ARTICLE pubs.acs.org/est
Elucidation of the Transformation Pathway of the Opium Alkaloid Codeine in Biological Wastewater Treatment Arne Wick,† Manfred Wagner,‡ and Thomas A. Ternes*,† † ‡
Federal Institute of Hydrology (BfG), Koblenz, Germany Max Planck Institute for Polymer Research (MPI-P), Mainz, Germany
bS Supporting Information ABSTRACT: Codeine, an opium alkaloid, was transformed in aerobic batch experiments with activated sludge into several transformation products (TPs). For eight TPs, the chemical structures were unambiguously identified by a multistep approach using results from high-resolution mass spectrometry (HR-MS) and 1D and 2D nuclear magnetic resonance (NMR) experiments. For an additional 10 TPs, tentative structures were proposed. Most of the TPs identified exhibited only slightly modified molecular structures featuring double bond shifts, introduction of hydroxy groups, or amine demethylation. The transformation pathway of codeine in contact with activated sludge is characterized by a combination of biologically and chemically mediated reactions. Biological oxidation of codeine leads to the formation of the R,β-unsaturated ketone codeinone, which is the precursor for further abiotic and biotic transformation due to its high chemical reactivity. An analytical method based on solid-phase extraction and LC tandem MS detection was developed to confirm the formation of several TPs in wastewater treatment plants (WWTPs). The mass balances were comparable to those obtained from batch experiments. An HR-MS screening approach of TPs from dihydrocodeine and morphine revealed that the knowledge from the codeine transformation pathway can be extrapolated to the distinct transformation pathways of these structurally related opium alkaloids. In total, 17 TPs were proposed for morphine and 2 TPs for dihydrocodeine.
’ INTRODUCTION The natural opium alkaloid codeine is highly consumed, either by prescription or as an illicit drug. In Germany, approximately 2.4 t of codeine are annually used as an analgesic, mostly in combination with nonopioid substances or as an antitussive. Codeine has been detected in wastewater influents at a maximum concentration of 540 ng L1 in Germany and up to 10 μg L1 in Wales.1,2 Ecotoxicity data of opium alkaloids is rare but indicates a potential impact on the immune system of vertebrates.3 While sorption was shown to be of minor importance, biological transformation was identified to remove about 80% of the codeine load during biological treatment of a WWTP with a sludge age of 18 d.4 For the majority of the organic trace contaminants, microbial transformation does not lead to mineralization, but rather to the formation of a multitude of transformation products (TPs).5 These TPs might have a similar or in some cases an enhanced ecotoxicological potential relative to that of the parent compound.6,7 In the past, the elucidation of biotransformation and chemical transformation of micropollutants at environmental relevant concentrations was extremely time-consuming or even r 2011 American Chemical Society
impossible due to the limited analytical capabilities. In recent years, sophisticated hybrid mass spectrometry systems (e.g., triple quadrupole linear ion trap mass spectrometers, Qq-LITMS; quadrupole time-of-flight mass spectrometers, Qq-TOF-MS; linear ion trap quadrupole Fourier transformation mass spectrometers, LTQ-FT-MS) have been shown to be powerful tools for the structural identification of TPs.810 For example, Medana et al.11 identified 23 photochemical TPs of atenolol using the capability of a LTQ-Orbitrap-MS for analyzing the fragmentation pathways of the TPs by high-resolution MSn experiments. Kosjek et al.5 identified a TP of diclofenac and a TP of clofibric acid formed in activated sludge bioreactors using Qq-TOF-MS. Multistep analytical approaches are frequently crucial for the successful identification of TPs. Nuclear magnetic resonance (NMR) is one option for a complementary technique for identification and confirmation of TP structures as long as Received: October 15, 2010 Accepted: March 3, 2011 Revised: February 19, 2011 Published: March 23, 2011 3374
dx.doi.org/10.1021/es103489x | Environ. Sci. Technol. 2011, 45, 3374–3385
Environmental Science & Technology
ARTICLE
Table 1. Overview of the Different Batch Experiments compounds
spike conc
codeine
200 mg L1
matrix
duration
diluted sludge
35 days
(1:20, 0.2 gSS L1) codeine, codeine TPsa
200 μg L1
analysis HPLC-UV, Qq-LIT-MS, LTQ-Orbitrap-MS, NMR
7 days
diluted sludge
LC-Qq-LIT-MS (SIM mode)
(1:10, 0.38 gSS L1) TP 298 (codeinone)
5 mg L1
application isolation and identification elucidation of transformation pathway
diluted sludge
10 days
(1:20, 0.2 gSS L1), autoclaved diluted
LC-Qq-LIT-MS (full scan and SIM mode), DOC measurement
sludge, groundwater
quantification, mass balance, differentiation between chemical and microbial transformation
and ultrapure water 2 μg L1
codeine and TP 298 (codeinone)
70 h
diluted sludge (1:20, 0.18 gSS L1),
morphine, dihydrocodeine,
0.5 mg L1
hydrocodone
LC-Qq-LIT-MS (SIM mode, enrichment via SPE)
quantification, occurrence and mass
undiluted sludge
balance at environmental-
(3.6 gSS L1)
relevant concentrations
diluted sludge
6d
(1:20, 0.2 gSS L1),
LC-LTQ-Orbitrap-MS (full scan and MS2)
autoclaved diluted
screening for codeinelike TPs, identification of further TPs
sludge a
Codeine TPs used: TP 300(1), TP 314, TP 316, and TP 332.
isolation of sufficient quantities can be accomplished. The identification of a total of 46 microbial TPs of iodinated X-ray contrast media12,13 and 6 ozonation products of β-lactam antibiotics14 by multinuclear NMR experiments in addition to Qq-LIT-MS and LTQ-Orbitrap-MS, respectively, showed the successful application of NMR techniques for TP identification. The objective of this study was to identify TPs of codeine formed in laboratory batch experiments with activated sludge under aerobic conditions. Identification was accomplished using a multistep analytical approach including HR-MS and 1D and 2D-NMR techniques. Furthermore, we investigated the transformation pathway of codeine as well as the transferability of the results to full-scale wastewater treatment plants (WWTPs) and to other structurally related opium alkaloids such as morphine and dihydrocodeine.
’ EXPERIMENTAL SECTION Chemicals. Codeine and dihydrocodeine were purchased from Th. Geyer (Renningen, Germany), and hydrocodone was purchased from Sigma-Aldrich (Schnelldorf, Germany). Codeine-d6, used as the surrogate standard, and the analytical reference standards codeinone, 14-hydroxycodeinone, and pseudomorphine were purchased from LGC Promochem (Wesel, Germany). 14-Hydroxycodeine was purchased from Campro Scientific (Berlin, Germany). Reference standards of TP 300(1), TP 302, TP 332(1), TP 264, TP 346, and TP 348 were not commercially available, and therefore were isolated and purified as described below. The purity of these TPs were assessed by HPLC-UV and exceeded 90%. Information about the solvents used within the current study are provided in the Supporting Information (SI). Batch Systems with Activated Sludge. For the batch experiments, sludge was taken from the nitrification zone of
the activated sludge tank of a conventional German WWTP. The WWTP has a designed capacity of 320 000 population equivalents (PE) and a daily flow rate of 60 000 m3. The activated sludge unit is characterized by a hydraulic retention time (HRT) and a sludge retention time (SRT) of 6 h and 12 d, respectively. Details about the treatment processes at this WWTP can be found in Wick et al.15 Fresh sludge [total suspended solids (TSS), ∼4 gSS L1; total organic carbon (TOC), ∼0.3 g gSS1] was transported to the laboratory, and experiments were initiated on the same day. In general, the sludge was diluted with groundwater, free of all target compounds, to minimize the impact of sorption and analytical interference with matrix components. The sludge was continuously stirred in 500 mL amber gas wash bottles and flushed with a mixture of air and CO2 to establish aerobic conditions and a stable pH of 7.0 ( 0.2 (see SI). Depending on the specific application, different concentrations (between 2 μg L1 and 200 mg L1) of the respective target compounds were separately spiked into the sludge slurry after an equilibration time of 1 h (Table 1). The duration of the experiments was chosen in a way that at least 80% of the initially spiked concentration was transformed. Samples were taken just before and at specified time intervals after spiking. For direct injection, samples were filtered through 0.45 μm syringe filters made of regenerated cellulose (Spartan, Whatman, U.S.A.), and formic acid was added to a final concentration of 0.2% (see SI). The samples were kept frozen (25 °C) until analysis. In those cases where enrichment was necessary (i.e., batch systems with a spiking concentration of 2 μg L1), 50 mL samples were taken and acidified to pH 3 with 3.5 M sulfuric acid, filtered through glass fiber filters (GF/6, Whatman) and stored at 4 °C. The samples were subjected to solid-phase extraction (SPE) as described below within 24 h after sampling. Nonspiked sludge samples were always run in parallel. For certain experiments, spiked autoclaved diluted sludge as well as 3375
dx.doi.org/10.1021/es103489x |Environ. Sci. Technol. 2011, 45, 3374–3385
Environmental Science & Technology autoclaved groundwater and ultrapure water were used to confirm the occurrence of abiotic transformation processes. Samples from the autoclaved controls were taken with sterile pipet tips, and the gas stream used for aeration was directed through 0.2 μm sterile syringe filters to avoid any microbial contamination during incubation. An overview of the details about the different batch experiments is given in Table 1. Environmental Samples from WWTPs. Influent (after primary clarification) and effluent samples were collected from four conventional WWTPs (WWTP 14) to determine the formation of codeine TPs during wastewater treatment. In addition, grab samples of activated sludge were taken from the nitrification zone of the activated sludge tank of WWTP 1 to determine the sorbed concentrations. Details about each WWTP are provided in the SI. All samples were taken during dry weather periods in amber, solvent-rinsed glass bottles, acidified to pH 3 with 3.5 M sulfuric acid, filtered through glass-fiber filters (GF/6, Whatman), and stored at 4 °C until SPE. The activated sludge samples were centrifuged, and the aqueous phase was treated in the same manner as the wastewater samples. The solid phase was freeze-dried, spiked with 50 ng of the surrogate codeine-d6, and extracted with a Dionex ASE 200 instrument (Sunnyvale, CA) according to a procedure described in Wick et al.15 TP Isolation via Semipreparative HPLC-UV. After the concentrations of codeine were reduced by more than 90%, the batch samples were centrifuged and filtered through 1 μm glass fiber filters (GF/6, Whatman) and 0.45 μm cellulose sodium nitrate filters (Sartorius, G€ottingen, Germany). The filtrate was frozen and concentrated to a final volume of 2050 mL by freezedrying. To fractionate and collect the TPs, a Waters HPLC-UV system consisting of a Waters 717 plus autosampler, column oven, Waters 600 controller with quaternary pump, in-line degasser, and Waters 2487 dual-wavelength absorbance detector (operated at 213 and 280 nm) was used. Isolation of individual TPs was achieved by chromatographic separation on a semipreparative Synergi Polar-RP column (250 mm 10 mm, 4 μm) from Phenomenex (Aschaffenburg, Germany). Details about the chromatographic conditions are provided in the SI. Individual fractions were collected with an automated sample collector (Advantec SF-2120 super fraction collector, Techlab GmbH, Erkerode, Germany) on the basis of the retention time of the peaks in the chromatogram. The composition and purity of the fractions were determined by HPLC-UV and LC tandem MS. If the purity of a TP was found to exceed 90% in the collected fraction, it was freeze-dried to complete dryness. Otherwise, the volume of the solution was reduced by freeze-drying and purified again using HPLC-UV. The isolated TPs were used as analytical reference standards and for identification by mass spectrometry and NMR analysis. Identification of TPs via Mass Spectrometry. Qq-LIT-MS. Determination of the nominal masses as well as MS2 and MS3 spectra of codeine and its isolated TPs were performed on a QqLIT-MS (MDX Sciex 4000 Q Trap, Applied Biosystems, Langen, Germany) by direct injection using a syringe pump (10 μL min1). Electrospray ionization (ESI) was used in the positive ionization mode. For MS2 fragmentation, the collision energy was set to 40 eV, while for MS3 fragmentation different excitation energies in the range 20200 mV were applied. For the mass spectrometric analysis of nonisolated codeine TPs, the Qq-LITMS was interfaced with an Agilent HPLC system (Agilent 1200 Series, Agilent Technologies, Waldbronn, Germany). Chromatographic conditions and ESI source conditions were similar to
ARTICLE
those used for the analysis of samples from the batch experiments and WWTPs (see below). LTQ-Orbitrap-MS. The mass spectrometric information obtained from Qq-LIT-MS experiments were supplemented with high-resolution MS analyses using a LTQ-Orbitrap-MS (LTQ Orbitrap Velos, Thermo Scientific, Bremen, Germany) with ESI operated in positive and negative ionization mode. Measurement of exact masses (mass accuracy
