effective antibiotics against fish disease agents such as Pseudomonas spp., Vibrio spp. and Staphylococcus spp. in Turkey [42]. On the other hand, all P. fluorescence strains isolated from fish were resistant to piperacillin, ceftazidime, and cefepime in Egypt [43]. In the present study, psychrotrophic Pseudomonas strains were partially resistant (based on the EUCAST and CLSI) to antibiotics commonly used in fish farms. This fact should be carefully addressed in the context of the environmental spread of antibiotic resistance.

MICs of Psychrotrophic Pseudomonas spp.
Pseudomonas strains that had shown resistance to antibiotics in the disc diffusion assay were selected for examination using the E-Test (gradient diffusion method) to determine their MIC (Table 5). From the 13 strains that showed resistance to doripenem in the disc diffusion test, two had an MIC exceeding the threshold ≥8 g/mL for antibiotic resistance (12 and 125 g/mL; the latter isolate originated from farm A3 inİzmir). For imipenem, three out of six isolates resistant according to disc-diffusion assay were confirmed as resistant by E-test. The MIC of these three resistant strains was >32 µg/mL. All these isolates originated from farms A3 (in Izmir), C1, and D1 (both in Mugla). Similarly, isolates resistant to tobramycin, meropenem, or ciprofloxacin according to the disc diffusion assay, were classified as susceptible based on the E-test MIC [13,14]. Only one Pseudomonas strain from C1 fish farms (Sample no. 24) was resistant to doripenem and imipenem, as assessed by MIC determination. The different results obtained by the gradient diffusion (E-test) and the disc diffusion methods for Enterobacterales and Pseudomonas aeruginosa strains are not unexpected since the E-test generally performs better [45]. Despite the different outcomes from different methods, our results are in line with reports on antimicrobial resistance in Pseudomonas and Escherichia coli in general. The European Antimicrobial Resistance Surveillance Network reported on samples from human patients in 2017, of which, 30.8% of the Pseudomonas aeruginosa strains isolated were resistant to at least one of the antimicrobial groups under regular surveillance (fluoroquinolones, aminoglycosides, and carbapenems) [46]. Moreover, the European Centre for Disease Prevention and Control has shown significant increments in the percentage of antibiotic-resistance among pathogenic bacteria, such as carbapenemresistance in Pseudomonas aeruginosa and Acinetobacter spp. in several countries in the European region of concern [47].
With respect to fish, a study from Egypt reported that Pseudomonas aeruginosa and E. coli strains were resistant to third-generation cephalosporin and last-resort carbapenems isolated from Nile tilapia [41]. Interestingly, 29.7% of P. fluorescens strains isolated showed MDR, especially to penicillin and cephalosporin groups [41].

Conclusions
Results from this study show that psychrotrophic Pseudomonas were the dominant bacterial species in seabass skin samples from 12 selected fish farms in the Aegean Sea.
Ninety-six fish were sampled by skin swab, and in all samples, NGS analysis indicated the presence of Pseudomonas. Viable isolates were cultured from 46 of these samples. Testing the isolates against 11 different antibiotics (five main groups), showed that all samples were susceptible to piperacillin-tazobactam, gentamicin, amikacin, levofloxacin, norfloxacin, and tetracycline. Based on the CLSI, the isolates from across the farms showed various resistance patterns to the carbapenem group [doripenem (28.3%), imipenem (13%), and meropenem (2.3%)] and aminoglycosides [tobramycin (4.3%)]. Using the EUCAST standard, there was additional resistance to doripenem (28.3%), imipenem (13%), meropenem (2.3%), and ciprofloxacin (2.2%). MDR was found among three Pseudomonas strains from Mugla (D = 2) based on the CLSI and five Pseudomonas strains based on the EUCAST criteria (disc diffusion method). Three farms with six isolates showed no antibiotic resistance based on EUCAST and CLSI criteria.
This study has shown that resistance to a broad range of antibiotics prevails in Pseudomonas from the selected farms. As the farms were chosen without looking at their histories of disease and antibiotic use, our results may indicate a representative situation for the industry in the region. This should, however, be confirmed in a broader study, including records of antibiotic use at the farm level.
The use of antibiotics is generally regarded as the main driver for developing resistance. Exposure to antibiotics may be due to own use or external exposure. The industry uses antibiotics for prophylactic and therapeutic treatments to keep farmed fish free of diseases. Prudent use of antibiotics is therefore essential also for the aquaculture industry to minimize antibiotic resistance and the spread of resistant bacteria or genes to the environment. Ultimately, this will serve consumer protection and lead to a more efficient application of antibiotics in human therapy.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.