GABAA receptor targetof tetramethylenedisulfotetramineChunqing Zhaoa,1, Sung Hee Hwangb, Bruce A. Buchholzc,2, Timothy S. Carpenterd, Felice Lightstoned,2, Jun Yangb,Bruce D. Hammockb,2, and John E. Casidaa,2

aEnvironmental Chemistry and Toxicology Laboratory, Department of Environmental Science, Policy, and Management, University of California, Berkeley, CA94720; bDepartment of Entomology, College of Agricultural and Environmental Science, University of California, Davis, CA 95616; and cCenter for AcceleratorMass Spectrometry and dBiosciences and Biotechnology Division, Lawrence Livermore National Laboratory, Livermore, CA 94551

Contributed by John E. Casida, April 24, 2014 (sent for review March 22, 2014)

Use of the highly toxic and easily prepared rodenticide tetrame-thylenedisulfotetramine (TETS) was banned after thousands ofaccidental or intentional human poisonings, but it is of continuedconcern as a chemical threat agent. TETS is a noncompetitiveblocker of the GABA type A receptor (GABAAR), but its molecularinteraction has not been directly established for lack of a suitableradioligand to localize the binding site. We synthesized [14C]TETS(14 mCi/mmol, radiochemical purity >99%) by reacting sulfamidewith H14CHO and s-trioxane then completion of the sequential cycli-zation with excess HCHO. The outstanding radiocarbon sensitivity ofaccelerator mass spectrometry (AMS) allowed the use of [14C]TETS inneuroreceptor binding studies with rat brain membranes in compar-ison with the standard GABAAR radioligand 4′-ethynyl-4-n-[3H]pro-pylbicycloorthobenzoate ([3H]EBOB) (46 Ci/mmol), illustrating the useof AMS for characterizing the binding sites of high-affinity 14C radio-ligands. Fourteen noncompetitive antagonists of widely diverse che-motypes assayed at 1 or 10 μM inhibited [14C]TETS and [3H]EBOBbinding to a similar extent (r2 = 0.71). Molecular dynamics simula-tions of these 14 toxicants in the pore region of the α1β2γ2 GABAARpredict unique and significant polar interactions for TETS with α1T1′and γ2S2′, which are not observed for EBOB or the GABAergic insec-ticides. Several GABAAR modulators similarly inhibited [14C]TETS and[3H]EBOB binding, including midazolam, flurazepam, avermectin Ba1,baclofen, isoguvacine, and propofol, at 1 or 10 μM, providing an invitro system for recognizing candidate antidotes.

neurotoxicity | convulsant | molecular modeling

Severe poisonings in a German furniture factory in the 1940swere traced to wool impregnated with the resinous reaction

product of sulfamide (NH2SO2NH2) and formaldehyde (HCHO).The causative agent was identified as tetramethylenedisulfotetr-amine (TETS, also known as tetramine) which was then developedas a rodenticide (now illegal) and continues to be of concern asa chemical threat agent. The chronology of TETS chemistryand toxicology is given briefly here and more extensively in SIAppendix, section S1.TETS was first synthesized more than 80 y ago (1–3). Structure–

activity studies showed that any structural modification greatlyreduces the toxicity (4). The need to understand the distributionand fate of TETS led to 14C radiosynthesis in 1967 (5) by an un-disclosed method, but the product was only 80% pure, limiting theinterpretation of biological experiments. Analysis is achieved byliquid chromatography/MS (6) or when ultrahigh sensitivity is re-quired and [14C]TETS is available by accelerator mass spectrom-etry (AMS) (7), as reported here.TETS is highly toxic to mammals with an i.p. LD50 of 0.11–0.22

mg/kg in mice and rats, leading to its use as a rodenticide until itwas banned worldwide in the early 1990s (2, 8, 9). However, it isstill available illegally and responsible for accidental or intentionalpoisonings in China and other countries. The estimated lethal doseof 7–10 mg in adult humans coupled with its ease of synthesis andstability serve as the basis for the chemical threat concern (10–15).Neurotoxicity is sometimes alleviated or antidoted by compounds

modulating the target site to reduce disruption by the toxicant.TETS toxicity is reported to be alleviated in rodents or humansby diazepam, barbiturates, allopregnanolone, and sodium 2,3-dimercapto-1- propanesulfonate (NaDMPS), some of which areGABAAR modulators (16–24) (SI Appendix, section S1). TETS isone of several small-cage convulsants, a group that also includesbicyclophosphorus compounds, such as the even more toxic t-butyl-bicyclophosphate (TBPO) and t-butylbicyclophosphorothionate(TBPS) (25) (Fig. 1).TETS is a noncompetitive antagonist of the GABA type A re-

ceptor (GABAAR) based on multiple physiological and toxicolog-ical criteria* (20, 26–29) (SI Appendix, section S1) and GABAARassays with two trioxabicyclooctane radioligands, [35S]TBPS (30),and 4-n-[3H]propyl-4′-ethynylbicycloorthobenzoate ([3H]EBOB)(31) (Fig. 1). TETS is a competitive inhibitor of [3H]EBOBbinding with a potency in rat brain GABAAR consistent withits toxicity in mice (4, 32). However, it does not inhibit humanGABAAR recombinant β3 homopentamer assayed with [3H]EBOB(33), which has a structure–activity relationship for inhibitorssimilar to that for the housefly GABAR (34). These deductionsare based on the use of [35S]TBPS and [3H]EBOB to assay theaction of TETS.Direct observation of the TETS binding site requires the use of

TETS as the radioligand. Radioligand binding studies for neu-roreceptors as toxicant targets normally require high specific ac-tivities (>10 Ci/mmol) such as 3H labeling, which is not availableto date for TETS. [14C]TETS reported here has a specific activity

Significance

Tetramethylenedisulfotetramine (TETS) is a feared chemical threatagent because of its high convulsant toxicity, ease of synthesis,and availability even though it is banned as a rodenticide. Earlierphysiological evidence indicating action as a GABA receptor an-tagonist and inhibitor of [35S]TBPS and [3H]EBOB binding is con-firmed here by radiosynthesis of [14C]TETS and defining its bindingsite in rat brain membranes by accelerator mass spectrometry andtoxicant specificity studies on inhibition of [14C]TETS and [3H]EBOBbinding. TETS undergoes specific and unique polar interactionsinside the 1′2′ ring pore region instead of the 2′,6′, and 9′ site forinsecticides. This study helps define GABAAR sites for potentialantidotes acting to prevent TETS binding or displace it from itsbinding site.

Author contributions: B.D.H. and J.E.C. designed research; C.Z., S.H.H., B.A.B., and J.Y.performed research; T.S.C. and F.L. analyzed data; and B.D.H. and J.E.C. wrote the paper.

The authors declare no conflict of interest.1Present address: College of Science, China Agricultural University, Beijing 100193, China.2To whom correspondence may be addressed. E-mail: buchholz2@llnl.gov, lightstone1@llnl.gov, bdhammock@ucdavis.edu, or ectl@berkeley.edu.

This paper contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1407379111/-/DCSupplemental.

*Zolkowska D, et al., American Epilepsy Society Annual Meeting, December 2–6, 2011,Baltimore, abstr 3.069.

www.pnas.org/cgi/doi/10.1073/pnas.1407379111 PNAS | June 10, 2014 | vol. 111 | no. 23 | 8607–8612

NEU

ROSC

IENCE

of 14 mCi/mmol, which is only about 0.02–0.1% the level normallyused for radioligand binding assays. The required detection couldonly be achieved with [14C]TETS by greatly enhanced sensitivityresulting from AMS, which to our knowledge has never been usedbefore in neuroreceptor radioligand binding assays.This study characterizes the [14C]TETS binding site in rat brain

GABAAR. TETS and EBOB are compared as an unknown versusa standard cage convulsant radioligand that may have some com-mon features in their binding sites. Rat brain is used because TETSis primarily a rodenticide. The same experiments are run with [14C]TETS and [3H]EBOB so that only the radioligand and analyticalmethod are varied to best evaluate the utility of the new radio-ligand and the AMS analysis technique in defining the mechanismof TETS toxicity. The [14C]TETS binding assay allows verificationof the mode of action, definition of the pharmacological profile,localization of the binding site, and characterization of potentialantidotes or alleviating agents.

Results and DiscussionSynthesis of [13C]TETS and [14C]TETS. Unlabeled TETS is readilysynthesized from sulfamide by reacting with HCHO (37% wt/vol

in water) or its equivalent such as s-trioxane or paraformaldehydein acidic condition (2, 4). However, the low concentration ofcommercially available H14CHO (1–3% in water) seriously delayedthe final ring cyclization step to form [14C]TETS. To overcomethis, the procedure was modified by stepwise cyclization (Fig. 2).Reaction conditions were optimized through tests with H13CHO asthe H14CHO mimic. H13CHO (20% aqueous solution, 0.25equivalents relative to sulfamide) and s-trioxane (source of 1.75equivalents of HCHO as a solid form) ensured that all H13CHOwas incorporated into the product owing to the slower release ofunlabeled HCHO from s-trioxane. An additional treatment withunlabeled HCHO completed the final cyclization reaction. Theseconditions were then used to prepare the [14C]TETS as follows. Toa chilled solution of sulfamide (1.9 mg, 20 μmol) and s-trioxane(1.1 mg, 12 μmol) in 21 μL of concentrated hydrochloric acid(conc. HCl, 250 μmol) was added 250 μL of H14CHO (5 μmol, 3%in water, 250 μCi, specific activity 50 mCi/mmol, 99% pure byHPLC) at 0 °C. The reaction mixture was slowly warmed to roomtemperature and stirred for 1 d. After adding acetonitrile (250 μL),the azeotrope was evaporated under a stream of dry air at roomtemperature. To the remaining reaction mixture were added 21 μLof conc. HCl and 3.8 μL of unlabeled HCHO (37% in water). After1 h, acetonitrile (100 μL) was added and the azeotrope was againremoved under a stream of dry air at room temperature. Theremaining solid was dissolved by adding 100 μL of acetone then500 μL of dichloromethane, yielding a white precipitate. Afterfiltration, the filtrate was evaporated under a stream of dry air andthe residual crude TETS was purified by column chromatographyon silica gel with an eluent (dichloromethane: n-hexane 3:1, Rf =0.54). The [14C]TETS was obtained on evaporation as a whitesolid: 310 μg, 12.9% chemical yield, 7.2% radiochemical yield,specific activity 14 mCi/mmol, and >99% radiochemical purity (SIAppendix, section S2). The product was dissolved in 1 mL acetoneand stored in a sealed amber glass ampoule at −20 °C. GC-MSanalysis data for the final [13C]TETS and [14C]TETS revealedslightly less label incorporation with [14C]TETS (Table 1).

[14C]TETS Binding Parameters. Neuroreceptor binding assays witha 14C-labeled compound require an ultrasensitive analytical methodprovided by the use of tandem HPLC and AMS with a typical limitof quantification of 2–20 amol, which proved to be adequate in thepresent studies.[14C]TETS undergoes specific binding to rat brain membranes

at 37 °C with half saturation at 0.08 μM (Fig. 3A). TETS is also apotent inhibitor of [3H]EBOB binding under the same conditionswith an IC50 of 0.79 μM (Fig. 3B). Nonspecific binding was de-termined with unlabeled TETS at 10 μM. Total, nonspecific, andspecific binding with [14C]TETS were 1,390, 454, and 1,019 fgTETS/μg protein, respectively (i.e., 67 ± 1% specific binding).The corresponding values with [3H]EBOB were 2,013, 994, and1,019 dpm/125 μg protein, respectively, corresponding to 50 ± 5%specific binding. GABA at 0.3, 1, and 10 μM inhibited [14C]TETSbinding by 15 ± 3, 55 ± 1, and 79 ± 3%, respectively, and thecorresponding values for [3H]EBOB were 42 ± 8, 61 ± 4, and

Fig. 1. Structures of three radioligands and unlabeled insecticides or con-vulsants with number designations. Two numbers for TETS (1/2) and flur-alaner (10/11) refer to different concentrations considered later.

Fig. 2. Synthesis of [13C]- and [14C]TETS by serial cyclization steps in con-centrated hydrochloric acid. The percentages of mono-, di-, tri-, and tetra-labeling are given in Table 1.

8608 | www.pnas.org/cgi/doi/10.1073/pnas.1407379111 Zhao et al.

104 ± 2%, respectively. L-glutamic acid did not inhibit bindingof either radioligand at 1 μM.

Convulsants and Insecticides Compete Similarly for [14C]TETS and[3H]EBOB Binding Sites. If [14C]TETS and [3H]EBOB bind at thesame site in the same way (i.e., are superimposable), they shouldbe similarly inhibited by a series of convulsants and insecticidesselected for their widely varied chemotypes and assayed at 1 or10 μM. The results for 16 compounds or concentrations (Fig. 1)are presented in Fig. 4 and SI Appendix, section S3. Earlier publishedfindings on the [3H]EBOB and [35S]TBPS sites are given in SI Ap-pendix, section S4. TETS with an IC50 of 0.08 μM (Fig. 3A) wasconsiderablymorepotent than three of its analogs assayedwith eitherradioligand. The TETS-type compounds (1–5) and several insecti-cides or cage convulsants (6–16) inhibit [14C]TETS and [3H]EBOBbinding to a similar extent (r2 = 0.71). TETS, EBOB, and the otherconvulsants and insecticides therefore compete with each other atcomparable binding sites, prompting atomistic, structural examina-tion in the GABAAR pore.

Different Binding Positions for TETS and EBOB.Much is known aboutthe binding sites for EBOB, picrotoxinin (PTX), lindane, 3,3-bis-trifluoromethyl-bicyclo[2,2,1]heptane-2,2-dicarbonitrile, fipronil, andα-endosulfan in the α1β2γ2 GABAAR (33–40). TETS, TBPO, andEBOB were initially molecularly docked in the pore region of

the α1β2γ2 GABAAR model. After 40 ns of molecular dy-namics (MD) simulations (described in SI Appendix, sectionS5), their optimized and equilibrated positions (Fig. 5) illus-trate that EBOB and TBPO overlap the proposed TETSbinding site (located around the 1′2′ region of the pore). MDstimulations of all of the insecticides and convulsants in thisstudy predict a partially overlapping binding site with a com-mon region at the 1′2′ position (the 2′ “contact zone”) (SIAppendix, section S6). However, the interactions with specificresidues in the 1′2′ position can differ. TETS and EBOB, forexample, both make a substantial proportion of all their cal-culated contacts to the 1′2′ residues (61% and 45%, respectively).However, whereas EBOB contacts the α/γ subunits at the 1′2′ site68% of the time (only slightly more than the 60% expected fornonspecific subunit binding), TETS contacts the α/γ subunits at the1′2′ site 98% of the time with an α1:β2 ratio of 31:1, suggestinga specific α1 interaction. Moreover, 69% of all of the simulatedTETS contacts with GABAAR are made to just the α1T1′ andγ2S2′ residues. In contrast, EBOBmakes only 6% of its contacts tothese residues. Extracting detailed interactions from our simu-lations posits four ways to interact with the 1′2′ residues (SI Ap-pendix, section S7): specific polar interactions (TETS), bothsignificant polar and hydrophobic interactions (TBPO and TBPS),general hydrophobic interactions (such as EBOB), and nonspecificor nonsignificant interactions that imply that other residues withinthe GABAAR pore are more important for binding (e.g., PTXinteracts with 6′ as a key residue).The predicted TETS and EBOB residue contact differences

and binding interactions correspond with the sensitivity andspecificity observed in expressed human β3 (hydrophobic at 1′2′)homopentameric GABAARs (33). In agreement with mutationstudies that show that changes to the β3 homopentameter 2′residue from hydrophobic to polar (A→S) decreases the affinityof EBOB for the receptor (36, 37, 39), our binding pattern showsEBOB makes significant hydrophobic interactions at the 1′2′region during the simulation. Conversely, our simulated TETSmakes polar interactions at this 1′2′ region, suggesting that polarresidues are needed for TETS binding. In the β3 subunit valineand alanine have replaced the α1T1′ and γ2S2′ residues, abol-ishing the necessary polar residues that TETS is predicted tobind, and could explain why α1β2γ2 is sensitive to TETS but thehomopentamer is not. Thus, a β3 homopentamer 1′V→T or 2′A→S mutant (similar to the α1 or γ2 residues) may show in-creased affinity for TETS.

Table 1. Labeling of [13C]TETS and [14C]TETS from 20% aqueousH13CHO and 3% aqueous H14CHO (50 mCi/mmol) and specificactivity of [14C]TETS

No. of labeledcarbons

[14C]TETS

[13C]TETS, %* %* mCi/mmol† Contribution‡

0 67.3 76.2 0 01 25.4 19.8 50 9.92 5.7 3.4 100 3.43 1.2 0.6 150 0.94 0.4 0 200 0Total 100 100 14.2§

*Percentage isotopic distribution of TETS determined by GC-MS in full-scanmode for [13C]TETS and selected ion monitoring mode for [14C]TETS (SI Ap-pendix, section S2).†Theoretical specific activity of TETS with the indicated number of 14C-la-beled carbon(s).‡Contribution of components with one, two, and three 14C-labeled carbon(s)summated to give the total specific activity (mCi/mmol) of the final [14C]TETS.§Experimental specific activity (14.0 mCi/mmol) of the final [14C]TETS wasdetermined by liquid scintillation counting of 1 μL aliquot from 1 mL acetonesolution of 310 μg [14C]TETS.

Fig. 3. TETS target assayed as (A) displacement of [14C]TETS and (B) in-hibition of [3H]EBOB binding in rat brain membranes.

Fig. 4. Inhibition of [14C]TETS and [3H]EBOB binding in rat brain mem-branes by convulsants and insecticides at 10 μM (2 and 11) or 1 μM (all otherdata). Plotted from data in SI Appendix, section S3.

Zhao et al. PNAS | June 10, 2014 | vol. 111 | no. 23 | 8609

NEU

ROSC

IENCE

Types A and B Toxic Action Relative to Binding Positions. TETS andEBOB fall into two different types (A and B) on comparingtoxicity with target site potency assayed as inhibition of either[35S]TBPS binding in brain membranes or 36Cl uptake in mem-brane vesicles of the cerebral cortex (20, 32). Type A compoundsinclude EBOB and many insecticides with large substituents orextended structures, and the type B set includes TETS, TBPS,TBPO, and other small compact molecules, some of very high i.p.toxicity to mice (LD50 36 μg/kg for TBPO) (25) (SI Appendix,section S8), although much less toxic to injected houseflies (LD5090 mg/kg for TETS) (4). The target site mapping studies abovesuggest a molecular distinction between the binding of type Acompounds and Type B cage convulsants (32). Type B antagonists(TETS, TBPO, and TBPS) bind with significant polar interactions,whereas type A antagonists (PTX, lindane, 12-ketoendrin, andEBOB) do not (SI Appendix, section S7). In confirmation, distinctdifferences appear between types A and B compounds on com-paring native, α1β3γ2, and (β3)5 GABAARs (33). Whereas thetype A compounds are exceptionally potent on the β3 homo-pentamer, the type B TBPS acts similarly on all three receptor

types and TETS is a poor inhibitor of (β3)5 (33, 34). Consideringthese relationships, we propose that the type B compact set in-cluding TETS and TBPS undergoes significant polar interactionsin the 1′2′ ring, whereas the type A elongated compounds such asEBOB do not. Interestingly, the insecticidal activity of the iso-xazoline fluralaner (10) seems to result from action at a distinctGABA receptor site (41, 42) not considered here.

TETS Candidate Antidotes. TETS was the first and because of manypoisoning cases is now the best known of the small-cage con-vulsants, but some bicyclophosphorus compounds are much moretoxic and probably act in the same way (19, 25). After a halfcentury of search, there are still no adequate antidotes for TETS-induced poisoning, either accidental or intentional. The candidateshave come from anticonvulsants used to counteract convulsantaction, trials in rats and mice, and mechanism studies in animals,cells, and in vitro systems (SI Appendix, section S1). Cell and nervestudies confirm action on GABA-induced signals and chloride flux.Diazepam and Na phenobarbital increase the mouse i.p. LD50 ofTETS by severalfold (19) and diazepam and midazolam inhibit[3H]EBOB and [35S]TBPS binding (SI Appendix, sections S3and S4). The highest inhibitory effect among the benzodiaze-pines and barbiturates examined at 1 or 10 μM was 30–40% formidazolam and flurazepam (SI Appendix, section S3). SeveralGABAAR modulators that alter [35S]TBPS binding (30) arealso allosteric inhibitors of [14C]TETS or [3H]EBOB binding.Allopregnanolone is known to be active in [35S]TBPS bindingassays (30) and alleviating TETS toxicity (11, 27, 28). NaDMPS,a chelating agent normally used for treating heavy metal poi-soning, is effective as a TETS antidote in rodent models andhuman poisonings proposed to be due to elevating GABA levelsrather than as a chelator (SI Appendix, section S1). GABA levelsare elevated by TETS poisoning in rats and GABA administra-tion relieves the convulsions (SI Appendix, section S1). Theseizures induced by acute and repeated exposure to TETS arecharacterized as actions at both GABA and NMDA receptors(28, 29). TETS inhibition of NMDA-induced Ca2+ signaling incultured hippocampal neurons is partially reversed by either, orboth, NaDMPS and allopregnanolone (28). Binding of [14C]TETSor [3H]EBOB, or both, is inhibited by avermectin at 1 μM and bypreganolone, isoguvacine, NMDA, propofol, and pyridoxinebut not by NaDMPS at 1 or 10 μM, whereas bicuculline at 1 μMstimulated [14C]TETS and [3H]EBOB binding by 59–68% (SIAppendix, section S3). Baclofen at 1 μM and ethanol at 300 mMhad apparently somewhat different effects with the two radio-ligands (SI Appendix, section S3). However, TETS poisoningcases in humans have been treated with diazepam, allopreg-nanolone, and NaDMPS with little or no benefit (9–15).The GABAAR is the target of many toxicants for mammals

(TETS) and insects (insecticides) and exists in a multiplicity ofsubunit and interface combinations (43, 44), allowing high toxicitythat reaches its extreme for mammals with TETS and some othersmall-cage convulsants. In the search for antidotes the GABAAR invitro assays described here may provide a rapid means of limitingthe number of compounds for animal experimentation and ultimatetesting in cases of human poisoning. Further test of this hypothesisrequires a larger dataset for inhibition of [14C]TETS and [3H]EBOB binding versus toxicity.

Materials and MethodsChemicals and Chromatography. H14CHO (1 mCi/mL, 250 μCi) was purchasedfrom American Radiolabeled Chemicals Inc. H13CHO (99 atom % 13C, 20%aqueous solution) was obtained from Sigma-Aldrich. [3H]EBOB (26 Ci/mmol)was from Perkin-Elmer Inc.. All other reagents and solvents were obtainedfrom commercial suppliers and used without further purification. The syn-thesized products were characterized by TLC comparisons on Merck silica gel60 F254 plates detected for unlabeled and [13C]TETS by potassium perman-ganate and for [14C]TETS by radio TLC using a Bioscan System 200 ImagingScanner. Purification involved column chromatography using Spe-ed SPE

Fig. 5. The equilibrated positions of TETS (1), TBPO (6), and EBOB following40 ns of MD simulation in the pore region of the α1β2γ2 GABAAR model, withviews from the side (Left, the front two M2 helices have been removed forclarity), and views from the bottom of the pore (Right). The red dashed linessignify the common binding region around the 2′ “contact zone.” Thecontact zone is the region where a compound can make contacts to 2′, eitherfrom above or below the residue. The view down the pore shows that at the1′2′ region, TETS makes primarily polar interactions to the α and γ subunits(hydrogen bonds shown as dashed magenta lines), TBPO makes both polarand hydrophobic interactions, and EBOB makes nonspecific hydrophobicinteractions with this 1′2′ ring of residues.

8610 | www.pnas.org/cgi/doi/10.1073/pnas.1407379111 Zhao et al.

Cartridges (Super Spe-ed silica gel, 5101; Applied Separations). Radioactivitywas determined by liquid scintillation analysis using a Tri-Carb 2810 TR. GC-MS data were recorded on a HP 6890 GC with the 5973 MS instrument.

GABAAR Membrane Preparation. The preparation method was modified fromthat of Squires et al. (30). Whole rat brains from Pel-Freez Biologicals storedat −80 °C were thawed and homogenized in 50 volumes of ice-cold 1 mMEDTA using a Brinkmann Polytron Homogenizer. The homogenate wascentrifuged at 1,000 × g for 10 min, and the supernatant was then centri-fuged at 25,000 × g for 30 min. The resulting pellets were suspended in50 volumes of 1 mM EDTA, packed into cellophane tubing, and dialyzedagainst distilled/deionized water in an ice-bath (1–2 L, three times for 2 h).The dialyzed suspension was then centrifuged at 25,000 × g for 30 min andthe pellets were stored at −80 °C.

Binding Assays. The rat brain membrane pellets from storage at −80 °C weresuspended in ice-cold buffer B [10 mM phosphate buffer (pH 7.5) containing300 mM NaCl]. Incubation mixtures consisted of membranes (125 μg protein)(45) and 0.5 nM [3H]EBOB or 1.5 nM [14C]TETS in 1.0 mL of buffer B. Afterincubation with shaking for 90 min at 37 °C, the mixtures were filteredthrough GF/C filters and rapidly rinsed three times with 5 mL of cold bufferB using a Brandel M-24 cell harvester. Tritium from bound [3H]EBOB wasquantitated by liquid scintillation counting (31). Rabiocarbon from [14C]TETSwas analyzed by AMS. The filter papers were collected, put in Eppendorftubes, and held up to 4 wk at 4 °C. Then, each filter loaded with protein wasplaced with 1 μL tributyrin carbon carrier in a quartz tube (∼6 × 30 mm,4 mm i.d.) nested inside two borosilicate glass culture tubes (10 × 75 mm in12 × 100 mm) and dried overnight in a vacuum centrifuge. An excess of CuO(∼40 mg) was added and the inner quartz vials were transferred to quartzcombustion tubes, evacuated, and sealed with a torch. The samples werecombusted at 900 °C for 3.5 h to oxidize all organic carbon to CO2 and thenreduced to filamentous carbon as previously described (46). Carbon sampleswere packed into aluminum sample holders, and carbon isotope ratios

were measured on the compact 1-MV AMS spectrometer at the LawrenceLivermore National Laboratory. Typical AMS measurement times were 3–5 min per sample, with a counting precision of 0.6–1.4% and a SD among3–10 measurements of 1–3%. The 14C/13C ratios of the protein sampleswere normalized to measurements of four identically prepared standardsof known isotope concentration (IAEA C-6, also known as ANU sucrose)and converted to units of femtograms TETS per microgram protein (47).Each experiment was performed in triplicate and repeated three times indetermining the mean and SEs. Curve fitting used the nonlinear (Fig. 3)or linear (Fig. 4) regression program with Prism Software Version 5.0(GraphPad Software Inc.).

Modeling the GABAAR Binding Sites. The GABAAR α1β2γ2 homology model wasbuilt with a GluCl template (PDB ID code 3RHW) (48) using previously pub-lished protocols (36, 37). Small molecules were parameterized with thePRODRG server (49) and docked into the pore using VinaLC (50). The protein–ligand system was embedded in a lipid bilayer and solvated. Atomistic sim-ulations were performed using GROMACS (51). For more details, see SI Ap-pendix, section S5.

ACKNOWLEDGMENTS. C.Z. thanks Prof. Lihong Qiu (China AgriculturalUniversity) for academic counsel and Berkeley laboratory colleagues AmandaLy, Breanna Ford, and Madhur Garg for assistance in manuscript preparation.S.H.H. and B.D.H. thank Jai Woong Seo for the [14C]TETS radio-TLC analysis.We thank the Livermore Computing Grand Challenge for computer time. Thiswork was supported in part by State Scholarship Fund 2011635139 providedby the China Scholarship Council (to C.Z.), National Institutes of Health Officeof the Director and the CounterACT Program National Institute of Neurolog-ical Disorders and Stroke Grant U54 NS079202 (to S.H.H. and B.D.H.), NationalInstitute of General Medical Sciences Grant 8P41GM103483 (to B.A.B.), andLaboratory Directed Research and Development Grant 13-LW-085 (to T.S.C.and F.L.). Portions of this work were performed under the auspices of the USDepartment of Energy by Lawrence Livermore National Laboratory underContract DE-AC52-07NA27344, Release LLNL-JRNL-649601.

1. Wood FC, Battye AE (1933) The condensation of sulphamide, dimethylsulphamide,and aniline-p-sulphonamide with formaldehyde. J Soc Chem Ind 56:346–349.

2. Hecht G, Henecka H (1949) Uber ein hochtoxisches kondensationsprodukt von sul-famid und formaldehyde. Angew Chem 61:365–366.

3. Hagen J (1950) Schwere vergiftungen in liner polstermöbelfabrik durch einen neuartigenhoch toxischen giftstoff (tetramethylendisulfotetramin). Dtsch Med Wochenschr75:183–184.

4. Esser T, Karu AE, Toia RF, Casida JE (1991) Recognition of tetramethylenedisulfo-tetramine and related sulfamides by the brain GABA-gated chloride channel anda cyclodiene-sensitive monoclonal antibody. Chem Res Toxicol 4(2):162–167.

5. Radwan M (1967) Translocation and metabolism of C14-labeled tetramine by douglas-fir, orchard grass, and blackberry. For Sci 13:265–273.

6. Owens J, Hok S, Alcaraz A, Koester C (2009) Quantitative analysis of tetramethyle-nedisulfotetramine (tetramine) spiked into beverages by liquid chromatography-tan-dem mass spectrometry with validation by gas chromatography-mass spectrometry.J Agric Food Chem 57(10):4058–4067.

7. Shan G, et al. (2000) Isotope-labeled immunoassays without radiation waste. Proc NatlAcad Sci USA 97(6):2445–2449.

8. Hecht G, Henecka H, Meisenheimer H (1949) US Patent 2,650,186.9. Croddy E (2004) Rat poison and food security in the People’s Republic of China: Focus

on tetramethylene disulfotetramine (tetramine). Arch Toxicol 78(1):1–6.10. Zhang SL, Ding MB (2001) Diagnosis and treatment of tetramine poisoning. Chin J Ind

Med 14:163–165.11. Whitlow KS, Belson M, Barrueto F, Nelson L, Henderson AK (2005) Tetramethylene-

disulfotetramine: Old agent and new terror. Ann Emerg Med 45(6):609–613.12. Jett DA, Yeung DT (2010) The CounterACT Research Network: Basic mechanisms and

practical applications. Proc Am Thorac Soc 7(4):254–256.13. Zhang Y, Su M, Tian DP (2011) Tetramine poisoning: A case report and review of the

literature. Forensic Sci Int 204(1-3):e24–e27.14. Li JM, Gan J, Zeng TF, Sander JW, Zhou D (2012) Tetramethylenedisulfotetramine in-

toxication presenting with de novo Status Epilepticus: A case series. Neurotoxicology33(2):207–211.

15. Zhang JS, Xiang P, Zhuo XY, Shen M (2014) Acute poisoning types and prevalence inShanghai, China, from January 2010 to August 2011. J Forensic Sci 59(2):441–446.

16. Voss E, Haskell AR, Gartenberg L (1961) Reduction of tetramine toxicity by sedativesand anticonvulsants. J Pharm Sci 50:858–860.

17. Bowery NG, Brown DA, Collins JF (1975) Tetramethylenedisulphotetramine: An in-hibitor of γ-aminobutyric acid induced depolarization of the isolated superior cervicalganglion of the rat. Br J Pharmacol 53(3):422–424.

18. Dray A (1975) Tetramethylenedisulphotetramine and amino acid inhibition in the ratbrain. Neuropharmacology 14(9):703–705.

19. Casida JE, et al. (1976) Structure-toxicity relationships of 2,6,7-trioxabicyclo(2.2.2)octanesand related compounds. Toxicol Appl Pharmacol 36(2):261–279.

20. Huang J, Casida JE (1996) Characterization of [3H]ethynylbicycloorthobenzoate([3H]EBOB) binding and the action of insecticides on the γ-aminobutyric acid-gated

chloride channel in cultured cerebellar granule neurons. J Pharmacol Exp Ther 279(3):1191–1196.

21. Qiu Z, Lan H, Zhang S, Xia Y, Huang S (2002) [Antidotal effects of vitamin B(6) andsodium dimercaptopropane sulfonate on acute poisoning with tetramethylene di-sulphotetramine in animals]. Zhonghua Nei Ke Za Zhi 41(3):186–188, Chinese.

22. Chen ZK, Lu ZQ (2004) Sodium dimercaptopropane sulfonate as antidote against non-metallic pesticides. Acta Pharmacol Sin 25(4):534–544.

23. Wright DW, et al. (2007) ProTECT: A randomized clinical trial of progesterone foracute traumatic brain injury. Ann Emerg Med 49(4):391–402.

24. Xie H, et al. (2010) The therapeutic effects of combination of γ-aminobutyric acid,sodium dimercaptopropane sultanate and vitamin B6 in large doses on liver and heartin rats with acute tetramine intoxication. Chin J Emergency Med 19:703–707.

25. Milbrath DS, Engel JL, Verkade JG, Casida JE (1979) Structure—toxicity relationships of1-substituted-4-alkyl-2,6,7-trioxabicyclo[2.2.2.]octanes. Toxicol Appl Pharmacol 47(2):287–293.

26. Wang J, Luo Y, Wang YA (2011) Treatment drugs against tetramine-induced seizure:Research advances. Int J Pharm Res 38:284–287.

27. Zolkowska D, et al. (2012) Characterization of seizures induced by acute and repeatedexposure to tetramethylenedisulfotetramine. J Pharmacol Exp Ther 341(2):435–446.

28. Cao Z, et al. (2012) Tetramethylenedisulfotetramine alters Ca²⁺ dynamics in culturedhippocampal neurons: Mitigation by NMDA receptor blockade and GABA(A) receptor-positive modulation. Toxicol Sci 130(2):362–372.

29. Shakarjian MP, Velíšková J, Stanton PK, Velíšek L (2012) Differential antagonism of tetra-methylenedisulfotetramine-induced seizures by agents acting at NMDA and GABA(A)receptors. Toxicol Appl Pharmacol 265(1):113–121.

30. Squires RF, Casida JE, Richardson M, Saederup E (1983) [35S]t-Butylbicyclophosphor-othionate binds with high affinity to brain-specific sites coupled to γ-aminobutyricacid-A and ion recognition sites. Mol Pharmacol 23(2):326–336.

31. Cole LM, Casida JE (1992) GABA-gated chloride channel: Binding site for 4′-ethynyl-4-n-[2,3-3H2]propylbicycloorthobenzoate ([3H]EBOB) in vertebrate brain and insecthead. Pestic Biochem Physiol 44:1–8.

32. Palmer CJ, Casida JE (1988) Two types of cage convulsant action at the GABA-gatedchloride channel. Toxicol Lett 42(2):117–122.

33. Ratra GS, Kamita SG, Casida JE (2001) Role of human GABA(A) receptor β3 subunit ininsecticide toxicity. Toxicol Appl Pharmacol 172(3):233–240.

34. Ratra GS, Casida JE (2001) GABA receptor subunit composition relative to insecticidepotency and selectivity. Toxicol Lett 122(3):215–222.

35. Law RJ, Lightstone FC (2008) Gaba receptor insecticide non-competitive antagonistsmay bind at allosteric modulator sites. Int J Neurosci 118(5):705–734.

36. Carpenter TS, Lau EY, Lightstone FC (2012) A role for loop F in modulating GABAbinding affinity in the GABA(A) receptor. J Mol Biol 422(2):310–323.

37. Carpenter TS, Lau EY, Lightstone FC (2013) Identification of a possible secondarypicrotoxin-binding site on the GABA(A) receptor. Chem Res Toxicol 26(10):1444–1454.

38. Hisano K, Ozoe F, Huang J, Kong X, Ozoe Y (2007) The channel-lining 6′ amino acid inthe second membrane-spanning region of ionotropic GABA receptors has more

Zhao et al. PNAS | June 10, 2014 | vol. 111 | no. 23 | 8611

NEU

ROSC

IENCE

profound effects on 4′-ethynyl-4-n-propylbicycloorthobenzoate binding than the 2′amino acid. Invert Neurosci 7(1):39–46.

39. Chen L, Durkin KA, Casida JE (2006) Structural model for γ-aminobutyric acid receptor

noncompetitive antagonist binding: Widely diverse structures fit the same site. Proc

Natl Acad Sci USA 103(13):5185–5190.40. Charon S, Taly A, Rodrigo J, Perret P, Goeldner M (2011) Binding modes of non-

competitive GABA-channel blockers revisited using engineered affinity-labeling re-

actions combined with new docking studies. J Agric Food Chem 59(7):2803–2807.41. Ozoe Y, Asahi M, Ozoe F, Nakahira K, Mita T (2010) The antiparasitic isoxazoline

A1443 is a potent blocker of insect ligand-gated chloride channels. Biochem Biophys

Res Commun 391(1):744–749.42. García-Reynaga P, Zhao C, Sarpong R, Casida JE (2013) New GABA/glutamate receptor

target for [³H]isoxazoline insecticide. Chem Res Toxicol 26(4):514–516.43. Olsen RW, Sieghart W (2008) International Union of Pharmacology. LXX. Subtypes of

γ-aminobutyric acid(A) receptors: Classification on the basis of subunit composition,

pharmacology, and function. Update. Pharmacol Rev 60(3):243–260.44. Ozoe Y (2013) γ-Aminobutyrate- and glutamate-gated chloride channels as targets of

insecticides. Adv Insect Physiol 44:211–286.

45. Bradford MM (1976) A rapid and sensitive method for the quantitation of microgramquantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72:248–254.

46. Ognibene TJ, Bench G, Vogel JS, Peaslee GF, Murov S (2003) A high-throughputmethod for the conversion of CO2 obtained from biochemical samples to graphite insepta-sealed vials for quantification of 14C via accelerator mass spectrometry. AnalChem 75(9):2192–2196.

47. Vogel JS, Love AH (2005) Quantitating isotopic molecular labels with accelerator massspectrometry. Methods Enzymol 402:402–422.

48. Hibbs RE, Gouaux E (2011) Principles of activation and permeation in an anion-selective Cys-loop receptor. Nature 474(7349):54–60.

49. Schüttelkopf AW, van Aalten DM (2004) PRODRG: A tool for high-throughput crys-tallography of protein-ligand complexes. Acta Crystallogr D Biol Crystallogr 60(Pt 8):1355–1363.

50. Zhang X, Wong SE, Lightstone FC (2013) Message passing interface and multithreadinghybrid for parallel molecular docking of large databases on petascale high perfor-mance computing machines. J Comput Chem 34(11):915–927.

51. Van Der Spoel D, et al. (2005) GROMACS: Fast, flexible, and free. J Comput Chem26(16):1701–1718.

8612 | www.pnas.org/cgi/doi/10.1073/pnas.1407379111 Zhao et al.

1

Supporting Information

The GABAAR Receptor Target of Tetramethylenedisulfotetramine

Title Authors Page

S1 Chronology of TETS chemistry and toxicology John E. Casida and

Chunqing Zhao

2

S2 Analysis of [13C]- and [14C]TETS Sung Hee Huang, Jun

Yang and Bruce D,

Hammock

5

S3 New findings from this study on effects of convulsants,

insecticides, anticonvulsants and other GABAAR

modulators on [14C]TETS and [3H]EBOB binding in rat

brain membranes

John E. Casida and

Chunqing Zhao

7

S4 Earlier studies on effects of convulsants, insecticides,

anticonvulsants and other GABAAR modulators on

[35S]TBPS and [3H]EBOB binding in rat brain membranes

and β3 homopentamer

John E. Casida and

Chunqing Zhao

8

S5 SI Methods: Molecular Dynamics simulation parameters Timothy S. Carpenter

and Felice C. Lightstone

9

S6 Binding site models for α1β2γ2 GABAAR with TETS,

EBOB, TBPS, TBPO, HEXS, TE3, TE5, PTX,

t-butyl-4-bromophenyldithiane, 4-ethynylphenylsilatrane,

fluralaner, lindane, BIDN, 12-ketoendrin, fipronil, and

-endosulfan

Timothy S. Carpenter

and Felice C. Lightstone

11

S7 Analysis of compound interactions with the 1’2’ ring of

residues

Timothy S. Carpenter

and Felice C. Lightstone

13

S8 Toxicity relative to GABAAR potency and types of cage

convulsant action

John E. Casida and

Chunqing Zhao

15

2

S1. Chronology of TETS chemistry and toxicology

John E. Casida and Chunqing Zhao

System and Observation Year and Reference

Chemistry

Synthesis 1933 (1) , 1950 (2), 1953 (3), Analogs 1991 (4)

Radiosynthesis 1967 (5), 2014 (this study)

Analysis 1993 (6), 2009 (7), 2014 (this study) Organismal toxicity

Toxicity recognized 1949 (8) Potent convulsant 1950 (2), 1957 (9) 1976 (10), 2012 (11) Type B cage convulsant 1988 (12)

Rodenticide 1953 (3), 2004 (13) Human poisoning and fatalities 1949 (8), 2005 (14), 2011 (15), 2012 (16), 2012 (17), 2013 (18)

Chemical threat agent 2001 (19), 2005 (14), 2010 (20),2011 (15)

Physiological disruptions

GABA evoked nerve currents 1975 (21-23), 1981(24) GABA stimulated chloride flux 1988 (25) NMDA induced Ca2+ signaling 2012 (26, 27)

Target site binding and modulator action

[35S]TBPS binding 1983 (28), 1996 (29) [3H]EBOB binding 1992 (30), 2005 (31) [14C]TETS binding 2014 (this study) GABAAR native but not (β3)5 2001 (32)

Candidate antidotes

GABA 2007 (33), 2010 (34) Diazepam, midazolam, phenobarbital 1961 (35), 1976 (10), 2007 (36), 2012 (27)

Allopregnanolone 1996 (29), 2005 (14), 2012 (11, 26) Na DMPS 2001 (37), 2002 (38) , 2004 (39), 2005 (14), 2011 (15, 40) Pyridoxine base, propofol, perampanel 2002 (38), 2007 (36), 2009 (41), 2011 (42)

1. Wood FC ,Battye AE (1933) The condensation of sulphamide, dimethylsulphamide, and

aniline-p-sulphonamide with formaldehyde. J Soc Chem Ind 56:346-349.

2. Hagen J (1950) Schwere vergiftungen in liner polstermöbelfabrik durch einen neuartigen hoch

toxischen giftstoff (tetramethylendisulfotetramin). Dtsch Med Wochenschr 75:183-184.

3. Hecht G, Henecka H ,Meisenheimer H (1949) A.P. 2 650 186, angemeldet: 10, 12, veroffentlicht:

25, 8, 1953, Farbenfabriken Bayer AG,.

4. Esser T, Karu A, Toia R ,Casida JE (1991) Recognition of tetramethylenedisulfotetramine and

related sulfamides by the brain GABA-gated chloride channel and a cyclodiene-sensitive

monoclonal antibody. Chem Res Toxicol 4:162-167.

5. Radwan M (1967) Translocation and metabolism of C14-labeled tetramine by douglas-Fir,

orchard grass, and blackberry. Forest Sci 13:265-273.

6. Guan F-Y, et al. (1993) GC/MS identification of tetramine in samples from human alimentary

intoxication and evaluation of artificial carbonic kidneys for the treatment of the victims. J Anal Toxicol 17:199-201.

7. Owens J, Hok S, Alcaraz A ,Koester C (2009) Quantitative analysis of

tetramethylenedisulfotetramine (tetramine) spiked into beverages by liquid

chromatography-tandem mass spectrometry with validation by gas chromatography−mass

spectrometry. J Agric Food Chem 57:4058-4067.

3

8. Hecht G ,Henecka H ( 1949) Uber ein hochtoxisches kondensationsprodukt von sulfamid und

formaldehyd. Angew Chem 61:365-366.

9. Haskell AR ,Voss E (1957) The pharmacology of tetramine (tetraethylenedisulfotetramine). J Am

Pharm Assoc (Baltim) 46:239-242.

10. Casida JE, et al. (1976) Structure-toxicity relationships of 2,6,7-trioxabicyclo[2.2.2]octanes and

related compounds. Toxicol Appl Pharmacol 36:261-279.

11. Zolkowska D, et al. (2012) Characterization of seizures induced by acute and repeated exposure

to tetramethylenedisulfotetramine. J Pharmacol Exp Ther 341:435-446.

12. Palmer CJ ,Casida JE (1988) Two types of cage convulsant action at the GABA-gated chloride

channel. Toxicol Lett 42:117-122.

13. Croddy E (2004) Rat poison and food security in the People's Republic of China: focus on

tetramethylene disulfotetramine (tetramine). Arch Toxicol 78:1-6.

14. Whitlow KS, Belson M, Barrueto F, Nelson L ,Henderson AK (2005)

Tetramethylenedisulfotetramine: old agent and new terror. Ann Emerg Med 45:609-613.

15. Zhang Y, Su M ,Tian D-P (2011) Tetramine poisoning: A case report and review of the literature.

Forensic Sci Int 204:e24-e27.

16. Li J-M, Gan J, Zeng T-F, Sander JW ,Zhou D (2012) Tetramethylenedisulfotetramine

intoxication presenting with de novo Status Epilepticus: A case series. NeuroToxicol 33:207-211.

17. Deng X, Li G, Mei R, Sun S (2012) Long term effects of teramine poisoning: an observational

study. Clin Toxicol 50: 172-175.

18. Zhang JS, Xiang P, Zhuo XY ,Shen M (2013) Acute poisoning types and prevalence in Shanghai,

China, from January 2010 to August 2011. J Forensic Sci:DOI: 10.1111/1556-4029.12334.

19. Zhang SL ,Ding MB (2001) Diagnosis and treatment of tetramine poisoning. Chin J Ind Med

14:163-165.

20. Jett DA ,Yeung DT (2010) The CounterACT Research Network: basic mechanisms and practical

applications. Proc Am Thoracic Soc 7:254-256.

21. Bowery NG, Brown DA ,Collins JF (1975) Tetramethylenedisulphotetramine: an inhibitor of

γ-aminobutyric acid induced depolarization of the isolated superior cervical ganglion of the rat.

Br J Pharmacol 53:422-424.

22. Dray A (1975) Tetramethylenedisulphotetramine and amino acid inhibition in the rat brain.

Neuropharmacol 14:703-705.

23. Large WA (1975) Effect of tetramethylenedisulphotetramine on the membrane conductance

increase produced by γ-aminobutyric acid at the crab neuromuscular junction. Br J Pharmacol 53:598-599.

24. Roberts CJ, James VA, Collins JF, Walker RJ (1981) The action of seven convulsants as

antagonists of the GABA response of Limulus neurons. Comp Biochem Physiol C Pharmacol Toxicol Endocrinol 244:802-806.

25. Obata T, Yamamura HI, Malatynska E, Ikeda M, Laird H, Palmer CJ, Casida JE (1988)

Modulation of γ-aminobutyric acid-stimulated chloride influx by bicycloorthocarboxylates,

bicyclophosphorus esters, polychlorocycloalkanes and other cage convulsants. J Pharmacol Exp

Ther 279:1191-1196.

26. Cao Z, et al. (2012) Tetramethylenedisulfotetramine alters Ca2+ dynamics in cultured

hippocampal neurons: mitigation by NMDA receptor blockade and GABAA receptor-positive

modulation. Toxicol Sci 130:362-372.

27. Shakarjian M, Velíšková J, Stanton P ,Velíšek L (2012) Differential antagonism of

tetramethylenedisulfotetramine- induced seizures by agents acting at NMDA and GABAA

receptors. Toxicol Appl Pharmacol 265:113-121.

28. Squires RF, Casida JE, Richardson M ,Saederup E (1983) [35S]t-Butylbicyclophosphorothionate

binds with high affinity to brain-specific sites coupled to γ-aminobutyric acid-A and ion

recognition sites. Mol Pharmacol 23:326-336.

29. Concas A, et al. (1996) Functional correlation between allopregnanolone and [35S]-TBPS binding in the brain of rats exposed to isoniazid, pentylenetetrazol or stress. Br J Pharm 118:839-846.

4

30. Cole LM ,Casida JE (1992) GABA-gated chloride channel: Binding site for

4′-ethynyl-4-n-[2,3-3H2]propylbicycloorthobenzoate ([3H]EBOB) in vertebrate brain and insect

head. Pestic Biochem Physiol 44:1-8.

31. Maksay G ,Bíró T (2005) High affinity, heterogeneous displacement of [3H]EBOB binding to

cerebellar GABAA receptors by neurosteroids and GABA agonists. Neuropharmacol 49:431-438.

32. Ratra G, Kamita S ,Casida JE (2001) Role of human GABAA receptor β subunit in insecticide

toxicity. Toxicol Appl Pharmacol 172:233-240.

33. Sun P, Han J, Weng Y (2007) The antidotal effects of high-dosage γ-aminobutyric acid on acute

teramine poisoning as compared with sodium dimercaptopropane sulfonate. J Huanzhong Uni Sci Tech 27:419-421

34. Xie H et al. (2010) The therapeutic effects of combination of γ-aminobutyric acid, sodium

dimercaptopropane sultanate and vitamin B6 in large doses on liver and heart in rats with acute

tetramine intoxication. Chin J Emergency Med 19:703-707.

35. Voss E, Haskell AR ,Gartenberg L (1961) Reduction of tetramine toxicity by sedatives and

anticonvulsants. J Pharm Sci 50:858-860.

36. Zhou ZQ, et al. (2007) Causes of mortality and emergency treatments in patients with

tetramethylene disulfotetramine poisoning. J Clin Anesthesiol 23 3:204-205.

37. Zhang CY, et al (2001) Effect of sodium dimercaptopropanesulfonate on antagonism of

tetramethylenedisulphotetramine to GABA receptor. Acta Pharmacol Sin 22:435-439

38. Qiu Z, Lan H, Zhang S, Xia Y ,Huang S (2002) Antidotal effects of vitamin B (6) and sodium

dimercaptopropane sulfonate on acute poisoning with tetramethylene disulphotetramine in

animals. Zhonghua nei ke za zhi [Chin J Internal Med] 41:186-188.

39. Chen Z ,Lu Z (2004) Sodium dimercaptopropane sulfonate as antidote against non-metallic

pesticides. Acta Pharmacol Sinica 25:534-544.

40. Wang J, Luo Y ,Wang YA (2011) Treatment drugs against tetramine-induced seizure: research

advances. Int J Pharm Res 38:284-287.

41. Wang H-L, et al. (2009) Inhibitory effects of propofol on tetramine-induced epilepsy in rats and

its mechanism. Acad J Chin PLA Postgrad Med Sch 2:200-202.

42. Zolkowska D, et al (2011) Perampanel, a potent AMPA receptor antagonist, protects against

tetramethylenedisulfotetramine-induced seizures. American Epilepsy Society Abst. 3.069.

5

S2. Analysis of [13C]- and [14C]TETS

Sung Hee Hwang, Jun Yang and Bruce D. Hammock

1. GC-MS conditions

GC-MS data were recorded on a HP 6890 GC with 5973 MS instrument (Santa Clara,

CA). The sample (1 μL) was injected into a DB-5 MS column (30 m x 0.25 mm ID x 0.25 µm,

Folsom, CA). The injector port was at 250 oC with splitless mode (5 min solvent delay).

Helium was used as the carrier gas at 0.8 mL/min. Initial temperature was 40 oC for 3 min and

ramped at 8 oC/min to 300 oC for 3 min.

2. GC/MS detection

Full scan mode was used for [13C}TETS and selected ion monitoring (SIM) for [14C]TETS.

3. Analysis of unlabeled TETS and [13C]TETS by GC-MS in full-scan mode.

A. unlabeled TETS. B. [13C]TETS (mono-, di-, tri- and tetra-13C labelled TETS).

4. Radio-TLC analysis of [14C]TETS using a Bioscan System 200 Imaging Scanner

(Bioscan, Washington, DC)

6

5. Sensitivity of detection methods for TETS

Current analytical methods (e.g. LC-MS/MS = 0.10 μg/mL and GC-MS = 0.15 μg/mL,

respectively) are not sensitive enough to quantify TETS from many biological matrices and at

toxicologically relevant concentrations. However tandem use of an HPLC with AMS has a

typical limit of quantification of 2 - 20 attomoles for the 14C fractions Therefore, LC-AMS was

used in our study.

7

S3. New findings from this study on effects of convulsants, insecticides,

anticonvulsants and other GABAAR modulators on [14C]TETS and [3H]EBOB

binding in rat brain membranes

John E. Casida and Chunqing Zhao

Inhibition of binding (%±SE) b

Compounds a Concentration (µM) [14C]TETS [3H]EBOB

convulsants and insecticides

1 TETS 1 55 ± 9 57 ± 2

2 TETS 10 100 b 100 b

3 HEXS 1 26 ± 6 47 ± 1

4 TE3 1 7 ± 3 20 ± 1

5 TE5 1 11 ± 5 23 ± 1

6 TBPO 1 41± 7 78 ± 3

7 picrotoxinin 1 64 ± 1 51 ± 5

8 t-butyl-4-bromophenyldithiane 1 100 ± 6 108 ± 4

9 4-ethynylphenylsilatrane 1 53 ± 5 67 ± 2

10 Isoxazoline fluralaner 1 10 ± 4 5 ± 2

11 Isoxazoline fluralaner 10 47 ± 4 40 ± 3

12 lindane 1 87 ± 3 66 ± 4

13 dinitrile BIDN 1 67 ± 4 72 ± 6

14 12-ketoendrin 1 96 ± 1 107 ± 6

15 fipronil 1 56 ± 7 69 ± 2

16 α-endosulfan 1 55 ± 5 103 ± 3

anticonvulsants

17 amobarbital 1 21 ± 2 5± 1

18 Na pentobarbital 1 14 ± 3 19 ± 1

19 Na pentobarbital 10 28 ± 4 26 ± 1

20 diazepam 1 10 ± 1 18 ± 1

21 diazepam 10 21 ± 5 20 ± 2

22 midazolam 1 30 ± 6 18 ± 1

23 midazolam 10 40 ± 2 30 ± 2

24 flurazepam 1 31 ± 3 38 ± 6

others

25 5α-pregnan-3β-ol-20-one 1 18 ± 1 12 ± 1

26 5α-pregnan-3β-ol-20-one 10 62 ± 4 16 ± 3

27 avermectin Ba1 1 66 ± 7 47 ± 4

28 baclofen 1 -17 ± 6 19 ± 1

29 isoguvacine hydrochloride 1 46 ± 4 28 ± 3

30 NMDA 1 59 ± 4 12 ± 3

31 propofol 1 31 ± 8 28 ± 5

32 propofol 10 57 ± 6 54 ± 1

33 pyridoxine base 1 36 ± 8 2 ± 2

34 pyridoxine base 10 45 ± 8 3 ± 2

35 ethanol 300 mM -70 ± 3 34 ± 6

36 Na DMPS 1 -17 ± 5 2 ± 3

37 Na DMPS 10 -16 ± 5 9 ± 5

38 bicuculline 1 -68 ± 8 -59 ± 5 a Structures for 1-16 are given in Fig. 1 and for the other compounds in Merck Index 15th Ed. (1)

b Nonspecific binding was always determined with 10 µM TETS.

1. Merck (2013) The Merck Index, 15th Ed. The Royal Society of Chemistry. 1896 pp

No.

8

S4. Earlier studies on effects of convulsants, insecticides, anticonvulsants and other GABAAR modulators on [35S]TBPS and [3H]EBOB binding in rat

brain membranes and β3 homopentamer John E. Casida and Chunqing Zhao

IC50 (µM)

Compounds Brain membranes (β3)5

Name [35S]TBPS a [35S]TBPS b [3H]EBOB cd [3H]EBOB e

GABA 0.34 0.89 1.12 TBPS 0.062 0.115 0.059

phenobarbital 58 (-); 110 (+)

1/2 TETS 0.82 >10

6 TBPO 0.073

7 picrotoxinin 0.19 0.34 0.032

8 t-butyl-4-bromophenyldithiane 0.064

9 4-ethynylphenylsilatrane >10

12 lindane 0.39 0.83

14 12-ketoendrin 0.036 0.014 0.00045

15 fipronil 1.66 f 0.0024

16 α-endosulfan 0.1 – 0.2 0.11 0.00047

20/21 diazepam ~10

27 avermectin B1a 1.77

28 baclofen >100

29 isoguvacine hydrochloride 0.93

35 ethanol 5.1×105

38 bicuculline >100 0.028

a. Squires RF, Casida JE, Richardson M ,Saederup E (1983) [35S]t-Butylbicyclophosphorothionate binds with high

affinity to brain-specific sites coupled to -aminobutyric acid-A and ion recognition sites. Mol Pharmacol

23:326-336.

b. Casida JE ,Lawrence LJ (1985) Structure-activity correlations for interactions of bicyclophosphorus esters and some

polychlorocycloalkane and pyrethroid insecticides with the brain-specific t-butylbicyclophosphorothionate receptor.

Environ Health Perspect 61:123-132.

c. Cole LM ,Casida JE (1986) Polychlorocycloalkane insecticide-induced convulsions in mice in relation to disruption

of the GABA-regulated chloride ionophore. Life Sci 39:1855-1862.

d. Cole LM ,Casida JE (1992) GABA-gated chloride channel: Binding site for

4′-ethynyl-4-n-[2,3-3H2]propylbicycloorthobenzoate ([3H]EBOB) in vertebrate brain and insect head. Pestic Biochem

Physiol 44:1-8.

e. Ratra G, Kamita S ,Casida J (2001) Role of human GABAA receptor β subunit in insecticide toxicity. Toxicol Appl

Pharmacol 172:233-240.

f. Ikeda T, et al. (2001) Fipronil modulation of -aminobutyric acidA receptors in rat dorsal root ganglion neurons. J

Pharmacol Exp Ther 296:914-921.

No

9

S5. SI Methods: Molecular Dynamics simulation parameters Timothy S. Carpenter and Felice C. Lightstone

The 122 GABAAR homology model was built based upon the GluCl (1) template using

previously published methodologies (2, 3). The GluCl structure is reported to have an open

conformation of the TM pore region. The model was inserted into a preformed and equilibrated

POPC (palmitoyl-oleoyl phosphatidylcholine) lipid bilayer. The system was solvated and had

counter-ions added to neutralize the charge. Further Na+/Cl− ions were added to create an

effective concentration of 0.15 M. The final system consists of ~210,000 atoms. This apo

GABAAR system was energy minimized using steepest descents. The system was equilibrated

during a 30 ns simulation. This equilibrated GABAAR structure was used for molecular docking

of the small compounds [TETS (1), TBPS, HEXS (3), TE3 (4), TE5 (5), and TBPO (6)], and for

the remaining EBOB, PTX (7), t-butyl-4-bromophenyldithiane (8), 4-ethynylphenylsilatrane (9),

fluralaner (10), lindane (12), BIDN (13), 12-ketoendrin (14), fipronil (15), and -endosulfan

(16)). Each system had the compound molecularly docked into the pore using VinaLC (4). The

ligands were prepared for docking using MGLTools (5, 6). The docking was performed on

30x30x30 Å grid with 0.33 Å resolution that was centered on the pore region to examine

potential binding sites at this vicinity. Twelve exhaustiveness docking runs were carried out per

compound. Any water molecules in the pore that overlapped the small compound were

removed, and the system was again energy minimized. The ligand-bound,

membrane-embedded protein system was then subjected to 40 ns MD simulation production

to identify the equilibrated ligand position and the protein-ligand interactions.

All MD simulations were run using GROMACS 4.5.5 (7). The Berger et al. force field was used

for the POPC molecules (8), the gromos53a6 force field was used for the protein and the small

compounds (9). The parameters for the small compounds were generated using the PRODRG

server (10). The SPC model was used to represent the water (11). Simulations were

performed at 323 K using the Nosé-Hoover thermostat (12) with τT = 0.5 ps. The pressure was

maintained at 1 bar using a semi-isotropic Parrinello-Rahman barostat (13) with τP = 1 ps and

a compressibility of 4.5x10-5 bar-1. Bonds lengths were constrained using the LINCS algorithm

(14), allowing a 2-fs time step to be used. Non-bonded interactions were truncated at 14 Å, and

the neighbor list updated every 10 ps. The long-range electrostatic interactions were

calculated using the particle mesh Ewald method (15). Images were prepared and analyses

were carried out using GROMACS, locally written scripts, and VMD (16).

1. Hibbs R and Gouaux E (2011) Principles of activation and permeation in an anion-selective

Cys-loop receptor. Nature 474:54-60.

10

2. Carpenter TS, Lau EY ,Lightstone FC (2012) A role for loop F in modulating GABA binding

affinity in the GABAA receptor. J Mol Biol 422:310-323.

3. Carpenter TS, Lau EY ,Lightstone FC (2013) Identification of a possible secondary

picrotoxin-binding site on the GABAA receptor. Chem Res Toxicol 26:1444-1454.

4. Zhang X, Wong S, and Lightstone F (2013) Message passing interface and multithreading hybrid

for parallel molecular docking of large databases on petascale high performance computing

machines. J. Comput. Chem. 34:915–927.

5. Sanner F. (1999) Python: A programming language for software integration and development. J.

Mol. Graphics Mod. 17:57-61

6. Morris, G. M., Huey, R., Lindstrom, W., Sanner, M. F., Belew, R. K., Goodsell, D. S. and Olson,

A. J. (2009) Autodock4 and AutoDockTools4: automated docking with selective receptor

flexiblity. J. Comput. Chem. 16:2785-91

7. Van Der Spoel D, et al. (2005) GROMACS: Fast, flexible, and free. J Comput Chem

26:1701–1718.

8. Berger, O., O. Edholm, and F. Jahnig, (1997) Molecular dynamics simulations of a fluid bilayer of

dipalmitoylphosphatidycholine at full hydration, constant pressure and constant temperature.

Biophys. J., 72:2002-2013.

9. Oostenbrink, C., et al., (2004) A biomolecular force field based on the free enthalpy of hydration

and solvation: The GROMOS force-field parameter sets 53A5 and 53A6. J. Comput. Chem.

25:1656-1676.

10. Schüttelkopf A and van Aalten D (2004) PRODRG: a tool for high-throughput crystallography of

protein-ligand complexes. Acta Crystallogr D60, 1355–1363.

11. Berendsen, H.J.C., et al., (1981) Interaction models for water in relation to protein hydration, in

Intermolecular Forces, B. Pullman, Editor, D. Reidel: Dordrecht, The Netherlands. p. 331-342.

12. Hoover, W.G., (1985) Canonical dynamics: equilibrium phase-space distributions. Phys. Rev.,

A31:1695-1697.

13. Parrinello, M. and A. Rahman, (1981) Polymorphic transitions in single-crystals - a new

molecular-dynamics method. J. Appl. Physics, 52:7182-7190.

14. Hess, B., et al., (1997) LINCS: A linear constraint solver for molecular simulations. J. Comp.

Chem., 18:1463-1472.

15. Darden, T., D. York, and L. Pedersen, (1993) Particle mesh Ewald - an N.log(N) method for Ewald

sums in large systems. J. Chem. Phys., 98(12):10089-10092.

16. Humphrey, W., Dalke, A. and Schulten, K. (1996). VMD: visual molecular dynamics. J. Mol. Graphics, 14:33-38.

11

S6. Binding site models for α1β2γ2 GABAAR with TETS, EBOB, TBPS, TBPO, HEXS, TE3, TE5, PTX, t-butyl-4-bromophenyldithiane,

4-ethynylphenylsilatrane, fluralaner, lindane, BIDN, 12-ketoendrin, fipronil,

and -endosulfan Timothy S. Carpenter and Felice C. Lightstone

The equilibrated conformations of TETS (1), EBOB , TBPS, TBPO (6), HEXS (3), TE3 (4),

TE5 (5), PTX (7), t-butyl-4-bromophenyldithiane (8), 4-ethynylphenylsilatrane (9), fluralaner

(10), lindane (12), BIDN (13), 12-ketoendrin (14), fipronil (15), and -endosulfan (16) are

shown following their 40 ns MD simulation. The overlapping conformations display

consistency with the experimental results. All of the compounds that inhibit TETS binding

12

show overlap of the 2’ “contact zone”, which is the TETS-binding ring of residues (the 2’

“contact zone” is indicated as between the red dashed lines). All compounds fall within the

region defined by the 2’ to 9’ residues. Lindane was observed to persist in three separate

sites with approximately equal frequency. Multiple lindane LGIC sites have previously been

suggested (1). The lowest two of these sites overlap the TETS binding region.

1. Alqazzaz M, Thompson AJ, Price KL, Breitinger HG, & Lummis SC (2011) Cys-loop receptor

channel blockers also block GLIC. Biophys. J. 101:2912-2918.

13

S7. Analysis of compound interactions with the 1’2’ ring of residues Timothy S. Carpenter and Felice C. Lightstone

The number of contacts that each compound made to either the polar or hydrophobic residues of the 1’2’

ring were calculated as a percentage of the total contacts that each compound made to the GABAAR. These

percentages were then normalized to the size of the compound. A normalized contact percentage of >1.4%

generally equates to ‘significant’ contacts, signifying an interaction that is far beyond a non-specific

occurance. Dashed lines representing significant normalized contacts of 1.4% or greater divide the plot into

four quadrants.

The numbering and chemical structures of the compounds are displayed in text Fig. 1.

The interactions of the compounds were examined with the residues at the 1’2’ positions of

the M2 helices. The percentage of each compound’s total contacts that were made to either

the polar residues (1T1’ and 2S2’) or the hydrophobic residues (1V2’ and 2V1’/A2’)

were measured, normalized to compound size, and plotted against each other.

Four categories of interactions at the 1’2’ region are shown in the figure and are defined as

quadrants Q1-Q4:

Q1 – Potentially significant hydrophobic interactions are between the compound and the

1’2’ residues. For example, EBOB and BIDN (13) fall into this quadrant.

Q2 – Significant hydrophobic and polar interactions are between the compound and the

1’2’ residues. TBPO (6) and TBPS fall into this quadrant.

Q3 – Some interactions exist between the compound and the 1’2’ residues, but the most

specific/important interactions are with other residues within the pore and outside the 1’2’

rings [e.g., PTX (7)].

Q4 – Significant/specific polar interactions are between the compound and the 1’2’

residues [i.e., TETS (1) and TE3 (4)].

14

The results of classifying the compounds in the different quadrants are consistent with

previously published data on compound binding to specific subunit configurations. In

general, TETS (Q4) requires polar interactions for binding and explains why the TETS

binding is abolished in the 3 homopentamer (where all the 1’2’ residues are hydrophobic

[1]). TBPO and TBPS, like TETS, also require polar interactions. However, hydrophobic

interactions are also required for TBPO and TBPS binding. The requirement of having both

hydrophobic and hydrophilic residues in the 1’2’ ring is consistent with TBPS only binding

significantly when both and subunit subtypes make up the GABAAR.[2,3] BIDN (and

to a lesser extent EBOB) relies on mainly hydrophobic interactions at this region. BIDN

binding to the insect RDL receptor is reduced when hydrophobic A2’ is mutated to polar

S2’ [4], showing the need for hydrophobic residues for binding. EBOB also has its binding

reduced when A2’ is mutated to S2’ in the 3 homopentamer, though it appears that the

more crucial interaction is outside of the 1’2’ ring at the 6’ residue [5]. Hence, the 1’2’

percentage hydrophobic contacts are lower for EBOB than BIDN. As the compounds in

Q3 do not appear to make large percentage contacts to either polar or hydrophobic residues,

we hypothesize that their major interactions with the GABAAR pore take place outside of

the 1’2’ ring. This is true for PTX, which has interactions with 6’ [6]. However, several of

the Q3 compounds are greatly extended in length and make interactions with two, if not

three, of the pore-lining rings of M2 residue.

Furthermore, the two established types of cage convulsants [7] (‘elongated’ Type A, and

‘compact’ Type B) fall into two distinct categories (S8). Type A convulsants are shown in

blue, and Type B are shown in red. While both Type A and Type B cage convulsants may

act at a similar site within the pore, Type B compounds make a significant number of polar

interactions, and Type A compounds do not.

1. Ratra GS, Kamita S ,Casida JE (2001) Role of human GABAA receptor β3 subunit in insecticide

toxicity. Toxicol Appl Pharmacol 172:233-240.

2. Zezula J, Slany A, and Sieghart W (1996) Interaction of allosteric ligands with GABAA receptors

containing one, two, or three different subunits. Eur J Pharmacol 301:207-214.

3. Jursky F, Fuchs K, Buhr A, Tretter V, Sigel E, and Sieghart W (2000) Identification of amino acid

residues of GABAA receptor subunits contributing to the formation and affinity of the tert--butylbicyclophosphorothionate binding site. J Neurochem 74:1310-1316

4. Buckingham SD, Biggin PC, Sattelle BM, Brown LA and Sattelle DB (2005) Insect GABA

receptors: splicing, editing, and targeting by antiparasitics and insecticides. Mol Pharmacol

68:942-951

5. Hisano K, Ozoe F, Huang J, Kong X, and Ozoe Y (2007) The channel-lining 6′ amino acid in the

second membrane-spanning region of ionotropic GABA receptors has more profound effects on

4′-ethynyl-4-n-propylbicycloorthobenzoate binding than the 2′ amino acid. Invert Neurosci 7:39-46

6. Gurley D, Amin J, Ross PC, Weiss DS, White G (1994) Point mutations in the M2 region of the

alpha, beta, or gamma subunit of the GABAA channel that abolish block by picrotoxin. Receptor

Channels 3:13-20.

7. Palmer CJ and Casida JE (1988) Two types of cage convulsant action at the GABA-gated chloride

channel. Toxicol Lett 42:117-122.

15

S8. Toxicity relative to GABAAR potency and types of cage convulsant action John E. Casida and Chunqing Zhao

GABAAR radioligand specific binding

Inhibition at 1 µM (%) b IC50 (µM) [3H]EBOB IC50 (µM) e Cl- flux IC50

(µM)d Type of action and compound LD50 mg/ kg a [14C]TETS

[3H]EBOB

[35S]TBPS c [3H]EBOB d

native α1β3γ2 (β3)5

Type B - small or compact

TETS 0.24 55 57 0.82

1.14 1.27 > 10

TBPS 0.053

0.017 0.77 0.11 0.055 0.059 6

TBPO 0.036 41 78 0.073

Type A - large or elongated

EBOB 0.11f

0.001 f 0.0012

0.29

t-Butyl-4-bromophenyldithiane 1.2 100 108

0.018 0.36 0.16 0.87 0.65 g

α-Endosulfan 76 55 103

0.017 0.0073 0.016 0.00047 0.99

Fipronil 41 56 69

3.3 2.5 0.033 0.0024 >50

a Mouse ip data from various publications of the Berkeley laboratory.

b Rat brain membranes, this study.

c .Squires RF, Casida JE, Richardson M ,Saederup E (1983) [35S]t-Butylbicyclophosphorothionate binds with high affinity to brain-specific sites coupled to γ-aminobutyric acid-A

and ion recognition sites. Mol Pharmacol 23:326-336. Study with rat brain membranes.

d. Huang J ,Casida JE (1996) Characterization of [3H]ethynylbicycloorthobenzoate ([3H]EBOB) binding and the action of insecticides on the γ-aminobutyric acid-gated chloride

channel in cultured cerebellar granule neurons. J Pharmacol Exp Ther 279:1191-1196.

e. Ratra GS, Kamita SG ,Casida JE (2001) Role of human GABAA receptor β3 subunit in insecticide toxicity. Toxicol Appl Pharmacol 172:233-240.

f. Palmer CJ, Cole LM, Larkin JP, Smith IH ,Casida JE (1991) 1-(4-Ethynylphenyl)-4-substituted-2,6,7-trioxabicyclo[2.2.2]octanes: effect of 4-substituent on toxicity to houseflies

and mice and potency at the GABA-gated chloride channel. J Agric Food Chem 39:1329-1334. Study with mouse brain membranes.

g t-Butyl-4-ethynylphenyldithiane.