Gas chromatographic–mass spectrometric fragmentation study of phytoestrogens as their trimethylsilyl derivatives: Identification in soy milk and wastewater samples

Journal of Chromatography A, 1216 (2009) 6024–6032

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Gas chromatographic–mass spectrometric fragmentation study of phytoestrogens as their trimethylsilyl derivatives: Identification in soy milk and wastewater samples Imma Ferrer a,∗ , Larry B. Barber b , E. Michael Thurman a,b a b

Center for Environmental Mass Spectrometry, Department of Environmental Engineering, University of Colorado, ECOT 441, Boulder, CO 80309, USA U.S. Geological Survey, Water Resources Division, Boulder, CO, USA

a r t i c l e

i n f o

Article history: Received 16 March 2009 Received in revised form 7 June 2009 Accepted 12 June 2009 Available online 18 June 2009 Keywords: Phytoestrogens Ion-trap GC–MS Wastewater Mass spectrometry

a b s t r a c t An analytical method for the identification of eight plant phytoestrogens (biochanin A, coumestrol, daidzein, equol, formononetin, glycitein, genistein and prunetin) in soy products and wastewater samples was developed using gas chromatography coupled with ion trap mass spectrometry (GC/MS–MS). The phytoestrogens were derivatized as their trimethylsilyl ethers with trimethylchlorosilane (TMCS) and N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA). The phytoestrogens were isolated from all samples with liquid–liquid extraction using ethyl acetate. Daidzein-d4 and genistein-d4 labeled standards were used as internal standards before extraction and derivatization. The fragmentation patterns of the phytoestrogens were investigated by isolating and fragmenting the precursor ions in the ion-trap and a typical fragmentation involved the loss of a methyl and a carbonyl group. Two characteristic fragment ions for each analyte were chosen for identification and confirmation. The developed methodology was applied to the identification and confirmation of phytoestrogens in soy milk, in wastewater effluent from a soy-milk processing plant, and in wastewater (influent and effluent) from a treatment plant. Detected concentrations of genistein ranged from 50,000 ␮g/L and 2000 ␮g/L in soy milk and in wastewater from a soy-plant, respectively, to 20 ␮g/L and <1 ␮g/L for influent and effluent from a wastewater treatment plant, respectively. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Phytoestrogens are a group of nonsteroidal polyphenolic compounds that occur naturally in a wide range of plants and induce biological responses based on their ability to bind to estrogen receptors. Usually, they are present at high concentrations in legumes such as soy, clover, alfalfa, beans, and peas [1,2]. Beneficial effects of flavonoids (water-soluble plant pigments derived from the 2phenyl-1,4-benzopyrone structure) on humans have been reported in recent reviews [3–6] and include their ability to scavenge free radicals [5,7] and to have immuno-modulatory [4] effects. However, plant phytoestrogens have been also cited as possible endocrine disruptors in fish along with synthetic hormones and human estrogens from wastewater effluents [8]. Additionally, phytoestrogens, especially daidzein, genistein, and glycitein, have been correlated with low sperm counts in North American men with high and frequent soy food intake, more than twice per week [9]. Because phytoestrogens are excreted, not only by plants but also by humans

∗ Corresponding author. Tel.: +1 303 7354147. E-mail address: [email protected] (I. Ferrer). 0021-9673/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2009.06.042

and livestock via food consumption, it is important to evaluate their sources in food and their concentration in surface water and wastewater. Previous studies have found concentrations of biochanin A, daidzein, formononetin, and coumestrol ranging from 1 to 10 ng/L in rivers in Australia, Germany, and Italy [10–13], whereas 43 ␮g/L daidzein and 143 ␮g/L genistein were detected in a Japanese river [8], perhaps because of higher soy impact. Several analytical methodologies have been described for the identification and quantitation of flavonoids, including gas chromatography [14–16] and liquid chromatography [17–19]. In fact, an extensive recent review by de Rijke et al. [20] reports the advantages and disadvantages of these methodologies for the analysis of flavonoid compounds. Using LC–MS based methods, matrix interferences have been reported [19], which make quantitation of phytoestrogens difficult, especially in those cases where the matrix is complex, such as wastewater samples. Because of potential matrix interferences the use of selective ion-trap or tandem mass spectrometric methodologies (either LC/MS–MS or GC/MS–MS) is crucial to identify and confirm the presence of these analytes in complex samples. In general, GC-based methods provide high resolution and low detection limits for the phytoestrogens, although they require derivatization to form the trimethylsilylether

I. Ferrer et al. / J. Chromatogr. A 1216 (2009) 6024–6032

derivatives that increases volatility and improves thermal stability. Usually, MS–MS techniques that look at two or more characteristic fragments are required to fully characterize the silanized phytoestrogens since they provide higher selectivity than single ion monitoring techniques [21]. In this study, a GC/MS–MS methodology using ion-trap was developed for the identification and confirmation of eight phytoestrogens in soy products and water samples from a wastewater treatment plant in Boulder, CO. The wastewater from this plant has been under study for endocrine disrupting effects on fish downstream of the plant [22]. The concentration of natural plant phytoestrogens to wastewater in this study area is currently unknown and is a valuable data set for future biological studies of plant phytoestrogenic effects on native fish, especially given the wide occurrence of plant indicators in wastewater, such as beta sitosterols [23]. The phytoestrogens were analyzed both in soy milk and wastewater using the characteristic fragmentation of the molecular ions for each of the analytes studied. An extensive study on the fragmentation patterns of each analyte was carried out in this study and it is described in detail. Finally, the principal phytoestrogens occurring in a commercial soy milk and the influent and effluent wastewater from a soy-milk plant are reported and discussed. 2. Experimental 2.1. Chemicals and reagents

6025

methanol residues. Then, 200 ␮L of 10% TMCS/BSTFA derivatizing reagent was added to the vials and vortexed for 15 s. The vials were placed on a heating block at 60 ◦ C for 30 min, 1 h, 4 h and 20 h. Four hours was selected as the optimum derivatization condition. After this time the vials were removed from the heating block, set aside 15 min to cool, and the reagents were evaporated to dryness under nitrogen. Finally, the dry residue was dissolved in 200 ␮L of 5:1 BSTFA/pyridine injection solvent and vortexed for 30 s. Daidzein-d4 and genistein-d4 were used as internal standards before extraction and derivatization procedure and they were added together with the calibration standards. The derivatized extracts were transferred into autosampler vials using a clean silynized glass pipette and were analyzed by GC/MS–MS. 2.3. Sample preparation A simple and rapid procedure for the isolation of the phytoestrogens from soy milk and water was carried out using ethyl acetate as an extracting solvent. The soy milk sample (1 mL) and the water and wastewater samples (5 mL) were liquid–liquid extracted with 15 mL of ethyl acetate (two times with 7.5 mL each) and evaporated to dryness under a stream of nitrogen. The internal standards, daidzein-d4 and genistein-d4 were added before extraction to account for any loses during the extraction process. The residues were derivatized with TMCS and BSTFA, at 60 ◦ C for 4 h. 2.4. GC/MS–MS analyses

Biochanin A, coumestrol, daidzein, equol, formononetin, genistein, glycitein and prunetin were purchased from Sigma– Aldrich (St. Louis, MO, USA). Isotopic labeled compounds daidzein-d4 and genistein-d4 were obtained from Cambridge Isotopes (Cambridge, MA, USA). The derivatization reagents: N,Obis(trimethylsilyl)trifluoroacetamide (BSTFA), trimethylchlorosilane (TMCS) were obtained from Sigma–Aldrich (St. Louis, MO, USA) and 500 ␮L of TMCS was added to 5 mL of BSTFA in order to have a 10% TMCS derivatization reagent. In the same way, 2 mL of pyridine were mixed with 8 mL of BSTFA to have a BSTFA/Pyridine (5:1; v:v) solvent mixture that is added to the dry extracts before injection onto the GC. Fresh derivatizating reagents were prepared every two weeks. Methanol, ethyl acetate and dichloromethane were obtained from Burdick and Jackson (Muskegon, MI, USA). All materials and reagents were of analytical purity. Standards of individual phytoestrogens were prepared in methanol at concentrations of 500 mg/L. From these solutions, diluted working solutions were prepared to be derivatized and analyzed by GC/MS–MS. All standard solutions were stored at −20 ◦ C and allowed to equilibrate at ambient temperature for at least 2 h before use. Calibration samples were prepared by mixing different volumes of stock mixture (4–20 ␮L) and internal-standard mixture (20 ␮L).

The identification of phytoestrogens by GC/MS was carried out using a Trace GC 2000 gas chromatograph coupled to a GCQ/Polaris ion-trap mass spectrometer (Thermo-Finnigan, Austin, TX, USA) equipped with an AS2000 autosampler (ThermoElectron, Milan, Italy). The chromatographic separation was performed using a DB5MS (5% phenyl, 95% methylpolysiloxane), 25 m × 0.25 mm I.D., fused-silica capillary column (J&W Scientific, Folsom, CA, USA) of 0.25 ␮m film thickness. The carrier gas was helium at a constant flow-rate of 1.2 mL/min held by electronic pressure control. Injector temperature was 280 ◦ C, and splitless injection mode was used. The oven temperature program was 80 ◦ C (held for 2 min) to 230 ◦ C at 40 ◦ C/min (held for 1 min), to 300 at 4 ◦ C/min (held for 3 min) and finally to 310 ◦ C at 10 ◦ C/min (held for 1 min). The MS operating conditions were the following: positive electron ionization mode (EI+) using automatic gain control (AGC) with an electron energy of 70 eV. The ion source and transfer line temperatures were 200 ◦ C and 280 ◦ C, respectively. The instrument was tuned using perfluorotributylamine (FC-43) according to manufacturer’s recommendations in order to achieve the best sensitivity. Electron multiplier voltage was set to 1500 V by automatic tuning. EI full scan data acquisition was registered over the range m/z 50–500 at 0.66 s per scan. One microliter of the extracts was injected on the system. Xcalibur version 1.2 software was used for control, general operation, and data acquisition of the results.

2.2. Derivatization of phytoestrogens

3. Results and discussion

A modified version of a previously reported methodology for the derivatization of estrogen compounds [22] was used here. The analytes were derivatized to their trimethylsilyl ethers with N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) and trimethylchlorosilane (TMCS). The calibration standards were evaporated to dryness under a stream of nitrogen in silynized 5-mL reaction vials. The ethylacetate extracts from water samples were evaporated under the same procedure. Afterwards, 1 mL of methylene chloride was added to the reaction vials to form an azeotrope during evaporation to dryness under a stream of nitrogen. This step was repeated in order to assure complete removal of water and

3.1. Optimization of derivatization Phytoestrogens are polar, non-volatile compounds and therefore not well suited to GC analysis in their native form. In order to make these compounds more amenable to GC/MS analysis, they must first be derivatized to more volatile, less polar species. This is usually done by derivatizing the hydroxyl groups with suitable protecting groups. The derivatization procedure described in the experimental section and used in this work has been already reported for hormone compounds by Vadja et al. [22], and it was successfully applied here to the phytoestrogens. However, a slightly

6026

I. Ferrer et al. / J. Chromatogr. A 1216 (2009) 6024–6032

Table 1 Formula, precursor ion chosen, and chemical structures of the BSTFA derivatives of all the phytoestrogens and deuterated standards studied. Name

Formula

M+ or M+ •

Biochanin A

C22 H28 O5 Si2 + • , m/z 428

413

Coumestrol

C21 H24 O5 Si2 + • , m/z 412

412

Daidzein

C21 H26 O4 Si2 + • , m/z 398

398

Daidzein-d4

C21 H22 D4 O4 Si2 + • , m/z 402

402

Equol

C21 H30 O3 Si2 + • , m/z 386

386

Formononetin

C19 H20 O4 Si+ • , m/z 340

340

Genistein

C24 H34 O5 Si3 + • , m/z 486

471

Chemical structures

I. Ferrer et al. / J. Chromatogr. A 1216 (2009) 6024–6032

6027

Table 1 (Continued ) Name

Formula

M+ or M+ •

Genistein-d4

C24 H30 D4 O5 Si3 + • , m/z 490

475

Glycitein

C22 H28 O5 Si2 + • , m/z 428

428

Prunetin

C22 H28 O5 Si2 + • , m/z 428

413

modification was included in this work, which was the removal of O-methoxyamine hydrochloride that was not necessary because the chemical structure of the phytoestrogens does not contain an adjacent keto group to a double bond in the ring, so no blocking of the carboxyl groups is necessary for the derivatization of these compounds. The derivatization time needed for the complete formation of the trimethylsilyl derivatives also was optimized in this study. Several reaction times were studied and it was concluded that a step of 4-h derivatization yielded best efficiency (data not shown) in the least time for the formation of the trimethylsilyl derivatives for all the phytoestrogens. Table 1 shows the chemical structures for each of the trimethylsilyl derivatives of the phytoestrogens studied along with the m/z values for the molecular ions or main precursor ions detected by GC/MS scan. The deuterated standards, daidzein-d4 and genisteind4 , are also shown. All the compounds presented the molecular ion as a base peak in the spectrum except for biochanin A, genistein, and prunetin. In these cases the [M−CH3 ]+ ion was observed, probably due to the closer position of the trimethylsilyl to the keto group in the chemical moiety, making this methyl group more susceptible to be lost in the EI source. The rest of compounds exhibited the molecular ion as a major ion in the mass spectrum. The precursor ions shown in this table were the ones chosen for the subsequent selective MS–MS experiments, as detailed in the section below. 3.2. Chromatographic separation GC/MS conditions were optimized such that baseline resolution was achieved between all analytes while also keeping an adequate run time (25 min) so as to maximize the throughput of samples. Because of the similar polarity exhibited by all of the derivatized phytoestrogens, the mix of compounds was separated using a slow temperature program (4 ◦ C/min) in the GC/MS oven. Fig. 1 shows the separation of the eight compounds and the internal standards. Good chromatographic separation was obtained with the temperature program reported in the experimental section. As can be observed

Chemical structures

in this figure, the BSTFA derivatives eluted in an order depending on the number of trimethylsilyl groups substituted, and the position of other groups in the molecule such as methoxy and carbonyl groups. In general, the compounds eluted in the order of 1, 2, and 3 trimethylsilyl groups substituted with the exception of coumestrol, which has a different chemical structure with four rings and thus presents a later retention time than the other phytoestrogens. Glycitein was the other exception being the last eluting compound of the mix since the ortho/meta position of the trimethylsilyl and the methoxy groups gives this molecule a highly hydrophobic nature. Equol eluted first because it was the only compound lacking the keto group in the middle aromatic ring. Formononetin, with only one trimethylsilyl group eluted second, followed by biochanin A, prunetin and daidzein with two trimethylsilyl groups, and finally, genistein with 3 trimethylsilyl groups, which was the last one of this group of three member ring phytoestrogens. 3.3. Fragmentation study of phytoestrogens by GC/MS–MS Fragmentation patterns were studied for all the phytoestrogens under different voltages. An optimization of the collision voltage in the trap was carried out ranging from 3 V to 10 V. The majority of compounds fragmented easily yielding at least two characteristic fragment ions, at a voltage of 6 V. However, daidzein, genistein and prunetin required a higher fragmentation voltage of 9 V to achieve good sensitivity for the fragment ions. Only one compound, equol, did not fragment under any conditions, due to the high stability of this molecule related to the symmetry of this phytoestrogen. Table 2 shows the precursor ions and the main fragment ions for all the analytes studied. In some cases the fragmentation of the trimethylsilyl group is involved such as for genistein and prunetin. In other cases either a methyl or a methyl and a carbonyl group are lost and re-arrangement of the structure occurs (Fig. 2). Fig. 2 shows the detailed fragmentation pattern of the group of phytoestrogens. From an analytical point of view it is of primary importance to select those fragment ions which can be used

6028

I. Ferrer et al. / J. Chromatogr. A 1216 (2009) 6024–6032

Fig. 1. GC/MS–MS chromatogram of the mixture of phytoestrogens (concentration 0.5 ␮g/mL). Compound labels: 1 = equol, 2 = formononetin, 3 = biochanin A, 4 = prunetin, 5 = daidzein + daidzein-d4, 6 = genistein + genistein-d4, 7 = coumestrol, 8 = glycitein. Table 2 Precursor ions and main product ions obtained by GC/MS–MS for the phytoestrogens studied. See Fig. 2 for complete fragmentation pathways from the molecular ion. Name Biochanin A Coumestrol Daidzein Daidzein-d4 Equol Formononetin Genistein Genistein-d4 Glycitein Prunetin

Formula +•

C22 H28 O5 Si2 , m/z 428 C21 H24 O5 Si2 + • , m/z 412 C21 H26 O4 Si2 + • , m/z 398 C21 H22 D4 O4 Si2 + • , m/z 402 C21 H28 O3 Si2 + • , m/z 386 C19 H20 O4 Si+ • , m/z 340 C24 H34 O5 Si3 + • , m/z 486 C24 H30 D4 O5 Si3 + • , m/z 490 C22 H28 O5 Si2 + • , m/z 428 C22 H28 O5 Si2 + • , m/z 428

M+ or M+ •

Fragment ions

413 412 398 402 386* 340 471 475 428 413

398 [M+ −(• CH3 )]+ • , 370* [M+ −(• CH3 )–CO]+ • 397* [M+ • −(• CH3 ]+ , 369 [M+ • −(• CH3 )–CO]+ 383 [M+ • −(• CH3 )]+ , 355* [M+ • −(• CH3 )–CO]+ 387 [M+ • −(• CH3 )]+ , 359* [M+ • −(• CH3 )–CO]+ No ions detected in MS–MS 325 [M+ • −(• CH3 )]+ , 296* [M+ • −(• CH3 )–• HCO]+ • 399* [M+ −Si(CH3 CH3 CH2 )]+ , 327 [M+ −(Si(CH3 CH3 CH2 )2 ]+ 403* [M+ −Si(CH3 CH3 CH2 )]+ , 331 [M+ −(Si(CH3 CH3 CH2 )2 ]+ 413 [M+ • −(• CH3 )]+ , 398* [M+ • −(• CH3 )2 ]+ • 370* [M+ −(• CH3 )–CO]+ • , 341 [M+ −Si(CH3 CH3 CH2 )]+

T˛ˇε 0. Ion used for quantitation.

for quantitative and confirmatory purposes that typically do not involve the loss of the TMS (trimethylsilyl) group. For this reason and based on the relative abundance, two selective fragment ions were chosen for each analyte. One of them (having the highest abundance) was used for quantitation purposes (shown in asterisk in Table 2). And the second one was used as a qualifier ion for identification and confirmatory purposes. Only genistein incurred two losses of 72 mass units consistent with two losses of the TMS group and did not show the characteristic loss of a methyl radical and a CO group. Because of the non-characteristic double loss of 72 mass units (other derivatized compounds could exhibit this same loss), care was taken in the exact assignment of retention time for a correct assignment of genistein. Equol was the only compound that would not fragment and no fragments could be obtained, so in this case only the molecular ion was used for both quantitation and confirmation. Thus, caution must be used in interpretation of the analysis for this compound more so than any of the other phytoestrogens. 3.4. Analytical performance Calibration charts were plotted for concentrations of 0.05, 0.1, 0.25, 0.5, 1 and 2.5 ␮g/mL of standards with constant amounts of internal standards. The charts were found to be linear between 0.1 and 2.5 ␮g/mL for all the analytes. Values for the coefficient

of determination, R2 , were >0.99. Calculated instrumental limits of detection (LODs) for all analytes are shown in Table 3. The lowest LODs were for genistein and biochanin A. Prunetin showed the lowest sensitivity (i.e., signal strength/concentration) and consequently produced a higher LOD. Equol had to be quantitated in full-scan conditions (as explained in the last section) and for this reason a high LOD value of 100 ␮g/L was achieved. Intraand interassay coefficients of variation (CV, n = 3) for the standards ranged from 5 to 15%, showing a good reproducibility of the methodology.

Table 3 Instrumental limits of detection (LODs) for the phytoestrogens studied in methanol solutions of standards. Name

Calibration curve

R2

LODs (␮g/L)

Biochanin A Coumestrol Daidzein Equol Formononetin Genistein Glycitein Prunetin

0.751x − 0.1705 0.425x − 0.0322 0.317x − 0.0338 0.856x − 0.0566 0.310x − 0.0519 1.107x − 0.3071 0.545x − 0.5433 0.721x − 0.2342

0.992 0.993 0.994 0.993 0.992 0.998 0.995 0.997

15 25 20 100a 20 10 15 30

a

LOD of compound in full-scan.

I. Ferrer et al. / J. Chromatogr. A 1216 (2009) 6024–6032

Fig. 2. Fragmentation pattern of the phytoestrogens studied by GC/MS–MS.

6029

6030

I. Ferrer et al. / J. Chromatogr. A 1216 (2009) 6024–6032

Fig. 3. GC/MS–MS chromatogram corresponding to the analysis of a soy milk sample from a local grocery store.

Table 4 Concentrations of phytoestrogens (in ␮g/L) in various samples. Compound Genistein Daidzein Glycitein

Soy milk 50,000 15,000 200

Soy processing plant effluent

Wastewater treatment plant influent

Wastewater treatment plant effluent

2000 500 50

20 20 <1

<1 <1 <1

3.5. Analysis of food and environmental samples Fig. 3 shows the chromatographic analysis of commercially available soy milk purchased from a local grocery store. Genistein (50,000 ␮g/L) and daidzein (15,000 ␮g/L) were the most common phytoestrogens identified in the soy milk sample (Table 4). Dilution of the extract had to be performed due to the high concentration of these two analytes in soy milk to avoid overloading. During a full-scan analysis of the soy milk sample, an unknown peak was discovered in the chromatogram at masses of m/z 428 and 413. This

compound chromatographed in the vicinity of the coumestrol and contained losses of 15 and 43, which made it appear as a possible phytoestrogen. The GC/MS–MS analysis shows the spectrum in Fig. 4. We suspected it was an isomer of biochanin A and a literature search revealed that glycitein has the same mass as biochanin A at m/z 428 when derivatized. We obtained a standard of glycitein and verified that it had the same retention time and same mass fragments as the unknown peak in the full-scan analysis of the soy milk sample. The concentration was 200 ␮g/L in the soy milk (Table 4). The literature search also revealed that glycitein has been reported

Fig. 4. GC/MS spectrum of an unknown peak at 21.6 min in soy milk.

I. Ferrer et al. / J. Chromatogr. A 1216 (2009) 6024–6032

6031

Fig. 5. GC/MS–MS chromatogram corresponding to the analysis of the effluent of a soy product making plant in Boulder, CO.

in soy products. Thus, we added this compound to our list of important phytoestrogens to look for in the wastewater samples. Previous studies have not reported this compound in water. It is somewhat difficult to identify unknowns by GC/MS–MS when doing derivatization as TMS derivatives because of the problem of not knowing how many TMS derivatives are present. We addressed this problem by using the ion trap in MS3 mode which allows us to check for the losses of 72 for the TMS group. This turned out to be a useful and powerful procedure for determining the unknown phytoestrogen glycitein, in the soy milk sample. After determining the major phytoestrogens in the soy milk, we next examined waste and wastewater that was receiving influent from a soy-milk plant. Fig. 5 shows the GC/MS–MS analysis of wastewater (WW) from a soy-product plant. Genistein and daidzein were identified at concentrations of 2000 and 500 ␮g/L, respectively. In this case, 5 mL of

wastewater sample were extracted and pre-concentrated to a final volume of 200 ␮L. Glycitein was below the limit of quantitation, but it was present in this sample as shown in the chromatogram at an estimated concentration of ∼50 ␮g/L. This experiment showed the applicability of the technique for the analysis of phytoestrogens in wastewater samples that are directly discharged from a food processing plant. This is a specific location in Colorado and should not be extrapolated beyond other untreated wastewater locations. Fig. 6a shows the GC/MS–MS chromatogram for an influent to the Boulder wastewater treatment plant. Genistein and daidzein were about the same concentration of ∼20 ␮g/L, which is 100× less than the effluent (Fig. 5) from the soy wastewater. In Boulder wastewater treated effluent, phytoestrogens were less than 1 ␮g/L (see Fig. 6b). The fact that the influent of the Boulder plant was 20 ␮g/L suggests that the soy wastewater is not the only source for the genistein and daidzein in the influent. The volume of

Fig. 6. GC/MS–MS chromatogram corresponding to the analysis of an (a) influent and an (b) effluent of the Boulder wastewater treatment plant.

6032

I. Ferrer et al. / J. Chromatogr. A 1216 (2009) 6024–6032

wastewater is estimated at 10 millions gallons per day and the soy wastewater plant makes a small contribution to this total, approximately 0.1% of the flow. Thus, at maximum there is no more than 5% of the total phytoestrogens that are coming from the soy wastewater and it makes a small contribution to the total phytoestrogen pool. Thus, there must be a considerable input from other sources to the Boulder wastewater-plant influent. Obviously there is the input from human consumption of soy products, this would include not only soy milk, but many other products that use soy, including many soy product additives and beverages. It is interesting to note that the other phytoestrogens (biochanin A, formononetin, coumestrol, prunetin, and equol) were not detected. These results suggest that these phytoestrogens are not sufficiently present in the wastewater influent or in the soy milk effluent to reach the LODs of the method. Furthermore, they are apparently not part of the diet of the city of Boulder; whereas, genistein, daidzein, and glycitein are. This is also an important negative result for future studies on the impact of phytoestrogens on fish studies of Boulder Creek. 4. Conclusions For the first time, an analytical method for the identification of eight plant phytoestrogens (biochanin A, coumestrol, daidzein, equol, formononetin, glycitein, genistein and prunetin) in soy products and wastewater samples was developed using gas chromatography coupled with ion trap mass spectrometry (GC/MS–MS) using trimethylsilyl derivatives and deuterated labeled standards. Furthermore, the fragmentation patterns of all the phytoestrogens were investigated by isolating and fragmenting the precursor ions in the ion-trap and a typical fragmentation involving the loss of a methyl and a carbonyl group was discovered. Two characteristic fragment ions for each analyte were chosen for identification and confirmation. Finally, the developed methodology was applied to the identification and confirmation of phytoestrogens in soy milk, in wastewater effluent from a soy-milk processing plant, and in wastewater (influent and effluent) from a treatment plant. Detected concentrations of genistein ranged from 50,000 ␮g/L and 2000 ␮g/L in soy milk and in wastewater from a soy-plant, respectively, to 20 ␮g/L and <1 ␮g/L for influent and effluent from a wastewater treatment plant, respectively. Finally, future endocrine disruption studies on fish will focus on the effects of genistein and daidzein since they are the main compounds occurring in the influent to this wastewater plant.

Acknowledgements The use of brand, trade, or firm names in this paper is for identification purposes only and does not constitute endorsement by the U.S. Geological Survey. I.F. acknowledges travel support from the Ministerio de Ciencia y Tecnología (“José Castillejo” Program) in Spain. This work was partly supported by the Toxics Substances Hydrology Program of the U.S. Geological Survey and authors want to thank our colleagues, Mike Focazio and Herb Buxton of the USGS. References [1] K.R. Price, G.R. Fenwick, Food Addit. Contam. 2 (1985) 73. [2] K. Reinli, G. Block, Nutr. Cancer 26 (1996) 123. [3] J.J. Peterson, G.R. Beecher, S.A. Bhagwat, J.T. Dwyer, S.E. Gebhardt, D.B. Haytowitz, J.M. Holden, J. Food Comp. Anal. 19 (2006) S74. [4] J.Y. Lin, C.Y. Tang, Food Chem. 101 (2007) 140. [5] M.A. Anagnostopoulou, P. Kefalas, V.P. Papageorgiou, A.N. Assimopolou, D. Boskou, Food Chem. 94 (2006) 19. [6] K. Gao, S.M. Henning, Y. Niu, A.A. Youssefian, N.P. Seeram, A. Xu, D. Heber, J. Nutr. Biochem. 17 (2006) 89. [7] M.M. Saleh, F.A.E.M. Hashem, K.W. Glombitza, Food Chem. 63 (1998) 397. [8] M. Kawanishi, T. Takamura-Enya, R. Ermawati, C. Shimohara, M. Sakamoto, K. Matsukawa, T. Matsuda, T. Murahashi, S. Matsui, K. Wakamayashi, T. Watanabe, Y. Tashiro, T. Yagi, Environ. Sci. Technol. 38 (2004) 6424. [9] J.E. Chavarro, T.L. Toth, S.M. Sadio, R. Hauser, Hum. Reprod. 23 (2008) 2584. [10] P. Spengler, W. Korner, J.W. Metzger, Environ. Toxicol. Chem. 20 (2001) 2133. [11] A. Lagana, A. Bacaloni, I. De Leva, A. Faberi, G. Fago, A. Marino, Anal. Chim. Acta 501 (2004) 79. [12] A. Bacaloni, C. Cavaliere, A. Faberi, P. Foglia, R. Samperi, A. Lagana, Anal. Chim. Acta 531 (2005) 229. [13] J. Kang, W.E. Price, L.A. Hick, Rapid Commun. Mass Spectrom. 20 (2006) 2411. [14] Z. Fuzfai, I. Molnar-Perl, J. Chromatogr. A 1149 (2007) 88. [15] S. Moors, M. Blaszkewicz, H.M. Bolt, G.H. Degen, Mol. Nutr. Food Res. 51 (2007) 787. [16] P.B. Grace, J.I. Taylor, N.P. Botting, T. Fryatt, M.F. Oldfield, S.A. Bingham, Anal. Biochem. 315 (2003) 114. [17] M. Erbs, C.C. Hoerger, N. Hartmann, T.D. Bucheli, J. Agric. Food Chem. 55 (2007) 8339. [18] G.G.C. Kuhnle, C. Dell’Aquila, Y.L. Low, M. Kussmaul, S.A. Bingham, Anal Chem. 79 (2007) 9234. [19] J. Kang, L.A. Hick, W.E. Price, Rapid Commun. Mass Spectrom. 21 (2007) 4065. [20] E. de Rijke, P. Out, W.M.A. Niessen, F. Ariese, C. Gooijer, U.A.Th. Brinkman, J. Chromatogr. A 1112 (2006) 31. [21] H.W. Kuo, W.H. Ding, J. Chromatogr. A 1027 (2004) 67. [22] A.M. Vadja, L.B. Barber, J.L. Gray, E.M. Lopez, J.D. Woodling, D.O. Norris, Environ. Sci. Technol. 42 (2008) 3407. [23] M.J. Focazio, D.W. Kolpin, K.K. Barnes, E.T. Furlong, M.T. Meyer, S.D. Zaugg, L.B. Barber, E.M. Thurman, Sci. Total Environ. 402 (2008) 201.