17β-Estradiol Degradation by TiO2 Photocatalysis as a Means of

Aug 21, 2002 - Kawasaki, Kanagawa 213-0012, Japan, Department of. Urology, School of Medicine, University of Yokohama City,. 3-9 Fukuura, Kanazawa-ku,...
0 downloads 0 Views 167KB Size
Download PDF
Environ. Sci. Technol. 2002, 36, 4175-4181

17β-Estradiol Degradation by TiO2 Photocatalysis as a Means of Reducing Estrogenic Activity Y O S H I H I S A O H K O , †,‡ K E N - I C H I R O I U C H I , † C H I S A N I W A , ‡,§ TETSU TATSUMA,| TETSUTO NAKASHIMA,⊥ TAISEN IGUCHI,# Y O S H I N O B U K U B O T A , ‡,§ A N D A K I R A F U J I S H I M A * ,†,‡ Department of Applied Chemistry, School of Engineering, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan, Kanagawa Academy of Science and Technology, KSP Bldg. East 412, 3-2-1 Sakado, Takatsu-ku, Kawasaki, Kanagawa 213-0012, Japan, Department of Urology, School of Medicine, University of Yokohama City, 3-9 Fukuura, Kanazawa-ku, Yokohama, Kanagawa 236-0004, Japan, Institute of Industrial Science, University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan, Nagareyama Higashi High School, Nazukari, Nagareyama, Chiba 270-0145, Japan, Center for Integrative Bioscience, Okazaki National Research Institutes, 38 Nishigonaka, Myoudaiji, Okazaki, Aichi 444-8585, Japan, and CREST, JST, Japan

The degradation of 17β-estradiol (E2) in water by TiO2 photocatalysis was investigated; concurrently the estrogenic activity of the treated water was evaluated during the photocatalytic reactions by an estrogen screening assay. As a result, 10-6 M of E2 was totally mineralized to CO2 in 1.0 g L-1 TiO2 suspension under UV irradiation for 3 h. 10-17β-Dihydroxy-1,4-estradien-3-one and testosteronelike species were elucidated as intermediate products by GC/MS analysis. The mechanisms of E2 degradation by TiO2 photocatalysis were discussed not only experimentally but also theoretically by calculating the frontier electron densities of the E2 molecule. On the basis of the results obtained, it was concluded that the phenol moiety of the E2 molecule, one of the essential functional groups to interact with the estrogen receptor, should be the starting point of the photocatalytic oxidation of E2. This means that the estrogenic activity should be almost lost concurrently with the initiation of the photocatalytic degradation. Actually, the estrogenic activities of the intermediate products were negligible. TiO2 photocatalysis could be applied to water treatment to effectively remove natural and synthetic estrogens without producing biologically active intermediary products.

* Corresponding author phone: +81-3-3812-9276; fax: +81-33812-6227; e-mail: [email protected]. † Department of Applied Chemistry, School of Engineering, University of Tokyo. ‡ Kanagawa Academy of Science and Technology. § University of Yokohama City. | Institute of Industrial Science, University of Tokyo. ⊥ Nagareyama Higashi High School. # Center for Integrative Bioscience, Okazaki National Research Institutes and CREST, JST. 10.1021/es011500a CCC: $22.00 Published on Web 08/21/2002

 2002 American Chemical Society

Introduction Recently, several types of chemicals in the environment have been suggested to affect human and wildlife health (1); for example, abnormal sexual development of animals (24) and the decrease in the average numbers of human spermatozoa (5, 6) are widely reported. These chemicals are assumed to disrupt normal endocrine functions through interaction with steroid hormone receptors, even at a low concentration, due to their steroid-like structures. We have recently reported for the first time that bisphenol A (BPA), which is suspected to act as an endocrine disruptor (ED), was decomposed to carbon dioxide (CO2) to lose its estrogenic activity by TiO2 photocatalytic reactions (7). We have also identified several intermediate products by liquid chromatography/mass spectrometric (LC/MS) analysis. Although these retain the phenol group, which is the most important structure for BPA to exhibit its estrogenic property, we verified that they did not possess strong estrogenic activities against recombinant yeast cells with the human estrogen receptor (hER). In the present work, we aimed at investigating TiO2 photocatalytic degradation of 17β-estradiol (E2), which is one of the basic natural estrogens and well-known to exhibit very potent estrogenic activity even at a very low concentration (∼10-9 M, in vitro (8, 9)). This concentration is about 2 orders of magnitude lower than the concentration region in which BPA exhibits the estrogenic activity (8, 9). Up to 60 µg of E2 is excreted in the urine of woman per day (10, 11). This increases to 200-400 µg during pregnancy (10, 11). Recently, E2 has been reported to be an important endocrine disruptor in the aquatic environment (12-15) and may be associated with increased incidences of hermaphrodite carp and trout in British rivers that receive significant inputs of domestic effluents (12, 13). Desbrow and colleagues reported levels of E2 in domestic sewage effluents between 1 and 50 ng L-1 (14); concentrations which were also shown to induce significant levels of vitellogenin production in male fish (15), a sensitive biomarker of exposure to estrogenic chemicals (12, 16). Also, synergistic (9) and additive (17) effects of estrogenic activity of E2 with other coexisting estrogens have been proposed and argued. Thus, it is crucially important to develop a method to reduce the E2 concentration in environmental water to an insignificant level. It is well-known that most organic compounds can be oxidized to CO2 by TiO2 photocatalytic reactions at ambient temperature and pressure because photogenerated holes exhibit strong oxidizing power (∼3.0 V vs SHE, which is ca. 1 V more positive than that of ozone) (18-20). However, the types of intermediate products that will be generated during these reactions are very important, especially for E2 degradation, because even a small structural change may cause an increase in the estrogenic activity. In the case of a highly estrogenic species such as E2, such an unexpected increase of the estrogenic activity during the degradation may result in a serious problem. However, we cannot predict whether this is the case or not on the basis of the previous investigation of BPA, because the structure of E2 is quite different from that of BPA. Thus, we must investigate anew the pathway of the photocatalytic E2 degradation by identifying the types of intermediate products to examine the applicability of TiO2 photocatalysis to the removal of E2 from, for instance, sewage. As a result, fortunately, we have verified in the present work that the series of E2 degradation reactions starts with the oxidation of the phenol moiety, which plays the most important role in the interaction with hER (21). This is VOL. 36, NO. 19, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4175

indicative that the estrogenic activity should be lost almost completely in the first step of the photocatalytic degradation of the E2 molecule. Consequently, in the present paper, we were able to demonstrate that TiO2 photocatalysis could be a useful technology for the purification of sewage containing E2, without generating any serious secondary pollution. Other natural estrogens, including estrone and estriol, and some synthetic estrogens such as ethynylestradiol (EE2) also have steroid structures, with a phenol moiety similar to E2. EE2 is contained in oral contraceptives (22, 23), and the frequent use of oral contraceptives, followed by the excretion of EE2 into urine, is assumed to lead to increases in the environmental concentration of EE2 as an ED (14, 15). However, such estrogens must be degraded in a manner similar to E2, so that the information obtained in this work should be important for removing those estrogens. Recently, King et al. (24) degraded E2 in water by TiO2 photocatalysis. Although they observed changes in the amount of E2, CO2 and other intermediate products as well as the estrogenic activity were not monitored. In the present study, we have investigated the degradation of E2 to CO2 and have identified some of the intermediates by gas chromatography/mass spectrometric (GC/MS) analysis. Molecular orbital (MO) calculations were also carried out to ascertain the position of the E2 molecule at which the photocatalytic oxidation is initiated. On the basis of the results, we discuss the interaction between the hER and the intermediates both theoretically and experimentally. Concurrently, we have monitored the estrogenic activity of the photocatalytically treated water to verify our proposed mechanism of E2 degradation.

Experimental Section Photocatalytic Oxidation of E2. E2 was purchased from Tokyo Kasei (GC grade >99%). The E2 was initially dissolved in 0.5 M aqueous NaOH and then diluted to 10-6 M followed by neutralization with equimolar HNO3. A 50-mL portion of the E2 test solution (10-6 M) was put into a Pyrex reaction vessel (100 mL). TiO2 powder (Degussa P-25, Japan Aerosil Co.) was added into the solution to give a concentration of 1.0 g L-1. Instantly, the reaction vessel was irradiated from outside with a 365-nm band-pass filtered light from a 200-W Hg-Xe lamp (Luminar Ace 210, Hayashi Tokei). The irradiation intensity was 6 mW cm-2, as determined by use of a UV radiometer (UVR-36, Topcon). The solution was stirred during UV irradiation time. Analysis of the Treated Solution. The amounts of E2 in the test solutions were evaluated with a high-performance liquid chromatograph (HPLC) equipped with an ODS column (a reversed-phase TSKgel, ODS-80Ts; 250 mm long, 4.6 mm i.d.) at 40 °C. The eluent was an acetonitrile-water mixture (6/5, v/v). E2 was detected with a Hitachi F2000 fluorescence spectrometer (excitation, 278 nm; emission, 320 nm). After the photocatalytic reaction for a given time, 1 mL of the treated suspension was subjected to HPLC analysis after removal of the TiO2 powders by centrifugation. The amount of CO2 generated during the reaction was determined by gas chromatography (GC) (Model GC-8A, Shimadzu), equipped with a Porapak-Q column, a methanizer, and a flame ionization detector, with N2 as the carrier gas. The concentration of CO2 was estimated by comparing the peak area appearing in the GC trace with that of a standard gas of CO2 (Takachiho) whose concentration was alreadyknown to be 506.2 ppmv. Since we used the methanizer after the gas components were separated through the column, the amounts of CO2 and CH4 gas generated during the reaction could be evaluated separately and accurately. To detect some intermediates by using the HPLC equipped with a UV absorption detector (UV-8020, Tosoh Co.; monitored at 275 nm), the solution was concentrated 1000-fold 4176

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 19, 2002

by solid-phase extraction with filteration disks (C18 FF, Empore Disk, 3M Co.) and eluted with 1:1 v/v acetonemethanol followed by N2 gas bubbling. In the intermediate analyses and the following bioassay for the estrogenic activity, E2 test solutions were initially prepared with ethanol, followed by dilution to 10-6 M with pure water. The initial volume of the solution was 1 L. The intermediates were analyzed with a GC/MS analyzer (QP-5050A, Shimadzu), equipped with a fused silica capillary column (DB-5, J&W Scientific; 30 m long, 0.25 mm i.d., 0.4 µm film thickness). A split-splitless injection port was used in the splitless mode at high pressure (350 kPa). The column temperature was programmed as follows: 2 min at 50 °C, 20 °C min-1 to 130 °C, 10 °C min-1 to 300 °C, and 10 min at 300 °C. The helium gas flow rate was 1.63 mL min-1 (at 50 °C). Electron impact was used for ionization of samples. Some of the intermediates were identified by use of an identification program of the U.S. National Institute of Standards and Technology (NIST) library (Shimadzu). Calculation of the Frontier Electron Density. The ab initio molecular orbital (MO) calculations were carried out by using Gaussian 98 program (Gaussian, Inc.) on a personal computer. Structures were optimized with the STO-3G basis set at the level of the unrestricted Hartree-Fock (UHF) for all calculations, and then the frontier electron densities (FEDs) of the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) were calculated. We obtained the values of 2FEDHOMO2 and (FEDHOMO2 + FEDLUMO2) to predict the reaction sites for electron extraction and radical attack, respectively (25, 26). Evaluation of Estrogenic Activities. Transcriptional estrogenic activities in response to a recombinant yeast-based estrogen assay (27, 28) were evaluated for the treated waters as follows. A single yeast colony (S. cerevisiae BJ3505) (29) was grown overnight (18-20 h) at 30 °C on a growth medium (0.67% yeast nitrogen base, 2% dextose) supplemented with lysine and histidine. Then 250 µL of cell suspension, 50 µL of 10 mM CuSO4, and 50 µL of E2 sample solution were added to 10 mL of the fresh medium. After the incubation treatment at 30 °C (for 18-20 h), a 150-µL aliquot of the yeast cell suspension was added to 250 µL of saline, and then the cells were collected by centrifugation (14000 rpm for 10 min) and permeabilized by the addition of 700 µL of Z-buffer (60 mM Na2HPO4, 40 mM NaH2PO4, 10 mM KCl, 1 mM MgSO4, and 35 mM β-mercaptoethanol), 25 µL of 0.1% sodium dodecyl sulfate, and 50 µL of CHCl3, and mixed for 30 s. It was then brought to room temperature with the addition of 200 µL of o-nitrophenyl β-D-galactopyranoside (4 mg mL-1 in Z-buffer) for coloring. After 20 min, the coloring reactions were terminated by the addition of 1 M Na2CO3, and the absorbance at 420 nm was measured. β-Galactosidase activity was determined by the formula [A420/(A600 of cells × reaction time (min) × volume of cells (mL))] × 1000.

Results and Discussion HPLC/Fluorescence Analysis of E2. Photocatalytic degradation of E2 in the TiO2 suspension was monitored by HPLC/ fluorescence analysis. More than 99% of the E2 was degraded by TiO2 photocatalysis after 30 min of UV irradiation with first-order kinetics, as shown in Figure 1. Except for E2, any other components were not detected with the fluorescence spectrometer during the photocatalytic reactions. In a control experiment, the solution was irradiated without TiO2 photocatalyst, and the disappearing of E2 was negligible. In another control experiment, in which the solution containing TiO2 photocatalyst was not irradiated, a half amount of E2 was adsorbed on TiO2 in 1 h. Although some intermediate species were produced during the photocatalytic reactions as we discuss later, the fact that the intermediate products did not fluoresce indicates that they did not have the phenol

E2 plus generated CO2 were also plotted in the same figure. This index reached the maximum value (∼60% of N0) after about 20 min of UV irradiation. If the photoinduced total mineralization of E2 is given by ref (30)

C18H24O2 + 34H2O + 92h+ f 18CO2 + 92H+

FIGURE 1. Changes in the mass balance of carbon atom during photocatalytic degradation of E2 (initial concentration, 1 µM) in a vigorously stirred TiO2 (1.0 g L-1) suspension (50 mL). UV intensity was 6 mW cm-2. The amounts of carbon corresponding to E2 (NE2; b) and CO2 (NCO2; 9) were measured by HPLC/fluorescence analysis and GC analysis, respectively. The initial amount of carbon (N0) was 0.9 µmol (18 × 1 µM × 50 mL). Carbon balances (N0 - [NE2 + NCO2]; 2) were also plotted. moiety in their structures. Hence, it follows that the phenol moiety but not the other positions of E2 molecule was probably oxidized at first by TiO2 photocatalysis. It is well-known that the phenol moiety (or A-ring) of E2 plays a key role in the interaction with the hER (21) as shown in Figure 2a. In practice, most EDs with potent estrogenic activity, such as EE2 and diethylstilbestrol, have phenol groups. In contrast, other compounds without phenol groups generally exhibit lower estrogenic activity, if any. For example, testosterone (TS), which has a similar steroid structure with a cyclohexenone moiety instead of the phenol moiety, exhibits much weaker estrogenic activity than E2 (8). Thus, the result allows us to predict that the photocatalytically generated products are also less active than E2. GC Analysis of E2. Next, we measured the amount of CO2 generated as a result of the E2 degradation. As shown in Figure 1, of which the vertical axis corresponds to the amounts of each compound expressed as µmol of carbon, the amount gradually increased, and a constant value was finally reached after 3 h. The final carbon amount of CO2 agreed with the initial carbon amount of E2 (N0); i.e., the E2 was completely decomposed to CO2. The carbon balances, differences between N0 and the sum of the amounts of carbon corresponding to undecomposed

(1)

the apparent quantum yield of CO2 generation was ca. 6.7% in the first 10 min of UV irradiation. Coleman and colleagues reported an apparent quantum yield of 0.41% for photocatalytic degradation of E2 over TiO2 powder immobilized on a Ti-6Al-4V alloy plate (24). The difference can be explained in terms of the greater TiO2 surface area and more rapid mass transfer in our suspension system. HPLC/UV Analysis of Intermediate Products. Since no products were detected by the fluorescence detector, we combined the HPLC system with a UV absorption detector. In this experiment, 1 L of the TiO2 suspension containing E2 was subjected to UV-irradiation and the irradiated solutions were 1000-fold concentrated, because the sensitivity of the UV detector is much lower than that of the fluorescence detector. As Figure 3 shows, a longer time was needed to degrade E2 than the corresponding data in Figure 1, due to the 20-fold larger suspension volume, which decreases the number of photons irradiated per unit amount of the TiO2 particles. Two peaks (referred to as a and b based on their order of elution) appeared in the chromatogram. Peak a appeared at 9.2 min, the peak for E2 at 11.0 min, and peak b at 33.0 min, respectively, indicating that the compound for peak a is more hydrophilic, and the compound for peak b is less hydrophilic than E2. Figure 3 depicts changes in the peak heights relative to the initial E2 peak in the course of the photocatalytic reactions. All of the relative peak heights for the products were lower than 0.04, and all of the peaks disappeared during 3.5 h of UV irradiation. Considering the mass balance shown in Figure 1 (∼60% of the carbon atoms exist as intermediates at the most) and the very low relative peak heights in Figure 3, there may be many other components insensitive to UV. However, the undetectable components should not consist of such chromophores as phenol groups or carboxyl groups, so that they are not expected to have potent estrogenic activity. GC/MS Analysis of Intermediate Products. To elucidate the structures of the intermediates, the concentrated sample solution was subjected to GC/MS analysis. Three compara-

FIGURE 2. Schematic illustration of binding models of (a) E2 (ref 30) and (b) 10E-17β-dihydroxy-1,4-estradien-3-one (DEO) with the human estrogen receptor. DEO is one of the intermediates produced during the photocatalytic reactions as shown in Figure 4c. VOL. 36, NO. 19, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4177

FIGURE 3. Time courses of the three peaks appearing in the HPLC trace detected at 275 nm with a UV absorption detector during the TiO2 photocatalytic degradation of E2 in water (0, peak a; 4, peak b; listed in order of elution; b, E2). Each height of peaks a and b was normalized by that of the initial E2 peak. The initial volume of the E2 test solution was 1 L, and the treated solution was 1000fold concentrated by solid-phase extraction before the measurements. tively large peaks appeared in the chromatogram (referred to as c, d, and e based on their order of elution). However, the peak heights relative to the initial E2 height were low, at most ∼5% for peak c, ∼3% for peak d, and ∼1% for peak e. Retention times were as follows: peak c, 21.7 min; peak d, 22.1 min; E2, 24.0 min; peak e, 24.7 min. Mass spectra corresponding to peak c-e are shown in Figure 4. Interestingly, all of the mass spectra look similar

to each other. The identification program of the NIST library suggested that the spectra of peaks c, d, and e can be ascribed to 10-17β-dihydroxy-1,4-estradien-3-one (DEO), androsta4,16-dien-3-one (ADO), and testosterone (TS), respectively. The similarity values evaluated by the program were relatively high: 82 and 71 for peaks c and d, respectively; 63 for peak e. Their original mass spectra are shown in Figure 4 with their molecular structures. The spectra had a base peak at m/z ) 147 in common, which is assignable to a fragment generated by splitting the original molecules at dashed line (i) (31). The mass peak at m/z ) 124 in spectrum c is assignable to another fragment generated by splitting at dashed line (i) (32). The peak at m/z ) 246 (M - 42; where M is the mass number of the original molecule) in spectrum c can be explained by the loss of CH2dCdO (mass number ) 42) fragment by scission at dashed line (ii). This peak is typical of a steroid having an A-ring (see Figure 2a) with an unsaturated carbonyl structure (32). Comparing the molecular structure of E2 and DEO, it can be deduced that the quinone-like moiety of DEO was generated by oxidation of the phenol moiety of E2 in the first step of the photocatalytic reaction. This corresponds to the result obtained by the HPLC/fluorescence analysis. It is reasonable to consider that the interaction between DEO and hER must be weakened due to the conversion of the phenol moiety of E2 to a quinone-like moiety, so that DEO should exhibit much weaker estrogenic activity than that of E2 (Figure 2b). Spectrum d and e also show mass peaks at m/z ) M - 42, indicating that the phenolic OH group of E2 was oxidized to the carbonyl group (>CdO). Thus, these inter-

FIGURE 4. Mass spectra corresponding to peaks c-e appearing in the GC/MS trace and those suggested by the NIST identification program with corresponding molecular structures. Dashed lines indicate the splitting positions in electron impact ionization of MS measurements (see text). 4178

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 19, 2002

SCHEME 1. Suggested Mechanism of E2 Degradation by TiO2 Photocatalysis

mediates should also exhibit weaker estrogenic activities than E2. Incidentally, the A-ring of E2 should be reduced partially to generate ADO and TS. However, such a reduction is unlikely to occur during the photocatalytic oxidation. Since almost the same mass spectra as spectra d and e should be obtained even if the C1-C2 bonds (see Figure 2a for atom numbering) of ADO and TS, respectively, are double bonds, the intermediates corresponding to spectra d and e may still have C1dC2 double bonds. Among DEO, ADO, and TS, only TS was available, and it was subjected to GC/MS analysis. As a result, we found that the retention time of TS was 24.7 min, slightly different from that of spectrum c (24.5 min). This difference might reflect the difference in the molecular structure: C1-C2 single or C1dC2 double bond. In addition, the partly hydrophobic structure of E2 molecule is also important for the interaction with the estrogen receptor (Figure 2a) (21). Hence, the OH group of DEO at C10 atom would also work to decrease the estrogenic activity by increasing hydrophilicity in this region of the molecule (Figure 2b). Frontier Electron Density of E2. To support our conclusion that the photocatalytic degradation of E2 is initiated with the phenol-moiety oxidation, theoretical MO calculations were carried out for the E2 molecule. The evaluated frontier electron densities (FEDs) are listed in Table 1. We predicted the first reaction site of E2 molecule at which an electron would be extracted by a hole on TiO2 (h+) or •OH radical (generated by the reaction: H2O + h+ f •OH + H+), on the basis of 2FEDHOMO2 values (see Experimental). These values for E2 were found to be high at the phenol moiety, especially at the C10 and C3 atoms. Consequently, we concluded that the C10 or C3 atoms should be the sites at which the first electron is extracted. DEO will be produced in the subsequent reactions, as shown in Scheme 1 (E2 f 1 f 2 f 3 f DEO). A similar reaction route to the hole-oxidation reactions (E2 f 1 f 2) has been reported by other groups for the photocatalytic oxidation of phenols (33, 34), in which the details of the reaction with O2•- (2 f 3 f DEO) were reported by von Sonntag et al. (35). On the other hand, the reaction may be initiated by addition of the photocatalytically generated •OH to E2. The site for the first addition was discussed on the basis of FEDHOMO2 + FEDLUMO2 values (see Experimental). These values

TABLE 1. Frontier Electron Densities (FED) on Atoms of E2 Calculated by Using Gaussian 98 Program with the STO-3G Basis Set at the Level of the Unrestricted Hartree-Fock

a

atoma

2FEDHOMO2

FEDHOMO2 + FEDLUMO2

C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 O3 O17

0.100 0.210 0.364 0.153 0.110 0.003 0.001 0.009 0.015 0.508 0.014 0.002 0.000 0.002 0.000 0.000 0.000 0.000 0.374 0.000

0.333 0.448 0.189 0.345 0.431 0.003 0.002 0.005 0.008 0.258 0.007 0.001 0.000 0.001 0.000 0.000 0.000 0.000 0.188 0.000

See Figure 2a for atom numbering.

were found to be high at the phenol moiety, especially at the C2 and C5 atoms. If the first addition of •OH radical occurs at the C2 atom, the corresponding o-hydroquinone could be produced by a subsequent reaction with O2, as shown in Scheme 1 (E2 f 4 f 5 f 6) (36). However, such an intermediate was not detected in the GC/MS analysis. On the other hand, if a dehydration reaction follows the •OH attack (37), DEO could be produced, as shown in Scheme 1 (E2 f 4 f 2 f 3 f DEO). Besides the reaction with O2•-, DEO could be produced by direct attack of •OH radical at the C10 atom of the resonance structure 7 (37). In this case, whichever carbon atoms of the A-ring are attacked by an •OH radical, the same product (DEO) can be expected. In either case, the calculation supported our conclusion that a series of E2 degradation reactions started with the oxidation of the phenol moiety. Incidentally, the electronic structure of E2 molecule calculated in the present work was not so different from results obtained by another MO calculation method (38). VOL. 36, NO. 19, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4179

tions did not exhibit any potent estrogenic activity in the treated water; and (4) the phenol moiety of E2 molecule, one of the essential functional groups to interact with the hER, should be the starting point of the photocatalytic oxidation of E2, so that the estrogenic activity should be almost lost concurrently with the initiation of the photocatalytic degradation. On the basis of the results obtained in the present work, it is expected that other steroidal estrogens will be degraded in a similar manner. Thus, we conclude that TiO2 photocatalysis could be applied to water treatment as a novel method for removing natural and synthetic estrogens effectively without generating biologically active intermediates.

Acknowledgments

FIGURE 5. Estrogenic activity of aqueous solutions containing various amounts of E2. The estrogenic activities were measured by means of the yeast hybrid assay (see Experimental). (O) E2 solutions without photocatalytic treatment; (b) E2 solutions treated photocatalytically for 20 min and diluted. Experimental conditions of photocatalysis were the same as those for Figure 3. See text for further detail. The OH group at C17 atom may also be oxidized easily to a carbonyl group. The reaction rate of secondary-alcohol oxidation by an •OH radical (e.g. 2.1 × 109 M-1 s-1: 3-pentanol) is similar to that of phenol oxidation by an •OH radical (1.41.8 × 1010 M-1 s-1) (39). In this case, the estrogenic activity should be suppressed also, because the estrogen acceptor may not interact very well with the carbonyl group. Actually, estrone, having a carbonyl group at the C17 atom, exhibits much smaller estrogenic activity than E2 (8, 40). Although such an intermediate was not detected, the mass spectrum shown in Figure 3d suggests that the C17 site was also attacked. Estrogenic Activity. To ensure further the conclusion derived experimentally and theoretically that the estrogenic activity of E2 is lost in the first step of the photocatalytic reactions, we have evaluated the transcriptional estrogenic activities in response to hER in a yeast assay system. First, the activity-concentration relationship (doseresponse curve) for E2 was investigated (Figure 5, O). A peak of the dose response curve for E2 was obtained at around 10-9 M. This is in good agreement with other reports (8, 9). Next, we prepared an E2 solution treated photocatalytically for 20 min. The amounts of intermediate products reach almost the maximum values in the 20-min treatment, as can be seen in Figure 3. The treated solution, which contains about 0.26 µM E2, was diluted by about 2-6 orders of magnitude, and the estrogenic activities of the diluted solutions were evaluated and plotted in Figure 5 (b) against the E2 concentrations in the diluted solutions. If the intermediate products in the treated solution have appreciable estrogenic activities, the plot should exhibit higher values than the original dose-response curve in a certain concentration region. However, we found that the curve obtained for the treated solution (Figure 5, b) was in good agreement with the doseresponse curve. This means that the estrogenic activities of the intermediate products are negligible and that there is no secondary risk to increase the estrogenic activity as a result of the photocatalytic degradation of E2 in water under the present experimental conditions. In summary, we confirmed the following in the present work: (1) E2 in water was totally mineralized as a result of the photocatalytic reactions; (2) 10-17β-dihydroxy-1,4estradien-3-one (DEO) and testosterone-like species were deduced as intermediate products by GC/MS analysis; (3) the intermediates produced during the photocatalytic reac4180

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 19, 2002

The authors are grateful to Mr. H. Sakashita for help with GC/MS measurements. The authors thank Dr. D. A. Tryk for carefully reading the manuscript. This work was supported by a Grant-in-Aid for Scientific Research on Priority Areas (Area No. 417 for A.F., Y.K., and T.T.), Tokyo Ohka Foundation for the Promotion of Science and Technology (for Y.K.), and Joint Research Projects for Regional Intensive of Kanagawa Prefecture.

Literature Cited (1) ) Colborn, T.; Dumanoski, D.; Myers, J. P. Our Stolen Future; Dutton: New York, 1996. (2) Guillette, L. J.; Gross, T. S.; Gross, D.; Rooney, A. A.; Percival, H. F. Environ. Health Perspect. 1994, 102, 680. (3) Guillette, L. J.; Gross, T. S.; Gross, D.; Rooney, A. A.; Percival, H. F. Environ. Health Perspect. 1995, 103, 31. (4) Fry, D. M. Environ. Health Perspect. 1995, 103, 106. (5) Carlsen, E.; Giwercman, A.; Keiding, N.; Skakkebaek, N. E. Br. Med. J. 1992, 305, 609. (6) Auger, J.; Kunstmann, J. M.; Czyglik, F.; Jouannet, P. N. Engl. J. Med. 1995, 332, 281. (7) Ohko, Y.; Ando, I.; Niwa, C.; Tatsuma, T.; Yamamura, T.; Nakashima, T.; Kubota, Y.; Fujishima, A. Environ. Sci. Technol. 2001, 35, 2365. (8) Gaido, K. W.; Leonard, L. S.; Lovell, S.; Gould, J. C.; Babai, D.; Portier, C. J.; McDonnell, D. P. Toxicol. Appl. Pharmacol. 1997, 143, 205. (9) Arnold, S. F.; Klotz, D. M.; Collins, B. M.; Vonier, P. M.; J., G. L.; McLachlan, J. A. Science 1996, 272, 1489. (10) Adlercreutz, H.; Fostis, T.; Bannwart, C.; Hamalainen, E.; Bloigu, S.; Ollus, A. J. Steroid Biochem. 1992, 17, 639. (11) Pharmacentical Affairs and Food Sanitation Council, Ministry of Health, Labor and Welfare, http://www1.mhlw.go.jp/houdou/ 1103/h0303-1_a_15.html (accessed June 2002) (in Japanese). (12) Purdom, C. E.; Hardiman, P. A.; Bye, V. J.; Eno, N. C.; Tyler, C. R.; Sumpter, J. P. Chem. Ecol. 1994, 8, 275. (13) Guillette, L. J.; Arnold, S. F.; MacLachlan, J. A. Anim. Reprod. Sci. 1996, 42, 13. (14) Desbrow, C.; Routledge, E. J.; Brighty, G. C.; Sumpter, J. P.; Waldock, M. Environ. Sci. Technol. 1998, 32, 1549. (15) Routledge, E. J.; Sheahan, D.; Desbrow, C.; Brighty, G. C.; Waldock, M.; Sumpter, J. P. Environ. Sci. Technol. 1998, 32, 1559. (16) Harries, J. E.; Sheahan, D. A.; Jobling, S.; Matthiessen, P.; Neall, P.; Sumpter, J. P.; Tylor, T.; Zaman, N. Environ. Toxicol. Chem. 1997, 16, 534. (17) Rajapakse, N.; Ong, D.; Kortenkamp, A. Toxicol. Sci. 2001, 60, 296. (18) Ohko, Y.; Tryk, D. A.; Hashimoto, K.; Fujishima, A. J. Phys. Chem. B 1998, 102, 2699. (19) Fujishima, A.; Rao, T. N.; Tryk, D. A. J. Photochem. Photobiol. C 2000, 1, 1 and references therein. (20) Fujishima, A.; Rao, T. N.; Tryk, D. A. Electrochim. Acta 2000, 45, 4683 and references therein. (21) Brzozowski, A. M.; Pike, A. C. W.; Dauter, Z.; Hubbard, R. E.; Bonn, T.; Engstrom, O.; Ohman, L.; Green, G. L.; Gustafsson, J. A.; Carlquist, M. Nature (London) 1997, 389, 753. (22) Rathner, M.; Sonneborn, M. Forum Staedte-Hyg. 1979, 30, 45. (23) The Top 200 Prescriptions* Ranked/Alpha, www.rxlist.com/ top200.htm. (accessed June 2002). (24) Coleman, H. M.; Eggins, B. R.; Byrne, J. A.; Palmer, F. L.; King, E. Appl. Catal. B: Environ. 2000, 24, L1.

(25) Fukui, K.; Yonezawa, T.; Shingu, H. J. Chem. Phys. 1952, 20, 722. (26) Fukui, K.; Yonezawa, T.; Nagata, C. J. Chem. Phys. 1954, 22, 1433. (27) Coldham, N. G.; Dave, M.; Sivapathasundaram, S. Environ. Health Perspect. 1997, 105, 734. (28) Collins, B. M.; McLachlan, J. A.; Arnold, S. F. Steroids 1997, 62, 365. (29) The recombinant yeast of S. cerevisiae BJ3505 was expressed as [Mat R, pep4::HIS3, prb-∆1, 6R, his3-∆200, lys2-801, trpl-∆101, ura3-52(can1)] with an expression plasmid containing the complementary DNA of hER and a reporter plasmid containing estrogen response elements linked to the lacZ gene. (30) Sopyan, I.; Watanabe, M.; Murasawa, S.; Hashimoto, K.; Fujishima, A. J. Photochem. Photobiol. A: Chem. 1996, 98, 79. (31) Friedland, S. S.; Lane, G. H. Anal. Chem. 1959, 31, 169. (32) Budzikiewicz, H.; Dierassi, C.; Williams, D. H. In Structure elucidation of natural products by mass spectrometry vol. II: steroids, terpenoids, sugars, and miscellaneous classes; HoldenDay, Inc.: San Francisco, 1964; pp 50 (Chapters 19 and 20). (33) Richard, C.; Boule, P. New J. Chem. 1994, 18, 547.

(34) d'Hennezel, O.; Pichat, P.; Ollis, D. F. J. Photochem. Photobiol. A: Chem. 1998, 118, 197. (35) Jin, F.; Leitich, J.; von Sonntag, C. J. Chem. Soc., Perkin Trans. 2 1993, 1583. (36) Mvula, E.; Schuchmann, M. N.; von Sonntag, C. J. Chem. Soc., Perkin Trans. 2 2001, 264. (37) Peiro, A. M.; Ayllon, J. A.; Peral, J.; Domenech, X. Appl. Catal. B: Environ. 2001, 30, 359. (38) Kubli-Garfias, C. J. Mol. Struct. (Thermochem) 1998, 452, 175. (39) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. J. Phys. Chem. Ref. Data 1988, 17, 744. (40) Arnold, S. F.; Bergeron, J. M.; Tran, D. Q.; Collins, B. M.; Vonier, P. M.; Crews, D.; Toscano, W. A.; McLachlan, J. A. Biochem. Biophys. Res. Commun. 1997, 235, 336.

Received for review December 28, 2001. Revised manuscript received July 3, 2002. Accepted July 8, 2002. ES011500A

VOL. 36, NO. 19, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4181