http://www.diva-portal.org

This is the published version of a paper published in Environmental Science and Technology.

Citation for the original published paper (version of record):

Zhang, J., Begum, A., Brännström, K., Grundström, C., Iakovleva, I. et al. (2016)Structure-based Virtual Screening Protocol for in silico Identification of Potential ThyroidDisrupting Chemicals Targeting Transthyretin.Environmental Science and Technology, 50(21): 11984-11993https://doi.org/10.1021/acs.est.6b02771

Access to the published version may require subscription.

N.B. When citing this work, cite the original published paper.

Permanent link to this version:http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-125630

Structure-Based Virtual Screening Protocol for in Silico Identificationof Potential Thyroid Disrupting Chemicals Targeting TransthyretinJin Zhang,† Afshan Begum,† Kristoffer Brannstrom,‡ Christin Grundstrom,† Irina Iakovleva,‡

Anders Olofsson,‡ A. Elisabeth Sauer-Eriksson,† and Patrik L. Andersson*,†

†Department of Chemistry and ‡Department of Medical Biochemistry and Biophysics, Umea University, SE-901 87 Umea, Sweden

*S Supporting Information

ABSTRACT: Thyroid disruption by xenobiotics is associatedwith a broad spectrum of severe adverse outcomes. Onepossible molecular target of thyroid hormone disruptingchemicals (THDCs) is transthyretin (TTR), a thyroidhormone transporter in vertebrates. To better understandthe interactions between TTR and THDCs, we determinedthe crystallographic structures of human TTR in complex withperfluorooctanesulfonic acid (PFOS), perfluorooctanoic acid(PFOA), and 2,2′,4,4′-tetrahydroxybenzophenone (BP2). Themolecular interactions between the ligands and TTR werefurther characterized using molecular dynamics simulations. Astructure-based virtual screening (VS) protocol was developedwith the intention of providing an efficient tool for thediscovery of novel TTR-binders from the Tox21 inventory. Among the 192 predicted binders, 12 representatives were selected,and their TTR binding affinities were studied with isothermal titration calorimetry, of which seven compounds had bindingaffinities between 0.26 and 100 μM. To elucidate structural details in their binding to TTR, crystal structures were determined ofTTR in complex with four of the identified compounds including 2,6-dinitro-p-cresol, bisphenol S, clonixin, and triclopyr. Thecompounds were found to bind in the TTR hormone binding sites as predicted. Our results show that the developed VS protocolis able to successfully identify potential THDCs, and we suggest that it can be used to propose THDCs for future toxicologicalevaluations.

■ INTRODUCTION

Thyroid hormone disrupting chemicals (THDCs) are anthro-pogenic compounds that disrupt the homeostasis of the thyroidhormone (TH) system1 and can induce disorders inphysiological processes, including macronutrient metabolism,energy balance, brain development, and reproduction.2

Exposure to THDCs poses a significant threat to humanhealth, especially during fetal development. Prenatal hypo-thyroxinemia induced by THDCs can impair the developmentof the embryonic brain and lead to neurologic deficits such asvisuo-spatial processing difficulties3 and reduced learning andmemory.4

Transthyretin (TTR), a homotetrameric serum protein thattransports L-thyroxine (T4),5 has been suggested to be apossible molecular target of THDCs such as per- andpolyfluoroalkyl substances (PFASs)6 and hydroxylated poly-chlorinated biphenyls (OH-PCBs).7 TTR is the primary THtransporter in developing rodents,8 fish, birds, and amphibians,9

and THDCs can disrupt TH transport by competitively bindingto the two identical thyroxine binding sites (TBSs) situated atthe dimeric interface of TTR.10,11 TTR can also mediatedelivery of THDCs across the placental and blood-brain barrierinto the fetus and brain.12−14 Most TTR in human plasmaresides in the apo form, and the concentration of tetrameric

TTR in plasma (4−7 μM)15,16 is significantly higher than thatof T4 (0.112 μM).17 This means that low affinity TTR-binderscan, without competing with T4, bind to TTR and betransported to vital organs where they may subsequentlycause adverse health effects including developmental disor-ders.18,19 Inuit populations that are chronically exposed torelatively high levels of TTR-binding contaminants showednormal T4 levels, but TTR-assisted accumulation of THDCscould potentially cause developmental disorders.20−22 It is thuscritical to understand the molecular interactions betweenTHDCs and TTR and more importantly to identify andreduce the exposure to THDCs that target TTR.Virtual screening (VS) using molecular docking is an efficient

means to reduce costs and animal testing by identifying themost hazardous compounds in risk assessment processes orprioritizing the most promising candidates in drug discov-ery.23,24 The method has been used to propose novel TTRamyloid inhibitors25 and to identify environmental pollutantstargeting estrogen receptors.26,27 Molecular docking can give

Received: June 3, 2016Revised: September 23, 2016Accepted: September 26, 2016Published: September 26, 2016

Article

pubs.acs.org/est

© 2016 American Chemical Society 11984 DOI: 10.1021/acs.est.6b02771Environ. Sci. Technol. 2016, 50, 11984−11993

This is an open access article published under an ACS AuthorChoice License, which permitscopying and redistribution of the article or any adaptations for non-commercial purposes.

insights into ligand-protein binding conformations, and this wasused to reveal interactions between PFASs and human liverfatty-acid binding proteins.28 Such studies provide a betterunderstanding of critical molecular interactions betweenhazardous compounds and their targets.Crystal structures of ligand-protein complexes are an

essential component in the development of VS protocols.Over 200 crystal structures of TTR have been reported in theProtein Data Bank (PDB),29,30 but only a few of these are incomplex with environmental pollutants or their metabolites.These include TTR complexes with pentabromophenol31 andOH-PCBs.10 In addition, we recently determined the crystalstructure of TTR with the brominated flame retardanttetrabromobisphenol A (TBBPA).32 Additional crystal struc-tures of TTR complexes with environmental pollutants withlarge chemical variation would improve the development of aVS protocol for environmental pollutants and increase ourunderstanding of their interactions with TTR at the molecularlevel.In this study, we determined the X-ray structures of human

TTR in complex with three emerging contaminants that wereshown to bind to TTR: perfluorooctanoic acid (PFOA),perfluorooctanesulfonic acid (PFOS), and 2,2′,4,4′-tetrahydrox-ybenzophenone (BP2).6,33 PFOA and PFOS are commonlyused as stain, water, and grease repellents in food packaging,carpets, and textiles as well as in fire-fighting foams.34 BP2 isused as a UV absorber in personal care products and plasticmaterials.35 Exposure to PFOS and PFOA is associated withhypothyroidism and an increasing incidence of osteoporo-sis.36−38 BP2 might cause the development of endometriosis.39

These ligands together with TBBPA cover a wide chemicalvariation of TTR ligands. Their TTR binding activities inhuman plasma were measured with a previously establishedplasma assay.32,40 The ligand-TTR interactions in the X-raystructures were further studied by performing moleculardynamics (MD) simulations, and residues important for theinteractions were proposed by decomposition of ligand-bindingfree energies using the Molecular Mechanics-Generalized BornSurface Area (MM-GBSA) method. Based on the novel X-raystructures and the molecular interaction information, astructure-based VS protocol was developed to identifyTHDCs targeting TTR from the Tox21 inventory of 7,849organic compounds.41 Twelve representative industrial com-pounds were selected for in vitro studies using isothermaltitration calorimetry (ITC). Of these, four hits have beencocrystallized with TTR to study their binding conformations.

■ MATERIALS AND METHODSCrystallization of the TTR-THDC Complexes. Human

wild-type TTR was purified and crystallized in complex withPFOA, PFOS, BP2, and with four confirmed hits (bisphenol S(BPS), clonixin, 2,6-dinitro-p-cresol (DNPC), and triclopyr)following the procedures described previously.42,43 The purifiedTTR was dialyzed against 10 mM Na-phosphate buffer with100 mM KCl (pH 7.6) and concentrated to 5 mg/mL using anAmicon Ultra centrifugal filter device (Millipore, 3 kDamolecular-weight cutoff) and cocrystallized at room temper-ature with a 5-molar excess of the compounds using the vapor-diffusion hanging drop method. A drop containing a 3 μLprotein solution was mixed with 3 μL of precipitant andequilibrated against a 1 mL reservoir solution containing 1.3−1.6 M sodium citrate and 3.5% v/v glycerol at pH 5.5 in 24-wellLinbro plates. Crystals grew to dimensions of 0.2 × 0.2 × 0.3

mm3 after 5 days. The crystals were cryoprotected with 12% v/v glycerol.

Crystallographic Data Collection, Integration, andStructure Determination. The X-ray diffraction data of thecomplexes were collected at the European SynchrotronRadiation Facility (Grenoble, France) on beamline ID29 orID23-1 and in-house instruments (Tables S1 and S2). Thestructure of human TTR (PDB ID: 1F415) and X-ray data from38.2−2.5 Å resolution were used in molecular replacementsearches with the program PHASER.44 The models wererefined against all of the diffraction data using PHENIX.45 Atthe end of structure refinement, anisotropic B-factors wererefined. Manual map inspection was performed with COOT.46

The details of the refinement statistics are shown in Tables S1and S2. Molecular graphics were produced using CCP4 mg.47

Structure factors and coordinates of the TTR complexes havebeen deposited with the Protein Data Bank (PDB ID: 5JID(PFOA), 5JIM (PFOS), 5JIQ (BP2), 5L4F (DNPC), 5L4I(clonixin), 5L4J (BPS), and 5L4M (triclopyr)).

Molecular Dynamics Simulations. The MD simulationswere prepared based on the three structures of TTR complexeswith BP2, PFOA (sharing similar binding profiles with PFOS),and TBBPA.32 Each system was prepared following theprocedures described in the Supporting Information. TheMD simulations of each prepared system were performed usingAmber 1448 with the following steps: 1) each system wasminimized with two-stage energy minimizations, 2) eachsystem was gradually heated to 310 K and then equilibratedunder NPT conditions for 1 ns, and 3) the 20 ns productionMD was performed under NPT conditions using a Langevinthermostat and Berendsen barostat. Details of the simulationare described in the Supporting Information.

MM-GBSA Free-Energy Decomposition. We calculatedthe binding free energies of BP2, PFOA, and TBBPA to TTRusing the MM-GBSA method in AmberTools 1548 based on thelast 10 ns of the MD trajectory. The contributions of eachresidue in the TBS to the ligand binding were determined bythe pairwise decomposition method. The five residues thatcontributed the most to the enthalpic binding free energy ofeach ligand were considered to be important for ligand-TTRinteractions (Table S3). The entropy contributions wereneglected due to low prediction accuracy.49 The proceduresand parameters are described in the Supporting Information.

Collection and Preparation of the Ligand Data Set.Two ligand sets were used in the study − the TTRbenchmarking set and the Tox21 inventory.41 The TTRbenchmarking set (Table S4) contained 155 TTR-binders and10,197 in silico generated decoys (inactives).50 The TTR-binders were collected from the scientific literature using anactivity cutoff of Ki or IC50 ≤ 100 μM as suggested by the USEPA in the studies of THDCs51 and estrogen disruptors26,52

and used in the Tox21Challenge in silico model-developingexercise (more explanations given in the SupportingInformation).53 The benchmarking set was used in thedevelopment and evaluation of the VS protocol. The Tox21data set contains 7,849 industrial compounds and pharmaceut-icals.54 For each compound in the two sets, a maximum of 32low-energy conformations were generated by considering allchiral atoms using the Schrodinger LigPrep module under theOPLS_2005 force field,55 and their ionization and tautomericstates were determined at pH 7.0 ± 1.0.

Molecular Docking. Molecular docking models weredeveloped using the TTR complex structures with BP2,

Environmental Science & Technology Article

DOI: 10.1021/acs.est.6b02771Environ. Sci. Technol. 2016, 50, 11984−11993

11985

PFOA, and TBBPA. The TBS situated at the BB′ interface wasdescribed by a grid box centered on its coligand. The preparedcompounds in the TTR benchmarking set were docked intoeach structure using the Schrodinger Glide module understandard precision.56

For the docking results, we assessed their VS performanceand enrichment. The VS performance of each model wasdescribed with the area under curve (AUC) value of thereceiver operating characteristic (ROC) curves. The AUC valuewas used because it is insensitive to the ratio of actives versusdecoys in the benchmarking set.57 The enrichment wascharacterized by two enrichment factors (EFs) that wereevaluated at 10% and 20% of the ranked database, referred to asEF10% and EF20%, respectively. The AUC values and EFs aregiven in Table S5.Ligand-TTR Interaction Analysis. Residues important for

ligand-TTR interactions were identified using the MM-GBSAbinding free-energy decomposition. The W188 water moleculein the TTR-TBBPA complex was also considered because itwas suggested to be critical for mediating hydrogen bonds (H-bonds) between TBBPA and Ser117.32 For each initial hitidentified by molecular docking, we counted the number ofelectrostatic interactions (H-bonds and salt bridges) with polarresidues (including W188) and hydrophobic interactions withnonpolar residues based on their docking poses. The number ofinteractions together with the docking score was used to refinethe initial hits and identify the bioactive compounds from theTox21 inventory. The refined hits and their interactions with

TTR are shown in Table S6. The known TTR-bindersidentified among the refined hits are given in Table S7.

Measuring Compound Binding Activity in HumanPlasma. Binding activity of BP2, PFOS, and PFOA to TTR inthe presence of human plasma was measured with a previouslyestablished plasma assay.40 Details of experimental proceduresare given in the Supporting Information.

Isothermal Titration Calorimetry. The binding affinitiesof 12 selected potential TTR-binders were measured using anAuto-iTC200 (MicroCal, Malvern, UK) at 25 °C. The chemicalinformation and properties of the 12 compounds are shown inTable S8. Stock solutions were prepared in dimethyl sulfoxide(DMSO) and diluted with phosphate-buffered saline (PBS, pH= 7.4) to a final level of 5% DMSO immediately before the ITCexperiments.40 Aliquots of each compound were titrated intothe TTR buffer solution (30 or 90 μM) to reach a 10-foldmolar excess. For the control experiment, the compounds weretitrated into the cell with only buffer. Raw data were collected,subtracted for compound heats of dilution, and processed usingthe MicroCal Origin software. Calorimetric data (Table S9)were plotted and fitted using the standard single-site bindingmodel to yield the binding affinities (Kd), enthalpy changes,and entropy changes.

■ RESULTS AND DISCUSSIONCrystal Structures of the TTR-THDC Complexes. We

cocrystallized human wild-type TTR in complex with PFOA,PFOS, and BP2 and determined their structures at 1.2−1.4 Åresolution (Table S1). The structures showed that all three

Figure 1. (A) The TTR monomers in the dimer structure are shown as ribbons and are labeled A and B. The symmetry-related monomers arelabeled A′ and B′. Locations of the thyroxine binding sites (TBSs) are shown as gray surfaces. (B) Close-up view of the crystallographic bindingconformations of (a) BP2, (b) PFOA, and (c) PFOS in the TBS. The refined (2|Fo|−|Fc|) electron density is shown in gray mesh at the 1σ level, andthe anomalous difference map showing the position of the sulfur atoms in PFOS is shown in violet mesh at the 3σ level. The water molecules presentaround the ligands are shown as red spheres.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.6b02771Environ. Sci. Technol. 2016, 50, 11984−11993

11986

ligands bind at the deep end of the TBS (Figure 1). Comparedwith T4, the three ligands have the ability to bind deeper in thepocket due to their smaller molecular size (Figure S1).Of particular interest are the TTR complex structures with

PFOA and PFOS. The structures share three unique features.1) The two ligands bind with their hydrophobic tail inside thepocket, and their hydrophilic acid groups are surface exposed(Figure S2), which is different from our previous assumption.33

2) The hydrophobic tail bends at the second to last carbonatom of the ligand to accommodate the volume between twoSer117 residues and to allow the acid groups to form saltbridges with Lys15. 3) The bent hydrophobic tail forces thehydroxyl groups on Ser117 to rotate away from the ligands(Figure 1), which creates a more hydrophobic pocket thatfavors the ligand-TTR interactions.BP2 interacts with TTR similarly as luteolin in the deep end

of the TBS (Figure S2).58 Two hydroxyl groups form hydrogenbonds with Ser117 residues, and the carbonyl group interactswith Thr119. Interactions between the aromatic ring of BP2and TTR involve van der Waal contacts with the side chain ofLeu17.MD Simulations for Ligand-TTR Interaction Studies.

We investigated the molecular interactions between TTR andTHDCs by performing MD simulations on the TTR complexstructures with BP2, PFOA, and TBBPA.32 All three ligandsremained bound in the TBS, and the root-mean-squared-deviation (RMSD) of the protein and ligands was under 2 Åthroughout the simulations (Figure S3 and Figure S4).The ligand-TTR binding free energies were calculated using

the MM-GBSA method based on the last 10 ns of the MDtrajectory. The calculated energies of the ligands correlate wellwith their previously measured potencies (Figure S5).6,33,59

The energies were further decomposed to reveal five residues ofsignificance for the ligand-TTR interactions in each of the threecomplex structures (Table S3). Two types of ligand-TTRinteractions were identified electrostatic interactions with thepolar residues B-Thr119, B-Ser117, B′-Ser117, B-Lys15, and B′-Lys15 and hydrophobic interactions with the nonpolar residuesB-Leu110, B-Leu17, B′-Leu17, and B-Ala108. The electrostaticinteractions account for specific ligand-TTR recognition and alarge proportion of the binding free energies,60−62 whereas thehydrophobic interactions were less significant for ligandbinding. The results also revealed that the three ligands interactwith the residues deeper in the TBS, e.g. Ser117, whereas T4interacts with residues at the opening of the TBS, e.g. Glu54.63

Information on the molecular interactions with the identifiedresidues (Table S3) was used to better understand the ligand-TTR interaction and to refine potential TTR-binders for betteridentifying of bioactive compounds.Molecular Docking Model. The accuracy of the molecular

docking using Glide was assessed by comparing the dockingconformations of BP2, PFOA, and TBBPA with their X-raystructures.33 The RMSD values for the coligands were lowerthan 2 Å (Table S5), indicating that the molecular docking wasable to reproduce the actual binding conformation. Theprepared compounds in the benchmarking set were dockedinto each TTR complex, and the docking results werecompared in terms of AUC values and enrichment factors(Table S5). The TTR-TBBPA structure showed the best resultswith AUC of 0.75 and good early enrichment (EF10% = 3.76and EF20% = 2.95), and 141 binders were identified among the155 binders in the benchmark set (Table S5). The dockingmodels based on the TTR-PFOA and TTR-BP2 structures had

lower performances in identifying TTR-binders, probablybecause TBBPA best reflects the majority of the halogenatedand hydroxylated aromatic compounds in the benchmark set.However, the two structures provide useful information formolecular interaction studies, and the TTR-PFOA structurereveals the correct binding orientations of PFASs in the TBSwhere the acidic group was found to interact with Lys15 ratherthan Ser117.The impact of water in the TBS of the TTR-TBBPA complex

was studied by evaluating the model performance of thestructure with and without water molecules in the TBS. Thestructure with water in the TBS outperformed the one without(Table S5 and Figure S6). After examining the ligand-proteininteractions in the TTR-TBBPA structure, the W188 watermolecule was found to be critical for mediating hydrogen bondsbetween the ligand and the Ser117 residues, which agrees wellwith previously reported data.32

Binding Activity of the Studied THDCs to TTR inHuman Plasma. We compared the binding activity of BP2,PFOA, and PFOS to TTR in human plasma in terms of theirinhibitory concentrations at 50% (IC50) to prevent tetramericdissociation in the established assay.32,40 BP2 bound strongly toTTR with an IC50 of 4 μM, while PFOA (IC50 = 175 μM) andPFOS (IC50 = 103 μM) had poor binding activity to TTR inplasma (Figure S7). Their differences in binding activity couldbe caused by the hydrophobic tails of the PFASs giving limitedelectrostatic contributions to their TTR binding, which arecritical for molecular recognition.40,62 TBBPA has also shownstrong binding to TTR (IC50 = 3.3 μM).32 Clearly, BP2 sharecharacteristics with TBBPA in terms of molecular interactionswith TTR. It is thus of interest to identify contaminants sharingsimilar molecular interactions with TTR as found in BP2 andTBBPA.

Virtual Screening of Industrial Compounds. Thedocking model developed using the TTR-TBBPA structurewas used as the basis for the VS of the Tox21 inventory becauseof its high VS performance and the strong binding of TBBPA toTTR in human plasma. We applied a two-step procedure toidentify potential TTR-binders from the Tox21 inventory. Inthe first step, compounds were docked into the TTR-TBBPAstructure followed by ranking based on their standard precisiondocking score. For each industrial compound, all protonationstates in the range of pH 7.0 ± 1.0 were considered during themolecular docking. As suggested by previous studies,64,65 weselected its top scored protonation form (Tables S4 and S6),since docking score considers both binding affinity andtautomeric ratio.66,67 The top 20% scored compounds covering1,282 hits were considered as initial hits for TTR, and thesewere extracted for further analysis. In the second step, the initialhits were further refined by selecting those that showed at leastthree electrostatic interactions with the identified polar residues(including W188) and two hydrophobic interactions with thenonpolar residues. The criteria were based on the interactionsbetween TTR and TBBPA that binds to TTR with a Kd of 20nM.32 The electrostatic interactions have been suggested to becritical for specific recognition of TTR.60−62 A total of 192substances among the top 20% scored compounds also fulfilledthe molecular interaction criteria (Table S6) and thus werepredicted to be potential TTR binders by the VS protocol.Eleven known TTR-binders were identified (Table S7),including five pharmaceuticals (including diflunisal andaceclofenac),40 two flame retardants (TBBPA and tetrachlor-obisphenol A), one herbicide (2,4,5-trichlorophenoxyacetic acid

Environmental Science & Technology Article

DOI: 10.1021/acs.est.6b02771Environ. Sci. Technol. 2016, 50, 11984−11993

11987

(2,4,5-T)),33 one UV filter BP2, and one T4 derivative(3,3′,5,5′-tetraiodothyroacetic acid).68 The recognition ofthese known TTR binders by the VS protocol indicates thatit is capable of identifying potential TTR-binders for furthertoxicological investigations.Binding Affinities of Selected Compounds. We selected

12 representative compounds (Table 1) for in vitro validationfrom the VS results based on their various applications,structural diversity, and environmental relevance. Thesecompounds mainly represent herbicides and polymer additives(plasticizers and polymerization inhibitors) and includeprimarily halogenated and/or aromatic substances as well as

nitro compounds. The thermodynamic profiles of the ligand-TTR binding were investigated using ITC (Figure 2). OneTTR tetramer can bind two ligands. However, we fitted ITCthermograms (Figure 2) using a one-site binding model asbinding of the first ligand dominates the total binding energyand its affinity is approximately 100 times stronger than thesecond ligand.69,70 The ITC results showed that seven of the 12compounds bound to TTR with Kd values below 100 μM(Table 1) indicating that our VS protocol gave a reasonable hitrate of approximately 60%, which can be compared withcommonly practiced VS campaigns that generally have a hit rateof 20−40%.71,72 The thermodynamic data (Table S9) revealed

Table 1. Studied Compounds with Abbreviations, Chemical Abstract Services (CAS) Registry Numbers, Usage, and BindingAffinities (Kd) Determined Using Isothermal Titration Calorimetry

compounds abbreviations CAS usage Kd (μM)a

3,3′,5,5′-tetraiodo-L-thyronine T4 51-48-9 positive control 0.09b

2-nitro-5-(2-chloro-4-trifluoromethylphenoxy)benzoic acid acifluorfen 50594-66-6 herbicide ND2,2′-dihydroxy-4,4′-dimethoxybenzophenone BP6 131-54-4 UV absorber ND4,4′-dihydroxydiphenyl sulfone BPS 80-09-1 plasticizer 522-(3-chloro-2-methylanilino)pyridine-3-carboxylic acid clonixin 17737-65-4 pharmaceutical 0.262,6-dinitro-p-cresol DNPC 609-93-8 polymerization inhibitor 1.3diphenolic acid DPA 126-00-1 plasticizer ND2-[(4-amino-3,5-dichloro-6-fluoro-2-pyridinyl)oxy]acetic acid fluroxypyr 69377-81-7 herbicide 45L-γ-glutamyl-p-nitroanilide GPNA 67953-08-6 food additive ND2-[4-(methylsulfonyl)-2-nitrobenzoyl]-1,3-cyclohexanedione mesotrione 104206-82-8 herbicide 993,5,6-trichloro-4-aminopicolinic acid picloram 1918-02-1 herbicide 631-(4-sulfophenyl)-3-carboxy-5-pyrazolone PyT 118-47-8 dye ND3,5,6-trichloro-2-pyridinyloxyacetic acid triclopyr 55335-06-3 herbicide 4.6

aND refers to nondetected affinity. bThe binding affinity of T4 is only an approximate value due to its thermal instability and poor solubility in theisothermal titration calorimetry buffer. The approximated binding affinity is similar to the value of 84 nM reported previously.98

Figure 2. Structures of active compounds and thermograms of their binding to TTR as determined by isothermal titration calorimetry. Theabbreviations of the compounds are given in Table 1.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.6b02771Environ. Sci. Technol. 2016, 50, 11984−11993

11988

that the binding of BPS, clonixin, fluroxypyr, mesotrione, andtriclopyr was largely driven by the enthalpic component,whereas the binding energy of DNPC and picloram was mainlyentropy driven. The affinities of the active compounds were notcorrelated with their docking scores (Figure S8), which was notsurprising since docking scores generally give poor estimationsof the binding affinities.73,74 However, the trend is correct(except DNPC), as more negative docking scores are associatedwith higher affinities. Notably, five predicted binders showed noactivities in the ITC experiments (Figure S9), and we provide athorough discussion on the potential reasons for theirmispredictions in the Supporting Information.Crystal Structures of the TTR-Confirmed Hits Com-

plexes. Four confirmed TTR-binders (clonixin, DNPC,triclopyr, and BPS) were selected for cocrystallization withhuman wild-type TTR, considering their high binding affinities,diverse structures, and various applications. The crystalstructures were determined at 1.5−1.6 Å resolution (TableS2). Compared with T4, the four ligands bind deeper in theTBS due to their smaller molecular size (Figure 3). Clonixininteracts with TTR similarly as flufenamic acid.75 It forms H-bonds and salt bridges with Lys15 (Figure S10). DNPCpresents different binding conformations in the two TBSs.Furthermore, in the AA′ interface, two DNPC molecules bindinto the TBS simultaneously where one DNPC interacts withSer117 at the deep end of TBS, whereas the other DNPC bindsto Lys15 at the cavity entrance (Figure S10). In the BB′interface, DNPC also binds in two different orientations;however, these conformations are different from the ones inAA′, and both cannot be occupied at the same time as theirbinding sites overlap (Figure 3). The multiple conformations ofDNPC give better knowledge on interactions between TTR

and nitro compounds, which provides guidance for moleculardesign and risk assessment. Triclopyr interacts with TTR byhaving electrostatic interactions with Lys15. BPS binds in theTBS similarly as TBBPA.32 Its hydroxyl groups interact withLys15 and form water-bridged H-bonds with Thr119 andSer117 (Figure S8). We compared the docking pose of eachligand with its crystal conformation in the BB′ interface (FigureS11), and the results showed that we correctly predicted thebinding conformations for clonixin, DNPC, and triclopyr. Theinaccurate prediction of BPS may be due to its weaker affinityand high mobility in the TBS. Besides DNPC, EMD21388 wasalso reported to adapt to distinctive binding-modes in the twoTBSs.76 Ensemble docking has been proposed as a suitablestrategy for predicting conformation of such ligands,77,78 whichhowever was not completed here due to lack of TTR structureswith similar binding-modes to DNPC.

Environmental Implications of the Potential THDCs.Seven compounds with diverse structures and variousapplications were identified as TTR-binders (Table 1). Thesecompounds warrant further toxicological investigations consid-ering the following two reasons: (1) humans may be chronicallyexposed to the compounds at doses that are similar orsignificantly higher than the normal T4 level (0.112 μM).17 Invivo levels of the compounds have been reported to be 35844μM (9.416 mg/mL) for clonixin,79 0.312 μM (0.08 μg/mL) fortriclopyr,80 3.72 μM (0.93 μg/mL) for BPS,81 and 9.19 μM(2.22 μg/mL) for picloram.82 (2) Low-affinity binders could beeffectively transported by TTR and accumulated in vulnerableorgans in vivo,12−14 since most TTR in human plasma resides inthe apo form.15−17 Their accumulations could subsequentlyinduce adverse effects including developmental disorders.18,19

Figure 3. Crystallographic binding conformations of the selected active compounds and T4 in the thyroid-binding site of TTR: (a) clonixin, (b)DNPC (in the AA′ interface), (c) DNPC (in the BB′ interface), (d) triclopyr, (e) T4 (PDB ID: 2ROX, for comparison), and (f) BPS. The refined(2|Fo|−|Fc|) electron density is shown in gray mesh at the 1σ level. The water molecules present around the ligands are shown as red spheres. Theelectron density of the second ring of BPS was poorly defined probably due to high mobility.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.6b02771Environ. Sci. Technol. 2016, 50, 11984−11993

11989

Interestingly, BPS, a replacement for bisphenol A (BPA),bound to TTR with a Kd of 52 μM. BPS is used extensively as aplasticizer in BPA-free consumer products and as ananticorrosive agent in canned food. Despite claims that it is asafer alternative to BPA, BPS has been reported to pose similarpotential health hazards as BPA in a number of in vitro and invivo studies.83 Embryonic exposure to BPS is associated withneurogenesis within the hypothalamus, causing gestation andhyperactivity in juvenile zebrafish84 and nonassociative learningimpairments in adult zebrafish.85 BPS can also decrease plasmaTH levels and impair the reproductive potential of zebrafish.86

It was also found in dust samples in 12 countries, and thehighest median concentration was in Greece (860 ng/g).87

Clonixin was the strongest TTR binder identified in theTox21 inventory and had a Kd of 0.26 μM. Clonixin is anonsteroidal anti-inflammatory drug, and it has a similarstructure and TTR binding affinity as meclofenamic acid.40 Toour knowledge, its thyroid effects have not been studied.DNPC is used as an inhibitor of styrene polymerization and a

dye intermediate, and this molecule showed the secondstrongest interaction with TTR. DNPC is an uncoupler ofoxidative phosphorylation in humans, and it can cause acutehyperthermia, kidney and liver failure, and cerebral edema.88

DNPC has a high potential for accumulation in biota andhumans.88

The herbicides fluroxypyr, mesotrione, picloram, andtriclopyr were identified as TTR-binders, with triclopyr beingthe strongest binder. Triclopyr is used as an alternative to 2,4,5-T, and it has been reported to cause thyroid-relatedneurotoxicity89 and developmental toxicity in rats.90 Themajor metabolite of triclopyr is 3,5,6-trichloro-2-pyridinol,91

which is also a TTR-binder,33 and it has been reported todisrupt the levels of T4 and thyroid stimulating hormone inhumans.92 Exposure to triclopyr poses a significant threat to itsapplicators, and the amount in their urine samples was reportedto be 1−13 mg/day.93 The compound has been detected inwater samples from Australia,94 France,95 and multiplelocations in the US.96,97 Environmental implications of theother three herbicides are given in the Supporting Information.In summary, we have determined the binding conformations

of PFOS, PFOA, and BP2 in the TBS of TTR. The structuresof the complexes provide valuable information for improvingour understanding of the molecular interactions between TTRand emerging contaminants. PFOS and PFOA were previouslyreported as strong TTR-binders, but our results indicate thatthey are weak binders in plasma, whereas BP2 showed strongbinding activity. BP2 is thus more likely to be transported byTTR and accumulated in vulnerable organs. Based on the novelTTR complex structures and the information on molecularinteractions obtained from MD simulations, a VS protocol wasdeveloped for the identification of potential THDCs targetingTTR from the Tox21 inventory. The protocol proposed 192industrial compounds as potential binders to TTR. Among the12 compounds tested using in vitro experiments, we identifiedseven novel TTR-binders including clonixin, DNPC, triclopyr,fluroxypyr, BPS, picloram, and mesotrione. Most of these havebeen detected in the environment and have shown adversethyroid-related effects in animal models. We further cocrystal-lized TTR with PBS, clonixin, DNPC, and triclopyr, and theirstructures showed that the compounds bind in the TBS aspredicted by the VS protocol. We suggest further in vivo studieson these TTR-targeting compounds especially as they mayaccumulate in vulnerable organs.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.est.6b02771.

Details of the X-ray structures, plasma assay, and VSperformance; the identified potential TTR-binders andthe compounds tested using ITC; and procedures andresults of the MD simulations and MM-GBSAcalculations (PDF)

■ AUTHOR INFORMATIONCorresponding Author*Phone: +46-90-786-5266. Fax: +46-90-786-7655. E-mail:patrik.andersson@umu.se.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis study was financed by the MiSSE project through grantsfrom the Swedish Research Council for the Environment,Agricultural Sciences and Spatial Planning (Formas) (210-2012-131) and by the Swedish Research Council (VR) (521-2011-6427). The MD simulations were conducted using theHigh Performance Computing Center North.

■ REFERENCES(1) Murk, A. J.; Rijntjes, E.; Blaauboer, B. J.; Clewell, R.; Crofton, K.M.; Dingemans, M. M.; Furlow, J. D.; Kavlock, R.; Kohrle, J.; Opitz,R.; Traas, T.; Visser, T. J.; Xia, M.; Gutleb, A. C. Mechanism-basedtesting strategy using in vitro approaches for identification of thyroidhormone disrupting chemicals. Toxicol. In Vitro 2013, 27 (4), 1320−46.(2) Crofton, K. M. Thyroid disrupting chemicals: mechanisms andmixtures. Int. J. Androl. 2008, 31 (2), 209−23.(3) Haddow, J. E.; Palomaki, G. E.; Allan, W. C.; Williams, J. R.;Knight, G. J.; Gagnon, J.; O’Heir, C. E.; Mitchell, M. L.; Hermos, R. J.;Waisbren, S. E.; Faix, J. D.; Klein, R. Z. Maternal thyroid deficiencyduring pregnancy and subsequent neuropsychological development ofthe child. N. Engl. J. Med. 1999, 341 (8), 549−55.(4) Gilbert, M. E.; Rovet, J.; Chen, Z.; Koibuchi, N. Developmentalthyroid hormone disruption: prevalence, environmental contaminantsand neurodevelopmental consequences. NeuroToxicology 2012, 33 (4),842−52.(5) Hornberg, A.; Eneqvist, T.; Olofsson, A.; Lundgren, E.; Sauer-Eriksson, A. E. A comparative analysis of 23 structures of theamyloidogenic protein transthyretin. J. Mol. Biol. 2000, 302 (3), 649−69.(6) Weiss, J. M.; Andersson, P. L.; Lamoree, M. H.; Leonards, P. E.;van Leeuwen, S. P.; Hamers, T. Competitive binding of poly- andperfluorinated compounds to the thyroid hormone transport proteintransthyretin. Toxicol. Sci. 2009, 109 (2), 206−16.(7) Cheek, A. O.; Kow, K.; Chen, J.; McLachlan, J. A. Potentialmechanisms of thyroid disruption in humans: interaction of organo-chlorine compounds with thyroid receptor, transthyretin, and thyroid-binding globulin. Environ. Health Perspect. 1999, 107 (4), 273−8.(8) Lans, M. C.; Spiertz, C.; Brouwer, A.; Koeman, J. H. Differentcompetition of thyroxine binding to transthyretin and thyroxine-binding globulin by hydroxy-PCBs, PCDDs and PCDFs. Eur. J.Pharmacol., Environ. Toxicol. Pharmacol. Sect. 1994, 270 (2−3), 129−36.(9) Schreiber, G. The evolutionary and integrative roles oftransthyretin in thyroid hormone homeostasis. J. Endocrinol. 2002,175 (1), 61−73.(10) Purkey, H. E.; Palaninathan, S. K.; Kent, K. C.; Smith, C.; Safe,S. H.; Sacchettini, J. C.; Kelly, J. W. Hydroxylated polychlorinated

Environmental Science & Technology Article

DOI: 10.1021/acs.est.6b02771Environ. Sci. Technol. 2016, 50, 11984−11993

11990

biphenyls selectively bind transthyretin in blood and inhibit amyloido-genesis: rationalizing rodent PCB toxicity. Chem. Biol. 2004, 11 (12),1719−28.(11) Eneqvist, T.; Lundberg, E.; Karlsson, A.; Huang, S.; Santos, C.R.; Power, D. M.; Sauer-Eriksson, A. E. High resolution crystalstructures of piscine transthyretin reveal different binding modes fortriiodothyronine and thyroxine. J. Biol. Chem. 2004, 279 (25), 26411−6.(12) Kim, S. Y.; Choi, E. S.; Lee, H. J.; Moon, C.; Kim, E.Transthyretin as a new transporter of nanoparticles for receptor-mediated transcytosis in rat brain microvessels. Colloids Surf., B 2015,136, 989−96.(13) Meerts, I. A.; Assink, Y.; Cenijn, P. H.; Van Den Berg, J. H.;Weijers, B. M.; Bergman, A.; Koeman, J. H.; Brouwer, A. Placentaltransfer of a hydroxylated polychlorinated biphenyl and effects on fetaland maternal thyroid hormone homeostasis in the rat. Toxicol. Sci.2002, 68 (2), 361−71.(14) Diamanti-Kandarakis, E.; Bourguignon, J. P.; Giudice, L. C.;Hauser, R.; Prins, G. S.; Soto, A. M.; Zoeller, R. T.; Gore, A. C.Endocrine-disrupting chemicals: an Endocrine Society scientificstatement. Endocr. Rev. 2009, 30 (4), 293−342.(15) Ritchie, R. F.; Palomaki, G. E.; Neveux, L. M.; Navolotskaia, O.;Ledue, T. B.; Craig, W. Y. Reference distributions for the negativeacute-phase serum proteins, albumin, transferrin and transthyretin: apractical, simple and clinically relevant approach in a large cohort. J.Clin. Lab. Anal. 1999, 13 (6), 273−9.(16) Ritchie, R. F.; Palomaki, G. E.; Neveux, L. M.; Navolotskaia, O.Reference distributions for the negative acute-phase proteins, albumin,transferrin, and transthyretin: a comparison of a large cohort to theworld’s literature. J. Clin. Lab. Anal. 1999, 13 (6), 280−6.(17) Hollowell, J. G.; Staehling, N. W.; Flanders, W. D.; Hannon, W.H.; Gunter, E. W.; Spencer, C. A.; Braverman, L. E. Serum TSH, T(4),and thyroid antibodies in the United States population (1988 to1994): National Health and Nutrition Examination Survey (NHANESIII). J. Clin. Endocrinol. Metab. 2002, 87 (2), 489−99.(18) Boas, M.; Feldt-Rasmussen, U.; Main, K. M. Thyroid effects ofendocrine disrupting chemicals. Mol. Cell. Endocrinol. 2012, 355 (2),240−248.(19) Porterfield, S. P. Thyroidal dysfunction and environmentalchemicals–potential impact on brain development. Environ. HealthPerspect. 2000, 108 (s3), 433−8.(20) Dallaire, R.; Muckle, G.; Dewailly, E.; Jacobson, S. W.; Jacobson,J. L.; Sandanger, T. M.; Sandau, C. D.; Ayotte, P. Thyroid hormonelevels of pregnant inuit women and their infants exposed toenvironmental contaminants. Environ. Health Perspect. 2009, 117 (6),1014−20.(21) Audet-Delage, Y.; Ouellet, N.; Dallaire, R.; Dewailly, E.; Ayotte,P. Persistent organic pollutants and transthyretin-bound thyroxin inplasma of Inuit women of childbearing age. Environ. Sci. Technol. 2013,47 (22), 13086−92.(22) Plusquellec, P.; Muckle, G.; Dewailly, E.; Ayotte, P.; Begin, G.;Desrosiers, C.; Despres, C.; Saint-Amour, D.; Poitras, K. The relationof environmental contaminants exposure to behavioral indicators inInuit preschoolers in Arctic Quebec. NeuroToxicology 2010, 31 (1),17−25.(23) Rabinowitz, J. R.; Goldsmith, M. R.; Little, S. B.; Pasquinelli, M.A. Computational molecular modeling for evaluating the toxicity ofenvironmental chemicals: prioritizing bioassay requirements. Environ.Health Perspect. 2008, 116 (5), 573−7.(24) Muralidharan, A. R.; Selvaraj, C.; Singh, S. K.; Sheu, J. R.;Thomas, P. A.; Geraldine, P. Structure-Based Virtual Screening andBiological Evaluation of a Calpain Inhibitor for Prevention of Selenite-Induced Cataractogenesis in an in Vitro System. J. Chem. Inf. Model.2015, 55 (8), 1686−97.(25) Simoes, C. J. V.; Mukherjee, T.; Brito, R. M.; Jackson, R. M.Toward the discovery of functional transthyretin amyloid inhibitors:application of virtual screening methods. J. Chem. Inf. Model. 2010, 50(10), 1806−20.

(26) Rabinowitz, J. R.; Little, S. B.; Laws, S. C.; Goldsmith, M. R.Molecular modeling for screening environmental chemicals forestrogenicity: use of the toxicant-target approach. Chem. Res. Toxicol.2009, 22 (9), 1594−602.(27) McRobb, F. M.; Kufareva, I.; Abagyan, R. In silico identificationand pharmacological evaluation of novel endocrine disruptingchemicals that act via the ligand-binding domain of the estrogenreceptor alpha. Toxicol. Sci. 2014, 141 (1), 188−97.(28) Zhang, L.; Ren, X. M.; Guo, L. H. Structure-based investigationon the interaction of perfluorinated compounds with human liver fattyacid binding protein. Environ. Sci. Technol. 2013, 47 (19), 11293−301.(29) RCSB-PDB, RCSB Protein Data Bank (PDB), 2015. http://www.rcsb.org/pdb/ (accessed Sept 30, 2016).(30) Palaninathan, S. K. Nearly 200 X-ray crystal structures oftransthyretin: what do they tell us about this protein and the design ofdrugs for TTR amyloidoses? Curr. Med. Chem. 2012, 19 (15), 2324−42.(31) Ghosh, M.; Meerts, I. A.; Cook, A.; Bergman, A.; Brouwer, A.;Johnson, L. N. Structure of human transthyretin complexed withbromophenols: a new mode of binding. Acta Crystallogr., Sect. D: Biol.Crystallogr. 2000, 56 (9), 1085−95.(32) Iakovleva, I.; Begum, A.; Brannstrom, K.; Wijsekera, A.; Nilsson,L.; Zhang, J.; Andersson, P. L.; Sauer-Eriksson, A. E.; Olofsson, A.Tetrabromobisphenol A Is an Efficient Stabilizer of the TransthyretinTetramer. PLoS One 2016, 11 (4), e0153529.(33) Zhang, J.; Kamstra, J. H.; Ghorbanzadeh, M.; Weiss, J. M.;Hamers, T.; Andersson, P. L. In Silico Approach To Identify PotentialThyroid Hormone Disruptors among Currently Known DustContaminants and Their Metabolites. Environ. Sci. Technol. 2015, 49(16), 10099−107.(34) US-EPA Emerging Contaminants-Perfluorooctane Sulfonate(PFOS)and Perfluorooctanoic Acid (PFOA); The United StatesEnvironmental Protection Agency: 2014.(35) Spyropoulos, D. V. Stability testing of the plastics additives 2,4-dihydroxybenzophenone, 2,2 ′-dihydroxy-4-methoxybenzophenone, 2-hydroxy-4-n-octyloxybenzophenone, 2-ethoxy-2-ethyloxanilide (tinn-vin 312) and 2,2 ′-methylenebis(4-methyl-t-tert-butyl phenol) inaqueous and fatty food simulants. Food Addit. Contam. 1998, 15 (3),362−369.(36) Melzer, D.; Rice, N.; Depledge, M. H.; Henley, W. E.; Galloway,T. S. Association between serum perfluorooctanoic acid (PFOA) andthyroid disease in the U.S. National Health and Nutrition ExaminationSurvey. Environ. Health Perspect. 2010, 118 (5), 686−92.(37) Lopez-Espinosa, M. J.; Mondal, D.; Armstrong, B.; Bloom, M.S.; Fletcher, T. Thyroid function and perfluoroalkyl acids in childrenliving near a chemical plant. Environ. Health Perspect. 2012, 120 (7),1036−41.(38) Khalil, N.; Chen, A.; Lee, M.; Czerwinski, S. A.; Ebert, J. R.;DeWitt, J. C.; Kannan, K. Association of Perfluoroalkyl Substances,Bone Mineral Density, and Osteoporosis in the U.S. Population inNHANES 2009−2010. Environ. Health Perspect. 2016, 124 (1), 81−7.(39) Kunisue, T.; Chen, Z.; Buck Louis, G. M.; Sundaram, R.;Hediger, M. L.; Sun, L.; Kannan, K. Urinary concentrations ofbenzophenone-type UV filters in U.S. women and their associationwith endometriosis. Environ. Sci. Technol. 2012, 46 (8), 4624−32.(40) Iakovleva, I.; Brannstrom, K.; Nilsson, L.; Gharibyan, A. L.;Begum, A.; Anan, I.; Walfridsson, M.; Sauer-Eriksson, A. E.; Olofsson,A. Enthalpic Forces Correlate with the Selectivity of Transthyretin-Stabilizing Ligands in Human Plasma. J. Med. Chem. 2015, 58 (16),6507−15.(41) US-NRC Toxicity Testing in the 21st Century: A Vision and aStrategy; National Research Council, National Academies Press.Washington, DC, 2007.(42) Johnson, S. M.; Connelly, S.; Wilson, I. A.; Kelly, J. W. Towardoptimization of the second aryl substructure common to transthyretinamyloidogenesis inhibitors using biochemical and structural studies. J.Med. Chem. 2009, 52 (4), 1115−25.(43) Lindhagen-Persson, M.; Vestling, M.; Reixach, N.; Olofsson, A.Formation of cytotoxic transthyretin is not dependent on inter-

Environmental Science & Technology Article

DOI: 10.1021/acs.est.6b02771Environ. Sci. Technol. 2016, 50, 11984−11993

11991

molecular disulphide bridges commonly found within the amyloidform. Amyloid 2008, 15 (4), 240−5.(44) McCoy, A. J.; Grosse-Kunstleve, R. W.; Adams, P. D.; Winn, M.D.; Storoni, L. C.; Read, R. J. Phaser crystallographic software. J. Appl.Crystallogr. 2007, 40 (4), 658−674.(45) Adams, P. D.; Afonine, P. V.; Bunkoczi, G.; Chen, V. B.; Davis,I. W.; Echols, N.; Headd, J. J.; Hung, L. W.; Kapral, G. J.; Grosse-Kunstleve, R. W.; McCoy, A. J.; Moriarty, N. W.; Oeffner, R.; Read, R.J.; Richardson, D. C.; Richardson, J. S.; Terwilliger, T. C.; Zwart, P. H.PHENIX: a comprehensive Python-based system for macromolecularstructure solution. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2010, 66(2), 213−21.(46) Emsley, P.; Cowtan, K. Coot: model-building tools formolecular graphics. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2004,60 (12), 2126−32.(47) McNicholas, S.; Potterton, E.; Wilson, K. S.; Noble, M. E.Presenting your structures: the CCP4mg molecular-graphics software.Acta Crystallogr., Sect. D: Biol. Crystallogr. 2011, 67 (4), 386−94.(48) Case, D. A.; Darden, T. A.; Gusarov, S.; Kovalenko, A.; Kollman,P. A. Amber 2015 (AmberTools15 and Amber14); University ofCalifornia: San Francisco, 2015.(49) Hou, T.; Wang, J.; Li, Y.; Wang, W. Assessing the performanceof the MM/PBSA and MM/GBSA methods. 1. The accuracy ofbinding free energy calculations based on molecular dynamicssimulations. J. Chem. Inf. Model. 2011, 51 (1), 69−82.(50) Mysinger, M. M.; Carchia, M.; Irwin, J. J.; Shoichet, B. K.Directory of useful decoys, enhanced (DUD-E): better ligands anddecoys for better benchmarking. J. Med. Chem. 2012, 55 (14), 6582−94.(51) Freitas, J.; Miller, N.; Mengeling, B. J.; Xia, M.; Huang, R.;Houck, K.; Rietjens, I. M.; Furlow, J. D.; Murk, A. J. Identification ofthyroid hormone receptor active compounds using a quantitative high-throughput screening platform. Curr. Chem. Genomics Transl. Med.2014, 8, 36−46.(52) Browne, P.; Judson, R. S.; Casey, W. M.; Kleinstreuer, N. C.;Thomas, R. S. Screening Chemicals for Estrogen Receptor BioactivityUsing a Computational Model. Environ. Sci. Technol. 2015, 49 (14),8804−14.(53) Huang, R.; Xia, M.; Nguyen, D.-T.; Zhao, T.; Sakamuru, S.;Zhao, J.; Shahane, S. A.; Rossoshek, A.; Simeonov, A. Tox21Challengeto build predictive models of nuclear receptor and stress responsepathways as mediated by exposure to environmental chemicals anddrugs. Front. Environ. Sci. 2016, DOI: 10.3389/fenvs.2015.00085.(54) Tice, R. R.; Austin, C. P.; Kavlock, R. J.; Bucher, J. R. Improvingthe human hazard characterization of chemicals: a Tox21 update.Environ. Health Perspect. 2013, 121 (7), 756−65.(55) Schrodinger LigPrep, version 3.4; Schrodinger, LLC: New York,NY, 2015.(56) Schrodinger Glide, version 6.7; Schrodinger, LLC: New York,NY, 2015.(57) Jain, A. N.; Nicholls, A. Recommendations for evaluation ofcomputational methods. J. Comput.-Aided Mol. Des. 2008, 22 (3−4),133−9.(58) Iakovleva, I.; Begum, A.; Pokrzywa, M.; Walfridsson, M.; Sauer-Eriksson, A. E.; Olofsson, A. The flavonoid luteolin, but not luteolin-7-O-glucoside, prevents a transthyretin mediated toxic response. PLoSOne 2015, 10 (5), e0128222.(59) Hamers, T.; Kamstra, J. H.; Sonneveld, E.; Murk, A. J.; Kester,M. H.; Andersson, P. L.; Legler, J.; Brouwer, A. In vitro profiling of theendocrine-disrupting potency of brominated flame retardants. Toxicol.Sci. 2006, 92 (1), 157−73.(60) Trivella, D. B.; dos Reis, C. V.; Lima, L. M.; Foguel, D.;Polikarpov, I. Flavonoid interactions with human transthyretin:combined structural and thermodynamic analysis. J. Struct. Biol.2012, 180 (1), 143−53.(61) Yokoyama, T.; Kosaka, Y.; Mizuguchi, M. Inhibitory activities ofpropolis and its promising component, caffeic acid phenethyl ester,against amyloidogenesis of human transthyretin. J. Med. Chem. 2014,57 (21), 8928−35.

(62) Alhamadsheh, M. M.; Connelly, S.; Cho, A.; Reixach, N.;Powers, E. T.; Pan, D. W.; Wilson, I. A.; Kelly, J. W.; Graef, I. A.Potent kinetic stabilizers that prevent transthyretin-mediated cardio-myocyte proteotoxicity. Sci. Transl. Med. 2011, 3 (97), 97ra81.(63) Wojtczak, A.; Cody, V.; Luft, J. R.; Pangborn, W. Structures ofhuman transthyretin complexed with thyroxine at 2.0 A resolution and3′,5′-dinitro-N-acetyl-L-thyronine at 2.2 A resolution. Acta Crystallogr.,Sect. D: Biol. Crystallogr. 1996, 52 (4), 758−65.(64) Pospisil, P.; Ballmer, P.; Scapozza, L.; Folkers, G. Tautomerismin computer-aided drug design. J. Recept. Signal Transduction Res. 2003,23 (4), 361−371.(65) Martin, Y. C. Let’s not forget tautomers. J. Comput.-Aided Mol.Des. 2009, 23 (10), 693−704.(66) Shelley, J. C.; Cholleti, A.; Frye, L. L.; Greenwood, J. R.; Timlin,M. R.; Uchimaya, M. Epik: a software program for pKa prediction andprotonation state generation for drug-like molecules. J. Comput.-AidedMol. Des. 2007, 21 (12), 681−91.(67) Friesner, R. A.; Banks, J. L.; Murphy, R. B.; Halgren, T. A.;Klicic, J. J.; Mainz, D. T.; Repasky, M. P.; Knoll, E. H.; Shelley, M.;Perry, J. K.; Shaw, D. E.; Francis, P.; Shenkin, P. S. Glide: A newapproach for rapid, accurate docking and scoring. 1. Method andassessment of docking accuracy. J. Med. Chem. 2004, 47 (7), 1739−1749.(68) Muziol, T.; Cody, V.; Luft, J. R.; Pangborn, W.; Wojtczak, A.Complex of rat transthyretin with tetraiodothyroacetic acid refined at2.1 and 1.8 A resolution. Acta Biochim. Pol. 2001, 48 (4), 877−84.(69) Ferguson, R. N.; Edelhoch, H.; Saroff, H. A.; Robbins, J.;Cahnmann, H. J. Negative cooperativity in the binding of thyroxine tohuman serum prealbumin. Preparation of tritium-labeled 8-anilino-1-naphthalenesulfonic acid. Biochemistry 1975, 14 (2), 282−9.(70) Cheng, S. Y.; Pages, R. A.; Saroff, H. A.; Edelhoch, H.; Robbins,J. Analysis of thyroid hormone binding to human serum prealbumin by8-anilinonaphthalene-1-sulfonate fluorescence. Biochemistry 1977, 16(16), 3707−13.(71) Schneider, G. Virtual screening: an endless staircase? Nat. Rev.Drug Discovery 2010, 9 (4), 273−6.(72) Zhu, T.; Cao, S.; Su, P. C.; Patel, R.; Shah, D.; Chokshi, H. B.;Szukala, R.; Johnson, M. E.; Hevener, K. E. Hit identification andoptimization in virtual screening: practical recommendations based ona critical literature analysis. J. Med. Chem. 2013, 56 (17), 6560−72.(73) Park, M. S.; Dessal, A. L.; Smrcka, A. V.; Stern, H. A. Evaluatingdocking methods for prediction of binding affinities of small moleculesto the G protein betagamma subunits. J. Chem. Inf. Model. 2009, 49(2), 437−43.(74) Warren, G. L.; Andrews, C. W.; Capelli, A. M.; Clarke, B.;LaLonde, J.; Lambert, M. H.; Lindvall, M.; Nevins, N.; Semus, S. F.;Senger, S.; Tedesco, G.; Wall, I. D.; Woolven, J. M.; Peishoff, C. E.;Head, M. S. A critical assessment of docking programs and scoringfunctions. J. Med. Chem. 2006, 49 (20), 5912−31.(75) Peterson, S. A.; Klabunde, T.; Lashuel, H. A.; Purkey, H.;Sacchettini, J. C.; Kelly, J. W. Inhibiting transthyretin conformationalchanges that lead to amyloid fibril formation. Proc. Natl. Acad. Sci. U. S.A. 1998, 95 (22), 12956−60.(76) Muziol, T.; Cody, V.; Wojtczak, A. Comparison of bindinginteractions of dibromoflavonoids with transthyretin. Acta Biochim. Pol.2001, 48 (4), 885−92.(77) Totrov, M.; Abagyan, R. Flexible ligand docking to multiplereceptor conformations: a practical alternative. Curr. Opin. Struct. Biol.2008, 18 (2), 178−84.(78) Frimurer, T. M.; Peters, G. H.; Iversen, L. F.; Andersen, H. S.;Møller, N. P. H.; Olsen, O. H. Ligand-induced conformationalchanges: improved predictions of ligand binding conformations andaffinities. Biophys. J. 2003, 84 (4), 2273−81.(79) Bica, A.; Farinha, A.; Blume, H.; Barbosa, C. M. Determinationof clonixin in plasma and urine by reversed-phase high-performanceliquid chromatography. J. Chromatogr. A 2000, 889 (1−2), 135−41.(80) Marin-Water-Agency, Vegetation Management Plan 2010. www.marinwater.org/DocumentCenter/ (accessed Aug 6, 2016).

Environmental Science & Technology Article

DOI: 10.1021/acs.est.6b02771Environ. Sci. Technol. 2016, 50, 11984−11993

11992

(81) Liao, C.; Liu, F.; Alomirah, H.; Loi, V. D.; Mohd, M. A.; Moon,H. B.; Nakata, H.; Kannan, K. Bisphenol S in urine from the UnitedStates and seven Asian countries: occurrence and human exposures.Environ. Sci. Technol. 2012, 46 (12), 6860−6.(82) Hall, J. C.; Deschamps, R. J. A.; Krieg, K. K. Immunoassays forthe Detection of 2,4-D and Picloram in River Water and Urine. J. Agric.Food Chem. 1989, 37 (4), 981−984.(83) Rochester, J. R.; Bolden, A. L. Bisphenol S and F: A SystematicReview and Comparison of the Hormonal Activity of Bisphenol ASubstitutes. Environ. Health Perspect. 2015, 123 (7), 643−50.(84) Kinch, C. D.; Ibhazehiebo, K.; Jeong, J. H.; Habibi, H. R.;Kurrasch, D. M. Low-dose exposure to bisphenol A and replacementbisphenol S induces precocious hypothalamic neurogenesis inembryonic zebrafish. Proc. Natl. Acad. Sci. U. S. A. 2015, 112 (5),1475−80.(85) Mersha, M. D.; Patel, B. M.; Patel, D.; Richardson, B. N.;Dhillon, H. S. Effects of BPA and BPS exposure limited to earlyembryogenesis persist to impair non-associative learning in adults.Behav. Brain Funct. 2015, 11, 27.(86) Naderi, M.; Wong, M. Y.; Gholami, F. Developmental exposureof zebrafish (Danio rerio) to bisphenol-S impairs subsequentreproduction potential and hormonal balance in adults. Aquat. Toxicol.2014, 148, 195−203.(87) Wang, W.; Abualnaja, K. O.; Asimakopoulos, A. G.; Covaci, A.;Gevao, B.; Johnson-Restrepo, B.; Kumosani, T. A.; Malarvannan, G.;Minh, T. B.; Moon, H. B.; Nakata, H.; Sinha, R. K.; Kannan, K. Acomparative assessment of human exposure to tetrabromobisphenol Aand eight bisphenols including bisphenol A via indoor dust ingestionin twelve countries. Environ. Int. 2015, 83, 183−91.(88) ATSDR Toxicological Profile for Dinitrocresols; USA Agency forToxic Substances and Disease Registry (ATSDR): 1995.(89) Reddy, T. P.; Manczak, M.; Calkins, M. J.; Mao, P.; Reddy, A.P.; Shirendeb, U.; Park, B.; Reddy, P. H. Toxicity of neurons treatedwith herbicides and neuroprotection by mitochondria-targetedantioxidant SS31. Int. J. Environ. Res. Public Health 2011, 8 (1),203−21.(90) Carney, E. W.; Billington, R.; Barlow, S. M. Developmentaltoxicity evaluation of triclopyr butoxyethyl ester and triclopyrtriethylamine salt in the CD rat. Reprod. Toxicol. 2007, 23 (2), 165−74.(91) Schmidt, L.; Muller, J.; Goen, T. Simultaneous monitoring ofseven phenolic metabolites of endocrine disrupting compounds(EDC) in human urine using gas chromatography with tandemmass spectrometry. Anal. Bioanal. Chem. 2013, 405 (6), 2019−29.(92) Fortenberry, G. Z.; Hu, H.; Turyk, M.; Barr, D. B.; Meeker, J. D.Association between urinary 3, 5, 6-trichloro-2-pyridinol, a metaboliteof chlorpyrifos and chlorpyrifos-methyl, and serum T4 and TSH inNHANES 1999−2002. Sci. Total Environ. 2012, 424, 351−5.(93) Gosselin, N. H.; Brunet, R. C.; Carrier, G.; Dosso, A. Workerexposures to triclopyr: risk assessment through measurements in urinesamples. Ann. Occup. Hyg. 2005, 49 (5), 415−22.(94) Tran, A. T.; Hyne, R. V.; Doble, P. Determination of commonlyused polar herbicides in agricultural drainage waters in Australia byHPLC. Chemosphere 2007, 67 (5), 944−53.(95) Botta, F.; Fauchon, N.; Blanchoud, H.; Chevreuil, M.; Guery, B.Phyt’Eaux Cites: application and validation of a programme to reducesurface water contamination with urban pesticides. Chemosphere 2012,86 (2), 166−76.(96) Battaglin, W. A.; Rice, K. C.; Focazio, M. J.; Salmons, S.; Barry,R. X. The occurrence of glyphosate, atrazine, and other pesticides invernal pools and adjacent streams in Washington, DC, Maryland,Iowa, and Wyoming, 2005−2006. Environ. Monit. Assess. 2009, 155(1−4), 281−307.(97) Tagert, M. L.; Massey, J. H.; Shaw, D. R. Water quality survey ofMississippi’s Upper Pearl River. Sci. Total Environ. 2014, 481, 564−73.(98) Purkey, H. E.; Dorrell, M. I.; Kelly, J. W. Evaluating the bindingselectivity of transthyretin amyloid fibril inhibitors in blood plasma.Proc. Natl. Acad. Sci. U. S. A. 2001, 98 (10), 5566−71.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.6b02771Environ. Sci. Technol. 2016, 50, 11984−11993

11993