Science in China Series B: Chemistry © 2009
www.scichina.com chem.scichina.com www.springerlink.com
SCIENCE IN CHINA PRESS
Springer
QSAR study on the non-monotonic dose-response curve of PCBs in chicken embryo hepatocyte bioassay MU YunSong, ZHANG AiQian†, GAO ChangAn, PENG SuFen & WANG LianSheng State Key Laboratory of Pollution Control and Resource Reuse, School of Environment, Nanjing University, Nanjing 210093, China
Endocrine disrupting chemicals (EDCs) in the natural environment exhibit a unique non-monotonic dose-response curve and it is impossible to select one simple index to characterize the bilogogical activity of these compounds. Quantitative structure-activity relationship (QSAR) study on non-monotonic dose-response curve has become a real challenge presently. In order to explore the possible mechanism for the non-monotonic dose-response curve of polychlorinated biphenyls congeners (PCBs) in chicken embryo hepatocyte bioassay, AM1 method of ChemOffice was adopted to calculate necessary structure descriptors for PCBs, while the interactions between PCBs and simulated AhR ligand binding domain (LBD) were analyzed by using FlexX in SYBYL7.0. Different binding modes for PCBs have been distinguished not only from aligned conformation but also from free binding energy. Some QSAR models were established separately for both low and high doses ranges, revealing that receptor binding may predominate in the interference of the physiological function of cytochrome P4501A-P4501A in the low doses range. But with the higher doses range, the EROD suppression might be related to acute toxicity owing to molecular polarity or distribution of charges and consequently damage structure and function of chicken embryo hepatocyte. polychlorinated biphenyls congeners (PCBs), non-monotonic dose-response curve, quantitative structure-activity relationship (QSAR)
1
Introduction
In recent years, it has been found that certain endocrine disruptors tested at low or high doses evoke different responses in the toxicological experiments. Saal et al.[1] suggested that low doses of diethylstilbestrol (DES) would promote the mouse’s prostate hyperplasia while high doses would inhibit prostate from growing. Other chemicals with non-monotonic dose-response relationship have also been confirmed in endocrine disrupting toxicological tests, such as bisphenol A[2], methyl chloride and DDT[3]. Therefore, a single linear model for toxicity evaluation is insufficient within the whole range of concentrations[4]. Besides, it would be a great challenge to the threshold model in the traditional toxicology since the concentration level at low doses could be lower than the conventional no-observable-adverse-effect level (NOAEL) in breeding or growth testing protocols.
Quantitative structure-activity relationship (QSAR) study is an effective tool for toxicological mechanism research, which has been widely used in the evaluation and estimation of various toxic effects. However, the scientific assumption that the same mode of action exhibits within the entire doses range is still doubtful. Thus, it is urgent to develop the simulation method for solving this problem. Polychlorinated biphenyls congeners (PCBs) have attracted more and more people’s attention as global environmental pollutants and representative endocrine disruptors. They can cause adverse biological effects by interacting with aryl hydrocarbon receptor (AhR), such Received August 10, 2008; accepted October 3, 2008 doi: 10.1007/s11426-009-0023-1 † Corresponding author (email:
[email protected]) Supported by the National Natural Science Foundation of China (Grant Nos. 20777035 & 20737001) and 863 Advanced Research Project (Grant Nos. 2007AA06Z416, 2006AA06Z424 & 2007AA06A405)
Sci China Ser B-Chem | May 2009 | vol. 52 | no. 5 | 662-669
as thymus atrophy, immune toxicity, acute death and cytochrome P4501A1 expression, among others. The structure-activity relationship (SAR) research on PCBs started a decade ago and has been a hot issue up to now[5]. However, these investigations have been limited because of the novel dose-response curve of PCBs. Welshons et al. [6] , who were interested in the biological effects of environmental estrogens on MCF-7 cell proliferation, found that some EDCs at low doses played an indispensable role in the increased biological responses with the dose rising through the receptormediated process, while the chemicals failed to induce physiological alternations by the interaction with receptors when the concentrations were much higher than their toxicity level. Such phenomena about the nonmonotonic dose-response relationship are derived from different molecular mechanisms between high and low doses ranges. The possible reason for non-monotonic dose-response relationship of PCBs-induced CEH-EROD activities ought to be that PCBs with specific stereo conformation and electrostatic field at low doses induced EROD activity by the binding interaction with AhR, while receptor-binding saturation and acute toxicity occurred with the increased doses of the compounds. In this paper, the binding energy between PCBs and AhR is calculated using molecular docking simulation to evaluate the inhibitory potency of different PCBs. QSAR analysis on the CEH-EROD activity at different dose ranges is also emphasized to explore the reasonable biochemical mechanism.
2
Theoretical calculations
2.1 Modeling dataset In previous studies, Kennedy et al.[7] had reported PCBs-induced EROD activities in chicken embryo hepatocyte primary cultures. Moreover, other datasets about the binding affinity between PCBs and AhR were － derived from Safe’s studies[8 11]. Eleven PCB compounds as the potential EDCs were selected for this study, the IUPAC PCB numbers of which were 77, 78, 79, 81, 105, 118, 126, 138, 156, 157 and 169. All the dose-EROD activity curves within the same concentration level of 10−3－104 nmol/L were reported recently. The dataset was fitted to a fourparameter logical regression model (eq. (1)) empirically
so as to obtain the EC50H values at high doses and the EC50L values at low doses[7]. ⎛ ( x − xc ) ⎞ (1) ECA = θ min + Aexp ⎜ ⎟, 2 w2 ⎠ ⎝ where ECA is the effective activity, x is the logarithm of PCB concentration, θmin is the EROD activity in control group, A is the maximum value of the EROD activity, xc is the curve-oriented parameter and w is the slope for curve. 2.2 Preparation of the receptor structure AhR is one of the bHLH-PAS members in the transcription regulatory proteins. The ligand binding domain (LBD) from 232 to 402 amino acid residues is located in the most conservative district of PAS region. Bacteria photoactive yellow protein (PYP) was used as the three-dimensional template of Per-ARNT-Sim (PAS) region. Other preserved crystal structures such as the human-erg potassium channel (HERG) and the bacteria CN-bound BjFixL heme domain (FixL) were also － detected in this study[12 14]. The crystal structure of AhR is absent till now. In this case, PYP, HERG and FixL PAS domains were selected as AhR’s substitutes in the molecular simulations, since they presented a high level of conservative in the long-term evolution. The three crystal structures for molecular docking were acquired from protein data bank (http://www.pdb.org/), where the PDB entries were 2PYP, 1BYW and 1BV5. 2.3 Structural descriptors Semiempirical AM1 method in ChemOffice 2005 was adopted to calculate the necessary structural descriptors for PCBs after energy optimization, and the binding interactions between PCBs and simulated AhRs were analyzed by using FlexX module in SYBYL7.0. Some critical structure parameters such as heat of formation (HF), HOMO energy (EHOMO), LUMO energy (ELUMO) and dipole moment (μ) were reckoned for QSAR analysis. Tripos standard molecular field, Gasteiger-Hückel charge and Powell energy optimation strategy were utilized for the preparation of ligand structures. Energy convergence criterion was fixed at 0.01 kcal/(mol·Å), and other parameters rarely designated were default values. After analyzing the active center of receptor by SOLV technique in SiteID program, molecular docking research was performed at the DELL Precision370 station.
MU YunSong et al. Sci China Ser B-Chem | May 2009 | vol. 52 | no. 5 | 662-669
663
2.4 Statistical analysis Monte Carlo method was used for validating the robustness of the QSAR model by analyzing the liner regression equation. Firstly, median and standard error of the predictive sample were calculated, and then one hundred pseudo predictive values randomly originated depending on the normal distribution with the above median and standard error. Furthermore, correlation coefficient between every pseudo predictive value and the independent variable was acquired. Finally, pseudo predictive coefficient of the correlation (R*) was built up correspondingly when the confidence of this normal distribution was 0.95. The prediction potency was estimated by comparison between real correlation coefficient (R) and R* values.
program. It is shown in Figure 1 that PAS domain of FixL protein had the best homology to chicken’s AhR. Furthermore, the secondary structures like α-helix and β-sheet were mainly semblable. As shown in Figure 2, active centers of PYP, HERG Table 1
Fitting values for critical parameters and typical effective ones
Compounds
3 Results and discussion 3.1 Identification of decisive biological indicators Eleven dose-EROD activity curves had been fitted through four-parameter logical regression model (eq. (1)) with searching method of Newton's law. The estimated values of every parameter and significant level (P) are present in Table 1. According to these equations, typical effective values in high and low doses ranges (pEC50H and pEC50L) were achieved for QSAR studies.
θmin
A
Xc
w
P
pEC50H
pEC50L
PCB77
20.64 112.67 0.69 1.03 0.0001
−1.850
0.30
PCB78
22.80 152.40 3.23 0.86 0.0001
−3.703
−2.15
PCB79
22.62
57.59 1.82 0.86 0.0001
−2.736
−0.55
PCB81
12.30 171.05 0.55 1.30 0.0001
−1.579
1.02
PCB105
22.31
18.62 1.68 0.88 0.0001
−3.326
−0.47
PCB118
24.93
47.08 2.37 0.66 0.0001
−3.020
−1.25
PCB126
19.00 116.43 0.24 1.28 0.0001
−1.794
1.29
PCB138
20.65
14.77 2.15 0.88 0.0001
−3.130
−0.94
PCB156
22.56
69.95 1.97 0.74 0.0001
−2.806
−1.05
PCB157
27.33
50.36 1.68 0.73 0.0001
−2.462
−0.81
PCB169
22.67
92.60 1.04 0.89 0.0001
−2.062
0.10
3.2 Confirmation of AhR’s substitutive protein Procopioe et al.[15] had modeled mouse’s AhR LBD structure by PAS domain of FixL protein and set up a model for recognition of 2,3,7,8-TCDD. In this paper, sequence alignment is firstly performed among PAS domains of proteins (FixL, PYP and HERG) and chicken’s AhR LBD (Gallus gallus) using Clustal X
FixL Figure 2
664
Figure 1 Sequence alignment between LBD of chicken AhR and FixL protein and secondary structure analysis. Sequences shadowed are α-helix but the ones framed represent β-sheet.
PYP
Active centers of FixL, PYP and HERG by SiteID program.
MU YunSong et al. Sci China Ser B-Chem | May 2009 | vol. 52 | no. 5 | 662-669
HERG
and FixL were simulated through SiteID program in SYBYL7.0. The most obvious difference between these three structures was a spiral connecting department, in which FixL had GLY224 and GLY251 residues, PYP possessed PHE96 and VAL120 residues but HERG had PHE98 and LEU127 residues. Compared with other PAS domains, the active center of FixL protein was closest to AhR structure but the size of its pocket was much bigger than that of AhR. In this study, the binding interaction between FixL and 39 PCB congeners emphasized in the previous literature had been simulated by FlexX program. As a result, all of these PCBs could be accessed to the target active pocket successfully. But owing to its pocket size, the combining process and conformations of PCBs appear to be different in stereo orientation so that the combination mode needs further distinction. As shown in Figure 3, the key amino acid residues around the active pocket of FixL had been modeled. Green ribbon represented three-dimensional structure of FixL and purple space was the shape of active pocket through SiteID program. The key amino acid residues around co-crystallized ligand within 6 Å scope were mainly hydrophobic ones, including LEU, VAL, TYR, ILE and PHE. Therefore, hydrophobic interaction played an indispensable role in ligand-receptor binding. Compared to the 3D-QSAR study only based on structures of PCBs, it is better to stimulate binding process and validate research findings in the past. For example, it had been proven from CoMFA results that the halogenated aromatic compounds with 3.0 Å×10.0 Å
Figure 4
Figure 3
The active center of FixL and key amino acid residues.
rectangular dimension exhibited high biological activity and the molecular size of 6.8 Å×13.7 Å was optimal when considering van der Waals atomic radius [16]. The maximum volume of AhR reported in the literature was 14.0 Å×12.0 Å×5.0 Å[17], so molecules with the same orientation of active pocket can achieve good matching performance. 3.3 Comparison between alien binding modes of PCBs Based on toxicological datum of PCBs, the mode for interaction has usually been classified by the number and position of substituted chlorine. Although this classification might be in accordance with bioactive datum to some extent, the combination mechanism for different types of PCBs lacks microcosmic evidence to support. The molecular docking results for 39 PCBs are presented in Figure 4 for further mode differentiation. As shown in this figure, coplanar PCBs and non-coplanar
Two binding modes for different kinds of PCBs, coplanar (a) or non-coplanar (b).
MU YunSong et al. Sci China Ser B-Chem | May 2009 | vol. 52 | no. 5 | 662-669
665
Table 2
Activity datum with docking results for 39 PCB congeners No
Structure
pEC50
(SE, n)
IEF b)
TS (solv)
Non-ortho-PCB 2
3-CB
5.84
10, 3
0.00001
−15.16
3
4-CB
−
−
0.0
−14.34
11
3,3′-DiCB
−
−
0.0
−15.05
12
3,4-DiCB
7.72
0.02, 3
0.0008
−15.48
35
3,3′,4-TriCB
7.66
0.06, 3
0.0007
−15.35
37
3,4,4′-TriCB
7.40
0.03, 3
0.0004
−14.66
77
3,3′,4,4′-TeCB
9.30
0.06, 20
0.03
−15.84
78
3,3′,4,5-TeCB
6.85
0.08, 2
0.0001
−14.16
79
3,3′,4,5′-TeCB
8.45
0.23, 2
0.004
−15.04
80
3,3′,5,5′-TeCB
6.76
0.24, 2
0.00008
−13.13
81
3,4,4′,5′-TeCB
10.02
0.01, 3
0.2
−16.00
126
3,3′,4,4′,5-PeCB
10.29
0.03, 5
0.3
−16.15
127
3,3′,4,5,5′-PeCB
8.49
0.13, 5
0.005
−14.65
169
3,3′4,4′,5,5′-HxCB
9.10
0.04, 3
0.02
−15.14
66
2,3′,4,4′-TeCB
8.08
0.14, 6
0.002
−14.66
70
2,3′,4′,5-TeCB
7.44
0.06, 3
0.0004
−15.53
Mono-ortho-PCB
105
2,3,3′4,4′-PeCB
8.53
0.21, 6
0.005
−15.33
118
2,3′,4,4′,5-PeCB
7.75
0.17, 6
0.001
−14.54
122
2′,3,3′,4,5-PeCB
6.08
0.03, 3
0.00002
−14.04
156
2,3,3′,4,4′,5-HxCB
7.95
0.0, 4
0.001
−14.37
157
2,3,3′,4,4′,5-HxCB
8.19
0.01, 2
0.002
−14.59
167
2,3′,4,4′,5,5′-HxCB
8.16
0.07, 3
0.002
−13.54
Di-ortho-PCB 4
2,2′-DiCB
−a)
−
0.0
−10.38
47
2,2′,4,4′-TeCB
−
−
0.0
−8.728
52
2,2′,5,5′-TeCB
−
−
0.0
−9.787
101
2,2′,4,5,5′-PeCB
−
−
0.0
−10.13 −10.68
110
2,3,3′,4′,6-PeCB
6.55
0.15, 3
0.00005
128
2,2′,3,3′,4,4′-HxCB
7.90
0.10, 4
0.001
−10.45
138
2,2′,3,4,4′,5′-HxCB
8.06
0.42, 4
0.001
−13.85
153
2,2′,4,4′,5,5′-HxCB
−
−
0.0
−11.81
170
2,2′,3,3′,4,4′,5-HpCB
7.05
0.14, 3
0.0002
−12.25
180
2,2′,3,4,4′,5,5′-HpCB
6.70
0.06, 3
0.0002
−6.986c)
194
2,2′,3,3′,4,4′,5,5′-OcPCBs
6.49
0.05, 3
0.00005
−7.119
Tri-ortho-PCB 95
−
−
0.0
−9.955
7.57
0.17, 3
0.006
−8.662
2,2′,3,4′,5,5′,6-HpCB
−
−
0.0
−11.08
2,2′,3,5′,6-PeCB
139
2,2′,3,4,4′,6-HxCB
187 Tetra-ortho-PCB
2,2′,6,6′-TeCB
−
−
0.0
−8.416
136
2,2′,3,3′,6,6′-HxCB
−
−
0.0
−6.883
209
2,2′,3,3′,4,4′,5,5′,6,6′-DeCB
−
−
0.0
−5.545
54
a) No induced activity, b) induction equivalency factors (IEFs) are EROD induction potencies relative to 2,3,7,8-TCDD, and c) PCBs 180 and 194 are not docked into the active pocket.
666
MU YunSong et al. Sci China Ser B-Chem | May 2009 | vol. 52 | no. 5 | 662-669
ones had different binding conformations in the active center. The coplanar PCBs (non-ortho and mono-ortho) could be aligned with TCDD and located in the same plane. While the non-coplanar PCBs (di-ortho, tri-ortho and tetra-oratho) turned out to be in the same orientation with temples of PCBs 95 and 170 expressed in purple ball and stick structure. Only by focusing on binding free energy in Table 2, can PCBs with scant chlorine perform poor selectivity to AhR. The possible reason is that scant chlorine PCBs have variable conformations with respect to the bigger size of FixL’s active pocket. For example, PCBs with one or two chlorine atoms (PCBs 2, 3, 11, etc.) have higher predictive activity than actual biological ones. But for the non-ortho-substituted congeners and mono-ortho-PCBs, molecular simulation results show that all of them turned to be with high affinity to AhR, suggesting a good relationship between induced EROD activity and binding free energy in low doses range. Researches reported previously on toxic equivalent non-ortho-PCBs and mono-ortho-PCBs provided the evidence that both types of PCBs have common coplanar characteristic[9,18]. By contrast, PCBs containing chlorine atoms in 2 and 2′ positions simultaneously displayed less or no activities with the total score values (TS) all beyond −10 kJ·mol−1. Edwards et al.[19] had compared symmetrical structure of PCB77 with PCBs 47, 52 and 54 within low doses of EROD activity and found that coplanar PCB77 is more active than other non-coplanar PCBs. These achievements fully demonstrate that coplanar and non-coplanar structures have great differences especially in the Ah receptor-mediated process. 3.4 QSAR study on PCB-induced CHE-EROD activity in both doses ranges In order to study the relationship between PCBs structure and CHE-EROD activity in both high and low doses ranges, some meaningful models had been established including pEC50L and pEC50A, pEC50A and Total Score (TS), pEC50L and TS as well as pEC50H and pEC50L (eq. (2)－(5)), where n represents sample size, R coefficient of correlation, SE standard error, p level of significance and TS free energy in molecular docking. The sample sizes in eqs. (2)－(4) are 7, 12, 11 and 11. So Monte Carlo simulation testing might be needed for solving small sample problem. As shown in Table 3, coefficients
of correlation in eqs. (2)－(5) are higher than the ones of R* deprived from Monte Carlo simulation within 95% confidence interval. pEC50L = (1.190±0.145)pEC50A+(–7.005±0.810), n=7, R2=0.931, SE=0.26, p=0.00,
(2)
pEC50A = (–2.145±0.011)TS +(2.737±0.075), n=12, R2=0.895, SE=0.11, p=0.00,
(3)
pEC50L = (–0.916±0.229)TS+(–14.144±3.41), n=11, R2=0.641, SE=0.36, p=0.00,
(4)
pEC50H = (0.614±0.1)pEC50L+(–2.336±0.106), n=11, R2=0.806, SE=0.31, p=0.00,
(5)
pEC50H = (–0.701±0.195)Dipole+ (–1.876±0.21), n=10, R2=0.619, SE=0.40, p=0.00
(6)
Table 3
Monte Carlo simulating results in eqs. (2)－(5)
Equations
R
R*
(2)
0.965
0.713
(3)
0.946
0.557
(4)
0.800
0.484
(5)
0.898
0.573
In eq. (2), low-dose induced EROD activity has high correlation with affinity datum to AhR (R = 0.965). In other words, PCBs with specific stereo conformation and electrostatic field can increase EROD activity by interacting with CEH-AhR. This hypothesis has been fully confirmed from molecular simulation that the TS values had significant correlation with PCBs-AhR binding values (pEC50A) and low-dose-induced concentrations (pEC50L). But the correlation coefficient R of eq. (4) is far below the one of eq. (2) while SE is higher comparably. It was mainly due to different amino acid residues in active site of two receptors, FixL and AhR, which led to the warp between simulation results and experimental evidence. Although there are still particular cases with respect to docking results, the number and location of substituted chlorine directly affected ligand-receptor binding capacity. Over all, the more dispersedly the chlorines are located, the lower the polarizabilty it has resulting in hydrophobic combination. Correspondingly, FixL possessed a typical hydrophobic cavity and was inclined to binding with more non-polar elements. Similar studies about more simple congeners of the PCDFs have been reported and QSAR models on the polarization had been obtained satisfactorily[20].
MU YunSong et al. Sci China Ser B-Chem | May 2009 | vol. 52 | no. 5 | 662-669
667
The experimental study about EROD activity for low-dose PCBs has been widely carried out and explanation for relevant mechanism was comparatively systematic. It has been reported that low doses of four chlorine-substituted dibenzo-dioxins, for example, PCBs 126 and 169, could induce expression of CYP1A1 in a vicious human breast cancer cell named MCF-7 and further influence formation of 2-hydroxyestrogen-monomethylethers (2-MeOE) and EROD activity. However, the reason for inducing EROD activity at high doses is still a bone of contention. Some people thought that in many cases the high-dose-halogenated hydrocarbons (HAHs) might engender cell toxicity or competitive inhibition to enzyme, but Hahn et al.[21] proposed query for this explanation. Previous findings had shown that high doses of PCBs will not directly inhibit the EROD activity but indirectly interfere with AhR-mediated process and expression of CYP1AmRNA[22]. The possible reason to speculate was that the CYP1A1 gene in its catalytic cycle could lead to the formation of active oxygen free radical, which was supposed to be the antagonist against expression of CYP1A1 gene. In eq. (5), pEC50H values present certain relativity with pEC50L values. The PCBs-induced EROD effect still has been concerned with Ah receptor-mediated action. Another reason why a turning point turned up on the dose-response curve was probably that when the concentration of PCBs was far higher than the physiological activity level, AhR’s binding capability would reach saturation, thus resulting in the appearance of reactive oxygen species. In order to explore possible constraints on EROD activity in high doses range, the 1
Saal F S, Timms B G, Montano M M, Palanza P, Thayer K A, Nagel S
linear regression analysis was performed between pEC50H and correlative structure or property parameters. As shown in eq. (6), pEC50H possesses a good correlation with molecular dipole moment. Based on the Monte Carlo testing, the pseudo coefficient of correlation (R*) is 0.491 (95% confidence interval), less than that of original regression equation (R = 0.787), which demonstrates that eq. (6) has a good ability to predict. Molecular dipole moment of PCBs is influenced by location and quantity of substituted chlorine atoms. Therefore, the endocrine disrupting effect of PCBs at high doses level is to a large extent determined by distribution of molecular charge as well as complementarity for positive electricity of AhR. These assumptions need to be verified by more systematic experimental data.
4 Conclusions In high and low doses ranges, the mechanisms to induce EROD activity are different and ultimately cause the non-monotonic dose-response relationship. Based on the QSAR model of CEH-EROD activity in both dose ranges, it is revealed that binding with receptor may predominate in interference with physiological function of cytochrome P4501A-P4501A in low dose range and binding mode is determined by structure of PCBs. While the dose level increases to a certain extent, the mechanism will be attributed to acute toxicity owing to molecular polarity or distribution of charges and consequently damage structure and function of chicken embryo hepatocyte. 5
C, Dhar M D, Ganjam V K, Parmigiani S, Welshons W S. Prostate
ture-activity relation ships for polychlorinated biphenyls. Prog Chem
enlargement in mice due to fetal exposure to low doses of estradiol or diethylstilbestrol and opposite effects at high doses. Proc Natl Acad 2
6
Sci USA, 1997, 94: 2056－2061 Wetherill Y B, Petre C E, Monk K R, Puga A, Knudsen K E. The
3
crine-disrupting chemicals with estrogenic activity. Environ Health 7
Cancer Ther, 2002, 1: 515－524 Saal F S, Nagel S C, Palanza P, Boechler M, Parmigiani S, Welshons
4
tetrachloro dibenzopdioxin equivalent concentrations in environ8
mental samples. Environ Sci Technol, 1996, 30: 706－715 Safe S, Bandiera S, Sawyer T, Zmudzka B, Mason G, Romkes M.
behaviour in male mice. Toxicol Lett, 1995, 77: 343－350 Zala S M, Penn D J. Abnormal behaviours induced by chemical pol-
Effects of structure on binding to the 2,3,7,8-TCDD receptor protein
lution: A review of the evidence and new challenges. Anim Behav,
spect, 1985, 61: 21－23
2004, 68: 649－664
668
Perspect, 2003, 111(8): 994－1006 Kennedy S W, Lorenzen A, Norstrom R J. Chicken embryo hepatocyte bioasay for measuring cytochrome P4501A-based 2,3,7,8-
W S. Estrogenic pesticides: Binding relative to estradiol in MCF-7 cells and effects of exposure during fetal life on subsequent territorial
(in Chinese), 1996, 8(4): 293－300 Welshons W V, Thayer K A, Judy B M, Taylor J A, Curran E M, Saal F S. Large effects from small exposures. I. Mechanisms for endo-
xenoestrogen bisphenol A induces inappropriate androgen receptor activation and mitogenesis in prostatic adenocarcinoma cells. Mol
Miao X S, Chu S G, Xu X B. Advances in the research of struc-
and AHH induction——Halogenated biphenyls. Environ Health Per9
Safe S. Polychlorinated biphenyls (PCBs), dibenzo-p-dioxins
MU YunSong et al. Sci China Ser B-Chem | May 2009 | vol. 52 | no. 5 | 662-669
(PCDDs), dibenzofurans (PCDFs), and related compounds: Envi-
16
ronmental and mechanistic considerations which support the development of toxic equivalency factors (TEFs). Crit Rev Toxicol, 1990, 21: 51－88 10
17
Bandiera S, Sawyer T W, Campbell M A, Fujita T, Safe S. Competi-
ture-activity relationships of dioxins and dioxin-like compounds: Model validation and Ah receptor characterization. Chem Res Toxicol,
biphenyls —— A QSAR analysis. Biochem Pharmacol, 1983, 32:
18
3803－3813 Bandiera S, Safe S, Okey A B. Binding of polychlorinated biphenyls mixed-type inducers to cytosolic Ah receptor. Chem Biol Interact,
transformed rat cytosolic arylhydrocarbon receptor. Chem Res Toxi-
13
19
1982, 39: 259－277 White R, Jobling S, Hoare S A, Sumpter J P, Parker M G. Environmentally persistent alkyl-phenolic compounds are estrogenic. Endo-
hepatic microsomal alkoxyresorufin O-dealkylation activities by tet20
Shoji H, Tomoko I, Totaro Imasaka. Chlorine substitution pattern,
crinology, 1994, 135 (1): 175－182
molecular electronic properties, and the nature of the ligand-receptor
Shipley J M, Waxman D J. Aryl hydrocarbon receptor-independent
interaction: Quantitative property-activity relationships of poly-
cholanthrene. Toxicol Appl Pharmacol, 2006, 213(2): 87－97
21
chlorinated dibenzofurans. Chem Res Toxicol, 2005, 18: 232－238 Hahn M E, Lamb T M, Schultz M E, Smolowitz R M, Stegeman J J.
Denison M S, Pandini A, Nagy S R, Baldwin E P, Bonati L. Ligand
Cytochrome P4501A induction and inhibition by 3,3′,4,4′-tetrach-
binding and activation of the Ah receptor. Chem Biol Interact, 2002,
lorobiphenyl in an Ah receptor-containing fish hepatoma cell line
141: 3－24 15
col, 1994, 7: 544－550 Edwards P R, Hrycay E G, Bandiera S M. Differential inhibition of rachlorobiphenyls. Chem Biol Interact, 2007, 169(1): 42－52
activation of estrogen receptor-dependent transcription by 3-methy14
1995, 8: 847－858 Santostefano M, Liu H, Wang X H, Chaloupka K, Safe S. Effect of ligand structure on formation and DNA binding properties of the
clasified as either phenobarbitone-, 3-methylchlolanthrene- or
12
lorodibenzo-p-dioxin in rat liver. Mol Pharmaco, 1985, 28: 357－363 Waller C L, McKinney J D. Three-dimensional quantitative struc-
tive binding to the cytosolic 2,3,7,8-tetrachlorodibenzo-p-dioxin receptor: effects of structure on the affinities of substituted halogenated
11
Gillner M, Bergman J, Cambillau C, Fernström B, Gustafson J R. Interactions of indols with specific binding sites for 2,3,7,8-tetrach-
Procopio1 M, Lahm A, Tramontano A, Bonati1 L, Pitea1 D. A model
22
(PLHC-1). Aquatic Toxicology, 1993, 26(3-4): 185－208 Head J A, Kennedy S W. Differential expresion, induction, and sta-
for recognition of polychlorinated dibenzo-p-dioxinsby the aryl hy-
bility of CYP1A4 and CYP1A5 mRNA in chicken and herring gull
drocarbon receptor. Eur J Biochem, 2002, 269: 13－18
embryo hepatocytes. Comp Biochem Phys C, 2007, 145: 617－624
MU YunSong et al. Sci China Ser B-Chem | May 2009 | vol. 52 | no. 5 | 662-669
669
