- Email: [email protected]
Biodegradation and biosorption of tetracycline and tylosin antibiotics in activated sludge system
Process Biochemistry 44 (2009) 1302–1306
Contents lists available at ScienceDirect
Process Biochemistry journal homepage: www.elsevier.com/locate/procbio
Short communication
Biodegradation and biosorption of tetracycline and tylosin antibiotics in activated sludge system Nolwenn Prado a, Juan Ochoa b, Abdeltif Amrane a,* a b
Sciences Chimiques de Rennes, UMR 6226 CNRS-ENSCR, Avenue du Ge´ne´ral Leclerc, 35 700 Rennes, France Veolia Environnement Research Center, Chemin de la Digue, B.P. 76, 78 603 Maisons-Laffitte Cedex, France
A R T I C L E I N F O
A B S T R A C T
Article history: Received 19 March 2009 Received in revised form 4 August 2009 Accepted 10 August 2009
The aim of this study was to determine the fate of veterinary antibiotics entering biological treatment process. Due to the prevalence of their respective antibiotic family usage in livestock, tetracycline and tylosin were selected. Using modified Sturm test (OECD 301-B), their biodegradation were compared to that of a referent pollutant, sodium benzoate, well-known for its high biodegradability. Biodegradation rates were 28 and 35% for tetracycline and 4 and 5% for tylosin showing an absence of biodegradability. OECD 301-B inhibition tests showed a potential toxicity of both molecules on activated sludge inoculum derived from membrane bioreactor. Tetracycline presented good adsorbability while tylosin remained mostly present in the soluble phase. The Langmuir maximum adsorption capacity (Cs,max) was found to be 72 and 7.7 mg g1 for tetracycline and tylosin, respectively. Adsorption was therefore the most favourable fate for tetracycline entering a biological process. Conclusions on tylosin case were more controversial. ß 2009 Elsevier Ltd. All rights reserved.
Keywords: Biodegradation Biosorption Tetracycline Tylosin Activated sludge
1. Introduction Largely used in livestock, veterinary pharmaceutically active compounds are now identified as environmental issue. Public concern on their environmental fate and occurrence increased in recent years. Among veterinary pharmaceuticals, antibiotics are widely prescribed with a prevalence of the tetracycline family; 50% of the antibiotics sold in France in 2004 [1] were derived from tetracycline. The other antibiotics frequently used in livestock are aminoside, b-lactamine, macrolide, polypeptide, sulfamide and trimetoprime. Tylosin antibiotic belong to the macrolide family representing 7.8% of the total veterinary antibiotic sold in France in 2004 [1]. Tetracycline and tylosin are both broad-spectrum antibiotics widely used in pig production, tetracycline is used in case of respiratory affection and tylosin is recommended in case of diarrhoea. Antibiotics are proscribed as growth factor in France since 2006. However, their usage as therapeutics is still very large as far as they are nearly always used in group [2]. Moreover the dosage of veterinary antibiotics is high with among 40 mg kg1 day1 during 10 days and 10 mg kg1 day1 during 15 days for tetracycline and tylosin, respectively. In many cases, only part of the treatment is actually metabolised [3,4], the part left is found back as active form in pigs’ excreta.
* Corresponding author. Tel.: +33 223238155; fax: +33 223238120. E-mail address: [email protected] (A. Amrane). 1359-5113/$ – see front matter ß 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.procbio.2009.08.006
The concentration of pig production in a restricted area leads to exceed cropland capacity to receive piggeries wastewater as agronomic rates. As a results pigs industry involves environmental burdens on water and air quality, which finally increased pressure for the installation of water treatment plants. Swine wastewater is commonly treated by biological way with activated sludge processes and anaerobic or aerobic digestion processes [5]. Treatments by system including activated sludge such as membrane bioreactors have proved their efficiency for the decrease of organic and nitrogen loads of swine wastewater [6]; however, antibiotics excreted as active form may enter the system and results in an acute contamination during a therapeutic usage. Different studies on antibiotics fate [7–10] in bioreactor highlighted three different ways for antibiotics disappearance, the most favourable case is biodegradation, another way is accumulation on biomass defined as biosorption, which leads to the release of molecules after biomass death, and a third way is hydrolysis. CO2-evolution tests (OECD 301-B), formerly known as modified Sturm tests [11], are commonly used for the evaluation of the biodegradation potential of non-volatile molecules via the measurement of the produced carbon dioxide. In this test, the ratio molecules/biomass is high and the biodegradation qualified as ‘‘easy’’. Further tests could be done in order to evaluate the ‘‘inherent’’ biodegradability [12], which means the potential biodegradation in a specific environment but would not be relevant to simulate an acute pollution of a bioreactor. Biosorption of non-biodegradable molecules on activated sludge can be compared to adsorption on an inactive adsorbent
N. Prado et al. / Process Biochemistry 44 (2009) 1302–1306
and Freundlich and Langmuir absorption models can be used to describe the phenomenon. The Freundlich model (1939) assumes a heterogeneous adsorbent surface with an exponential distribution of adsorption sites. It leads to the determination of Kf and n, Freundlich coefficient and linearity parameter, respectively and no restriction due to possible molecule–adsorbent interactions exist. As far as the adsorption capacity has no upper limit, this model is confined to diluted solutions. Kf and n are limited to the couple molecule–adsorbent at a given temperature. The Langmuir model (1918) describes the molecule–adsorbent interactions with a limited number of adsorption sites and no interaction between the molecules. It leads to the determination of Kl and Cs,max, the Langmuir coefficient and the maximum solid concentration. Freundlich coefficients for tylosin and oxytetracycline can be found in the available literature [13–15], contrarily to Langmuir coefficients for relevant environmental conditions. Previous work showed that nearly 90% of tetracycline was eliminated from the soluble phase in a membrane bioreactor; however, no conclusion on the tetracycline fate was possible at this stage [16]. This paper proposes to evaluate biodegradation and biosorption of tetracycline and tylosin in presence of activated sludge at lab-scale. Modified Sturm test was used for biodegradation evaluation. Biosorption was evaluated by batch tests in order to determine Langmuir and Freundlich constants for each molecule. The purpose was to predict the most probable route for tetracycline and tylosin antibiotics through activated sludge processes facing an acute contamination.
1303
with 0.0125 M Ba(OH)2 (Fig. 1). The CO2-free air was passed on to an air sparger with one input and six output channels and through PE-tubes to the SVU. Two Ba(OH)2 traps were connected to each SVU. The six reactors were prepared as described in Fig. 1 using the mineral medium given in OECD 301-B for dilutions. The mass of CO2 produced (mCO2) and the biodegradation ratio (nCO2) were determined as follows: nHCl V ¼ ðC BaðOHÞ2 V BaðOHÞ2 Þ C HCl HCl 2 2 44 M CO2 ¼ nCO2 12
nCO2 ¼ nBaðOHÞ2
(1)
M CO2 ¼ nCO2
(2)
SVU1 called blank inoculum test allowed to evaluate the CO2 production if no carbon was added. CO2 production in SVU1 corresponded to the measured endogenous CO2 production (mCO2endo). SVU3 corresponded to the abiotic test to evaluate the CO2 production if no activated sludge was added to the reactor. CO2 production in SVU1 and SVU3 were taken into account for CO2 production calculations in the reactors 2, 4, 5 and 6. The ultimate biodegradation (%) of the target compound was calculated as follows:
Biodð%Þ ¼
mCO2 100 ThCO2 Tot
ThCO2 ¼ mC i
44 12
(3)
(4)
With ThCO2 corresponding to the total theoretical CO2 formation produced by total oxidation of the material and mCi corresponding to the initial carbon mass in the reactor (mg TOC L1). The ratio 44/12 was the conversion factor of carbon to carbon dioxide. 2.3. Total organic carbon (TOC) determination
2. Materials and methods 2.1. Chemical and reagents Tetracycline hydrochloride (>96%, HPLC) and tylosin tartrate (>98% UV) were obtained from Fluka–Sigma–Aldrich (St. Quentin Fallavier, France). Citric acid (anhydrous, 99%), Na2EDTA (99%), trifluoroacetic acid (99.5%) and Na2HPO4 (99%) were purchased from Acroˆs Organics (Noisy-le-grand, France). Methanol and acetonitrile (ACN) were both HPLC grade from Fisher Scientific (Illkirch, France) and tetrahydrofuran (THF–HPLC grade, 99.7%) was purchased from Prolabo VWR (Fontenay-sous-Bois, France). Standards were made with ultra pure water (Purelab Options–Q7/15, Elga, 18.2 MV cm). Conditioning and extraction buffer was EDTA–McIlvaine buffer (50:50), prepared by mixing 150 mL of 0.1 M EDTA (ethylenediaminetetraacetic acid), 90 mL 0.2 M citric acid, 60 mL 0.4 M Na2HPO4. Extraction buffer pH was adjusted to 4 by adding H3PO4 if needed. 2.2. Biodegradation tests Modified Sturm tests (28 days aerobic degradation) were conducted according to the OECD 301-B method. To determine heterotrophic activity the CO2 uptake rate test was applied. Biomass to be characterized was placed into a sealed vessel unit (SVU); the inoculum concentration was 30 mg L1. The SVU was a 1 L vessel continuously aerated with air CO2-free (the system was composed by six units, Fig. 1). Gas from SVU was continuously transferred in a CO2 trap unit. CO2 produced by microbial activity was trapped in a solution of barium hydroxyde (Ba(OH)2) which precipitate as barium carbonate in presence of CO2. The remaining barium hydroxide was titrated with 0.05N standard HCl in the presence of phenolphtaleine. The CO2-free air production system consisted of an air compressor, two 200 mL gas wash bottle filled with 4 M NaOH, followed by one 200 mL gas wash bottle filled
TOC was analysed via a TOC-meter (OI-Analytical 1010). Sample analysis included several steps. At the first step, the sample was acidified with sulfuric acid to reach a pH lower than 2 and served with gas to remove the inorganic carbon. Carbon was then oxidized and released as CO2, which was then determined by an infrared detector. 2.4. Biosorption tests For biosorption, a common isotherm experiment was applied. A known amount of activated sludge (about 0.5 g) was suspended in 500 mL identical batch reactors where various concentrations of the target compounds (tetracycline and tylosin) were added. All reactors were placed in thermostated baths at 25 8C. Initial samples and samples after 3 and 24 h were taken and analysed by HPLC with UV detection. The equilibrium was considered reached after 24 h. Freundlich coefficients were determined by means of the Frendlich model: n C s ¼ K f Cw
(5)
With Cs and Cw the equilibrium target compound concentrations on the biomass (mg g1) and in solution (mg L1) respectively, Kf the Freundlich parameter (L g1) and n the linearity parameter. Langmuir coefficients were determined using the Langmuir equation:
Cs ¼
C s;max K l C w 1 þ K l Cw
(6)
With Cs and Cw the equilibrium target compound concentrations on the biomass (mg g1) and in solution (mg L1), respectively, Kl the Langmuir parameter (L mg1) and Cs,max the maximum adsorption capacity.
Fig. 1. Diagram of Sturm tests experimental design. Sealed vessel unit (SVU) 1, blank inoculum; SVU 2, reference (sodium benzoate, 47 mg L1 TOC); SVU 3, abiotic test; SVU 4 and 5, target molecule (40 mg L1 TOC); SVU 6, inhibition control test (sodium benzoate 47 mg L1 TOC + target molecule 40 mg L1 TOC).
N. Prado et al. / Process Biochemistry 44 (2009) 1302–1306
1304
Table 1 Initial and final TOC values measured during tetracycline and tylosin biodegradation tests. SVU
Sample
TOC (mg L1) Tetracycline
1 2 3 4 5 6
Inoculum Referent Abiotic Target compound Target compound Inhibition control
Tylosine
Initial theoretical
Initial measured
Final measured
Initial theoretical
Initial measured
Final measured
0 47 40 40 40 87
0.84 48.52 39.12 38.42 37.64 85.81
0.72 4.23 36.22 5.81 7.31 6.25
0 47 40 40 40 87
0.96 45.01 39.62 37.96 39.07 83.15
1.13 0.61 12.43 1.20 7.33 8.51
2.5. Tylosin and tetracycline analysis Samples were centrifuged (3600 rpm, 15 min) immediately after collection and filtered to 0.7 mm through Cloup glass microfibre filters (Champigny s/ Marne, France) under vacuum. 5 mL of extraction buffer and 2 mL of methanol were added per 100 mL of sample and then were thoroughly mixed prior to solid phase extraction (SPE) clean-up. Strong anion exchange (SAX) cartridge (Sep-Pak Vac 3cc, Waters, France) was preconditioned with methanol (5 mL) and SPE conditioning buffer (5 mL), and then the activated sludge extract or standard activated sludge extract was filtered through the cartridge at a rate of 100 mL h1. Analyses were performed using an HPLC system including a WatersTM 600 controller pump, a Waters 717plus autosampler and a Spectra-Physics UV2000 detector. Separations were performed on a Waters Symmetry C18 5 mm column. A gradient elution was carried out with tetrahydrofuran (solvent A), acetonitrile (solvent B) and 0.05% trifluroacetic acid in water (solvent C) as follow: 3, 92 and 5% from 0 to 4 min, 75, 20 and 5% from 4 to 18 min and 3, 92 and 5% from 18 to 20 min for solvents A, B and C, respectively. The flow rate was 1 mL mn1. The injection volume was 50 mL. Tetracycline and tylosin detections were performed at 360 and 270 nm.
3. Results and discussion 3.1. Biodegradation tests 3.1.1. Sturm test validation Two different modified Sturm test series were carried out, one for tetracycline and one for tylosin. In both tests an initial organic carbon concentration of 40 mg L1 was introduced into target compounds reactors (Table 1). The modified Sturm test lasted at least 28 days; experiments were carried out during 36 and 30 days for tetracycline and tylosin, respectively. According to the OECD 301-B guidelines, the total CO2 evolution in the inoculum blank at the end of the test should not normally exceed 40 mg L1. The total CO2 evolution into the blank inoculum tests were 147 and 65 mg L1 for tetracycline and tylosin tests, respectively. These high results were due to a lack of activated sludge preparation. In order to minimise the effect of the initial carbon content of the activated sludge matrix on CO2 production, activated sludge were washed and centrifuged before modified Sturm tests, the supernatant was discarded and replaced by mineral media. This operation was repeated several times. In our case the carbon content of the matrix led to exceed 40 mg L1 for the total CO2 evolution of the inoculum blank. However, the matrix was the same in each reactor during a given test, therefore the bias was counterbalanced since SVU1 results were taken into account in the calculation. The abiotic control with both target compounds (SVU3) displayed very low CO2 production (14 mg in 36 days and 7 mg in 30 days for tetracycline and tylosin, respectively). 3.1.2. Test time-courses The ‘‘lag phase’’ which corresponded to the beginning of the pollutant biodegradation was deduced for each test from CO2 timecourses (not shown). In both case, sodium benzoate (SB) was degraded nearly immediately (0.25 day) with a CO2 production rate of 28 and 20 mg day1 L1 of CO2 during the first 5 days. These results confirmed its status as reference pollutant because of its high
biodegradability. Biodegradation of both target compounds started rapidly (after about 1 day). However, the initial biodegradation rates differed, since a relatively higher rate of CO2 production seemed to be observed for tetracycline (15 and 2 mg day1 L1 of CO2 for TC and TYL) even if a possible bias due to the carbon content of the activated sludge matrix can account for this result. In this case, the biodegradation level, which took into account this bias, appeared therefore a better tool to characterize the biodegradation. 3.1.3. Biodegradation results For each test, maximum biodegradation level was achieved after 15 days (Fig. 2a and b) and final biodegradation of SB was 62 and 78% (Fig. 2a and b) despite its hypothetical complete biodegradation. 100% of biodegradation cannot be achieved due to its partial use for biosynthesis and not only for energetic metabolism after oxidation into CO2. The results obtained for tetracycline (28 and 35%; Fig. 2a) and tylosin (4 and 5%; Fig. 2b) indicated that both antibiotics can be classified as nonbiodegradable compounds. These results are in agreement with Gartiser and co-workers [12] concerning the tetracycline case. The negative results indicated a lower CO2 production than that of the inoculum blank showing a toxicity of the molecule on activated sludge. Inhibition test helped to determine molecule toxicity.
Fig. 2. Tetracycline (a) and tylosin (b) biodegradation kinetics during 28 days Sturm tests (inoculum source: MBR).
N. Prado et al. / Process Biochemistry 44 (2009) 1302–1306
1305
Fig. 4. Calculated data for tetracycline (TC) and tylosin (TYL) by means of the Langmuir model.
Fig. 3. Validation of the Freundlich model (linear form: ln Cs = ln Kf + n ln Cw) (a) and the Langmuir model (linear form: 1/Cs = 1/Cs,max KlCw + 1/Cs,max) (b) for tetracycline (~) and tylosin (*) adsorption on activated sludge. The linearities n were 0.772 and 0.812 for TC and TYL and the Freundlich parameters Kf were 4.13 and 0.39 L g1 for TC and TYL; the maximum adsorption capacities Cs,max were 72 and 7.75 mg g1 for TC and TYL and the Langmuir parameters Kl were 0.061 and 0.048 L mg1 for TC and TYL.
3.1.4. Inhibition test results SVU6 was designed in order to evaluate the biodegradation of the reference compound (SB) in presence of a target molecule, illustrating the toxicity of the molecule on activated sludge activity. The final biodegradation of inhibition controls were 18 and 30% at 30 days for tetracycline and tylosin, respectively. Even if both compounds showed an inhibitory effect on SB biodegradation, tetracycline toxicity seemed to be more important. 3.1.5. TOC mass balance TOC measurements in the soluble phase of reactor samples at the beginning and at the end of biodegradation tests (Table 1) gave an indication on antibiotics fate into activated sludge reactors. No significant CO2 formation occurred during biodegradation tests in SVU4 and 5 (Fig. 2a and b), showing an absence of TC and TYL biodegradation. However, an important decrease of the TOC concentrations was observed at the end of experiments (Table 1). Antibiotics adsorption on the solid phase (biomass) can be therefore assumed. 3.2. Biosorption results The equilibrium concentration ranges investigated were 2–100 and 2–15 mg L1 for tetracycline and tylosin, respectively. The observed sorption kinetics were rapid; only few hours were needed to reach equilibrium. However, 24 h were considered in order to compare with the available literature. Indeed, during the first day, the activated sludge may be considered as a biologically inactive material owing to the ‘‘lag phase’’ biodegradation (about 1 day). Freundlich model was therefore applied to tetracycline and tylosin at equilibrium (Fig. 3a) with satisfactory correlation coefficients (0.97 and 0.99 for tetracycline and tylosin). The
resulting n coefficients were 0.772 and 0.812 (Fig. 3) for tetracycline and tylosin showing an absence of linearity between the concentrations in the soluble and on the solid phases (Henry law’s case). Therefore, Langmuir model had to be considered. The Langmuir model application led to good correlation coefficients (0.994 and 0.999 for tetracycline and tylosin; Fig. 3b). The observed Langmuir coefficients and Cs,max gave interesting indications on antibiotics fate (Fig. 3). Tetracycline was highly adsorbable on activated sludge with a high Cs,max of 72 mg g1, while tylosin Cs,max was about 10 time lower (7.75 mg g1). Furthermore, the Langmuir adsorption coefficients calculated showed that for a concentration upper than 10 mg L1, the concentration of tetracycline adsorbed on the solid phase was higher than that of tylosin and this difference was accentuated when the concentration in the solution rose as illustrated in Fig. 4. The high adsorbability of tetracycline on activated sludge tends to prove that biosorption is the most favourable route for tetracyclines fate into a biological system. The partitioning of tylosin between soluble and solid phases tended to promote soluble phase. However, these results should be handled with care as far as the activated sludge were washed before sorption test and organic content of environmental real matrix may modify sorption capacities. 4. Conclusion The biodegradation yields were 28 and 35% for tetracycline and 4 and 5% for tylosin, showing that neither of the two molecules was biodegradable. The strong tetracycline adsorption on activated sludge showed that the partition between soluble and solid phases was widely favourable to the solid phase. Moreover, the maximum capacity adsorption using the Langmuir model was high (Cs,max = 72 mg g1). The important TOC decrease during experiments confirmed the high tetracycline biosorption. The conclusions were more controversial in the case of tylosin, since the biosorption results showed that the partition was favourable to the mobile phase, while a low amount of TOC was found in the soluble phase after 28 days of experiment. A modification of the tylosin conformation for a more hydrophobic form, consequently to environmental modifications or bacterial activity can account for this contradictory result, owing to their various molecular conformations [17]. The fate of antibiotics is therefore closely linked to the operating conditions of the biological process. In contrast, the introduction of an antibiotic in a biological system can alter the operating conditions, particularly in case of an inhibitory capacity as shown during inhibition tests of tetracycline and tylosin.
1306
N. Prado et al. / Process Biochemistry 44 (2009) 1302–1306
References [1] AFSSA. Usages ve´te´rinaires des antibiotiques, re´sistance bacte´rienne et conse´quences pour la sante´ humaine; 2006. [2] Chauvin C, Beloeil P-A, Orand J-P, Sanders P, Madec F. A survey of group-level antibiotic prescriptions in pig production in France. Prev Vet Med 2002;55:109–20. [3] Feinman SE, Matheson J-C. Environmental impact statement subtherapeutic antibacterial agents in animal feeds. Rockville, MD: Department of Health, Education and Welfare; 1978. [4] Arikan OA, Sikora LJ, Mulbry W, Khan SU, Rice C, Foster GD. The fate and effect of oxytetracycline during the anaerobic digestion of manure from therapeutically treated calves. Process Biochem 2006;41:1637–43. [5] Lee S-I, Park J-H, Ko K-B, Koopman B. Effect of fermented swine wastes on biological removal in sequencing batch reactors. Water Res 1997;31:1807–12. [6] Prado N, Ochoa J, Audic J-L, Amrane A, Meinhold J. Semi-industrial-scale process for dilute swine wastewater treatment using a submerged membrane bioreactor (MBR) with direct reuse of treated water. Int J Chem React Eng 2007;5. [7] Samuel Suman Raj D, Anjaneyulu Y. Evaluation of biokinetic parameters for pharmaceutical wastewaters using aerobic oxidation integrated with chemical treatment. Process Biochem 2005;40:165–75. [8] Xu W, Zhang G, Li X, Zou S, Li P, Hu Z, et al. Occurence and elimination of antibiotics at four sewage treatment plants in the Pearl River Delta (PRD), South China. Water Res 2007;2007:4526–34.
[9] Watkinson AJ, Murby EJ, Constanzo SD. Removal of antibiotics in conventional and advanced wastewater treatment: implications for environmental discharge and wastewater recycling. Water Res 2007;41:4164–76. [10] Kim S, Eichhorn P, Jensen JN, Weber AS, Aga DS. Removal of antibiotics in wastewater: effect of hydraulic and solid retention times on the fate of tetracycline in the activated sludge process. Environ Sci Technol 2005;39: 5816–23. [11] Sturm RN. Biodegradability of nonionic surfactants screening test for predicting rate and ultimate biodegradation. J Am Oil Chem Soc 1973;50: 159–67. [12] Gartiser S, Urich E, Alexy R, Ku¨mmerer K. Ultimate biodegradation and elimination of antibiotics in inherent tests. Chemosphere 2007;67:604–13. [13] Pouliquen H, Le Bris H. Sorption of oxolinic acid and oxytetracycline to marine sediments. Chemosphere 1996;33:801–15. [14] Halling-Sorensen B, Nors Nielsen S, Lanzky PF, Ingerslev F, Holten Lu¨tzhoft HC, Jorgensen SE. Occurence, fate and effects of pharmaceutical substances in the environment—a review. Chemosphere 1998;36:357–93. [15] Rabolle M, Spliid NH. Sorption and mobility of metronidazole, olaquindox, oxytetracycline and tylosin in soil. Chemosphere 2000;40:715–22. [16] Prado N, Ochoa J, Amrane A. Biodegradation by activated sludge and toxicity of tetracycline into a semi-industrial membrane bioreactor. Bioresour Technol 2009;100:3769–74. [17] Prado N. Semi-industrial scale membrane bioreactor for the overall treatment of dilute swine wastewater—impact and fate of veterinary antibiotics in the process. PhD thesis, University of Rennes 1, France; 2008.
Contents lists available at ScienceDirect
Process Biochemistry journal homepage: www.elsevier.com/locate/procbio
Short communication
Biodegradation and biosorption of tetracycline and tylosin antibiotics in activated sludge system Nolwenn Prado a, Juan Ochoa b, Abdeltif Amrane a,* a b
Sciences Chimiques de Rennes, UMR 6226 CNRS-ENSCR, Avenue du Ge´ne´ral Leclerc, 35 700 Rennes, France Veolia Environnement Research Center, Chemin de la Digue, B.P. 76, 78 603 Maisons-Laffitte Cedex, France
A R T I C L E I N F O
A B S T R A C T
Article history: Received 19 March 2009 Received in revised form 4 August 2009 Accepted 10 August 2009
The aim of this study was to determine the fate of veterinary antibiotics entering biological treatment process. Due to the prevalence of their respective antibiotic family usage in livestock, tetracycline and tylosin were selected. Using modified Sturm test (OECD 301-B), their biodegradation were compared to that of a referent pollutant, sodium benzoate, well-known for its high biodegradability. Biodegradation rates were 28 and 35% for tetracycline and 4 and 5% for tylosin showing an absence of biodegradability. OECD 301-B inhibition tests showed a potential toxicity of both molecules on activated sludge inoculum derived from membrane bioreactor. Tetracycline presented good adsorbability while tylosin remained mostly present in the soluble phase. The Langmuir maximum adsorption capacity (Cs,max) was found to be 72 and 7.7 mg g1 for tetracycline and tylosin, respectively. Adsorption was therefore the most favourable fate for tetracycline entering a biological process. Conclusions on tylosin case were more controversial. ß 2009 Elsevier Ltd. All rights reserved.
Keywords: Biodegradation Biosorption Tetracycline Tylosin Activated sludge
1. Introduction Largely used in livestock, veterinary pharmaceutically active compounds are now identified as environmental issue. Public concern on their environmental fate and occurrence increased in recent years. Among veterinary pharmaceuticals, antibiotics are widely prescribed with a prevalence of the tetracycline family; 50% of the antibiotics sold in France in 2004 [1] were derived from tetracycline. The other antibiotics frequently used in livestock are aminoside, b-lactamine, macrolide, polypeptide, sulfamide and trimetoprime. Tylosin antibiotic belong to the macrolide family representing 7.8% of the total veterinary antibiotic sold in France in 2004 [1]. Tetracycline and tylosin are both broad-spectrum antibiotics widely used in pig production, tetracycline is used in case of respiratory affection and tylosin is recommended in case of diarrhoea. Antibiotics are proscribed as growth factor in France since 2006. However, their usage as therapeutics is still very large as far as they are nearly always used in group [2]. Moreover the dosage of veterinary antibiotics is high with among 40 mg kg1 day1 during 10 days and 10 mg kg1 day1 during 15 days for tetracycline and tylosin, respectively. In many cases, only part of the treatment is actually metabolised [3,4], the part left is found back as active form in pigs’ excreta.
* Corresponding author. Tel.: +33 223238155; fax: +33 223238120. E-mail address: [email protected] (A. Amrane). 1359-5113/$ – see front matter ß 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.procbio.2009.08.006
The concentration of pig production in a restricted area leads to exceed cropland capacity to receive piggeries wastewater as agronomic rates. As a results pigs industry involves environmental burdens on water and air quality, which finally increased pressure for the installation of water treatment plants. Swine wastewater is commonly treated by biological way with activated sludge processes and anaerobic or aerobic digestion processes [5]. Treatments by system including activated sludge such as membrane bioreactors have proved their efficiency for the decrease of organic and nitrogen loads of swine wastewater [6]; however, antibiotics excreted as active form may enter the system and results in an acute contamination during a therapeutic usage. Different studies on antibiotics fate [7–10] in bioreactor highlighted three different ways for antibiotics disappearance, the most favourable case is biodegradation, another way is accumulation on biomass defined as biosorption, which leads to the release of molecules after biomass death, and a third way is hydrolysis. CO2-evolution tests (OECD 301-B), formerly known as modified Sturm tests [11], are commonly used for the evaluation of the biodegradation potential of non-volatile molecules via the measurement of the produced carbon dioxide. In this test, the ratio molecules/biomass is high and the biodegradation qualified as ‘‘easy’’. Further tests could be done in order to evaluate the ‘‘inherent’’ biodegradability [12], which means the potential biodegradation in a specific environment but would not be relevant to simulate an acute pollution of a bioreactor. Biosorption of non-biodegradable molecules on activated sludge can be compared to adsorption on an inactive adsorbent
N. Prado et al. / Process Biochemistry 44 (2009) 1302–1306
and Freundlich and Langmuir absorption models can be used to describe the phenomenon. The Freundlich model (1939) assumes a heterogeneous adsorbent surface with an exponential distribution of adsorption sites. It leads to the determination of Kf and n, Freundlich coefficient and linearity parameter, respectively and no restriction due to possible molecule–adsorbent interactions exist. As far as the adsorption capacity has no upper limit, this model is confined to diluted solutions. Kf and n are limited to the couple molecule–adsorbent at a given temperature. The Langmuir model (1918) describes the molecule–adsorbent interactions with a limited number of adsorption sites and no interaction between the molecules. It leads to the determination of Kl and Cs,max, the Langmuir coefficient and the maximum solid concentration. Freundlich coefficients for tylosin and oxytetracycline can be found in the available literature [13–15], contrarily to Langmuir coefficients for relevant environmental conditions. Previous work showed that nearly 90% of tetracycline was eliminated from the soluble phase in a membrane bioreactor; however, no conclusion on the tetracycline fate was possible at this stage [16]. This paper proposes to evaluate biodegradation and biosorption of tetracycline and tylosin in presence of activated sludge at lab-scale. Modified Sturm test was used for biodegradation evaluation. Biosorption was evaluated by batch tests in order to determine Langmuir and Freundlich constants for each molecule. The purpose was to predict the most probable route for tetracycline and tylosin antibiotics through activated sludge processes facing an acute contamination.
1303
with 0.0125 M Ba(OH)2 (Fig. 1). The CO2-free air was passed on to an air sparger with one input and six output channels and through PE-tubes to the SVU. Two Ba(OH)2 traps were connected to each SVU. The six reactors were prepared as described in Fig. 1 using the mineral medium given in OECD 301-B for dilutions. The mass of CO2 produced (mCO2) and the biodegradation ratio (nCO2) were determined as follows: nHCl V ¼ ðC BaðOHÞ2 V BaðOHÞ2 Þ C HCl HCl 2 2 44 M CO2 ¼ nCO2 12
nCO2 ¼ nBaðOHÞ2
(1)
M CO2 ¼ nCO2
(2)
SVU1 called blank inoculum test allowed to evaluate the CO2 production if no carbon was added. CO2 production in SVU1 corresponded to the measured endogenous CO2 production (mCO2endo). SVU3 corresponded to the abiotic test to evaluate the CO2 production if no activated sludge was added to the reactor. CO2 production in SVU1 and SVU3 were taken into account for CO2 production calculations in the reactors 2, 4, 5 and 6. The ultimate biodegradation (%) of the target compound was calculated as follows:
Biodð%Þ ¼
mCO2 100 ThCO2 Tot
ThCO2 ¼ mC i
44 12
(3)
(4)
With ThCO2 corresponding to the total theoretical CO2 formation produced by total oxidation of the material and mCi corresponding to the initial carbon mass in the reactor (mg TOC L1). The ratio 44/12 was the conversion factor of carbon to carbon dioxide. 2.3. Total organic carbon (TOC) determination
2. Materials and methods 2.1. Chemical and reagents Tetracycline hydrochloride (>96%, HPLC) and tylosin tartrate (>98% UV) were obtained from Fluka–Sigma–Aldrich (St. Quentin Fallavier, France). Citric acid (anhydrous, 99%), Na2EDTA (99%), trifluoroacetic acid (99.5%) and Na2HPO4 (99%) were purchased from Acroˆs Organics (Noisy-le-grand, France). Methanol and acetonitrile (ACN) were both HPLC grade from Fisher Scientific (Illkirch, France) and tetrahydrofuran (THF–HPLC grade, 99.7%) was purchased from Prolabo VWR (Fontenay-sous-Bois, France). Standards were made with ultra pure water (Purelab Options–Q7/15, Elga, 18.2 MV cm). Conditioning and extraction buffer was EDTA–McIlvaine buffer (50:50), prepared by mixing 150 mL of 0.1 M EDTA (ethylenediaminetetraacetic acid), 90 mL 0.2 M citric acid, 60 mL 0.4 M Na2HPO4. Extraction buffer pH was adjusted to 4 by adding H3PO4 if needed. 2.2. Biodegradation tests Modified Sturm tests (28 days aerobic degradation) were conducted according to the OECD 301-B method. To determine heterotrophic activity the CO2 uptake rate test was applied. Biomass to be characterized was placed into a sealed vessel unit (SVU); the inoculum concentration was 30 mg L1. The SVU was a 1 L vessel continuously aerated with air CO2-free (the system was composed by six units, Fig. 1). Gas from SVU was continuously transferred in a CO2 trap unit. CO2 produced by microbial activity was trapped in a solution of barium hydroxyde (Ba(OH)2) which precipitate as barium carbonate in presence of CO2. The remaining barium hydroxide was titrated with 0.05N standard HCl in the presence of phenolphtaleine. The CO2-free air production system consisted of an air compressor, two 200 mL gas wash bottle filled with 4 M NaOH, followed by one 200 mL gas wash bottle filled
TOC was analysed via a TOC-meter (OI-Analytical 1010). Sample analysis included several steps. At the first step, the sample was acidified with sulfuric acid to reach a pH lower than 2 and served with gas to remove the inorganic carbon. Carbon was then oxidized and released as CO2, which was then determined by an infrared detector. 2.4. Biosorption tests For biosorption, a common isotherm experiment was applied. A known amount of activated sludge (about 0.5 g) was suspended in 500 mL identical batch reactors where various concentrations of the target compounds (tetracycline and tylosin) were added. All reactors were placed in thermostated baths at 25 8C. Initial samples and samples after 3 and 24 h were taken and analysed by HPLC with UV detection. The equilibrium was considered reached after 24 h. Freundlich coefficients were determined by means of the Frendlich model: n C s ¼ K f Cw
(5)
With Cs and Cw the equilibrium target compound concentrations on the biomass (mg g1) and in solution (mg L1) respectively, Kf the Freundlich parameter (L g1) and n the linearity parameter. Langmuir coefficients were determined using the Langmuir equation:
Cs ¼
C s;max K l C w 1 þ K l Cw
(6)
With Cs and Cw the equilibrium target compound concentrations on the biomass (mg g1) and in solution (mg L1), respectively, Kl the Langmuir parameter (L mg1) and Cs,max the maximum adsorption capacity.
Fig. 1. Diagram of Sturm tests experimental design. Sealed vessel unit (SVU) 1, blank inoculum; SVU 2, reference (sodium benzoate, 47 mg L1 TOC); SVU 3, abiotic test; SVU 4 and 5, target molecule (40 mg L1 TOC); SVU 6, inhibition control test (sodium benzoate 47 mg L1 TOC + target molecule 40 mg L1 TOC).
N. Prado et al. / Process Biochemistry 44 (2009) 1302–1306
1304
Table 1 Initial and final TOC values measured during tetracycline and tylosin biodegradation tests. SVU
Sample
TOC (mg L1) Tetracycline
1 2 3 4 5 6
Inoculum Referent Abiotic Target compound Target compound Inhibition control
Tylosine
Initial theoretical
Initial measured
Final measured
Initial theoretical
Initial measured
Final measured
0 47 40 40 40 87
0.84 48.52 39.12 38.42 37.64 85.81
0.72 4.23 36.22 5.81 7.31 6.25
0 47 40 40 40 87
0.96 45.01 39.62 37.96 39.07 83.15
1.13 0.61 12.43 1.20 7.33 8.51
2.5. Tylosin and tetracycline analysis Samples were centrifuged (3600 rpm, 15 min) immediately after collection and filtered to 0.7 mm through Cloup glass microfibre filters (Champigny s/ Marne, France) under vacuum. 5 mL of extraction buffer and 2 mL of methanol were added per 100 mL of sample and then were thoroughly mixed prior to solid phase extraction (SPE) clean-up. Strong anion exchange (SAX) cartridge (Sep-Pak Vac 3cc, Waters, France) was preconditioned with methanol (5 mL) and SPE conditioning buffer (5 mL), and then the activated sludge extract or standard activated sludge extract was filtered through the cartridge at a rate of 100 mL h1. Analyses were performed using an HPLC system including a WatersTM 600 controller pump, a Waters 717plus autosampler and a Spectra-Physics UV2000 detector. Separations were performed on a Waters Symmetry C18 5 mm column. A gradient elution was carried out with tetrahydrofuran (solvent A), acetonitrile (solvent B) and 0.05% trifluroacetic acid in water (solvent C) as follow: 3, 92 and 5% from 0 to 4 min, 75, 20 and 5% from 4 to 18 min and 3, 92 and 5% from 18 to 20 min for solvents A, B and C, respectively. The flow rate was 1 mL mn1. The injection volume was 50 mL. Tetracycline and tylosin detections were performed at 360 and 270 nm.
3. Results and discussion 3.1. Biodegradation tests 3.1.1. Sturm test validation Two different modified Sturm test series were carried out, one for tetracycline and one for tylosin. In both tests an initial organic carbon concentration of 40 mg L1 was introduced into target compounds reactors (Table 1). The modified Sturm test lasted at least 28 days; experiments were carried out during 36 and 30 days for tetracycline and tylosin, respectively. According to the OECD 301-B guidelines, the total CO2 evolution in the inoculum blank at the end of the test should not normally exceed 40 mg L1. The total CO2 evolution into the blank inoculum tests were 147 and 65 mg L1 for tetracycline and tylosin tests, respectively. These high results were due to a lack of activated sludge preparation. In order to minimise the effect of the initial carbon content of the activated sludge matrix on CO2 production, activated sludge were washed and centrifuged before modified Sturm tests, the supernatant was discarded and replaced by mineral media. This operation was repeated several times. In our case the carbon content of the matrix led to exceed 40 mg L1 for the total CO2 evolution of the inoculum blank. However, the matrix was the same in each reactor during a given test, therefore the bias was counterbalanced since SVU1 results were taken into account in the calculation. The abiotic control with both target compounds (SVU3) displayed very low CO2 production (14 mg in 36 days and 7 mg in 30 days for tetracycline and tylosin, respectively). 3.1.2. Test time-courses The ‘‘lag phase’’ which corresponded to the beginning of the pollutant biodegradation was deduced for each test from CO2 timecourses (not shown). In both case, sodium benzoate (SB) was degraded nearly immediately (0.25 day) with a CO2 production rate of 28 and 20 mg day1 L1 of CO2 during the first 5 days. These results confirmed its status as reference pollutant because of its high
biodegradability. Biodegradation of both target compounds started rapidly (after about 1 day). However, the initial biodegradation rates differed, since a relatively higher rate of CO2 production seemed to be observed for tetracycline (15 and 2 mg day1 L1 of CO2 for TC and TYL) even if a possible bias due to the carbon content of the activated sludge matrix can account for this result. In this case, the biodegradation level, which took into account this bias, appeared therefore a better tool to characterize the biodegradation. 3.1.3. Biodegradation results For each test, maximum biodegradation level was achieved after 15 days (Fig. 2a and b) and final biodegradation of SB was 62 and 78% (Fig. 2a and b) despite its hypothetical complete biodegradation. 100% of biodegradation cannot be achieved due to its partial use for biosynthesis and not only for energetic metabolism after oxidation into CO2. The results obtained for tetracycline (28 and 35%; Fig. 2a) and tylosin (4 and 5%; Fig. 2b) indicated that both antibiotics can be classified as nonbiodegradable compounds. These results are in agreement with Gartiser and co-workers [12] concerning the tetracycline case. The negative results indicated a lower CO2 production than that of the inoculum blank showing a toxicity of the molecule on activated sludge. Inhibition test helped to determine molecule toxicity.
Fig. 2. Tetracycline (a) and tylosin (b) biodegradation kinetics during 28 days Sturm tests (inoculum source: MBR).
N. Prado et al. / Process Biochemistry 44 (2009) 1302–1306
1305
Fig. 4. Calculated data for tetracycline (TC) and tylosin (TYL) by means of the Langmuir model.
Fig. 3. Validation of the Freundlich model (linear form: ln Cs = ln Kf + n ln Cw) (a) and the Langmuir model (linear form: 1/Cs = 1/Cs,max KlCw + 1/Cs,max) (b) for tetracycline (~) and tylosin (*) adsorption on activated sludge. The linearities n were 0.772 and 0.812 for TC and TYL and the Freundlich parameters Kf were 4.13 and 0.39 L g1 for TC and TYL; the maximum adsorption capacities Cs,max were 72 and 7.75 mg g1 for TC and TYL and the Langmuir parameters Kl were 0.061 and 0.048 L mg1 for TC and TYL.
3.1.4. Inhibition test results SVU6 was designed in order to evaluate the biodegradation of the reference compound (SB) in presence of a target molecule, illustrating the toxicity of the molecule on activated sludge activity. The final biodegradation of inhibition controls were 18 and 30% at 30 days for tetracycline and tylosin, respectively. Even if both compounds showed an inhibitory effect on SB biodegradation, tetracycline toxicity seemed to be more important. 3.1.5. TOC mass balance TOC measurements in the soluble phase of reactor samples at the beginning and at the end of biodegradation tests (Table 1) gave an indication on antibiotics fate into activated sludge reactors. No significant CO2 formation occurred during biodegradation tests in SVU4 and 5 (Fig. 2a and b), showing an absence of TC and TYL biodegradation. However, an important decrease of the TOC concentrations was observed at the end of experiments (Table 1). Antibiotics adsorption on the solid phase (biomass) can be therefore assumed. 3.2. Biosorption results The equilibrium concentration ranges investigated were 2–100 and 2–15 mg L1 for tetracycline and tylosin, respectively. The observed sorption kinetics were rapid; only few hours were needed to reach equilibrium. However, 24 h were considered in order to compare with the available literature. Indeed, during the first day, the activated sludge may be considered as a biologically inactive material owing to the ‘‘lag phase’’ biodegradation (about 1 day). Freundlich model was therefore applied to tetracycline and tylosin at equilibrium (Fig. 3a) with satisfactory correlation coefficients (0.97 and 0.99 for tetracycline and tylosin). The
resulting n coefficients were 0.772 and 0.812 (Fig. 3) for tetracycline and tylosin showing an absence of linearity between the concentrations in the soluble and on the solid phases (Henry law’s case). Therefore, Langmuir model had to be considered. The Langmuir model application led to good correlation coefficients (0.994 and 0.999 for tetracycline and tylosin; Fig. 3b). The observed Langmuir coefficients and Cs,max gave interesting indications on antibiotics fate (Fig. 3). Tetracycline was highly adsorbable on activated sludge with a high Cs,max of 72 mg g1, while tylosin Cs,max was about 10 time lower (7.75 mg g1). Furthermore, the Langmuir adsorption coefficients calculated showed that for a concentration upper than 10 mg L1, the concentration of tetracycline adsorbed on the solid phase was higher than that of tylosin and this difference was accentuated when the concentration in the solution rose as illustrated in Fig. 4. The high adsorbability of tetracycline on activated sludge tends to prove that biosorption is the most favourable route for tetracyclines fate into a biological system. The partitioning of tylosin between soluble and solid phases tended to promote soluble phase. However, these results should be handled with care as far as the activated sludge were washed before sorption test and organic content of environmental real matrix may modify sorption capacities. 4. Conclusion The biodegradation yields were 28 and 35% for tetracycline and 4 and 5% for tylosin, showing that neither of the two molecules was biodegradable. The strong tetracycline adsorption on activated sludge showed that the partition between soluble and solid phases was widely favourable to the solid phase. Moreover, the maximum capacity adsorption using the Langmuir model was high (Cs,max = 72 mg g1). The important TOC decrease during experiments confirmed the high tetracycline biosorption. The conclusions were more controversial in the case of tylosin, since the biosorption results showed that the partition was favourable to the mobile phase, while a low amount of TOC was found in the soluble phase after 28 days of experiment. A modification of the tylosin conformation for a more hydrophobic form, consequently to environmental modifications or bacterial activity can account for this contradictory result, owing to their various molecular conformations [17]. The fate of antibiotics is therefore closely linked to the operating conditions of the biological process. In contrast, the introduction of an antibiotic in a biological system can alter the operating conditions, particularly in case of an inhibitory capacity as shown during inhibition tests of tetracycline and tylosin.
1306
N. Prado et al. / Process Biochemistry 44 (2009) 1302–1306
References [1] AFSSA. Usages ve´te´rinaires des antibiotiques, re´sistance bacte´rienne et conse´quences pour la sante´ humaine; 2006. [2] Chauvin C, Beloeil P-A, Orand J-P, Sanders P, Madec F. A survey of group-level antibiotic prescriptions in pig production in France. Prev Vet Med 2002;55:109–20. [3] Feinman SE, Matheson J-C. Environmental impact statement subtherapeutic antibacterial agents in animal feeds. Rockville, MD: Department of Health, Education and Welfare; 1978. [4] Arikan OA, Sikora LJ, Mulbry W, Khan SU, Rice C, Foster GD. The fate and effect of oxytetracycline during the anaerobic digestion of manure from therapeutically treated calves. Process Biochem 2006;41:1637–43. [5] Lee S-I, Park J-H, Ko K-B, Koopman B. Effect of fermented swine wastes on biological removal in sequencing batch reactors. Water Res 1997;31:1807–12. [6] Prado N, Ochoa J, Audic J-L, Amrane A, Meinhold J. Semi-industrial-scale process for dilute swine wastewater treatment using a submerged membrane bioreactor (MBR) with direct reuse of treated water. Int J Chem React Eng 2007;5. [7] Samuel Suman Raj D, Anjaneyulu Y. Evaluation of biokinetic parameters for pharmaceutical wastewaters using aerobic oxidation integrated with chemical treatment. Process Biochem 2005;40:165–75. [8] Xu W, Zhang G, Li X, Zou S, Li P, Hu Z, et al. Occurence and elimination of antibiotics at four sewage treatment plants in the Pearl River Delta (PRD), South China. Water Res 2007;2007:4526–34.
[9] Watkinson AJ, Murby EJ, Constanzo SD. Removal of antibiotics in conventional and advanced wastewater treatment: implications for environmental discharge and wastewater recycling. Water Res 2007;41:4164–76. [10] Kim S, Eichhorn P, Jensen JN, Weber AS, Aga DS. Removal of antibiotics in wastewater: effect of hydraulic and solid retention times on the fate of tetracycline in the activated sludge process. Environ Sci Technol 2005;39: 5816–23. [11] Sturm RN. Biodegradability of nonionic surfactants screening test for predicting rate and ultimate biodegradation. J Am Oil Chem Soc 1973;50: 159–67. [12] Gartiser S, Urich E, Alexy R, Ku¨mmerer K. Ultimate biodegradation and elimination of antibiotics in inherent tests. Chemosphere 2007;67:604–13. [13] Pouliquen H, Le Bris H. Sorption of oxolinic acid and oxytetracycline to marine sediments. Chemosphere 1996;33:801–15. [14] Halling-Sorensen B, Nors Nielsen S, Lanzky PF, Ingerslev F, Holten Lu¨tzhoft HC, Jorgensen SE. Occurence, fate and effects of pharmaceutical substances in the environment—a review. Chemosphere 1998;36:357–93. [15] Rabolle M, Spliid NH. Sorption and mobility of metronidazole, olaquindox, oxytetracycline and tylosin in soil. Chemosphere 2000;40:715–22. [16] Prado N, Ochoa J, Amrane A. Biodegradation by activated sludge and toxicity of tetracycline into a semi-industrial membrane bioreactor. Bioresour Technol 2009;100:3769–74. [17] Prado N. Semi-industrial scale membrane bioreactor for the overall treatment of dilute swine wastewater—impact and fate of veterinary antibiotics in the process. PhD thesis, University of Rennes 1, France; 2008.