Stereoselective Conjugation of Prostaglandin A2 and Prostaglandin

Stereoselective Conjugation of Prostaglandin A2 and. Prostaglandin J2 with Glutathione, Catalyzed by the. Human Glutathione S-Transferases A1-1, A2-2,...
1 downloads 0 Views 277KB Size
Download PDF
310

Chem. Res. Toxicol. 1997, 10, 310-317

Stereoselective Conjugation of Prostaglandin A2 and Prostaglandin J2 with Glutathione, Catalyzed by the Human Glutathione S-Transferases A1-1, A2-2, M1a-1a, and P1-1 Jan J. P. Bogaards,*,† Joke C. Venekamp,‡ and Peter J. van Bladeren† Toxicology Division and Division Analytical Sciences, TNO Nutrition and Food Research Institute, P.O. Box 360, 3700 AJ Zeist, The Netherlands Received October 16, 1996X

Prostaglandins containing an R,β-unsaturated keto group, such as prostaglandin A2 (PGA2) and prostaglandin J2 (PGJ2), inhibit cell proliferation. These cyclopentenone prostaglandins may be conjugated with GSH chemically or enzymatically via glutathione S-transferases, and this has been suggested to result in inhibition of the antiproliferative mode of action. In the present study, the role of the major human GSTs in the conjugation of PGA2 and PGJ2 with GSH was investigated with purified enzymes, i.e., the Alpha-class enzymes GST A1-1 and GST A2-2, the Mu-class enzyme GST M1a-1a, and the Pi-class enzyme GST P1-1. The GSH conjugates were separated from the parent compound by HPLC and identified by fast atom bombardment mass spectrometry and 1H-NMR. Two GSH conjugates were found for both PGA2 and PGJ2, the R- and S-GSH conjugates of both prostaglandins. Incubation experiments with PGA2 and PGJ2 (70-600 µM) clearly showed the role of individual GSTs in the conjugation of PGA2 and PGJ2. Compared to the chemical reaction, enzyme activities towards PGA2 were up to 5.4 times as high (GSTA1-1) at the lowest concentration (70 µM), while at the highest concentration (600 µM) enzyme activities were up to 3.0 times as high (GST P1-1). For PGJ2, enzyme activities were up to 4.3 (GSTM1a-1a, 70 µM) and up to 3.1 (GSTM1a-1a, 600 µM) times as high. As expected, similar amounts of the R- and S-conjugates of both prostaglandins were found in the chemical reaction. Striking stereoselectivities in conjugating activities were observed for GST A1-1 and GST P1-1. GST A1-1 favors the formation of the R-GSH conjugates of both prostaglandins. GST P1-1 showed a clear selectivity with regard to the formation of the S-GSH conjugate of PGA2. However, this selectivity was not found for the formation of the S-GSH conjugate of PGJ2. GSTM1a-1a showed no stereoselectivity with regard to the GSH conjugation of both PGA2 and PGJ2. GSTA2-2 only showed some minor formation of the R-GSH conjugate of PGJ2. The possible implications of the observed stereoselectivity on the effects of PGA2 and PGJ2 are discussed.

Introduction Prostaglandins containing an R,β-unsaturated keto group, such as prostaglandin A2 (PGA2), prostaglandin J2 (PGJ2), and ∆12-PGJ2, inhibit cell proliferation (1-6). PGA2 and PGJ2 are dehydration products of PGE2 and PGD2, respectively, which are presumably formed in a non-enzymatic process (2). It has also been reported that ∆12-PGJ2 is formed non-enzymatically from PGJ2 in the presence of serum albumin (2). The R,β-unsaturated keto group of the cyclopentenone prostaglandins is essential for their cytotoxic activity because it allows covalent binding to macromolecules such as proteins and DNA via a Michael addition (2, 3, 7-9). In addition to macromolecules, R,β-unsaturated ketones may also bind to the tripeptide glutathione (GSH). Conjugation of cyclopentenone prostaglandins with GSH may occur chemically or enzymatically via glutathione S-transferases (GSTs) (3-5, 10), perhaps resulting in inhibition of the antiproliferative mode of action. However, the role of GSH and * Corresponding author. Telephone: +31 30 694 4418. FAX: +31 30 696 0264. E-mail: [email protected]. † Toxicology Division. ‡ Division Analytical Sciences. 1 Abbreviations: PGA , prostaglandin A ; PGJ , prostaglandin J ; 2 2 2 2 GST, glutathione S-transferase; FAB-MS, fast atom bombardment mass spectrometry; COSY, 2D proton-proton correlation spectra; GS-X pump, glutathione S-conjugate export pump. X Abstract published in Advance ACS Abstracts, February 15, 1997.

S0893-228x(96)00177-4 CCC: $14.00

GST and their relation to cell proliferation is rather unclear to date. For instance, Ohno and co-workers (11) have found that the growth inhibitory activity of PGA2 in L-1210 murine leukemia cells was not influenced by the GSH status, while Atsmon et al. (3) have concluded that intracellular GSH may modulate the antiproliferative activity of ∆12-PGJ2 in Chinese hamster ovary and hepatoma tissue cells. Furthermore, Sato et al. (12) have suggested the involvement of GST in K562 cell proliferation, while GST is not involved in this process according to Ohno and Hirata (5). As a first step in clarifying this controversy, the role of the major human GSTs in the conjugation of PGA2 and PGJ2 with GSH was investigated with purified enzymes, i.e., the Alpha-class enzymes GST A1-1 and GST A2-2, the Mu-class enzyme GST M1a-1a, and the Pi-class enzyme GST P1-1. GSH conjugates were separated from the parent compounds by HPLC analysis. Metabolites were identified by fast atom bombardment mass spectrometry (FAB-MS) and 1H-NMR.

Materials and Methods Chemicals. PGA2 and PGJ2 were obtained from Cascade Biochem Ltd. (Reading, England), and [35S]glutathione was from DuPont (’s-Hertogenbosch, The Netherlands). All other chemicals were of the highest available quality. Origin of Human Tissue. Human liver and placenta were from healthy subjects. The liver tissue was obtained at autopsy

© 1997 American Chemical Society

Stereoselective Conjugation of PGA2 and PGJ2 from kidney donors. Blood circulation was maintained until the moment of removal of the organ. After removal, tissues were stored immediately on ice and frozen at -80 °C within 8 h of clinical death. Enzyme Purification. GST A1-1, GST A2-2, and GST M1a1a were purified from liver and GST P1-1 from placenta using affinity chromatography as described by A° lin et al. (13). The separation of the GSTs was achieved by chromatofocusing on PBE 94 (Pharmacia, Uppsala, Sweden). PBE 94 was equilibrated with 0.025 M ethanolamine-HCl, pH 9.5 (GST A1-1, GST A2-2, GST M1a-1a), or 0.025 M Tris-HCl, pH 7.5 (GST P1-1). Elution was performed using polybuffer 96 (dilution factor 1:10, pH 6.0) and polybuffer 74 (dilution factor 1:8, pH 4.0), and the purity was judged by reversed-phase HPLC analysis (14) using a Vydac 218TP54 (250 × 4.6 mm) column (The Separations Group, Hesperia, CA) and a flow of 1 mL/min. The protein content of the purified GSTs was determined with the Lowry method, using BSA as standard. Specific activities of GST A11, GST A2-2, GST M1a-1a, and GST P1-1 toward 1-chloro-2,4dinitrobenzene (15) were 46.4, 9.8, 67.2, and 53.0 µmol/min/mg of protein, respectively. The purity of the GSTs was >98%. Incubation Experiments. PGA2 and PGJ2 (70-600 µM) were incubated with human GST A1-1, GST A2-2, GST M1a1a, and GST P1-1 (65 µg/mL, corresponding to 1.3 µM) in incubation mixtures containing 0.05 M potassium phosphate buffer (pH 7.0), BSA (1 mg/mL), and [35S]glutathione (1 mM, specific activity 10 000 dpm/nmol) in a final volume of 100 µL. Incubations were performed at 37 °C. The reactions were terminated after 5 min by the addition of 2 µL of 3 M HCl. Subsequently, 75 µL was subjected to HPLC analysis. Blanks consisted of incubation mixtures without GST. HPLC analysis was performed using a Zorbax reversed-phase C18 column (250 × 4.6 mm) with 50 mM aqueous ammonium acetate (pH 3.4)/ acetonitrile (75:25, v/v) run isocratically at 1 mL/min (3) for 30 min, followed by a linear gradient from 25 to 50% acetonitrile in 30 min. Radioactivity was detected using on-line radiochemical detection (Canberra Packard A500). Isolation and Identification of Metabolites. 5 mg of PGA2 in 500 µL of acetone and 2.5 mg of PGJ2 in 250 µL of methyl acetate were added to 5 and 2.5 mL 5 mM GSH solutions in water, adjusted to pH 7.0, respectively. After 10 min at 37 °C, the reaction was terminated by the addition of 300 or 150 µL of formic acid. Subsequently, the solutions were freeze-dried, and the residues were dissolved in 1 mL of water adjusted to pH 2.2 with HCl and subjected to HPLC analysis using the conditions described above. Peak detection was performed at 200 nm. Peak fractions were collected and freeze-dried. The isolated peaks were analyzed by FAB-MS and 1H-NMR analysis. FAB-MS and 1H-NMR Analysis of Prostaglandin Conjugates. FAB-MS analysis was performed on a FinniganMAT 900 mass spectrometer using glycerol as a matrix. The proton NMR spectra were recorded on a 400 MHz Varian Unity 400 spectrometer, using standard pulse programs. The spectra were recorded at a temperature of 30 °C. The isolated prostaglandin HPLC fractions were dissolved in D2O and acidified to pH 2-3 with DCl. PGA2/PGJ2 and GSH were dissolved in acetone-d6 and D2O, respectively. The CH3 signal of the prostaglandins was used as internal reference for the chemical shift (0.88 ppm). For GSH, the sodium salt of trimethylsilylpropionic acid (0 ppm) was used as internal reference. The 1D proton spectra were recorded with 1024 transients using 64K digital points over a 8000 Hz spectral width and using a 8.2 µs (circa 45°) rf pulse. With a pulse repetition time of 6.1 s the total acquisition time of a typical proton spectrum was 105 min. The 2D proton-proton correlation spectra (COSY) were obtained from the same solutions. Typical parameters were 2200 Hz spectral width with 1024 data points and 256 increments, Fourier transformed to a 1K × 1K data matrix. With 256 transients and a pulse repetition time of 1.45 s the total acquisition time was 26-28 h. The COSY spectra were symmetrized over the diagonal.

Chem. Res. Toxicol., Vol. 10, No. 3, 1997 311

Results Identification of Metabolites. The UV-200 profiles of the incubation mixtures of PGA2 and PGJ2, when incubated with glutathione at pH 7.0 (Figure 1), showed a number of signals which were identified as GSH conjugates by incubation experiments with [35S]glutathione. Chromatograms of PGA2 showed a maximum of four radioactive peaks, those of PGJ2 a maximum of 5. Incubations with PGA2 and PGJ2 showed two major signals in each chromatogram (i.e., PGA2-M3 and -M4; PGJ2-M2 and -M3) at short incubation times (less than 10 min). If the incubation experiments were continued for longer periods, the other signals were increased. FAB-MS analysis (Figure 2) of the isolated peak fractions showed mass spectra with molecular masses of 641, the molecular weight of the PGA2- and PGJ2-GSH conjugates, with exception of the signal in the PGJ2 chromatogram at 48 min (peak number 5), which showed a molecular mass of 623. This latter compound is most likely formed due to the loss of an H2O molecule at C15-C16 (Figure 3). The UV spectrum showed an absorption maximum at 305 nm (data not shown), which was not observed in the other peaks, thus confirming the formation of a conjugated π-system. The 305 nm absorption maximum, however, suggests extended conjugation. Most likely, rearrangement of the double bonds toward the C11-carboxyl group takes place to give a conjugated dienone. The results of NMR/COSY analysis of PGA2 and PGJ2 and their two GSH conjugates (PGA2-M3/M4 and PGJ2M2/M3) are presented in the Tables 1 and 2, respectively. Both the 1D proton and the 2D-COSY spectra were measured, with exception of PGJ2-M2, for which only the COSY spectrum is presented. The spectra of the minor peak fractions could not be obtained or could not be used for structure elucidation, because of the limited amount of material or because the fractions were a mixture of compounds which caused ambiguities in the spectra. The minor peak fractions are presumably the result of chemical breakdown of the two initially formed GSH conjugates of PGA2 and PGJ2. The addition of glutathione to the double bond is obvious: the resonances of the unsaturated protons H10 and H11 for PGA2 and of H9 and H10 for PGJ2 (Figure 4) in the region from 6 to 8 ppm disappeared and were accompanied by the appearance of three new resonances between 2.3 and 3.8 ppm. Although these resonances partly overlap with resonances from other protons of the conjugates, the assignment of these protons is evident from the correlations found in the COSY spectra. The COSY spectra showed correlations between these resonances corresponding to proton-proton couplings over two or three bonds as observed in the 1D proton spectra. The enantiomeric structures of the GSH conjugates were elucidated by using the coupling constants and observed correlations between H11/H12 and H11/H10a,b for the PGA2-GSH conjugates and between H9/H8 and H9/ H10a,b for the PGJ2-GSH conjugates. The other proton resonances of GSH and of PGA2 and PGJ2 do not change much on addition of GSH to the double bond, with exception of the proton shift of H8 and H12. PGA2-M3/PGA2-M4: Two almost similar coupling constants with an average value of 10.5 Hz and an additional coupling constant of 7.8 Hz were measured on proton H11 in PGA2-M3. In PGA2-M4 these coupling constants were 6.2 (average) and 1.7 Hz. In both cases, the measured coupling constants on H11 from H10a,b

312 Chem. Res. Toxicol., Vol. 10, No. 3, 1997

Bogaards et al.

Figure 1. Reversed-phase HPLC UV-200 profiles of incubations of PGA2 (a) and PGJ2 (b) with glutathione. PGA2 and PGJ2 (1 mM) were incubated with glutathione (10 mM) at 37 °C and pH 7.5. Elution was performed as described in Materials and Methods. Unchanged PGA2 and PGJ2 both eluted around 60 min.

and H12 were larger for PGA2-M3 than for PGA2-M4. One of these coupling constants is due to coupling with H12. This indicates a large dihedral angle (trans position) between H11 and H12 for PGA2-M3 (R-GSH conjugate of PGA2) and a small dihedral angle (cis position) for PGA2-M4 (S-GSH conjugate of PGA2). PGJ2-M3: The COSY spectrum of PGJ2-M3 was rather similar to the COSY spectrum of PGA2-M4 as relatively small coupling constants were found on H9. The multiplet found on H9 was caused by two almost similar coupling constants of 6.7 Hz from H8 and H10b. In addition, a very weak correlation, corresponding to a very small coupling constant in the 1D spectrum, was observed between H9 and H10a. The proposed structure for PGJ2-M3 is a cis position between H9 and H8 (SGSH conjugate of PGJ2). Estimation of Enzymatic Activities. Representative chromatograms obtained from incubation experi-

ments with PGA2/PGJ2 and purified glutathione S-transferases are shown in Figure 5. The enzymatic activities of GST A1-1, GST A2-2, GST M1a-1a, and GST P1-1 (enzyme concentration: 1.3 µM) with regard to the formation of the two metabolites of PGA2, representing about 90% of the radioactive peaks, are presented in Figure 6. GST M1a-1a resulted in the formation of both metabolites, while GST A1-1 and GST P1-1 were clearly selective with regard to the formation of PGA2-M3 and PGA2-M4, respectively. GST A2-2 seemed to be inactive with regard to the formation of both metabolites. In the chemical reaction (incubations without GST), similar amounts of PGA2-M3 and PGA2-M4 were formed. In general, the enzymatic formation of the conjugates compared to the chemical reaction, was more dominant at the lower prostaglandin concentrations. At the lowest concentration (70 µM), GST A1-1 (PGA2-M3), GST M1a1a (PGA2-M3 and M4), and GST P1-1 (PGA2-M4) showed

Stereoselective Conjugation of PGA2 and PGJ2

Chem. Res. Toxicol., Vol. 10, No. 3, 1997 313

Figure 2. FAB-MS spectra of (a) PGJ2-M3 and (b) PGJ2-M5.

Figure 3. Structures of the glutathione conjugates of PGA2 and PGJ2.

enzymatic activities which were 5.4, 3.9, and 4.5 times as high as in the chemical reaction. At the highest concentration (600 µM), these ratios were still 2.2, 2.1, and 3.0 times as high, respectively. Figure 7 shows the enzymatic activities of GST A1-1, GST A2-2, GST M1a-1a, and GST P1-1 with regard to the formation of the two metabolites of PGJ2, representing about 85% of the radioactive peaks. GST M1a-1a and GST P1-1 catalyzed the formation of both metabolites, while GST A1-1 again was clearly selective with regard to the formation of one of the metabolites, i.e., PGJ2-M2. PGJ2-M2 seemed to be formed to a small extent by GST A2-2. As expected, almost similar amounts of PGJ2-M2 and PGJ2-M3 were formed in the chemical reaction. At the lowest concentration, GST A1-1 (PGJ2-M2), GST M1a-1a (PGJ2-M2 and M3), and GST P1-1 (PGJ2-M2 and M3) showed enzymatic activities which were 2.7, 4.3, and 2.7 times as high as in the chemical reaction, respectively.

Figure 4. Structures of the R- and S-glutathione conjugates of PGA2 and PGJ2.

At the highest concentration these ratio’s were 2.0, 3.1, and 1.9, respectively. The kinetic parameters obtained for the GSH conjugation of PGA2 and PGJ2 are presented in Table 3. In general, the affinity of GST M1a-1a > GST A1-1 > GST P1-1 with regard to the formation of PGA2-M3 and PGJ2M2, the R-GSH conjugates of PGA2 and PGJ2, respectively. GST A1-1 showed the highest Vmax values with respect to the formation of these metabolites, although more pronounced for PGA2-M3 than for PGJ2-M2. With regard to the formation of PGA2-M4 and PGJ2-M3, GSTP1-1 showed a higher affinity than GSTM1a-1a, although much more obvious for PGA2-M3. Kinetic parameters could not be determined for the formation of PGJ2-M3 by GST M1a-1a due to its first-order behavior at the used concentration range (70-600 µM). GST M1a-

314 Chem. Res. Toxicol., Vol. 10, No. 3, 1997

Bogaards et al.

Table 1. NMR Data of PGA2 and the Glutathione Conjugates of PGA2a PGA2-M3 H

n

δ (ppm)

2 3 4 5 6 7a, 7b 8 10a 10b 11 12 13 14 15 16a, 16b 17 18 19 20

2 2 2 1 1 2 1 1 1 1 1 1 1 1 2 2 2 2 3

2.40 t 1.68 m 2.09 dt 5.51 m 5.35 m 2.36 m 2.44 m cosy 2.31 m 2.98 dd 3.29 dt 2.53 m 5.62 dd 5.68 dd 4.18 dt 1.51, 1.61 m 1.31 m 1.31 m 1.31 m 0.88 t

GSH part glu-R glu-β glu-γ cys-R cys-β cys-β gly-R

1 2 2 1 1 1 2

4.00 t 2.23 dd 2.59 m 4.60 dd 2.99 dd 3.15 dd 4.03 s

PGA2-M4 J (Hz)

δ (ppm)

6.4 7.4 cosy 7.4 cosy 11.5 11.5, 7.8 10.5 av, 7.8 cosy 15.2, 7.5 15.2, 5.7 6.4 av

6.4 6.6 7.3, 2.9

2.40 t 1.67 m 2.09 dt 5.53 m 5.36 m 2.36 m 2.58 m 2.58 m 2.82 dd 3.78 dt 3.07 m 5.82 dd 5.68 dd 4.19 dt 1.54, 1.61 m 1.31 m 1.31 m 1.31 m 0.88 t

7.4 7.1 7.1

19.0 19.0, 6.9 6.2 av, 1.7 cosy 15.4, 8.1 15.4, 7.0 6.5 av

6.4

4.04 t 2.24 dd 2.58 m 4.61 dd 2.94 dd 3.09 dd 4.03 s

8.6, 5.3 13.8, 7.8 13.8, 5.4

PGA2 J (Hz)

6.5 7.3, 2.9 cosy 7.9, 6.0 13.7, 7.8 13.7, 5.4

δ (ppm)

J (Hz)

2.30 t 1.66 m 2.14 m 5.44 m 5.40 m 2.28, 2.48 m 2.08 m 7.59 dd

7.0 7.0, 7.4

6.11 dd 3.29 m 5.62 m 5.64 m 4.04 m 1.48 m 1.40 m 1.30 m 1.30 m 0.88 t

5.6, 1.9

cosy 5.6, 5.2

6.3

GSH 3.78 t 2.17 dt 2.56 dt 4.57 dd 2.97 dd 2.93 dd 3.78 s

6.4 7.6, 6.4 7.6 7.0, 5.4 14.2, 7.0 14.2, 5.4

a Abbreviations: av ) average of two; cosy ) additional correlations observed inn the COSY spectrum, the coupling constants, however, were to small to be determined; if no coupling constants are presented, overlapping proton signals were observed. The coupling constants presented are the observed couplings.

Table 2. NMR Data of PGJ2 and the Glutathione Conjugates of PGJ2a PGJ2-M2

a

PGJ2-M3

PGJ2

H

n

δ (ppm)

δ (ppm)

J (Hz)

δ (ppm)

J (Hz)

2 3 4 5 6 7a, 7b 8 9 10a 10b 12 13 14 15 16a, 16b 17 18 19 20

2 2 2 1 1 2 1 1 1 1 1 1 1 1 2 2 2 2 3

2.40 c 1.69 c 2.13 c 5.58 c 5.56 c 2.45 c 2.15 c 3.28 c 2.46 c 3.05 c 2.95 c 5.48 c 5.62 c 4.14 c 1.44, 1.52 c 1.31 c 1.31 c 1.31 c 0.88 c

2.42 t 1.71 m 2.14 m 5.57 m 5.60 m 2.35 m 2.50 m 3.82 t 2.69 d 2.90 dd 2.89 dd 5.49 dd 5.65 dd 4.16 dt 1.53, 1.61 m 1.31 m 1.31 m 1.31 m 0.88 t

7.4 7.4 av 7.4 av 10.4 10.4 cosy 5.9 cosy 6.7 av, cosy 19.1 1.91, 6.7

2.30 t 1.66 m 2.14 m 5.52 dt 5.49 dt 2.36, 2.40 m 2.85 m 6.08 dd 7.72 dd

7.3 7.4 av 7.6 av 10.6, 4.8 10.6, 4.9 14 2.4 5.9, 2.4 5.9, 2.4

2.62 dd 5.55 dd 5.65 dd 4.03 dt 1.48 m 1.40 m 1.30 m 1.30 m 0.88 t

7.3, 3.0 15.1, 7.3 15.1, 6.2 6.3

GSH part glu-R glu-β glu-γ cys-R cys-β cys-β gly-R

1 2 2 1 1 1 2

4.00 c 2.23 c 2.59 c 4.60 c 2.99 c 3.15 c 4.03 c

4.04 t 2.24 m 2.61 dt 4.67 m 2.95 dd 3.10 dd 4.05 s

6.4 cosy 7.2, 3.2 cosy 13.5, 7.5 13.5, 7.5

15.0, 7.9 15.0, 7.3 7.0 av

6.8

GSH 3.78 t 2.17 dt 2.56 dt 4.57 dd 2.97 dd 2.93 dd 3.78 s

6.5 6.4 7.6, 6.4 7.6 7.0, 5.4 14.2, 7.0 14.2, 5.4

For abbreviations see Table 1.

1a clearly showed a much higher affinity with respect to the formation of PGA2-M3/PGJ2-M2 than for PGA2-M4/ PGJ2-M3.

Discussion In the present study, the conjugation of PGA2 and PGJ2 with glutathione was shown to result in two major

conjugates of each prostaglandin. On longer incubation times, HPLC analysis showed a total of four peaks for PGA2, while five peaks were found in incubation with PGJ2. Mass spectroscopic analysis revealed that all peaks were glutathione conjugates of the investigated prostaglandins, of which one metabolite for PGJ2 is probably formed by dehydration at C15-C16 after GSH

Stereoselective Conjugation of PGA2 and PGJ2

Chem. Res. Toxicol., Vol. 10, No. 3, 1997 315

Figure 5. Radioactivity profiles of PGA2 and PGJ2 (250 µM) incubated with purified GSTs (65 µg/mL): (a) PGA2 with GST A1-1 and (b) PGJ2 with GST M1a-1a. Incubations and HPLC analyses were performed as described in Materials and Methods.

conjugation and rearrangement into a conjugated dienone. Both for PGA2 and PGJ2, the structures of two metabolites, representing about 85-90% of the radioactive peaks, could be elucidated by NMR/COSY analysis. The proposed structures for the two GSH-conjugates of PGA2, PGA2-M3 and PGA2-M4, are the R- and S-GSH conjugates of PGA2, respectively. In the three-dimensional structures of the R- and S-GSH conjugates of PGA2 (as calculated with the software program Alchemy III, V2.0, Tripos Associates, Inc.), the calculated dihedral angles (φ) between H11 and H12, H10a, and H10b in the R-enantiomer were found to be 166, -38, and -163°. According to the Karplus equation, which relates the vicinal coupling constants to the torsion angle between the coupling protons (17), the corresponding 3J protonproton couplings are in the range of 9-14, 4.5-8.5, and 9-14 Hz. The actual coupling constants found on H11 for PGA2-M3 were 10.5 (average of two) and 7.8 Hz (Table 1). The dihedral angles between H11 and H12, H10a, and H10b in the S-enantiomer were found to be 49, 79, and -45°, corresponding to expected values for the 3J coupling constants in the range of 3-6, 0-1, and 4-7.5 Hz. The actual coupling constants were 6.2 (average of two) and 1.7 Hz. The determined coupling constants for PGA2-M3 and PGA2-M4 were in the same order as the expected ones, thus confirming the proposed structures of the GSH conjugates of PGA2. The proposed structure for one of the two GSH conjugates of PGJ2, PGJ2-M3, is the S-GSH conjugate of PGJ2. The dihedral angles between H9 and H8, H10a, and

H10b in the S-enantiomer of PGJ2 were found to be 42, -46, and 79°, corresponding to expected 3J values in the range of 4.5-8.5, 4-7.5, and 0-1 Hz. The actual coupling constants were 6.7 (average of two) and a very small coupling (based on the observation of a very weak correlation in the COSY spectrum), thus confirming the proposed structure of this GSH conjugate of PGJ2. No structure could be proposed for PGJ2-M2 on the basis of obtained NMR data. However, based on the observation that the PGA2 and PGJ2 conjugates are formed in similar amounts in the non-enzymatic reactions, and in line with the structures found for PGA2, the most likely structure for PGJ2-M2 is the R-GSH conjugate of PGJ2. Incubation experiments clearly showed the role of individual GSTs in the conjugation of PGA2 and PGJ2. Striking stereoselectivities in conjugating activities were observed for GST A1-1 and GST P1-1. While GST A1-1 prefers the formation of the R-GSH conjugates of both prostaglandins, GST P1-1 showed a clear selectivity with regard to the formation of the S-GSH conjugate of PGA2. Glutathione conjugates are transported out of the cell by the ATP-dependent glutathione S-conjugate export pump (GS-X pump) (18). Since product inhibition is a wellknown phenomenon for GSTs, the GS-X pump seems to play an important role in determining the rate of conjugation of certain substrates. The R- and S-GSH conjugates of PGA2 and PGJ2 could be transported out of the cell at different rates since the GS-X pump may very well display a different selectivity toward the two enantiomers of PGA2 and PGJ2. Thus, the observed

316 Chem. Res. Toxicol., Vol. 10, No. 3, 1997

Bogaards et al.

Figure 6. Enzymatic activities of GST A1-1, GST A2-2, GST M1a-1a, and GST P1-1 (65 µg/mL) toward the formation of the glutathione conjugates of PGA2: (a) PGA2-M3 and (b) PGA2-M4. Incubations and HPLC analyses were performed as described in Materials and Methods. Activities are expressed as nmol of product/5 min/100 µL of incubation mixture.

Figure 7. Enzymatic activities of GST A1-1, GST A2-2, GST M1a-1a, and GST P1-1 (65 µg/mL) toward the formation of the glutathione conjugates of PGJ2: (a) PGJ2-M2 and (b) PGJ2-M3. Incubations and HPLC analyses were performed as described in Materials and Methods. Activities are expressed as nmol of product/5 min/100 µL of incubation mixture.

stereoselective conjugation of the prostaglandins might have possible consequences with respect to modulation of the antiproliferative activity of PGA2 and PGJ2.

In the present study, incubations were performed with a GST concentration of 1.3 µM. At this GST concentration, PGA2 and PGJ2 were conjugated up to 5.4-fold faster

Stereoselective Conjugation of PGA2 and PGJ2

Chem. Res. Toxicol., Vol. 10, No. 3, 1997 317

Table 3. Kinetic Parameters for the Formation of the GSH Conjugates of PGA2 and PGJ2a PGA2 conjugates PGA2-M3 GST A1-1 GST A2-2 GST M1a-1a GST P1-1

PGA2-M4

PGJ2 conjugates PGJ2-M2

PGJ2-M3

(4)

Km

Vmax

Km

Vmax

Km

Vmax

Km

Vmax

(5)

160 26 -

121 27 -

765 395

345 267

270 95 450

102 72 86

∞ 210

∞ 105

(6)

a Kinetic parameters were calculated with the curve-fitting program for the analysis of enzyme kinetic data EZ-FIT (16). Km and Vmax are expressed as µM and nmol/min/mg of protein, respectively.

compared to the chemical reaction at a prostaglandin concentration of 70 µM and up to 3.1-fold faster at 600 µM. Because the enzymatic conjugations usually follow Michaelis-Menten kinetics (Table 3), the ratio between enzymatic and chemical conjugation at lower (physiological) concentrations, will be at least as high as observed at 70 µM. In addition, the ratio between enzymatic and chemical conjugation will increase with higher GST concentrations, which are for example found in tumor cells (19). In proliferating cells like tumor cells, overexpression of GST P1-1 and to a lesser extent GST A1-1 is observed. Overexpression of GSTs has been associated with drug resistance in chemotherapy (8, 20). However, increased conjugation of prostaglandins due to overexpression of GSTs may also accelerate cell proliferation because of inhibition of the antiproliferative mode of action. In addition, GST P1-1 expression has been related to proliferative stages of development: GST P1-1 is predominant in fetal liver in early gestation, while in infants and adults GST P1-1 is only weakly expressed (21-23). Increased levels of GSTP1-1 and to a lesser extent GSTA1-1, as observed in proliferative cells, and the results of the present study in which the role of GSTP1-1 and GSTA1-1 in the GSH-conjugation of PGA2 and PGJ2 was clearly demonstrated, support the conclusions of Atsmon et al. (3) and Sato et al. (12) that GSH and GST may indeed be involved in cell proliferation, although direct evidence still has to be provided. In the literature, the reversibility of GSH-conjugation of R,β-unsaturated ketones and aldehydes has been described (10, 24, 25). This reversibility increases rapidly with pH (25). In the present study, the NMR spectra were measured in acidified solutions to prevent deconjugation of GSH. At physiological pH, deconjugation of the glutathione conjugates of cyclopentenone prostaglandins may however be a distinct possibility. The fact that GSH conjugation may not be a final step in this case may help interpret the conflicting results as described in literature.

References (1) Kato, T., Fukushima, M., Kurozumi, S., and Noyori, R. (1986) Antitumor activity of ∆7-prostaglandin A1 and ∆12-prostaglandin J2 in vitro and in vivo. Cancer Res. 46, 3538-3542. (2) Ito, S., Narumiya, S., and Hayaishi, O. (1989) Prostaglandin D2: A biochemical perspective. Prostaglandins, Leukotrienes Essent. Fatty Acids 37, 219-234. (3) Atsmon, J., Freeman, M. L., Meredith, M. J., Sweetman, B. J., and Roberts, L. J. (1990) Conjugation of 9-deoxy-∆9,∆12(E)-

(7)

(8) (9)

(10)

(11)

(12) (13)

(14)

(15) (16) (17) (18) (19)

(20) (21)

(22)

(23) (24) (25)

prostaglandin D2 with intracellular glutathione and enhancement of its antiproliferative activity by glutathione depletion. Cancer Res. 50, 1879-1885. Parker, J., and Ankel, H. (1992) Formation of a prostaglandin A2-glutathione conjugate in L1210 mouse leukemia cells. Biochem. Pharmacol. 43, 1053-1060. Ohno, K., and Hirata, M. (1993) Characterization of the transport system of prostaglandin A2 in L-1210 murine leukemia cells. Biochem. Pharmacol. 46, 661-670. Kim, I., Lee, J., Sohn, H., Kim, H., and Kim, S. (1993) Prostaglandin A2 and ∆12-prostaglandin J2 induce apoptosis in L1210 cells. FEBS Lett. 321, 209-214. Honn, K. V., and Marnett, L. J. (1985) Requirement of a reactive R,β-unsaturated carbonyl for inhibition of tumor growth and induction of differentiation by “A” series of prostaglandins. Biochem. Biophys. Res. Commun. 129, 34-40. Mannervik, B., and Danielson, U. H. (1988) Glutathione transferasessStructure and catalytic activity. CRC Crit. Rev. Biochem. 23, 283-337. Chien, C., Kirollos, K. S., Linderman, R. J., and Dauterman, W. C. (1994) R,β-Unsaturated carbonyl compounds: Inhibition of rat liver glutathione S-transferase isozymes and chemical reaction with reduced glutathione. Biochem. Biophys. Acta 1204, 175180. Atsmon, J., Sweetman, B. J., Baertschi, S. W., Harris, T. M., and Roberts, L. J. (1990) Formation of thiol conjugates of 9-deoxy∆9,∆12(E)-prostaglandin D2 and ∆12(E)-prostaglandin D2. Biochemistry 29, 3760-3765. Ohno, K., Hirata, M., Narumiya, S., and Fukushima, M. (1992) Effect of glutathione content on cellular uptake and growth inhibitory activity of prostaglandin A2 in L-1210 cells. Eicosanoids 5, 81-85. Sato, Y., Fujii, S., Fujii, Y., and Kaneko, T. (1990) Antiproliferative effects of glutathione S-transferase inhibitors on the K562 cell line. Biochem. Pharmacol. 39, 1263-1266. A° lin, P., Jensson, H., Guthenberg, C., Danielson, U. H., Tahir, M. K., and Mannervik, B. (1985) Purification of major basic glutathione transferase isoenzymes from rat liver by use of affinity chromatography and fast liquid chromatofocusing. Anal. Biochem. 146, 313-320. Bogaards, J. J. P., Van Ommen, B., and Van Bladeren, P. J. (1989) An improved method for the separation and quantification of glutathione S-transferase subunits in rat tissue using highperformance liquid chromatography. J. Chromatogr. 474, 435440. Habig, W. H., Pabst, M. J., and Jakoby, W. B. (1974) Glutathione S-transferases. The first step in mercapturic acid formation. J. Biol. Chem. 249, 7130-7139. Perella, F. W. (1988). EZ-FIT: A practical curve-fitting microcomputer program for the analysis of enzyme kinetic data on IBMPC compatible computers. Anal. Biochem. 174, 437-447. Karplus, M. (1959) Contact electron-spin coupling of nuclear magnetic moments. J. Chem. Phys. 30, 11-15. Ishikawa, T. (1992) The ATP-dependent glutathione S-conjugate export pump. Trends Biochem. Sci. 17, 463-468. Dirven, H. A. A. M., van Ommen, B., and van Bladeren, P. J. (1994). Involvement of human glutathione S-transferase isoenzymes in the conjugation of cyclophosphamide metabolites with glutathione. Cancer Res. 54, 6215-6220. Tsuchida, S., and Sato, K. (1992) Glutathione transferases and cancer. CRC Crit. Rev. Biochem. Mol. Biol. 27, 337-384. Strange, R. C., Faulder, C. G., Davis, B. A., Hume, R., Brown, J. A. H., Cotton, W., and Hopkinson, D. A. (1984) The human glutathione S-transferases: Studies on the tissue distribution and genetic variation of the GST1, GST2 and GST3 enzymes. Ann. Hum. Genet. 48, 11-20. Strange, R. C., Davis, B. A., Faulder, C. G., Cotton, W., Bain, A. D., Hopkinson, D. A., and Hume, R. (1985) The human glutathione S-transferases: Developmental aspects of the GST1, GST2, and GST3 loci. Biochem. Genet. 23, 1011-1028. van Ommen, B., Bogaards, J. J. P., Peters, W. H. M., Blaauboer, B., and van Bladeren, P. J. (1990) Quantification of human hepatic glutathione S-transferases. Biochem. J. 269, 609-613. Witz, G. (1989) Biological interactions of R,β-unsaturated aldehydes. J. Free Radicals Biol. Med. 7, 333-349. Ploemen, J. H. T. M., van Schanke, A., van Ommen, B., and van Bladeren, P. J. (1994) Reversible conjugation of ethacrynic acid with glutathione and human glutathione S-transferase P1-1. Cancer Res. 54, 915-919.

TX9601770