ARTICLE pubs.acs.org/est
Isolation and Identification of Ligands for the Goldfish Testis Androgen Receptor in Chemical Recovery Condensates from a Canadian Bleached Kraft Pulp and Paper Mill Philip D. Scott,† Craig B. Milestone,†,‡ D. Scott Smith,† Deborah L. MacLatchy,† and L. Mark Hewitt*,‡ † ‡
Wilfrid Laurier University, Waterloo, Ontario, Canada Environment Canada, Burlington, Ontario, Canada
bS Supporting Information ABSTRACT: This study builds on a series of investigations characterizing substances in kraft mill chemical recovery condensates that depress sex steroids in fish. Here, incubations of goldfish testis androgen receptors (AR) with condensate extracts were used to investigate the potential role of androgens in hormone depressions. Condensates contained variable levels of AR ligands, with the highest amounts in nonpolar extracts of filtered solids prior to solid phase extraction (SPE). High pressure liquid chromatography (HPLC) fractionation recovered the majority of activity in one fraction, with ligands detected in three additional fractions. Gas chromatography mass spectrometry analysis of the most active fraction confirmed the two most abundant components as the diterpenes manool and geranyl linalool. Manool exhibited a relative affinity for the AR that was 300 fold less than testosterone and accounted for 26% of total filtered solids activity. Geranyl linalool exhibited no affinity for the AR. Three additional diterpenoid families were tentatively identified as principal components of the three other androgenic HPLC fractions. Compared to condensates, final effluent had 3000 fold less androgenic activity, with 0.05; Figure 3). Three additional HPLC fractions (4, 6 and 9; Figure 3) also demonstrated statistically significant androgenic activities that ranged from 15 to 1500 fold less active than HPLC fraction 7.
GC-MS-MS Analysis of Active HPLC Fractions. The total ion chromatogram (TIC) from a full scan GC-MS analysis of HPLC fraction 7 is given in Figure 4. The chromatogram is dominated by peak 6, which exhibits an apparent molecular ion of m/z 272 and fragment ions consistent with a diterpene (inset, Figure 4). For example, m/z 272 represents loss of H2O from a molecular ion of m/z 290, while m/z 257 represents a subsequent loss of CH3 (more fragmentation information in Table S1).38 The mass spectra closely resembled that of bicyclic diterpene manool, previously identified in condensates by Belknap et al.17 A positive confirmation of manool was obtained through mass spectral and GC retention time matches with an authentic standard. The presence of manool in HPLC fraction 7 and FP-NP condensate extracts is consistent with its hydrophobicity, as this compound would be expected to partition to solids in the condensates (log Kow = 7.09) and is not readily soluble in water (0.0395 mg/L).39 Closer inspection of the remaining seven major peaks evident in the TIC of HPLC fraction 7 reveal mass spectra consistent with a series of diterpenes (Table S1). All have apparent molecular ions of 272 m/z and prominent fragment ions at m/z 257, 137, 95, and 81, similar to manool. Since the molecular ion for manool is weak to nonexistent, this is likely also the case for these unknowns. Peak 5 was subsequently confirmed with an 10229
dx.doi.org/10.1021/es202732c |Environ. Sci. Technol. 2011, 45, 10226–10234
Environmental Science & Technology
ARTICLE
Figure 3. HPLC fractionation of active FP-NP Condensate 3 extract and androgenic activity of HPLC fractions and extracts generated from condensate solid phase extraction protocol. TOP: Testosterone equivalents (TEq) detected in each fraction. Alphabetical superscripts indicate treatments different than each other (p < 0.05). Error bars represent standard error of the mean of two cytosolic AR preparations. DAD: diode array detector, ELSD: evaporative light scattering detector, FLD: fluorescence detector. BOTTOM: Dotted lines in chromatograms indicate HPLC fraction collections. All data generated with Condensate 3.
Figure 4. Total ion chromatogram from a full scan positive ion electron impact ionization analysis of Condensate 3 FP-NP HPLC fraction 7, with the major detectable components labeled. Peak 5 was confirmed as the acyclic diterpene geranyl linalool (left inset), and peak 6 was confirmed as the cyclic diterpene manool (right inset); the latter was found to account for 25% the androgenic activity associated with combined solids extract.
authentic standard as geranyl linalool, an acyclic diterpene (inset, Figure 4). HPLC fraction 4 was the second most androgenic fraction. Inspection of the seven major peaks evident in its TIC revealed mass spectra consistent with a different series of diterpenes (Table S1). Peak 5 was tentatively identified as dehydroabietal, a cyclic diterpene previously identified in Serbian Spruce (Picea omorika).40 Spectral analysis of all other peaks, with the exception of peak 7, suggests cyclic diterpene structures. A TIC of HPLC fraction 6 (the third most androgenic HPLC fraction) revealed 12 major detectable constituents, many of which had mass spectra consistent with a third series of diterpenes (Table S1). The mass spectrum of one peak was confirmed as manool carryover during fraction collections. Upon comparison, it
was determined that peaks 1, 6, and 7 from HPLC fraction 6 were the same as peaks 1, 3, and 4, respectively, in HPLC fraction 7, also evidence of carryover. GC-MS analysis revealed 14 major detectable constituents in HPLC fraction 9, the fourth most androgenic HPLC fraction (Table S1). Peak 12 had an apparent molecular ion of m/z 286 and prominent fragment ions at m/z 271, 253, 173, among others. This compound was tentatively identified as the cyclic diterpene dehydroepiabietal, a compound previously found in softwood.41,42 Spectral comparisons of other constituents insinuate cyclic diterpene structures. Androgenic Activities of Confirmed Compounds. Both manool and geranyl linalool, confirmed in the most active HPLC fraction (HPLC fraction 7; Figure 3), were evaluated for androgenic 10230
dx.doi.org/10.1021/es202732c |Environ. Sci. Technol. 2011, 45, 10226–10234
Environmental Science & Technology
ARTICLE
Figure 5. AR displacement curves of testosterone, manool, and geranyl linalool. The testosterone IC50 value was 0.23 μM for testosterone and 71.61 μM for manool (∼300 fold less androgenic). Geranyl linalool did not interact with the AR, as indicated by the nonparallel slope. Error bars represent standard error of the mean of two cytosolic AR preparations.
activity using the AR binding assay. Manool’s displacement curve was parallel to that of testosterone, with a relative AR affinity of 0.003 (300 fold less potent than testosterone, IC50 of 71.61 μM versus 0.23 μM for testosterone) (Figure 5). This value was used to determine the amount of androgenic activity attributable to manool, in extracts and fractions. Manool accounted for 29, 22, and 28% of the total androgenic activity in FP-NP extracts from Condensates 1, 2, and 3, respectively. Out of the total androgenic activity of HPLC fractions 6 and 7, manool accounted for 44% and 14%, respectively. The dilution series of geranyl linalool produced a binding curve with a nonparallel slope to that of testosterone, exhibiting no affinity for the AR (data not shown). All compounds previously found in condensates17 that were confirmed in the three condensate samples of the present study (Table 1) were also evaluated for their affinities to the AR and showed no androgenic activity (data not shown). Analysis of Final Effluent. Treated final effluent was sampled and evaluated chemically for the presence of manool and biologically to determine its androgenic potency. Using the condensate extraction SPE method, the total effluent androgenic activity was 5706 ng/L TEq, where the greatest proportion was found in the FP-NP extract (3400 ng/L). This FP-NP activity was approximately 3000 fold less than the average activity of the condensate FP-NP extracts. Manool was confirmed by GC-MSMS in the FP-NP and SPE-NP effluent extracts in small quantities (18.8 and 10.5 μg/L, respectively) and was found to account for only 0.05% of the androgenic activity present in the FP-NP effluent extracts.
’ DISCUSSION In this study, the androgenic effects hypothesized for an in-mill waste stream associated with hormone depressions in fish was
studied in detail using binding assays with goldfish testis AR. Chemical recovery condensates exhibited significant androgenic activity, and this was found to vary substantially (4 6 fold), along with condensate extractives (1.5 6 fold). High levels of activity were consistently associated with nonpolar extracts of filtered solids (FP-NP), while lower amounts were recovered by SPE. HPLC fractionations of an FP-NP extract recovered the activity in one fraction, with multiple AR ligands evidenced in three additional fractions. The cyclic diterpene manool was confirmed in the most active fraction and accounted for approximately 25% of the FP-NP extract activity and