In-stream attenuation of neuro-active pharmaceuticals and their

Transcription

1 In-stream attenuation of neuro-active pharmaceuticals and their metabolites Supporting Information Analytical Methods and Figures (10 pages) Jeffrey H. Writer, Ronald C. Antweiler, Imma Ferrer, Joseph N. Ryan, E. Michael Thurman Supporting Information Table SI Table 1. Measured concentrations (ng L -1 ) of neuro-active compounds in Boulder Creek. Supporting Information Figures Figure SI1. Locations of sampling points on Boulder Creek, CO upstream and downstream from the City of Boulder wastewater treatment plant (WWTP). Figure SI2. Targeted neuro-active compounds and major metabolites, compound name followed by diagnostic ions, underlined ion used for quantification; DiOH-CBZ = 10,11,-dihydro-10,11,- dihydroxycarbamazepine, 10-OH-CBZ = 10-hydroxycarbamazepine. Figure SI3. Concentration profiles of neuro-active pharmaceuticals and associated metabolites in Boulder WWTP effluent after being passed through Boulder Creek bed sediment material based on composite samples, concentrations corrected to represent the Lagrangian parcel; error bars representing standard error for parent compounds based on triplicate analyses at BC2; error bars for metabolites and Citalopram correspond to analytical variability relative to concentration at BC2. Figure SI4. Column breakthrough curves; C/C 0 represents measured concentration divided by initial concentration in effluent; no breakthrough was observed for fluoxetine, citalopram, and N- desmethyl-citalopram. SI-1

2 Analytical. All samples collected for trace metal and cation analysis were analyzed in triplicate on both an ELAN-DRC II inductively coupled plasma-mass spectrometry (ICP-MS) and an Optima 3300 DV inductively coupled plasma-atomic emission spectrometry (ICP-AES). Details of the analytical settings and protocols can be found in Taylor (2001), Mitko and Bebek (1999), and Garbarino and Taylor (1994). Gadolinium and boron were analyzed in triplicate on the ICP-MS. Samples for chloride were analyzed in duplicate on a Dionex DX600 ion chromatograph. Accuracy and precision for all of the elements in this study were determined by standard reference water samples provided by the U.S. Geological Survey. The following neuro-active pharmaceutical compounds were initially targeted for analysis (bupropion, hydroxy-bupropion, carbamazepine, fluoxetine, lamotrigine, and venlafaxine) using high-performance liquid chromatography/tandem mass spectrometry (LC/MS/MS) as previously described (Ferrer and Thurman, 2010; Ferrer and Thurman, 2012). Briefly, each water sample was spiked with the labeled internal standard d 10 -carbamazepine (Cambridge Isotope Laboratories, Andover, MA.), concentrated using solid phase extraction (SPE), target compounds eluted with methanol, evaporated to 0.5 ml, and injected (20 µl) onto a high performance liquid chromatography system (HPLC) connected to a triple quadrupole mass spectrometer (Model 6460; Agilent Technologies, Santa Clara, CA). To identify neuroactive compounds not initially targeted for analyses, extracts previously analyzed by LC/MS/MS were subsequently injected (within 1 d) onto an HPLC system connected to a time-of-flight mass spectrometer (LC-TOF-MS, model 6520, Agilent Technologies, Santa Clara, CA). The following compounds were identified by LC-TOF-MS; erythro-bupropion, threohydrobupropion, 10,11,-dihydro-10,11,-dihydroxycarbamazepine (DiOH-CBZ), 10-hydroxy- SI-2

3 carbamazepine (10-OH-CBZ), citalopram, N-desmethylcitalapram (DMCit), 2-N-glucuronidelamotrigine (Gluc-LMG), oxcarbazepine, and O-desmethyl-venlafaxine (DMV). Compounds were identified by evaluation of the accurate mass (± m/z) of prominent peaks, diagnostic ions extracted from the full scan LC-TOF-MS chromatogram, and fragment ions indicative of neuro-active metabolites (Figure SI2). Matrix effects on ion response will be similar between samples due to Lagrangian sampling. After initial identification, authentic standards were used to verify retention times and mass spectra; bupropion, carbamazepine, DiOH-CBZ, 10-OH-CBZ, citalopram, and gabapentin, purchased from Santa Cruz Biotechnologies (Santa Cruz, CA); fluoxetine, lamotrigine, oxcarbazepine, DMV (purchased from Sigma Aldrich (St. Louis, MO) verified identification by LC-TOF-MS analysis. Authentic standards were not readily available for the metabolites threo-hydrobupropion, erythro-bupropion, DMCit, Gluc-LMG. Compounds identified by LC/MS/MS (bupropion, carbamazepine, lamotrigine, fluoxetine, and venlafaxine) were quantified by a seven point calibration curve using authentic standards and the internal standard d 10 -carbamazepine. Metabolites and compounds not initially identified by LC-MS/MS were quantified using methods that slightly varied based on availability of authentic standards. Concentrations of DiOH-CBZ, 10-OH-CBZ, Gluc-LMG oxcarbazepine, and DMV were calculated relative to parent compound concentrations determined by LC- MS/MS, using LC-TOF-MS diagnostic ion areas for the metabolite and parent compound relative to the internal standard d 10 -carbamazepine, accounting for differences in ion response based on analysis of individual standards in a surface water sample collected upstream from the WWTP to compensate for effects of matrix suppression. Relative concentrations of erythrohydrobupropion and threo-hydrobupropion were determined by a single point calibration comparing LC-TOF-MS diagnostic ion areas to the buproprion concentration determined using SI-3

4 LC-MS/MS. Because standards for erythro-hydrobupropion and threo-hydrobupropion were not available, we assumed the response factor for each the metabolite ions and parent ion were similar. Concentrations of citalopram, DMCit, lamotrigine, were determined by a single point calibration comparing LC-TOF-MS diagnostic ion areas relative to the internal standard d 10 - carbamazepine, accounting for differences in ion response based on analysis of individual standards in an uncontaminated surface water sample. The method detection limit was defined as the lowest concentration of the chemical that yielded minimum ion signal-to-noise ratios of 3:1 for both the quantitation and the confirmatory ions. To evaluate the effectiveness of the LC-MS/MS and LC-TOF-MS methods, water samples (n=2, LC-MS/MS; n=1, LC-TOF-MS) were spiked with authentic standards (40 ng L -1 LC-MS/MS, 80 ng L -1 LC-TOF-MS ), and the concentrations determined; recovery ranged from %, mean 96% for LC-MS/MS and %, mean 81% for LC-TOF-MS. Analytical variability for neuro-active compounds quantified by LC/MS/MS (bupropion, carbamazepine, fluoxetine, lamotrigine, and venlafaxine) was evaluated by replicate analyses consisting of one sample analyzed in duplicate, and one sample analyzed in triplicate and averaged 3% (Table SI1). Analytical variability for neuro-active metabolites and compounds not quantified by LC- MS/MS (erythro-hydrobuprion, threo-hydrobupropion, 10-OH-CBZ, DiOH-CBZ, oxcarbazepine, citalopram, DMCit, Gluc-LMG, DMV) was evaluated by analyzing one sample in duplicate and averaged 16%. Targeted compounds were not detected in any of the blank samples. SI-4

5 References Garbarino, J.R. and Taylor, H.E. Inductively coupled plasma-mass spectrometric method for the determination of dissolved trace elements in natural water. U.S. Geological Survey Open-File Report , 1995, 88 p. Mitko, K. and Bebek, M. ICP-OES determination of trace elements in salinated water. Atomic Spectroscopy, 1999, 20, Taylor, H.E. Inductively Coupled Plasma-Mass Spectrometry: Practices and Techniques; Academic Press, San Diego, 294 p., Ferrer, I.; Thurman, E. M. Identification of a new antidepressant and its glucuronide metabolite in water samples using liquid chromatography/quadrupole time-of-flight mass spectrometry. Anal. Chem. 2010, 82, Ferrer, I.; Thurman, E. M. Analysis of 100 pharmaceuticals and their degradates in water samples by liquid chromatography/quadropole time of flight mass spectrometry. J. Chromatogrphy A. 2012, 1259, SI-5

6 SI Table 1. Measured concentrations (ng L -1 ) of neuro-active compounds in Boulder Creek. Measured Discharge (m 3 s -1 ) t med a (h) Bupropion b Hydrobupropion c Hydrobupropion c Carbamazepine b erythro- threo- 10-hydroxy- Carbamazepine c DiOH-CBZ c Oxcarbazepine c Citalopram c N -desmethyl- Citalopram c Fluoxetine b Lamotrigine b 2-N -Glucuronide LMG c Venlafaxine b O -desmethyl- Venlafaxine c Method Detection Limit (ng L -1 ) Method Reproducibility (%) d ND e November 2, 2010 sampling BC US NM NA <5 <10 <10 <5 <10 <10 <10 <10 <10 <5 <5 <100 <5 <10 Effluent at 10:25 f 0.47 NA 193 ± 4 42 ± ± ± ± ± ± ± ± ± ± ± ± 51 BC 100 m DS NM NM BC3 (4.05 km) NM NM May 4, 2011 Tracer Study BC-US 0.47 g NA <5 <10 <10 <5 <10 <10 <10 <10 <10 <5 5 <100 <5 <10 Effluent at 5: NA Effluent at 9: NA Effluent at 13: NA theoretical mixing point (56 % Effluent) h 1.07 NA BC1 (2.29 km DS) (-0.25) 100 ± 4 i ± 2 i ± 0.3 i 1232 ± 52 i ± 1 i 57 BC2 (3.28 km DS) (-0.53) BC3 (4.05 km DS) (-0.02) BC4 (5.43 km DS) (+0.12) Summer 2011 Sampling j BC-US ( ) NM NA <5 <10 <10 <5 <10 <10 <10 <10 <10 <5 5 <100 <5 <10 BC-US (7-1-11) NM NA <5 <10 <10 <5 <10 <10 <10 <10 <10 <5 <5 <100 <5 <10 BC US (7-7-11) NM NA <5 <10 <10 <5 <10 <10 <10 <10 <10 <5 <5 <100 <5 <10 Effluent ( ) NM NA interference k Effluent (7-1-11) NM NA Effluent (7-7-11) NM NA Nov 1, 2011 Column Study Effluent at 13:00 NM NA Effluent (after 5 d) NM NA April 4, 2012 Sampling BC-US NM NA <5 <10 <10 <5 <10 <10 <10 <10 <10 <5 <5 <100 <5 <10 Effluent at 9: NA Stream Biofilm (ng kg -1 ) NA NA 4200 NM NM NM NM NM a values in paranthesis indicate time in hours that center of composite sample was offset from t med (negative values corresond to earlier time); b compound quantified by LC/MS/MS; c compound quantified by LC-TOF-MS; d method reproducibillity defined by analyses of field replicates; e quant. ion interference preculded determination of duplicate sample f average of field duplicates ± relative difference; g upstream gage measurement, less water diverted by City of Lafayette; h based on discharge and concentration measurements at 9:00; i average of field triplicates ± standard deviation j date of sampling in parenthesis; k parent ion detected but interference with confirmatory ion precludes quantitation; BC = Boulder Creek; Effluent = wastewater effluent; US = Boulder Creek upstream from effluent outfall; DS = Boulder Creek downstream from effluent outfall; DiOH-CBZ = 10,11 dihydro, 10,11 dihydroxy-carbamazepine; NA = not applicable; NM = not measured. SI-6

7 Figure SI1. Map of study area. Composite sample sites (black dots) were: Boulder Creek 50 m upstream from the WWTP effluent outfall (BC-US), the WWTP effluent outfall (EFF), BC km downstream, BC km downstream, BC km downstream, Dry Creek (DC), and BC km downstream. SI-7

8 Figure SI2. Targeted neuro-active compounds and major metabolites, compound name followed by diagnostic ions, underlined ion used for quantification; DiOH-CBZ = 10,11,-dihydro-10,11,- dihydroxycarbamazepine, 10-OH-CBZ = 10-hydroxycarbamazepine. SI-8

9 mol L Bupropion erythro-hydrobupropion threo-hydrobupropion Carbamazepine 10-OH-CBZ DiOH-CBZ Oxcarbazepine Citalopram N-desmethyl-Citalopram Fluoxetine Lamotrigine 2-N-Glucuronide Lamotrigine Venlafaxine O-desmethyl-Venlafaxine Travel Time (h) Travel time (h) Figure SI3. Concentration profiles of neuro-active pharmaceuticals and associated metabolites in Boulder WWTP effluent after being passed through Boulder Creek bed sediment material based on composite samples, concentrations corrected to represent the Lagrangian parcel; error bars representing standard error for parent compounds based on triplicate analyses at BC2; error bars SI-9

10 for metabolites and Citalopram correspond to analytical variability relative to concentration at BC Br Venlafaxine DMVX Fluoxetine and others C/C Bupropion erythro-bp threo-bp Lamotrigine 2N-gluc-LMG Carbamazepine 10-OH-CBZ DiOH-CBZ Oxcarbazepine Approximate Pore Volumes Figure SI4. Column breakthrough curves; C/C 0 represents measured concentration divided by initial concentration in effluent; no breakthrough was observed for fluoxetine, citalopram, and N-desmethyl- Citalopram. SI-10

11 SI-11