Measuring the Kinetics of the Binding of ... - ACS Publications

Article pubs.acs.org/est

Measuring the Kinetics of the Binding of Xenoestrogens and Estrogen Receptor Alpha by Fluorescence Polarization Kwok-Wing Yiu, Chi-Kin Lee, Ka-Cheung Kwok, and Nai-Ho Cheung* Department of Physics, Hong Kong Baptist University, Kowloon Tong, Hong Kong, People’s Republic of China S Supporting Information *

ABSTRACT: The mechanism of endocrine disruption by environmental xenoestrogens is unclear. Bisphenol-A (BPA) is an example. Its concentration in human serum is low, and its binding with estrogen receptor (ER) is weak. Yet its effect on prostate and mammary gland development was observed. We investigated whether this effect could be explained in terms of binding kinetics. We used the method of fluorescence polarization anisotropy to measure the kinetic rate constants of the binding of ERα with 19 xenoestrogens. Relative binding affinities (RBA) were also deduced from the kinetics. We drew three observations. First, our RBAs were consistent with published values, thus establishing the validity of our results. Second, our method allowed the determination of low RBAs (∼10−4) of lipophilic ligands, such as dibutyl phthalate. They could not be measured by steady-state IC50 assays because of their low solubility. Third, we found that BPA had a surprisingly high kon > 104 M−1 s−1. While its RBA was 1500 times lower than that of 17β estradiol (E2), its kon was >1/90 that of E2. As a result, a 10 min surge of BPA from pM to nM could drive the fraction of BPA-activated ERα to a potent 0.1%. This might help to explain the observable estrogenic impacts of BPA.

■

INTRODUCTION

This kind of kinetic argument may be applicable to endocrine disruptors as well. Unfortunately, the kinetics of xenoestrogenER binding has not been reported. We therefore measured, in this study, the binding kinetics of the alpha form of ER and 19 xenoestrogens using a technique we recently developed.13 We correlated the kinetic rate constants and the molecular structure of the ligands. We will show that binding kinetics offers a useful perspective for understanding endocrine disruptors.

How environmental xenoestrogens disrupt endocrine function is unclear.1 For example, the estrogenic impact of bisphenol A (BPA) remains controversial despite decades of study.2 Its chronic amount in human serum was contested. One school puts it at the pM level, while another school puts it at the nM level.3,4 Its low affinity for estrogen receptor (ER) also argued against ER-mediated effects.3 Yet low doses of BPA during fetal, perinatal, and puberty stages were shown to affect both prostate and mammary gland development.3 In order to explain the tangible downstream effect of BPA, several explanations were offered: (1) that BPA acts via nonclassical pathways mediated by membrane ER,5 or receptors other than ER;2 (2) that metabolites of BPA, especially 4methyl-2,4-bis(4-hydroxyphenyl)pent-1-ene, bind to ER with high affinity to elicit estrogenic effects;6 and (3) that BPA induces nonmonotonic dose−response.3 The first two explanations assumed that BPA could not elicit ER-mediated response and alternative mechanisms were required. This assumption was certainly reasonable given the generally positive correlation between binding affinity and downstream potencies.7 Interestingly, however, this correlation was recently challenged in the field of drug discovery. It was argued that drug reactions in live cells were hardly equilibrium reactions. Steady-state binding affinity was therefore a poor indicator of drug potency. Instead, the residence time of the drug in the receptor would be a better indicator.8−12 © 2014 American Chemical Society

■

MATERIALS AND METHODS We measured the kinetics by fluorescence polarization (FP) and total internal reflection fluorescence (TIRF). The setup was previously described.13,14 A brief account is given below. Experimental Setup. A fluorescent hormone analog F was introduced into a sample flow cell to bind with tethered ERα (see Figure 1).15 Both bound and free F in the illuminated volume was excited to fluoresce by a 473 nm laser beam. The fluorescence emissions were collected with a microscope objective and directed through a polarizing beam splitter. The beams with respective polarization parallel (∥) and perpendicular (⊥) to that of the excitation beam were imaged onto an electron-multiplying charge-coupled detector (emCCD). Comparison of the intensities of the two arms gave the FP Received: Revised: Accepted: Published: 11591

May 26, 2014 August 28, 2014 September 2, 2014 September 2, 2014 dx.doi.org/10.1021/es503801c | Environ. Sci. Technol. 2014, 48, 11591−11599

Environmental Science & Technology

Article

Ligands. Green fluorescent hormone F was diluted in 50% methanol and 50% DMSO as stock.15 High purity (≥95%) diethylstilbestrol (DES),17β-estradiol (E2), ICI 182 780, raloxifene, tamoxifen, α-zearalenol, coumestrol, zearalenone, resveratrol, apigenin, bisphenol A, perfluoro-nonanoic acid, perfluoro-octane-sulfonic acid, perfluoro-octanoic acid, pp′dichlorodiphenytrichloroethane, and pp′-dichlorodiphenyldichloroethylene (all 16 ligands above from Sigma); as well as analytical grade diethylhexyl phthalate and dibutyl phthalate (both phthalates from Supelco) were dissolved in DMSO as stocks. For kinetic experiments, all reagents were finally diluted in the vendor’s binding buffer but with DMSO maintained at 2%. The vendor’s binding buffer contained about 1% DMSO.15 We doubled that percentage to ensure dissolution of lipophilic ligands such as DBP, DEHP, DDT, and DDE. We found that the kinetics rates were not sensitive to this small increase in [DMSO], at least for the five standard ligands measured in this work as well as in Kwok and Cheung 2010.13 Data Analysis and Kinetics Modeling. When F was injected into the flow cell at t = 0, it remained free and therefore tumbled in solution to give a low polarization anisotropy A. As F started to bind with R, it ceased tumbling, and A increased. The bound F fraction f was deduced from A using the following equation,

f= Figure 1. Sample flow cell was made from a quartz slide Q coated with receptors ERα on the underside, filled with buffer, sandwiched with a cover glass C and sealed with sealant S. Reagents were injected through the inlet port I and ejected through the outlet port O. Receptors ERα were tethered by successive links: positively charged poly lysine PLL electrostatically adsorbed on negatively charged quartz Q, a primary antibody IgG bonded covalently to PLL, and a secondary antibody Abα captured by IgG. Abα specifically targeted the AF1 region of ERα. A 473 nm laser beam was coupled by a prism P and total-internally reflected (TIR) at the quartz-solvent interface to excite the bound green fluormone F. A second 473 nm laser beam was directed through the sample cell to excite both bound and free F. Fluorescence emissions were collected by a microscope objective M, split into ∥ and ⊥ polarized arms, and finally imaged on an emCCD. A typical image is shown in the inset. The uppermost bright spot was due to the ∥ arm of the through beam. The second spot was the ∥ arm of the TIR beam. The bottom two spots were the corresponding images of the ⊥ arm.

A − A̲ A̅ − A̲

(1)

where A is the fluorescence anisotropy, given by the following:16

A=

I// − I⊥ I// + 2·I⊥

(2)

Maximum A̅ was determined by flushing out all free F. Minimum A was determined by injecting an excess of a competitor ligand such as DES or E2. The time-dependence of f(t) can be modeled as follows. kon

With reference to the binding scheme, R+F HooI RF, the koff

population dynamics obeys the following differential equation, d[RF] = kon[R][F] − koff [RF] dt

(3)

where standard notations are used. We define the fraction of bound F, f, by the following:

anisotropy signal. A second laser beam at 473 nm was totalinternally reflected off the quartz-solvent interface of the flow cell to excite the bound F. The TIRF signal was also directed through the polarizing beam splitter and imaged onto the emCCD. A typical image is shown in the inset of Figure 1. Analyte ligand L was subsequently injected into the flow cell to displace the bound F and movies of the FP and TIRF images were captured. ERα Tethering. ERα was tethered on the quartz substrate as shown in Figure 1. Details are explained in the figure caption. Each link of the tether was checked for strength and specificity to guard against false positives and negatives.14 The entire tether was firm enough to allow multiple flushing so up to three different ligands could be measured sequentially in one flow cell. Typical ERα coverage was about 30 molecules per μm2; it was halved for weak binders to intensify the competition with F for better assay sensitivity. At the lower coverage, loss of ERα with flushing became apparent so ERα coverage was remeasured between ligand injections.

f=

[RF] [F] + [RF]

(4)

and eq 3 can be rewritten as follows: •

f = kon(1 − f )(ρtotal − f [F]total ) − koff f

(5)

where [F]total = [F] + [RF] is the concentration of free and complexed F. Among the four parameters that appear in eq 5, [F]total could be deduced from the total fluorescence emissions while kon, koff, and ρtotal (number of ERα molecules per volume if all tethered receptors were dissolved in solution) could be determined from curve-fitting f(t). As explained earlier, f(t) was deduced from the FP signal. It was cross-checked by the TIRF signal. To determine the binding kinetics of ERα with a nonfluorescent ligand L, we let L compete with F for ERα. The 11592

dx.doi.org/10.1021/es503801c | Environ. Sci. Technol. 2014, 48, 11591−11599


Article

Figure 2. ERα-BPA binding kinetics. At t = 0, 10 μL of 1 nM F was flowed into the sample cell (red crosses). At t = 400 s, another 10 μL of 1 nM F was flowed in and the equilibrium bound fraction was shown (blue crosses). At t = 600 s, 10 μL of 2 μM BPA was flowed in and the subsequent dissociation of F from ERα was shown (brown crosses). At t = 1000 s, 10 μL of 1 nM F was flowed into the cell again to calibrate the new ER coverage (light blue crosses). At t = 1200 s, 10 μL of 10 μM E2 was flowed in to induce complete dissociation of F and to generate the baseline (black crosses). Best fits are shown as solid traces. Misfits are also shown: 50% higher koff (curve labeled d), 50% lower koff (curve c), and both kon and koff 50% higher (curve b) and 50% lower (curve a). When the 2 μM BPA was replaced by 1 μM genistein, the equivalent displacement of F between t = 600 and 900 s is shown in the top inset. Similarly, the equivalent displacement of F by 50 μM DEHP is shown in the bottom inset.

explained in the figure caption. We used ligand concentrations comparable to their dissociation constants. We first determined the rate constants of F in three steps. (1) koff was directly gotten from the dissociation curve of F (t > 1,200 s in Figure 2) because rebinding of F was prevented by the excess E2. (2) KD and ρtotal were determined from the two steady-state values of f at t = 100−300 s and 400−550 s. (3) Finally, the values of the three parameters, koff, kon, and ρtotal were optimized to best fit f(t) using eq 5 for t = 0−550 s, and t > 1200 s. On the basis of numerous measurements, more than ten sets of koff and kon were determined. Their values were found to be highly consistent. The relative standard deviation was only 7% for koff and 6% for kon. ̃ of BPA as follows. We rẽ and kon We next determined koff evaluated the ERα coverage (ρtotal) using the new equilibrium f value at t = 1000−1200 s. This accounted for loss of receptors due to reagent flushes. The ERα coverage was taken to be the average of the initial and final ρtotal. We finally used eqs 9 and ̃ / 10 to fit the steady-state f at t = 800−900 s to give K̃ D (= koff ̃ ) and we fit the dissociation curve (t = 600−800 s) to give kon ̃ and kon ̃ . The best fit is shown in Figure 2 (bold solid brown koff curve). Uncertainty in the Fitted Values. The tolerance in the curve-fitting is illustrated by the a curve (Figure 2) when both ̃ and kon ̃ were decreased by 50%. The b curve was when both koff rates were increased by 50%. Evidently, the b curve is not significantly different from the best-fit, implying that the rates could be higher by 50% or more. For stronger binders such as genistein, the uncertainty would be less. This is illustrated by the clearly separated a, b and best-fitted curves in the top inset of Figure 2. The uncertainty became worse for weaker binders such as DEHP, as shown in the bottom inset. The reason for weak binders to cause large uncertainties is as follows. Imagine blank buffer was injected at t = 600 s. As the ERα-F complex dissociated, there would not be competitor ligands to occupy the vacated ERα site. As a result, F could

∼ kon

ERα-L binding is represented by R+L H∼ooI RL. The population koff

dynamics then obeys the following two differential equations, d[RF] = kon[R][F] − koff [RF] dt

(6)

and, d[RL] ̃ [R][L] − koff ̃ [RL] = kon dt

(7)

which are coupled by the constraint, [R] + [RF] + [RL] = constant

(8)

Equations 6 and 7 can be rewritten as follows: •

f = kon(1 − f )(ρtotal − f [F]total − f ̃ [L]total ) − koff f

(9)

and, •

∼ ̃ (1 − f ̃ )(ρ − f [F ]total − f ̃ [L]total ) − koff ̃ f̃ f = kon total (10)

where variables with tildes are the L analogs of F. L was not fluorescent so [L]total had to be determined from dilution ratios based on the known concentration of the stock. The five unknowns, ρtotal and the four rate constants in eqs 9 and 10, could be determined by solving the coupled differential equations numerically to best fit the measured f(t). More theoretical and modeling details were given elsewhere.13,14

■

RESULTS Determination of Kinetic Rate Constants. We will use the binding of ERα and BPA to explain how the rate constants were determined. A typical plot of bound fraction f(t) versus time is shown in Figure 2. Ligands of various volume and concentration were injected at specified time points, as 11593



Article

Table 1. On and off Rate Constants and Relative Binding Affinities (RBA) of ERα-Ligand Interactionsa kon (103 M−1 s−1)b ligand green fluormone (F) standard agonist diethylstilbestrol (DES) 17β-estradiol (E2) antiestrogen ICI 182,780 raloxifene (RAL) 4-hydroxytamoxifen (4OHT) tamoxifen (TAM) phytoestrogens α-zearalenol (α-ZEA) coumestrol (COUM) zearalenone (ZEA) genistein (GEN) resveratrol (RESV) apigenin (APIG) plasticizers bisphenol A (BPA) diethylhexyl phthalate (DEHP) dibutyl phthalate (DBP) perfluoroalkyls perfluoro-nonanoic acid (PFNA) perfluoro-octane-sulfonic acid (PFOS) perfluoro-octanoic acid (PFOA) pesticides pp′-dichlorodiphenytrichloroethane (DDT) pp′dichlorodiphenyldichloroethylene (DDE)

koff (10−3 s−1)c e

this work

K&C 2010

58 000 (3500)

57 000 (10 000)

1900 (200) 1000 (200) 130 (30) 500 (100) 110 (60)

1,000 (100) 1100 (300)

330 (50) 240 (90) 100 (30)

220 (20) 60 (10) 59 (6)

this work

K&C 2010

RBAd e

this work

K&C 2010

e

hERα-FPf

29 (2)

30 (5)

364 (26)

270 (6)

140 (70)

1.4 (3) 1.8 (2)

1.5 (3) 1.5 (4)

200 (40) 100

100 (20) 100

120 (30) 100

1.9 (2) 1.37 (6)

1.3 (5) 1.6 (4) 1.73 (6)

2.8 (3) 4.5 (6) 3.5 (5) 3.6 (3)

127 (34) 68 (14) 7 (3)

205 (240) 35 (17) 21 (9) 8 (3)

7.3 (9) 2.8 (7) 2.5 (3)

26 (8)

2.3 (5)

8 (4) 4 (1)

9 (1) 9 (1)

17 (8) 4 (3) 48

1.5 (6) 0.13 (7) 0.06 (1)

radiog

11.2 7.5 6 (9) 0.056 0.3

11 (3)h 0.20 (1)h 0.087 (20)h

26 (4)h 28 (3)h 28 (2)h

0.065 (10) 0.0010 (1) 0.000 46 (8)

1.3 (5)h 1.1 (1)h

27 (6)h 29 (1)h

0.0070 (16) 0.0056 (3)

0.0009 0.007

0.27 (8)h

29 (2)h

0.0013 (4)

0.0008

0.070 (10)h

28 (3)h

0.000 37 (3)

h

h

0.000 36 (9)

0.073 (20)

29 (1)

0.04 NDi NDi

0.03−0.09

0.0003 0.0003− 0.009

a

Standard deviations of three or more trials are given in parentheses. bAssociation rate. cDissociation rate. dRelative binding affinity. eRef 13. fRBA values of human full length ERα measured by FP, as reported in the literature.15,18,21,26,43,44 gRBA values measured by radio-assays, as reported in the ̃ ’s are comparable to that of the fluormone F. iIC50 > 10−3 M.22,47 literature.19−21,25,40,44−48 hReported values are lower bounds because koff

rebind and the bound fraction f would decay at the slowest ̃ and high koff ̃ limiting rate. A weakly binding ligand with low kon was like a noncompetitive ligand. The dissociation curve therefore approached the blank buffer limit. That explained why ̃ the fitted curve was no longer sensitive to further increase in koff ̃ or further decrease in kon. To circumvent this indecisiveness, we ̃ value that gave a good fit, that is, we quoted the minimum koff reported a lower bound. We should point out, however, that the uncertainty in K̃ D was still well within 50% even for the weakest binder. This is illustrated by the c and d curves (Figure 2) when 50% lower and higher K̃ D values are respectively plotted. As can be seen from the bottom inset, these misfits are well-separated from the best-fit even for the case of DEHP. In other words, while the kinetic rates could not be precisely measured for weak binders, the equilibrium dissociation constants could still be measured with high accuracy. Tabulating Results. The binding kinetics of ERα and BPA was measured in three separate trials. The average rate constants are given in Table 1. As explained above, these rates were lower bounds. The standard deviations are shown in parentheses as errors. On the basis of K̃ D, the relative binding affinity (RBA) was also computed and displayed;17 it was relative to an RBA of 100 for E2. In a similar way, we determined the rate constants and RBA of all 19 ligands. The average of three or more trials for each

ligand is listed in Table 1; the corresponding standard deviations are given as errors. The ligands were grouped under six categories, and ranked by RBA within each category. Published rate constants and RBA’s are also tabulated for easy comparison. For the top 12 ligands (DES through APIG) in Table 1, their ̃ ’s were significantly less than that of F. Their on and off rates koff could therefore be determined unambiguously from curvefitting (top inset of Figure 2). The values of their kinetic rate constants shown in Table 1 were therefore true rates. On the contrary, for the eight weak binders listed at the bottom of ̃ ’s were Table 1 (BPA through DDE), because their koff comparable to that of F, the rates reported were lower bounds.

■

DISCUSSION Agreement of Our Results with Published Data. With reference to Table 1, among the 19 ligands that we studied, only five of them had their kinetic rates reported in the literature.13 Our results agreed with them to within 40%. For the remaining 14 ligands, their RBAs were measured by various other groups. We selected those studies that used similar methods as ours for consistency. They are listed under the “hERα-FP” column in Table 1. The column title refers to human ERα and fluorescence polarization assay. As can be seen, except for pp′-DDT, our results agreed with all of them to within a factor of 4. A factor-of-four difference was considered 11594



Article

Figure 3. Ligands ranked by dissociative half-life of ERα-ligand complex. Their agonist−antagonist nature is indicated by red (agonist), green (antagonist), blue (mixed), and black (unknown) labels. Ligand abbreviations are explained in Table 1. For graphical clarity, pp′-DDE is not shown. It is similar to pp′-DDT in terms of half-life and molecular structure. For the same reason, PFOS and PFOA are not shown because they are similar to PFNA.

reasonable. Even when hERα-FP was adopted in all the studies, the discrepancy among published RBA’s could be large. For example, RBA for TAM ranged from 1.7 to 9.4, and that for GEN ranged from 0.13 to 16.15,18 Our measured RBA for pp′-DDT was 80× lower than the reported value. However, another group also measured that affinity using radio-labeled E2.19 Their result is listed under the “radio” column in Table 1. It agreed with ours. RBAs of the other nine ligands were measured by radio assays and are listed under the “radio” column as well. As can be seen, when compared against our results, four (ICI 182 780, PFOS, PFOA and pp′-DDE) agreed to within 40%, and three (α-ZEA, APIG, PFNA) were of the same order of magnitude. This kind of agreement was reassuring. Published affinities could differ by orders of magnitude when different assay methods were used.20,21 Finally, for the remaining two ligands DBP and DEHP, their RBAs were too low to be measurable by others.22

In summary, among the 19 ligands that we studied, 16 gave results consistent with published values, one disagreed with a reported result but agreed with another, and two had RBA too low to be measurable by others. Measuring Weak Binding. As the case of DBP showed, we could measure RBA as low as 10−4. Other methods would be handicapped because they determined RBA by ligand titration to determine IC50. For RBA in the 10−4 range, ligand concentration would need to exceed the solubility limit of lipophilic ligands such as DBP.22 We did not measure IC50 but simulated the binding kinetics instead. We found that a DBP concentration of 100 μM was already adequate to displace enough F from ERα for curve-fitting to yield an RBA of 4.6 × 10−4. Dissociative Half-Life and Molecular Structure. Much work had been done on correlating the structure of the ERαligand complex and their binding affinity using crystallographic studies,23 and in silico docking.6,24 However, the correlation of structure and kinetic rate constants had not been reported.9 11595



Article

̃ of ERα-ligand binding. The same label abbreviation and color code as Figure 3 is used. Labels underlined indicate Figure 4. Ligands ranked by kon ̃ is only a lower bound. that the corresponding value of kon

The half-lives spanned a 20× range, from 0.4 to 8.4 min. Apparently, the long-lived ones shared the common feature of a linear structure with two diametric phenolic rings. Potent antagonists had bulky side chains to prevent helix 12 of ERα from forming the coactivator docking pocket.30 The importance of diametric phenolic rings was further illustrated by the following example. RESV and APIG had structures similar to GEN except for two nondiametric OH groups. Their lifetimes were 4× shorter than that of GEN. Another example was BPA. Its structure was not linear and its half-life was among the shortest. It is interesting to note that GEN, COUM, and ZEA had the longest half-life among the phytoestrogens. While the chemoprotective effect of these compounds against breast cancer was debated,31,32 they might serve as lead compounds for the synthesis of selective estrogen receptor modulators.

Below, we will correlate the ligand structure with its dissociative half-life. The importance of dissociative half-life is best understood from the perspective of antiestrogens. Antagonists block agonist-induced transcription by occupying the ERα site. A longer lifetime of the ERα-antagonist complex would mean more effective blocking.8−12 The dissociative half-life τ1/2 of the receptor−ligand complex ̃ by the following: is related to koff ̃ τ1/2 = ln(2)/koff

(6)

Here, half-life is correlated with molecular structure for the first time. This is shown in Figure 3. We included ligands studied in this work as well as Kwok and Cheung 2010.13 The agonist− antagonist properties of most of these ligands were known.22,25−29 We labeled the agonists in red, antagonists in green, mixed in blue and unknowns in black. 11596



Article

̃ of BPA were minute after the surge, it increased to 0.03%. If kon 30× lower, as expected of weak binders such as DEHP and PFOA, even if its RBA were the same, then the fraction of receptors occupied would only be 0.002% after 1 min (gray curve). The threshold receptor occupancy for physiological effects was reported to be 0.1%.37 So the [BPA] surge to 1 nM may not cause harmful effects. However, the situation can be different if the ERα-BPA complex is stabilized by recruited coactivators (CoA).34 We modeled this process numerically (see SI for detail). We assumed [CoA] to be 10 nM,38 and its association rate with ERα-BPA to be comparable to that of the binding of E2 with ERα.39 We further assumed that the dissociation rate of the ERα-BPA-CoA complex to be 60× slower than the dissociation of ERα-BPA. A sixty-fold stabilization is not unreasonable.34 The result of our modeling is shown in Figure 5 (red curve) when the percentage of activated receptors (ERα-BPA-CoA) is plotted against time. As can be seen, that fraction reached the potent 0.1% if the surge persisted for 7 min. If the surge should subside after 10 min, then the fraction of activated receptors would remain ≥0.1% for another 10 min (not shown in Figure 5). On the contrary, if BPA were a slower binder, say by 30× , even if its binding affinity were the same, then the activated fraction would only be 0.01% at the 7 min point, and never more than 0.02% throughout the 10 min surge (green curve). We should also point out that the on-rate of 1.1 × 104 M−1 −1 s for BPA was only a lower bound. This is because its measured half-life of 0.4 min was an upper bound. For that reason, BPA could bind even faster. Reasoned this way, we might be able to explain the significant activation of transcription induced by BPA,27,40 as well as the marked enhancement in its transactivity when coactivator concentration was increased.41 The measurement limit of 0.4 min half-life also applied to a few other ligands; their labels were underlined in Figure 4. Three such ligands were PFOS, PFNA, and PFOA. Significant gene reporter activity induced by these perfluoroalkyls was already reported.25 In conclusion, we measured the kinetic rate constants of the binding of ERα with 19 xenoestrogens using fluorescence polarization. Their RBAs were shown to be consistent with published values. On and off rates were correlated with the molecular structure of the ligands. Interesting systematics was revealed. Low off-rate presumably signified potent antiestrogens. Among the phytoestrogens, genistein, coumestrol, and zearalenone had off-rates as low as that of tamoxifen. They might serve as lead compounds for antiestrogen synthesis. On the contrary, higher on-rate might correlate with stronger estrogenic effects. The plasticizer BPA had an on-rate higher than 104 M−1 s−1 which was comparable to that of strong binders such as the three phytoestrogens mentioned. This might explain why BPA at nM concentration could induce measurable gene transcription in cells. Our kinetic assay also allowed us to measure low RBA values, down to 10−4 range even for low solubility lipophilic ligands such as DBP, pp′-DDT, and pp′-DDE. Two further studies should be pursued. First, the binding of ERα and BPA in the presence of coactivators should be investigated. In our model, we assumed that coactivator recruitment would stabilize the ERα-BPA complex.34 This assumption needs to be validated. Second, the binding kinetics of additional xenoestrogens should be measured. For example,

As for the shortest lived ligands at the bottom, their half-life of 0.4 min was an upper bound. As explained previously, measurable half-life was limited by the half-life of F. On-Rate and Molecular Structure. We mentioned the importance of long antagonist residence time for effective blockage of ERα from binding with endogenous estrogens. To elicit estrogenic response, however, high on-rates could be more important.33 A fast binding agonist, even if short-lived, could exploit a transient surge of its concentration to induce downstream effects. Moreover, the recruitment of a coactivator might stabilize the ERα-ligand complex, thus making a shortlived complex longer lived.34 For these reasons, it is informative to correlate k̃on and molecular structure, especially for endocrine disruptors. The correlation is shown in Figure 4. The color coding of the ligand labels is the same as Figure 3. The on-rates spanned 5 orders of magnitude, ranging from 70 M−1 s−1 to 1.9 × 106 M−1 s−1. Evidently, the fastest binders were slim, linear stuctures. Side chains apparently slowed down the binding. Unlike the half-life ranking shown in Figure 3, E2 was now ahead of the antagonists. In a similar way, the more linear α-ZEA and ZEA were ahead of GEN. The case of BPA was the most interesting. Its RBA was 1500× lower than that of E2, yet its binding speed was only 90× lower. It was therefore conceivable that BPA could elicit estrogenic response whenever there was a surge in its concentration. We numerically modeled the effect of such a surge (see SI for modeling detail). The results are shown in Figure 5. A stepwise

Figure 5. Numerical modeling of ligand−receptor binding events upon a transient surge of BPA. Black trace: [BPA] surged from a constant 1 pM background to 1 nM at t = 0. Blue trace: % of the receptor ERα occupied by BPA, as a function of time. Gray trace: same ̃ of BPA-ERα interaction ̃ and koff as the blue trace but with kon decreased by 30×. Red trace: % ERα complexed with BPA and coactivator (CoA). CoA recruitment was assumed to lengthen the lifetime of the ERα-BPA-CoA complex by 60×. Green trace: same as ̃ of BPA-ERα interaction decreased by ̃ and koff the red trace but with kon 30×.

increase of [BPA] from 1 pM to 1 nM at time t = 0 was supposed (black trace). The subsequent change in the percentage of occupied ERα is plotted as a function of time for various scenarios. The blue trace represents the case of twobody interaction when BPA binds with ERα. The kinetic rate coefficients listed in Table 1 were used in the model. A [ERα] of nM range was assumed.35,36 As can be seen, before the surge, the fraction of receptors occupied was less than 0.0005%. One 11597



Article

(14) Kwok, K. C. Measuring binding kinetics of ligands with tethered receptors by fluorescence polarization complemented with total internal reflection fluorescence microscopy. PhD Dissertation, Hong Kong Baptist University, Hong Kong, 2010. Downloadable from http://libproject.hkbu.edu.hk/was40/detail?record=11&channelid= 48892&searchword= %28aTitle%3D%27receptor%27+or+aAuthor%3D%27receptor%27+ or+aDept%3D%27receptor%27+or+aSubject%3D%27receptor%27+ or+aDesc%3D%27receptor%27%29&sortfield=%2BaTitle. (15) Life Technologies product manual titled “PolarScreen ER Alpha Competitor Assay, Green”. Catalog nos. A15882, A15883. Downloadable from http://tools.lifetechnologies.com/content/sfs/manuals/ polarscreen_er_alpha_green_man.pdf. (16) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 3rd ed.; Springer: New York, 2006. (17) Cheng, Y.; Prusoff, W. H. Relationship between the inhibition constant (KI) and the concentration of inhibitor which causes 50% inhibition (I50) of an enzymatic reaction. Biochem. Pharmacol. 1973, 22, 3099−108. (18) Ohno, K.; Suzuki, S.; Fukushima, T.; Maeda, M.; Santa, T.; Imai, K. Study on interactions of endocrine disruptors with estrogen receptor using fluorescence polarization. Analyst 2003, 128, 1091− 1096. (19) Waller, C. L.; Oprea, T. I.; Chae, K.; Park, H.-K.; Korach, K. S.; Laws, S. C.; Wiese, T. E.; Kelce, W. R.; Gray, L.E., Jr. Ligand-based identification of environmental estrogens. Chem. Res. Toxicol. 1996, 9, 1240−1248. (20) Andersen, H. R.; Andersen, A.-M.; Arnold, S. F.; Autrup, H.; Barfoed, M.; Beresford, N. A.; et al. Comparison of short-term estrogenicity tests for identification of hormone-disrupting chemicals. Environ. Health. Perspect. 1999, 107 (Suppl 1), 89−108. (21) National Institute of Health. Current status of test methods for detecting endocrine disruptors: in vitro estrogen receptor binding assaysa background review document. NIH Publication No 034504, October 2002. (22) Lee, H. K.; Kim, T. S.; Kim, C. Y.; Kang, I. H.; Kim, M. G.; Jung, K. K.; Kim, H. S.; Han, S. Y.; Yoon, H. J.; Rhee, G. S. Evaluation of in vitro screening system for estrogenicity: comparison of stably transfected human estrogen receptor-α transcriptional activation (OECD TG455) assay and estrogen receptor (ER) binding assay. J. Toxicol. Sci. 2012, 37, 431−437. (23) Huang, P.; Chandra, V.; Rastinejad, F. Structural overview of the nuclear receptor superfamily: Insights into physiology and therapeutics. Annu. Rev. Physiol. 2010, 72, 247−272. (24) Celik, L.; Lund, J. D.; Schiott, B. Conformational dynamics of the estrogen receptor alpha: Molecular dynamics simulations of the influence of binding site structure on protein dynamics. Biochemistry 2007, 46, 1743−1758. (25) Benninghoff, A. D.; Bisson, W. H.; Kock, D. C.; Ehresman, D. J.; Kolluri, S. K.; Williams, D. E. Estrogen-like activity of perfluroalkyl acids in vivo and interaction with human and rainbow trout estrogen receptors in vitro. Toxicol. Sci. 2011, 120, 42−58. (26) Mueller, S.; Simon, S.; Chae, K.; Metzler, M.; Korach, K. S. Phytoestrogens and their human metabolites show distinct agonistic and antagonistic properties on estrogen receptor α (ERα) and ERβ in human cells. Toxicol. Sci. 2004, 80, 14−25. (27) Hiroi, H.; Tsutsumi, O.; Momoeda, M.; Takai, Y.; Osuga, Y.; Taketani, Y. Differential interactions of bisphenol A and 17β-estradiol with estrogen receptor α (ERα) and ERβ. Endocrine J. 1999, 46, 773− 778. (28) Seo, H.-S.; DeNardo, D. G.; Jacquot, Y.; Laios, I.; Vidal, D. S.; Zambrana, C. R.; Leclercq, G.; Brown, P. H. Stimulatory effect of genistein and apigenin on the growth of breast cancer cells correlates with their ability to activate ER alpha. Breast Cancer Res. Treat. 2006, 99, 121−134. (29) Li, J.; Li, N.; Ma, M.; Giesy, J. P.; Wang, Z. In vitro profiling of the endocrine disrupting potency of organochlorine pesticides. Toxicol. Lett. 2008, 183, 65−71.

the op′ version of the pesticides should be included. They were known to bind ERα with higher affinity than their pp′ isomers.42

■

ASSOCIATED CONTENT

S Supporting Information *

Effect of a transient surge of bisphenol A. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +852 3411-7034; fax: +852 3411-5813; e-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the General Research Fund of the Research Grants Council of Hong Kong under Grant No. HKBU 200508 and the Faculty Research Grants of Hong Kong Baptist University under Grant No. FRG1/12-13/020.

■

REFERENCES

(1) Zhang, W.; Luo, Y.; Zhang, L.; Cai, Q.; Pan, X. Known and emerging factors modulating estrogenic effects of endocrine-disrupting chemicals. Environ. Rev. 2014, 22, 87−98. (2) Vandenberg, L. N.; Maffini, M. V.; Sonnenschein, C.; Rubin, B. S.; Soto, A. M. Bisphenol-A and the great divide: A review of controversies in the field of endocrine disruption. Endocrine. Rev. 2009, 30, 75−95. (3) Teeguarden, J.; Hanson-Drury, S.; Fisher, J. W.; Doerge, D. R. Are ypical human serum BPA concentrations measurable and sufficient to be estrogenic in the general population? Food Chem. Toxicol. 2013, 62, 949−963. (4) Vandenberg, L. N.; Colborn, T.; Hayes, T. B.; Heindel, J. J.; Jacobs, D. R., Jr.; Lee, D.-H.; et al. Hormones and endocrinedisrupting chemicals: low-dose effects and nonmonotonic dose responses. Endocrine. Rev. 2012, 33, 1−78. (5) Alonso-Magdalena, P.; Ropero, A. B.; Soriano, S.; Garcia-Arevalo, M.; Ripoll, C.; Fuentes, E.; Quesada, I.; Nadal, A. Bisphenol-A acts as a potent estrogen via non-classical estrogen triggered pathways. Mol. Cell. Endocrinol. 2012, 355, 201−207. (6) Baker, M. E.; Chandsawangbhuwana, C. 3D models of MBP, a biologically active metabolite of bispehnol A, in human estrogen receptor α and estrogen receptor β. PLoS One 2012, 7, e46078. (7) Fang, H.; Tang, W.; Perkins, R.; Soto, A. M.; Prechtl, N. V.; Sheehan, D. M. Quantitative comparisons of in vitro assays for estrogenic activities. Environ. Health. Perspect. 2000, 108, 723−729. (8) Copeland, R. A.; Pompliano, D. L.; Meek, T. D. Drug-target residence time and its implications for lead optimization. Nat. Rev. Drug Discovery 2006, 5, 730−739. (9) Pan, A. C.; Borhani, D. W.; Dror, R. O.; Shaw, D. E. Molecular determinants of drug-receptor binding kinetics. Drug Discovery Today 2013, 18, 667−673. (10) Zhang, R.; Monsma, F. The importance of drug-target residence time. Curr. Opin. Drug Discovery Devel. 2009, 12, 488−496. (11) Copeland, R. A. The dynamics of drug-target interactions: Drugtarget residence time and its impact on efficacy and safety. Expert Opin. Drug Discovery 2010, 5, 305−310. (12) Lu, H.; Tonge, P. J. Drug-target residence time: critical information for lead optimization. Curr. Opin. Chem. Biol. 2010, 14, 467−474. (13) Kwok, K.-C.; Cheung, N.-H. Measuring binding kinetics of ligands with tethered receptors by fluorescence polarization and total internal reflection fluorescence. Anal. Chem. 2010, 82, 3819−3825. 11598



Article

determinants of tissue selectivity in estrogen receptor modulators. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 14105−14110.

(30) Pike, A. C. W. Lessons learnt from structural studies of the estrogen receptor. Best Pract. Res. Clin. Endocrinol. Metab. 2006, 20, 1− 14. (31) Naciff, J. M.; Overmann, G. J.; Torontali, S. M.; Carr, G. J.; Tiesman, J. P.; Daston, G. P. Impact of the phytoestrogen content of laboratory animal feed on the gene expression profile of the reproductive system in the immature female rat. Environ. Health Perspect. 2004, 112, 1519−1526. (32) Mense, S. M.; Hei, T. K.; Ganju, R. K.; Bhat, H. K. Phytoestrogens and breast cancer prevention: Possible mechanisms of action. Environ. Health Perspect. 2008, 116, 426−433. (33) Vauquelin, G. Rebinding: Or why drugs may act longer in vivo than expected from their in vitro target residence time. Expert Opin. Drug Discovery 2010, 5, 927−941. (34) Gee, A. C.; Carlson, K. E.; Martini, P. G. V.; Katzenellenbogen, B. S.; Katzenellenbogen, J. A. Coactivator peptides have a differential stabilizing effect on the binding of estrogens and antiestrogens with the estrogen receptor. Mol. Endocrinol. 1999, 13, 1912−1923. (35) Furlow, J. D.; Murdoch, F. E.; Gorski, J. High affinity binding of the estrogen receptor to a DNA response element does not require homodimer formation or estrogen. J. Biol. Chem. 1993, 268, 12519− 12525. (36) For [ERα] ranging from pM to tens of nM, the percentage of ERα occupied was the same to two significant figures. (37) Welshons, W. V.; Thayer, K. A.; Judy, B. M.; Taylor, J. A.; Curran, E. M.; vom Saal, F. S. Large effects from small exposures. I. Mechanisms for endocrine disrupting chemicals with estrogenic activity. Environ. Health Perspect. 2003, 111, 994−1006. (38) Kd of ERα-E2 + CoA was in the tens of nM range, see PanVera Era-CoA assay 2002. Downloadable from http://tools.lifetechnologies. com/content/sfs/manuals/L0977.pdf. It is plausible that endogenous [CoA] is also in that range. (39) Paal, K.; Baeuerle, P. A.; Schmitz, M. L. Basal transcription factors TBP and TFIIB and the viral coactivator E1A 13S bind with distinct affinities and kinetics to the transactivation domain of NF-κB p65. Nucleic Acid Res. 1997, 25, 1050−1055. (40) Kuiper, G. G. J. M.; Lemmen, J. G.; Carlsson, B.; Corton, J. C.; Safe, S. H.; Van Der Saag, P. T.; Van Der Burg, B.; Gustafsson, J.-A. Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor β. Endocrinol. 1998, 139, 4252−4263. (41) Jeyakumar, M.; Carlson, K. E.; Gunther, J. R.; Katzenellenbogen, J. A. Exploration of dimensions of estrogen potency: Parsing ligand binding and coactivator binding affinities. J. Biol. Chem. 2011, 286, 12971−12982. (42) Shanle, E. K.; Xu, W. Endocrine disrupting chemicals targeting estrogen receptor signaling: Identification and mechanisms of action. Chem. Res. Toxicol. 2011, 24, 6−19. (43) Bolger, R.; Wiese, T. E.; Ervin, K.; Nestich, S.; Checovich, W. Rapid screening of environmental chemicals for estrogen receptor binding capacity. Environ. Health Perspect. 1998, 106, 551−557. (44) Clegg, N. J.; Paruthiyil, S.; Leitman, D. C.; Scanlan, T. S. Differential response of estrogen receptor subtypes to 1,3-diarylindene and 2,3-diarylindene ligands. J. Med. Chem. 2005, 48, 5989−6003. (45) Matthews, J.; Celius, T.; Halgren, R.; Zacharewski, T. Differential estrogen receptor binding of estrogenic substances: A species comparison. J. Steroid Biochem. Mol. Biol. 2000, 74, 223−234. (46) Wijayaratne, A. L.; Nagel, S. C.; Paige, L. A.; Christensen, D. J.; Norris, J. D.; Fowlkes, D. M.; McDonnell, D. P. Comparative analysis of mechanistic differences among antiestrogens. Endocrinol. 1999, 140, 5828−5840. (47) Blair, R. M.; Fang, H.; Branham, W. S.; Hass, B. S.; Dial, S. L.; Moland, C. L.; Tong, W.; Shi, L.; Perkins, R.; Sheehan, D. M. The estrogen receptor relative binding affinities of 188 natural and xenochemicals: structural diversity of ligands. Toxicol. Sci. 2000, 54, 138−153. (48) Grese, T. A.; Sluka, J. P.; Bryant, H. U.; Cullinan, G. J.; Glasebrook, A. L.; Jones, C. D.; Matsumoto, K.; Palkowitz, A. D.; Sato, M.; Termine, J. D.; Winter, M. A.; Yang, N. N.; Dodge, J. A. Molecular 11599


Measuring the Kinetics of the Binding of ... - ACS Publications

Recommend Documents