Structural Requirements for the Interaction of 91 Hydroxylated Polychlorinated Biphenyls with Estrogen and Thyroid Hormone Receptors
http://www.100md.com
《毒物学科学杂志》
National Institute for Environmental Studies, 16-2 Onogawa, Tsukuba, Ibaraki 305-8506, Japan
Environmental Pollution Control Center, Osaka Prefectural Government, 1-3-62 Nakamichi, Higashinari-ku, Osaka 537-0025, Japan
EDC Analysis Center, Otsuka Pharmaceutical Co. Ltd., 224-18 Hiraishi-Ebisuno, Kawauchi-cho, Tokushima 771-0195, Japan
ABSTRACT
Estrogenic and thyroid activities of 91 monohydroxylated PCBs were measured with two-hybrid assays using yeast cells containing the human estrogen receptor ER or human thyroid receptor TR. Estrogenic activity of 30 of the 91 compounds, including all compounds active in the yeast two-hybrid assay, were also measured by a reporter gene assay employing Chinese hamster ovary cells. The mammalian cell assay was more sensitive than the yeast assay but the rank order of estrogenicities of the compounds were in broad agreement for the two assays. Results for estrogenicity and thyroid activity were analyzed by inspection and those for estrogenicity by a theoretical treatment. Inspection indicated para-hydroxyl was more likely to be estrogenically active than meta-, which was more likely to be active than ortho-; one ortho-chlorine was important for activity but additional ortho-chlorines did not increase activity; and 2 lateral chlorines or 2,4,6-chloro- substitution of the non-phenol ring were favorable. In contrast, thyroid activity appeared not to depend strongly on the position of the hydroxyl group although ortho-hydroxyls occurred in the most active compounds. Activity was usually associated with at least one ortho-chlorine, with 2 chlorines in the phenolic ring and, importantly, two chlorines in the non-phenolic ring, and with 1 or 2 chlorines ortho to the hydroxyl group. Examination of the torsion angle between the rings, in the theoretical examination of estrogenicity, suggested that perpendicular orientation (i.e., rigidity) was not essential for activity. Intramolecular hydrogen bonding of the phenolic groups to adjacent chlorines or to the -electron cloud of the non-phenol ring possibly decreased activity—the hydroxyl should be free of intramolecular interactions for maximum activity. It was difficult to predict the estrogenic activity of a congener from its obtained potential energy curve (PEC). In general, estrogenically active congeners had large values for the sum of the atomic charges on the carbon atoms of the hydroxylated ring, and on the oxygen atom. Hydroxyl substitution at the para-position allowed the compounds to become more polarizable in the x-axis (molecular axis), whereas OH substitution at the ortho-position made the congeners less polarizable in the same direction. However, no general statement about polarizability and estrogenic activity was possible.
Key Words: hydroxylated polychlorinated biphenyls; hydro-xylated PCBs; estrogenicity; thyroid hormone activity; yeast two-hybrid assay; Chinese hamster ovary cell assay; structure-activity relationships.
INTRODUCTION
Polychlorinated biphenyls (PCBs) are industrial chemicals that have had a variety of uses but have been banned in most countries for some years They are metabolized in vivo to hydroxyl and sulfur compounds (Letcher et al., 2000) and hydroxylated metabolites have been found in human serum (Bergman et al., 1994), whole blood (Sandau et al., 2000), and plasma (Hovander et al., 2002). Hydroxylated PCBs have been shown to inhibit mitochondrial oxidative phosphorylation (Narasimhan et al., 1991), thyroid hormone sulfation (Schuur et al., 1998a,b,c), estrogen sulfotranferase (Kester et al., 2000) and the sulfation and glucuronidation of 3-hydroxy-benzo[a]pyrene (van den Hurk et al., 2002), to affect thyroxine (T4) levels (Sinjari and Darnerud, 1998) and to exhibit estrogenic or antiestrogenic activity (Bergeron et al., 1994; Fielden et al., 1997; Korach et al., 1988; Moore et al., 1997; Schultz, 2002; Vakharia and Gierthy, 2000).
Monohydroxyl-PCBs are attracting increasing attention as potentially endocrinologically active metabolites of PCBs. Recently a range of the compounds was found in fish from the Great Lakes area (Campbell et al., 2003; Li et al., 2003). A full assessment of their possible endocrine disrupting effects will follow the synthesis of the reported compounds. However detailed structure-activity relationships will most likely be elucidated by the examination of as many congeners as possible, including those not encountered in environmental samples.
There is some information on the influence of OH and Cl position on the estrogenicity of hydroxylated PCBs. Korach et al. (1988) compared the competitive binding of 7 mono-hydroxylated PCBs and 3 di-hydroxylated PCBs with those of estradiol (E2, Fig. 1a). The hydroxylated group(s) was para-orientated in each case in the Korach study and there was no opportunity to investigate the importance of phenol OH position. They concluded that the compounds with the stronger affinities (more estrogenic) had 2 ortho-chlorines (none of their compounds had 3) and it was the (steroid-like) rigidity provided by the ortho-chlorines that was important rather than any specific conformation. McKinney and Waller (1994) also stressed the importance of the para-OH group and ortho-chlorines and also mentioned "hydrophobic bulk" with the implication that the more chlorines the higher the activity. A study of a small number of hydroxylated PCBs in a recombinant yeast estrogenicity assay indicated that para-hydroxyl substituted compounds had the highest activities, followed by meta- and then ortho-substituted compounds (Schultz, 2002). Connor et al. (1997) investigated estrogenic and antiestrogenic effects of several synthetic hydroxylated PCBs using a range of assays. Again all compounds were para-phenols and there was no opportunity to investigate the effect of hydroxyl position. They concluded that structure- estrogenicity/antiestrogenicity relationships for their compounds were complex and response-specific.
Comparative molecular field analysis (CoMFA; Cramer et al., 1988), a three-dimensional quantitative structure-activity (QSAR) paradigm has been used to examine the physicochemical properties of hydroxylated PCBs that underlie their estrogen binding affinities and thus their potential estradiol equivalency (Waller et al., 1995). The conclusions were that non-coplanarity, manifest through ortho substitution, increased the steroid-like (more rigid) structural nature of the PCBs, contributed to their overall hydrophobic bulk, and possibly served to inhibit any competitive coplanar-type binding activity (dioxin-like).
The properties of hydroxylated PCB structure that makes them possibly active as thyroid hormone (Fig. 1b) mimics have not been examined in detail although it is obvious that the molecular structural requirements for thyroid hormone activity are very different from those required for estrogenicity. However, the binding of PCBs to proteins relevant to thyroid action have been investigated and the need for "lateral" chlorines to facilitate involvement in stacking and cleft-type interactions has been mentioned (McKinney and Waller, 1994).
We here report the estrogenicity and thyroid hormone activity of 91 hydroxylated PCBs measured with a yeast two-hybrid assay incorporating the human estrogen receptor hER or the human thyroid hormone receptor TR. The estrogenicity of 30 of the compounds, including those found to be active in the yeast two-hybrid assay, were also tested for estrogen agonist activity by a highly sensitive reporter gene assay using Chinese hamster ovary cells incorporating the human estrogen receptor hER. Fifty-eight of the 91 congeners were synthesized (Okumura and Shibata, 1975) as part of an environmental study following the "yusho" PCB poisoning incident in Japan in 1968 in which 1800 people suffered the effects of PCBs and other chlorinated hydrocarbons from the consumption of contaminated rice oil (Kuratsune et al., 1996). The hydroxylated PCBs available to us from this source were biased towards low chorine rather than high chlorine numbers because this reflected the involvement of Kanechlor 400, containing predominantly tri-, tetra-, and penta-chlorinated PCBs, in the "yusho" incident. Some information on 15 of these compounds is given in our earlier article (Shiraishi et al., 2003). The remaining 33 congeners that we examined were commercially available.
We make some general observations on the relationships between structure and activity for both estrogenicity and thyroid hormone activity by a simple inspection of the results, and we have supplemented the observations for estrogenicity by a more rigorous theoretical approach.
MATERIALS AND METHODS
Compounds. There is a total of 209 possible PCB congeners including those with all numbers of chlorine subsituents from 1 to the maximum 10. There are, however, 837 mono-hydroxylated PCB congeners with all numbers of chlorines from 1 to maximum 9. We examined 91 of these. Fifty-eight compounds were synthesized by thermal diazo-coupling between a chlorophenol and a chloro-aniline diazonium salt (24), and 33 were obtained from Accu Standard, Inc. (New Haven, CT). Compounds are listed in Table 1 including general footnote. The nomenclature of hydroxylated PCBs has been a matter for discussion (Maervoet et al., 2004). Nomenclature based on the Ballschmiter and Zell (1980) system (BZ) for numbering PCBs has been generally adopted and active hydroxylated PCB numbers using this scheme are given in the first column of Tables 1 and 2. Also given in Tables 1 and 2 are the IUPAC names of the active compounds. Although conveniently short, the BZ numbers for the hydroxylated PCBs convey very little structural information. In this article it was necessary to include full structural information when discussing structure and activity and for this purpose we have used an abbreviated form of the IUPAC name. In this scheme, chlorobiphenyl is reduced to CB and mono-, di-, tri-, tetra-, and penta- reduced to M, D, Tr, T, and Pe, respectively. For example, 2',4',6'-trichlorobiphenyl-4-ol becomes 2',4',6'-TrCB-4-ol. We consider this to be readily understandable and to convey full structural information.
Yeast two-hybrid assays. The agonist activities of the 91 compounds were measured with (1) a yeast two-hybrid estrogenicity assay using yeast cells (Saccharomyces cervisiae Y190) into which the human estrogen receptor ER and the coactivator TIF2 had been introduced, and (2) a yeast two-hybrid thyroid hormone activity assay using S. cervisiae Y190 cells into which the human thyroid hormone receptor TR and the coactivator TIF2 had been introduced. Expression plasmids for each hormone receptor ligand binding domain and pGAAD424-TIF-2 were introduced into yeast cells that carried the -galactosidase reporter gene (Nishikawa et al., 1999). Both assays were adapted to a chemiluminescent reporter gene (for -galactosidase) method employing a 96-well culture plate (Shiraishi et al., 2000). Yeast cells were preincubated for 24 h at 30°C with shaking in modified SD medium (lacking trypophan and leucine, 0.86% dextrose) and the cell density adjusted to an absorbance of 1.75 to 1.85 at 595 nm. The medium (60 μl) was placed in the wells of the first row of a black 96-well culture plate for chemiluminescence measurement. Wells in rows 2 to 8 were charged with a solution of 2% DMSO in the medium (60 μl). A solution of test compound (1 mM in DMSO, 20 μl) was added to medium (480 μl) and aliquots of this mixture (60 μl) were also added to the wells of the first row of the plate. The test solution was serially diluted from rows 1 to 7 (each 2x) and then the yeast cell suspension (60 μl) was also added to each well (including those in row 8, which served as the blank). Thus the first row contained a 10 μM solution of test chemical, the second row a 5 μM solution, and so on. After the addition of the yeast suspension and vortex mixing, the plates were incubated at 30°C under conditions of high humidity for 4 h. A solution (80 μl) for inducing chemiluminescence from released -galactosidase, consisting of reaction buffer (30 μl) containing GalactLux substrate (AURORA GAL-XE, ICN Biomedicals, Inc., Irvine, CA) and zymolase 20T solution for enzymatic digestion (50 μl) was added to each well. The plate was incubated at 37°C for 1 h and then placed in a 96-well plate luminometer (Luminescencer-JNR AB-2100, ATTO, Tokyo, Japan) and a light emission accelerator solution (AURORA GAL-XE, 50 μl) added to each well using the luminometer pump. The chemiluminescence produced by released -galactosidase in each well was measured.
For estrogenicity, agonist activity was recorded as the REC20 (20% relative effective concentration)—that is, the concentration of the test compound showing 20% of the activity of 10–8 M estradiol. For thyroid hormone activity the REC20 was defined as the concentration of the test compound showing 20% of the activity of 10–6 M T3.
Chinese hamster ovary cell assay. The estrogenic agonist activities of 30 of the 91 hydroxylated PCBs (Table 1) were measured with a reporter gene assay using Chinese hamster ovary cells incorporating the human estrogen receptor hER as has been previously described (Kojima et al., 2004). Again, activity was recorded as the REC20 (the concentration of test compound showing 20% of the activity of 10–10 M estradiol).
Theoretical structure-activity investigation of estrogenicity. We performed structural studies on 26 selected hydroxylated PCBs by using density functional theory to explore the possible factors that determine the estrogenic effects of these compounds. The compounds were selected to include active and inactive compounds (including those with highest activities) and ranging from di-chlorinated- to penta-chlorinated. A popular three-parameter hybrid functional, B3LYP (Becke, 1993; Lee et al., 1988) with 6-31G(d) and 6-311G(d,p) basis sets were utilized for this purpose. All computations were performed with Gaussian 98 and Gaussian 03 programs (Frisch et al., 1998, 2003). The geometries of compounds were optimized first at the B3LYP/6-31G(d) level of theory followed by frequency calculations, which showed that all of the optimized structures were minima on the potential energy surface. Frequency calculations also helped to obtain the polarizabilities of the congeners. Four possible isomeric structures (two with respect to the rotation of the phenyl rings, and two with respect to the OH orientation) of all the selected hydroxylated PCBs were explored. Next, the geometries were optimized by using triple- type basis set, 6-311G(d,p), which also included polarization functions for all atoms. Potential energy curves (PECs) [relative energy vs. torsional angle ()] for all the selected 26 compounds were drawn. For this purpose, structures at various values (from = 40° to = 140° in steps of 20°) were optimized at B3LYP/6-31G(d) level of theory. In addition to this, structures at = 90° were also optimized at the same level of theory to include in the PECs. All geometrical parameters were allowed to optimize. The atomic charges [at the B3LYP/6-311G(d,p)], for the most stable isomers of all the selected compounds, were derived by using electrostatic potential (ESP)-driven charges according to the Merz-Singh-Kollman scheme (Besler et al., 1990; Singh and Kollman, 1984). The results obtained in the theoretical calculations were mainly used to analyze the estrogenicity of the hydroxylated PCBs.
RESULTS AND DISCUSSION
Results of the assays are shown in Tables 1 and 2. Agonist activities were recorded as REC20 values. Although the CHO cell estrogenicity assay was more sensitive than the yeast two-hybrid assay, the rank orders of activities of the compounds tested by both assays were in broad agreement (Fig. 2). Some of the tested compounds showed both estrogenicity and thyroid hormone activity, but it was evident that, in general, different structural features were important for the higher activities for each receptor.
Estrogenicity
Inspection of the results from both assays suggested that an important factor in determining estrogenicity is the position of the phenolic hydroxyl group with highest activities resulting from a para-hydroxyl group. This was apparently assumed by Korach and co-workers (1988) who looked only at para-hydroxylated compounds. Of the 27 compounds with measurable activity in the yeast two-hybrid assay in the current study, 17 (63%) had para-OHs, 7 (26%) had meta-OHs, and 3 (11%) had ortho-OHs. Of the 64 inactive compounds 13 (20.3%) had para-OHs, 17 (26.6%) had meta-OHs, and 34 (53.1%) had ortho-OHs. Although the compounds in the CHO cell assay were selected on the basis of the yeast assay, within the selected group compounds with para-hydroxyls again showed, in general, the highest activity, followed by meta- and ortho-hydroxyls. The importance of the position of the hydroxyl group is clearly illustrated in Figures 3a and b, which show plots of hydroxyl position against log activity for both assays.
Inspection also revealed that estrogenic activity is highly dependent on the number of substituents ortho to the phenolic hydroxyl group. Fourteen of the 27 active compounds in the yeast two-hybrid assay (including the seven most active) had no ortho substituents; 9 had one chlorine ortho to the phenolic hydroxyl, 2 had 1 ortho-phenyl group only, 1 had 1 ortho-phenyl and 1 ortho chlorine, and 1 had 2 ortho chlorines (Fig. 3c). It is notable that in the latter mentioned few cases, where apparently unpromising environments for the phenolic OH are involved, the non-phenolic ring has what appears to be highly favored chlorine substitution patterns (see below) and these apparently counteract, to some extent, the negative influences on the phenol group. Compounds examined in the CHO cell assay, selected on the basis of their activities in the yeast assay, paralleled the results obtained in the yeast assay in regard to the number of substituents ortho- to the phenolic hydroxyl (Fig. 3d) although the relationship was not as strong. The report (Gantchev et al., 1994) that halogenation of the phenolic A ring of estradiol does not diminish, and in certain cases, enhances, the estrogen receptor binding to the ligand was not paralleled by our results for hydroxylated PCBs.
Another factor stressed by McKinney and Waller (1994) was the need for ortho-chlorines to ensure sufficient (steroid-like) rigidity. Of the active 27 compounds in the yeast assay in this study, 26 had at least one ortho-chlorine. Of the 13 inactive compounds with para-OH groups, 4 had no ortho-chlorines. However, additional ortho-chlorines beyond 1 did not appear to confer additional activity.
To further illustrate the importance to estrogenic activity of the number of chlorine atoms in both rings and their substitution patterns, we have listed series of compounds in which one structural feature has been held constant and others varied to reveal differences in activity and the molecular properties apparently responsible for them (Tables 3 and 4). Table 3 lists all compounds in the study with no chlorine substituents in the phenol rings. Going across the columns from left to right (para to meta- to ortho-phenolic hydroxyl groups) reinforces the observation that the position of the hydroxyl group is of fundamental importance to estrogenic activity. However, descending the rows from top to bottom shows that certain patterns of chlorine substitution in the non-phenol ring are also important for activity. This can clearly be seen for para- and meta-phenols, which have the same rank order for activity and this order depends on the chlorine substitution pattern in the non-phenol ring. This is not simply a case of the greater number of ortho-chlorines; there is no consistent trend in that regard—the most active compound in each case has two ortho chlorines, the second and third have one only, and the fourth has two. It seems rather that some particular property governed by the substitution pattern of chlorines in the non-phenolic ring is important in determining activity; compounds with two ortho chlorines and one para chlorine are the most active for both para- and meta-phenols in both assays. There were small differences in the rank order of the compounds for the two assays but they were in general agreement. Of the six compounds having phenol rings lacking chlorine but with ortho-OH groups (Table 3, column 3), only two were active in the yeast assay and one of these was active in the CHO cell assay; both of these had rather low activity. Again the ortho-, ortho-, para- chlorinated non-phenolic ring seemed to confer activity, even on the unpromising ortho-phenol.
The series shown in Table 4, in which the apparently favored ortho-, ortho-, para-chlorinated non-phenol ring is constant, illustrates again the importance of the OH group position, and shows that additional chlorines in the phenolic ring diminish or eliminate activity.
In summary:
Para-hydroxyl is more likely to be active (and have high activity) than meta-; which in turn is more likely to be active than ortho-.
One ortho-chlorine is important but additional ortho-chlorines do not appear to increase activity or the chance of a compound being active.
The number and pattern of chlorine substituents in the non-phenol ring appear to be important for activity. This is not just a question of the more ortho-chlorines the better (conferring steroid-type rigidity), or, more generally, the more the better ("hydrophobic bulk") but a specific pattern seems to favor activity and to be associated with 2 or 3 chlorines in the non-phenol ring. For example, two "lateral" chlorines (only) seem particularly advantageous, i.e., an ortho- and a meta-chlorine on the same side. Also tri substitution in a particular pattern seem to favor activity, namely, 2 ortho- and 1 para-chlorine.
Theoretical Consideration of Estrogenicity
It is known that the halogenated aromatic hydrocarbons (HAHs) with the highest binding affinities for aryl hydrocarbon (Ah) receptors are those with planar structures (Denison et al., 2002). Although there is no such general rule for HAHs to have large binding affinities for estrogen receptors (ER), the existing assumption is that steroid-like rigidity is important rather than any specific conformation. In other words, perpendicular orientation of the phenyl rings is considered to be an important parameter (Korach et al., 1988). But our results were not always supportive of this assumption. For example, the most active hydroxylated PCB in our list had perpendicular phenyl rings, while the phenyl rings in 2',5'-DCB-4-ol (third most active) were around 32° away from the perpendicular orientation at B3LYP/6-311G(d,p) level of theory. The value calculated for 2',3',4',5'-TCB-4-ol, the most active hydroxylated PCB congener in most of the earlier studies (Korach et al., 1988; Waller et al., 1995), is 59.3° at the same level of theory. 2',4',6'-TrCB-4-ol, one of the two most active congeners in most previous studies and the second most active congener in our list had a value of 96.3°, but the closely related 2',4',6'-TrCB-2-ol, which had a perfect perpendicular phenyl rings orientation was the least active congener in our list of 27 active congeners. 2,3',4,4',6-PeCB-3-ol, which had a perpendicular phenyl rings orientation, is inactive. Hence it seems that the perpendicular orientation may not be an essential parameter for a congener to have a large estrogenic activity. Indirectly, this also supports the observation that additional ortho-chlorines do not increase the activity.
Of the 27 active hydroxylated PCBs in our experimental study, 26 had at least one ortho-chlorine. Results showed that additional ortho-chlorines beyond 1 did not seem to confer additional activity. These results revealed that one ortho-chlorine was essential for a congener to acquire estrogenic activity. Our calculations show that a single ortho- substitution takes the phenyl rings around 60° away from each other (from co-planar orientation). Additional chlorine substitutions at ortho- positions, obviously, increase the repulsive interaction between overlapping ortho-chlorine atoms and increase the torsional angle further. With three ortho-chlorine atoms, the phenyl rings become perpendicular to each other. The values calculated for hydroxylated PCBs lacking ortho-chlorine atoms are close to 40°. So it is concluded that the hydroxylated PCBs are not active in a co-planar (about 40°) conformation (and do not show estrogenic activity), but rather with a 60–90° twist between the phenyl rings.
Earlier studies demonstrated that chlorine substitution in the phenolic ring does not significantly affect the estrogenic activity of hydroxylated PCBs, whereas 2,4,6-trichloro- and 2,3,4,6-tetrachloro-substitution in the non-phenolic ring gave compounds with the highest estrogenic activity (Korach et al., 1988). It has been further reported (Gantchev et al., 1994) that halogenation of the phenolic ring, in certain cases, enhances the estrogenic activity. We found that estrogenic activity is highly dependent on the number of substituents ortho- to the phenolic hydroxyl group. It is worthwhile to analyze these observations through a theoretical approach. Consider chlorine substitutions at the 2 and 3 positions in 2',4',6'-TrCB-4-ol. Substitution at the 3-position gives rise to 2',3,4',6'-TCB-4-ol which might exhibit intramolecular hydrogen bonding between the hydrogen of the hydroxyl group and the nearby chlorine atom; however, such hydrogen bonding is not possible in the case of chlorine substitution at the 2-position, (giving 2,2',4',6'-TCB-4-ol). This is evidenced by the relative energies obtained for the possible conformers in both congeners. The conformer of 2',3,4',5'-TCB-4-ol with the OH orientation favorable for such a probable hydrogen bonding (the ClH non-bonded bond length is 2.418 ) is energetically more stable by 12.65 kJ/mol [at B3LYP/6–311G(d,p) level of theory] than the conformer with the OH orientation which is not favorable to hydrogen bonding. But hydrogen bonding is not possible in 2,2',4',6'-TCB-4-ol and so there is no energy difference between the various conformers of this congener. The same is true for 2',5'-DCB-4-ol; while there is no real energy difference between the possible conformers of 2,2',5'-TrCB-4-ol, the conformer with the probable hydrogen bonding orientation in 2',3,5'-TrCB-4-ol is energetically favorable by 12.85 kJ/mol when compared to the conformer with unfavorable hydrogen bonding orientation. The non-bonded ClH bond length in the hydrogen bonded 2',3,5'-TrCB-4-ol conformer is 2.416 . So it is clear that an intramolecular hydrogen bonding of about 12 kJ/mol forms, where it can, between the hydrogen of OH group and the neighboring chlorine atom in the hydroxylated PCBs. That means chlorine substitution ortho- to the hydroxyl group favors such hydrogen bonding. This assumption is also evidenced from the results obtained for other hydroxylated PCBs. It is worth analyzing whether this intramolecular hydrogen bonding has any influence on the estrogenic activities of the congeners. In both the above cases, the estrogenic activities of the 3-substituted congeners (2',3,4',5'-TCB-4-ol and 2',3,5',TrCB-4-ol) are lower than those of 2-substituted congener (2,2',4',6'-TCB-4-ol and 2,2',5'-TrCB-4-ol). So it seems that the intramolecular hydrogen bonding decreases the activity. This can be assumed from analyzing the experimental results obtained for other hydroxylated PCBs too; of 2,2',3',4',5'-PeCB-4-ol and 2',3,3',4',5'-PeCB-4-ol, the former is active while the latter is inactive; 2,3',4'-TrCB-4-ol is active but its 3-chlorinated congener, 3,3',4'-TrCB-4-ol is inactive; 2-MCB-4-ol is more active than 3-MCB-4-ol.
It is evident that a para-substituted hydroxyl group is a property of estrogenically active congeners. 2',4',6'-TrCB-4-ol, 2',3',4',5'-TCB-4-ol, and 2',5'-DCB-4-ol are examples of active hydroxyl-PCBs with para-hydroxyl groups. However, the consequences for estrogenicity of hydroxyl group substitution at the meta- or ortho-positions is, apart from indications from the limited study of Schultz (2002), largely unknown. We used our calculations to investigate this subject. Analyzing the results revealed that the hydrogen of the hydroxyl group weakly interacts with the neighboring phenyl group (either with the chlorine atom or with the -electron cloud) if the OH group is substituted at ortho-position. Thorough investigation of the calculated relative energies showed that this weak interaction is energetically around 3 to 5 kJ/mol strong. However, the OH group substituted at the 3-position is free from such intramolecular interactions. It should be interesting to see whether this weak interaction has any role in the estrogenic activity of the congeners. First we considered 2',4',6'- and 2',5'-chlorinated congeners for this purpose. Our calculations indicated that congeners with the hydroxyl group substituted at the ortho-position (2',4',6'-TrCB-2-ol and 2',5'-DCB-2-ol) have such intramolecular interactions (between OH group and the neighboring phenyl group) of about 3 to 5 kJ/mol. Our experimental studies showed that the compounds with the hydroxyl group substituted at the meta-position in both the cases, 2',4',6'-TrCB-3-ol and 2',5'-DCB-3-ol, are active while their ortho-substituted (OH group) counterparts, 2',4',6'-TrCB-2-ol and 2',5'-DCB-2-ol, are either very weakly active or inactive. This gives the impression that these weak intramolecular interactions play a role in the estrogenic activities of hydroxylated PCBs. This can be examined in other compounds too. For example, in the case of the 2',3',4',5' substituted hydroxylated PCBs, 2',3',4',5'-TCB-3-ol is active whereas 2',3',4',5'-TCB-2-ol is inactive. The latter might have OH-phenyl group intramolecular interaction. Hence it seems that the existence of such intramolecular interactions decreases the estrogenic activities of hydroxylated PCBs. It is evident from the results that even among the hydroxylated PCBs with a chlorine substitution in their hydroxylated phenyl rings, compounds with the OH group in the para- position are more active than those with the OH group substituted at ortho- or meta- positions: 2,2',5'-TrCB-4-ol is more active than 2,2',5'-TrCB-3-ol; 2',3,3'-TrCB-4-ol is more active than 2',3,3'-TrCB-2-ol; 2',3,4'-TrCB-4-ol is active while 2',3,4'-TrCB-2-ol is not.
The foregoing facts clearly reveal that the hydroxyl group should be in a position where it can be highly flexible and is available to make strong intermolecular interactions with the estrogen receptors in the cell. Obviously, this should be possible at the para-position but the OH group should not have any chlorine substitution ortho- to it. In other words, the hydroxyl group should be completely free from any intramolecular interactions. This would suggest, interestingly, that such a free hydroxyl group could help to make a strong hydrogen bonding between the hydroxylated PCBs and the estrogen receptors (particularly with amino acids). It can be assumed that if the hydroxyl group is involved in an intramolecular interaction, it cannot make a strong intermolecular interaction with the ER. It has been hypothesized (Gantchev et al., 1994) that hydroxyl groups at position 2 in estradiol (E2) and in E2 further substituted at position 3 may share, via hydrogen bonding, a common acceptor/donor site in the receptor cavity. Hence our results indirectly support the hypothesis that intermolecular hydrogen bonding between the hydroxylated PCBs and the estrogen receptors plays an important role in the estrogenic activity of these compounds.
Previous studies have concluded that conformational restriction about the twist bond of the overall structure of a hydroxylated PCB appeared to play an important role in estrogen receptor binding affinity. In other words, as mentioned earlier, steroid-like (more rigid) structure is believed to be an essential parameter for the active congeners. In these circumstances, PEC studies on the hydroxylated PCBs could be very useful for a deeper understanding of this subject. We investigated the PECs of all 26 selected hydroxylated PCBs, and it is clear from our results that most of these compounds are energetically flexible for C–C interring rotation (Fig. 4). The energy difference due to the rotation of the phenyl rings from = 60° to = 120° is very small. This energy change in 2',4',6'-TrCB-4-ol, the second most active congener, is around 1 kJ/mol. It is around 5 and 2 kJ/mol, respectively, for the most active 2,2',4',6'-TCB-4-ol and the fifth most active 2,2',5'-TrCB-4-ol. These three active congeners have no (or negligible) torsional barrier at = 90°. The same, very small change in energy due to the rotation of the phenyl rings from = 60° to = 120° and no barrier at = 90°, is true for two other active congeners, 2',3,4',6'-TCB-4-ol and 2',4',6'-TrCB-3-ol. 2',5'-DCB-4-ol and 2',3',4',5'-TCB-4-ol, the third and fourth most active congeners, respectively, had barriers of about 4 kJ/mol at = 90°. However, it should be noted that the energy change due to the rotation of phenyl rings from = 40° to = 140° in these two hydroxylated PCBs is just 4 kJ/mol. Again, this is the same for 2',3,5'-TrCB-4-ol and 2',5'-DCB-3-ol, both of which are active. Even in 2',4',6'-TrCB-4-ol, 2',3,4',6'-TCB-4-ol, and 2',4',6'-TrCB-3-ol, the energy changes due to the rotation of phenyl rings from = 40° to = 140° are around 10 kJ/mol. All these results reveal that hydroxylated PCBs are not conformationally very restrictive about the twist bond. The structure is not rigid for rotation of the phenyl bonds at least from = 60° to = 120°. Furthermore, by considering the fact that the free energy for the binding of 17 -estradiol to the estrogen receptor is around 50 kJ/mol (Wiese and Brooks, 1994), even a much weaker binding of hydroxylated PCBs to the ER could provide the energy necessary for hydroxylated PCBs to freely rotate (or to change its conformation!) from = 60° to = 120°. This study also helps to make two other conclusions: (1) it is difficult to predict whether or not a congener will be estrogenically active from its PEC; and (2) torsional barriers at co-planar or near co-planar structures are very high and hence for these hydroxylated PCBs, conformationally moving towards co-planar structures is energetically very difficult.
Studies in the past illustrated that the electron donating properties of the hydroxyl group (phenolic oxygen) and the analog in the hydroxylated PCBs of the aromatic component of the estradiol A ring were found to be correlated with higher ER binding affinity (Bradbury et al., 1996). Based on experimental and theoretical findings (Palomino et al., 1994; VanderKuur et al., 1993a,b; Wiese and Brooks, 1994) for the involvement of a negative isopotential surrounding the ER binding of the E2 A ring and the assumption that there are cationic sites (amino acid residues) inside the ER (Amouri et al., 1992; Pakdel and Katzenellenbogen, 1992), there is a hypothesis that hydroxylated PCBs with greater binding affinity should have higher electron donating abilities associated with their corresponding phenolic ring. Such chemicals have larger negative charges on the phenolic oxygen and the aromatic carbons in the A ring. Atomic charges obtained in this study (Table 5) are mixed and it is not easy to draw any concrete conclusions. For example, in 2',4',6'-TrCB-4-ol the sum of the atomic charges on the carbon atoms of the hydroxylated ring (qC-ring) is smaller (in negative scale) than those in 2',4',6'-TrCB-3-ol and in 2',4',6'-TrCB-2-ol. This trend continued in 2',5'-chlorine substituted hydroxylated-PCBs too. In the above two groups, compounds with OH groups at ortho- positions, the least active congeners in their respective groups, had the larger qC-ring values. Again, the qC-ring value obtained for 2,2',4',6'-TCB-4-ol is smaller than that obtained for 2',3,4',6'-TCB-4-ol. The value obtained for 2,2',5'-TrCB-4-ol is smaller than that obtained for 2',3,5'-TrCB-4-ol. The total atomic charges on the hydroxylated phenyl ring (qring) in 2',4',6'-TrCB-4-ol is smaller than those in 2',4',6'-TrCB-3-ol and in 2',4',6'-TrCB-2-ol. The qring value obtained for 2',5'-DCB-4-ol is smaller than that obtained for 2',5'-DCB-2-ol. In fact the qring value calculated for 2',5'-DCB-4-ol is in the positive scale. The present study shows atomic charge on oxygen atom (qO) is in the order 2',4',6'-TrCB-3-ol > 2',4',6'-TrCB-4-ol > 2',4',6'-TrCB-2-ol, whereas the estrogenic activity order is 2',4',6'-TrCB-4-ol > 2',4',6'-TrCB-3-ol > 2',4',6'-TrCB-2-ol. The same was found in the 2',5'-substituted group too. Although these results show a complicated picture, they also reveal that, in general, the active congeners have large qO and qC-ring values. Hence estrogenic activity appears to be related to the presence of an optimum range of qO and qC-ring values.
The most polarizable conformers are assumed to be closely associated with ER binding affinity (Bradbury et al., 1996) and so polarizability might be an important parameter for estrogenic activity. So we calculated polarizability tensors (XX, YY, and ZZ) along the principle axes to know whether there is any direct relationship between polarizability and activity. The average polarizability, ave, was calculated as ave = (1/3)(XX + YY + ZZ). It was true that in some cases that active congeners had large polarizabilities. The ave values calculated for 2',4',6'-TrCB-4-ol, 2',4',6'-TrCB-3-ol, and 2',4',6'-TrCB-2-ol were in the same order as their estrogenic activities. This was true for the 2',5'-DCB-4-ol, 2',5'-DCB-3-ol, and 2',5'-DCB-2-ol series too. Analysis of the results obtained for the above two sets of congeners revealed that hydroxyl group substitution at the para-position allows the congeners to become more polarizable in the x-axis (molecular axis) whereas OH substitution at the ortho-position makes the congeners less polarizable in the same direction. However, making a generalized statement such as the active congeners have larger polarizabilities than their less active or inactive counterparts is not possible.
In summary:
Perpendicular orientation of the two rings (i.e., rigidity) is not necessary for activity.
Intramolecular hydrogen bonding of the phenolic hydroxyl to adjacent chlorines or to the -electron cloud of the non-phenol ring decreased activity.
Prediction of the estrogenic activity of a congener from its obtained potential energy curve was difficult.
In general, estrogenically active congeners had large values for the sum of the atomic charges on the carbon atoms of the phenol ring, and on the oxygen atom.
No general statement about polarizability and estrogenic activity was possible although hydroxyl substitution at the para-position allowed the compounds to become more polarizable in the x-axis (molecular axis), whereas hydroxylation at the ortho-position made the congeners less polarizable in the same direction.
Thyroid Activity
The natural thyroid hormones T3 and T4 (Fig. 1b) obviously exhibit the necessary structural properties for activity. Iodine must be substituted at position 3 and 5 in the inner ring to maintain the outer ring perpendicular to the inner ring, i.e., to prevent rotation of the rings relative to each other. A spacer atom (which can be O as in T3 and T4, or S or C) is required between the two rings and should form a 120° angle. The carboxyl and the phenolic hydroxyl groups must be present to maximize receptor binding. Lipophilic groups (iodine in the case of T3 and T4) must be present at the 3' position of the outer ring to optimize binding (Norman and Litwack, 1997).
The activities of the natural thyroid hormones and their analogues measured by the yeast two-hybrid assay used in this study do not exactly parallel the activities of these compounds at the organism level. For example, T3 is 3 to 8x as active as T4 in the human body (Norman and Litwack, 1997) but in our assay T4 is almost as active as T3, and reverse T3, which is virtually inactive in the human organism, shows appreciable activity. Activities of these and related synthetic hormones are shown in Figure 5. This characteristic of the assay must be considered when seeking to make correlations between structure and activity.
Inspection of the results (Table 2) suggested that although the most active compounds 2',4,5',6-TCB-2-ol and 3',5,5',6-TCB-2-ol had their phenolic hydroxyl groups in the ortho- position, the actual position of the phenolic hydroxyl group might be of rather less importance than the number of substituents (chlorine or phenyl ring) ortho- to the phenolic hydroxyl. It appears that at least one substituent ortho- to the hydroxyl is necesary for activity, and, perhaps, if this is the other phenyl ring the chances of activity are enhanced. Thus it might be inferred that the non-phenolic ring with its substituents is serving as the necessary hydrophobic bulk ortho- to the phenolic hydroxyl, i.e., playing the role of an iodine atom in the natural hormones. Although there is a slight preponderance of ortho-phenols, activity is rather evenly distributed between ortho-, meta-, and para- phenols—of 24 active compounds 10 (42%) are ortho-phenols, 7 (29%) are meta-, and 7 (29%) are para-. With regard to substituents ortho- to the phenolic hydroxyl, of the 24 active compounds, 14 (58%) have 1 ortho-substituent, 8 (33%) have 2, and only 2 (8%) have none. The ortho- substituent can of course be the other phenyl ring as well as a chlorine atom.
It also appeared that the pattern and degree of substitution of the non-phenolic ring might be important. Of the 24 active compounds 19 (79%) had 2 chlorines in the non-phenolic ring, 4 had 1, and 1 had 3. None had 4 or 0. Of the 67 non-active compounds 36 (53%) have 2 Cls, 7 have 3, 14 have 4, 4 have 0, and 6 have 1. Numbers of compounds with 4 chlorines in the non-phenol ring are particularly striking—no active compounds have 4 chlorines but 14 (21%) of the non-active compounds do. Furthermore, all active compounds except 2 (22 of 24, 92%) have at least one meta-chlorine in the non-phenolic ring.
There are no active compounds without at least one ortho-chlorine (even for ortho-phenols). So coplanarity is, as expected, not favorable for thyroid hormone activity. Seventeen of 91 compounds (19%) have no ortho-chlorines and all are inactive. Of these 6 do have ortho-OH groups but that does not confer activity in the absence of at least one ortho-chlorine.
In summary:
Thyroid hormone activity appears not to depend strongly on the position of the phenolic hydroxyl although ortho-hydroxyl occurs in the most active compounds suggesting that the other (non-phenol) phenyl ring might serve as the required ortho- hydrophobic bulk (the role of the iodine ortho- to the phenolic hydroxyl in T3).
Activity is usually associated with at least one ortho-chlorine irrespective of whether the phenolic hydroxy is in the ortho- position or not.
Activity is usually associated with 2 chlorines in the phenolic ring and, importantly, two chlorines in the non-phenolic ring.
Activity is usually associated with one or two chlorines ortho to the phenolic hydroxyl.
Hydroxylated PCBs in the Environment
Studies on the identification and quantification of hydroxylated PCBs in environmental samples (Asplund et al., 1999; Campbell et al., 2003; Hoekstra et al., 2003; Klasson-Wehler et al., 1998; Li et al., 2003; Olsson et al., 2000) and in humans (Guvenius et al., 2002; Sandau et al., 2000) have revealed, among a total of some 15 or so compounds, the presence of four of the compounds included in our list of synthetic hydroxylated PCBs. Interestingly, the most estrogenically active compound (2,2',4',6'-TCB-4-ol) in our assays, was identified and quantified in lake trout from Lake Ontario and Lake Superior (Campbell et al., 2003). The results of our estrogenicity assays for 2',4',6'-TCB-4-ol (second most active) and 2',3',4',5'-TCB-4-ol (fourth most active) were consistent with previous experimental work on the endocrine disrupting properties of these compounds in fish (Carlson and Williams, 2001; Gerpe et al., 2000).
REFERENCES
Amouri, H. E., Vessiers, A., Vichard, D., Top, S., Grusselle, M., and Jaouen, G. (1992). Syntheses and affinities of novel organometallic-labeled estradiol derivatives: A structure-affinity relationship. J. Med. Chem. 35, 3130–3135.
Asplund, L., Athanasiadou, M., Sjdin, A., Bergman, ., and Brjeson, H. (1999). Organohalogen substances in muscle, egg and blood from healthy Baltic salmon (Salmo salar) with the M74 syndrome. Ambio 28, 67–76.
Ballschmiter, K., and Zell, M. (1980). Analysis of polychlorinated biphenyls (PCB) by glass capillary gas chromatography. Fresenius J. Anal. Chem. 302, 20–31.
Becke, A. D. (1993). Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 98, 5648–5652.
Bergeron, J. M., Crews, D., and McLachlan, J. A. (1994). PCBs as environmental estrogens: Turtle sex determination as a biomarker of environmental contamination. Environ. Health Perspect. 102, 780–781.
Bergman, ., Klasson Wehler, E., and Kuroki, H. (1994). Selective retention of hydroxylated PCB metabolites in blood. Environ. Health Perspect. 102, 464–469.
Besler, B. H., Merz, K. M., Jr., and Kollman, P. A. (1990). Atomic charges derived from semiempirical methods. J. Comput. Chem. 11, 431–439.
Bradbury, S. P., Mekenyan, O. G., and Ankley, G. T. (1996). Quantitative structure-activity relationships for polychlorinated hydroxybiphenyl estrogen binding affinity: An assessment of conformer flexibility. Environ. Toxicol. Chem. 15, 1945–1954.
Campbell, L. M., Muir, D. C. G., Whittle, D. M., Backus, S., Norstrom, R. J., and Fisk, A. T. (2003). Hydroxylated PCBs and other chlorinated phenolic compounds in lake trout (Salvelinus namaycush) blood plasma from the Great Lakes region. Environ. Sci. Technol. 37, 1720–1725.
Carlson, D. B., and Williams, D. E. (2001). 4-Hydroxy-2',4',6'-trichorobiphenyl and 4-hydroxy-2',3',4',5'-tetrachlorobiphenyl are estrogenic in rainbow trout. Environ. Toxicol. Chem. 20, 351–358.
Connor, K., Ramamoorthy, K., Moore, M., Mustain, M., Chen, I., Safe, S., Zacharewski, T., Gillesby, B., Joyeux, A., and Balaguer, P. (1997). Hydroxylated polychlorinated biphenyls (PCBs) as estrogens and antiestrogens: Structure-activity relationships. Toxicol. Appl. Pharmacol. 145, 111–123.
Cramer, R. D., III, Patterson, D. E., and Bumce, J. D. (1988). Comparative molecular field analysis (CoMFA). 1. Effect of shape on binding of steroids to carrier proteins. J. Am. Chem. Soc. 110, 5959–5967.
Denison, M. S., Pandini, A., Nagy, S. R., Baldwin, E. P., and Bonati, L. (2002). Ligand binding and activation of the Ah receptor. Chem. Biol. Interact. 141, 3–24.
Fielden, M. R., Chen, I., Chittim, B., Safe, S. H., and Zacharewski, T. R. (1997). Examination of the estrogenicity of 2,4,6,2'6'-pentachlorobiphenyl (PCB 104), its hydroxylated metabolite 2,4,6,2',6'-pentachloro-4-biphenylol (HO-PCB 104), and a further chlorinated derivative, 2,4,6,2',4',6'-hexachlorobiphenyl (PCB 155). Environ. Health Perspect. 105, 1238–1248.
Frisch, M. J. et al. (1998). Gaussian 98, Revision A.7 Gaussian, Inc., Pittsburgh, PA.
Frisch, M. J. et al. (2003). Gaussian 03, Revision B.01; Gaussian, Inc., Pittsburgh, PA.
Gantchev, T. V., Ali, H., and van Lier, J. E. (1994). Quantitative structure-activity relationships/comparative molecular field analysis (QSAR/CoMFA) for receptor-binding properties of halogenated estradiol derivatives. J. Med. Chem. 37, 4164–4176.
Gerpe, M., Kling, P., Berg, A. H., and Olsson, P.-E. (2000). Arctic char (Salvelinus alpinus) metallothionein: cDNA sequence, expression, and tissue-specific inhibition of cadmium-mediated metallothionein induction by 17-estradiol, 4-OH-PCB30, and PCB104. Environ. Toxicol. Chem. 19, 638–645.
Guvenius, D. M., Hassanzadeh, P., Bergman, ., and Norén, K. (2002). Metabolites of polychlorinated biphenyls in human liver and adipose tissue. Environ. Toxicol. Chem. 21, 2264–2269.
Hoekstra, P. F., Letcher, R. J., O'Hara, T. M., Backus, S. M., Solomon, K. R., and Muir, D. C. (2003). Hydroxylated and methylsulfone-containing metabolites of polychlorinated biphenyls in the plasma and blubber of bowhead whales (Balaena mysticetus). Environ. Toxicol. Chem. 22, 2650–2658.
Hovander, L., Malmberg, T., Athanasiadou, M., Athanassiadis, I., Rahm, S., Bergman, ., and Klasson Wehler, E. (2002). Identification of hydroxylated PCB metabolites and other phenolic halogenated pollutants in human blood plasma. Arch. Environ. Contam. Toxicol. 42, 105–117.
Kester, M. H. A., Bulduk, S., Tibboel, D., Meinl, W., Glatt, H., Falany, C. N., Coughtrie, M. W. H., Bergman, ., Safe, S. H., Kuiper, G. G. J. M., Schuur, A. G., Brouwer, A., and Visser, T. J. (2000). Potent inhibition of estrogen sulfotransferase by hydroxylated PCB metabolites: A novel pathway explaining the estrogenic activity of PCBs. Endocrinology 141, 1897–1900.
Klasson-Wehler, E., Bergman, ., Athanasiadou, M., Ludwig, J. P., Auman, H. J., Kannan, K., Van Den Berg, M., Murk, A. J., Feyk, L. A., and Giesy, J. P. (1998). Hydroxylated and methylsulfonyl polychlorinated biphenyl metabolites in albatrosses from Midway Atoll, North Pacific Ocean. Environ. Toxicol. Chem. 17, 1620–1625.
Kojima, H., Katsura, E., Takeuchi, S., Niiyama, K., and Kobayashi, K. (2004). Screening for estrogen and androgen receptor activities in 200 pesticides by in vitro reporter gene assays using Chinese hamster ovary cells. Environ. Health Perspect. 112, 524–531.
Korach, K. S., Sarver, P., Chae, K., McLachlan, J. A., and McKinney, J. D. (1988). Estrogen receptor-binding activity of polychlorinated hydroxybiphenyls: Conformationally restricted structural probes. Mol. Pharmacol. 33, 120–126.
Kuratsune, M., Yoshimura, H., Hori, Y., Okumura, M., and Masuda, Y. (Eds.) (1996). Yusho. A Human Disaster Caused by PCBs and Related Compounds. Kyushu University Press, Fukuoka, Japan.
Lee, C., Yang, W., and Parr, R. G. (1988). Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density. Phys. Rev. B 37, 785–789.
Letcher, R. J., Klasson Wehler, E., and Bergman, . (2000). Methyl sulfone and hydroxylated metabolites of polychlorinated biphenyls. In New Types of Persistent Halogenated Compounds. (J. Paasivirta, Ed.), pp. 315–359. Springer-Verlag, Berlin.
Li, H., Drouillard, K. G., Bennett, E., Haffner, G. D., and Letcher, R. J. (2003). Plasma-associated halogenated phenolic contaminants in benthic and pelagic fish species from the Detroit River. Environ. Sci. Technol. 37, 832–839.
Maervoet, J., Covaci, A., Schepens, P., Sandau, C. D., and Letcher, R. J. (2004). A reassessment of the nomenclature of polychlorinated biphenyl (PCB) metabolites. Environ. Health Perspect. 112, 291–294.
McKinney, J. D., and Waller, C. L. (1994). Polychlorinated biphenyls as hormonally active structural analogues. Environ. Health. Perspect. 102, 290–297.
Moore, M., Mustain, M., Daniel, K., Chen, I., Safe, S., Zacharewski, T., Gillesby, B., Joyeux, A., and Balaguer, P. (1997). Antiestrogenic activity of hydroxylated polychlorinated biphenyl congeners identified in human serum. Toxicol. Appl. Pharmacol. 142, 160–168.
Narasimhan, T. R., Kim, H. L., and Safe, S. (1991). Effects of hydroxylated polychlorinated biphenyls on mouse liver mitochondrial oxidative phosphorylation. J. Biochem. Toxicol. 6, 229–236.
Nishikawa, J., Saito, K., Goto, J., Dakeyama, F., Matsuo, M., and Nishihara, T. (1999). New screening methods for chemicals with hormonal activities using interaction of nuclear hormone receptor with coactivator. Toxicol. Appl. Pharmacol. 154, 76–83.
Norman, A. W., and Litwack, G. (1997). Hormones, 2nd ed., Academic Press, San Diego.
Okumura, T., and Shibata, J. (1975). Investigation on the behavior of PCBs in the environment. Research Report of the Environmental Pollution Control Center, Osaka Prefectural Government, Japan (in Japanese).
Olsson, A., Ceder, K., Bergman, ., and Helander, B. (2000). Nestling blood of the white-tailed sea eagle (Haliaeetus albicilla) as an indicator of territorial exposure to organohalogen compounds-an evaluation. Environ. Sci. Technol. 34, 2733–2740.
Pakdel, F., and Katzenellenbogen, B. S. (1992). Human estrogen receptor mutants with altered estrogen and antiestrogen ligand discrimination. J. Biol. Chem. 267, 3429–3437.
Palomino, E., Heeg, M. J., Horwitz, J. P., Polin, L., and Brooks, S. C. (1994). Skeletal conformations and receptor binding of some 9,11-modified estradiols. J. Steroid Biochem. Mol. Biol. 50, 75–84.
Sandau, C. D., Ayotte, P., Dewailly, E., Duffe, J., and Norstrom, R. J. (2000). Analysis of hydroxylated metabolites of PCBs (OH-PCBs) and other chlorinated phenolic compounds in whole blood from Canadian Inuit. Environ. Health Perspect. 108, 611–616.
Schultz, T. W. (2002). Estrogenicity of biphenylols: Activity in the yeast gene activation assay. Bull. Environ. Contam. Toxicol. 68, 332–338.
Schuur, A. G., Brouwer, A., Bergman, ., Coughtrie, M. W. H., and Visser, T. J. (1998a). Inhibition of thyroid hormone sulfation by hydroxylated metabolites of polychlorinated biphenyls. Chem. Biol. Interact. 109, 293–297.
Schuur, A. G., Legger, F. F., van Meeteren, M. E., Moonen, M. J. H., van Leeuwen-Bol, I., Bergman, ., Visser, T. J., and Brouwer, A. (1998b). In vitro inhibition of thyroid hormone sulfation by hydroxylated metabolites of halogenated aromatic hydrocarbons. Chem. Res. Toxicol. 11, 1075–1081.
Schuur, A. G., van Leeuwen-Bol, I., Jong, W. M. C., Bergman, ., Coughtrie, M. W. H., Brouwer, A., and Visser, T. J. (1998c) In vitro inhibition of thyroid hormone sulfation by polychlorobiphenylols: Isozyme specificity and inhibition kinetics. Toxicol. Sci. 45, 188–194.
Shiraishi, F., Okumura, T., Nomachi, M., Serizawa, S., Nishikawa, J., Edmonds, J. S., Shiraishi. H., and Morita, M. (2003). Estrogenic and thyroid hormone activity of a series of hydroxy-polychlorinated biphenyls. Chemosphere 52, 33–42.
Shiraishi, F., Shiraishi, H., Nishikawa, J., Nishihara, T., and Morita, M. (2000). Development of a simple operational estrogenicity assay system using the yeast two-hybrid system. J. Environ. Chem. 10, 57–64.
Singh, U. C., and Kollman, P. A. (1984). An approach to computing electrostatic charges for molecules. J. Comput. Chem. 5, 129–145.
Sinjari, T., and Darnerud, P. O. (1998). Hydroxylated polychlorinated biphenyls: Placental transfer and erffects on thyroxine in the foetal mouse. Xenobiotica 28, 21–30.
Vakharia, D. D., and Gierthy, J. F. (2000). Use of a combined human liver microsome-estrogen receptor binding assay to assess potential estrogen modulating activity of PCB metabolites. Toxicol. Lett. 114, 55–65.
van den Hurk, P., Kubiczak, G. A., Lehmler, H.-J., and James, M. O. (2002). Hydroxylated polychlorinated biphenyls as inhibitors of the sulfation and glucuronidation of 3-hydroxy-benzo[a]pyrene. Environ. Health Perspect. 110, 343–348.
VanderKuur, J. A., Hafner, M. S., Christman, J. K., and Brooks, S. C. (1993a). Effects of estradiol-17 beta analogues on activation of estrogen response element regulated chloramphenicol acetyltransferase expression. Biochemistry, 32, 7016–7021.
VanderKuur, J. A., Wiese, T., and Brooks, S. C. (1993b). Influence of estrogen structure on nuclear binding and progesterone receptor induction by the receptor complex. Biochemistry 32, 7002–7008.
Waller, C. L., Minor, D. L., and McKinney, J. D. (1995). Using three-dimensional quantitative structure-activity relationships to examine estrogen receptor binding affinities of polychlorinated hydroxybiphenyls. Environ. Health. Perspect. 103, 702–707.
Wiese, T. E., and Brooks, S. C. (1994). Molecular modeling of steroidal estrogens: Novel conformations and their role in biological activity. J. Steroid Biochem. Mol. Biol. 50, 61–73.(Sundaram Arulmozhiraja, F)
Environmental Pollution Control Center, Osaka Prefectural Government, 1-3-62 Nakamichi, Higashinari-ku, Osaka 537-0025, Japan
EDC Analysis Center, Otsuka Pharmaceutical Co. Ltd., 224-18 Hiraishi-Ebisuno, Kawauchi-cho, Tokushima 771-0195, Japan
ABSTRACT
Estrogenic and thyroid activities of 91 monohydroxylated PCBs were measured with two-hybrid assays using yeast cells containing the human estrogen receptor ER or human thyroid receptor TR. Estrogenic activity of 30 of the 91 compounds, including all compounds active in the yeast two-hybrid assay, were also measured by a reporter gene assay employing Chinese hamster ovary cells. The mammalian cell assay was more sensitive than the yeast assay but the rank order of estrogenicities of the compounds were in broad agreement for the two assays. Results for estrogenicity and thyroid activity were analyzed by inspection and those for estrogenicity by a theoretical treatment. Inspection indicated para-hydroxyl was more likely to be estrogenically active than meta-, which was more likely to be active than ortho-; one ortho-chlorine was important for activity but additional ortho-chlorines did not increase activity; and 2 lateral chlorines or 2,4,6-chloro- substitution of the non-phenol ring were favorable. In contrast, thyroid activity appeared not to depend strongly on the position of the hydroxyl group although ortho-hydroxyls occurred in the most active compounds. Activity was usually associated with at least one ortho-chlorine, with 2 chlorines in the phenolic ring and, importantly, two chlorines in the non-phenolic ring, and with 1 or 2 chlorines ortho to the hydroxyl group. Examination of the torsion angle between the rings, in the theoretical examination of estrogenicity, suggested that perpendicular orientation (i.e., rigidity) was not essential for activity. Intramolecular hydrogen bonding of the phenolic groups to adjacent chlorines or to the -electron cloud of the non-phenol ring possibly decreased activity—the hydroxyl should be free of intramolecular interactions for maximum activity. It was difficult to predict the estrogenic activity of a congener from its obtained potential energy curve (PEC). In general, estrogenically active congeners had large values for the sum of the atomic charges on the carbon atoms of the hydroxylated ring, and on the oxygen atom. Hydroxyl substitution at the para-position allowed the compounds to become more polarizable in the x-axis (molecular axis), whereas OH substitution at the ortho-position made the congeners less polarizable in the same direction. However, no general statement about polarizability and estrogenic activity was possible.
Key Words: hydroxylated polychlorinated biphenyls; hydro-xylated PCBs; estrogenicity; thyroid hormone activity; yeast two-hybrid assay; Chinese hamster ovary cell assay; structure-activity relationships.
INTRODUCTION
Polychlorinated biphenyls (PCBs) are industrial chemicals that have had a variety of uses but have been banned in most countries for some years They are metabolized in vivo to hydroxyl and sulfur compounds (Letcher et al., 2000) and hydroxylated metabolites have been found in human serum (Bergman et al., 1994), whole blood (Sandau et al., 2000), and plasma (Hovander et al., 2002). Hydroxylated PCBs have been shown to inhibit mitochondrial oxidative phosphorylation (Narasimhan et al., 1991), thyroid hormone sulfation (Schuur et al., 1998a,b,c), estrogen sulfotranferase (Kester et al., 2000) and the sulfation and glucuronidation of 3-hydroxy-benzo[a]pyrene (van den Hurk et al., 2002), to affect thyroxine (T4) levels (Sinjari and Darnerud, 1998) and to exhibit estrogenic or antiestrogenic activity (Bergeron et al., 1994; Fielden et al., 1997; Korach et al., 1988; Moore et al., 1997; Schultz, 2002; Vakharia and Gierthy, 2000).
Monohydroxyl-PCBs are attracting increasing attention as potentially endocrinologically active metabolites of PCBs. Recently a range of the compounds was found in fish from the Great Lakes area (Campbell et al., 2003; Li et al., 2003). A full assessment of their possible endocrine disrupting effects will follow the synthesis of the reported compounds. However detailed structure-activity relationships will most likely be elucidated by the examination of as many congeners as possible, including those not encountered in environmental samples.
There is some information on the influence of OH and Cl position on the estrogenicity of hydroxylated PCBs. Korach et al. (1988) compared the competitive binding of 7 mono-hydroxylated PCBs and 3 di-hydroxylated PCBs with those of estradiol (E2, Fig. 1a). The hydroxylated group(s) was para-orientated in each case in the Korach study and there was no opportunity to investigate the importance of phenol OH position. They concluded that the compounds with the stronger affinities (more estrogenic) had 2 ortho-chlorines (none of their compounds had 3) and it was the (steroid-like) rigidity provided by the ortho-chlorines that was important rather than any specific conformation. McKinney and Waller (1994) also stressed the importance of the para-OH group and ortho-chlorines and also mentioned "hydrophobic bulk" with the implication that the more chlorines the higher the activity. A study of a small number of hydroxylated PCBs in a recombinant yeast estrogenicity assay indicated that para-hydroxyl substituted compounds had the highest activities, followed by meta- and then ortho-substituted compounds (Schultz, 2002). Connor et al. (1997) investigated estrogenic and antiestrogenic effects of several synthetic hydroxylated PCBs using a range of assays. Again all compounds were para-phenols and there was no opportunity to investigate the effect of hydroxyl position. They concluded that structure- estrogenicity/antiestrogenicity relationships for their compounds were complex and response-specific.
Comparative molecular field analysis (CoMFA; Cramer et al., 1988), a three-dimensional quantitative structure-activity (QSAR) paradigm has been used to examine the physicochemical properties of hydroxylated PCBs that underlie their estrogen binding affinities and thus their potential estradiol equivalency (Waller et al., 1995). The conclusions were that non-coplanarity, manifest through ortho substitution, increased the steroid-like (more rigid) structural nature of the PCBs, contributed to their overall hydrophobic bulk, and possibly served to inhibit any competitive coplanar-type binding activity (dioxin-like).
The properties of hydroxylated PCB structure that makes them possibly active as thyroid hormone (Fig. 1b) mimics have not been examined in detail although it is obvious that the molecular structural requirements for thyroid hormone activity are very different from those required for estrogenicity. However, the binding of PCBs to proteins relevant to thyroid action have been investigated and the need for "lateral" chlorines to facilitate involvement in stacking and cleft-type interactions has been mentioned (McKinney and Waller, 1994).
We here report the estrogenicity and thyroid hormone activity of 91 hydroxylated PCBs measured with a yeast two-hybrid assay incorporating the human estrogen receptor hER or the human thyroid hormone receptor TR. The estrogenicity of 30 of the compounds, including those found to be active in the yeast two-hybrid assay, were also tested for estrogen agonist activity by a highly sensitive reporter gene assay using Chinese hamster ovary cells incorporating the human estrogen receptor hER. Fifty-eight of the 91 congeners were synthesized (Okumura and Shibata, 1975) as part of an environmental study following the "yusho" PCB poisoning incident in Japan in 1968 in which 1800 people suffered the effects of PCBs and other chlorinated hydrocarbons from the consumption of contaminated rice oil (Kuratsune et al., 1996). The hydroxylated PCBs available to us from this source were biased towards low chorine rather than high chlorine numbers because this reflected the involvement of Kanechlor 400, containing predominantly tri-, tetra-, and penta-chlorinated PCBs, in the "yusho" incident. Some information on 15 of these compounds is given in our earlier article (Shiraishi et al., 2003). The remaining 33 congeners that we examined were commercially available.
We make some general observations on the relationships between structure and activity for both estrogenicity and thyroid hormone activity by a simple inspection of the results, and we have supplemented the observations for estrogenicity by a more rigorous theoretical approach.
MATERIALS AND METHODS
Compounds. There is a total of 209 possible PCB congeners including those with all numbers of chlorine subsituents from 1 to the maximum 10. There are, however, 837 mono-hydroxylated PCB congeners with all numbers of chlorines from 1 to maximum 9. We examined 91 of these. Fifty-eight compounds were synthesized by thermal diazo-coupling between a chlorophenol and a chloro-aniline diazonium salt (24), and 33 were obtained from Accu Standard, Inc. (New Haven, CT). Compounds are listed in Table 1 including general footnote. The nomenclature of hydroxylated PCBs has been a matter for discussion (Maervoet et al., 2004). Nomenclature based on the Ballschmiter and Zell (1980) system (BZ) for numbering PCBs has been generally adopted and active hydroxylated PCB numbers using this scheme are given in the first column of Tables 1 and 2. Also given in Tables 1 and 2 are the IUPAC names of the active compounds. Although conveniently short, the BZ numbers for the hydroxylated PCBs convey very little structural information. In this article it was necessary to include full structural information when discussing structure and activity and for this purpose we have used an abbreviated form of the IUPAC name. In this scheme, chlorobiphenyl is reduced to CB and mono-, di-, tri-, tetra-, and penta- reduced to M, D, Tr, T, and Pe, respectively. For example, 2',4',6'-trichlorobiphenyl-4-ol becomes 2',4',6'-TrCB-4-ol. We consider this to be readily understandable and to convey full structural information.
Yeast two-hybrid assays. The agonist activities of the 91 compounds were measured with (1) a yeast two-hybrid estrogenicity assay using yeast cells (Saccharomyces cervisiae Y190) into which the human estrogen receptor ER and the coactivator TIF2 had been introduced, and (2) a yeast two-hybrid thyroid hormone activity assay using S. cervisiae Y190 cells into which the human thyroid hormone receptor TR and the coactivator TIF2 had been introduced. Expression plasmids for each hormone receptor ligand binding domain and pGAAD424-TIF-2 were introduced into yeast cells that carried the -galactosidase reporter gene (Nishikawa et al., 1999). Both assays were adapted to a chemiluminescent reporter gene (for -galactosidase) method employing a 96-well culture plate (Shiraishi et al., 2000). Yeast cells were preincubated for 24 h at 30°C with shaking in modified SD medium (lacking trypophan and leucine, 0.86% dextrose) and the cell density adjusted to an absorbance of 1.75 to 1.85 at 595 nm. The medium (60 μl) was placed in the wells of the first row of a black 96-well culture plate for chemiluminescence measurement. Wells in rows 2 to 8 were charged with a solution of 2% DMSO in the medium (60 μl). A solution of test compound (1 mM in DMSO, 20 μl) was added to medium (480 μl) and aliquots of this mixture (60 μl) were also added to the wells of the first row of the plate. The test solution was serially diluted from rows 1 to 7 (each 2x) and then the yeast cell suspension (60 μl) was also added to each well (including those in row 8, which served as the blank). Thus the first row contained a 10 μM solution of test chemical, the second row a 5 μM solution, and so on. After the addition of the yeast suspension and vortex mixing, the plates were incubated at 30°C under conditions of high humidity for 4 h. A solution (80 μl) for inducing chemiluminescence from released -galactosidase, consisting of reaction buffer (30 μl) containing GalactLux substrate (AURORA GAL-XE, ICN Biomedicals, Inc., Irvine, CA) and zymolase 20T solution for enzymatic digestion (50 μl) was added to each well. The plate was incubated at 37°C for 1 h and then placed in a 96-well plate luminometer (Luminescencer-JNR AB-2100, ATTO, Tokyo, Japan) and a light emission accelerator solution (AURORA GAL-XE, 50 μl) added to each well using the luminometer pump. The chemiluminescence produced by released -galactosidase in each well was measured.
For estrogenicity, agonist activity was recorded as the REC20 (20% relative effective concentration)—that is, the concentration of the test compound showing 20% of the activity of 10–8 M estradiol. For thyroid hormone activity the REC20 was defined as the concentration of the test compound showing 20% of the activity of 10–6 M T3.
Chinese hamster ovary cell assay. The estrogenic agonist activities of 30 of the 91 hydroxylated PCBs (Table 1) were measured with a reporter gene assay using Chinese hamster ovary cells incorporating the human estrogen receptor hER as has been previously described (Kojima et al., 2004). Again, activity was recorded as the REC20 (the concentration of test compound showing 20% of the activity of 10–10 M estradiol).
Theoretical structure-activity investigation of estrogenicity. We performed structural studies on 26 selected hydroxylated PCBs by using density functional theory to explore the possible factors that determine the estrogenic effects of these compounds. The compounds were selected to include active and inactive compounds (including those with highest activities) and ranging from di-chlorinated- to penta-chlorinated. A popular three-parameter hybrid functional, B3LYP (Becke, 1993; Lee et al., 1988) with 6-31G(d) and 6-311G(d,p) basis sets were utilized for this purpose. All computations were performed with Gaussian 98 and Gaussian 03 programs (Frisch et al., 1998, 2003). The geometries of compounds were optimized first at the B3LYP/6-31G(d) level of theory followed by frequency calculations, which showed that all of the optimized structures were minima on the potential energy surface. Frequency calculations also helped to obtain the polarizabilities of the congeners. Four possible isomeric structures (two with respect to the rotation of the phenyl rings, and two with respect to the OH orientation) of all the selected hydroxylated PCBs were explored. Next, the geometries were optimized by using triple- type basis set, 6-311G(d,p), which also included polarization functions for all atoms. Potential energy curves (PECs) [relative energy vs. torsional angle ()] for all the selected 26 compounds were drawn. For this purpose, structures at various values (from = 40° to = 140° in steps of 20°) were optimized at B3LYP/6-31G(d) level of theory. In addition to this, structures at = 90° were also optimized at the same level of theory to include in the PECs. All geometrical parameters were allowed to optimize. The atomic charges [at the B3LYP/6-311G(d,p)], for the most stable isomers of all the selected compounds, were derived by using electrostatic potential (ESP)-driven charges according to the Merz-Singh-Kollman scheme (Besler et al., 1990; Singh and Kollman, 1984). The results obtained in the theoretical calculations were mainly used to analyze the estrogenicity of the hydroxylated PCBs.
RESULTS AND DISCUSSION
Results of the assays are shown in Tables 1 and 2. Agonist activities were recorded as REC20 values. Although the CHO cell estrogenicity assay was more sensitive than the yeast two-hybrid assay, the rank orders of activities of the compounds tested by both assays were in broad agreement (Fig. 2). Some of the tested compounds showed both estrogenicity and thyroid hormone activity, but it was evident that, in general, different structural features were important for the higher activities for each receptor.
Estrogenicity
Inspection of the results from both assays suggested that an important factor in determining estrogenicity is the position of the phenolic hydroxyl group with highest activities resulting from a para-hydroxyl group. This was apparently assumed by Korach and co-workers (1988) who looked only at para-hydroxylated compounds. Of the 27 compounds with measurable activity in the yeast two-hybrid assay in the current study, 17 (63%) had para-OHs, 7 (26%) had meta-OHs, and 3 (11%) had ortho-OHs. Of the 64 inactive compounds 13 (20.3%) had para-OHs, 17 (26.6%) had meta-OHs, and 34 (53.1%) had ortho-OHs. Although the compounds in the CHO cell assay were selected on the basis of the yeast assay, within the selected group compounds with para-hydroxyls again showed, in general, the highest activity, followed by meta- and ortho-hydroxyls. The importance of the position of the hydroxyl group is clearly illustrated in Figures 3a and b, which show plots of hydroxyl position against log activity for both assays.
Inspection also revealed that estrogenic activity is highly dependent on the number of substituents ortho to the phenolic hydroxyl group. Fourteen of the 27 active compounds in the yeast two-hybrid assay (including the seven most active) had no ortho substituents; 9 had one chlorine ortho to the phenolic hydroxyl, 2 had 1 ortho-phenyl group only, 1 had 1 ortho-phenyl and 1 ortho chlorine, and 1 had 2 ortho chlorines (Fig. 3c). It is notable that in the latter mentioned few cases, where apparently unpromising environments for the phenolic OH are involved, the non-phenolic ring has what appears to be highly favored chlorine substitution patterns (see below) and these apparently counteract, to some extent, the negative influences on the phenol group. Compounds examined in the CHO cell assay, selected on the basis of their activities in the yeast assay, paralleled the results obtained in the yeast assay in regard to the number of substituents ortho- to the phenolic hydroxyl (Fig. 3d) although the relationship was not as strong. The report (Gantchev et al., 1994) that halogenation of the phenolic A ring of estradiol does not diminish, and in certain cases, enhances, the estrogen receptor binding to the ligand was not paralleled by our results for hydroxylated PCBs.
Another factor stressed by McKinney and Waller (1994) was the need for ortho-chlorines to ensure sufficient (steroid-like) rigidity. Of the active 27 compounds in the yeast assay in this study, 26 had at least one ortho-chlorine. Of the 13 inactive compounds with para-OH groups, 4 had no ortho-chlorines. However, additional ortho-chlorines beyond 1 did not appear to confer additional activity.
To further illustrate the importance to estrogenic activity of the number of chlorine atoms in both rings and their substitution patterns, we have listed series of compounds in which one structural feature has been held constant and others varied to reveal differences in activity and the molecular properties apparently responsible for them (Tables 3 and 4). Table 3 lists all compounds in the study with no chlorine substituents in the phenol rings. Going across the columns from left to right (para to meta- to ortho-phenolic hydroxyl groups) reinforces the observation that the position of the hydroxyl group is of fundamental importance to estrogenic activity. However, descending the rows from top to bottom shows that certain patterns of chlorine substitution in the non-phenol ring are also important for activity. This can clearly be seen for para- and meta-phenols, which have the same rank order for activity and this order depends on the chlorine substitution pattern in the non-phenol ring. This is not simply a case of the greater number of ortho-chlorines; there is no consistent trend in that regard—the most active compound in each case has two ortho chlorines, the second and third have one only, and the fourth has two. It seems rather that some particular property governed by the substitution pattern of chlorines in the non-phenolic ring is important in determining activity; compounds with two ortho chlorines and one para chlorine are the most active for both para- and meta-phenols in both assays. There were small differences in the rank order of the compounds for the two assays but they were in general agreement. Of the six compounds having phenol rings lacking chlorine but with ortho-OH groups (Table 3, column 3), only two were active in the yeast assay and one of these was active in the CHO cell assay; both of these had rather low activity. Again the ortho-, ortho-, para- chlorinated non-phenolic ring seemed to confer activity, even on the unpromising ortho-phenol.
The series shown in Table 4, in which the apparently favored ortho-, ortho-, para-chlorinated non-phenol ring is constant, illustrates again the importance of the OH group position, and shows that additional chlorines in the phenolic ring diminish or eliminate activity.
In summary:
Para-hydroxyl is more likely to be active (and have high activity) than meta-; which in turn is more likely to be active than ortho-.
One ortho-chlorine is important but additional ortho-chlorines do not appear to increase activity or the chance of a compound being active.
The number and pattern of chlorine substituents in the non-phenol ring appear to be important for activity. This is not just a question of the more ortho-chlorines the better (conferring steroid-type rigidity), or, more generally, the more the better ("hydrophobic bulk") but a specific pattern seems to favor activity and to be associated with 2 or 3 chlorines in the non-phenol ring. For example, two "lateral" chlorines (only) seem particularly advantageous, i.e., an ortho- and a meta-chlorine on the same side. Also tri substitution in a particular pattern seem to favor activity, namely, 2 ortho- and 1 para-chlorine.
Theoretical Consideration of Estrogenicity
It is known that the halogenated aromatic hydrocarbons (HAHs) with the highest binding affinities for aryl hydrocarbon (Ah) receptors are those with planar structures (Denison et al., 2002). Although there is no such general rule for HAHs to have large binding affinities for estrogen receptors (ER), the existing assumption is that steroid-like rigidity is important rather than any specific conformation. In other words, perpendicular orientation of the phenyl rings is considered to be an important parameter (Korach et al., 1988). But our results were not always supportive of this assumption. For example, the most active hydroxylated PCB in our list had perpendicular phenyl rings, while the phenyl rings in 2',5'-DCB-4-ol (third most active) were around 32° away from the perpendicular orientation at B3LYP/6-311G(d,p) level of theory. The value calculated for 2',3',4',5'-TCB-4-ol, the most active hydroxylated PCB congener in most of the earlier studies (Korach et al., 1988; Waller et al., 1995), is 59.3° at the same level of theory. 2',4',6'-TrCB-4-ol, one of the two most active congeners in most previous studies and the second most active congener in our list had a value of 96.3°, but the closely related 2',4',6'-TrCB-2-ol, which had a perfect perpendicular phenyl rings orientation was the least active congener in our list of 27 active congeners. 2,3',4,4',6-PeCB-3-ol, which had a perpendicular phenyl rings orientation, is inactive. Hence it seems that the perpendicular orientation may not be an essential parameter for a congener to have a large estrogenic activity. Indirectly, this also supports the observation that additional ortho-chlorines do not increase the activity.
Of the 27 active hydroxylated PCBs in our experimental study, 26 had at least one ortho-chlorine. Results showed that additional ortho-chlorines beyond 1 did not seem to confer additional activity. These results revealed that one ortho-chlorine was essential for a congener to acquire estrogenic activity. Our calculations show that a single ortho- substitution takes the phenyl rings around 60° away from each other (from co-planar orientation). Additional chlorine substitutions at ortho- positions, obviously, increase the repulsive interaction between overlapping ortho-chlorine atoms and increase the torsional angle further. With three ortho-chlorine atoms, the phenyl rings become perpendicular to each other. The values calculated for hydroxylated PCBs lacking ortho-chlorine atoms are close to 40°. So it is concluded that the hydroxylated PCBs are not active in a co-planar (about 40°) conformation (and do not show estrogenic activity), but rather with a 60–90° twist between the phenyl rings.
Earlier studies demonstrated that chlorine substitution in the phenolic ring does not significantly affect the estrogenic activity of hydroxylated PCBs, whereas 2,4,6-trichloro- and 2,3,4,6-tetrachloro-substitution in the non-phenolic ring gave compounds with the highest estrogenic activity (Korach et al., 1988). It has been further reported (Gantchev et al., 1994) that halogenation of the phenolic ring, in certain cases, enhances the estrogenic activity. We found that estrogenic activity is highly dependent on the number of substituents ortho- to the phenolic hydroxyl group. It is worthwhile to analyze these observations through a theoretical approach. Consider chlorine substitutions at the 2 and 3 positions in 2',4',6'-TrCB-4-ol. Substitution at the 3-position gives rise to 2',3,4',6'-TCB-4-ol which might exhibit intramolecular hydrogen bonding between the hydrogen of the hydroxyl group and the nearby chlorine atom; however, such hydrogen bonding is not possible in the case of chlorine substitution at the 2-position, (giving 2,2',4',6'-TCB-4-ol). This is evidenced by the relative energies obtained for the possible conformers in both congeners. The conformer of 2',3,4',5'-TCB-4-ol with the OH orientation favorable for such a probable hydrogen bonding (the ClH non-bonded bond length is 2.418 ) is energetically more stable by 12.65 kJ/mol [at B3LYP/6–311G(d,p) level of theory] than the conformer with the OH orientation which is not favorable to hydrogen bonding. But hydrogen bonding is not possible in 2,2',4',6'-TCB-4-ol and so there is no energy difference between the various conformers of this congener. The same is true for 2',5'-DCB-4-ol; while there is no real energy difference between the possible conformers of 2,2',5'-TrCB-4-ol, the conformer with the probable hydrogen bonding orientation in 2',3,5'-TrCB-4-ol is energetically favorable by 12.85 kJ/mol when compared to the conformer with unfavorable hydrogen bonding orientation. The non-bonded ClH bond length in the hydrogen bonded 2',3,5'-TrCB-4-ol conformer is 2.416 . So it is clear that an intramolecular hydrogen bonding of about 12 kJ/mol forms, where it can, between the hydrogen of OH group and the neighboring chlorine atom in the hydroxylated PCBs. That means chlorine substitution ortho- to the hydroxyl group favors such hydrogen bonding. This assumption is also evidenced from the results obtained for other hydroxylated PCBs. It is worth analyzing whether this intramolecular hydrogen bonding has any influence on the estrogenic activities of the congeners. In both the above cases, the estrogenic activities of the 3-substituted congeners (2',3,4',5'-TCB-4-ol and 2',3,5',TrCB-4-ol) are lower than those of 2-substituted congener (2,2',4',6'-TCB-4-ol and 2,2',5'-TrCB-4-ol). So it seems that the intramolecular hydrogen bonding decreases the activity. This can be assumed from analyzing the experimental results obtained for other hydroxylated PCBs too; of 2,2',3',4',5'-PeCB-4-ol and 2',3,3',4',5'-PeCB-4-ol, the former is active while the latter is inactive; 2,3',4'-TrCB-4-ol is active but its 3-chlorinated congener, 3,3',4'-TrCB-4-ol is inactive; 2-MCB-4-ol is more active than 3-MCB-4-ol.
It is evident that a para-substituted hydroxyl group is a property of estrogenically active congeners. 2',4',6'-TrCB-4-ol, 2',3',4',5'-TCB-4-ol, and 2',5'-DCB-4-ol are examples of active hydroxyl-PCBs with para-hydroxyl groups. However, the consequences for estrogenicity of hydroxyl group substitution at the meta- or ortho-positions is, apart from indications from the limited study of Schultz (2002), largely unknown. We used our calculations to investigate this subject. Analyzing the results revealed that the hydrogen of the hydroxyl group weakly interacts with the neighboring phenyl group (either with the chlorine atom or with the -electron cloud) if the OH group is substituted at ortho-position. Thorough investigation of the calculated relative energies showed that this weak interaction is energetically around 3 to 5 kJ/mol strong. However, the OH group substituted at the 3-position is free from such intramolecular interactions. It should be interesting to see whether this weak interaction has any role in the estrogenic activity of the congeners. First we considered 2',4',6'- and 2',5'-chlorinated congeners for this purpose. Our calculations indicated that congeners with the hydroxyl group substituted at the ortho-position (2',4',6'-TrCB-2-ol and 2',5'-DCB-2-ol) have such intramolecular interactions (between OH group and the neighboring phenyl group) of about 3 to 5 kJ/mol. Our experimental studies showed that the compounds with the hydroxyl group substituted at the meta-position in both the cases, 2',4',6'-TrCB-3-ol and 2',5'-DCB-3-ol, are active while their ortho-substituted (OH group) counterparts, 2',4',6'-TrCB-2-ol and 2',5'-DCB-2-ol, are either very weakly active or inactive. This gives the impression that these weak intramolecular interactions play a role in the estrogenic activities of hydroxylated PCBs. This can be examined in other compounds too. For example, in the case of the 2',3',4',5' substituted hydroxylated PCBs, 2',3',4',5'-TCB-3-ol is active whereas 2',3',4',5'-TCB-2-ol is inactive. The latter might have OH-phenyl group intramolecular interaction. Hence it seems that the existence of such intramolecular interactions decreases the estrogenic activities of hydroxylated PCBs. It is evident from the results that even among the hydroxylated PCBs with a chlorine substitution in their hydroxylated phenyl rings, compounds with the OH group in the para- position are more active than those with the OH group substituted at ortho- or meta- positions: 2,2',5'-TrCB-4-ol is more active than 2,2',5'-TrCB-3-ol; 2',3,3'-TrCB-4-ol is more active than 2',3,3'-TrCB-2-ol; 2',3,4'-TrCB-4-ol is active while 2',3,4'-TrCB-2-ol is not.
The foregoing facts clearly reveal that the hydroxyl group should be in a position where it can be highly flexible and is available to make strong intermolecular interactions with the estrogen receptors in the cell. Obviously, this should be possible at the para-position but the OH group should not have any chlorine substitution ortho- to it. In other words, the hydroxyl group should be completely free from any intramolecular interactions. This would suggest, interestingly, that such a free hydroxyl group could help to make a strong hydrogen bonding between the hydroxylated PCBs and the estrogen receptors (particularly with amino acids). It can be assumed that if the hydroxyl group is involved in an intramolecular interaction, it cannot make a strong intermolecular interaction with the ER. It has been hypothesized (Gantchev et al., 1994) that hydroxyl groups at position 2 in estradiol (E2) and in E2 further substituted at position 3 may share, via hydrogen bonding, a common acceptor/donor site in the receptor cavity. Hence our results indirectly support the hypothesis that intermolecular hydrogen bonding between the hydroxylated PCBs and the estrogen receptors plays an important role in the estrogenic activity of these compounds.
Previous studies have concluded that conformational restriction about the twist bond of the overall structure of a hydroxylated PCB appeared to play an important role in estrogen receptor binding affinity. In other words, as mentioned earlier, steroid-like (more rigid) structure is believed to be an essential parameter for the active congeners. In these circumstances, PEC studies on the hydroxylated PCBs could be very useful for a deeper understanding of this subject. We investigated the PECs of all 26 selected hydroxylated PCBs, and it is clear from our results that most of these compounds are energetically flexible for C–C interring rotation (Fig. 4). The energy difference due to the rotation of the phenyl rings from = 60° to = 120° is very small. This energy change in 2',4',6'-TrCB-4-ol, the second most active congener, is around 1 kJ/mol. It is around 5 and 2 kJ/mol, respectively, for the most active 2,2',4',6'-TCB-4-ol and the fifth most active 2,2',5'-TrCB-4-ol. These three active congeners have no (or negligible) torsional barrier at = 90°. The same, very small change in energy due to the rotation of the phenyl rings from = 60° to = 120° and no barrier at = 90°, is true for two other active congeners, 2',3,4',6'-TCB-4-ol and 2',4',6'-TrCB-3-ol. 2',5'-DCB-4-ol and 2',3',4',5'-TCB-4-ol, the third and fourth most active congeners, respectively, had barriers of about 4 kJ/mol at = 90°. However, it should be noted that the energy change due to the rotation of phenyl rings from = 40° to = 140° in these two hydroxylated PCBs is just 4 kJ/mol. Again, this is the same for 2',3,5'-TrCB-4-ol and 2',5'-DCB-3-ol, both of which are active. Even in 2',4',6'-TrCB-4-ol, 2',3,4',6'-TCB-4-ol, and 2',4',6'-TrCB-3-ol, the energy changes due to the rotation of phenyl rings from = 40° to = 140° are around 10 kJ/mol. All these results reveal that hydroxylated PCBs are not conformationally very restrictive about the twist bond. The structure is not rigid for rotation of the phenyl bonds at least from = 60° to = 120°. Furthermore, by considering the fact that the free energy for the binding of 17 -estradiol to the estrogen receptor is around 50 kJ/mol (Wiese and Brooks, 1994), even a much weaker binding of hydroxylated PCBs to the ER could provide the energy necessary for hydroxylated PCBs to freely rotate (or to change its conformation!) from = 60° to = 120°. This study also helps to make two other conclusions: (1) it is difficult to predict whether or not a congener will be estrogenically active from its PEC; and (2) torsional barriers at co-planar or near co-planar structures are very high and hence for these hydroxylated PCBs, conformationally moving towards co-planar structures is energetically very difficult.
Studies in the past illustrated that the electron donating properties of the hydroxyl group (phenolic oxygen) and the analog in the hydroxylated PCBs of the aromatic component of the estradiol A ring were found to be correlated with higher ER binding affinity (Bradbury et al., 1996). Based on experimental and theoretical findings (Palomino et al., 1994; VanderKuur et al., 1993a,b; Wiese and Brooks, 1994) for the involvement of a negative isopotential surrounding the ER binding of the E2 A ring and the assumption that there are cationic sites (amino acid residues) inside the ER (Amouri et al., 1992; Pakdel and Katzenellenbogen, 1992), there is a hypothesis that hydroxylated PCBs with greater binding affinity should have higher electron donating abilities associated with their corresponding phenolic ring. Such chemicals have larger negative charges on the phenolic oxygen and the aromatic carbons in the A ring. Atomic charges obtained in this study (Table 5) are mixed and it is not easy to draw any concrete conclusions. For example, in 2',4',6'-TrCB-4-ol the sum of the atomic charges on the carbon atoms of the hydroxylated ring (qC-ring) is smaller (in negative scale) than those in 2',4',6'-TrCB-3-ol and in 2',4',6'-TrCB-2-ol. This trend continued in 2',5'-chlorine substituted hydroxylated-PCBs too. In the above two groups, compounds with OH groups at ortho- positions, the least active congeners in their respective groups, had the larger qC-ring values. Again, the qC-ring value obtained for 2,2',4',6'-TCB-4-ol is smaller than that obtained for 2',3,4',6'-TCB-4-ol. The value obtained for 2,2',5'-TrCB-4-ol is smaller than that obtained for 2',3,5'-TrCB-4-ol. The total atomic charges on the hydroxylated phenyl ring (qring) in 2',4',6'-TrCB-4-ol is smaller than those in 2',4',6'-TrCB-3-ol and in 2',4',6'-TrCB-2-ol. The qring value obtained for 2',5'-DCB-4-ol is smaller than that obtained for 2',5'-DCB-2-ol. In fact the qring value calculated for 2',5'-DCB-4-ol is in the positive scale. The present study shows atomic charge on oxygen atom (qO) is in the order 2',4',6'-TrCB-3-ol > 2',4',6'-TrCB-4-ol > 2',4',6'-TrCB-2-ol, whereas the estrogenic activity order is 2',4',6'-TrCB-4-ol > 2',4',6'-TrCB-3-ol > 2',4',6'-TrCB-2-ol. The same was found in the 2',5'-substituted group too. Although these results show a complicated picture, they also reveal that, in general, the active congeners have large qO and qC-ring values. Hence estrogenic activity appears to be related to the presence of an optimum range of qO and qC-ring values.
The most polarizable conformers are assumed to be closely associated with ER binding affinity (Bradbury et al., 1996) and so polarizability might be an important parameter for estrogenic activity. So we calculated polarizability tensors (XX, YY, and ZZ) along the principle axes to know whether there is any direct relationship between polarizability and activity. The average polarizability, ave, was calculated as ave = (1/3)(XX + YY + ZZ). It was true that in some cases that active congeners had large polarizabilities. The ave values calculated for 2',4',6'-TrCB-4-ol, 2',4',6'-TrCB-3-ol, and 2',4',6'-TrCB-2-ol were in the same order as their estrogenic activities. This was true for the 2',5'-DCB-4-ol, 2',5'-DCB-3-ol, and 2',5'-DCB-2-ol series too. Analysis of the results obtained for the above two sets of congeners revealed that hydroxyl group substitution at the para-position allows the congeners to become more polarizable in the x-axis (molecular axis) whereas OH substitution at the ortho-position makes the congeners less polarizable in the same direction. However, making a generalized statement such as the active congeners have larger polarizabilities than their less active or inactive counterparts is not possible.
In summary:
Perpendicular orientation of the two rings (i.e., rigidity) is not necessary for activity.
Intramolecular hydrogen bonding of the phenolic hydroxyl to adjacent chlorines or to the -electron cloud of the non-phenol ring decreased activity.
Prediction of the estrogenic activity of a congener from its obtained potential energy curve was difficult.
In general, estrogenically active congeners had large values for the sum of the atomic charges on the carbon atoms of the phenol ring, and on the oxygen atom.
No general statement about polarizability and estrogenic activity was possible although hydroxyl substitution at the para-position allowed the compounds to become more polarizable in the x-axis (molecular axis), whereas hydroxylation at the ortho-position made the congeners less polarizable in the same direction.
Thyroid Activity
The natural thyroid hormones T3 and T4 (Fig. 1b) obviously exhibit the necessary structural properties for activity. Iodine must be substituted at position 3 and 5 in the inner ring to maintain the outer ring perpendicular to the inner ring, i.e., to prevent rotation of the rings relative to each other. A spacer atom (which can be O as in T3 and T4, or S or C) is required between the two rings and should form a 120° angle. The carboxyl and the phenolic hydroxyl groups must be present to maximize receptor binding. Lipophilic groups (iodine in the case of T3 and T4) must be present at the 3' position of the outer ring to optimize binding (Norman and Litwack, 1997).
The activities of the natural thyroid hormones and their analogues measured by the yeast two-hybrid assay used in this study do not exactly parallel the activities of these compounds at the organism level. For example, T3 is 3 to 8x as active as T4 in the human body (Norman and Litwack, 1997) but in our assay T4 is almost as active as T3, and reverse T3, which is virtually inactive in the human organism, shows appreciable activity. Activities of these and related synthetic hormones are shown in Figure 5. This characteristic of the assay must be considered when seeking to make correlations between structure and activity.
Inspection of the results (Table 2) suggested that although the most active compounds 2',4,5',6-TCB-2-ol and 3',5,5',6-TCB-2-ol had their phenolic hydroxyl groups in the ortho- position, the actual position of the phenolic hydroxyl group might be of rather less importance than the number of substituents (chlorine or phenyl ring) ortho- to the phenolic hydroxyl. It appears that at least one substituent ortho- to the hydroxyl is necesary for activity, and, perhaps, if this is the other phenyl ring the chances of activity are enhanced. Thus it might be inferred that the non-phenolic ring with its substituents is serving as the necessary hydrophobic bulk ortho- to the phenolic hydroxyl, i.e., playing the role of an iodine atom in the natural hormones. Although there is a slight preponderance of ortho-phenols, activity is rather evenly distributed between ortho-, meta-, and para- phenols—of 24 active compounds 10 (42%) are ortho-phenols, 7 (29%) are meta-, and 7 (29%) are para-. With regard to substituents ortho- to the phenolic hydroxyl, of the 24 active compounds, 14 (58%) have 1 ortho-substituent, 8 (33%) have 2, and only 2 (8%) have none. The ortho- substituent can of course be the other phenyl ring as well as a chlorine atom.
It also appeared that the pattern and degree of substitution of the non-phenolic ring might be important. Of the 24 active compounds 19 (79%) had 2 chlorines in the non-phenolic ring, 4 had 1, and 1 had 3. None had 4 or 0. Of the 67 non-active compounds 36 (53%) have 2 Cls, 7 have 3, 14 have 4, 4 have 0, and 6 have 1. Numbers of compounds with 4 chlorines in the non-phenol ring are particularly striking—no active compounds have 4 chlorines but 14 (21%) of the non-active compounds do. Furthermore, all active compounds except 2 (22 of 24, 92%) have at least one meta-chlorine in the non-phenolic ring.
There are no active compounds without at least one ortho-chlorine (even for ortho-phenols). So coplanarity is, as expected, not favorable for thyroid hormone activity. Seventeen of 91 compounds (19%) have no ortho-chlorines and all are inactive. Of these 6 do have ortho-OH groups but that does not confer activity in the absence of at least one ortho-chlorine.
In summary:
Thyroid hormone activity appears not to depend strongly on the position of the phenolic hydroxyl although ortho-hydroxyl occurs in the most active compounds suggesting that the other (non-phenol) phenyl ring might serve as the required ortho- hydrophobic bulk (the role of the iodine ortho- to the phenolic hydroxyl in T3).
Activity is usually associated with at least one ortho-chlorine irrespective of whether the phenolic hydroxy is in the ortho- position or not.
Activity is usually associated with 2 chlorines in the phenolic ring and, importantly, two chlorines in the non-phenolic ring.
Activity is usually associated with one or two chlorines ortho to the phenolic hydroxyl.
Hydroxylated PCBs in the Environment
Studies on the identification and quantification of hydroxylated PCBs in environmental samples (Asplund et al., 1999; Campbell et al., 2003; Hoekstra et al., 2003; Klasson-Wehler et al., 1998; Li et al., 2003; Olsson et al., 2000) and in humans (Guvenius et al., 2002; Sandau et al., 2000) have revealed, among a total of some 15 or so compounds, the presence of four of the compounds included in our list of synthetic hydroxylated PCBs. Interestingly, the most estrogenically active compound (2,2',4',6'-TCB-4-ol) in our assays, was identified and quantified in lake trout from Lake Ontario and Lake Superior (Campbell et al., 2003). The results of our estrogenicity assays for 2',4',6'-TCB-4-ol (second most active) and 2',3',4',5'-TCB-4-ol (fourth most active) were consistent with previous experimental work on the endocrine disrupting properties of these compounds in fish (Carlson and Williams, 2001; Gerpe et al., 2000).
REFERENCES
Amouri, H. E., Vessiers, A., Vichard, D., Top, S., Grusselle, M., and Jaouen, G. (1992). Syntheses and affinities of novel organometallic-labeled estradiol derivatives: A structure-affinity relationship. J. Med. Chem. 35, 3130–3135.
Asplund, L., Athanasiadou, M., Sjdin, A., Bergman, ., and Brjeson, H. (1999). Organohalogen substances in muscle, egg and blood from healthy Baltic salmon (Salmo salar) with the M74 syndrome. Ambio 28, 67–76.
Ballschmiter, K., and Zell, M. (1980). Analysis of polychlorinated biphenyls (PCB) by glass capillary gas chromatography. Fresenius J. Anal. Chem. 302, 20–31.
Becke, A. D. (1993). Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 98, 5648–5652.
Bergeron, J. M., Crews, D., and McLachlan, J. A. (1994). PCBs as environmental estrogens: Turtle sex determination as a biomarker of environmental contamination. Environ. Health Perspect. 102, 780–781.
Bergman, ., Klasson Wehler, E., and Kuroki, H. (1994). Selective retention of hydroxylated PCB metabolites in blood. Environ. Health Perspect. 102, 464–469.
Besler, B. H., Merz, K. M., Jr., and Kollman, P. A. (1990). Atomic charges derived from semiempirical methods. J. Comput. Chem. 11, 431–439.
Bradbury, S. P., Mekenyan, O. G., and Ankley, G. T. (1996). Quantitative structure-activity relationships for polychlorinated hydroxybiphenyl estrogen binding affinity: An assessment of conformer flexibility. Environ. Toxicol. Chem. 15, 1945–1954.
Campbell, L. M., Muir, D. C. G., Whittle, D. M., Backus, S., Norstrom, R. J., and Fisk, A. T. (2003). Hydroxylated PCBs and other chlorinated phenolic compounds in lake trout (Salvelinus namaycush) blood plasma from the Great Lakes region. Environ. Sci. Technol. 37, 1720–1725.
Carlson, D. B., and Williams, D. E. (2001). 4-Hydroxy-2',4',6'-trichorobiphenyl and 4-hydroxy-2',3',4',5'-tetrachlorobiphenyl are estrogenic in rainbow trout. Environ. Toxicol. Chem. 20, 351–358.
Connor, K., Ramamoorthy, K., Moore, M., Mustain, M., Chen, I., Safe, S., Zacharewski, T., Gillesby, B., Joyeux, A., and Balaguer, P. (1997). Hydroxylated polychlorinated biphenyls (PCBs) as estrogens and antiestrogens: Structure-activity relationships. Toxicol. Appl. Pharmacol. 145, 111–123.
Cramer, R. D., III, Patterson, D. E., and Bumce, J. D. (1988). Comparative molecular field analysis (CoMFA). 1. Effect of shape on binding of steroids to carrier proteins. J. Am. Chem. Soc. 110, 5959–5967.
Denison, M. S., Pandini, A., Nagy, S. R., Baldwin, E. P., and Bonati, L. (2002). Ligand binding and activation of the Ah receptor. Chem. Biol. Interact. 141, 3–24.
Fielden, M. R., Chen, I., Chittim, B., Safe, S. H., and Zacharewski, T. R. (1997). Examination of the estrogenicity of 2,4,6,2'6'-pentachlorobiphenyl (PCB 104), its hydroxylated metabolite 2,4,6,2',6'-pentachloro-4-biphenylol (HO-PCB 104), and a further chlorinated derivative, 2,4,6,2',4',6'-hexachlorobiphenyl (PCB 155). Environ. Health Perspect. 105, 1238–1248.
Frisch, M. J. et al. (1998). Gaussian 98, Revision A.7 Gaussian, Inc., Pittsburgh, PA.
Frisch, M. J. et al. (2003). Gaussian 03, Revision B.01; Gaussian, Inc., Pittsburgh, PA.
Gantchev, T. V., Ali, H., and van Lier, J. E. (1994). Quantitative structure-activity relationships/comparative molecular field analysis (QSAR/CoMFA) for receptor-binding properties of halogenated estradiol derivatives. J. Med. Chem. 37, 4164–4176.
Gerpe, M., Kling, P., Berg, A. H., and Olsson, P.-E. (2000). Arctic char (Salvelinus alpinus) metallothionein: cDNA sequence, expression, and tissue-specific inhibition of cadmium-mediated metallothionein induction by 17-estradiol, 4-OH-PCB30, and PCB104. Environ. Toxicol. Chem. 19, 638–645.
Guvenius, D. M., Hassanzadeh, P., Bergman, ., and Norén, K. (2002). Metabolites of polychlorinated biphenyls in human liver and adipose tissue. Environ. Toxicol. Chem. 21, 2264–2269.
Hoekstra, P. F., Letcher, R. J., O'Hara, T. M., Backus, S. M., Solomon, K. R., and Muir, D. C. (2003). Hydroxylated and methylsulfone-containing metabolites of polychlorinated biphenyls in the plasma and blubber of bowhead whales (Balaena mysticetus). Environ. Toxicol. Chem. 22, 2650–2658.
Hovander, L., Malmberg, T., Athanasiadou, M., Athanassiadis, I., Rahm, S., Bergman, ., and Klasson Wehler, E. (2002). Identification of hydroxylated PCB metabolites and other phenolic halogenated pollutants in human blood plasma. Arch. Environ. Contam. Toxicol. 42, 105–117.
Kester, M. H. A., Bulduk, S., Tibboel, D., Meinl, W., Glatt, H., Falany, C. N., Coughtrie, M. W. H., Bergman, ., Safe, S. H., Kuiper, G. G. J. M., Schuur, A. G., Brouwer, A., and Visser, T. J. (2000). Potent inhibition of estrogen sulfotransferase by hydroxylated PCB metabolites: A novel pathway explaining the estrogenic activity of PCBs. Endocrinology 141, 1897–1900.
Klasson-Wehler, E., Bergman, ., Athanasiadou, M., Ludwig, J. P., Auman, H. J., Kannan, K., Van Den Berg, M., Murk, A. J., Feyk, L. A., and Giesy, J. P. (1998). Hydroxylated and methylsulfonyl polychlorinated biphenyl metabolites in albatrosses from Midway Atoll, North Pacific Ocean. Environ. Toxicol. Chem. 17, 1620–1625.
Kojima, H., Katsura, E., Takeuchi, S., Niiyama, K., and Kobayashi, K. (2004). Screening for estrogen and androgen receptor activities in 200 pesticides by in vitro reporter gene assays using Chinese hamster ovary cells. Environ. Health Perspect. 112, 524–531.
Korach, K. S., Sarver, P., Chae, K., McLachlan, J. A., and McKinney, J. D. (1988). Estrogen receptor-binding activity of polychlorinated hydroxybiphenyls: Conformationally restricted structural probes. Mol. Pharmacol. 33, 120–126.
Kuratsune, M., Yoshimura, H., Hori, Y., Okumura, M., and Masuda, Y. (Eds.) (1996). Yusho. A Human Disaster Caused by PCBs and Related Compounds. Kyushu University Press, Fukuoka, Japan.
Lee, C., Yang, W., and Parr, R. G. (1988). Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density. Phys. Rev. B 37, 785–789.
Letcher, R. J., Klasson Wehler, E., and Bergman, . (2000). Methyl sulfone and hydroxylated metabolites of polychlorinated biphenyls. In New Types of Persistent Halogenated Compounds. (J. Paasivirta, Ed.), pp. 315–359. Springer-Verlag, Berlin.
Li, H., Drouillard, K. G., Bennett, E., Haffner, G. D., and Letcher, R. J. (2003). Plasma-associated halogenated phenolic contaminants in benthic and pelagic fish species from the Detroit River. Environ. Sci. Technol. 37, 832–839.
Maervoet, J., Covaci, A., Schepens, P., Sandau, C. D., and Letcher, R. J. (2004). A reassessment of the nomenclature of polychlorinated biphenyl (PCB) metabolites. Environ. Health Perspect. 112, 291–294.
McKinney, J. D., and Waller, C. L. (1994). Polychlorinated biphenyls as hormonally active structural analogues. Environ. Health. Perspect. 102, 290–297.
Moore, M., Mustain, M., Daniel, K., Chen, I., Safe, S., Zacharewski, T., Gillesby, B., Joyeux, A., and Balaguer, P. (1997). Antiestrogenic activity of hydroxylated polychlorinated biphenyl congeners identified in human serum. Toxicol. Appl. Pharmacol. 142, 160–168.
Narasimhan, T. R., Kim, H. L., and Safe, S. (1991). Effects of hydroxylated polychlorinated biphenyls on mouse liver mitochondrial oxidative phosphorylation. J. Biochem. Toxicol. 6, 229–236.
Nishikawa, J., Saito, K., Goto, J., Dakeyama, F., Matsuo, M., and Nishihara, T. (1999). New screening methods for chemicals with hormonal activities using interaction of nuclear hormone receptor with coactivator. Toxicol. Appl. Pharmacol. 154, 76–83.
Norman, A. W., and Litwack, G. (1997). Hormones, 2nd ed., Academic Press, San Diego.
Okumura, T., and Shibata, J. (1975). Investigation on the behavior of PCBs in the environment. Research Report of the Environmental Pollution Control Center, Osaka Prefectural Government, Japan (in Japanese).
Olsson, A., Ceder, K., Bergman, ., and Helander, B. (2000). Nestling blood of the white-tailed sea eagle (Haliaeetus albicilla) as an indicator of territorial exposure to organohalogen compounds-an evaluation. Environ. Sci. Technol. 34, 2733–2740.
Pakdel, F., and Katzenellenbogen, B. S. (1992). Human estrogen receptor mutants with altered estrogen and antiestrogen ligand discrimination. J. Biol. Chem. 267, 3429–3437.
Palomino, E., Heeg, M. J., Horwitz, J. P., Polin, L., and Brooks, S. C. (1994). Skeletal conformations and receptor binding of some 9,11-modified estradiols. J. Steroid Biochem. Mol. Biol. 50, 75–84.
Sandau, C. D., Ayotte, P., Dewailly, E., Duffe, J., and Norstrom, R. J. (2000). Analysis of hydroxylated metabolites of PCBs (OH-PCBs) and other chlorinated phenolic compounds in whole blood from Canadian Inuit. Environ. Health Perspect. 108, 611–616.
Schultz, T. W. (2002). Estrogenicity of biphenylols: Activity in the yeast gene activation assay. Bull. Environ. Contam. Toxicol. 68, 332–338.
Schuur, A. G., Brouwer, A., Bergman, ., Coughtrie, M. W. H., and Visser, T. J. (1998a). Inhibition of thyroid hormone sulfation by hydroxylated metabolites of polychlorinated biphenyls. Chem. Biol. Interact. 109, 293–297.
Schuur, A. G., Legger, F. F., van Meeteren, M. E., Moonen, M. J. H., van Leeuwen-Bol, I., Bergman, ., Visser, T. J., and Brouwer, A. (1998b). In vitro inhibition of thyroid hormone sulfation by hydroxylated metabolites of halogenated aromatic hydrocarbons. Chem. Res. Toxicol. 11, 1075–1081.
Schuur, A. G., van Leeuwen-Bol, I., Jong, W. M. C., Bergman, ., Coughtrie, M. W. H., Brouwer, A., and Visser, T. J. (1998c) In vitro inhibition of thyroid hormone sulfation by polychlorobiphenylols: Isozyme specificity and inhibition kinetics. Toxicol. Sci. 45, 188–194.
Shiraishi, F., Okumura, T., Nomachi, M., Serizawa, S., Nishikawa, J., Edmonds, J. S., Shiraishi. H., and Morita, M. (2003). Estrogenic and thyroid hormone activity of a series of hydroxy-polychlorinated biphenyls. Chemosphere 52, 33–42.
Shiraishi, F., Shiraishi, H., Nishikawa, J., Nishihara, T., and Morita, M. (2000). Development of a simple operational estrogenicity assay system using the yeast two-hybrid system. J. Environ. Chem. 10, 57–64.
Singh, U. C., and Kollman, P. A. (1984). An approach to computing electrostatic charges for molecules. J. Comput. Chem. 5, 129–145.
Sinjari, T., and Darnerud, P. O. (1998). Hydroxylated polychlorinated biphenyls: Placental transfer and erffects on thyroxine in the foetal mouse. Xenobiotica 28, 21–30.
Vakharia, D. D., and Gierthy, J. F. (2000). Use of a combined human liver microsome-estrogen receptor binding assay to assess potential estrogen modulating activity of PCB metabolites. Toxicol. Lett. 114, 55–65.
van den Hurk, P., Kubiczak, G. A., Lehmler, H.-J., and James, M. O. (2002). Hydroxylated polychlorinated biphenyls as inhibitors of the sulfation and glucuronidation of 3-hydroxy-benzo[a]pyrene. Environ. Health Perspect. 110, 343–348.
VanderKuur, J. A., Hafner, M. S., Christman, J. K., and Brooks, S. C. (1993a). Effects of estradiol-17 beta analogues on activation of estrogen response element regulated chloramphenicol acetyltransferase expression. Biochemistry, 32, 7016–7021.
VanderKuur, J. A., Wiese, T., and Brooks, S. C. (1993b). Influence of estrogen structure on nuclear binding and progesterone receptor induction by the receptor complex. Biochemistry 32, 7002–7008.
Waller, C. L., Minor, D. L., and McKinney, J. D. (1995). Using three-dimensional quantitative structure-activity relationships to examine estrogen receptor binding affinities of polychlorinated hydroxybiphenyls. Environ. Health. Perspect. 103, 702–707.
Wiese, T. E., and Brooks, S. C. (1994). Molecular modeling of steroidal estrogens: Novel conformations and their role in biological activity. J. Steroid Biochem. Mol. Biol. 50, 61–73.(Sundaram Arulmozhiraja, F)