Glucuronidation of Hydroxylated Polychlorinated Biphenyls (PCBs

Graduate Center for Toxicology, University of Kentucky Chandler Medical Center, 306 ..... Acta Crystallographica Section E Structure Reports Online 20...
2 downloads 0 Views 148KB Size
Chem. Res. Toxicol. 2002, 15, 1259-1266

1259

Glucuronidation of Hydroxylated Polychlorinated Biphenyls (PCBs) Nilufer Tampal,† Hans-Joachim Lehmler,† Parvaneh Espandiari,† Tina Malmberg,‡ and Larry W. Robertson*,† Graduate Center for Toxicology, University of Kentucky Chandler Medical Center, 306 Health Sciences Research Building, Lexington, Kentucky 40536-0305, and Department of Environmental Chemistry, Stockholm University, 10691 Stockholm, Sweden Received March 13, 2002

Polychlorinated biphenyls (PCBs) may be metabolized to hydroxylated compounds. While many of these metabolites are further converted to either the glucuronic acid or the sulfate conjugates by phase II enzymes, which facilitates their excretion, some hydroxylated PCBs persist in the body. This may reflect their inability to be conjugated. A possible role of uridine diphosphate glucuronosyl transferase (UGT) in the elimination of hydroxylated metabolites of PCBs was therefore investigated. Glucuronidation studies of PCB metabolites included ones which are eliminated with relative ease and also ones which are reported to be retained in blood. Liver microsomes, prepared from male Wistar rats treated by intraperitoneal injections of phenobarbital for 3 days (400 µmol/kg/day), were used as the source of UGT. Enzyme kinetics (Vmax and Km) were determined for each of the metabolites. The efficiency of glucuronidation (Vmax/Km) was found to vary from 99% by GC-MS analysis (Mass Spectrometry Facility of the University of Kentucky, Lexington, KY), combustion analysis (Atlantic Microlab, Atlanta, GA), and thin-layer chromatography. 4-OH2,2′,3,3′,4′,5-hexaCB, 4-OH-2,2′,3,4′,5,5′-hexaCB, 4-OH-2,2′,3,4′, 5,5′,6-heptaCB, and 4-OH-2,2′,3,3′,4,5,5′-heptaCB were synthesized and characterized as described elsewhere (12, 27). Caution: PCBs and their metabolites are reasonably anticipated to be human carcinogens and should therefore be handled in an appropriate manner. 3,3′,4′,5-Tetrachloro-4-methoxybiphenyl. mp ) 108-109 °C (>99% by GC-MS). 1H NMR (CDCl3, 200 MHz): δ 3.95 (s, OCH3, 3H), 7.32 (d, Jortho ) 8.4 Hz, d, Jmeta ) 2.2 Hz, H-6′), 7.45 (s, H-2,6), 7.50 (d, Jortho ) 8.4 Hz, H-5′), 7.58 (d, Jmeta ) 2.2 Hz, H-2′). 13C NMR (CDCl3, 50 MHz): δ 60.85 (OCH3), 126.04 (CH), 127.25 (2 CH), 128.68 (CH), 129.96 (CH), 130.91, 133.20, 136.07, 138.17, 152.19. IR (cm-1): 3071, 2942, 2834, 1464, 1427, 1360, 1301, 1259, 1136, 1029, 993. MS m/z (relative intensity, %): 320 (77, C13H8Cl4O•+), 305 (59, M-CH3), 277 (20, M-CO-CH3), 207 (35, M-CO-CH3-Cl2). Elemental analysis calcd for C13H8Cl4O: C 48.49, H 2.50; found: C 48.45, H 2.51. 3,3′,4′,5-Tetrachloro-4-hydroxybiphenyl. mp ) 187 °C (>99%). 1H NMR (CDCl3, 200 MHz): δ 3.98 (br s, -OH, 1H), 7.61-7.64 (m, ArH, H-3′,4′), 7.69 (s, ArH, H-2,6), 7.84-7.88 (m, ArH, H-2′). 13C NMR (CDCl3, 50 MHz): δ 123.26, 127.34, 127.76, 129.29, 131.75, 131.87, 132.23, 133.24, 139.55, 150.04. IR (cm-1): 3501 (ν(OH)), 1496, 1462, 1412, 1362, 1326, 1305, 1233, 1179, 1136, 1029, 870, 819, 799, 714, 693, 547, 511, 408. MS m/z (relative intensity, %): 306 (80, C12H6Cl4O•+), 207 (45, M-99). Elemental analysis calcd for C12H6Cl4O: C 46.80, H 1.96; found: C 46.76, H 2.01. Methods. 1. Animal Treatment and Preparation of Hepatic Microsomes. Male Wistar rats from Harlan Sprague Dawley Inc. (Indianapolis, IN) 6-7 weeks old were housed in a controlled environment maintained at 22 °C with a 12 h lightdark cycle, and with free access to food and water. Animals were treated intraperitoneally with sodium phenobarbital in saline (400 µmol/kg) for 3 days to increase the specific activity of

Conjugation of PCB Metabolites by UGT microsomal UGT. Animals were euthanized under carbon dioxide asphyxiation, 24 h after the last injection. Livers were excised and washed twice with ice-cold 0.25 M sucrose solution containing 0.1 mM EDTA, pH 7.4, then minced and homogenized in 40 mL of the above sucrose solution, at low speed using a Tekmar tissue homogenizer. All further operations were performed at 4 °C. The homogenate was centrifuged at 10000g for 20 min. The resulting supernatant was then centrifuged at 100000g for 1 h. Microsomal pellets were washed twice with cold sucrose/EDTA solution and resuspended in that solution to a final protein concentration of 10 mg/mL after pooling the microsomes from 6 rats. Microsomal preparations were stored at -80 °C until further use. Protein concentrations were determined by the method of Lowry (28). 2. Enzyme Assay. Enzyme activity was measured using the method of Bansal and Gessner (29) with modifications. Briefly, microsomes (200 µg of protein) were treated with Brij 58 (Brij/ protein, 0.25 w/w) for 20 min at 0 °C. Incubation mixtures were prepared containing 100 mM Tris-HCl, pH 7.4, 5 mM MgCl2, 4 mM nonradiolabeled UDPGA, 0.1 µCi of [14C]UDPGA, 5 mM saccharolactone, and the aglycon (final concentrations ranging from 0.04 to 4 mM) in 2 µL of ethyl acetate. The mixtures and the activated microsomes were preincubated for 2 min at 37 °C, and the reaction was started by addition of the activated microsomes to the incubation mixture. The final volume of reaction mixture was 100 µL. After incubation for 10 min at 37 °C in a shaking water bath, the reaction was stopped by adding 200 µL of absolute alcohol at -20 °C. Tubes were centrifuged, protein pellet was discarded, and the supernatant was further analyzed. Zero time blanks were conducted by adding 200 µL of cold alcohol to the reaction mixture prior to addition of microsomes. Control blanks were also carried out to account for background activity due to glucuronidation of endogenous compounds in the microsomal preparation by substituting the aglycon in the incubation mixture with 2 µL of ethyl acetate. 3. Thin-Layer Chromatography (TLC). Aliquots (50 µL) of alcoholic supernatant were applied to precoated silica gel plates which were developed in tanks saturated with a developing solvent system consisting of 1-butanol/acetone/glacial acetic acid/ammonia (30%)/water (70:50:18:1.5:60). Plates were airdried. 4. Electronic Autoradiography. Radioactivity in the chromatographed sample was detected and quantitated using a Packard Instantimager. The spots were visualized on a monitor screen, and areas of the spots were traced using the mouse. This provided the counts per minute (cpm) for the traced areas. The fraction of UDPGA converted to the product was determined as the ratio of cpm obtained from the spot corresponding to the glucuronide in the reaction mixture, to the cpm obtained from the spot corresponding to UDPGA in the zero time blank. Radioactivity of the product was converted to concentration units by considering that 1 mol of the aglycon combines with 1 mol of UDPGA to give 1 mol of the product. 5. Hydrolysis of Glucuronides. To confirm that the radioactive products were indeed the glucuronides of the aglycon, reactions were performed as described above, but without saccharolactone in the reaction mixture. After incubation for 10 min, 50 µL aliquots were removed and treated with 75 µL of β-glucuronidase (50 000 units/mL) in 0.2 M sodium acetate, pH 4.6. The mixture was further incubated at 37 °C for 6 h. The reaction was stopped by adding 125 µL of absolute alcohol at -20 °C and the reaction product analyzed as before after separation using TLC. Any nonspecific hydrolysis of the glucuronide product was measured by simultaneously running controls where 50 µL aliquots were incubated with 75 µL of 0.2 M sodium acetate only, without β-glucuronidase. 6. Structural Characteristics. For each of the compounds tested, structural characteristics such as the dihedral angle, molecular volume, and molecular surface area were calculated with MM2* using GB/SA water solvent continuum as implemented by MacroModel 5.0 (30). Estimates for pKa (an estimate of the composition of a mixture of molecules and ions), log P

Chem. Res. Toxicol., Vol. 15, No. 10, 2002 1261 Table 1. Effect of Saccharolactone on Glucuronide Formationa compound

in absence of saccharolactone

in presence of saccharolactone

2-OH-4′-CB 4-OH-3,5-diCB 4-OH-2′,3,5-triCB 4-OH-2′,3,4′,5-tetraCB

36.5 ( 2.41 62.4 ( 1.44 28.4 ( 3.71 6.90 ( 1.10

36.5 ( 1.09 62.9 ( 2.83 32.9 ( 3.46 11.6 ( 0.36

a Values represent the amount of product formed with different PCB metabolites (4 mM), expressed as nmol ( SD (n ) 3), in the presence and in the absence of 5 mM saccharolactone. Differences were not statistically significant at p < 0.05, as determined by ANOVA.

Table 2. Enzyme Activity as a Function of UDPGAa concentration of UDPGA (mM)

glucuronide formed (nmol)

1 2 3 4

38 ( 7.0 52* ( 7.7 53* ( 9.6 51* ( 2.3

a Saturating concentrations of the cofactor UDPGA required for the assay were determined by measuring glucuronide formation (expressed in nmol ( SEM, n ) 3) in the presence of increasing concentrations of UDPGA, and a fixed concentration (1 mM) of the substrate, 4-OH-biphenyl. An asterisk indicates differences are not statistically significant at p < 0.05.

(an estimate of the octanol/water partition coefficient of neutral molecules), and log D (an estimate of the partition coefficient for a mixture of neutral molecules and ions) were obtained using the software program from Advanced Chemistry Development Inc. (Ontario, Canada). 7. Statistical Calculations. Analysis of Variance (ANOVA) was conducted to determine statistical significance. Comparisons among treatment groups, wherever necessary, were carried out using Bonferroni post-hoc procedures. All statistical tests were performed using SYSTAT software (version 8.0) for Windows (SPSS Inc., Chicago, IL), and the level of significance, R, was 0.05 for all statistical testing. Regression analysis was performed using the SAS program (PROC General Linear Model).

Results Reaction conditions were determined for optimal product formation in the assay. Incubation conditions were not optimized to linearity for every PCB tested, but 4-OHbiphenyl was used as a surrogate for this purpose. Plots of glucuronide formed vs time and protein were linear under the assay conditions; 5 mM saccharolactone was used to inhibit the breakdown of the glucuronide. The inhibitory effect of saccharolactone on enzyme activity was tested for four compounds, picked randomly, but possessing a mono-, di-, tri-, and tetrachloro substitution pattern. The amount of product formed was comparable or slightly greater in the presence of saccharolactone (Table 1). However, this difference was not statistically significant as determined by ANOVA. For evaluation of the optimal concentration of the cofactor UDPGA, enzyme activity as a function of UDPGA concentration at a constant aglycon concentration was measured. The aglycon used was 4-OH-biphenyl. As seen in Table 2, the amount of glucuronide formed remained almost constant at concentrations of UDPGA 2 mM and above. No statistically significant difference in activity was observed at UDPGA concentrations greater than 1 mM; 4 mM UDPGA was used in all incubation mixtures to ensure a saturated concentration of the cofactor.

1262

Chem. Res. Toxicol., Vol. 15, No. 10, 2002

Tampal et al.

Figure 2. Separation of glucuronide formed in the reaction mixture from the unreacted UDPGA for 4-OH-4′-CB using the TLC procedure. At each concentration, measurements were performed in triplicate. Control blanks were performed in the absence of aglycon, while zero time blanks were performed using inactivated microsomes.

Activation of UGT by two nonionic detergents (Triton X-100 and Brij 58) was investigated also using 4-OHbiphenyl as the aglycon. The detergent:protein ratio used for activation ranged from 0 to 0.5 (w/w). Brij 58 produced greater activation at all ratios, compared to Triton X-100. Brij 58 was deemed to be the better of the two detergents and therefore used in the assay at an optimal ratio of 0.25 Brij/protein (w/w). Figure 2 shows the separation of the reaction mixtures on a TLC plate. UGT activity for each PCB was determined in triplicate at all concentrations. The low background enzyme activity calculated from the assay blanks for control incubations performed without the aglycon in the incubation mixture was subtracted from the total enzyme activity at each concentration. No spot corresponding to the glucuronide in the zero blank was observed, as the microsomes were deactivated by the addition of absolute alcohol prior to incubation. Spots corresponding to the glucuronide from the reaction mixture showed a decrease in radioactivity when treated with β-glucuronidase (Figure 3A). A decrease in the amount of glucuronide after hydrolysis with β-glucuronidase in buffer and a lack of decrease after hydrolysis with buffer alone confirmed that the reaction product was a glucuronide (Figure 3B). Vmax and apparent Km were calculated from the Michaelis-Menten plots using the Prism Graphpad (version 3.0) software, and the efficiency of the enzyme for glucuronidation was compared after determining the Vmax/Km ratios of each of the PCBs. Figure 4 compares the efficiency for glucuronidation of study compounds. For the higher chlorinated compounds (4-7 chlorine atoms) that have been detected in mammalian tissues, the Vmax ranged from 0.3 to 6 nmol/min/mg. Lower chlorinated compounds (1-3 chlorine atoms) exhibited a Vmax in the range of 6-31 nmol/min/mg. The Km for these compounds did not vary to the same extent as their Vmax (Table 3), so their efficiency for glucuronidation (i.e., Vmax/Km) followed the same pattern as their Vmax. In an attempt to establish structure-activity relationships, we compared the effect of the glucuronidation rate on the position of hydroxyl group on the ring structure. As seen in Figure 5, the presence of a hydroxyl group meta or para to the dihedral bond was not as favorable to enzyme activity as the presence of a hydroxyl group ortho to this bond. The effect of the chlorine substitution pattern of the nonphenolic ring on enzyme efficiency was studied next. Substitution in this ring was found to greatly lower the Vmax. 4-OH-3,5-diCB with no substitution on the nonphenolic ring exhibited the greatest

Figure 3. Method validation for glucuronide conjugate formation by hydrolysis of the glucuronide product (G) in the presence of β-glucuronidase in sodium acetate buffer (G1), and in the presence of sodium acetate buffer alone (G2) for two PCB metabolites, 2-OH-4′-CB and 4-OH-3,5-diCB. (A) TLC pattern. (B) Radioactivity of spots corresponding to glucuronides, expressed as nanomoles of product formed.

efficiency for conjugation. As chlorine atoms were introduced in that ring, a pattern similar to that observed for the position of the hydroxyl group became apparent; i.e., the rate decreased in the order 4′ < 3′ < 2′ (Figure 6). 4-OH-2′,3,5-triCB with a chlorine atom in the ortho position was glucuronidated almost twice as efficiently as 4-OH-3,3′,5-triCB and 4-OH-3,4′,5-triCB, that possess a chlorine atom in the unfavorable meta and para positions, respectively. When both these positions were occupied as in 4-OH-3,3′,4′,5-tetraCB, efficiency of the enzyme dropped dramatically, to5 times lower than that of 4-OH-3,5-diCB and 4 times lower than 4-OH-2′,3,5triCB. The differences were statistically significant at p < 0.05. The efficiency improved upon moving the chlorine atom from the meta position to the ortho position of the nonphenolic ring as in 4-OH-2′,3,4′,5-tetraCB; however, it was not statistically different from that of 4-OH3,3′,4′,5-tetraCB. Figure 7 shows the effect of steric hindrance around the hydroxyl group. 4-OH-3,3′,4′-triCB and 4-OH-3,3′,5′triCB, both having only one chlorine atom adjacent to the hydroxyl group, were glucuronidated about 4 times faster than 4-OH-3,3′,4′,5-tetraCB in which the hydroxyl group is flanked by two adjacent chlorine atoms. However, their rates were almost the same as for 4-OH-4′CB which has no steric hindrance around the hydroxyl group. On the other hand, 4-OH-3,5-diCB with the

Conjugation of PCB Metabolites by UGT

Chem. Res. Toxicol., Vol. 15, No. 10, 2002 1263

Figure 5. Effect of hydroxyl group position on enzyme activity. Experiments were performed in triplicate as described under Materials and Methods, using increasing concentrations of the aglycon in the presence of 4 mM UDPGA. Results are means ( standard deviations.

Figure 4. Efficiency of glucuronidation of some hydroxylated PCBs. Experiments were performed in triplicate as described under Materials and Methods, using increasing concentrations of the aglycon in the presence of 4 mM UDPGA. Results are means ( standard deviations. Table 3. Kinetic Constants for Glucuronidation of Hydroxylated PCBsa hydroxy PCB metabolites

Vmax (nmol/min/mg)

Km (mM)

2-OH-4′-CB 3-OH-4′-CB 4-OH-4′-CB 4-OH-3,5-diCB 4-OH-2′,3,5-triCB 4-OH-3,3′,5-triCB 4-OH-3,4′,5-triCB 4-OH-3,3′,4′-triCB 4-OH-3,3′,5′-triCB 4-OH-3,3′,4′,5-tetraCB 4-OH-2′,3,4′,5-tetraCB 4-OH-2,2′,3,3′,4′,5-hexaCB 4-OH-2,2′,3,4′,5,5′-hexaCB 4-OH-2,2′,3,3′,4′,5,5′-heptaCB 4-OH-2,2′,3,4′,5,5′,6-heptaCB

19.3 ( 0.54 14.1 ( 0.66 9.51 ( 0.56 31.3 ( 1.65 15.2 ( 0.60 10.4 ( 0.41 6.25 ( 0.39 8.40 ( 0.51 9.03 ( 0.47 2.11 ( 0.09 5.57 ( 0.14 1.89 ( 0.31 0.62 ( 0.06 0.97 ( 0.06 0.31 ( 0.02

0.25 ( 0.03 0.14 ( 0.03 0.26 ( 0.03 0.27 ( 0.05 0.17 ( 0.03 0.22 ( 0.03 0.15 ( 0.03 0.09 ( 0.01 0.15 ( 0.03 0.09 ( 0.01 0.14 ( 0.01 0.17 ( 0.10 0.20 ( 0.02 0.10 ( 0.01 0.13 ( 0.01

Figure 6. Effect of substitution of the chlorine atom on the nonphenolic ring of the PCB metabolite on enzyme activity. Experiments were performed in triplicate as described under Materials and Methods, using increasing concentrations of the aglycon in the presence of 4 mM UDPGA. Results are means ( standard deviations. An asterisk indicates statistically significant from 4-OH-3,5-diCB at p < 0.05.

a Kinetic constants were determined using hepatic microsomal UGTs from phenobarbital-treated male Wistar rats. Increasing concentrations of the aglycon were incubated with 2 mg/mL activated proteins, in Tris-HCl buffer, pH 7.4, at 37 °C for 10 min in the presence of 4 mM UDPGA. Vmax and Km were calculated from the Michaelis-Menten plot, using Prism Graphpad (version 3.0) software. Values are means ( standard deviations for n ) 3.

hydroxyl group completely flanked by two chlorines had the fastest rate of glucuronidation of all the compounds investigated. To gain insight into other factors affecting the rate of glucuronidation, structural characteristics such as the bond angle, surface volume, and surface area of the compounds were estimated (Table 4), and any possible relationship with the constants Km and Vmax was probed using linear regression analysis. Km failed to show a significant relationship with any of the parameters (r2 e 0.3 at p g 0.5). Vmax, on the other hand, related inversely with all three parameters. While surface area and surface volume related strongly with Vmax (r2 ) 0.57 and 0.54, respectively, at p e 0.05), the dihedral angle, which is a measure of the planarity of the molecule,

Figure 7. Effect of steric hindrance around the hydroxyl group on enzyme activity. Experiments were performed in triplicate as described under Materials and Methods, using increasing concentrations of the aglycon in the presence of 4 mM UDPGA. Results are means ( standard deviations.

showed a weak relationship with Vmax (r2 ) 0.26 at p < 0.05). We next calculated the pKa (an estimate of the composition of a mixture of molecules and ions at pH 7.0), log P (an estimate of the octanol/water partition coef-

1264

Chem. Res. Toxicol., Vol. 15, No. 10, 2002

Tampal et al.

Table 4. Calculated Structural Characteristics of Hydroxylated PCBs Used in the Study

hydroxylated PCBs

dihedral angle (deg)

molecular suface area (Å2)

molecular surface volume (Å3)

log P

log D at pH 7.0

pKa

2-OH-4′-CB 3-OH-4′-CB 4-OH-4′-CB 4-OH-3,5-diCB 4-OH-2′,3,5-triCB 4-OH-3,3′,5-triCB 4-OH-3,4′,5-triCB 4-OH-3,3′,4′-triCB 4-OH-3,3′,5′-triCB 4-OH-3,3′,4′,5-tetraCB 4-OH-2′,3,4′,5-tetraCB 4-OH-2,2′,3,3′,4′,5-hexaCB 4-OH-2,2′,3,4′,5,5′-hexaCB 4-OH-2,2′,3,3′,4′,5,5′-heptaCB 4-OH-2,2′,3,4′,5,5′,6-heptaCB

53 38 37 38 58 38 38 38 38 38 57 78 74 78 90

395 398 398 415 436 439 439 438 441 459 459 490 494 510 504

179 179 179 194 208 210 210 210 210 225 225 255 254 269 270

3.51 3.80 3.77 4.27 4.74 4.83 4.84 4.74 4.87 5.28 5.33 6.17 6.23 6.55 6.77

3.50 3.80 3.80 3.90 4.20 4.40 4.40 4.70 4.80 4.50 4.60 4.30 4.40 4.40 4.20

9.48 9.62 9.71 6.94 6.57 6.72 6.81 7.82 7.86 6.36 6.30 5.04 5.04 4.73 4.08

ficient of neutral molecules), and log D (an estimate of the partition coefficient for a mixture of neutral molecules and ions at pH 7.0) for these compounds (Table 4). log P and log D related inversely and pKa related directly with Vmax. Again, a poor relationship between these parameters and Km was found (r2 ) 0.02, 0.002, and 0.33, respectively, at p g 0.5). A weak relationship was observed for pKa and log D with Vmax (r2 ) 0.38 and 0.31, respectively, at p < 0.05), while a relatively stronger relationship between log P and Vmax was observed (r2 ) 0.52 at p < 0.05). However, when the combined influence of all the parameters on Vmax was assessed using multiple regression analysis, only the surface area/surface volume of the molecule related significantly with Vmax (r2 ) 0.59 at p < 0.05).

barbital is known to induce the UGT2B family of isozymes responsible for glucuronidation of bulky phenols such as 4-OH-biphenyl (36). Boutin and co-workers reported UGT activity for this aglycon to be comparable in Wistar rats and man (37). To ensure maximum product formation, microsomes were pretreated with a nonionic detergent, Brij 58, to perturb the phospholipid environment of the enzyme which otherwise exerts a constraint on the enzyme structure (38, 39). Using intact membranes for kinetic studies would limit the significance as transport of substrates across the membrane would be ratelimiting. The detergent:protein ratio has a bearing on the Vmax; hence, it needs to be assessed before use in the assay. Higher concentrations of the detergent can inactivate the enzyme (40). A ratio of 0.25 Brij:protein (w/w) was found to be optimal in our studies.

Discussion

Although the majority of the β-glucuronidase in a cell is present in the lysozymes, a small amount is also found in the endoplasmic reticulum (41); hence, saccharolactone, a known inhibitor of this enzyme, was added to the incubation mixture. Comparing the efficiency for glucuronidation among the PCBs tested, those metabolites which are not reported to persist in mammals with a free hydroxyl group had a much higher enzyme efficiency (35-116 µL/min/mg) as compared to some PCBs such as the 4-hydroxy metabolites of PCB 146 (4-OH2,2′,3,4′,5,5′-hexaCB), PCB 172 (4-OH-2,2′,3,3′,4′,5,5′heptaCB), and PCB 187 (4-OH-2,2′,3,4′,5,5′,6-heptaCB) found in mammalian blood (