Spectroscopic Characterization of the 4-Hydroxy Catechol Estrogen

Specifically, catechol estrogen quinones (CEQs) derived from the ... CEQs are also conjugated with GSH, a reaction that prevents damage to DNA, provid...
3 downloads 0 Views 127KB Size
Download PDF
304

Chem. Res. Toxicol. 2003, 16, 304-311

Spectroscopic Characterization of the 4-Hydroxy Catechol Estrogen QuinonessDerived GSH and N-Acetylated Cys Conjugates Ryszard Jankowiak,*,† Yuri Markushin,† Ercole L. Cavalieri,‡ and Gerald J. Small† Ames Laboratory, U.S. Department of Energy and Department of Chemistry, Iowa State University, Ames, Iowa 50011, and Eppley Institute for Research in Cancer, University of Nebraska Medical Center, Omaha, Nebraska 68198 Received September 27, 2002

Estrogens, including the natural hormones estrone (E1) and estradiol (E2), are thought to be involved in tumor induction. Specifically, catechol estrogen quinones (CEQs) derived from the catechol estrogens 4-hydroxyestrone (4-OHE1) and 4-hydroxyestradiol (4-OHE2) react with DNA and form DNA adducts (Cavalieri, E. L., et al. (1997) Proc. Natl Acad. Sci. U.S.A. 94, 10037). CEQs are also conjugated with GSH, a reaction that prevents damage to DNA, providing biomarkers of exposure to CEQs. Current detection limits for these analytes by HPLC with multichannel electrochemical detection are in the picomole range (Devanesan, P., et al. (2001) Carcinogenesis 22, 489). To improve the detection limit of CEQ-derived conjugates, spectrophotometric monitoring was investigated. Fluorescence and/or phosphorescence spectra of the 4-OHE1, 4-OHE2, Cys, N-acetylcysteine (NAcCys), 4-OHE1-2-SG, and 4-OHE2-2-SG conjugates and their decomposition products 4-OHE1-2-NAcCys and 4-OHE2-2-NAcCys were obtained at 300 and 77 K. It is shown that (i) 4-OHE1- and 4-OHE2-derived SG and NAcCys conjugates are weakly fluorescent at 300 K (with the emission maximum at 332 nm) but strongly phosphorescent at 77 K; (ii) Cys and NAcCys exhibit fluorescence and phosphorescence only at 77 K; and (iii) 4-OHE1 and 4-OHE2 are weakly fluorescent at 300 and 77 K and not phosphorescent. The phosphorescence spectra of SG and NAcCys conjugates are characterized by a weak origin band at ∼383 nm and two intense vibronic bands at 407 and 425 nm. After they are cooled from 300 to 77 K, the total luminescence intensity of SG and NAcCys conjugates increases by a factor of ∼150 predominantly due to phosphorescence enhancement. Theoretical calculations revealed, in agreement with the experimental data, that the lowest singlet (S1) and triplet (T1) states of 4-OHE2-2-NAcCys are of n,π* and π,π* character, respectively, leading to a large intersystem crossing yield and strong phosphorescence. The limit of detection (LOD) for CEQ-derived conjugates, based on phosphorescence measurements, is in the low femtomole range. The concentration LOD is approximately 10-9 M. Therefore, we propose that capillary electrophoresis interfaced with low temperature phosphorescence detection can be used to test for human exposure to CEQs by analyzing urine.

1. Introduction Recent studies suggest that estrogens can become endogenous carcinogens capable of tumor initiation by mutation of critical genes (1, 2). It has been shown that estrogens are converted to catechol estrogens, which, in turn, can be metabolized to ultimate carcinogenic forms, i.e., CEQs1 (3), which can induce mammary, pituitary, * To whom correspondence should be addressed. E-mail: jankowiak@ ameslab.gov. † Iowa State University. ‡ University of Nebraska Medical Center. 1 Abbreviations: CE, capillary electrophoresis; CEQ, catechol estrogen quinone(s); DMSO, dimethyl sulfoxide; E1, estrone; E2, estradiol; E2-3,4-Q, estradiol-3,4-quinones; GSH, glutathione reduced form; 2-OH-CE, 2-hydroxy catechol estrogen; 4-OHE1, 4-hydroxyestrone; 4-OHE2, 4-hydroxyestradiol; 4-OHE1-2-SG, 4-hydroxyestrone-derived glutathione conjugate; 4-OHE2-2-SG, 4-hydroxyestradiol-derived glutathione conjugate; 4-OHE1-2-NAcCys, 4-hydroxyestrone-derived Nacetylcysteine conjugate; 4-OHE2-2-NAcCys, 4-hydroxyestradiolderived N-acetylcysteine conjugate; MM, molecular mechanics; NAcCys, N-acetylcysteine; QM, quantum mechanics; S1, lowest excited singlet state; T1, lowest excited triplet state.

cervical, and uterine tumors in rats, mice, and guinea pigs (4). Exposure to estrogens is also a risk factor for breast and other human cancers (5). The natural estrogens, E1 and E2, are metabolized at the 2- or 4-position with formation of catechol estrogens. Major metabolic pathways of E1 and E2 are discussed in detail in ref 1. Catechol estrogens are oxidized to CEQ by peroxidases and cytochrome P-450 (3, 6-8). Liehr and Roy (9) and Nutter et al. (10) have shown that redox cycling of CEQ and the corresponding semiquinones can also generate hydroxyl radicals that cause DNA damage. CEQ derived from the catechol estrogens, 4-OHE1 and 4-OHE2, reacts with deoxyribonucleosides and DNA to form depurinating adducts at the N7 position of Gua and N3 position of adenine (1, 11-14). CEQ-derived GSH (γ-glutamyl-L-cysteinylglycine) conjugates were also identified in in vivo experiments (1518) and are considered to be potentially useful biomarkers for catechol estrogen-induced DNA damage and risk of breast and other cancers. Conjugation with GSH

10.1021/tx020088p CCC: $25.00 © 2003 American Chemical Society Published on Web 01/31/2003

Derived GSH and N-Acetylated Cys Conjugates

Figure 1. Chemical structures of 2-OHE1-, 2-OHE2-, 4-OHE1-, 4-OHE2-, and 4-OHE1/4-OHE2-derived SG and NAcCys conjugates.

prevents damage to DNA (19), such conjugation being one of the most important detoxification pathways in biological systems. A large number of electrophilic compounds conjugate with GSH nonenzymatically or, more effectively, via S-transferase-catalyzed reactions (1, 19). Therefore, the reaction of CEQs with various sulfur nucleophiles, RHS, in which R is the Cys, the NAcCys, or the GSH moiety, is of great interest in carcinogenesis. Once CEQ conjugates are formed, catabolism occurs via mercapturic acid biosynthesis. First, the glutamyl moiety of the GSH conjugate is removed by transpeptidation, catalyzed by γ-glutamyl transpeptidase. Then, the cysteinylglycine derivative is hydrolyzed to yield the Cys conjugate. The final step consists of acetylation to NAcCys conjugate and excretion in urine (1, 16, 20). Therefore, identification and quantitation of CEQ conjugates in urine have potential for assessment of the level of CEQ formed. Schematic structures of 2-OHE1-, 2-OHE2-, 4-OHE1-, 4-OHE2-, and 4-OHE1/4-OHE2-derived SG and NAcCys conjugates are shown in Figure 1. Natural estrogens possess an unsaturated ring and phenolic substitution that results in UV absorption and very weak fluorescence at room temperature (21, 22). As a result, identification of 4-OHE1- and 4-OHE2-derived DNA adducts has been accomplished following derivatization with a suitable fluorescent marker. For example, selectively excited fluorescence line-narrowing spectroscopy has been used for spectral characterization of fluorescently labeled CEQ-derived N7-Gua adducts and their identification in rat mammary gland tissue (13). Measurements of Cys, GSH, and homocysteine in human plasma were done by HPLC after derivatization with monobromobimane (23) or 7-fluoro-2,1,3-benzodiazole-4sulfonamide (24) and conventional fluorescence detection. Recent analysis of potential biomarkers of estrogeninitiated cancer in urine and the kidney of Syrian golden hamsters treated with 4-OHE2 revealed that HPLC with electrochemical detection (with picomole detection limit) also provides high specificity (1, 15, 16). Nagakomi and Suzuki developed a protocol for the quantitation of NAcCys conjugates in urine of rats and hamsters using an immunoaffinity column (20).

Chem. Res. Toxicol., Vol. 16, No. 3, 2003 305

This paper presents, for the first time, fluorescence and phosphorescence spectra of the NAcCys, 4-OHE1-2-SG, 4-OHE2-2-SG, 4-OHE1-2-NAcCys, and 4-OHE2-2-NAcCys analytes. It is shown that the above molecules are strongly phosphorescent when excited with a UV laser at 257 nm at low temperatures, which allows for sensitive detection and quantitation at low concentrations. For completeness, fluorescence spectra of 4-OHE1 and 4-OHE2 are briefly discussed and compared with the literature data obtained for E1 and E2 (21, 22). The main goals of this research were to provide a library of spectral fingerprints for future in vivo and in vitro studies of 4-OHE1-2-NAcCys, 4-OHE2-2-NAcCys, 4-OHE1-2-SG, and 4-OHE2-2-SG conjugates by CE with on-line detection by low temperature phosphorescence and to provide more insight on the origin of strong phosphorescence emission. We have focused on the characterization of 4-OHE1-2-NAcCys and 4-OHE2-2NAcCys conjugates that could be used as urinary biomarkers for CEQ formation. Detection limits for the analytes of interest, obtained by low temperature laserexcited luminescence, are compared with those obtained by HPLC interfaced with multichannel electrochemical detection. The potential of a phosphorescence-based approach for the determination of CEQ-derived conjugates in human urine and/or human tissue extracts is briefly addressed.

2. Materials and Methods Caution: CEQs are hazardous chemicals and should be handled carefully in accordance with NIH guidelines. 2.1. Chemicals and Analyte Purity. 4-OHE1 and 4-OHE2 were synthesized according to Dwivedy et al. (3). The 4-OHE1and 4-OHE2-derived SG and NAcCys conjugate standards were synthesized as previously described (19). Structural analysis of the standard conjugates was accomplished via NMR spectroscopy (19). MS was used in two ways: first, exact mass measurement at 10 000 resolving power was used to confirm molecular formulas and second, tandem four sector MS confirmed the structures determined by NMR (19). Cys, NAcCys, and spectrophotometric grade ethanol were purchased from Aldrich. Ultrapure grade glycerol was obtained from Spectrum Chemical (Gardena, CA). The purity of standards for CEQ-derived conjugates, originally separated by HPLC, was verified in our laboratory by CE, which possesses higher separation power than HPLC. Separation conditions and data obtained by CE with field-amplified sample stacking (FASS) and absorbance detection and CE interfaced with low temperature, laser-excited, phosphorescence detection will be published elsewhere (25). Here, it suffices to say that CE verified the high purity of the standards; thus, one can be confident that the luminescence spectra correspond to pure analytes. All CEQ-derived congugates were kept for longer storage at -80 °C under inert atmosphere (N2 or Ar) since they are heat and oxygen sensitive. Special care was taken since the above conjugates are susceptible to oxidation in air in the presence of small amounts of cations to give (via a mercaptide) disulfides (2). Therefore, samples were dissolved in methanol:buffer (80:20), with the following buffer content: 0.1 M NH4CH3COO and 1 mg/L ascorbic acid in Nanopure water, pH 4.5. 2.2. Luminescence and Absorption Spectroscopy. Luminescence spectra were obtained using an excitation wavelength of 257 nm of a Lexel 95-SHG-257 CW laser. Emission was dispersed by a model 218 0.3 m monochromator (McPherson, Acton, MA), equipped with a 300 G/mm grating, providing a resolution of ∼1 nm and a spectral window of approximately 200 nm. Spectra were detected with an intensified CCD camera (Princeton Instruments, Trenton, NJ) using gated and nongated modes of detection. A fast shutter, operated by a Uniblitz driver

306

Chem. Res. Toxicol., Vol. 16, No. 3, 2003

control (model SD-12 2B), was synchronized with the CCD camera (ICCD-1024 MLDG-E1) and used for time-resolved phosphorescence measurements. Using this setup, time-resolved phosphorescence spectra (∼10-4-10-8 M analyte concentrations) could be measured in 0.5 s intervals with a gate width of 0.5 s. Phosphorescence lifetimes and phosphorescence excitation spectra (for more concentrated samples; ∼10-4 M) were measured at 77 K using a Cary Eclipse fluorescence spectrophotometer (Varian, Inc., Palo Alto, CA). To ensure good glass formation, glycerol (50 vol %) was added to the samples in buffer just prior to cooling to 77 K in a liquid nitrogen optical cryostat with suprasil optical windows. Samples (ca. 20 µL) were contained in suprasil tubes (2 mm i.d.). In addition, ethanol and water matrices were used to study the shift of the fluorescence band(s) in solvents of different polarity. Luminescence spectra of 4-OHE1- and 4-OHE2-derived NAcCys and SG conjugates were obtained for 10-5-10-8 M concentrations. All spectra were corrected for background luminescence. Absorption spectra were measured at room temperature using a UV/vis spectrometer (Perkin-Elmer Lambda 18; Perkin-Elmer Instruments, Wellesley, MA). 2.3. Molecular Modeling and Theoretical Calculation. Conformational analysis of the 4-OHE2-2-NAcCys conjugate was carried out using methods of MM, and energy calculations were performed with HyperChem’s molecular modeling program (Release 7.0 for Windows, Hypercube Inc.). HyperChem’s force field (MM+) developed for organic molecules (26, 27) was employed utilizing default parameters. As the starting structure for the 4-OHE2-2-NAcCys conjugate, a model-built structure was used. The Polak-Ribiere algorithm (in vacuo) was used for MM optimization; the structure was refined until the rms gradient was less than 0.001 kcal/mol. Electrostatic contributions were evaluated by defining a set of bond dipole moments for polar bonds. To calculate thermodynamically favored conformations of the 4-OHE2-2-NAcCys conjugate, separated from MM structures by energy barriers, quenched dynamics (simulated annealing) was used to explore the conformational space. No constraints were used during high temperature searches of the conformational space. The starting structure was minimized using MM and then subjected to 50 ps of molecular dynamics at various temperatures between 300 and 400 K. Starting and final temperatures in a dynamic run were set to 0 K, and the heating and cooling times were set to 5 ps; the step size was 0.0005 ps. At various time points during the simulation, several randomly selected structures were also annealed to 0 K and optimized. A final structural optimization of the 4-OHE2-2NAcCys conjugate was performed in a box (25 Å× 25 Å× 25 Å) with equilibrated water molecules to simulate behavior in aqueous solution. The optimized structure was subsequently used as a starting point for further semiempirical QM calculations. To aid in the assignment of the nature of the lowest excited singlet and triplet states, QM calculations (for the lowest energy conformation determined above) were performed. HyperChem’s ZINDO/S (a modified INDO/S) method, parametrized to reproduce the UV/vis spectroscopic transitions using a configuration interaction (CI) treatment with 155 orbitals, was applied. The CI space was truncated by considering only the 10 lowest singly excited configurations. For spectra and orbital eigenvalues, the overlap weighting factors of 1.267 for σ-σ (28) and 0.585 for π-π (29, 30) were used. The excited state in a CI calculation is a mixture of singly excited determinants. Only the leading configurations (with largest contribution to the final state) were taken into account to examine whether a transition is of the n,π* or π,π* type.

3. Results and Discussion 3.1. Luminescence Spectra of 4-OHE1, 4-OHE2, Cys, and N-Acetylated Cys. Curve a in Figure 2 is the 77 K fluorescence spectrum of 4-OHE1 obtained in glycerol/water glass (pH 3.5, λex ) 257 nm). The fluores-

Jankowiak et al.

Figure 2. Luminescence spectra of 4-OHE1 (spectrum a, ×5) and pure NAcCys (spectrum b) in glycerol/water glass obtained at 77 K with an excitation wavelength of 257.0 nm. Both spectra were obtained with CW detection. The long and short arrows label the maximum of the fluorescence band at 300 and 77 K, respectively. The phosphorescence origin band at 77 K of NAcCys is at 372 nm. The bands at 387, 398, and 423 nm are vibronic bands (see text); concentration ∼10-5 M, pH 3.5.

cence band maximum is near 312 nm. An identical spectrum was obtained for 4-OHE2. At room temperature, the major fluorescence band of 4-OHE1 and 4-OHE2 is shifted to ∼320 nm (data not shown), in good agreement with ref 22, in which it was first shown that both E1 and E2 are also weakly fluorescent (maximum at 325-330 nm) upon excitation at 280 nm. As in the case of E2 and E1 (21, 22), fluorescence of 4-OHE2 is also several times stronger than that of 4-OHE1. However, neither 4-OHE1 nor 4-OHE2 is phosphorescent at room temperature or 77 K. Although under our experimental conditions room temperature fluorescence was not observed for Cys and/ or NAcCys, weak fluorescence with a maximum near 300 nm was observed at 77 K. In addition, and in contrast with 4-OHE1 and 4-OHE2, both Cys and NAcCys are phosphorescent at 77 K. An example is spectrum b in Figure 2, which is the luminescence spectrum of NAcCys (nongated detection, pH 3). The weak band near 300 nm corresponds to the fluorescence origin (as revealed by time-resolved luminescence spectroscopy). The band at 372 nm is the origin band of the phosphorescence spectrum, and the bands at 387 and 398 nm are assigned to vibronic transitions at ∼1040 and ∼1760 cm-1. The 423 nm band at ∼3240 cm-1 is probably due to combination and overtone transitions involving modes near 1700 cm-1. At this point, three remarks are pertinent as follows: (i) the phosphorescence lifetime of NAcCys is much longer (1.10 s) than that expected for n,π* triplet states (i.e., 0.01-0.1 s (31-35) and (ii) the phosphorescence lifetime, in the presence of potassium iodide (KI), decreases by a factor of 4 due to the heavy atom effect (31). We note that the phosphorescence lifetime of a n,π* state is only weakly affected by the external heavy atom effect (36). These observations strongly indicate that the lowest triplet state of NAcCys is π,π* in character. Cys exhibited a 77 K fluorescence spectrum that is similar to that of NAcCys with a maximum near 300 nm, but the phosphorescence origin band was slightly redshifted to 375 nm. Again, in this case, no carbonyl stretching mode was observed in the structured phosphorescence spectrum (not shown), strongly indicating that the lowest triplet state in both Cys and NAcCys is

Derived GSH and N-Acetylated Cys Conjugates

Chem. Res. Toxicol., Vol. 16, No. 3, 2003 307

Table 1. Room and Low Temperature Spectral Properties of 4-OHE1-, 4-OHE2-, Cys, NAcCys, and 4-OHE1/4-OHE2-Derived SG and NAcCys Conjugatesa absorbance max ( 1 nm analyte 4-OHE1 or 4-OHE2 Cys NAcCys 4-OHE1-2-NAcCys or 4-OHE2-2-NAcCys 4-OHE1-2-SG or 4-OHE2-2-SG

300 K

fluorescence max ( 1 nm 300 K

77 K

320

240, 280 b b 255, 292

332

312 300 300 328

255, 280, 323

332

328

phosphorescence max ( 1 nm 300 K

τphos (s)

77 K

77 K

375, 401 372, 398, 423 383, 407, 425

c 1.10 ( 0.02 0.22 ( 0.02

383, 407, 425

0.22 ( 0.02

a Luminescence spectra were obtained with an excitation wavelength of 257 nm in glycerol/water matrix (1/1). The matrix contained 10 mM phosphate buffer; pH 3.5. A dash indicates that luminescence was not observed. b No absorption maximum in the 220-300 nm region was observed. However, the extinction coefficient of Cys and NAcCys at 257 nm is about 40 cm-1 M-1 (pH 3.5). c Not measured.

Figure 3. Luminescence spectra of 4-OHE1-2-SG conjugates obtained at 77 K, pH 7, λex ) 257.0 nm. Spectra a and b were obtained in the CW and gated (0.5 s delay time of the observation window) detection mode, respectively. The solid and dashed arrows locate the maximum of the fluorescence (328 nm) and the phosphorescence (∼383 nm) origin bands, respectively. The band at 407 nm corresponds to a vibronic transition involving a 1540 cm-1 vibrational mode(s). The 2580 cm-1 vibronic band near 425 nm is due to combination of overtone and transition modes (see text).

π,π* in character. We are unaware of previously published phosphorescence spectra of Cys and NAcCys. A likely reason is that in most published studies on steroids, including estrogens, room temperature fluorometric assays were based on derivatization with a fluorescent chromophore at room temperature or on procedures involving treatment with strong sulfuric acid at elevated temperatures to form unidentified fluorescent derivatives (37). The absorption, fluorescence, and phosphorescence band maxima and phosphorescence lifetimes of the above metabolites and CEQ-derived conjugates discussed in section 3.2 are summarized in Table 1. 3.2. Luminescence Spectra of 4-OHE1- and 4-OHE2Derived SG and NAcCys Conjugates at 77 K. Luminescence spectra for 4-OHE1-2-SG and 4-OHE2-2-NAcCys are shown in Figures 3 and 4, respectively. Spectra a and b in Figure 3 correspond to luminescence spectra of 4-OHE1-2-SG (in a Gly/H2O glass) obtained at pH 7 with an excitation wavelength of 257 nm. Spectrum a was obtained in the nongated mode and shows a contribution from both fluorescence and phosphorescence. The position of the phosphorescence origin band is at 383 nm (dashed arrow) as revealed in spectrum b, which was obtained in a gated mode with 0.5 s delay time of the observation window. As expected, the 328 nm fluorescence band is absent in spectrum b. It should be mentioned that due to the lack of vibrational structure in the fluorescence

Figure 4. Shaded band (curve a) corresponds to the room temperature fluorescence of 4-OHE2-2-NAcCys multiplied by a factor of 5. Spectra b-d are 77 K luminescence spectra of 4-OHE2-2-NAcCys obtained at pH values of 3, 6, and 11, respectively. All spectra were obtained in glycerol/H2O glass (10 mM phosphate buffer) with λex ) 257.0 nm.

spectrum and corresponding absorption spectrum (not shown), the assignment of the fluorescent state as n,π* (vide infra) cannot be confirmed. The vibronic structure in the phosphorescence spectrum (b) of Figure 3 for 4-OHE1-2-SG is different from that seen in Figure 2 for pure NAcCys. The band at 407 nm in Figure 3 is due to a vibronic transition involving a 1540 cm-1 mode(s). The 425 nm band corresponds to the combination of vibrational modes. Identical phosphorescence spectra were obtained for 4-OHE2-2-SG; no spectral changes were observed for the pH range of 3-8 (spectra not shown). The total luminescence intensity for the SG conjugates increased by a factor of ∼150 upon cooling from 300 to 77 K. The ratio of the integrated phosphorescence to fluorescence at 77 K is about 12. The phosphorescence lifetime (τphos) of 4-OHE1-2-SG and 4-OHE2-2-SG conjugates, based on integrated timeresolved 77 K spectra (obtained at several delay times), is 0.25 ( 0.05 s. This lifetime was confirmed using a Cary Eclipse spectrophotometer, which led to a value of 0.22 ( 0.02 s for τphos. We consider next luminescence spectra of the 4-OHE12-SG and 4-OHE2-2-SG breakdown products that (if present in urine) could serve as biomarkers of exposure to CEQs. Curve a (multiplied by a factor of 5) in Figure 4 is the room temperature fluorescence spectrum of 4-OHE2-2-NAcCys conjugate obtained with an excitation

308

Chem. Res. Toxicol., Vol. 16, No. 3, 2003

Figure 5. Normalized phosphorescence decay curves obtained at 77 K in glycerol/buffer glass for 4-OHE2-2-NAcCys (∼10-4 M) in the presence (curve a, τphos ) 0.14 s) and in the absence (curve b, τphos ) 0.22 s) of 0.75 M KI. Both decay curves were obtained with a Cary eclipse spectrophotometer (delay time, 0.1 ms; gate time, 5 ms; number of flashes, 50; λex ) 240 nm; λobs ) 415 nm; excitation and emission slit width, 10 nm). Fit to curve b is shown as a dashed line. Inset shows normalized (laser excited, λex ) 257 nm) integrated phosphorescence intensity of 4-OHE1-2-NAcCys (10-5 M, in the absence of KI) plotted as a function of the delay time (in s) of the observation window. The circles, crosses, and stars correspond to data obtained at pH 3, 6, and 8, respectively. The estimated phosphorescence lifetime (τphos) of 4-OHE1-2-NAcCys conjugate is 0.25 ( 0.05 s.

wavelength of 257 nm in Gly/H2O (pH 3). Spectra b-d were obtained at 77 K and three different pH values: 3, 6, and 11, respectively. For the sake of brevity, we do not present spectra of 4-OHE1-2-NAcCys since they are similar in structure to those obtained for 4-OHE2-2NAcCys. With reference to the spectra in Figure 3, we note that phosphorescence spectra of NAcCys conjugates are similar to those of the corresponding SG conjugates. However, distinction between these conjugates is readily achieved by CE interfaced with low temperature on-line phosphorescence detection (to be published). Importantly, as in the case of the SG conjugates, the luminescence intensity of the NAcCys conjugates increases by a factor of ∼150 upon cooling from 300 to 77 K. This marked increase is mainly associated with phosphorescence; at 77 K, the integrated phosphorescence intensity is 14 times higher than that of the integrated fluorescence intensity, which is consistent with the lowest singlet and triplet states being n,π* and π,π*, respectively. That is, the S1 f T1 intersystem crossing quantum yield is high, resulting in strong phosphorescence (38). An assignment of the lowest singlet state as n,π* is supported by the observation that fluorescence spectrum of 4-OHE2-2-NAcCys undergoes a 2 nm blue shift in more polar solvents (spectra not shown). It is well-established that a 1(π,π*) state shifts to the red with increasing solvent polarity (38-40). Figure 5 shows two normalized phosphorescence decay curves obtained for 4-OHE2-2-NAcCys (∼10-4 M) in the presence (curve a) and absence (curve b; solid line) of external heavy atoms (0.75 M KI). Exponential fits to the experimental data shown in curves a and b of Figure 5 led to values for τphos of 0.22 and 0.14 s, respectively. For simplicity, only the fit to curve b (see a dashed line) is shown. As in the case of pure NAcCys (vide supra), the shortening of τphos (by a factor of 1.6) indicates that the lowest triplet of 4-OHE2-2-NAcCys is also of π,π* character. The inset in Figure 5 shows the normalized

Jankowiak et al.

Figure 6. Diagram showing the calculated transition type and energy level ordering of the two lowest singlet and triplet states for the 4-OHE2-2-NAcCys conjugate. Energies of the S1, S2, T1, and T2 states are given in Table 2. Sn labels higher excited states. The right upper part of the figure shows an optimized 0 K ground state structure of 4-OHE2-2-NAcCys used in calculations (hydrogens and double bonds are not shown for clarity; molecular rendering based on sticks and dots gives an idea of the shape and volume occupied by the molecule). Intersystem crossing rate is denoted as kST (S1 f T1).

phosphorescence intensity obtained for 4-OHE1-2-NAcCys from time-resolved spectra measured at 0, 0.5, 1.0, 1.5, and 2.0 s delay times of the observation window. The circles, crosses, and stars correspond to data obtained at pH 3, 6, and 8, respectively. As for SG conjugates, the estimated τphos of 4-OHE1-2-NAcCys conjugates (at pH 3, 6, and 8) is also 0.22 ( 0.05 s. We conclude, therefore, that the triplet lifetime of CEQ-derived conjugates is not dependent on pH below 8. In the pH range studied (311), the phosphorescence intensity increases up to pH 8 (by a factor of ∼1.2) and decreases slightly at higher pH. Irreversible spectral changes were observed only at pH g 8; the phosphorescence spectrum at pH 11 is shown in Figure 4 (spectrum d). The relatively large shift in the position of the origin band suggests that spectrum d most likely corresponds to an oxidized product of 4-OHE2-2NAcCys. (No measurable fluorescence was observed at pH 11.) 3.3. Computer Modeling Studies of 4-OHE2-2NAcCys. To the best of our knowledge, no calculations have been performed for CEQ-derived conjugates. To guide the interpretation of the strong phosphorescence observed for NAcCys conjugates at 77 K (Figure 5), the nature (and ordering) of the lowest S1 and T1 energy levels, which play an important role in determining the S1 f T1 intersystem rate (41), was investigated. The calculated ground state structure of 4-OHE2-2NAcCys is shown in the inset of Figure 6; hydrogens and double bonds are not shown for clarity. This structure was optimized first in vacuo and subsequently in water at 0 K. The lowest calculated singlet state (S1) lies at 31 970 cm-1 (312.8 nm). The experimentally observed fluorescence band maximum at 328 nm is in good agreement with this value since the calculations do not include a Stokes shift. The second singlet state (S2) is calculated at 33 515 cm-1 (298.4 nm), ∼1500 cm-1 above

Derived GSH and N-Acetylated Cys Conjugates

Chem. Res. Toxicol., Vol. 16, No. 3, 2003 309

Table 2. Calculated Singlet and Triplet Energy Levels and Transition Type Assignments for the Low Temperature Optimized Structure of 4-OHE2-2-NAcCys electronic state

λcalcd (nm)

energy level type

oscillator strength

absorption maxa (nm)

S1 S2 S3 T1 T2

312.8 298.4 276.0 369.0 318.4

1(n,π*)

1.3 × 10-3 2 × 10-3 2 × 10-3 sfd sf

∼292b

1(π,π*)

c 3(π,π*) 3(n,π*)

255 sf sf

a Absorption spectra measured in Gly/buffer glass at 77 K. The bands at 292 and 255 nm were also observed in phosphorescence excitation spectrum (see Figure 7). b The lowest absorption band at 292 nm (T ) 77 K) is most likely a mixture of S1 and S2 states (see Figure 7). c Highly mixed character. d sf ) spin forbidden.

the S1 transition. The above energy levels were calculated with the π-π atomic overlap weighting factor of 0.585 commonly used for singlet states of organic molecules (30). However, a value of 0.64 (which slightly changes the relative contribution of σ vs π bonding) has been recommended for triplet states (42). Adjusting the π-π overlap to 0.64 resulted in the lowest triplet state (T1) lying at 369 nm (27 100 cm-1), in good agreement with the observed value of the phosphorescence origin band at ∼383 nm. The small discrepancy between the measured and the calculated energy levels may be due to the neglect of solvation effects, the ground state geometry used, as well as inadequate accounting of electron correlation effects in the QM calculations. The second triplet state (T2) lies at 31 348 cm-1 (319 nm), ∼ 4250 cm-1 higher in energy than T1. The assignment of the lowest S1 state as n,π* is supported by QM calculations, which revealed that the two major configurations in CI calculations have nonbonding orbitals localized on the carbonyl oxygen atom (indicated by a short dotted arrow in Figure 6) and that the lowest unoccupied π* orbital delocalized over the entire carbonyl and the aromatic ring of the 4-OHE2-2NAcCys. Recall that the experimental phosphorescence spectra and the value of τphos for the 4-OHE2-2-NAcCys discussed above strongly indicate that T1 is π,π* in character. This is also confirmed by QM calculations; in this case, however, both π and π* orbitals are delocalized over the aromatic ring 4-OHE2-2-NAcCys. Thus, we conclude that the T1 state is of π,π* type in agreement with experimental data. The S2 and T2 states are predominantly π,π* and n,π* in character; the higher excited singlet and triplet states are highly mixed in character. The calculated electronic energies of the S1, S2, S3, T1, and T2 states for the 4-OHE22-NAcCys conjugate are given in Table 2; their energy level ordering is depicted in Figure 6. For the energy order of the S and T states, as calculated and displayed in Figure 6, the intersystem crossing rate should be significant since the lowest singlet and triplet states have n,π* and π,π* character. This leads to increased spinorbit coupling between singlet and triplet states, and as a result intense phosphorescence, in agreement with experimental data. Finally, a comment on the low energy absorption bands of the 4-OHE2-2-NAcCys is in order. The extinction coefficient of the lowest energy 292 nm absorption band is 3.8 × 103 M-1 cm-1. Figure 7 shows the 77 K phosphorescence excitation spectrum obtained for 4-OHE2-2-NAcCys in Gly/buffer glass that in agreement

Figure 7. Phosphorescence excitation spectrum of 4-OHE2-2NAcCys obtained in glycerol/buffer glass at T ) 77 K. The thin vertical lines correspond to the calculated stick absorption spectrum. The thick bars correspond to the calculated stick spectrum shifted blue by 13 nm (see text for details).

Figure 8. Calibration curve showing the integrated phosphorescence intensity measured at 77 K in glycerol/H2O glass (10 mM phosphate buffer, pH 3.3, λex ) 257.0 nm) plotted vs concentration (in mol/L) of 4-OHE2-2-NAcCys.

with the absorption spectrum (see Table 1), has two bands with the maxima near 292 and 255 nm. The thin vertical lines shown in Figure 7 correspond to the three lowest energy singlet states, calculated for 4-OHE2-2NAcCys. Their transition energies and oscillator strengths are listed in Table 2. The stick spectrum has been normalized to fit the intensity of the phosphorescence excitation spectrum with the length of the bars being proportional to the calculated oscillator strength. The phosphorescence excitation spectrum in Figure 7 appears to be well-reproduced by our model calculations, when the normalized stick spectrum is blue-shifted by ∼13 nm, as shown by the thick vertical lines. This shifting is justified since the calculations did not include a solvent shift; our approximation, however, should describe 4-OHE2-2-NAcCys with sufficient accuracy in terms of the relative oscillator strength and the energy differences between various electronic transitions. Thus, the finding of particular interest is that the lowest energy band near 292 nm is contributed to by two closely lying singlet states S1(n,π*) and S2(π,π*), with the n,π* state being lowest in energy. This finding is in agreement with the observed blue shift of the fluorescence band in more polar solvents briefly discussed in section 3.2. 3.4. Detection Limit. A linear calibration curve (correlation coefficient r2 ) 0.997) obtained for 4-OHE12-NAcCys in Gly/H2O glass (pH 3.5) is shown in Figure 8, where integrated phosphorescence intensity is plotted vs concentration (10-5-10-8) in the unit of molarity M. The concentration limit of detection (LOD) for 4-OHE12-NAcCys based on phosphorescence measurements is ∼10-9 M (defined by a signal-to-noise ratio of 3). Because

310

Chem. Res. Toxicol., Vol. 16, No. 3, 2003

the laser-excited volume in our experiments was approximately 1 µL, the LOD for 4-OHE1-2-NAcCys is in the low femtomole range. Similar LOD values were obtained for 4-OHE2-2-NAcCys, 4-OHE1-2-SG, and 4OHE2-2-SG conjugates (data not shown). Preliminary experiments have shown that femtomole detection limits can also be obtained using CE interfaced with low temperature phosphorescence detection. We note that the above LOD for CEQ-derived conjugates is about 2 orders of magnitude lower than the LOD obtained by HPLC with multichannel electrochemical detection (18).

4. Concluding Remarks It has been demonstrated that the luminescence intensity of CEQ-derived conjugates (4-OHE1-2-SG, 4-OHE22-SG, 4-OHE1-2-NAcCys, and 4-OHE2-2-NAcCys) increases by a factor of ∼150 upon cooling from 300 to 77 K. The bulk of the signal at low temperatures, for both NAcCys and SG conjugates, originates from phosphorescence. A relatively long phosphorescence lifetime of 0.22 s and the observed vibronic bands of CEQ-derived NAcCys and SG conjugates strongly indicate that the lowest triplet state is π,π*, in agreement with semiempirical QM calculations. In contrast, the lowest singlet state of 4-OHE2-2-NAcCys was shown to possess n,π* character, as indicated experimentally by a blue shift of the fluorescence band in more polar solvents. The latter observation was confirmed by QM calculations, which showed that the lowest singlet state of 4-OHE2-2-NAcCys is n,π*, with two major nonbonding orbitals localized on the carbonyl oxygen atom labeled by an arrow in Figure 6 and the lowest unoccupied π* orbital delocalized over the entire carbonyl and the aromatic ring of 4-OHE2-2NAcCys. Such energy level types and energy level ordering lead to a large intersystem crossing rate and, as a result, an intense phosphorescence in agreement with experimental data. A library of luminescence spectra for the 4-OHE1-, 4-OHE2-, Cys-, N-acetylated Cys-, and 4-OHE1/4-OHE2-derived SG and SG decomposition products (i.e., 4-OHE1-2-NAcCys and 4-OHE2-2-NAcCys) has been assembled. An excellent concentration detection limit of ∼10-9 M has been demonstrated. Therefore, we conclude that future spectroscopic interrogation of CEQderived conjugates, present in breast/prostate tissue extracts and/or in human urine, can be performed by CE interfaced with low temperature on-line phosphorescence detection. The combination of the separation power of CE (and/or HPLC) and spectral selectivity of low temperature fluorescence spectroscopy for identification of closely related biological analytes at low levels has been demonstrated in our laboratory (44, 45).

Acknowledgment. Ames Laboratory is operated for the U. S. Department of Energy by Iowa State University under Contract No. W-7405-Eng-82. This work was supported by the grant from the NIH (Program Project Grant 2PO1 CA49210-12). We also acknowledge partial support from the Office of Health and Environmental Research, Office of Energy Research. We thank I. Adamovic, J. M. Hayes, and W. Hurtubise for useful discussions.

References (1) Cavalieri, E. L., Rogan, E. G., and Chakravarti, D. (2002) Initiation of cancer and other diseases by catechol ortho-quinones: a unifying mechanism. Cell. Mol. Life Sci. 59, 665-681.

Jankowiak et al. (2) Chakravarti, D., Mailander, P. C., Li, K.-M., Higginbotham, S., Zhang, H. L., Gross, M., Meza, J. L., Cavalieri, E. L., and Rogan, E. G. (2001) Evidence that a burst of DNA depurination in SENCAR mouse skin induces error-prone repair and froms mutations in the H-ras gene. Oncogene 20, 7945-7953. (3) Dwivedy, I., Devanesan, P. D., Cremonosi, P., Rogan, E. G., and Cavalieri, E. L. (1992) Synthesis and characterization of estrogen 2,3- and 3,4-quinons. Comparison of the DNA adducts formed by the quinones versus horseradish peroxidase-activated catechol estrogens. Chem. Res. Toxicol. 5, 828-833. (4) International Agency for Research on Cancer Group (1979) IARC Monographs on Evaluation of Carcinogenic Risk of Chemicals to Humans, Sex Hormones (II), Vol. 21, pp 173-221, International Agency for Research on Cancer, Lyon, France. (5) Liehr, J. G. (2000) Is estradiol a genotoxic mutagenic carcinogen? Endocr. Rev. 21, 4054-4058. (6) Kalyanaraman, B., Sealy, R. C., and Sivarajah, K. (1984) An electron spin resonance study of o-semiquinones formed during the enzymatic and auto-oxidation of catechol estrogens. J. Biol. Chem. 259, 14018-14022. (7) Kalyanaraman, B., Felix, C. C., and Sealy, R. C. (1985) Semiquinone anion radicals of catechol(amine)s, catechol estrogens, and their metal ion complexes. Environ. Health Perspect. 64, 185198. (8) Liehr, J. G., Ulubelen, A. A., and Strobel, H. W. (1986) Cytochrome P-450-mediated redox cycling of estrogens. J. Biol. Chem. 261, 16865-16870. (9) Liehr, J. G., and Roy, D. (1990) Free radical generation by redox cycling of estrogens. Free Radical Biol. Med. 8, 415-423. (10) Nutter, L. M., Wu, Y.-Y., Ngo, E. O., Sierra, E. E., Gutierrez, P. L., and Abul-Hajj, Y. J. (1994) An o-quinone form of estrogen produces free radicals in human breast cancer cells: Correlation with DNA damage. Chem. Res. Toxicol. 7, 23-28. (11) Li, K.-M., Devanesan, P. D., Rogan, E. G., and Cavalieri, E. L. (1998) Formation of the depurinating 4-hydroxyestradiol (4OHE2)-1-N7Gua and 4-OHE2-1-N3Ade adducts by reaction of E23,4-quinone with DNA. Proc. Am. Assoc. Cancer Res. 39, 636643. (12) Stack, D., Byun, J., Gross, M. L., Rogan, E. G., and Cavalieri, E. (1996) Molecular characteristics of catechol estrogen quinones in reactions with deoxyribonucleosides. Chem. Res. Toxicol. 9, 851859. (13) Jankowiak, R., Zamzow, D., Stack, D. E., Cavalieri, E. L., and Small, G. J. (1998) Spectral characterization of fluorescently labeled catechol estrogen-3,4-quinone derived-N7Gua adducts and their identification in rat mammary gland tissue. Chem. Res. Toxicol. 11, 1339-1345. (14) Cavalieri, E. L., Stack, D. E., Devanesan, P. D., Todorovic, R., Dwivedy, I., Higginbotham, S., Johanson, S., Patil, K., and Rogan, E. (1997) Molecular origin of cancer: catechol estrogen-3,4quinones as endogeneous tumor initators. Proc. Natl. Acad. Sci. U.S.A. 94, 10937-10942. (15) Devanesan, P., Todorovic, R., Zhao, J., Gross, M. L., Rogan, E. G., and Cavalieri, E. L. (2001) Catechol estrogen conjugates and DNA adducts in the kidney of male Syrian golden hamsters treated with 4-hydroxyestradiol: potential biomarkers for estrogeninitiated cancer. Carcinogenesis 22, 489-497. (16) Todorovic, R., Devanesan, P., Higginbotham, S., Zhao, J., Gross, M. L., Rogan, E. G., and Cavalieri, E. L. (2001) Analysis of potential biomarkers of estrogen-initiated cancer in the urine of Syrian golden hamsters treated with 4-hydroxyestradiol. Carcinogenesis 22, 905-911. (17) Cavalieri, E. L., Kumar, S., Todorovic, R., Higginbotham, S., Badawi, A. F., and Rogan, E. G. (2002) Imbalance of estrogen homeostasis in kidney and liver of hamsters treated with estradiol: implications for estrogen-induced initiation of renal tumors. Chem. Res. Toxicol. 14, 1041-1050. (18) Devanesan, P., Santen, R. J., Bocchinfuso, W. P., Korach, K. S., Rogan, E., and Cavalieri, E. (2001) Catechol estrogen metabolites and conjugates in mammary tumors with hyperplastic tissue from estogen receptor-R knock-out (ERKO)/Wnt-1 mice: implications for initiation of mammary tumors. Carcinogenesis 22, 1573-1576. (19) Cao, K., Devanesan, P. D., Ramanathan, R., Gross, M. L., Rogan, E., and Cavalieri, E. L. (1998) Covalent binding of catechol estrogens to glutathione catalyzed by horseradish peroxidase. Chem. Res. Toxicol. 11, 917-924. (20) Nakagomi, M., and Suzuki, E. (2000) Quantitation of catechol estrogens and their N-acetylcysteine conjugates in urine of rats and hamsters. Chem. Res. Toxicol. 13, 1208-1213. (21) Udenfriend, S. (1962) Fluorescence Assays in Biology and Medicine, Chapter 10, pp 349-365, Academic Press, New York.

Derived GSH and N-Acetylated Cys Conjugates (22) Roberts, H. R., and Siino, M. R. (1963) Stability Assays of Pharmaceutical Preparations by Quantitative Paper Chromatography II. J. Pharm. Sci. 52, 370. (23) Velury, S., and Howell, S. (1988) Measurements of plasma thiols after derivatization with monobromobimane. J. Chromatogr. 424, 141-146. (24) Houze´, P., Dussaucy, M., Courties, Y., and Bousquet, B. (1999) Evaluation of the bio-red high-performance liquid chromatographic method for plasma total homocysteine assay. Ann. Biol Clin. 57, 611-616. (25) Markushin, Y., Zhong, W., Small, G. J., Yeung, E. S., Cavalieri, E. L., and Jankowiak, R. (2003) Anal. Chem. (to be submitted). (26) Burkert, U., and Allinger, N. L. (1982) Molecular Mechanics, ACS Monograph, Vol. 177, American Chemical Society, Washington, DC. (27) Allinger, N. L. (1977) Conformational analysis 130. MM2: a hydrocarbon force field utilizing V1 and V2 torsional terms. J. Am. Chem. Soc. 99, 8127-8134. (28) Ridley, J. E., and Zerner, M. C. (1976) Triplet states via intermediate neglect of differential overlap: benzene, pyridine, and the diazines. Theor. Chim. Acta 42, 223-236. (29) Del Bene, J., and Jaffe´, H. H. (1968) Use of the CNDO method in spectroscopy. I. Benzene, pyridine, and the diazines. J. Chem. Phys. 48, 1807-1813. (30) Del Bene, J., and Jaffe´, H. H. (1968) Use of the CNDO method in spectroscopy. II. Five-membered rings. J. Chem. Phys. 48, 40504055. (31) Becker, R. S. (1969) Theory and Interpretation of Fluorescence and Phosphorescence, Chapter 12, pp 155-189, Wiley-Interscience, New York. (32) Shimada, J., and Goodman, L. (1965) Polarization of aromatic carbonyl spectra. J. Chem. Phys. 43, 2027-2041. (33) Sidman, J. (1958) Spin-orbit coupling in the 3A2-1A1 transition of formaldehyde. J. Chem. Phys. 29, 644-652.

Chem. Res. Toxicol., Vol. 16, No. 3, 2003 311 (34) Vanquickenborne, L., and McGlynn, S. (1966) Spin-orbit coupling in aza-aromatics and carbonyls. J. Chem. Phys. 45, 4755-4756. (35) Lim, E. C. (1986) Proximity Effect in Molecular Photophysics: Dynamical Consequences of Pseudo-Jahn-Teller Interaction. J. Phys. Chem. 90, 6770-6777. (36) Kearns, D., and Case, W. (1966) Investigation of singlet-triplet transitions by the phosphorescence excitation model. III. Aromatic ketones and aldehydes. J. Am. Chem. Soc. 88, 5087-5097. (37) Udenfriend, S. (1969) Fluorescence Assay in Biology and Medicine, Chapter 14, pp 655-472, Academic Press, New York. (38) Klesinger, M., and Michl, J. (1995) Excited States and Photochemistry of Organic Molecules, VCM Publishers, New York. (39) Reichardt, Ch. (1988) Solvents Effects in Organic Chemistry (2nd ed.) VCM, Weinheim. (40) McConnell, H. (1952) Effect of polar solvents on the absorption frequency of n f π* electronic transitions. J. Chem. Phys. 20, 700-704. (41) Kasha, M. (1960) Radiat. Res. 2, 243-248. (42) Hyperchem Computational Chemistry (2002) Publication HC-7000-04-00, Part 2, Theory and Methods, p 307, Hypercube, Inc. (43) El Sayed, M. A. (1963) Spin-orbit coupling and the radiationless processes in nitrogen heterocyclic. J. Chem. Phys. 38, 2834-2838. (44) Roberts, K. P., Jankowiak, R., and Small, G. J. (2001) Highperformance liquid chromatography interfaced with fluorescence line-narrowing spectroscopy for on-line analysis. Anal. Chem. 73, 951-956. (45) Roberts, K. P., Lin, C.-H., Singhal, M., Casale, G. P., Small, G. J., and Jankowiak, R. (2000) On-line identification of depurinating DNA adducts in human urine by capillary electrophoresisfluorescence line narrowing spectroscopy. Electrophoresis 21, 799-806.

TX020088P