The Science of the Total Environment, 114 (1992) 25-36 Elsevier Science Publishers B.V., Amsterdam

25

Long-range changes in oxytetracycline concentration and bacterial resistance towards

oxytetracycline in a fish farm sediment after medication

Ole Bent Samuelsen a, Vigdis Torsvik b and Arne Ervik c

"Department of Pharmacology and Toxicology', University of Bergen, Amauer Hanssens Hus, 5021 Bergen, Norway

hDepartment of Microbiology and Plant Physiology, University of Bergen, Jahnebakken 5, 5007 Bergen, Norway

"Institute of Marine Research, Department of Aquaculture, Strandgt. 229, 5024 Bergen, Norway

(Received October 5th, 1990; accepted December 10th, 1990)

ABSTRACT

Following 10 days medication with oxytetracycline, marine sediment was sampled beneath three selected cages (cages 1, 2 and 3) at a fish farm over a period of 18 mnd., in order to detect any change in the sediment oxytetracycline concentration, bacterial number and bacte- rial resistance towards the drug. The bulk of oxytetracycline disappeared during the first weeks, but it persisted in the sediment at lower concentrations for quite some time after the medication. Half-life (tu2) of oxytetracycline in the sediment was measured as: 125, 144 and 87 days under cages 1, 2 and 3, respectively. At the end of the medication, all three sediments had > 100% oxytetracycline-resistant bacteria. This value dropped to 20% after 72 days and stabilised at levels of between 10 and 50%. The change in bacterial numbers, described as total and plate counts, was due to seasonal variations rather than to the medication.

Key words." fishfarm; sediment; oxytetracycline; resistance

INTRODUCTION

Oxytetracycline is a commonly used antibiotic in Norwegian fish farming. It is administered in the form of medicated food-pellets and provided water currents are moderate, some of the antibiotic will inevitably sediment under the farm. Jacobsen and-Berglind (1988) found detectable concentrations of oxytetracycline in a fish farm sediment 3 months following feeding of the antibiotic. Controlled experiments using aquaria have recently been carried out to determine the half-life (tl/2) of oxytetracycline in fish farm sediments,

0048-9697/92/$05.00 © 1992 Elsevier Science Publishers B.V. All rights reserved

26 O,B. SAMUELSEN ET AL,

(Jacobsen and Berglind, 1988; Samuelsen, 1989). Depending on experimental set-up and run-time, tl/2 of oxytetracycline was determined to approximately 10 weeks (Jacobsen and Berglind, 1988) and 64 days (Samuelsen, 1989). The presence of oxytetracycline in the sediment for such an extended period may increase the chance of the bacteria present in the sediment developing resis- tance to oxytetracycline since tetracyclines are known to induce resistance to antibiotics in the aquatic microflora and bacteria pathogenic to fish (Toran- zo et al., 1984; Austin, 1985; Schlotfeldt et al., 1985; Jones et al., 1986). The resistance factor can also be transferred to potential human pathogens like Escherichia coli (Toranzo et al., 1984). Torsvik et al. (1988) found the percent oxytetracycline resistant bacteria in former medicated fish farm sediments to be 19-38%. In control sediments, not influenced by fish farming, the percent resistance were always < 1. Aquaria experiments do not always perfectly mimic the conditions found in natural systems. Therefore, the present study was undertaken in order to obtain information on the occurrence and persis- tence of oxytetracycline and oxytetracycline-resistant bacteria in the bottom sediment under a salmon farm. The effects of medication on the bacterial number in the sediment was also studied.

MATERIALS AND METHODS

A salmon farm outside Bergen on the West coast of Norway, with a weak water-flow for waste transport and hence considerable sedimentation under the cages, was treated with oxytetracycline for 10 days (23 March-Apri l 2). The dosage was 75 mg kg q fish -1 day -~. The sediment under three cages were selected for sampling; the cages contained approximately 1500 kg fish (cage 1 and 2) and 1150 kg (cage 3). The salmon in cage 1 took part in a vac- cination program against 'vibriosis and cold water vibriosis' directed by the Institute of Marine Research (Bergen, Norway) and was therefore not given the antibiotic. At day 23 the fish in cage 3 were removed. Therefore, no feeding took place and further sedimentation of excess foodpellets and faeces under this cage was minimized. At day 51 another 5 days of medication with oxytetracycline took place in cage 2, using the same dosage as described previously.

Sampling

Sediment samples, six cores from under each of the three selected cages were taken by divers. The plexiglass coring tube used was 4.5 cm i.d. and was 25 cm long. The tube was forced into the sediment and a rubber stopper was placed on the top of the tube to prevent the sample from falling out on with- drawal from the sediment. The lower end of the coring tube was sealed with

OXYTETRACYCLINE IN FISH FARM SEDIMENT 27

a rubber stopper immediately after withdrawal from the sediment. One meter long white sticks placed in the sediment as markers (at day 12), made sure that the sampling was taking place within the same area of the sediment each time. The sediment samples used for the bacterial analysis were kept cold during transport and analysed immediately after arrival at the laboratory while the sediment samples used for drug residue analysis were kept at -30°C until analysed.

The sedimentation, (particulate organic material) was measured by placing plexiglass coring tubes in the center of cage 1 (I), mid position (II) and at the edge of the cage (III).

Sample preparation

For the measurement of oxytetracycline residues in the sediment, the core tube was gently heated for 1 min. with a current of warm air before the frozen sediment and superficial liquid was pushed out with a piston. Slices of 30 g, (wet weight), corresponding to 1.5-2 cm of the sediment core, were homogenized by vigorous shaking for 5 min. The homogenate (1 g) was transferred to a polypropylene tube (8 ml) with a screwcap and extracted three times with 4, 4 and 2 ml of 0.1 M Na2EDTA (Triplex III) in McI1- vaine buffer (pH 4.0). (Oka et al., 1985). After centrifugation at 2500 x g for 10 min at 4°C in a Sorwall RC-5B refrigerated superspeed centrifuge (Du Pont Instruments, Newtown, CT, USA), 1 ml of the combined supernatants was transferred to a 1.5 ml polypropylene microcentrifuge test tube, cen- trifuged at - 14930 × g for 2 min in a Beckman Microfuge B (Beckman In- struments, Palo Alto, CA, USA) and was ready for analysis. The superficial liquid was weighed and a 1 ml portion was pipetted, centrifuged (14930 x g) and analysed on HPLC. The amount injected into the HPLC for both sedi- ment and seawater samples was 20/zl. For bacterial analysis samples were taken from the top 0-5 cm of the sediment. Ten grams of sediment were homogenized with 90 ml ice-cold Rila Marine solution (Rila Products, Teaneck, N J, USA) in a Waring blender for 1 min at low speed. From this suspension serial tenfold dilutions were made in ice-cold Rila Marine solution.

Analysis

Acetonitrile (HPLC-grade), disodiumhydrogen-phosphate monohydrate, sodium chloride, potassium nitrate, nitric acid, citric acid monohydrate, Kieselgel G (type 60) and Na2EDTA-dihydrate (Triplex III) were all from Merck (Darmstadt, Germany). Oxytetracycline hydrochloride (84% pure) was obtained from Norsk Medisinaldepot (Bergen, Norway).

28 o.a , SAMUELSEN ET AL.

For the quantitative high-performance liquid chromatographic (HPLC) analysis of oxytetracycline in the sediment and superficial liquid, a Spectra- Physics SP-8800 ternary HPLC-pump connected to a Spectra-Physics SP-8780 XR Autosampler was used. The analytical column, 100 x 5 mm, was packed in our laboratory using a Shandon packing machine with ODS- Hypersil 3 ~m particle size packing material for HPLC (Shandon Southern Products, Cheshire, UK), and the guard column with ODS-Hypersil 40/~m pellicular material. The UV-detector was a Spectra-Physics model SP-8480 XR Scanning detector, and the integrator a model SP-4270, also from Spectra-Physics. The mobile phase used was acetonitrile:water (14:86, v/v) containing 1 x 10 -3 M Na2-EDTA, 0.1 M potassium nitrate and diluted nitric acid giving pH 3.2 (Knox and Jurand, 1979). The buffer (860 ml) and acetonitrile (140 ml) were mixed, filtered through a HA 0.45/~m Millipore filter and degassed with a stream of helium, (1 1 min -~) for 2 min. A small stream of He, (5 ml min -~) was passed through the mobile phase throughout the analyses. The detection was made at 365 nm, and the flow-rate was 1.5 ml min -1.

A set of standard curves was prepared using seawater to which oxy- tetracycline was added in amounts corresponding to 0.02, 0.1 and 1/~g ml -~, and 1 g sediment homogenate to which oxytetracycline ranging from 10 to 400/~g g-~ was added. By using a standard method with respect to sample preparation and injected sample volume, the peak areas from the chromatograms from the sediment and seawater samples could be directly compared with those from the standard curves to give the value of the amount of oxytetracycline present in the superficial liquid and sediment samples.

Plate counts of bacteria were determined on Marine Agar (MA; 37.4 g Marine Broth (Difco), 1.5% Difco Agar, 1000 ml distilled water) and on Tryptone Soya Citrate Agar (TSCA; 20 g Tryptone Soya Broth (Difco), 10 g Na-citrate, 1.5% Difco Agar, 1000 ml aged 70% seawater). The number of drug resistant bacteria was determined on TSCA medium with 25 #g ml -~ oxytetracycline added. The percent resistant bacteria were determined from the numbers of bacteria growing on TSCA. Samples (0.1 ml) of appropriate dilutions of the sediments were spread on the agar plates. The plates were incubated aerobically at 15°C and counted after 7 days.

Total counts were determined by epifluorescence microscopy. One- millilitre samples of a suitable dilution (normally I:1000) of the sediment were filtered on 0.2 ~m Unipore filters and stained with DAPI (4,6-diamidino-2-phenylindole) using the technique of Porter and Feig (1980). The bacteria were counted in a Leitz Orthoplane microscope with a Ploemopak A filter.

O X Y T E T R A C Y C L I N E I N F I S H F A R M S E D I M E N T 29

z ~a

0

E

150

100

50

a - - - - CAGE 1

100 200 300 400 500

DAYS

600

F ig . 1. Persistence of oxytetracycline in the sediment under cage l, presenting the calculated S.D. values as vertical bars.

R E S U L T S

The sediment was black coloured, smelled strongly of hydrogen sulphide and was partly covered with a white carpet of Beggiatoa. The oxytetracycline standard curves had a high,xlegree oflinearity (R = 1.0 for both the sediment and seawater curves). The mean recovery of oxytetracycline from the sedi- ment and seawater was 98.4 ± 2.1 and 99.5 :t: 1.8%, respectively. The persis- tence of oxytetracycline in the sediments under the three cages are shown in Figs. l, 2 and 3. The sedimentation, calculated tw2 and the vertical distribu- tion o f oxytetracycline in the sediment at day 12 and 245 are shown in Tables 1, 2 and 3, respectively.

500 -

z .a 400

300 E-- Lxa

200

O E 100

0

; CAGE2

0 100 200 300 400 500

DAYS

600

Fig . 2. Persistence of oxytetracycline in the sediment under cage 2, presenting the calculated S.D. values as vertical bars.

30 O.13. SAMUELSEN ET AL.

~a 300 z

200

100

-- CAGE 3

!

0 500 600 I I I I

1 O0 200 300 400

DAYS

Fig. 3. Persistence of oxytetracycline in the sediment under cage 3, presenting the calculated S.D. values as vertical bars.

The total number of bacteria ranged from 1 × 10 9 to 1.2 x 10 ~° g-~ (Fig. 4), and the plate count from 3 x 105 to 4 x 10 7 g-1 sediment (Fig. 5). All the three sediments showed seasonal variations in plate counts and sediment from cage 1 and 2 also showed some seasonal variations in total counts, with the highest numbers in the spring (day 12 and 373). Figure 6 illustrates the percent of bacteria resistant to oxytetracycline in the sediment at different time intervals after the medication. At the end of the medication all the sediments had more than 100% oxytetracycline-resistant bacteria compared to the control numbers. The percent resistant bacteria dropped to 20 and 10% after 72 and 150 days, and then the values stabilized more or less at levels between 10 and 50%.

T A B L E 1

Sedimentation under cage 1, measured in gram particulate organic material per m 2 per day (g m -2 d -1)

Sampling period Sampling site

I II III

14 A u g . - 1 7 Oct . 99.1 42.6 53.2 17 O c t . - 1 6 Nov . - - 155.9 58.0

16 N o v . - 4 Jan . - - 43.2 15.9

- - , no measurement; Aug . , A u g u s t ; Oct., October; Nov., November; Jan., January.

OXYTETRACYCLINE IN FISH FARM SEDIMENT 31

TABLE 2

Half-life (fi/2) in days of oxytetracycline in the sediment under three cages at a salmon farm. The fi/2 calculations are based on measurements for 496 days (cages 1 and 2) and 556 days (cage 3)

Cage 1 2 3

tl/2 125 142 89 R-values 0.90 0.96 0.98

DISCUSSION

T h e s e d i m e n t is a n o x i c f r o m t o p to b o t t o m . A s l igh t e n r i c h m e n t o f t he

s u p e r f i c i a l l i q u i d w i t h d i s s o l v e d o x y t e t r a c y c l i n e p r o b a b l y o c c u r r e d as a r e su l t

o f s a m p l i n g a n d t r a n s p o r t o f t he c o r e s to t he l a b o r a t o r y . T h e r e f o r e i t was

n e c e s s a r y a l s o to a n a l y s e t h e s u p e r f i c i a l l i q u i d f o r o x y t e t r a c y c l i n e in o r d e r

to f i n d t h e c o r r e c t c o n c e n t r a t i o n o f t he d r u g in t he s e d i m e n t s a m p l e s . T h e

S .D . c a l c u l a t i o n s , s h o w n as v e r t i c a l b a r s in F i g s 1, 2 a n d 3 a r e b a s e d o n 4

m e a s u r e m e n t s a n d i n d i c a t e t h a t t h e a n t i b i o t i c , e s p e c i a l l y in c a g e 2 a n d 3

TABLE 3

Percentage vertical distribution of oxytetracycline in three sediments under a fish farm at day 12 and 245 following medication

Day Slice Cage

1 2 3

12

245

0-2 cm 53 38 54 2 -4 cm 23 20 3 4 -6 cm 19 18 10 6-8 cm 5 16 1 8-10 cm - - 8 - -

10-12 cm - - - - - -

0 -2 cm 7 5 27 2 -4 cm 31 13 37 4 -6 cm 46 27 19 6-8 cm 16 32 15 8-10 cm 2 22 3

10-12 cm - - 1 - -

Each slice was 30 g (wet weight), corresponding to approximately 2 cm of the core; - - , no drug residue found.

32 O,B. S A M U E L S E N ET AL.

1,20e+10-

1,00e+10 - .

< 8,00e+9-

0 6,00e+9 - ~ .

m 4,00e+9 -

Z 2,00e+9 -

i i I i i

100 200 300 400 500

DAYS

O,OOe+O 0 600

t:l----.- CAGE 1 • C A G E 2 = CAGE 3

F ig . 4. Total microscopic counts at different times after antibiotic treatment. All numbers are per g wet sediment.

shortly after the medication, is not evenly distributed in the sediment. This patchiness indicates a need to take several cores at each sampling site in order to provide a representative measurement.

Samuelsen (1989) showed, in experiments using aquaria, that the persis- tence of oxytetracycline in a fish farm sediment was highly dependent on the sedimentation after medication. Based on measurements for 224 days and with the oxytetracycline containing layer (crushed medicated pellets) placed on top of the sediment, tl/2 was calculated to 72 days. When the oxytetracycline-containing layer immediately was covered with an additional

ga

,<

O

ua o~

Z 2! 0 100 200 300 400 500 600

DAYS

C A G E 1 C A G E 2 C A G E 3

F ig . 5. Plate counts (the numbers of aerobic/facultative anaerobic bacteria), determined o n

M A medium at different times after antibiotic treatment. All numbers are per g wet sediment, and were determined after 7 days incubation a t 15°C.

O X Y T E T R A C Y C L I N E I N F I S H F A R M S E D I M E N T 33

ee~

b,

trJ

200

150 '

100 -

50

or---- CAGE 1 a CAGE 2

i i ' i i i

0 100 200 300 400 500

DAYS

600

Fig. 6. Percent oxytetracycline resistant bacteria in the sediment at different times after an- tibiotic treatment. The percent resistant bacteria were determined from the numbers of bacte- ria growing on TSCA + oxytetracycline relative to the numbers of bacteria growing on TSCA.

4 cm of sediment, the calculated t,/2 was nearly doubled (135 days). Table 3, the vertical distribution of oxytetracycline in the sediment, shows that the oxy- tetracycline first disappears in the upper layer of the sediment. No reports concerning possible degradation processes of oxytetracycline in marine sediments have yet been published. It is therefore likely that since oxytetracycline-hydrochloride, the form of the drug used in aquaculture, is very soluble in seawater, a solvation and diffusion process is the most likely mechanism whereby oxytetracycline escapes from the sediment. A gradually covering of the oxytetracycline-containing layer with additional faeces and food residues inevitably will result in a reduced contact with the overlying seawater and a decreased solvation and diffusion process occurs. At the salmon farm the sedimentation in cages 1 and 2 was larger than in cage 3 due to removal of the fish. Therefore, considering the sedimentation both in the aquaria and the fish farm, Table 1 shows that the tl/2 calculated for cages I, 2 and 3 (125, 144 and 87 days, respectively) are close to the values calculated in the corresponding aquaria (135 and 72 days, respectively). Both Figs 2 and 3 and former exper!ments (Samuelsen, 1989) show that the bulk of oxy- tetracycline disappears from the sediment during the first few weeks, but a rest concentration persists in the sediment for months after medication. However, decreased diffusion due to sedimentation may not be the only reason for oxytetracycline to persist for such an extended period in the sedi- ment. Concentrated aqueous solutions, including seawater, of oxytetracycl- ine-hydrochloride hydrolyse on standing at neutral pH and deposit insoluble crystals of oxytetracycline (The Merck Index, 1983). It is also known that

34 O.B. SAMUELSEN El" AL.

oxytetracycline easily forms complexes, for example with cations such as Ca 2+ and Mg 2÷ (Lunestad and Goks0yr, 1990). It is therefore possible that at least some of the oxytetracycline in the sediment is bound to cations or is present as crystals. Solubility experiments (Samuelsen, unpublished results) and the high recovery of oxytetracycline from the sediments indicate that the buffer used for extraction dissolves both possible crystals and oxytetracycline-cation complexes present in the sediment. Therefore, the results shown in Figs 1, 2 and 3 are most likely a measure of the total ox- ytetracycline residue in the sediment. Ca 2÷ and Mg 2÷ oxytetracycline com- plexes are shown less potent than the free form (Lunestad and Goks0yr, 1990). The oxytetracycline present in the sediment may, therefore to an extent be in an antibacterial inactive form.

The plate counts were determined on MA medium as this medium gave the highest number of colonies. MA medium could not be used for testing oxytetracycline resistance as the antibacterial effect of oxytetracycline was suppressed in this medium due to the high concentrations of magnesium and calcium. It is recommended that media with 70% seawater should be used as a standard for testing antibiotic resistance of marine bacteria (Lunestad and GoksCb'r, 1990). Therefore TSCA was used to determine the relative numbers of oxytetracycline-resistant bacteria.

The plate counts represent the numbers of aerobic/facultative anaerobic colony forming bacteria, and therefore only a minor part of the bacterial flora. However, this group of bacteria may be of special interest as it includes most of the fish pathogenic bacteria and bacteria which are primary decom- posers of organic wastes in the sediments. Our investigation shows that changes in bacterial numbers (total counts and plate counts) due to seasonal variations were more pronounced than changes due to medication.

The overall variations in percent oxytetracycline resistance were quite similar in the three sediments. The sediment under cage 1 which had not been medicated did not differ significantly from the other sediments. Cages 1 and 2 were located close to each other and the sediment bacteria under the two cages seemed to be influenced almost identically by the medication. The ex- tremely high percent of oxytetracycline-resistant bacteria (100-165%) at the end of the medication may reflect an adaptation to oxytetracycline due to repeated treatments with this antibiotic. Addition of oxytetracycline during medication might inhibit part of the bacterial flora and stimulate some of the bacteria which has already developed oxytetracycline resistance due to previous exposure to the drug. It has been observed that the growth of a drug resistant strain of Vibrio anguillarum can be stimulated in the presence of low drug concentrations (Helland, 1988; Sandaa 1990).

The percent of resistant bacteria decreased during the first 5 months, but then it levelled off and even increased slightly. Torsvik et al. (1988) found

OXYTETRACYCLINE IN FISH FARM SEDIMENT 35

that the percent of oxytetracycline-resistant bacteria in fish farm sediments were about 19-38%. In control sediments not influenced by fish farming, the percent resistant bacteria were always < 1%. In the investigated sediments the percent of resistant bacteria were maintained at level which varied be- tween 10 and 50% for at least 18 months. The resistance may be maintained in the presence of traces of oxytetracycline left in the sediments. However, the variation and increase in resistance which was observed during this long term investigation may also indicate that antibiotic resistance can be pro- moted and maintained due to other factors than the drug concentration. In the fish farm sediments toxic end products from anaerobic decomposition may be accumulated (Samuelsen et al., 1988). Baya et al. (1986) found a strong correlation between water pollution and resistance to toxic chemicals and antibiotics.

ACKNOWLEDGEMENTS

The technical assistance of Mrs. Eli Tepstad is highly appreciated. Thanks also to The Norwegian Research Council for Fisheries Science and Statens Forurensingstilsyn for financial support. A special thanks to the owners and workers at the fish farm for valuable information and help during sampling.

REFERENCES

Austin, B., 1985. Antibiotic pollution from fish farms: effect on aquatic microflora. Microbiol. Sci., 2: 113-117.

Baya, A.M., P.R. Brayton, V.L. Brown, D.J. Grime, E. Russek-Choen and R.R. Colwell, 1986. Coinsident plasmids and antimicrobial resistance in marine bacteria isolated from polluted and unpolluted Atlantic ocean samples. Appl. Environ. Microbiol., 51: 1285-1292.

Helland, H., 1988. Development of resistance to vibriostat 0/129 by a strain of Vibrio an- quillarum. Thesis in General Microbiology at The University of Bergen, Norway.

Jacobsen, P. and L. Berglind, 1988. Persistence of oxytetracycline in sediments from fish farms. Aquaculture, 70: 365-370.

Jones, J.G., S. Gardener, B.M. Simon and R.W. Pickup, 1986. Factors affecting the measure- ment of antibiotic resistance in bacteria isolated from lake water. J. Appl. Bacteriol., 60: 455-462.

Knox, J.H. and J. Jurand, 1979. Mechanism of reversed-phase separation of tetracyclines by high-performance liquid chromatography. J. Chromatogr., 186: 760-782.

Lunestad, B.T. and J. Goks~yr, 1990. Reduction in the antibacterial effect of oxytetracycline in sea water by complex formation with magnesium and calcium. Dis. Aquat. Org.,

9: 67-72. The Merck Index, 10th edn., 1983. Merck and Co., Inc., Rahway, NJ, USA. Oka, H., H. Matsumoto and K. Uno, 1985. Improvement of chemical analysis of antibiotics.

VIII. Application of prepacked C-18 cartridge for the analysis of tetracycline residues in animal liver. J. Chromatogr., 325: 265-274.

36 O B . S A M U E L S E N ET AL.

Porter, K.G. and Y.S. Feig, 1980. The use of DAPI for identifying and counting aquatic microflora. Limnol. Oceanogr., 25: 943-948.

Samuelsen, O.B., A. Ervik and E. Solheim, 1988. A qualitative and quantitative analysis of the sediment gas and diethylether extract of the sediment from salmon farms. Aquaculture, 74: 277-285.

Samuelsen, O.B., 1989. Degradation of oxytetracycline in seawater at two different temperatures and light intensities, and the persistence of oxytetracycline in the sediment from a fish farm. Aquaculture, 83: 7-16.

Sandaa, R-A., 1990. Spredning av antibiotikaresistens i et marint oppdrettsmilj0. Thesis in General Microbiology at the University of Bergen, Norway.

Schlotfeldt, H.J., W. Neumann, H. Fuhrman, K. Pfortmueller and H. Boehm, 1985. Remarks on increasing resistance of fish pathogenic and facultative fish pathogenic bacteria in lower Saxony (FRG). Fish Pathol., 20: 85-91.

Toranzo, A.E., P. Combarro, M.L. Lemos and J.L. Barja, 1984. Plasmid coding for transferable drug resistance in bacteria isolated from cultured rainbow trout. Appl. En- viron. Microbiol., 48: 872-877.

Torsvik, V.L., R. S0rheim and J. GoksOyr, 1988. Antibiotic resistance of bacteria from fish farm sediments. ICES-report, C.M. 1988/F:10.