Biotransformation of the Antibiotic Agent Flumequine by Ligninolytic

Article pubs.acs.org/est

Biotransformation of the Antibiotic Agent Flumequine by Ligninolytic Fungi and Residual Antibacterial Activity of the Transformation Mixtures Monika Č vančarová,†,‡ Monika Moeder,*,§ Alena Filipová,† Thorsten Reemtsma,§ and Tomás ̌ Cajthaml†,‡ †

Institute of Microbiology, Academy of Sciences of the Czech Republic, v.v.i., Vídeňská 1083, CZ-142 20 Prague 4, Cech Republic Institute of Environmental Studies, Faculty of Science, Charles University, Benátská 2, CZ-128 01 Prague 2, Czech Republic § Department of Analytical Chemistry, Helmholtz Centre for Environmental ResearchUFZ, Permoserstrasse 15, 04318 Leipzig, Germany ‡

S Supporting Information *

ABSTRACT: Flumequine, a fluoroquinolone antibiotic, is applied preferably in veterinary medicine, for stock breeding and treatment of aquacultures. Formation of drug resistance is a matter of general concern when antibiotics such as flumquine occur in the environment. Thus, biodegradation of flumequine in solution was investigated using five different ligninolytic fungi. Irpex lacteus, Dichomitus squalens, and Trametes versicolor proved most efficient and transformed more than 90% of flumequine within 6 or even 3 days. Panus tigrinus and Pleurotus ostreatus required up to 14 days to remove >90% of flumequine. Analyses of the metabolites by liquid chromatography−mass spectrometry suggest different transformation pathways for the different fungal strains. Structure proposals were elaborated for 8 metabolites. 7Hydroxy-flumequine and flumequine ethyl ester were identified as common metabolites produced by all ligninolytic fungi. The largest variety of metabolites was formed by D. squalens. Residual antibacterial activity of the metabolite mixtures was tested using gram-positive and gram-negative bacteria. While for the less efficient P. tigrinus and P. ostreatus cultures the antibacterial activities corresponded to the residual concentrations of flumequine, a remarkable antibacterial activity remained in the D. squalens cultures although flumequine was transformed to more than 90%. Obviously, antibacterially active transformation products were formed by this fungal strain.

■

L.12 Although the acute toxicity of flumequine toward aquatic organisms is relatively low, recent studies confirmed that flumequine can be bioaccumulated and exhibits genotoxic and carcinogenic properties.13−15 As reported previously, FLU is not degraded in water/ sediment slurries from aquacultural ponds under exclusion of light 16 and, photolysis was emphasized as the main abiotic transformation process of flumequine in water.17 Sirtori et al. 18 identified 14 photodegradation products of FLU which exhibited a higher toxicity toward Daphnia magna and Vibrio f ischeri than FLU. Thus, transformation of flumequine does not ensure a loss of antimicrobial activity. Microbial degradation of fluoroquinolones has been less frequently reported although, selected bacterial strains are able to degrade this type of antibiotics as shown for danofloxacin.19 White rot fungi with their extracellular enzymes possess a high potential to transform a large variety of anthropogenic compounds even when adsorbed on sediments or soil.20,21

INTRODUCTION Flumequine (FLU) is a broad-spectrum antimicrobial agent of the fluoroquinolone family and is commonly used in veterinary medicine especially against gram-negative bacteria. The bactericidal action of quinolones is based on inhibition of bacterial growth by interfering with the enzyme DNA-gyrase and thus, terminating the normal DNA synthesis.1 Flumequine and its metabolites such as the corresponding glucuronide can enter surface water and soil via feces and urine of treated animals.2,3 Furthermore, flumequine is directly applied as feed additive to aquacultures,4,5 often in too high amounts due to unawareness about the low bioavailability of FLU. Consequently, flumequine retains in circumjacent water and is adsorbed on sediment causing an increase of its persistence.6 In sediments concentrations of 32−579 μg/kg have been reported.7−9 Already subtherapeutic concentrations of antibiotics as found in the environment can trigger specific transcriptional responses in bacteria and increase the risk to develop drug resistance in pathogenic strains (subinhibitory concentrations of FLU = 0.03−8 mg/L seawater).10,11 Several ecotoxicity tests indicated effects on growth and reproduction of cyanobacteria and fish species (fathead minnow) at concentrations ranged from 8 μg/L to 23 mg/ © 2013 American Chemical Society

Received: Revised: Accepted: Published: 14128

August 5, 2013 November 21, 2013 November 21, 2013 November 21, 2013 dx.doi.org/10.1021/es403470s | Environ. Sci. Technol. 2013, 47, 14128−14136

Environmental Science & Technology

Article

determined at 0.03 μg/mL with mean precision of ±17% (n = 6). All samples were measured twice and the data were averaged for further data processing. Blank analyses were included regularly to check memory effects. Standards were analyzed also multiply within analysis series to monitor instrumental performance. Identification of Main Metabolites. After 4 days of cultivation by ligninolytic fungi, FLU and the metabolites were extracted from the liquid medium with ethyl acetate by liquid− liquid extraction.21 The extraction was repeated five times with 20 mL portions of fresh solvent. The extracts were combined and concentrated to 10 mL by a vacuum evaporator (RVO 200A, Ingos, Czech Republic). To support structure identification, H-acidic functional groups in the metabolites were labeled by H−D exchange reactions. For this purpose, 2 mL of the ethyl acetate extracts were evaporated to dryness and dissolved in deuterated water. This procedure was repeated twice to ensure a quantitative H−D exchange. For these experiments, D2O instead of H2O was used as eluent. Subsequent LC-MS/MS analysis let recognize a defined mass increase pointing to the number of H-acidic substituents in their molecules. The extracts were analyzed by HPLC-ESI-MS using an Agilent 1100 series HPLC instrument (Agilent Technologies, Waldbronn, Germany) coupled to a triple stage quadrupole mass spectrometer (API 2000, AB Sciex, Darmstadt, Germany) with electrospray ionization (ESI) interface. Twenty microliters of the extract were injected and separated on a TSKgel ODS100 V column (150 × 4.6 mm; particle size 5 μm; Tosoh Bioscience). The temperature of the column was set to 25 °C, and the flow rate was set to 0.2 mL/min. Two different eluent systems were used in order to evaluate the chemical nature of the formed products. At acidic conditions (pH 2), eluent A consisted of 0.1% formic acid in water/ACN (v/v = 95/5) and eluent B contained 0.1% formic acid in water/ACN (v/v = 5/ 95). At basic separation conditions (pH 8), the eluent A comprises of aqueous ammonium hydroxide solution and B 100% acetonitrile. At both eluent systems the gradient was as follows: 0 min 20% B, 1 min 20% B, 20 min 80% B, 22 min 100% B. The ESI source was operated at 200 °C and the spray capillary voltage was set to 4 kV (ESI pos). The quadrupole scanned from 150 to 650 u in 1 s. The collision energy applied for product ion scans was 30 V. Additional precursor ion scan analyses were used to confirm or identify the molecular ions of the metabolites. Negative ESI was found to be less efficient for metabolite identification due to low signal intensity and interferences by matrix molecules (e.g., carbohydrates). Quantification of the amount of metabolites relative to FLU applied initially was performed using the signal intensities of the respective (M + H)+ ion traces extracted from the full scan analysis of the samples. Limit of detection at 0.06 ng/mL was calculated for the determination of FLU by HPLC-MS. The mean precision of the method was ±9% (at 1 ng/mL) for 5 replicates at two days. To support metabolite identification, analyses with a UPLC-QTOF mass spectrometer were performed, using an Agilent 1290 series LC system and an Agilent 6530 Accurate-Mass QTOF-mass spectrometer with electrospray ionization (Agilent Technologies). Five microliters of the extracts were injected by the autosampler and the analytes were separated on a ZORBAX Eclipse Plus C18 column (2.1 × 50 mm, particle size 1.8 μm,

For instance, Trametes versicolor transforms the fluoroquinolone antibiotics ciprofloxacin and norfloxacin primarily via ringopening at the piperazine moiety.22 However, the 3-ringannulated molecule of FLU differs remarkably from other fluoroquinolone structures; thus, different transformation pathways are also expected. Williams et al. have reported that the fungus Cunninghamella elegans can transform FLU into 7hydroxyflumequine and 7-oxoflumequine.23 Our investigation combines a comparative biodegradation study with the elucidation of metabolite structures and potential toxicological consequences. A broader range of fungal strains was evaluated in respect to their ability to transform the rather persistent FLU. Liquid chromatography−mass spectrometry (LC-MS) was used to detect and identify metabolites and to provide insight into the degradation pathway of the different strains. Moreover, the residual antimicrobial activity toward selected gram-positive and gram-negative bacteria was determined to verify whether fungal transformation of FLU reduces antibacterial activity and thus the risk to develop drug resistant bacteria in the environment.

■

EXPERIMENTAL SECTION Chemicals. The analytical standard of flumequine (IUPAC 7-fluoro-12-methyl-4-oxo-1-azatricyclo[7.3.1.0 5,13 ]trideca2,5,7,9(13)-tetraene-3-carboxylic acid, CAS No. 42835-25-6) of 99% purity was obtained from Sigma Aldrich (Steinheim, Germany). All the solvents were purchased from Merck (Darmstadt, Germany) and Chromservis (Prague, Czech Republic) and were of p.a. quality, trace analysis quality or gradient grade. D-(+)-glucose of p.a. quality and malt extract broth were obtained from Sigma Aldrich (Steinheim, Germany). A stock solution of FLU (2 mg/mL in DMSO) was used to spike the malt extract−glucose growth medium. Degradation Experiment. FLU was degraded by selected white rot fungi (Irpex lacteus 617/93, Panus tigrinus 577.79, Dichomitus squalens CCBAS 750, Trametes versicolor 167/93, and Pleurotus ostreatus 3004 CCBAS 278) in malt extractglucose medium as described elsewhere.20 Briefly: Fungal cultures were prepared from an inoculum seven days old which was spiked by 100 μL of FLU stock solution. The initial FLU concentration in the growth medium was 12 μg/mL. The cultures were cultivated in 20 mL liquid media (in 250 mL Erlenmeyer flasks) in the dark at 28 °C under static conditions. Sampling was performed after 3, 6, 10, and 14 days. The degradation process was validated by heat killed and abiotic controls. The residual concentrations of FLU in the liquid medium were determined by liquid chromatography (HPLC). All experiments were performed as triplicates. Quantitative Analysis by HPLC. HPLC consisting of a 2695 Separations Module (Waters, Milford, MA), a 2996 diode-array detector and a column X-Bridge C18 (250 mm ×4.6 mm; particle size 3.5 μm; Waters) was used for quantitative analysis of FLU after degradation. Ten μL of the growth liquid medium were directly injected by an autosampler. Separation was achieved with a gradient program, using (A) 0.1% trifluoroacetic acid in ACN/water (v/v = 10/90) and (B) 100% ACN. The elution program started with 30% (B) and was followed by a linear gradient to 100% (B) in 10 min. The column oven temperature was set to 25 °C and the flow rate to 0.7 mL/min. FLU was identified on the basis of the UV spectrum with maxima at 232 and 323 nm. Retention time of FLU was determined at 7.9 min. Calibration curve was built in the concentration range 0.1−50 μg/mL. Limit of detection was 14129

dx.doi.org/10.1021/es403470s | Environ. Sci. Technol. 2013, 47, 14128−14136


Article

Figure 1. Residual concentrations of flumequine in the liquid media during biodegradation by different ligninolytic fungi over 14 days. HK: Heat killed control.

Agilent Technologies) at a flow rate of 0.8 mL/min and a temperature of 25 °C. At isocratic conditions, the acidic eluents as mentioned above were applied with 20% B. Full scan data were acquired in the mass range from 100 to 500 u at 1.02 s at positive electrospray mode as the corresponding Product ion spectra. Instrumental parameters were as follows: capillary spray voltage 3500 V, nebulizer 138 kPa, drying gas 10 L/min, gas temperature 325 °C, skimmer voltage 60 V, octapole rf peak 750 V, collision energy 20 V. Simultaneous mass calibration in each analysis was realized by adding the ES TOF reference mix (Agilent Technologies) to the spray. Mass accuracy varied by ±2−3 ppm. Preparation of the Methyl- and Ethylesters of FLU. Preparation of the methyl- and ethylesters of FLU was performed in methanolic and ethanolic HCl (each at 6 N) solution, respectively. After heating at 60 °C for 30 min, evaporation to dryness and reconstitution in ACN/water (v/v = 20:80, each 0,1% formic acid), the yields detected with HPLCMS were 75% of FLU methyl ester and 50% of FLU ethyl ester. Residual Antibacterial Activity. Residual antibacterial activity of FLU together with its transformation products was evaluated with a variety of environmental gram-positive and gram-negative bacteria (Bacillus subtilis CCM 1999, Rhodococus erythropolis 2595, Citrobacter koseri CCM 2535, and Serratia marcescens CCM 303). The residual activity was determined using the Kirby-Bauer disk diffusion susceptibility test.24 The plates with nutrient agar were inoculated by 0.5 mL of bacteria suspension. Optical density of the suspension at 650 nm ranged 0.28−0.33. After the agar surface became dry, five holes with a diameter of 10 mm were made into the agar. Each hole was spiked with 200 μL of the sample and the plates were incubated 20 h at 37 °C. The resistance of the bacteria was evaluated according to the size of the sharply marginated zone of bacterial growth around the holes.

Data Analysis. The results of the biological tests were evaluated by principal component analysis (PCA) and t test using Minitab 16 version 2.2.0 (Minitab, Inc., PA, U.S.A.).

■

RESULTS AND DISCUSSION Removal of FLU from Liquid Media. Strains of five different ligninolytic fungi were exposed to FLU at initial concentration of 12 μg/mL. The residual concentration of FLU was determined directly from the solution by HPLC/UV after 3, 6, 10, and 14 days of cultivation (Figure 1). Abiotic degradation was not observed over the cultivation period and adsorption of FLU to the mycelium was found to be negligible which was proved by analyzing the exposed, heat killed mycelium after extraction. As Figure 1 indicates, FLU was removed rapidly by T. versicolor, D. squalens, and I. lacteus after 14 days of cultivation. Within 3 days T. versicolor transformed more than 90% of FLU, followed by I. lacteus and D. squalens which required 6 days for removing >90% of FLU. The slowest removal of FLU was observed in cultures with P. tigrinus and P. ostreatus where 8% and 2% remained after 10 days. Previous investigations on the biodegradation of the fluoroquinolones ciprofloxacin and norfloxacin by T. versicolor22,23 indicated a comparable time frame (10 days) for the removal above 90%. The biodegradation of FLU carried out by Williams et al. 23 using the filamentous fungus Cunninghamella elegans found that 77% of the initially applied FLU was removed within 7 days of cultivation. Identification of Main Metabolites. After 4 days of cultivation with the ligninolytic fungi, the extracts of the liquid media were analyzed by LC-MS/MS (ESI+). Eight major metabolites of FLU were detected in the different cultures. The general approach for metabolite identification used (i) LC retention order, (ii) the retention time shift at different pH of the eluent, (iii) exact mass data obtained by LC-QTOF-MS, (iv) product ion spectra, and (v) the number of exchangeable 14130



Article

Table 1. Metabolites Formed from FLU by Five Different White Rot Fungi Determined by UPLC-TOF-MS/MS (ESI pos)a metabolite

tR (min)

formula [M + H]+

calculated mass (m/z)

experimental mass (m/z)

error (ppm)

DBE

FLU

4.6

C14H13FNO3 C14H11FNO2 C11H5FNO2 C10H5FNO

262.0879 244.0774 202.0299 174.0350

262.0893 244.0792 202.0298 174.0341

1.3 1.8 −0.1 −0.9

9 10 10 9

F1

1.1, 1.2

C14H13FNO4 C14H11FNO3 C14H9FNO2 C11H5FNO3

278.0829 260.0717 242.0617 218.0248

278.0771 260.0669 242.0586 218.0228

−5.8 −4.8 −3.1 −2.0

9 10 11 10

F2b

2.1

C15H15FNO3 C14H11FNO2 C11H5FNO2 C10H5FNO

276.1030 244.0768 202.0299 174.0350

276.1036 244.0777 202.0229 174.0259

0.6 0.9 −7.0 −9.1

9 10 10 9

F3

0.5

C15H15FNO4 C14H11FNO3 C11H5FNO3

292.0985 260.0723 218.0248

292.0971 260.0708 218.0203

−1.4 −1.5 −4.5

9 10 10

F4b

4.0

C16H17FNO3 C14H11FNO2 C11H5FNO2

290.1192 244.0774 202.0299

290.1107 244.0760 202.0293

−8.5 −1.4 −0.6

9 10 10

F5

3.4

C16H17FNO3 C14H13FNO C12H9FNO C11H7FNO

290.1187 230.0976 202.0662 188.0506

290.1182 230.0962 202.0641 188.0492

−0.1 −1.6 −2.1 −1.4

9 9 8 9

F6

1.1

C14H15FNO2 C14H13FNO C11H7FNO

248.1081 230.0976 188.0506

248.1013 230.0943 188.0500

−6.8 −3.3 −0.6

8 9 9

F7

1.9

C14H13FNO2 C11H7FNO2

246.0925 204.0455

246.0926 204.0453

0.1 −0.2

9 9

F8

4.1

C16H19FNO2 C14H13FNO C12H9FNO C11H7FNO C10H7FN

276.1394 230.0976 202.0662 188.0506 160.0557

276.1354 230.0960 202.0645 188.0492 160.0547

−4.0 −1.6 −1.7 −1.4 −1.0

8 9 8 9 8

a Abbreviation, retention time, exact mass of the protonated molecule (underlined) and of the product ions, difference between measured and calculated mass (error), and double bond equivalent (DBE). bStructures confirmed by reference compounds.

protons obtained by H-D exchange experiments. Mass spectrometric data and structures proposed for these metabolites are provided in Table 1 and Figure 2. The respective product ion spectra of the molecular ions of the metabolites are shown as Supporting Information (SI). Unfortunately, attempts to isolate the metabolites in amounts high enough for NMR analysis failed but the methyl- and ethyl ester of FLU were synthesized as reference compounds for structure confirmation. FLU and its metabolite F1 reacted very sensitively with pH changes during LC separation and responded with longer retention times under acidic conditions. Peak tailing and broadening of the peaks at basic conditions have already been described for other fluoroquinolones,25 which reflects the amphoteric nature of FLU and its metabolite F1. The separation of the other metabolites was not influenced by varying pH-conditions suggesting that these compounds lost

the amphoteric character by transformation processes, such as esterification or biochemical reduction of the carboxylic group (Figure 2). The two chromatographic signals at 19.5 and 19.7 min showed identical mass spectra with molecular ions at m/z 278. Both substances possess two exchangeable protons (SI Figure S1, S3, and S4) and were identified as hydroxyflumequine diastereomers named metabolite F1 (Figure 2). Product ion spectra suggest a hydroxylation at position C7 or C6 of flumequine (SI Figure S3). Further structural isomers were not detected which implies a specific reaction pathway probably supported by the fluorine substitution at position C-9, which directs the OH attack to position C-7. This structure proposal agrees with the formation of a diastereomeric pair of 7-hydroxyflumequine during the biotransformation of FLU by C. elegans.23 The structures of these metabolites were confirmed by NMR spectroscopy. 14131



Article

Figure 2. Proposed transformation of flumequine with the different ligninolytic fungi (related from data at day 4 of cultivation).

tection mechanism of the organism. The availability of a methylating system can also prevent futile redox cycling of peroxidases and allows an oxidation of substances.21 Metabolite F3 consists of two diastereomers with retention times at 14.1 and 14.2 min and a molecular ion of m/z 292.0971. F3 was produced only by I. lacteus and represents probably the methyl ester of hydroxyflumequine F1. The fragmentation pattern points to the loss of methanol as detected also for F2 (SI Figure S6). One exchangeable proton (spectrum not shown) supports the proposed structure of F3 (Figure 2). Metabolite F4 with a molecular ion at m/z 290.1107 was produced by all fungal strains investigated. The lack of exchangeable protons and a longer retention time than the methyl ester of FLU (F2) supports the assumption that a larger moiety than methyl was introduced. In the product ion spectrum of the molecular ion, the neutral loss of 46.0347 u corresponds to the loss of ethanol (46.0413 u), analogue to the loss of methanol observed for F2 and F3 (SI Figure S7a). Thus, the ethyl ester of FLU was proposed as structure of F4 which was finally confirmed by a synthesized reference (SI Figure S7b). An esterification during sample preparation can be excluded because the extracts of the corresponding controls were free of both methyl and ethyl esters of FLU. Esterification to ethyl esters has previously been reported for the yeast strain Saccharomyces cerevisiae that transforms fatty acids to the corresponding ethyl esters.30 Metabolites F5, F6, and F8 were generated by D. squalens, only. These three metabolites are characterized by similar key fragments at m/z 230 (C 14 H 13 FNO) and m/z 188 (C11H7FNO) (exact masses see Table 1, SI Figures S8 and

Another study approved that hydroxyflumequine is preferentially excreted when calves were treated with FLU.2 All fungal strains produced the hydroxyflumequine (F1) thus, it seems to represent a general transformation pathway of the white rot fungi tested (Figure 2). Indeed, hydroxylation is a common reaction initiated by ligninolytic fungi and has been frequently reported, for example, polycylic aromatic hydrocarbons,26 polychlorinated biphenyls,27 and 17α-ethinylestradiol.28 The presence of hydroxylated compounds suggests the involvement of the monooxygenase system of cytochrome P-450 in the transformation reactions.29 Metabolite F2, another dominant product formed by I. lacteus, is characterized by a molecular ion at m/z 276.1036 and a neutral loss of 32.0259 u in the product ion spectrum. This was interpreted as elimination of methanol (exact mass 32.0262 u) suggesting that F2 was produced by methylation of the carboxylic acid of FLU (SI Figure S5). This view is supported by pH-independent chromatographic retention and the lack of exchangeable protons in F2. Williams et al. 23 identified in their experiments with C. elegans a product with molecular ion at m/z 276 (no exact mass reported) as 7-oxoflumequine, the oxidation product of 7hydroxy-flumequine. However, the product ions reported of that compound (m/z 258 and 216) were not detected for metabolite F2 in our investigations. Moreover, the measured mass of m/z 276.1036 (Table 1) differed by +0.036 u from the calculated mass for 7-oxoflumequine thus, the methylester of FLU is the more probable structure which was finally confirmed with the synthesized reference. In general, esterification decreases water solubility of polar compounds and in case of an exposure to potentially toxic compounds, methylation can also be considered an autopro14132



Article

Table 2. Relative Amounts of the Metabolites Produced by Individual Fungal Strains after 4 Days of Cultivationa

a

metabolite produced by

FLU

F1_I

F1_II

F2

F3_I

F3_II

F4

F5

F6_I

F6_II

F7

F8

I. lacteus P. tigrinus T. versicolor D. squalens P. ostreatus

8.8 74.2 n.d. 10.0 97.0

8.0 13.6 36.9 12.6 0.4

n.d. 8.2 31.8 11.8 0.4

79.1 0.5 n.d. 0.8 n.d.

1.3 n.d. n.d. n.d. n.d.

1.1 n.d. n.d. n.d. n.d.

0.3 0.9 0.8 0.8

Biotransformation of the Antibiotic Agent Flumequine by Ligninolytic

Recommend Documents