Effects of Cage Farming on Antimicrobial and Heavy Metal Resistance of Escherichia coli,Enterococcus faecium,and Lactococcus garvieae

Abstract

Objective:

To characterize antibiotic resistance genes (ARGs) and heavy metal resistance genes (HMRGs) of Escherichia coli and Enterococcus faecium isolated from the sediment and Lactococcus garvieae isolated from fish.

Materials and Methods:

The isolated bacteria were identified by sequencing 16S rRNA genes. After identification of the bacteria, tetracycline (tetA, tetB, tetD), erythromycin (ereA, ereB), sulfonamides (sulI, sulII), trimethoprim (dhfrA1), β-lactam (bla_TEM, bla_CTX, ampC), florfenicol (floR), and class 1 integron (Int1) resistance gene were then determined. The presence of HMRGs, including copper (copA), mercury (mer), cadmium, zinc, cobalt (czc), and nickel, cobalt cadmium (ncc), was also analyzed by PCR. All strains were checked for the presence of ARGs and/or HMRGs on the plasmid.

Results:

The frequency of the β-lactam resistance gene was highest and ranged from 49.7% to 62.3%, followed by sulfonamides, tetracyclines, phenicols, and macrolide resistance genes. The cage culture fish farming practice showed significant effects on ARG frequency of bacteria isolated from the sediment, whereas it had no effect on the frequency of HMRGs. The most prevalent HMRG was determined as mercury-resistant mer gene in all bacteria. All four of the HMRGs were located on plasmids with frequency ranging from 1.20% to 32.53%. The presence of ARGs on plasmids ranged between 2.2% (Dhfr1) and 75% (AmpC, blactx, tetB), and plasmids did not contain tetD and ereB genes.

Conclusion:

The results of this study indicate that fish farming can significantly influence the antimicrobial resistance properties of bacteria isolated from sediment samples.

Introduction

Aquaculture production has substantially increased in Turkey using open net cages. The most extensively cultured fish species in cages are rainbow trout (Oncorhynchus mykiss), sea bream (Sparus aurata), and sea bass (Dicentrarchus labrax), and their total production is ∼138,876 tons per year.¹

The development of fish production brings about fish diseases that result in significant economic losses. Bacterial diseases are one of the major problems in aquaculture. During disease outbreak, fish are usually treated with antibiotics that can accumulate and release into the aquatic environment and might trigger the development of antimicrobial resistance (AMR) in bacteria. Consequently, after a major bacterial disease outbreak antibacterial therapy may fail due to AMR of the pathogen.²

In addition to antibiotics, industrial and agricultural discharges create heavy metals such as nickel, cobalt, cadmium, zinc, and copper leading to aquatic pollution. Moreover, different kinds of human activities such as the release of wastewater can contaminate the aquatic environment with heavy metals.³ Although some of the low concentrations of heavy metals such as zinc, nickel, cobalt, and copper are crucial for the bacterial metabolic process, high concentrations of them are toxic to aquatic organisms. In addition, some of them are generally used as bactericides and fungicides in aquaculture. However, some heavy metals such as chromium, lead, mercury, and cadmium are toxic to bacteria even at low concentrations.⁴ Heavy metal pollution triggers bacteria to change their genetic and physiological structure to survive.⁵ Consequently, the bacteria can develop heavy metal resistance genes (HMRGs) in response to these pollutants. The HMRG can be situated on the chromosome, as well as plasmids of bacteria.^5,6

In recent years, there has been an increasing awareness regarding the occurrence and potential transfer of antibiotic resistance genes (ARGs) to aquaculture (e.g., cage farming) increasing the risk for ARG transmission to the aquatic environment,^7,8,9 hospital,¹⁰ river and sediments,¹¹ and other environments.¹² Mobile or chromosomal genetic elements such as conjugative plasmids, integrons, and transposons can carry antibiotics and the metal resistance gene. Bacteria can develop different strategies against antibiotics and metals to survive, including drug inactivation (precipitation or alteration of the toxic metal), efflux pump, mutation of cellular targets, and reduction of membrane permeability of bacteria.^13,14,15 According to our knowledge, studies focused on ARGs and HMRGs in bacteria isolated from fish and sediment are insufficient, especially related to fish culture. Therefore, the aim of this study was to elaborate and compare (1) the phenotypic antibacterial susceptibility of Escherichia coli and Enterococcus faecium isolated from sediments underneath the fish cages and reference sediments, while Lactococcus garvieae was isolated from fish cultured in freshwater and sea cages; (2) the presence of ARGs and HMRGs in isolated bacteria; and (3) the presence of ARGs and HMRGs in plasmids found in isolated bacteria. Since L. garvieae is one of the most isolated bacteria from trout cultured in fresh water and marine water in Turkey,¹⁶ E. coli, Gram-negative, and E. faecium, Gram-positive, are essential indicators of fecal pollution¹⁷; these bacterial species were chosen in the present study.

Materials and Methods

Sampling and bacterial examination

Fish and the bottom sediments were sampled from freshwater and sea cages during fall and spring seasons. The freshwater cages were located in Samsun (a lake) and Gumushane (a dam), while the sea cages were located in Ordu and Trabzon, at the Black Sea. A total of 30 fish were randomly sampled from three different cages for each area. Similarly, the bottom sediments from every cage were sampled. The bottom sediments were obtained with Beckman grap (15 × 15 cm). The reference sediments were sampled 1 km away from the cage farming sites. In these farms, the most commonly used antibiotics were oxytetracycline, florfenicol, and enrofloxacin (Table 1).

Table 1.

Information About the Sampling Area

Sampling point	No. of farms	No. of cages	Production (ton/year)	Sampling depth (m)	Fish species	Antibiotic used in the farm
Samsun	13	300	3500	17	Rainbow trout	TC, ENR
Gumushane	11	220	826	60	Rainbow trout	FFC, TC, EM
Trabzon	4	140	838	45	Rainbow trout	EM
Ordu	3	250	2600	47	Sea bass	TC, EM, FFC

According to fish farmers and government officials, fish farmers are commonly using tetracycline (TC), florfenicol (FFC), enrofloxacin (ENR), and erythromycin (EM).

The liver, kidney, and spleen of sampled fish were aseptically streaked on Tryptic soy agar (Merck) and incubated at 25°C for 3 days. L. garvieae was isolated and identified as described by Ture et al.¹⁸ and also confirmed by rapid ID 32 Strep (Biomerieux).

Beckman grap (15 × 15 cm) was used to sample sediment underneath three cages from each station, and reference sediment samples were collected 1 km away from the fish culture sites. Five grams of the sediments were diluted with 45 mL Luria Bertani Broth (Merck) and incubated for 24 hr at 37°C. After incubation, 0.1 mL of the broth was transferred into Chromocult Coliform ES Agar (Merck) and Bile Esculin Azide Agar (Merck) for isolation of E. coli and E. faecium, respectively. Further identification of presumptive E. faecium and E. coli colonies was performed as described by Bagcigil et al.¹⁹ and Byamukama et al.,²⁰ respectively. The isolated bacteria were identified as described below by sequencing 16S rRNA genes.

Antimicrobial susceptibility tests

After the biochemical and genetic identification, all isolates were then tested for susceptibility to seven antibiotics commonly used in aquaculture. Minimum inhibitory concentration (MIC) levels for all antibiotics were determined using the Epsilometer procedure (E test, Biotests) on Mueller Hinton agar (Oxoid). In this study, the commercial antibiotic test strep (concentration range, μL/mL) used included: tetracycline (TC; 256-0.016), amoxicillin/clavulanic acid (AMC; 256-0.016), enrofloxacin (ENR; 256-0.016), florfenicol (FFC; 32-0.002), erythromycin (EM; 245-0.016), penicillin (PG; 256-0.016), and trimethoprim/sulfamethoxazole (TS; 32-0.002). The breakpoints used for all strains were as defined by the CLSI.²¹

Genomic DNA preparation and sequencing of bacteria

Genomic DNA of bacteria was extracted from pure culture with the QIAamp DNA mini kit (Qiagen). RNA/DNA calculator (QIAexpert) was used to measure the DNA concentration (average, A260/280 was 1.85, and DNA concentration was 40 ng/μL). The 16S rRNA gene specific universal 27F and 1392R primers were synthesized by Integrated DNA Technologies. PCR was performed with PCR master mix (Qiagen) in a thermal cycler (Applied Biosystems). PCR product was directly sequenced according to the protocol of BigDye v3.1 (Applied Biosystems) and read on an Applied 3500 genetic analyzer (Applied Biosystems). The derived nucleotide sequences were aligned and analyzed using BioEdit.²² Pairwise genetic distances among strains were calculated in MEGA 7.0 using the Kimura 2-parameter distance model.

PCR assays for detection of ARGs and HMRGs

The presence of ARGs, including tetracycline (tetA, tetB, tetD), erythromycin (ereA, ereB), sulfonamides (sulI, sulII), trimethoprim (dhfrA1), β-lactam (bla_TEM, bla_CTX, ampC), florfenicol (floR), and class 1 integron (Int1), was analyzed. In addition, the presence of HMRGs, including copper (copA), mercury (mer), cadmium, zinc, cobalt (czc), nickel, and cobalt cadmium (ncc), was analyzed by PCR. Each PCR mix (25 μL) contained 50 ng (1 μL) of sample DNA, 10 pmol (1 μL) of each primer (Table 2), 12.5 μL of 2× PCR master mix solution (Qiagen), and 9.5 μL of distilled water. As a negative control, no template DNA was used, while for the positive control known DNA of bacteria that contained different resistance genes was used.¹¹ DNA amplification was carried out in a thermocycler (Applied Biosystems) under the following conditions: after initial denaturation at 94°C for 5 min, denaturation at 94°C for 30 sec, annealing for 30 sec (see Table 2 for different annealing temperature), extension at 72°C for 45 sec for 35 cycles, and final extension at 72°C for 10 min. Ten microliters of each PCR product was subjected to electrophoresis in 1.5% (w/v) agarose gel prepared with 1× Tris-Boric acid-EDTA (TBE) buffer and run at 100 V for 1 hr. The PCR products were then visualized by UV transillumination. Three PCR products belong to different ARGs, and HMRGs were also confirmed by DNA sequence.

Table 2.

Primers Used in the PCR

Primer	Sequence (5′-3′)	Target gene	Fragment length (bp)	References	Annealing temperature (°C)
TetA-f	GCTACATCCTGCTTGCCTTC	tetA	210	Marshall et al.⁴¹	55
TetA-r	CATAGATCGCCGTGAAGAGG	tetA	210		55
TetB-f	TTGGTTAGGGGCAAGTTTTG	tetB	659		54
TetB-r	GTAATGGGCCAATAACACCG	tetB	659		54
TetD-f	AAACCATTACGGCATTCTGC	tetD	787		54
TetD-r	GACCGGATACACCATCCATC	tetD	787		54
ereA-f	AACACCCTGAACCCAAGGGACG	ereA	420	Ounissi and Courvalin⁴²	58
ereA-r	CTTCACATCCGGATTCGCTCGA	ereA	420	Ounissi and Courvalin⁴²	58
ereB-f	AGAAATGGAGGTTCATACTTACCA	ereB	546	Arthur et al.⁴³	52
ereB-r	CATATAATCATCACCAATGGCA	ereB	546	Arthur et al.⁴³	52
Sul1-f	CGGCGTGGGCTACCTGAACG	sul1	433	Kerrn et al.⁴⁴	60
Sul1-r	GCCGATCGCGTGAAGTTCCG	sul1	433		60
Sul2-f	GCGCTCAAGGCAGATGGCATT	suI2	293		60
Sul2-r	GCGTTTGATACCGGCACCCGT	suI2	293		60
dhfr1f	CTGATATTCCATGGAGTGCCA	dhfr1	433	Schmidt et al.⁴⁵	56
dhfr1r	CGTTGCTGCCACTTGTTAACC	dhfr1	433	Schmidt et al.⁴⁵	56
TEM-f	ATGAGTATTCAACATTTCCG	bla _TEM	859	Olesen et al.⁴⁶	45
TEM-r	CAATGCTTAATCAGTGAGG	bla _TEM	859	Olesen et al.⁴⁶	45
CTXm-f	GGTTAAAAAATCACTGCGTC	bla _CTX	863	Saladin et al.⁴⁷	50
CTXm-r	TTGGTGACGATTTTAGCCGC	bla _CTX	863	Saladin et al.⁴⁷	50
AmpC-f	TTCTATCAAMACTGGCARCC	ampC	550	Schwartz et al.⁴⁸	49
AmpC-r	CCYTTTTATGTACCCAYGA	ampC	550	Schwartz et al.⁴⁸	49
Flor-f	TATCTCCCTGTCGTTCCAG	floR	399	Van et al.⁴⁹	50
Flor-r	AGAACTCGCCGATCAATG	floR	399	Van et al.⁴⁹	50
Int-f	GCCACTGCGCCGTTACCACC	Int	898	Kerrn et al.⁴⁴	62
Int-r	GGCCGAGCAGATCCTGCACG	Int	898	Kerrn et al.⁴⁴	62
coprun-f	GGSASDTACTGGTRBCAC	copA	1200	Lejon et al.³¹	55
coprun-r	TGNGHCATCATSGTRTCRTT	copA	1200	Lejon et al.³¹	55
czcA-f	GGS GCG MTS GAY TTC GGC	czcA	252	Kaci et al.¹⁴	55
czcA-r	GCC ATY GGN YGG AAC AT	czcA	252	Kaci et al.¹⁴	55
ncc-f	GCGTGGAAGGCAAGATGTTC	ncc	457	Shoeb⁵⁰	54
ncc-r	ACGTCCACCAACGTTGGC	ncc	457	Shoeb⁵⁰	54
mer-1	GAGATCTAAAGCACGCTAAGGC	mer	1011	Misra⁵¹	57
mer-2	GGAATCTTGACTGTGATCGGG	mer	1011	Misra⁵¹	57

Presence of ARGs and HMRGs in plasmid

All strains were checked for the presence of ARGs and/or HMRGs on plasmid. The plasmids were extracted with QIAprep Mini Kit (Qiagen) following the manufacturer's instructions. This procedure was modified with lysozyme treatment for Gram-positive bacteria. Presence and size of plasmids in each sample were detected by electrophoresis in 0.9% agarose gels in 1× TBE with known plasmid weight standards (Supercoiled DNA Ladder; Invitrogen). Presence of ARGs and/or HMRGs on the plasmid was tested using gene specific primers as described above; however, plasmid was used as a template instead of bacterial DNA in the PCR.

Statistical analysis

Chi-square and Fisher's exact tests were used to analyze the data on the distributions of ARGs, HMRGs and ARGs and HMRGs carrying plasmids found in bacteria isolated from sediment underneath the cages and reference sediment using IBM SPSS Statistics version 22. The Pearson Correlation test was used to describe the correlation between ARGs and AMR of the isolates. Correlations between ARGs and HMRGs from different sediments were considered significant when p ≤ 0.05.

Results

Isolates from fish and sediment samples were identified to species level through 16S rRNA gene sequence analysis. L. garvieae, E. coli, and E. faecium were ≥99% similar to reference strains from GenBank (Accession nos: NC 017490.1, NC 000913.3, and NC 017960.1, respectively) (Fig. 1). L. garvieae (24/183) was isolated from fish, while E. coli (74/183) and E. faecium (85/183) were isolated from sediments located under the fish cages or reference sediment (Table 3). The season did not affect the number of isolated sediment bacteria, but isolation of L. garvieae on fish was significantly affected by season. A total of 19 strains were isolated in the fall, while 5 strains were isolated in the spring (Table 3).

FIG. 1.

Maximum likelihood tree based on mitochondrial 16S rRNA gene sequences of Lactococcus garvieae, Escherichia coli, and Enterococcus faecium strains. 16S rRNA gene sequences of L. garvieae, E. coli, and E. faecium strains were demonstrated to have ≥99% similarity with reference strains.

Table 3.

Bacteria Isolated from Sediment Under the Cages (A), Reference Sediment Taken from Far Away Fish Culture Sites (B), and Fish in Freshwater Cage (FC) and in Seawater Cage (SC)

		Sampling time		Sampling location
Bacterial species	n	Fall	Spring	Gumushane	Samsun	Trabzon	Ordu
Sediment
Escherichia coli	74	35	39	13(A)/10(B)	5(A)/5(B)	13(A)/10(B)	10(A)/8(B)
Enterococcus faecium	85	45	40	18(A)/14(B)	5(A)/5(B)	13(A)/9(B)	14(A)/7(B)
Fish pathogen
Lactococcus garvieae	24	19	5	15(FC)	ND	9(SC)	ND
Total	183	99	84	70	20	54	39

ND, L. garvieae was not isolated from Samsun and Ordu.

Thirty strains of E. coli were isolated from the reference sediment. None of them was resistant to ENR, TS, TC, EM, and FFC, but three of them were resistant to AMC, while all of them were resistant to PG. In contrast, 31 strains of E. faecium were isolated from the reference sediment, and some of these strains were resistant to AMC, PG, ENR, TS, and TC (Table 4). According to MIC test results, all of the E. coli strains were resistant to penicillin (MIC ≥256 μg/mL). Up to 13% of the E. faecium isolated from the sediment underneath the cages were resistant to FFC, while the rest of the bacteria isolated from sediment and fish were sensitive to FFC (MIC ≤8 μg/mL). Furthermore, 46.7% of the L. garvieae isolated from fish were resistant to TS, followed by PG (20%). The most effective antibiotics were ER and FFC for sediment borne bacteria and L. garvieae (Table 4). Bacteria isolated from sediment underneath the cages had a significantly higher antibacterial resistance level compared with bacteria isolated from the control station. A high correlation was found between AMR and antimicrobial resistant genes of bacteria (r ranged 0.82–0.93, p < 0.01) except florfenicol. No correlation was found between florfenicol resistance and florfenicol resistance gene (r = 0.12, p > 0.05).

Table 4.

Antimicrobial Resistance of the Bacteria (n = 183) Isolated from Sediment Under the Cages (A), Reference Sediment Taken from Far Away Fish Culture Sites (B), and Fish in Freshwater Cage (FC) and in Seawater Cage (SC)

			Antimicrobial category
			β-lactams		Aminoglycosides	Sulfonamides	Tetracyclines	Macrolide	Phenicols
Bacterial species	Sampling area	Isolates	AMC	PG	ENR	TS	TC	EM	FFC
E. coli	A/B	44/30	—/3	44/30	—/—	3/—	7/—	4/—	—/—
E. faecium	A/B	54/31	5/6	27/6	36/10	5/3	14/3	5/—	13/—
%			5.1/14.8	72.5/59.1	36.7/16.4	8.2/4.9	21.4/4.9	9.2/—	13.3/—
L. garvieae	FC/SC	15/9	—/—	3/—	—/—	7/4	—/—	—/—	—/—
%			—	20/—	—/—	46.7/44.4	—/—	—/—	—/—

AMC, amoxicillin/clavulanic acid; PG, penicillin; TS, trimethoprim/sulfamethoxazole.

A total of 13 different ARGs were determined in 183 isolates, and all of the bacteria contained at least one or more ARGs (Table 5). The majority of isolates (95.62%) had more than two ARGs (multiresistant bacteria). The most prevalent ARGs were determined as β-lactam related genes (bla_CTX-M, bla_TEM, and ampC), and the presence of bla_CTX-M, bla_TEM, and ampC gene ratios was determined as 62.84%, 62.29%, and 49.72%, respectively. Furthermore, in E. coli strains the number of ARGs was significantly higher compared with other bacterial strains. E. coli and E. faecium bacteria contained integron class I gene, while none of the L. garvieae contained it (Table 5).

Table 5.

Antibiotic Resistance Genes of Bacteria (n = 183) Isolated from Sediment Under the Cages (A), Reference Sediment Taken from Far Away Fish Culture (B), Fish in Freshwater Cage (FC), and Fish in Seawater Cage (SC) and Fish Pathogenic Bacteria

			ARGs
			Tetracyclines			β-lactams			Sulfonamides			Macrolide		Phenicols	Integron
Bacterial species	Sampling area	n	tetA	tetB	tetD	ampC	bla_TEM	Bla_CTX-M	sulI	sulII	dhfrI	ereA	ereB	floR	Class 1
E. coli	A/B	44/30	6/5	8/1	—/4	43/30	35/8	42/26	15/5	5/4	2/2	3/—	—/—	18/8	11/2
E. faecium	A/B	54/31	16/—	3/5	3/—	12/2	37/13	27/2	8/12	3/16	3/—	5/—	2/—	15/—	2/4
	%A/%B		22.6/8.2	11.2/9.8	3.1/6.6	56.1/52.5	73.5/34.4	68.4/45.9	23.5/27.9	8.2/32.9	5.1/3.3	8.2/—	2.1/—	33.7/13.1	13.3/9.8
	%AB		16.9	10.7	4.4	54.7	58.5	61.0	25.2	17.6	4.4	5.0	1.3	25.8	11.9
L. garvieae	FC/SC	15/9	—/—	11/9	—/—	4/—	12/9	10/7	1/—	8/—	—/—	—/—	—/—	—/—	—/—
	%FC/%SC		—/—	73.3/100	—/—	26.7/—	80/100	66.7/77.8	6.7/—	53.3/—	—/—	—/—	—/—	—/—	—/—
	%FC-SC		—	83.3	—	16.7	87.5	70.8	4.2	33.3	—	—	—	—	—
Total%			14.8	20.2	3.8	49.7	62.3	62.3	22.4	19.7	3.8	4.4	1.1	22.4	10.4

ARG, antibiotic resistance gene.

Bacteria isolated from sediment underneath the cages had ARGs. More than 55 strains contained at least one β-lactam resistant gene followed by sulfonamides, tetracyclines, phenicols, and macrolide-resistant genes. All ARGs, except macrolide-resistant ARGs, were also detected from the reference sediments, but the frequency was low compared to bacteria isolated from sediment under the cages (Table 5). L. garvieae isolated from rainbow trout held in freshwater cages had tetracycline (tetB, 73.3%), β-lactam (ampC 26.7%, bla_TEM 80%, bla_CTX-M 66.7%), and sulfonamide (sulI 6.7%, sulII 53.3%) resistant genes while it only had tetB (100%), bla_TEM (100%), and bla_CTX-M (77.8%) genes (Table 5).

All HMRGs were amplified in bacteria isolated from sediments under the cages and reference sediments. The percentage of HMRGs containing bacteria isolated from sediments was similar in both sampling locations. Fish culture strategy did not affect the HMRG frequency of the sediment borne bacteria. The most prevalent HMRGs were determined as the mercury resistant (mer) gene in all bacteria, whereas L. garvieae strains isolated from fish only harbor mer gene (Table 6).

Table 6.

Heavy Metal Resistance Genes of Bacteria Isolated from Sediment Under the Cages (A), Reference Sediment Taken from Far Away Fish Culture Sites (B), Fish in Freshwater Cage (FC), and Fish in Seawater Cage (SC)

			HMRGs
Bacterial species	Sampling area	n	copA	Mer	czc	ncc
E. coli	A/B	44/30	—/3	19/9	4/—	5/4
E. faecium	A/B	54/31	2/—	19/20	4/4	16/9
Total		88/61	2/3	38/29	8/4	21/13
%			2.3/4.9	43.2/47.5	9.1/6.6	23.9/21.3
L. garvieae	FC/SC	15/9	—/—	9/3	—/—	—/—
%			—/—	60/33.3	—/—	—/—

HMRG, heavy metal resistance gene.

About 60% (44/74) of E. coli, 38% (32/85) of E. faecium, and 29% (7/24) of L. garvieae had plasmids with sizes ranging from 4 to 20 kb. All four of the HMRGs were located on plasmids with different frequency ranging from 1.20% to 32.53% (Table 7). The presence of ARGs on plasmids ranged between 51.81% (ampC) and 2.41% (Dhfr1 and ereA), and plasmids did not contain tetD and ereB genes. Prevalence of ARGs and HMRGs located on the plasmids of bacteria isolated from sediments under the cages and from reference sediment was almost identical. In contrast, L. garvieae isolated from fish cultured in fresh water had more ARGs and HMRGs on their plasmids (Table 8).

Table 7.

Presence of Plasmid in Bacteria Isolated from Sediment Under the Cages (A), Reference Sediment Taken from Far Away Fish Culture Sites (B), Fish in Freshwater Cage (FC), and Fish in Seawater Cage (SC) and Fish Pathogenic Bacteria

Bacterial species	Sampling area	n	Presence of plasmid	Plasmid, %
E. coli	A/B	44/30	26/18	59.1/60.0
E. faecium	A/B	54/31	20/12	37.0/38.7
Total		98/61 (159)	46/30 (76)	46.9/49.2 (47.8)
L. garvieae	FC/SC	15/9	4/3	26.7/33.3
Total		24	7	29.2

Table 8.

Percentage of Antibiotic Resistance Genes and Heavy Metal Resistance Genes on Plasmids Isolated from Sediment Under the Cages (A, n = 46), Reference Sediment Taken from Far Away Fish Culture Sites (B, n = 30), Fish in Freshwater Cage (FC, n = 4), and Fish in Seawater Cage (SC, n = 3)

Genes	A	B	FC	SC
Sul1	19.6	20.0	25.0	33.3
Sul2	10.9	23.3	50.0	33.3
AmpC	52.2	53.3	75.0	—
Bla _Tem	39.1	30.0	50.0	33.3
Dhfr1	2.2	3.3	—	—
bla _ctx	45.7	36.7	75.0	33.3
ereA	4.3	—	—	—
tetA	13.0	3.3	25.0	—
tetB	15.2	10.0	75.0	33.3
floR	17.4	10.0	50.0	33.3
mer	30.4	30.0	75.0	33.3
copA	2.2	—	—	—
czc	6.5	6.7	25.0	—
ncc	13.0	10.0	25.0	—

Discussion

Antimicrobial compounds have been widely used to treat or prevent bacterial diseases. The therapeutic and prophylactic use of antimicrobial agents may lead to the acquirement of AMR and ARGs in the aquaculture environment.²³ ARGs can be transferred horizontally from bacteria to pathogenic bacteria from fish. Consequently, ARGs found in bacteria may restrict the efficiency of antibacterial treatment in fish diseases, thus causing a serious problem.^24,25 In this study, ARGs and HMRGs were present in bacteria isolated from fish and sediment samples. The majority of bacterial isolates investigated in this study had two or more ARGs. Some of the extended-spectrum β-lactamase enzyme members are CTX-M, which were detected in the present study. The bla_CTX-M_, bla_TEM, and ampC gene encoded β-lactams were defined more frequently than any other ARGs. These results are in line with the findings of previous studies.^8,11

Tetracycline-resistant genes identified in many Gram-positive and Gram-negative microorganisms are mainly found in multiresistant bacteria. The rapid spread of these resistant genes is due to the location of new genes on mobile elements. Oxytetracycline is the most commonly used antibiotic in fish farms (Table 1). In a previous study, Capkin et al.¹¹ reported that bacteria isolated from water and fish in the Black Sea region of Turkey had tetA, tetB, tetC, and tetD genes at low frequency. In the present study, among tet genes, tetA was the most prevalent (16.9%), followed by tetB (10.7%) and tetD (4.4%). The occurrence of tet genes except tetD gene was higher in sediment samples under the cages compared with reference sediment samples. Furthermore, 13% and 15.2% of the plasmids present in bacteria isolated from sediment under the cages contained tetA and tetB genes, respectively. The occurrence of ARG percentages was low in bacteria isolated from the reference sediment. Therefore, the presence of ARGs containing plasmids may play an important role for the frequency of ARGs.

Some antibiotics, including florfenicol, oxytetracycline, enrofloxacin, trimethoprim+sulfadiazine, amoxicillin, and oxolinic acid, are permitted for use in aquaculture in Turkey, as well as in some European countries.²⁶ Florfenicol is a relatively new antibiotic, which was introduced to aquaculture few years ago. It has been intensively used for the treatment of bacterial fish diseases in Turkey. Ture and Alp⁹ reported that the most effective antibiotic was florfenicol for the treatment of cultured fish. In the present study, a small number of E. faecium was resistant to florfenicol, while the majority of the other sediment and fish pathogenic bacteria were sensitive to florfenicol (MIC ≤8 μg/mL). However, floR ARG was detected in E. coli and E. faecium strains in both sediments under the cages and reference sediments, but a number of floR ARGs were significantly higher under cage sediment compared with the reference sediment. Although some of the strains have florR ARG, its expression is not enough to resist more than 8 μg/mL of florfenicol.

Erythromycin is especially effective against L. garvieae and E. faecium. The majority of E. faecium and L. garvieae strains were sensitive or intermediate to erythromycin (MIC ≤1 μL/mL). In addition, ereA and ereB genes were only detected in five and two E. faecium strains, respectively. There are different erythromycin resistance determinants, including Erm methylase, efflux pumps, and inactivating enzymes.²⁷ In the present study, presence of ereA and ereB genes responsible for drug inactivation²⁸ was found with low frequency (7/183).

Although copper is an essential nutrient for many organisms, higher concentrations of copper have toxic effects on cell metabolism.²⁹ In the presence of copper, bacteria can develop resistance to it to survive. Bacterial resistance to copper is related to the copA gene.³⁰ Abundance of the copA gene depends on physicochemical characteristics of the environment.³¹ In the present study, the copA gene was detected in a small number of E. coli and E. faecium. Copper could be found in the aquatic environments, especially at the bottom of cage sediments due to the application of CuSO₄ to control parasitic and bacterial fish diseases.

The environmental contaminants pose an excessive threat to bacterial populations in many areas. Studies have shown that bacteria have developed resistance in the presence of high metal concentration.³ The resistance of bacteria to mercury is common in many areas, including aquatic environments.³² Mercury resistance determinant, mer gene, is frequently found on plasmids, transposons, as well as chromosomes.³³ Under the heavy metal contaminated area, mercury (mer) and copper (pco) resistance genes were found in bacteria.³ In our study, mer was the most prevalent gene among HMRGs. High prevalence of this gene might be due to feed ingredients that contained mercury, which can be released by uneaten feed or feces and settle on sediment.³⁴ Thus, mercury resistance of bacteria can be triggered.

Bacteria can gain resistance against Co, Zn, and Cd by developing efflux system to remove Cd, Zn, and Co.³⁵ Schmidt and Schlegel³⁶ investigated the nickel-cobalt-cadmium resistance genes in Alcaligenes xylosoxidans and E. coli. They found ncc and nre genes in Alcaligenes eutrophus under the cadmium, cobalt, and nickel contaminated areas, while they did not detect ncc gene in E. coli strains. In the present study, ncc gene–encoded cadmium, -cobalt, and -nickel resistance was found in bacteria isolated from both sediments under the cages and reference sediment. In addition, czc gene was also present in both freshwater and saltwater sediment. However, the number of ncc and czc gene containing bacteria was significantly higher in sediment under the cages compared with reference sediment which may be due to the presence of heavy metals in fish feed.³⁴

Plasmids are extrachromosomal DNA molecules that can be found in both Gram-positive and Gram-negative bacteria. The plasmid encodes some traits such as ARGs, specific virulence genes, and HMRGs. In contrast, bacterial resistance is usually associated with the occurrence of plasmids.³⁷ In the present study, plasmid frequencies in E. coli, E. faecium, and L. garvieae strains were 59.45%, 37.64%, and 29.16%, respectively. Fish culture did not affect the presence of plasmids in bacteria. The presence of ARGs on plasmids ranged between 2.2% (Dhfr1 and ereA) and 75% (AmpC, blactx, tetB). Plasmids did not contain tetD and ereB genes. In all four HMRGs (copA, mer, czc, and ncc), the HMRGs were located on plasmids with different frequency ranging from 1.20% to 32.53%.

Although sediment samples were taken from the bottom of the cage and 1-km away from the cages as a reference, in both sampling areas E. coli and E. faecium were isolated. Forty-four strains of E. coli isolated from sediment underneath the cages and 30 strains isolated from reference sediment were resistant to penicillin, which means that they are naturally resistant to it. In contrast, 31 strains of E. faecium were isolated from reference sediment, and some of the strains were resistant to AMC, PG, ENR, TS, and TC. ARGs except eraA and eraB were present in all strains of bacteria. The presence of ARGs was higher compared with phenotypic resistance. In addition, some of the bacteria isolated from the reference sediment that was not related to fish culture had some phenotypic AMR and ARGs. These resistances may be stimulated by deep-sea discharges of wastewater. Coastal cities discharge their wastewater into 18 m depth of the Black Sea, and one sampling station was very close to the discharge point. This may be the reason why some of the strains isolated from the control point were resistant to antibiotics. Mere detection of AMR in aquaculture systems does not imply misuse of antimicrobials in aquaculture. Source attribution of AMR in aquaculture-associated bacteria is complex, and caution is needed in the interpretation of data. AMR may be naturally present in the environment, from other sectors or from aquaculture itself. Over time, AMR occurs naturally by genetic changes. However, this process can be accelerated by the misuse and overuse of antibiotics.

Gumushane samples were taken from a dam with 108-million cubic meter volume of water. Tens of fish cages are located on the dam in which wastewater of nearby cities has been discharged. This explains a higher number of bacteria isolated from sediment and the presence of L. garvieae ARGs in the dam compared to other sampling areas.

It is hard to find a direct link between the antibiotic resistance profile and antimicrobial usage. In South Korea, Vibrio parahaemolyticus (n = 71) isolated from oysters was resistant to ampicillin and vancomycin, while more than 50% of them were resistant to cephalothin (52.2%), rifampicin (50.7%), and streptomycin (50.7%),³⁸ but there was no linkage found on the use of these antibiotics in aquaculture. Furthermore, studies in Baltic Sea showed the presence of resistance genes encoding in several antibiotics, although these antibiotics (tetracyclines, chloramphenicol, etc.) are not used in these areas.²⁴ Vibrio vulnificus isolated in Dutch eel farms showed resistance to cefoxitin although this antibiotic was not used in eel aquaculture.³⁹ Antibiotic resistant marine bacteria have been found as far as 522 km offshore and as deep as 8,200 m.⁴⁰ Therefore, our results are consistent with previous studies and explain why some of the bacteria isolated from sediment resistant to β-lactam antibiotics, which are not used in fish farm or some of the bacteria isolated from control sediment, are AMR and carry ARGs.

Conclusion

The results of this study revealed phenotypic and genotypic antibiotic resistance and the presence of some HMRGs in sediment and fish pathogenic bacteria, providing insights into understanding resistance mechanisms of bacteria in aquatic environments. Bacteria isolated from sediment samples underneath the cages had more resistant strains compared with the reference sediment samples. The majority of bacterial isolates had two or more ARGs. AmpC, bla_TEM, bla_CTX-M, and mer genes were the most prevalent antibiotic, and HMRGs were determined in all bacteria. In addition, ARGs were also detected within plasmids indicating potential dissemination of the corresponding ARGs in the environment. As a result, fish farming significantly affects AMR of bacteria isolated in the sediment. Therefore, antibiotic usage should be limited, and other environment friendly alternative treatment or prevention, such as probiotic, should be used.

Footnotes

Acknowledgments

This project was funded by Republic of Turkey Ministry of Food, Agriculture and Livestock (Project no: TAGEM/HSGYDA/16/A11/P03/84). The authors thank Cemil Altuntas, Haci Savas, Ilyas Kutlu, and Murat Erbay for their help during fish and sediment sampling.

Disclosure Statement

No competing financial interests exist.

References

(IWG-SCC). 2015. Available at www.turkstat gov.tr/Start.do (accessed October 15, 2015).

Aoki

, and Takahashi

. 1987. Class-D tetracycline resistance determinants of R-plasmids from the fish pathogens aeromonas-hydrophila, edwardsiella-tarda, and pasteurella-piscicida. Antimicrob. Agents. Chemother., 31:1278–1280.

Trajanovska

, Britz

M.L.

, and Bhave

. 1997. Detection of heavy metal ion resistance genes in Gram-positive and Gram-negative bacteria isolated from a lead-contaminated site. Biodegradation, 8:113–124.

Nies

D.H.

2003. Efflux-mediated heavy metal resistance in prokaryotes. Fems. Microbiol. Rev., 27:313–339.

Wuertz

, and Mergeay

. 1997. The impact of heavy metals on soil microbial communities and their activities. In van Elsas

J.D.

, Wellington

E.M.H.

, and Trevors

J.T.

(eds.), Modern soil microbiology. Marcel Decker, New York, pp. 1–20.

Cervantes

, and Gutierrezcorona

. 1994. Copper resistance mechanisms in bacteria and fungi. Fems. Microbiol. Rev., 14:121–137.

Boran

, Terzi

, Altinok

, Capkin

, and Bascinar

. 2013. Bacterial diseases of cultured Mediterranean horse mackerel (Trachurus mediterraneus) in sea cages. Aquaculture, 396:8–13.

Capkin

, Ozdemir

, Ozturk

R.C.

, and Altinok

. 2017. Determination and transferability of plasmid-mediated antibiotic resistance genes of the bacteria isolated from rainbow trout. Aquac. Res., 48:5561–5575.

Ture

, and Alp

. 2016. Identification of bacterial pathogens and determination of their antibacterial resistance profiles in some cultured fish in Turkey. J. Vet. Res., 60:141–146.

10.

Wilberger

M.S.

, Anthony

K.E.

, Rose

, McClain

, and Bermudez

L.E.

. 2012. Beta-lactam antibiotic resistance among Enterobacter spp. isolated from infection in animals. Adv. Microbiol., 2:129–137.

11.

Capkin

, Terzi

, and Altinok

. 2015. Occurrence of antibiotic resistance genes in culturable bacteria isolated from Turkish trout farms and their local aquatic environment. Dis. Aquat. Organ., 114:127–137.

12.

Guglielmetti

, Korhonen

J.M.

, Heikkinen

, Morelli

, and von Wright

. 2009. Transfer of plasmid-mediated resistance to tetracycline in pathogenic bacteria from fish and aquaculture environments. Fems. Microbiol. Lett., 293:28–34.

13.

Baker-Austin

, Wright

M.S.

, Stepanauskas

, and McArthur

J.V.

. 2006. Co-selection of antibiotic and metal resistance. Trends Microbiol., 14:176–182.

14.

Kaci

, Petit

, Lesueur

, Boust

, Vrel

, and Berthe

. 2014. Distinct diversity of the czcA gene in two sedimentary horizons from a contaminated estuarine core. Environ. Sci. Pollut. Res. Int., 21:10787–10802.

15.

Levy

S.B.

, and Marshall

. 2004. Antibacterial resistance worldwide: causes, challenges and responses. Nat. Med., 10:S122–S129.

16.

Ozturk

R.C.

, and Altinok

. 2014. Bacterial and viral fish diseases in Turkey. Turk. J. Fisheries Aquat. Sci., 14:275–297.

17.

Lunestad

B.T.

, Frantzen

, Svanevik

C.S.

, Roiha

I.S.

, and Duinker

. 2016. Time trends in the prevalence of Escherichia coli and enterococci in bivalves harvested in Norway during 2007–2012. Food Control, 60:289–295.

18.

Ture

, Altinok

, and Capkin

. 2015. Comparison of pulsed-field gel electrophoresis and enterobacterial repetitive intergenic consensus PCR and biochemical tests to characterize Lactococcus garvieae. J. Fish Dis., 38:37–47.

19.

Funda Bagcigil

, Ikiz

, Ak

, and Yakut Ozgur

. 2015. Isolation of vancomycin resistant enterococci from animal faeces, detection of antimicrobial resistance profiles and vancomycin resistant genes. Kafkas Univ. Vet. Fak., 21:87–94.

20.

Byamukama

, Kansiime

, Mach

R.L.

, and Farnleitner

A.H.

. 2000. Determination of Escherichia coli contamination with chromocult coliform agar showed a high level of discrimination efficiency for differing fecal pollution levels in tropical waters of Kampala, Uganda. Appl. Environ. Microb., 66:864–868.

21.

Clinical and Laboratory Standards Institute.. 2014. Performance Standards for Antimicrobial Susceptibility Testing. Twenty-Fourth Informational Supplement. CLSI document M100-S24. Clinical and Laboratory Standards Institute, Wayne, PA.

22.

Hall

T.A.

1999. A user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp. Ser., 41:95–98.

23.

Cabello

F.C.

, Godfrey

H.P.

, Tomova

, Ivanova

, Dolz

, Millanao

, and Buschmann

A.H.

. 2013. Antimicrobial use in aquaculture re-examined: its relevance to antimicrobial resistance and to animal and human health. Environ. Microbiol., 15:1917–1942.

24.

Muziasari

W.I.

, Managaki

, Parnanen

, Karkman

, Lyra

, Tamminen

, Suzuki

, and Virta

. 2014. Sulphonamide and trimethoprim resistance genes persist in sediments at baltic sea aquaculture farms but are not detected in the surrounding environment. PLoS One, 9:e92702.

25.

Shah

S.Q.A.

, Colquhoun

D.J.

, Nikuli

H.L.

, and Sorum

. 2012. Prevalence of antibiotic resistance genes in the bacterial flora of integrated fish farming environments of Pakistan and Tanzania. Environ. Sci. Technol., 46:8672–8679.

26.

Baydan

, Yurdakok

, and Aydin

F.G.

. 2012. Antibiotic use in fish. Turkiye Klinikleri J. Vet. Sci., 3:45–52.

27.

Sutcliffe

, Grebe

, Tait-Kamradt

, and Wondrack

. 1996. Detection of erythromycin-resistant determinants by PCR. Antimicrob. Agents. Chemother., 40:2562–2566.

28.

Schmitz

F.J.

, Sadurski

, Kray

, Boos

, Geisel

, Kohrer

, Verhoef

, and Fluit

A.C.

. 2000. Prevalence of macrolide-resistance genes in Staphylococcus aureus and Enterococcus faecium isolates from 24 European university hospitals. J. Antimicrob. Chemother., 45:891–894.

29.

Brown

N.L.

, Rouch

D.A.

, and Lee

B.T.O.

. 1992. Copper resistance determinants in bacteria. Plasmid, 27:41–51.

30.

Voloudakis

A.E.

, Reignier

T.M.

, and Cooksey

D.A.

. 2005. Regulation of resistance to copper in Xanthomonas axonopodis pv. vesicatoria. Appl. Environ. Microb., 71:782–789.

31.

Lejon

D.P.H.

, Nowak

, Bouko

, Pascault

, Mougel

, Martins

J.M.F.

, and Ranjard

. 2007. Fingerprinting and diversity of bacterial copA genes in response to soil types, soil organic status and copper contamination. Fems. Microbiol. Ecol., 61:424–437.

32.

Nakamura

, and Silver

. 1994. Molecular analysis of mercury-resistant bacillus isolates from sediment of minamata bay, Japan. Appl. Environ. Microb., 60:4596–4599.

33.

Bogdanova

E.S.

, Mindlin

S.Z.

, Kalyaeva

E.S.

, and Nikiforov

V.G.

. 1988. The diversity of mercury reductases among mercury-resistant bacteria. Febs. Lett., 234:280–282.

34.

Liang

, Shao

D.D.

, Wu

S.C.

, Shi

J.B.

, Sun

X.L.

, Wu

F.Y.

, Lo

S.C.L.

, Wang

W.X.

, and Wong

M.H.

. 2011. The influence of mariculture on mercury distribution in sediments and fish around Hong Kong and adjacent mainland China waters. Chemosphere, 82:1038–1043.

35.

Trevors

J.T.

, Stratton

G.W.

, and Gadd

G.M.

. 1986. Cadmium transport, resistance, and toxicity in bacteria, algae, and fungi. Can. J. Microbiol., 32:447–464.

36.

Schmidt

, and Schlegel

H.G.

. 1994. Combined nickel-cobalt-cadmium resistance encoded by the ncc locus of Alcaligenes xylosoxidans 31a. J. Bacteriol., 176:7045–7054.

37.

Actis

L.A.

, Tolmasky

M.E.

, and Crosa

J.H.

. 1999. Bacterial plasmids: replication of extrachromosomal genetic elements encoding resistance to antimicrobial compounds. Front. Biosci., 3:43–62.

38.

Kang

C.H.

, Shin

, Kim

, Song

, Oh

E.G.

, Kim

, Yu

, and So

J.S.

. 2016. Prevalence and antimicrobial susceptibility of Vibrio parahaemolyticus isolated from oysters in Korea. Environ. Sci. Pollut. Res. Int., 23:918–926.

39.

Haenen

O.L.M.

, van Zanten

, Jansen

, Roozenburg

, Engelsma

M.Y.

, Dijkstra

, Boers

S.A.

, Voorbergen-Laarman

, and A.V.M. Moller. 2014. Vibrio vulnificus outbreaks in Dutch eel farms since 1996: strain diversity and impact. Dis. Aquat. Organ., 108:201–209.

40.

Aminov

R.I.

2011. Horizontal gene exchange in environmental microbiota. Front. Microbiol., 2:158.

41.

Marshall

, Tachibana

, and Levy

S.B.

. 1983. Frequency of tetracycline resistance determinant classes among lactose-fermenting coliforms. Antimicrob. Agents Chemother., 24:835–840.

42.

Ounissi

, and Courvalin

. 1985. Nucleotide-sequence of the gene erea encoding the erythromycin esterase in Escherichia coli. Gene, 35:271–278.

43.

Arthur

, Autissier

, and Courvalin

. 1986. Analysis of the nucleotide-sequence of the ereb gene encoding the erythromycin esterase type-ii. Nucleic Acids Res., 14:4987–4999.

44.

Kerrn

M.B.

, Klemmensen

, Frimodt-Moller

, and Espersen

. 2002. Susceptibility of Danish Escherichia coli strains isolated from urinary tract infections and bacteraemia, and distribution of sul genes conferring sulphonamide resistance. J. Antimicrob. chemother., 50:513–516.

45.

Schmidt

A.S.

, Bruun

M.S.

, Larsen

J.L.

, and Dalsgaard

. 2001. Characterization of class 1 integrons associated with R-plasmids in clinical Aeromonas salmonicida isolates from various geographical areas. J. Antimicrob. Chemother., 47:735–743.

46.

Olesen

, Hasman

, and Aarestrup

F.M.

. 2004. Prevalence of beta-lactamases among ampicillin-resistant Escherichia coli and Salmonella isolated from food animals in Denmark. Microbial. Drug Resist., 10:334–340.

47.

Saladin

, Cao

V.T.B.

, Lambert

, Donay

J.L.

, Herrmann

J.L.

, Ould-Hocine

, Verdet

, Delisle

, Philippon

, and Arlet

. 2002. Diversity of CTX-M beta-lactamases and their promoter regions from Enterobacteriaceae isolated in three Parisian hospitals. Fems. Microbiol. Lett., 209:161–168.

48.

Schwartz

, Kohnen

, Jansen

, and Obst

. 2003. Detection of antibiotic-resistant bacteria and their resistance genes in wastewater, surface water, and drinking water biofilms. Fems. Microbiol. Ecol., 43:325–335.

49.

Van

T.T.H.

, Chin

, Chapman

, Tran

L.T.

, and Coloe

P.J.

. 2008. Safety of raw meat and shellfish in Vietnam: an analysis of Escherichia coli isolations for antibiotic resistance and virulence genes. Int. J. Food Microbiol., 124:217–223.

50.

Shoeb

2006. Genetic basis of heavy metal tolerance in bacteria. PhD dissertation; Centre for Molecular Genetics, University of, Karachi, Pakistan.

51.

Misra

T.K.

1992. Bacterial resistances to inorganic mercury salts and organomercurials. Plasmid, 27:4–16.