rates and product identification for trenbolone acetate...

RATES AND PRODUCT IDENTIFICATION FOR TRENBOLONE ACETATE METABOLITEBIOTRANSFORMATION UNDER AEROBIC CONDITIONS

EMILY A. COLE,y SAMANTHA A. MCBRIDE,y KAITLIN C. KIMBROUGH,y JAEWOONG LEE,y ERIC A. MARCHAND,yDAVID M. CWIERTNY,z and EDWARD P. KOLODZIEJ*xk

yDepartment of Civil Environmental Engineering, University of Nevada, Reno, Nevada, USAzDepartment of Civil and Environmental Engineering, University of Iowa, Iowa City, Iowa, USAxInterdisciplinary Arts and Sciences, University of Washington, Tacoma, Washington, USA

kDepartment of Civil and Environmental Engineering, University of Washington, Seattle, Washington, USA

(Submitted 8 June 2014; Returned for Revision 31 July 2014; Accepted 25 February 2015)

Abstract: Trenbolone acetate metabolites are endocrine-active contaminants discharged into the aquatic environment in runofffrom agricultural fields, rangelands, and concentrated animal feeding operations. To investigate the environmental fate of thesecompounds and their biotransformation mechanisms, the authors used inocula from a variety of different water sources and dosedbiologically active microcosms with approximately 1400 ng/L of trenbolone acetate metabolites, including 17b-trenbolone, trendione,and 17a-trenbolone. To investigate aerobic biotransformation rates and interconversions between known trenbolone acetate metabolites,gas chromatography–tandem mass spectrometry was used to measure concentrations and assess product distributions as a function oftime. High-resolution liquid chromatography–tandem mass spectrometry (LC-MS/MS) was used to characterize novel transformationproducts and potential transformation pathways. Kinetic analysis yields observed half-lives of approximately 0.9 d, 1.3 d, and 2.2 d for17b-trenbolone, trendione, and 17a-trenbolone, respectively, at 20 8C, although colder conditions increased half-lives to 8.5 d andbiphasic transformation was observed. Relative to reported faster attenuation rates in soils, trenbolone acetate metabolites are likelymore persistent in aqueous systems. Product distributions indicate an enzymatic preference for biotransformation between trendione and17b-trenbolone. The LC-MS/MS characterization indicates dehydrogenation products as themajor detectable products and demonstratesthat major structural elements responsible for bioactivity in steroids are likely retained during biotransformation. Environ Toxicol Chem2015;34:1472–1484. # 2015 SETAC

Keywords: Trenbolone acetate Product characterization Aerobic biotransformation Rate constant

INTRODUCTION

Agricultural runoff and municipal wastewater effluentare major contributors to the occurrence of endocrine-active steroidal contaminants capable of inducing potentialadverse effects on aquatic organisms in surface waters [1–8].Given its large geographic footprint and limited treatment,agricultural runoff may be of special concern as a source ofendogenous and exogenous steroids. For example, most beefcattle in the United States are implanted with trenboloneacetate, a potent synthetic steroid and anabolic growthpromoter [9]. Trenbolone acetate metabolites excreted by cattleinclude 17a-trenbolone (representing >95% of the knowntrenbolone acetate metabolite mass), 17b-trenbolone, andtrendione, although >90% of the trenbolone acetate implantdose remains uncharacterized [10,11]. Thus, in addition tonormal excretion of endogenous steroids from vertebrates,animal agriculture can act as point or nonpoint sources ofmetabolites of synthetic pharmaceuticals (such as trenboloneacetate) to aquatic environments.

Given their potency, which is usuallymuch higher than that ofendogenous steroids, steroid pharmaceuticals may be especiallyimportant contributors to ecosystem risks. Exposure to trenbo-lone acetate metabolites can significantly reduce fecundity infish at 10 ng/L to 30 ng/L concentrations for 17a-trenbolone and

17b-trenbolone, respectively; and trenbolone acetatemetabolitescan induce partial to complete sex reversal at similar concen-trations [9,12,13]. Trenbolone acetate metabolites can alterfemale fish behavior at low concentrations [14], and genotoxictrends in fish populations also are reported, although atsubstantially higher exposure concentrations [15].

Synthetic steroids usually are more environmentally persis-tent than their endogenous counterparts [4,6] and often areselected specifically for attributes such as reduced potential forin vivo enzymatic degradation (thus increasing their bioactivi-ty), which also may affect their environmental fate. Data forendogenous estrogens and androgens indicate that steroid-degrading microorganisms utilize highly site- and ring-specificenzymes for biotransformation, a selectivity that can limittheir utility for synthetic steroids whose structures divergefrom those of endogenous steroids at key reaction cen-ters [16,17]. Mass balance calculations suggest that up to80 mg/d of 17a-trenbolone is excreted per implanted animal,potentially leading to concentrations of several thousandnanograms per liter in agricultural runoff and indicating theimportance of subsequent environmental fate processes ingoverning their ecological risk [11,18].

The available biotransformation data for trenbolone acetatemetabolites are limited to agricultural soils and suggest half-lives near 4 h to 12 h for 17a-trenbolone and 17b-trenbolone,with longer (1–4 d) half-lives observed for trendione [18–21]. Incontrast to expectations derived fromMichaelis-Menten kineticmodels, some biotransformation data suggest that trenboloneacetate metabolite transformation rates exhibit concentrationdependence, demonstrating slower transformation rates at

All Supplemental Data may be found in the online version of this article.* Address correspondence to [email protected] online 28 February 2015 in Wiley Online Library

(wileyonlinelibrary.com).DOI: 10.1002/etc.2962

Environmental Toxicology and Chemistry, Vol. 34, No. 7, pp. 1472–1484, 2015# 2015 SETAC

Printed in the USA

1472

higher concentrations [19,21]. Also, similar to data observed forestrogens, biologically mediated reactivity within families ofmetabolites, such as interconversions between more potent 17-hydroxy and less potent 17-keto steroids mediated byhydroxysteroid dehydrogenase, can preserve the steroidbackbone and conserve bioactive functional groups despitetransformation [22]. Because biotransformation data in aqueoussystems or characterizations of transformation products beyond17a-trenbolone, 17b-trenbolone, or trendione are lacking, theselimitationsmerit additional investigation of trenbolone acetate’smetabolite fate.

To address these limitations, the objective of the presentstudy was to use kinetic modeling to determine biotransforma-tion rates of trenbolone acetate metabolites in aerobic batchmicrocosms inoculated from different sources. Two differentincubation temperature regimes and sampling campaigns wereconducted to estimate the dependence of kinetic model resultson seasonal and temperature variation. We also characterizedthe potential for interconversions between known trenboloneacetate metabolites to occur and used liquid chromatography–high-resolution tandemmass spectrometry (LC-HRMS/MS) foradditional biotransformation product identification andcharacterization.

EXPERIMENTAL PROCEDURES

Chemicals

Trenbolone acetate metabolites (17a-trenbolone, 17b-trenbolone, and trendione) were purchased from Steraloids.We used 17b-trenbolone-16,16,17-d3 as an isotopic standardfor quantitation (BDG Synthesis). High-performance liquidchromatography (HPLC)–grade methanol, acetone, anddichloromethane were purchased from Fisher Scientific. Allother chemicals were purchased from Sigma Aldrich.Following published protocols, N-methyl-N-(trimethylsilyl)-trifluoro-acetamine and iodine (99.999% pure) were usedto derivitize samples prior to gas chromatographic(GC) analysis [18,23]. Fluorescein diacetate (3060-diacetyl-fluorescein), fluorescein disodium salt, and sodium pyrophos-phate were used in microbial activity measurements(see Supplemental Data and Adam and Duncan [24]).Glassware was silanized with dichlorodimethylsiloxane. Toaddress nutrient limitations, a phosphate-buffered mineralmedium was prepared, which contained 0.85 g/L KH2PO4,2.175 g/L K2HPO4, 3.34 g/L Na2HPO4 �H2O, 0.05 g/LNH4Cl, 0.275 g/L CaCl2, 0.225 g/L MgSO4 � 7H2O, and0.002 g/L FeCl3 � 4H2O.

Inocula sources and sampling

All studies used to quantify biodegradation constants usedinocula gathered from Steamboat Creek (39825008.2500N,119844025.9100W), a mixed-use urban agricultural watershedin Reno, Nevada, USA. Inocula were gathered in April 2012 andJanuary 2013, respectively, representing seasonally dependent“active” and “dormant” inocula from the same site. Studiesexamining the possible influence of inocula source ontransformation product formation were conducted by collectinginocula from 1) wastewater effluent from the secondaryclarifiers at the Truckee Meadows Water Reclamation Facilityin Sparks, Nevada; 2) the Sparks Marina in Sparks, Nevada; 3)the Truckee River at Mayberry Park in Reno, Nevada(39830009.3000N, 119853049.9500W); 4) a pond in IdlewildPark in Reno, Nevada (39831017.6900N, 119849055.9000W);and 5) Steamboat Creek.

Mixed culture microbial inocula were obtained by agitatingthe top 3 cm of sediment at the sediment–water interface andcollecting the resulting high-turbidity solution. Inocula wereimmediately transported to the laboratory and allowed toacclimate to experimental conditions for up to 12 h prior to use.Inocula were characterized by measuring total organic carbon(TOC) and total carbon using a Shimadzu Total Organic CarbonAnalyzer (TOC-VCSH). Inorganic carbon was calculated as thedifference between total carbon and TOC.

Microcosm design

Biotransformation studies were conducted as batch micro-cosms (in triplicate) in 73-mL amber glass vials filled with a 1:1mineral medium and inocula mixture. To allow for mixing andgas exchange, each microcosm had approximately 1 cm ofheadspace with approximately 65 mL to 68 mL liquid volume.Trenbolone acetate metabolites (17a-trenbolone, 17b-trenbo-lone, or trendione) were added to microcosms by spiking 100mL of a 1-g/L trenbolone acetate metabolite standard inmethanol to obtain nominal concentrations of 1400 ng/L for allkinetic studies. Control microcosms were prepared identically,though autoclaved (121 8C, 15 psi) for 2 h and cooled to roomtemperature prior to trenbolone acetate metabolite addition [19].Microcosms were loosely covered with aluminum foil andincubated in the dark as they were gently shaken (100 rpm) topromote gas exchange and maintain aerobic conditions at 20 8Cin a climate-controlled lab. Sample evaporation was negligibleat these experimental conditions. Seasonal microcosms usingJanuary inocula also were incubated at 5 8C in a walk-inrefrigerator or at 35 8C in a large incubator to investigatetemperature effects.

Experiments employed sacrificial analysis of microcosms at0 d, 1 d, 3 d, 6 d, 10 d, and 15 d after incubation. For analysis,microcosms were spiked to 100 ng/L with 17b-trenbolone-d3 asan isotopic standard; then, the pH of each sample was recorded,and 3 mL were removed from each sample for qualitativesterility analysis using fluorescein diacetate hydrolysis analy-sis [24]. Next, the remaining approximately 65 mL of eachsample were extracted onto preconditioned (following manu-facturer protocols) Restek C-18 solid phase extraction car-tridges [18]. If necessary, solid-phase extraction cartridges werestored at –18 8C prior to processing. Additional experimentaldetails and results are found in Supplemental Data.

Gas chromatography–tandem mass spectrometry (GC-MS/MS)analysis

The GC-MS/MS analysis of trenbolone acetate metabolitesis described in detail elsewhere [18,23]. Briefly, elution of solid-phase extraction cartridges was performed with a 95:5 (v/v)methanol:water mixture in three 3-mL aliquots at 1 mL/min to5 mL/min. The eluent was dried in a vacuum oven at 25 8C,resuspended in 1 mL of a 95:5 (v/v) dichloromethane:methanolsolution, transferred to silanized GC vials, and dried withnitrogen. Derivitization was performed by adding 50 mL N-methyl-N-(trimethylsilyl)-trifluoro-acetamine-I2 (1.44 mg I2/mL N-methyl-N-[trimethylsilyl]-trifluoro-acetamine) to eachvial, vortexing, and then drying under N2. Samples wereresuspended in 100 mL N-methyl-N-(trimethylsilyl)-trifluoro-acetamine, capped prior to incubation (40 min) at 60 8C, thencooled and analyzed by GC-MS/MS with isotopic dilutionanalysis. At these microcosm volumes, the method limit ofquantification was 14 ng/L, and quality assurance and qualitycontrol metrics (near quantitative spike recovery, precisiongenerally within 10%, no false positives detected in blanks) are

Rates and product identification for trenbolone Environ Toxicol Chem 34, 2015 1473

similar to those we reported in detail for concurrentstudies [11,25,26]. Samples that failed either to meet this limitof quantification or to meet a signal-to-noise ratio of 10 wereexcluded from quantitative analysis. Control microcosms wereexamined for sterility and mass conservation; data sets withoutsterile controls were excluded from analysis.

Transformation products

To facilitate identification, microcosms for transformationproduct studies were sometimes spiked to 5 mg/L or 20 mg/L.All other experimental conditions and sample processing wereidentical to those described above. Samples (2 mL in amberglass HPLC vials) collected for LC-HRMS/MS analysis wereusually analyzed directly by partially evaporating (�90%)samples with a gentle stream of nitrogen gas, then addingHPLC-grade methanol to make a 90:10 (v/v) methanol:watersolution for analysis. In addition to direct analysis, solid-phaseextraction (described above) was sometimes employed toconcentrate samples where methanolic extracts were adjusted tothe same 90:10 (v/v) methanol:water composition. The LC-HRMS/MS analysis employed an LTQ-Orbitrap XL (Thermo-Electron) mass spectrometer to characterize products usingconditions and criteria described elsewhere [27,28].

Kinetic modeling

Tomaintain consistency and comparisonwith published datafor trenbolone acetate metabolites and other steroids, we appliednot only the standard pseudo–first order decaymodel [19,22,29–31] but also 2 biphasic models to estimate transformation ratesto account for observed biphasic characteristics. April andJanuary inocula data (t¼ 5 8C, 20 8C, and 35 8C) were used inthis biphasic analysis, which is an analysis similar to thosepublished by Stowe et al. [32,33] for polychlorinated biphenyl(PCB) kinetic rate modeling under conditions where residualcontaminants inhibited transformation rates. Biotransformationof the cyanobacterial hepatotoxinmicrocystin also has exhibitedbiphasic kinetics arising from the existence of multiplecommunities of microcystin-decaying microorganisms: oneutilizing rapid direct metabolism at high concentrations andanother responsible for slower cometabolism at low concen-trations [34–36]. Other examples of systems that diverge fromexponential decay or exhibit biphasic characteristics includebiotransformation of trenbolone acetate metabolites and estro-gens in soils [21,37], PCB transformation [32,33], and theMichaelis-Menten and Monod kinetic models.

Kinetic models included the pseudo–first order model(model 1); a “non-zero asymptote” model (model 2), whichassumes the presence of both a rapidly degrading contaminantpool and a persistent residual; and a “double-exponent” model(model 3), which assumes that the persistent residual alsodecays [32,33].

C ¼ C1e�k1t ð1Þ

C ¼ C2e�k2t þ C3 ð2Þ

C ¼ C4e�k4t þ C5e

�k5t ð3Þ

In the 3 models, C represents the observed metaboliteconcentration at time t and k1, k2, k4, and k5 all represent thecorresponding first-order rate constants for the different models.In model 1, C1 represents the initial metabolite concentration.For model 2, C2 represents the labile metabolite concentration,andC3 represents a residual recalcitrant concentration. In model

3, C4 and C5 represent the concentrations amenable to rapid orslow transformation, respectively. For models 2 and 3, summingC2 and C3 or C4 and C5 yields initial metabolite concentrations.

RESULTS AND DISCUSSION

Aerobic microcosms

Microbial activity measurements, including pH, TOC, andfluorescein diacetate analysis, indicate limited or no microbialactivity in controls relative to biologically active microcosmswhere microbial activity increased during the experiments(Figures 1D and 2D; Supplemental Data, Figures S1 and S2). Asa result of autoclaving, pH in control microcosms was generallyslightly higher than for active microcosms, with the differencein pH between aerobic and control samples being statisticallysignificant for data taken at 0 d, 1 d, 6 d, and 10 d afterincubation. The Steamboat Creek inocula contained approxi-mately 7 mg/L TOC and 13 mg/L inorganic carbon initially,which was subsequently diluted with mineral medium. Thus,TOC levels were initially low (<10 mg/L) in microcosms, butmethanol addition (100 mL in spikes) increased TOC levels to390� 18mg/L (Supplemental Data, Figure S1). Consistent withYang et al. [31], preliminary experiments conducted with andwithout glucose addition as a carbon source demonstrated nosignificant effect on transformation rates; thus, no manipulationof labile organic carbon was used.

For autoclaved controls with April inocula, 17a-trenbolone,17b-trenbolone, and trendione exhibited 85% to 100%, 77% to95%, and 69% to 110% recoveries of initial 1400 ng/Lconcentrations over 15 d, respectively (Figure 1). Observedslight deviations between nominal and observed concentrationsin controls are consistent with normal variations in methodperformance. Over the 15 d, 17a-trenbolone controls weregenerally stable, whereas controls for 17b-trenbolone andtrendione exhibited losses, up to 30% in one case, althoughusually smaller. Although sorption, abiotic transformation, orresidual enzymatic activity may explain slight losses in controls,the phenomenon was not observed in other trials. However it isquite consistent with similar, though unexplained, lossesobserved for trenbolone acetate metabolites in autoclavedsoil–water systems [19].

In biologically active microcosms spiked with 1400 ng/L17a-trenbolone at 20 8C (Figure 1A), 17a-trenbolone concen-trations in initial samples (extracted �2 h after trenboloneacetate metabolites were spiked) were not significantly differentfrom controls. Over the next 6 d, concentrations decreased to170 ng/L, with no further decreases apparent to 15 d. Lag phaseswere not observed for any trenbolone acetate metabolite,indicating that inherent biotransformation capabilities existedfor these steroids in the inocula. Consistent with results reportedfor estrogens and trenbolone acetate metabolites in soil systems,interconversions between trenbolone acetate metabolites wereimportant aspects of transformation [19,22,29,37]. In the 17a-trenbolone microcosms, trendione concentrations increased to170 ng/L over the first 3 d, corresponding to 12% of the massdosed, then decreased. 17b-Trenbolone was first observed at 3 d(36 ng/L) and increased to 89 ng/L at 15 d, a 7% mass yield. Nosimilar trenbolone acetate metabolites were observed in any17a-trenbolone controls, indicating that these interconversionswere biologically mediated.

In microcosms spiked with 1400 ng/L 17b-trenbolone(Figure 1B), a rapid 25% loss relative to controls apparentlyoccurred over the initial 2-h period after 17b-trenbolone spikingbut prior to extraction (processing period for 0 d samples). In the

1474 Environ Toxicol Chem 34, 2015 E.A. Cole et al.

same samples, 120 ng/L of trendione was present with nosimilar trendione detections observed in controls. Although notstatistically significant, similar trends of rapid initial attenuationrelative to controls were observed for 17a-trenbolone andtrendione microcosms. This phenomenon cannot be explainedby sorption to particles or other abiotic loss processes because itwas not observed in controls. The present data likely indicatethat enzymatic activity, such as 17-steroid dehydrogenase, caninduce rapid transformations (i.e., conversion of the 17b-hydroxy to a 17-keto) of 5% to 25% of the dosed mass over sub–2-h time scales; and they are consistent with results we haveobserved for steroid spikes in field studies [18]. Over the next3 d, 17b-trenbolone concentrations decreased to 120 ng/L andcontinued to slowly degrade to 14 ng/L by 15 d, representing a99% overall metabolite loss. Trendione concentrations were220 ng/L, a 16%yield, after 1 d and decreased to 34 ng/L by 15 d.

In biologically active samples spiked with 1400 ng/Ltrendione, concentrations decreased to 14 ng/L over 15 d(Figure 1C). Conversion to 17b-trenbolone (90 ng/L) occurred

within the first day, which then decreased slowly to no-detectconcentrations at 15 d. In these samples, 17a-trenbolonewas detected only once at 26 ng/L after 10 d of trendioneincubation. Together with the 17b-trenbolone data, theseobservations indicate that steroid transformations are dominatedinitially by interconversion between 17b-hydroxy and 17-ketospecies and likely exhibit steric or enzymatic inhibition forthe 17-keto to 17a-hydroxy conversion. These observationsare consistent with reports for estrogens, where preferences for17-keto to the 17b isomer instead of the 17a isomer pathway areobserved, and agree with trenbolone acetate metabolite data forsoils [19,29].

Because trendione conversion to 17a-trenbolone or 17b-trenbolone represents C17 reduction, these reactions shouldbe inhibited by aerobic conditions, whereas the reversereaction (converting 17a-trenbolone or 17b-trenbolone totrendione) might be favored by aerobic conditions. Whenfocusing on aerobic biotransformation of trenbolone acetatemetabolites in agricultural soils, Khan et al. [19] reported

Figure 1. Aerobic biodegradation of (A) 17a-trenbolone, (B) 17b-trenbolone, and (C) trendione at 20 8C with April inocula. Hollow symbols represent sterilecontrol samples (autoclaved). All samples were spiked to a nominal concentration of 1400 ng/L. (D) pH data corresponding to control samples from (A) through(C). (E) pH data corresponding to aerobic samples from (A) through (C). Error bars represent standard deviations of triplicate samples. TBA¼ trenboloneacetate; TBO¼ trendione; TBOH¼ trenbolone.


conversion of 17a-trenbolone and 17b-trenbolone into tren-dione at approximately 40% and 60% yields, respectively.Consistent with these expectations, although the differenceswere small, less 17a-trenbolone and 17b-trenbolone persistedin microcosms spiked with trendione compared with thetrendione concentrations observed after spiking with 17a-trenbolone and 17b-trenbolone (Figure 1). Relative to publishedsoil data, the lower maximum trendione yield observed (16%)implies that the aqueous conditions or inoculum used likelyenhanced subsequent trendione transformation to uncharacter-ized products.

Experimental conditions

In addition to a number of preliminary trials, selectexperiments were repeated using inocula collected duringdifferent seasons or from different source waters. For example,17a-trenbolone experiments were repeated with SteamboatCreek inocula collected in January. Biologically active micro-cosms with this inocula at 20 8C exhibited a 17a-trenbolonedecrease to 140 ng/L (90% loss) over 15 d (Figure 2). Maximumtrendione concentrations were 210 ng/L, a 12% yield, whereas

17b-trenbolone was detected only once, at 6 d (65 ng/L). Thepresent data indicate slower 17a-trenbolone transformationwith January inocula, but similar trends in overall loss andmetabolite interconversion were evident between all inoculacollected from this source.

Incubation temperature was varied in some microcosms. TheJanuary inocula described above also were incubated at either5 8C or 35 8C. In 5 8C samples, 17a-trenbolone concentrationsdecreased to 920 ng/L after 1 d and to 600 ng/L over 6 d with nofurther transformation evident after 6 d (Figure 2A). Trendionewas detected at 3 d and steadily increased to 370 ng/L at 15 d,corresponding to a 26% yield, although no corresponding 17a-trenbolone loss was observed. 17b-Trenbolonewas not detected,perhaps implying some temperaturedependence for the trendioneto 17b-trenbolone pathway. In 35 8C samples, 17a-trenboloneconcentration rapidly decreased to 740ng/L after 1 d (Figure 2C),which is statistically similar to the loss observed at 5 8C, thentransformationunexpectedly ceased. Trendionewas not detected,and 17b-trenbolone was observed only once at low concentra-tion, also consistent with the substantial inhibition of biotrans-formation at this experimental condition.

Figure 2. Aerobic biodegradation of 17a-trenbolone with January inocula at (A) 5 8C, (B) 20 8C, and (C) 35 8C. Hollow symbols represent sterile controlsamples (autoclaved). All samples were spiked to a concentration of 1400 ng/L. (D) pH data corresponding to (A) through (C). Error bars represent standarddeviations of triplicate samples. TBA¼ trenbolone acetate; TBO¼ trendione; TBOH¼ trenbolone.


Kinetic modeling

All 3models were evaluated with the data, and their accuracywas determined by comparing R2 values, which were calculatedby subtracting from 1 the ratio of the model’s residual sum ofsquares to the data’s total sum of squares. Half-lives are onlyreported formodel 1, and all othermodel parameters (C1–C5 andk1–k5) are presented in Table 1.

No trenbolone acetate metabolite exhibited completeremoval to no-detect levels in these 15-d experiments,particularly for 17a-trenbolone, the most common trenboloneacetate metabolite, whose concentrations stabilized (i.e., nofurther transformation was apparent) in many experiments.Biphasic behavior (i.e., variation in kinetic rates with time orconcentration) suggests that initial rates observed at higherconcentrations may not always accurately represent subsequentdata. The pseudo–first order decay model and associated half-life concepts implicitly assume that attenuation rates areindependent of the experimental time scale or concentrationpresent, and this assumption can result in overestimation oftransformation rates and underestimation of environmentalpersistence. In addition to inhibitions arising from multiplepools of degradable contaminants, system characteristics suchas sorption to particles, oxygen concentrations, and nutrientavailability can inhibit biotransformation [37], although we donot believe these factors explain the present data.

The double-exponent model exhibited the best fit, with 2exceptions: the 35 8C (17a-trenbolone) data set was betterestimated by the non-zero asymptote model, and the trendione(20 8C) data set was equally well estimated by all models.Models were evaluated with and without accounting forobserved slight losses in controls; this analysis indicated thatany control losses were negligible. Despite the goodness-of-fitfor the double-exponent model, in some cases the “residual”concentration (C3 or C5) was negligible relative to the labileconcentration, implying that the standard exponential decaymodel was appropriate (Table 1). As the double-exponentmodel often explained the data well, it is reasonable to

conclude that slow cometabolic or other biphasic processescan inhibit trenbolone acetate metabolite transformation atlow concentrations and sometimes a persistent residual ismaintained.

The present data clearly indicate that biotransformation ratesexhibit stereoselectivity for steroids. Similar to reported resultsfor 17a-estrogens [38], slower transformation is expectedfor 17a-trenbolone in particular. Under identical conditions, a2-fold difference in transformation rates was evident for 17a-trenbolone (0.27 d–1) compared with 17b-trenbolone (0.57 d–1),implying that the most common trenbolone acetate metabolite,17a-trenbolone, also is the most persistent. With respect tointerconversion, because trendione is the only intermediateproduct for 17a-trenbolone and 17b-trenbolone interconver-sion, rates of 17-keto reduction should dominate overall rates oftransformation between known high-potency trenbolone acetatemetabolites, consistent with reported results for other ste-roids [21,37,39,40]. Inocula collection also affected kinetics,likely because of different microbial community compositionand relative activity; kinetic rates for April inocula were nearlytwice the values for January inocula (0.27 d–1 vs 0.15 d–1).Similarly, slower rates (0.099 d–1) and concentration stabiliza-tion were observed during temperature manipulations. Thepresent data suggest approximately 2-fold changes in transfor-mation rates at temperature extremes, implying increasedpersistence under both colder (winter-like) conditions andwarmer (summer-like) conditions, and are consistent withtemperature-dependent results for estrogens [29] and trenboloneacetate metabolites in soils [21].

Half-lives in aqueous systems for 17a-trenbolone, 17b-trenbolone, and trendione (2.2–4.1 d, 0.9 d, and 1.3 d,respectively), which likely underestimate trenbolone acetatemetabolite persistence, were compared with reported half-lives in soils (0.18–0.33 d, 0.18–0.31 d, and 0.54–2.0 d,respectively) [19]. Although these comparisons are alwayscomplicated by differing experimental conditions, in general,dissipation rates in soils were faster than those observed foraqueous systems [19]. This difference may arise from larger,

Table 1. Biotransformation kinetic parameters estimated for models 1, 2, and 3

Trenbolone acetate metabolite Inocula Temp. (8C) C1 (ng/L) k1 (d�1) t1/2 (d) R2

Model 1: Pseudo–first order17a-Trenbolone January 2013 5 1400 0.081 8.5 0.26

20 1400 0.17 4.1 0.9035 1400 0.099 7.0 0.04

17a-Trenbolone April 2013 20 1400 0.32 2.2 0.5517b-Trenbolone 1400 0.78 0.9 0.77Trendione 1400a 0.54a 1.3a 0.99a

Model 2: Non-zero asymptote Inocula Temp. (8C) C2 (ng/L) k2 (day–1) C3 (ng/L) R2

17a-Trenbolone January 2013 5 750 0.43 650 0.6020 1340 0.15 60 0.8835 730 2.0 670 0.74

17a-Trenbolone April 2013 20 1290a 0.38a 110a 0.57a

17b-Trenbolone 1220 0.57 180 0.53Trendione 1380a 0.55a 20a 0.99a

Model 3: Double-exponent Inocula Temp. (8C) C4 (ng/L) k4 (day–1) C5 (ng/L) k5 (day

–1) R2

17a-Trenbolone January 2013 5 400a 11a 1000a 0.036a 0.64a

20 70a 8.6a 1330a 0.16a 0.91a

35 590a 2600a 810a 0.025a 0.79a

17a-Trenbolone April 2013 20 1330 0.32 70 0.32 0.5517b-Trenbolone 1400a 0.78a 0.2a 0.015a 0.77a

Trendione 1360a 0.56a 40a 0.12a 0.99a

aThese parameters reflect the best-fit model for a given data set.C1¼ initial metabolite concentration; C2¼ labile metabolite concentration; C3¼ residual recalcitrant concentration; C4¼ concentration amenable to rapidtransformation; C5¼ concentrationamenable to slow transformation; k1, k2, k4, k5¼ correspondingfirst-order rate constant for thedifferentmodels; t1/2¼ half-life.


more diverse, and more active microbial communities in soilsfacilitating transformation and the contributions of sorption andabiotic mineral-promoted transformation to dissipation rates insoils. Approximately 20% of any steroid mass would typicallyremain dissolved in a typical soil–water subsurface system [20],and the higher soil transformation rates suggest that infiltrationof agricultural runoff can be an effective mechanism forenhancing attenuation rates.

Transformation products

To characterize the long-term fate of trenbolone acetatemetabolites, a major objective of the present experimentswas to investigate biotransformation products, as structuralconservation during transformation also implies the potentialconservation of bioactivity. Therefore, focusing largely on 17a-trenbolone and trendione, we investigated the presence of novelbiotransformation products and examined any effects of inoculasource on products. To facilitate the identification of potentiallow-yield products, some microcosms were spiked at higherconcentrations (up to 20mg/L), although solid-phase extraction,which concentrated solutions by approximately 35-fold, wassometimes used to facilitate detection. Also, without availablestandards for products and because of uncertainties arising frommatrix and structural effects on ionization efficiency, thisanalysis was not considered quantitative but qualitativeonly [27]. Because we cannot estimate yields, we largelyattempted to verify the presence of biotransformation productsand comparisons of concentration as a function of time werelimited to single products.

Consistent with expectations and published data, intercon-versions between 17a-trenbolone, trendione, and 17b-

trenbolone accounted for most of the detectable transformationproducts (Table 2). For 17a-trenbolone and 17b-trenbolonemicrocosms, trendione was usually the dominant detectableproduct. However, other transformation products wereevident, particularly mass-to-charge ratio (m/z) – 2-Da productsconsistent with microbially mediated dehydrogenation. Forexample, in an m/z 269 scan (consistent with trendioneformation at m/z 269.1530) of a 17a-trenbolone microcosm,we see not only the expected formation of trendione at9.15 min but also the formation of 3 other m/z 269 peaks inthe chromatogram, including the dominant (by area count)6.88-min peak, whose MS/MS spectra also were all consistentwith trendione fragmentation (Figure 3). Their mass accuracyand conservedMS/MS fragments indicate that all of these peakswere 17a-trenbolone transformation products. We note thatwith this analysis we cannot differentiatem/z 269 [MþH]þ ionsfrom the [MþH-H2O]

þ ions expected form/z 285monohydroxyproducts, although relative retention times (an indication ofpolarity) can be used to facilitate product assignments, withearlier elution times expected for hydroxy products [27].

By analogy to testosterone biotransformation productsobserved usingmanure-borne bacteria under aerobic conditions,where m/z 288 testosterone eluted at 9.02 min and themore polar m/z 286 1,2-dehydrotestosterone product eluted at7.82min [31],wepropose that themajor 6.88-minpeak (1.02minbefore 17a-trenbolone) is an [MþH]þ ion representing a17a-trenbolone 1,2-dehydrogenation product. Although it isimpossible to conclusively identify the structure without astandard or nuclear magnetic resonance analysis, 1,2-dehydro-genation of 3-keto steroids mediated by 1,2-dehydrogenase is acommon microbial transformation pathway observed in aquatic

Table 2. Characteristics of proposed biotransformation products of 17a-trenbolone, 17b-trenbolone, and trendione

CompoundRetention time

(min)Proposedproduct ion

Observedmass (Da)

Exact mass(Da) Comments

17a-Trenbolone (parent) 7.90 [MþH]þ 271.1684 271.1698 Orbitrap mass calibration consistently had a low biasof �0.0014 Da for accurate mass assessment

Trendione 9.15 [MþH]þ 269.1528 269.1542 Confirmed by pure standard, strong literaturesupport; preferred product

Trendione analog ordehydrogenation product

9.23 [MþH]þ 269.1528 269.1542 Unknown coeluting trendione analog or17a-trenbolone dehydrogenation product

Dehydrogenation product 6.88 [MþH]þ 269.1528 269.1542 Potential dehydrogenation productMonohydroxytrendione product 5.29 [MþH-H2O]

þ 269.1528 269.1542 or285.1491

We propose a monohydroxy structure of thedehydrogenation product above or amonohydroxytrendione product

Coeluting hydroxylated products Near 3–4 [MþH-H2O]þ 271.1684 287.1647 A family of early eluting, presumably mono- or,

more likely, dihydroxylated products17b-Trenbolone (parent) 8.15 [MþH]þ 271.1684 271.1698Trendione 9.15 [MþH]þ 269.1528 269.1542 Preferred productDehydrogenation product 6.82 [MþH]þ 269.1528 269.1542 Potential dehydrogenation productMonohydroxytrendione product 5.09 [MþH-H2O]

þ 269.1528 269.1542 or285.1491

Analogous product to 17a-trenbolone productsabove

Trendione (parent) 11.05 [MþH]þ 269.1528 269.1542 These analyses used a slower chromatographygradient for improved separation

17b-Trenbolone 9.15 [MþH]þ 271.1684 271.1698 Preferred product17a-Trenbolone 8.75 [MþH]þ 271.1684 271.1698 Minor relative to 17b-trenboloneTrenbolone dehydrogenationproduct

10.05 [MþH]þ 267.1371 267.1385 Dehydrogenation product, possible 1,2- or 6,7- or15,16-dehydro location; preferred and mostpersistent dehydro product of trendione

Trenbolone dehydrogenationproduct

6.75 [MþH-H2O]þ 267.1371 267.1385 Dehydrogenation product, possible 1,2- or 6,7- or

15,16-dehydro locationTrenbolone dehydrogenationproduct

13.64 [MþH]þ 267.1371 267.1385 Dehydrogenation product, possible 1,2- or 6,7- or15,16-dehydro location

17b-Trenbolonedehydrogenation product

8.07 [MþH]þ 269.1528 269.1542 Potential secondary dehydro product of17b-trenbolone formed from trendione

17b-Trenbolonedehydrogenation product

5.90 [MþH-H2O]þ 269.1528 269.1542 Potential secondary dehydro product of

17b-trenbolone formed from trendione


environments and is consistentwith expectations based on steroidmetabolism [16,31,41]. Other possibilities include 6,7- or 15,16-dehydrogenation [16]. As a result of its earlier elution time andbecause them/z 269 product at 5.29min often growswith time aslater elutingm/z 269 products decrease, we propose that the 5.29-min peak is an [MþH-H2O]

þ ion of a monohydroxy product ofeither trendione or the presumptive 1,2-dehydrotrenboloneproduct. In conjunction with the formation of electrophilicmoieties resulting from 1,2-dehydrogenation or 6,7-dehydroge-nation, subsequent hydroxylation in the A or B ring is mostreasonable. We also note that mass spectra for these m/z 269products are more similar to results reported for A ring– and Bring–hydroxylated metabolites of 17b-trenbolone and unlike theD ring–hydroxylated metabolites reported by Pottier et al. [42]for in vivometabolites in cattle. Although all 3 of these unknownm/z 269 products exhibit similar MS/MS fragments as trendione,relative abundances of MS/MS fragments are substantiallydifferent from those observed for trendione (Figure 4). Inparticular, the 9.23-min peak observed to coelute with trendionemay be a structural analog or isomer forming from trendione ormay represent yet another dehydrogenation product of 17a-trenbolone of similar polarity to trendione.

Because dehydration of monohydroxy photoproducts oftrenbolone is possible to elicit product-to-parent reversion [28],

we also compared retention times and MS/MS spectra of thesebiotransformation products to those observed for monohydrox-ytrendione photoproducts. This analysis suggests that anybiotransformation products were distinct from presumptive 5-hydroxytrendione, 10-hydroxytrendione, or 12-hydroxytren-dione photoproducts capable of reversion; and no evidence ofproduct-to-parent reversion was observed for these biotransfor-mation products. Besides the m/z 269 products, other 17a-trenbolone products observed include a family of coeluting m/z287monohydroxy ([MþH]þ ions) or dihydroxy ([MþH-H2O]

þ

ion) products, present in most samples near 3 min to 4min.Withthe exception of the m/z 269 peak coeluting with trendione,similar biotransformation products were observed for 17b-trenbolone; and these product characterizations also wereconsistent with a number of similar 17a-trenbolone biotrans-formation experiments conducted at several different concen-trations with different inocula sources, including those reportedbelow (Supplemental Data, Figure S2).

To improve separation, subsequent LC-HRMS/MS analysesemployed a slower chromatography gradient, lengtheningelution times by approximately 1.8 min for trendione. Thisanalysis was used for 17a-trenbolone and trendione micro-cosms (1400 ng/L, 20 8C) that used 5 different inocula sources,with a focus on trendione biotransformation because, similar to

4HR_17a_SPE 8/24/2011 1:38:49 PM

RT: 0.00 - 15.00 SM: 15G

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Time (min)

0

20

40

60

80

100R

elat

ive

abun

danc

eRT: 6.88AA: 1065819

RT: 9.15AA: 825991

RT: 5.29AA: 103353

RT: 3.62AA: 69964

RT: 7.93AA: 81920

RT: 13.32AA: 36693

RT: 0.50AA: 34947

RT: 10.02AA: 34158

RT: 2.12AA: 21547

RT: 6.25AA: 24744

RT: 11.87AA: 20930

RT: 14.40AA: 22785

RT: 8.68AA: 23684

RT: 6.47AA: 8642

RT: 7.44AA: 6113

NL:1.51E5m/z= 268.50-269.50 MS ICIS 4HR_17a_SPE

4HR_17a_SPE #751-765 RT: 6.83-6.93 AV: 4 SB: 20 6.13-6.84 SM: 7B NL: 2.39E5T: FTMS + p ESI Full ms [100.00-350.00]

120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330m/z

0

20

40

60

80

100

Rel

ativ

e ab

unda

nce

269.1530

270.1563141.8599 196.2328171.1491 185.2863 217.1793129.8300 163.1332 315.7353202.3139 259.6114224.2872 285.6797252.1945 322.5979297.3591

4HR_17a_SPE #758 RT: 6.87 AV: 1 NL: 5.41E4T: ITMS + c ESI d Full ms2 [email protected] [60.00-280.00]

120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330m/z

0

20

40

60

80

100

Rel

ativ

e ab

unda

nce

251.2

241.2

252.2195.2 223.2209.2 225.3187.3159.2 242.32.3813.541 173.2133.3121.3 269.3

x2

17A-Trenbolonebioreactor 4 h SPEm/z 269 scan

Full scanRT = 6.88 min peak

MS/MS scanm/z 269 parentRT = 6.88 min peak

Trendione

Figure 3. Representative liquid chromatogram, high-resolution mass spectrometric full-scan mass spectrum, and tandemmass spectrometric spectra for a solid-phase extraction extract from a biologically active 5000 ng/L 17a-trenbolonemicrocosm after 4 h of incubation. The peak at 9.15min is trendione and consistentwith the expected mass-to-charge ratio of 269.1530 mass, although we propose that the dominant peak at 6.88 min represents a [MþH]þ dehydrogenationproduct of 17a-trenbolone. In addition, we propose that the more polar 5.29-min peak represents a monohydroxy-dehydrogenation product ([MþH-H2O]

þ ion),and we also note the presence of a product coeluting with trendione at 9.23 min. m/z¼mass-to-charge ratio; MS/MS¼ tandem mass spectrometry;RT¼ retention time; SPE¼ solid-phase extraction.


estrone, it probably is the most stable and environmentallyrelevant long-term steroidal product. Over the 13-d trial period,trendione peak area decreased by only 51% in the least activeinocula (Truckee River) microcosms, to decreases of well over99% in the most active inocula (Truckee Meadows WaterReclamation Facility) composed of mixed liquor from theaeration basins at the regional wastewater-treatment facility. Infact, in the Truckee Meadows Water Reclamation Facilitysamples, about 50% of the trendione transformed within theinitial 2-h solid-phase extraction processing interval, suggestingthat very rapid and efficient intrinsic steroid transformationcapabilities exist in wastewater-treatment plants.

Consistent with GC-MS/MS results, interconversion to17b-trenbolone dominated trendione biotransformation, withonly limited 17a-trenbolone formation apparent. Usually, 17a-trenbolone peak areas were only 10% to 15% of 17b-trenbolonepeak areas (Supplemental Data, Figure S3). For example, after13 d, Idlewild inocula were dominated by 17b-trenboloneformation, with final 17b-trenbolone peak areas higher thanresidual trendione peak areas; and similar results were observedfor Truckee River inocula samples (Supplemental Data,Figure S2). In these 2 systems, limited overall trenboloneacetate metabolite degradation was evident over the 13 d, withpeak areas for trendione and 17b-trenbolone at d 13 accountingfor 83% to 96% of initial (0 d) trendione peak area. In contrast,after 2 d in the Truckee Meadows Water Reclamation Facility

samples, only trace amounts of any known trenbolone acetatemetabolites were evident. Sparks Marina and Steamboat Creekinocula were intermediate between these extremes. Efficienttransformation was typical for Steamboat Creek inocula, with<10% of the initial mass remaining after 13 d. Thesecomparisons suggest that the 1-d to 8-d half-lives estimatedfor trenbolone acetate metabolites with Steamboat Creekinocula may be among the faster rates expected for typicalaerobic receiving waters, although we anticipate that agricul-tural runoff will be especially microbially active because of itshigh suspended solids concentration.

Novel trendione biotransformation products were observedwith all inocula. In fact, every sample after day 0 contained 2distinctm/z 267.1371 products, 1 at 6.75min and 1 at 10.05min,consistent with trendione dehydrogenation and with dehydro-genation products observed for 17a-trenbolone and 17b-trenbolone (Figure 5). We also propose that these productsare 1,2-dehydrotrendione species. In fact, after 13 d in the veryefficient Truckee Meadows Water Reclamation Facility–inoculated microcosms, these m/z 267 dehydrogenationproducts were still readily detected at similar peak areas today 5 samples and were the only detectable transformationproducts. The present data may indicate that steroidaldehydrogenation products are more persistent than any knowntrenbolone acetate metabolites such as trendione, 17a-trenbo-lone, or 17b-trenbolone. Similar to results observed for

90HR_17a_SPE 8/24/2011 2:36:58 PM

RT: 0.00 - 15.00 SM: 15G

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Time (min)

0

20

40

60

80

100

Rel

ativ

e ab

unda

nce

RT: 7.01AA: 1755753

RT: 9.23AA: 933244RT: 5.67

AA: 361403

RT: 3.71AA: 80685

RT: 7.66AA: 56866

RT: 8.65AA: 56895 RT: 11.53

AA: 45403RT: 6.41AA: 27627

RT: 14.41AA: 25863

RT: 9.93AA: 38257

RT: 2.13AA: 19615

RT: 0.86AA: 17400

RT: 13.86AA: 23124

RT: 4.73AA: 22226

NL:2.63E5m/z= 268.50-269.50 MS ICIS 90HR_17a_SPE

90HR_17a_SPE #1020 RT: 9.21 AV: 1 SB: 58 8.49-9.17 NL: 7.35E3T: ITMS + c ESI d Full ms2 [email protected] [60.00-550.00]

120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330m/z

0

20

40

60

80

100

Rel

ativ

e ab

unda

nce

225.2

251.2

223.2226.2 233.3195.1 207.2

241.2 252.4209.2169.3 192.3 3.6912.1813.341135.3 159.2147.2 249.1 269.1129.3

90HR_17a_SPE #1032 RT: 9.32 AV: 1 NL: 1.09E4T: ITMS + c ESI d Full ms2 [email protected] [60.00-280.00]

120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330m/z

0

20

40

60

80

100

Rel

ativ

e ab

unda

nce

251.2

241.2225.2

252.3195.2 223.2197.2 209.2 226.3 233.3192.2185.4 242.2143.2 173.2157.2 254.3131.3 260.3

x2

17A-Trenbolonebioreactor 90 h SPEm/z 269 scan

MS/MS scanm/z 269 parent9.32 min time point

MS/MS scanm/z 269 parent9.21 min time point

Figure 4. Representative liquid chromatogram and tandem mass spectrometric (MS/MS) spectra for a solid-phase extraction extract from a biologically active5000 ng/L 17a-trenbolone microcosm after 90 h of incubation. The MS/MS spectra are presented both for the expected trendione product (mass-to charge ratio[m/z] 269.1530, 9.21 min MS/MS spectra) and for the coeluting structural analog or 17a-trenbolone dehydrogenation product (m/z 269.1530, 9.32 min MS/MSspectra). Note the conservation of MS/MS fragments for both products, although at different abundances. RT¼ retention time; SPE ¼ solid-phase extraction.


17a-trenbolone phototransformation, where a less polar, lateelutingm/z 271 photoproduct was observed [27], a thirdm/z 267product was observed at 13.64 min in the Truckee MeadowsWater Reclamation Facility samples only, 2.6 min after thetrendione parent eluted (Figure 5). Thus, biotransformation alsocan potentially result in the formation of less polar stereo-isomers or structural analogs, although the formation of morepolar products was favored. Although numerous rearrange-ments to less polar structures are possible conceptually [41], theliterature provides few clear explanations for products withthese characteristics.

Twom/z 269.1530 products were detected in some trendionemicrocosms (Figure 6). In light of the 17a-trenbolone results,we propose that these m/z 269 products, 1 at 8.07 min and 1 at5.90 min, are most consistent with [MþH]þ and [MþH-H2O]

þ

dehydrogenation products of 17b-trenbolone, respectively.These are likely secondary products formed as trendioneinterconverts primarily to 17b-trenbolone, followed by dehy-drogenation. More polar, poorly resolved hydroxylated prod-ucts also were observed.

Product yield is a major uncertainty in this analysis, and it iscertainly possible that these proposed products exist only atminor concentrations in these samples; but this is an issue wecannot easily address without standards. For most detections,

peak areas for dehydrogenation products were substantially lessthan those observed for trenbolone acetate metabolite intercon-version products, potentially suggesting low yields, althoughwenote that our limited analyses of pure standards of other similartrenbolone acetate metabolite transformation products haveexhibited poor ionization and very low peak areas with this LC-HRMS/MS analysis [27]. Thus, we have no real insight intopotential yields of these proposed products.

Environmental implications

Because 17a-trenbolone accounts for 95% of the totalexcreted trenbolone acetate metabolite mass, it is the mostenvironmentally relevant of the known trenbolone acetatemetabolites [10,11,42]. The present study reports highertrenbolone acetate metabolite persistence, with half-lives of afew days for 17a-trenbolone, in aquatic systems relative tostudies performed in agricultural soils, where half-lives of onlya few hours were reported, although we note that such resultscan be very system-specific [19]. Soil systems should thus bemore efficient at attenuating trenbolone acetate metabolitesrelative to the higher persistence and reduced transformationrates expected for dark, turbid aquatic systems excepting theefficient transformation expected in the photic zone [28,43].Comparing the known trenbolone acetate metabolites,

TMWRF_TBO_day_02 5/2/2013 12:47:10 AM

RT: 0.00 - 15.00 SM: 15G

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Time (min)

0

20

40

60

80

100R

elat

ive

abun

danc

eRT: 6.75AA: 702832

RT: 10.05AA: 266596RT: 0.84

AA: 207951RT: 1.99AA: 117531 RT: 4.35

AA: 146656 RT: 13.64AA: 98787RT: 3.43

AA: 45371RT: 8.73AA: 21679

RT: 7.63AA: 14419RT: 5.52

AA: 16727RT: 11.37AA: 11568

RT: 13.01AA: 3950

RT: 6.08AA: 3917

NL:1.01E5m/z= 266.50-267.50 MS ICIS TMWRF_TBO_day_02

TMWRF_TBO_day_02 #946-967 RT: 6.69-6.83 AV: 5 SB: 33 4.58-5.74 SM: 7B NL: 1.71E5T: FTMS + p ESI Full ms [200.00-350.00]

120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330m/z

0

20

40

60

80

100

Rel

ativ

e ab

unda

nce

267.1371

268.1405

285.1477 329.1925233.1742 305.1586245.1741223.1112 255.1560211.0749201.1117

266.1264

TMWRF_TBO_day_02 #952 RT: 6.74 AV: 1 NL: 1.36E4T: ITMS + c ESI d Full ms2 [email protected] [60.00-280.00]

120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330m/z

0

20

40

60

80

100

Rel

ativ

e ab

unda

nce

223.1

249.1

221.2 225.2 250.2195.2133.0 185.1 211.1 239.2157.1121.1 173.2147.0 159.1 258.8

x2

Trendione-TMWRFDay 2 SPE extractm/z 267 scan



Figure 5. Representative liquid chromatogram, high-resolution mass spectrometric full-scan mass spectrum, and tandem mass spectrometric (MS/MS) spectrafor a solid-phase extraction extract from a biologically active 1400 ng/L trendione microcosm after 2 d of incubation. Three mass-to-charge ratio 267.1371transformation products are observed; we propose that the 10.05-min peak is a [MþH]þ ion for a 1,2-dehydrotrendione product, whereas the 6.75-min peak is a[MþH-H2O]

þ ion for a monohydroxy species. The final peak at 13.64 min is present in only a few samples and represents another uncharacterized stereoisomeror structural analog. Note the conservation of MS/MS fragments for these products with other trenbolone acetate metabolite products, and all 3 products in thepresent study exhibited near identical full-scan and MS/MS spectra. m/z¼mass-to-charge ratio; RT¼ retention time; SPE¼ solid-phase extraction;TMWRF¼Truckee Meadows Water Reclamation Facility.


17a-trenbolone was the most persistent compound; and it isclear that stereochemistry governs transformation rates asrelatively small structural differences have unexpectedly largeeffects on contaminant fate. The present data are consistentwith other observations indicating unique fate outcomesfor 17a-trenbolone, as receptor binding alone would predictgreatly reduced potency for 17a-trenbolone relative to 17b-trenbolone [42], yet unexpectedly similar potency for 17a-trenbolone to 17b-trenbolone is observed in fish exposures [12].17a-Trenbolone exhibits far more rapid product-to-parentreversion kinetics from photoproducts [28], and both sorp-tion [20] and biotransformation (the present study) outcomesexhibit substantial stereoselectivity within trenbolone acetatemetabolites.

Metabolite interconversion patterns observed in well-studiedestrogen systems were usually consistent with the results weobserved for trenbolone acetate metabolites. Interconversionpathways for steroids are probably well conserved andwidespread, including a preference for C17 ketone to 17bconversion over 17a conversion [29,38]. These pathways arisefrom the activity of promiscuous, stereospecific enzymes inreceiving waters and soils that preferentially discriminateagainst 17a conversions [16,17] and likely similarly governthe transformations of the entire suite of potential steroidalcontaminants in the aquatic environment excepting stericallyhindered substrates such as ethinyl estradiol or progestins.

More importantly, although we cannot estimate productyields with our approaches, the present data indicate that anumber of closely related transformation products that conservekey steroidal structural attributes should be present in receivingwaters (Figure 7). In fact, many published studies imply theconservation of bioactivity, and even potential increases incertain aspects of bioactivity, after transformation of steroidsand other pharmaceuticals [12,16,28,29,31]. For example, Yanget al. [31] documented the formation of 1,2-dehydrotestoster-one, also known as boldenone, on the incubation of testosteronewithmanure-borne bacteria. This transformationwould actuallybe expected to conserve and even increase bioactivity [44], asthere exist many transformations that can retain and evenincrease certain aspects of bioactivity in steroids [45]. Thus,proposed metabolic pathways such as that presented in theenvironmental risk assessment for trenbolone which demon-strate hydrogenation and dehydrogenation products [41] shouldreally be interpreted more accurately in the context that thesestructural modifications do not necessarily reduce overallmixture bioactivity.

Observation of conserved structure may be particularlyimportant when considering the ecological relevance ofexposure to mixtures of closely related contaminants that retainkey functional groups and may play an important role inresolving any discrepancies observed between concurrent use ofchemical analysis and bioanalytical approaches for sample

Steam_TBO_day_13 5/1/2013 11:53:19 PM

RT: 0.00 - 15.00 SM: 15G

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Time (min)

0

20

40

60

80

100

Rel

ativ

e ab

unda

nce

RT: 11.05AA: 765809

RT: 5.90AA: 401287

RT: 8.07AA: 168745

RT: 2.31AA: 44854

RT: 0.84AA: 56707 RT: 12.82

AA: 6283RT: 6.74AA: 6074

RT: 11.37AA: 6555RT: 8.69

AA: 5883RT: 9.78AA: 4738

RT: 4.58AA: 9528

RT: 13.85AA: 7920

RT: 4.13AA: 5274

RT: 5.35AA: 4951

NL:9.13E4m/z= 268.50-269.50 MS ICIS Steam_TBO_day_13

Steam_TBO_day_13 #1121-1137 RT: 8.00-8.10 AV: 4 SB: 42 6.44-7.92 SM: 7B NL: 3.65E4T: FTMS + p ESI Full ms [200.00-350.00]

120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330m/z

0

20

40

60

80

100

Rel

ativ

e ab

unda

nce

269.1528

310.1251

226.1796

328.1356270.1562217.1044 244.1900 311.1285

255.0630 292.1146301.0698

227.1829208.1691 266.1719 315.0653201.0733 281.1349

Steam_TBO_day_13 #1127 RT: 8.05 AV: 1 NL: 4.57E3F: ITMS + c ESI d Full ms2 [email protected] [60.00-280.00]

120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330m/z

0

20

40

60

80

100

Rel

ativ

e ab

unda

nce

251.1

195.0 252.3225.2223.1197.2 241.3209.2 226.1185.1157.2 183.2173.2 256.5120.8 133.2 147.1 269.3

x2

Trendione-SteamboatDay 13 SPE extractm/z 269 scan



Figure 6. Representative liquid chromatogram, high-resolution mass spectrometric full-scan mass spectrum, and tandemmass spectrometric spectra for a solid-phase extraction extract from a biologically active 1400 ng/L trendione microcosm after 13 d of incubation. With this chromatography gradient, the peak at11.05 min is trendione, and we propose that the 8.07-min and 5.90-min mass-to-charge ratio 269.1528 peaks are consistent with [MþH]þ and [MþH-H2O]

þ

dehydrogenation products of 17b-trenbolone, respectively. m/z¼mass-to-charge ratio; MS/MS¼ tandem mass spectrometry; RT¼ retention time;SPE¼ solid-phase extraction.


analysis [46]. Fractionation approaches such as toxicityidentification evaluation used to identify bioactive constituentsin complex mixtures can fail to identify causative agents [47]and are limited by the necessity for sensitive analytical methodsdeveloped around the availability of pure standards. Ecologicalrisk and environmental hazard arise not only from contaminantpersistence but also from the relationships between potency andyields in both parents and products, and we can expect thatchemical analyses in isolation will often struggle to fully definethese relationships. The best holistic approach to quantifyingecological risks would thus employ bioanalytical tools toimprove the characterization of linkages between exposure,structure, and bioactivity in contaminants and their products.

SUPPLEMENTAL DATA

Figures S1–S3. (424 KB DOC).

Acknowledgment—We thank the US Department of Agriculture (grant2010-65102-20425) for funding the present study. We also thank D. Quilici(University of Nevada, Reno) and the Nevada Proteomics Center foranalytical instrumentation and assistance, as well as 3 anonymous reviewersfor their helpful suggestions.

Data availability—For researchers interested in any data associated with thepresent study, please contact E. Kolodziej at [email protected].

REFERENCES

1. Kidd KA, Blanchfield PJ, Mills KH, Palace VP, Evans RE, LazorchakJM, Flick RW. 2007. Collapse of a fish population after exposure to asynthetic estrogen. Proc Natl Acad Sci USA 104:8897–8901.

2. Purdom CE, Hardiman PA, Bye VVJ, Eno NC, Tyler CR, Sumpter JP.1994. Estrogenic effects of effluents from sewage treatment works.Chemistry and Ecology 8:275–285.

3. J€urgens MD, Holthaus KIE, Johnson AC, Smith JJL, Hetheridge M,Williams RJ. 2002. The potential for estradiol and ethinylestradioldegradation in english rivers. Environ Toxicol Chem 21:480–488.

4. Lange IG, Daxenberger A, Schiffer B, Witters H, Ibarreta D, MeyerHHD. 2002. Sex hormones originating from different livestockproduction systems: Fate and potential disrupting activity in theenvironment. Anal Chim Acta 473:27–37.

5. Soto AM, Calabro JM, Prechtl NV, Yau AY, Orlando EF, DaxenbergerA, Kolok AS, Guillette LJ Jr, le Bizec B, Lange IG, Sonnenschein C.

2004. Androgenic and estrogenic activity in water bodies receivingcattle feedlot effluent in eastern Nebraska, USA. Environ HealthPerspect 112:346–352.

6. Khan SJ, Roser DJ, Davies CM, Peters GM, Stuetz RM, Tucker R,Ashbolt NJ. 2008. Chemical contaminants in feedlot wastes: Concen-trations, effects and attenuation. Environ Int 34:839–859.

7. Khan B, Lee LS. 2012. Estrogens and synthetic androgens inmanure slurry from trenbolone acetate/estradiol implanted cattle andin waste-receiving lagoons used for irrigation. Chemosphere 89:1443–1449.

8. Zhang YP, Zhou JL. 2008. Occurrence and removal of endocrinedisrupting chemicals in wastewater. Chemosphere 73:848–853.

9. Ankley GT, Jensen KM, Makynen EA, Kahl MD, Korte JJ, HornungMW, Henry TR, Denny JS, Leino RL, Wilson VS, Cardon MC, HartigPC, Gray LE. 2003. Effects of the androgenic growth promoter 17-b-trenbolone on fecundity and reproductive endocrinology of the fatheadminnow. Environ Toxicol Chem 22:1350–1360.

10. Schiffer B, Daxenberger A, Meyer K, Meyer HHD. 2001. The fate oftrenbolone acetate and melengestrol acetate after applicationas growthpromoters in cattle: Environmental studies. Environ Health Perspect109:1145–1151.

11. Jones GD, Benchetler PV, Tate KW, Kolodziej EP. 2014. Mass balanceapproaches to characterizing the leaching potential of trenboloneacetate metabolites in agro-ecosystems. Environ Sci Technol 48:3715–3723.

12. Jensen KM, Makynen EA, Kahl MD, Ankley GT. 2006. Effectsof the feedlot contaminant 17a-trenbolone on reproductive endo-crinology of the fathead minnow. Environ Sci Technol 40:3112–3117.

13. Morthorst JE, Holbech H, Bjerregaard P. 2010. Trenbolone causesirreversible masculinization of zebrafish at environmentally relevantconcentrations. Aquat Toxicol 98:336–343.

14. SaaristoM, Tomkins P, AllinsonM, Allinson G,Wong BBM. 2013. Anandrogenic agricultural contaminant impairs female reproductivebehaviour in a freshwater fish. PLoS One 8:e62782.

15. Boettcher M, Kosmehl T, Braunbeck T. 2011. Low-dose effects andbiphasic effect profiles: Is trenbolone a genotoxicant? Mutat Res723:152–157.

16. Donova MV, Egorova OV, Nikolayeva VM. 2005. Steroid 17b-reduction by microorganisms—A review. Process Biochem 40:2253–2262.

17. Horinouchi M, Hayashi T, Kudo T. 2012. Steroid degradation inComamonas testosteroni. J Steroid Biochem Mol Biol 129:4–14.

18. Webster JP, Kover SC, Bryson RJ, Harter T, Mansell DS, Sedlak DL,Kolodziej EP. 2012. Occurrence of trenbolone acetate metabolites insimulated confined animal feeding operation (CAFO) runoff. EnvironSci Technol 46:3803–3810.

Figure 7. Structures of proposed microbial biotransformation products of trenbolone acetate metabolites. Although we propose 1,2-dehydrogenation productsand A-ring hydroxylation in the present study, other configurations and isomers also are possible.


19. Khan B, Lee LS, Sassman SA. 2008. Degradation of syntheticandrogens 17a- and 17b-trenbolone and trendione in agricultural soils.Environ Sci Technol 42:3570–3574.

20. Khan B, Qiao X, Lee LS. 2009. Stereoselective sorption by agriculturalsoils and liquid�liquid partitioning of trenbolone (17a and 17b) andtrendione. Environ Sci Technol 43:8827–8833.

21. Khan B, Lee LS. 2010. Soil temperature and moisture effects on thepersistence of synthetic androgen 17a-trenbolone, 17b-trenbolone andtrendione. Chemosphere 79:873–879.

22. Mashtare ML, Lee LS, Nies LF, Turco RF. 2013. Transformation of17alpha-estradiol, 17beta-estradiol, and estrone in sediments undernitrate- and sulfate-reducing conditions. Environ Sci Technol 47:7178–7185.

23. Parker JA, Webster JP, Kover SC, Kolodziej EP. 2012. Analysis oftrenbolone acetate metabolites and melengestrol in environmentalmatrices using gas chromatography-tandem mass spectrometry.Talanta 99:238–246.

24. Adam G, Duncan H. 2001. Development of a sensitive and rapidmethod for themeasurement of totalmicrobial activity usingfluoresceindiacetate (FDA) in a range of soils. Soil Biol Biochem 33:943–951.

25. Jones GD, Benchetler PV, Tate KW, Kolodziej EP. 2014. Trenboloneacetate metabolite transport in rangelands and irrigated pasture:Observations and conceptual approaches for agro-ecosystems. EnvironSci Technol 48:12569–12576.

26. Jones GD, Benchetler PV, Tate KW, Kolodziej EP. 2014. Surface andsubsurface attenuation of trenbolone acetate metabolites and manure-derived constituents in irrigation runoff on agro-ecosystems. Environ-mental Science 16:2507–2516.

27. Kolodziej EP, Qu S, Forsgren KL, Long SA, Gloer JB, Jones GD,Schlenk D, Baltrusaitis J, Cwiertny DM. 2013. Identification andenvironmental implications of photo-transformation products oftrenbolone acetate metabolites. Environ Sci Technol 47:5031–5041.

28. Qu S, Kolodziej EP, Long SA, Gloer JB, Patterson EV, Baltrusaitis J,Jones GD, Benchetler PV, Cole EA, Kimbrough KC, Tarnoff MD,Cwiertny DM. 2013. Product-to-parent reversion of trenbolone:Unrecognized risks for endocrine disruption. Science 342:347–351.

29. Zheng W, Li X, Yates SR, Bradford SA. 2012. Anaerobic transforma-tion kinetics and mechanism of steroid estrogenic hormones in dairylagoon water. Environ Sci Technol 46:5471–5478.

30. ZhengW, ZouY, Li X,MacheskyML. 2013. Fate of estrogen conjugate17a-estradiol-3-sulfate in dairy wastewater: Comparison of aerobic andanaerobic degradation and metabolite formation. J Hazard Mater 258–259:109–115.

31. Yang Y-Y., Borch T, Young RB, Goodridge LD, Davis JG. 2010.Degradation kinetics of testosterone by manure-borne bacteria:Influence of temperature, pH, glucose amendments, and dissolvedoxygen. J Environ Qual 39:1153–1160.

32. Stow CA, Carpenter SR, Eby LA, Amrhein JF, Hesselberg RJ. 1995.Evidence that PCBs are approaching stable concentrations in LakeMichigan fishes. Ecol Appl 5:248–260.

33. Stow CA. 1995. Great Lakes herring gull egg PCB concentrationsindicate approximate steady-state conditions. Environ Sci Technol29:2893–2897.

34. Jones GJ, Bourne DG, Blakeley RL, Doelle H. 1994. Degradation of thecyanobacterial hepatotoxin microcystin by aquatic bacteria. Nat Toxins2:228–235.

35. Jones GJ, Orr PT. 1994. Release and degradation of microcystinfollowing algicide treatment of a Microcystis aeruginosa bloom in arecreational lake, as determined by HPLC and protein phosphataseinhibition assay. Water Res 28:871–876.

36. Park H-D, Sasaki Y, Maruyama T, Yanagisawa E, Hiraishi A, Kato K.2001. Degradation of the cyanobacterial hepatotoxin microcystin by anew bacterium isolated from a hypertrophic lake. Environ Toxicol16:337–343.

37. MashtareML, Green DA, Lee LS. 2013. Biotransformation of 17a- and17b-estradiol in aerobic soils. Chemosphere 90:647–652.

38. Xuan R, Blassengale AA, Wang Q. 2008. Degradation ofestrogenic hormones in a silt loam soil. J Agric Food Chem56:9152–9158.

39. Zheng W, Li X, Yates SR, Bradford SA. 2012. Anaerobic transforma-tion kinetics and mechanism of steroid estrogenic hormones in dairylagoon water. Environ Sci Technol 46:5471–5478.

40. Mashtare ML, Lee LS, Nies LF, Turco RF. 2013. Transformationof 17a-estradiol, 17b-estradiol, and estrone in sediments undernitrate- and sulfate-reducing conditions. Environ Sci Technol47:7178–7185.

41. Syntex Animal Health. 1995. Synovex Plus (trenbolone acetate andestradiol benxoate) implant environmental assessment. Palo Alto, CA,USA.

42. Pottier J, Cousty C, Heitzman RJ, Reynolds IP. 1981. Differences in thebiotransformation of a 17-beta-hydroxylated steroid, trenboloneacetate, in rat and cow. Xenobiotica 11:489–500.

43. Qu S, Kolodziej EP, Cwiertny DM. 2012. Phototransformation ratesand mechanisms for synthetic hormone growth promoters used inanimal agriculture. Environ Sci Technol 46:13202–13211.

44. Bauer ERS, Daxenberger A, Petri T, Sauerwein H, Meyer HHD. 2000.Characterisation of the affinity of different anabolics and synthetichormones to the human androgen receptor, human sex hormonebinding globulin and to the bovine progestin receptor. APMIS 108:838–846.

45. Kicman AT. 2008. Pharmacology of anabolic steroids. Br J Pharmacol154:502–521.

46. Stavreva DA, George AA, Klausmeyer P, Varticovski L, Sack D, VossTC, Schiltz RL, Blazer VS, Iwanowicz LR, Hager GL. 2012. Prevalentglucocorticoid and androgen activity in US water sources. ScientificReports 2:937.

47. Lavado R, Loyo-Rosales JE, Floyd E, Kolodziej EP, Snyder SA, SedlakDL, Schlenk D. 2009. Site-specific profiles of estrogenic activity inagricultural areas of California’s inland waters. Environ Sci Technol43:9110–9116.


rates and product identification for trenbolone acetate...

Documents