Para- and Ortho-Substitutions Are Key Determinants of Polybrominated Diphenyl Ether Activity toward Ryanodine Receptors and Neurotoxicity

Background Polybrominated diphenyl ethers (PBDEs) are widely used flame retardants that bioaccumulate in human tissues. Their neurotoxicity involves dysregulation of calcium ion (Ca2+) signaling; however, specific mechanisms have yet to be defined. Objective We aimed to define the structure–activity relationship (SAR) for PBDEs and their metabolites toward ryanodine receptors type 1 (RyR1) and type 2 (RyR2) and to determine whether it predicts neurotoxicity. Methods We analyzed [3H]ryanodine binding, microsomal Ca2+ fluxes, cellular measurements of Ca2+ homeostasis, and neurotoxicity to define mechanisms and specificity of PBDE-mediated Ca2+ dysregulation. Results PBDEs possessing two ortho-bromine substituents and lacking at least one para-bromine substituent (e.g., BDE-49) activate RyR1 and RyR2 with greater efficacy than corresponding congeners with two para-bromine substitutions (e.g., BDE-47). Addition of a methoxy group in the free para position reduces the activity of parent PBDEs. The hydroxylated BDEs 6-OH-BDE-47 and 4′-OH-BDE-49 are biphasic RyR modulators. Pretreatment of HEK293 cells (derived from human embryonic kidney cells) expressing either RyR1 or RyR2 with BDE-49 (250 nM) sensitized Ca2+ flux triggered by RyR agonists, whereas BDE-47 (250 nM) had negligible activity. The divergent activity of BDE-49, BDE-47, and 6-OH-BDE-47 toward RyRs predicted neurotoxicity in cultures of cortical neurons. Conclusions We found that PBDEs are potent modulators of RyR1 and RyR2. A stringent SAR at the ortho and para position determined whether a congener enhanced, inhibited, or exerted nonmonotonic actions toward RyRs. These results identify a convergent molecular target of PBDEs previously identified for noncoplanar polychlorinated biphenyls (PCBs) that predicts their cellular neurotoxicity and therefore could be a useful tool in risk assessment of PBDEs and related compounds.

Worldwide use of chemically stable poly brominated diphenyl ethers (PBDEs) as flame retardants in consumer products has increased since the 1970s (Costa et al. 2008). Levels of PBDEs in the environment (Hale et al. 2003;Law et al. 2006;Oros et al. 2005) and tissues of invertebrates, fish, birds, and mammals have increased in the last 30 years (Yogui and Sericano 2009). In humans, PBDEs have been detected in numerous tis sues, with serum concentrations reported as high as 50 ng/g lipid (Petreas et al. 2003). PBDE concentrations in breast milk dou bled every 5 years between 1972 and 1997 (Rahman et al. 2001;She et al. 2007). The predominant routes of exposure are diet and inhalation (Allen et al. 2007;Costa et al. 2008;Rose et al. 2010;Schecter et al. 2005). PBDEs with less than five bromine substitu tions, such as 2,2´,4,4´tetrabromo diphenyl ether (BDE47) and 2,2´,4,4´,5pentabromo diphenyl ether (BDE99), have attracted the greatest concern because they show the great est bio accumula tion (Darnerud et al. 2001) and can be formed by debromina tion of high ersubstituted PBDEs (Gerecke et al. 2005;Stapleton et al. 2006;Stapleton and Dodder 2008).
PBDEs share structural similarities to a subset of the polychlorinated biphenyls (PCBs) that have more than one chlorosub stituent in the ortho position, in particu lar, non coplanarity of their phenyl rings. Like the orthosubstituted PCBs (Denison and Nagy 2003;Van den Berg et al. 2006), PBDEs (Peters et al. 2004) have very weak or negli gible activity at the aryl hydro carbon recep tor, but disrupt calcium ion (Ca 2+ ) signaling in several cell types (Dingemans et al. 2008(Dingemans et al. , 2010. One mechanism by which orthosubstituted PCBs disrupt Ca 2+ signaling and contribute to neuro toxicity is via inter actions with ryanodine receptors (RyRs) Wong and Pessah 1996). For example, 2,2´,3,5´,6pentachloro biphenyl (PCB95) at nano molar concentrations sen sitizes RyR Ca 2+ channels in rat hippocam pal microsomal preparations Wong et al. 1997a), and alters neuro plasticity in rat hippocampal slices ). Developmental exposure to PCB95 or Aroclor 1254 results in behavioral abnormalities in rats and altered patterns of RyR expression in brain (Schantz et al. 1997;Wong et al. 1997b;Yang et al. 2009). In rats exposed peri natally, PCB95 alters the balance of excitatory/inhibitory currents and disrupts the tono topic map of the primary auditory cortex (Kenet et al. 2007). Identification of RyRs as a relevant molecular target in PCB neuro toxicity may be generalized to related non coplanar persistent organic pollutants such as PBDEs. Evidence of a convergent mechanism by which PCBs and PBDEs dys regu late RyRs would be of toxicological significance considering that these tightly regulated Ca 2+ channels reside within spe cialized regions of sarcoplasmic/endoplasmic reticulum (SR/ER) membranes where they contribute to and regulate essential aspects of Ca 2+ signaling ).
In the present study we examined the structure-activity relationship (SAR) of PBDEs and their metabolites toward the two major RyR isoforms found in the brain and striated muscle (RyR1 and RyR2). Our findings indicate that PBDEs are potent modulators of both RyR isoforms. Para and orthosubstitutions are a key structural deter minant of this activity and are predictive of their neurotoxicity.
Background: Polybrominated diphenyl ethers (PBDEs) are widely used flame retardants that bioaccumulate in human tissues. Their neurotoxicity involves dysregulation of calcium ion (Ca 2+ ) signaling; however, specific mechanisms have yet to be defined. oBjective: We aimed to define the structure-activity relationship (SAR) for PBDEs and their metabolites toward ryanodine receptors type 1 (RyR1) and type 2 (RyR2) and to determine whether it predicts neurotoxicity. Methods: We analyzed [ 3 H]ryanodine binding, microsomal Ca 2+ fluxes, cellular measurements of Ca 2+ homeo stasis, and neurotoxicity to define mechanisms and specificity of PBDE-mediated Ca 2+ dysregulation. results: PBDEs possessing two ortho-bromine substituents and lacking at least one para-bromine substituent (e.g., BDE-49) activate RyR1 and RyR2 with greater efficacy than corresponding congeners with two para-bromine substitutions (e.g., BDE-47). Addition of a methoxy group in the free para position reduces the activity of parent PBDEs. The hydroxylated BDEs 6-OH-BDE-47 and 4´-OH-BDE-49 are biphasic RyR modulators. Pretreatment of HEK293 cells (derived from human embryonic kidney cells) expressing either RyR1 or RyR2 with BDE-49 (250 nM) sensitized Ca 2+ flux triggered by RyR agonists, whereas BDE-47 (250 nM) had negligible activity. The divergent activity of BDE-49, BDE-47, and 6-OH-BDE-47 toward RyRs predicted neurotoxicity in cultures of cortical neurons. conclusions: We found that PBDEs are potent modulators of RyR1 and RyR2. A stringent SAR at the ortho and para position determined whether a congener enhanced, inhibited, or exerted non monotonic actions toward RyRs. These results identify a convergent molecular target of PBDEs previously identified for non coplanar polychlorinated biphenyls (PCBs) that predicts their cellular neuro toxicity and therefore could be a useful tool in risk assessment of PBDEs and related compounds. Membrane preparations. Microsomal membrane preparations enriched in RyR1 or RyR2 were isolated from skeletal muscle and heart, respectively, from male New Zealand White rabbits (Charles River Laboratories, Hollister, CA) as described previously (Mack et al. 1992;Zimanyi and Pessah 1991). We prepared brain microsomes from mouse neo cortical tissue as described previously for the rat (Wong et al. 1997a).
[ 3 H]Ry binding assays. We determined the amount of specific [ 3 H]Ry binding to microsomal membranes in the absence and pres ence of PBDE or their metabolites, a measure of how RyR activity is affected, as previously described ). See Supplemental Material (doi:10.1289/ehp.1002728) for details. At least eight concentrations of each congener were tested in triplicate in at least three indepen dent binding assays. The concentration-response curves were analyzed by curve fitting using GraphPad Prism Software (version 5; GraphPad Software, La Jolla, CA).
Microsomal Ca 2+ flux measurements. We measured PBDEinduced release of Ca 2+ actively accumulated by microsomal vesicles in the presence of ATP as previously described  Ca 2+ imaging of HEK293 cells. We fol lowed the method for growing human embry onic kidney [human embryonic kidney 293 (HEK293); RyR null ] cells in culture and the protocols for transfection and selection pro tocols for HEK293 cells that stably express RyR1 (HEK293 RyR1 ) or RyR2 (HEK293 RyR2 ) as described previously  Primary culture of mouse and rat cortical neurons. Highdensity cultures of corti cal neurons were dissociated from post natal day 0-2 SpragueDawley rats (Charles River Laboratories) for electro physiological measure ments (Yang et al. 2009) or C57BL/6J (B6) mice (Charles River Laboratories) for neuronal viability measure ments as previously described (Chen et al. 2010). See Supplemental Material (doi:10.1289/ehp.1002728) for details.
Multielectrode array (MEA) recording and data analysis. MEAs with a central 0.88mm 2 recording matrix of 64 multi electrodes (Med P545A; AutoMate Scientific, Berkeley, CA) were precoated with polyllysine (0.5 mg/mL, Sigma Chemical Co., St. Louis, MO) and laminin (10 µg/mL, Invitrogen, Carlsbad, CA, USA). Dissociated rat cortical cells were plated onto the MEAs at a density of 1 × 10 5 cells/MEA. Cultures were maintained in NeurobasalA (Invitrogen) supplemented with B27 (Invitrogen) as previously described (Wayman et al. 2006). Half of the medium was replaced twice weekly with fresh NeurobasalA containing B27. At 21 days in vitro, cortical neurons grown on MEAs were placed into the MED 64CH Integrated Amplifier inter face and spike activity was recorded using Mobius software (both from AutoMate Scientific). Baseline activity was recorded for 10 min at 37°C. Cultures were then exposed to vehi cle by adding 1 µL DMSO into MEA cul tures containing 1 mL culture medium (final   0.1% DMSO), and activity was recorded for 10 min. Subsequently, cultures were exposed to either BDE47 or BDE49 by adding 1 µL 1,000× stock solution in DMSO to the well. Cultures were sequentially exposed to increas ing concentrations of PBDE, and activity was recorded for 10 min after each addition. We analyzed spontaneous spike activity using the Spike Sorting and DC filter applications in Mobius. Spikes greater than three times the baseline noise were scored. The number of spikes per 10min recording session was deter mined (for electrodes showing activity), and the mean spike number per MEA was calcu lated. Three MEAs from three independent dissections were analyzed per PBDE. Tetrazolium-based MTS assay. We performed the MTS [3(4,5dimethyl thiazol2yl)5(3carboxymethoxyphenyl)2 (4sulfophenyl)2Htetrazolium] assay to test cell viability in cultured mouse cortical neurons after 48hr exposures to BDE47, BDE49 or 6OHBDE47 using the CellTiter 96 Cell Proliferation Assay kit (Promega, Madison, WI) as previously described (Chen et al. 2010). See Supplemental Material (doi:10.1289/ ehp.1002728) for additional details. Experiments were performed on three indepen dent cultures, with each vehicle or PBDE con centration replicated in four wells. Differences from vehicle were tested using twoway analysis of variance (ANOVA) with post hoc analysis.

Results
Ortho-bromine substitution is required for PBDE activity toward RyR1. We first tested the minimal structural requirement for PBDE activity toward RyR1 using [ 3 H]Ry binding analysis to microsomes. Neither unsubstituted diphenyl ether nor BDE15 (< 20 µM) altered the amount of specific binding of [ 3 H]Ry to RyR1, a meas ure of receptor activation (Pessah et al. 1987). In contrast, BDE4 significantly enhanced specific binding of [ 3 H]Ry to RyR1, increasing occupancy 14fold at 20 µM when measured with sub optimal Ca 2+ in the assay buffer ( Figure 1A). This finding indicates that BDE4 is an efficacious activator of RyR1 when measured under buffer conditions that promote a closedchannel conformation.
AntipyrylazoIII is a lowaffinity mem brane impermeable dye that allows quantifi cation of Ca 2+ fluxes across vesicles (Baylor et al. 1983). Spectroscopic detection of anti pyrylazoIII absorbance measures changes in free Ca 2+ in the extra vesicular solution. Ca 2+ was actively loaded into microsomes via the SR/ER Ca 2+ ATPase (SERCA) by bolus addi tions of Ca 2+ to the assay buffer containing ATP (Ca 2+ loading phase). After the loading phase was complete (the dye signal reestab lished baseline), addition of BDE4 (1-20 µM) elicited a net release of Ca 2+ (net Ca 2+ efflux; Figure 1B). The PBDE EC 50 value was 12.0 ± 0.8 µM under conditions that mimic the resting state of most mammalian cells (i.e., ~ 100 nM free Ca 2+ at the cyto plasmic face of RyR1). BDE4-triggered Ca 2+ release was fully blocked by ruthenium red (RR; 1 µM), an RyR blocker ( Figure 1B). EC 50 values were calculated from plots of the initial rate of Ca 2+ release as a function of BDE4 concentration ( Figure 1C). BDE15 and unsubstituted diphe nyl ether at concentrations < 20 µM did not elicit detectable release of accumulated Ca 2+ in the microsomal transport assay (not shown).
Activation of the Ca 2+ channel by ortho substituted PCBs requires an intact association of RyR1 with its accessory protein FK506 binding protein 12 kDa (FKBP12) . Consistent with this mecha nism, enhancement of [ 3 H]Ry binding to RyR1 by BDE4 (5 µM) was inhibited by inclusion of rapamycin in the assay (2 min pre incubation) to disrupt the RyR1FKBP12 complex ( Figure 1D). The importance of the FKBP12 complex in mediating the efficacy of BDE4 toward RyR1 was further veri fied by measuring active uptake and release of Ca 2+ from the same membrane vesicles used for [ 3 H]Ry binding analysis. Addition of BDE4 (5 µM) to vesicles actively loaded with Ca 2+ caused release of accumulated Ca 2+ , and this response was inhibited in vesicles pretreated with rapamycin in a concentrationdependent manner ( Figure 1E,F). Rapamycin eliminated BDE4-triggered Ca 2+ release without inhibit ing caffeineinduced Ca 2+ release ( Figure 1E), indicating that PBDEs sensitize RyR1 via a mechanism previously reported for PCBs.
Para substitution of environmentally relevant PBDEs determines RyR1 and RyR2 activity. Because parachlorosubstitutions of PCBs significantly influence activity toward RyR1 ), we investigated whether parabromosubstitutions simi larly influence the activity of PBDEs toward RyRs, using congeners with higher bromina tion that are environmentally relevant. We used a more activating basal buffer condi tion in these experi ments to detect both acti vating and inhibiting actions of PBDEs on RyR activity. Figure 2A shows that BDE47 weakly enhanced binding of [ 3 H]Ry to   (Figure 2A). A similar SAR was observed with RyR2 ( Figure 2B). At con centrations that maximally enhance RyR1, BDE17 and BDE49 (10 µM Figure 2A (doi:10.1289/ehp.1002728)]. However, addi tion of methoxy to the 5 position of BDE47 did not alter activity toward RyR1. Addition of BDE49 (10 µM) subsequent to loading vesi cles with Ca 2+ in the presence of ATP elicited a very robust efflux of the accumulated Ca 2+ , and the release could be completely blocked by the RyR channel blocker RR (see Supplemental Material, Figure 2B, top). 4´Methoxylated BDE49 (10 µM) added to the Ca 2+ loaded vesicles under identical assay conditions elicited a much slower release of Ca 2+ than the parent congener, and these effects were blocked by RR (see Supplemental Material, Figure 2B, middle). In contrast, BDE47 (10 µM) did not sufficiently sensitize RyR1 to produce net Ca 2+ efflux from the vesicles in the presence of the strong SERCA pump activity present in this assay (see Supplemental Material, Figure 2B, bottom).
Hydroxylation can differentially influence PBDE activity toward RyR1. OHBDE metabo lites have been found in maternal and fetal blood (Qiu et al. 2009); therefore, we tested the activity of the environ mentally rele vant hydroxylated metabo lites 6OHBDE47 and 4´OHBDE49 toward RyR1 using [ 3 H] Ry binding analysis ( Figure 3A). Unexpectedly, 4´OHBDE49 exhibited a non monotonic concentration-effect relationship, activating RyR1 at approximately 350% of vehicle control at 5 µM but having reduced efficacy at higher concentrations (< 200% of vehicle control at 10 µM). Interestingly, 6OHBDE47 inhibited [ 3 H]Ry binding in a concentrationdependent manner, with complete inhibition observed at 10 µM ( Figure 3A). Because the rate of [ 3 H]Ry binding to RyRs is too slow to detect transient activation of RyR1 if it occurred, we proceeded to test whether 6OHBDE47 had tempo rally distinct actions, initially activating and subsequently blocking RyR1 channel activity, by measuring the actions of this congener on microsomal Ca 2+ fluxes. Indeed, microsomal vesicles rapidly released their accumulated Ca 2+ when acutely challenged with 6OHBDE47 (10 µM), and this effect was completely blocked by RR ( Figure 3B). In addition, microsomal vesicles preincubated with 6OHBDE47 at 37°C (< 2 hr) significantly inhibited 4CmCinduced Ca 2+ release ( Figure 3C), indicating that prolonged incubation does subsequently block RyR1, as predicted from [ 3 H]Ry binding studies ( Figure 3A).

Nanomolar BDE-49, but not BDE-47, sensitizes RyR-mediated Ca 2+ release in intact cells.
To determine whether the different effi cacies of BDE47 and BDE49 toward RyRs extend to RyRdependent signaling events in intact cells, we tested the activity of the two congeners toward HEK293 null cells (which lack any expression of RyRs) and HEK293 cells, which stably express either RyR1 (HEK293 RyR1 ) or RyR2 (HEK293 RyR2 ) [see Supplemental Material, Figure 4A,D (doi:10.1289/ehp.1002728)]. Cells of each geno type were pretreated with 250 nM BDE49 or BDE47 for 16 hr before loading them with the Ca 2+ sensitive dye Fluo4. Once loaded with Ca 2+ indicator, cells were imaged to detect changes in cytoplasmic Ca 2+ ([Ca 2+ ] i ) before and after challenge with RyR agonists caffeine or 4CmC (Fessenden et al. 2000). HEK293 RyR1 responded to brief (10sec) focal application of caffeine with a Ca 2+ transient whose amplitude was concentration dependent  Figure 4B). Pretreatment of HEK293 RyR1 cells with BDE49 enhanced caffeineinduced Ca 2+ release, resulting in larger transient amplitudes than vehicle control (p < 0.05) at lower caffeine concentrations (see Supplemental Material, Figure 4B,C). In contrast, BDE47 did not alter caffeine responses compared with vehicle control. HEK 293 null cells failed to respond to RyR1 agonists even when pretreated with BDE49 (see Supplemental Material, Figure 4B). HEK293 RyR2 cells responded vig orously to a brief puff of 4CmC (1 mM) in contrast to HEK293 null cells (see Supplemental Material, Figure 4E). HEK293 RyR2 cells pretreated for 16 hr with BDE49, but not BDE47, showed significantly larger Ca 2+ transient amplitudes (~ 180%; p < 0.05) com pared with vehicle controls (see Supplemental Material, Figure 4F). RyR activity predicts neurotoxic potential. Considering the widely divergent activi ties of BDE47 and BDE49 on RyR1 and RyR2, both of which are expressed in brain, we tested whether their differential effects on RyR activity predict neuro toxicity. First, we examined how these PBDEs influence spon taneous electrical spiking activity of cortical neurons cultured on microelectrode arrays (MEA). MEAs have been used increasingly to assess altered develop ment of neural networks and excito toxicity caused by xeno biotics (Johnstone et al. 2010). Cortical neurons cul tured on MEA probes exhibited spontaneous spike (action potential) activity at 21 days in vitro ( Figure 4A,B). Sequential exposure to increasing concentrations of BDE49 (0, 0.2, 20, and 200 nM) at 10min intervals signif icantly increased the number of spontane ous spikes, whereas BDE47 (< 200 nM) did not alter spike activity relative to the control period ( Figure 4B,C).  Representative raster plot of spike trains over a 6-sec period in neurons exposed acutely to vehicle, BDE-47, or BDE-49. (C) BDE-49, but not BDE-47, increases spontaneous spike activity in a concentration-dependent manner; data are presented as mean ± SE (n = three arrays per treatment group). (D) Micrographs of neurons after 48 hr exposure to vehicle, BDE-47 (10 µM), or BDE-49 (5 µM or 10 µM) between 6 and 8 days in vitro. Compared with neurons exposed to vehicle and BDE-47, BDE-49-exposed neurons showed pronounced morphological changes, including fasciculation and decreased soma diameter; bar = 150 µm. (E) BDE-49 (5 and 10 µM) and 6-OH-BDE-47 (10 µM), but not BDE-47, significantly decreased cell viability as assessed using the MTS assay. We next determined whether exposures to BDE47 and BDE49 at higher concentra tions (low micromolar) and for longer times (48 hr) influence the viability of cortical neu ron cultures. Compared with vehicle controls, cultures exposed to 5 and 10 µM BDE49 invariably showed loss of phasebright cell bodies, decreased soma diameter, extreme fas ciculation of processes, and a tendency for the monolayer to pull up off the substrate ( Figure 4D). In contrast, no overt morpho logical abnormalities were observed in neu rons exposed to BDE47 (5 µM, 48 hr). Cell viability assessed using the MTS assay con firmed that BDE49 (> 5 µM) significantly decreased neuronal cell viability by > 50% (p < 0.05). A potentially significant finding is that 6OHBDE47 (10 µM), a major metabo lite of BDE47, also caused significant loss of cell viability ( Figure 4E).

Discussion
The relatively high levels of PBDEs recently meas ured in young children in North America (Herbstman et al. 2010), and espe cially California (Rose et al. 2010), esca lates concerns about their effects on human and environmental health that were raised a decade ago (Darnerud et al. 2001). Levels of the most abundant congeners in serum-BDE47, BDE99, and BDE100-in chil dren < 72 months of age have been reported to be associated with lower scores on tests of cognitive, behavioral, and physical develop ment (Herbstman et al. 2010). PBDE levels were also correlated with impairments in fine psycho motor abilities and attention but improved coordination, visual perception, and behavior (Roze et al. 2009). Similarly, in utero or post natal exposure to PBDEs resulted in hyper activity in rats (Kuriyama et al. 2005;Suvorov et al. 2009) and mice (Viberg et al. 2004). Exposure to PBDEs in neonatal mice resulted in learning deficits in visual discrimi nation tasks (Dufault et al. 2005).
Despite the mounting evidence that PBDEs cause developmental neurotoxicity, the principal molecular targets responsible for this toxicity have not been identified. PBDEs are usually found in combination with other persistent organic pollutants such as PCBs, tri closan, and o,p´DDE (o,p´dichloro diphenyl dichloro ethylene), each of which has been shown to alter Ca 2+ homeo stasis, in part by dys regu lating RyR channels (Ahn et al. 2008;Morisseau et al. 2009;Pessah et al. 2010). RyRs, along with inositol 1,4,5trisphosphate receptors (IP 3 Rs), are a family of Ca 2+ chan nels in endoplasmic or sarco plasmic reticu lum, broadly expressed in both excitable cells (striated muscles, neurons) and non excitable cells (dendritic cells, T lympho cytes) . RyRs assemble as very large homo tetrameric structures with a molecular weight > 2 MDa and consist of a very large cytoplasmic assembly (the foot region) and a relatively small trans membrane assembly com posed of six putative transmembrane passes with a central ionconducting pore. All three isoforms of RyRs are expressed in the central nervous system but are differentially distrib uted among specific brain regions, cell types, and cell regions, reflecting their participation in specialized functions . Environmental toxicants that have a poten tial for altering RyR function can influence neuronal excitability, alter synaptic plasticity, and activate cytosolic and nuclear transcrip tional events implicated in activitydependent dendritic growth .
The SAR of PBDEs toward RyR1 and RyR2 indicate that the location of ortho and para-bromine substituents significantly influ ence the level of activation of RyRs, as high lighted by divergent activities of BDE47 and BDE49, and is unrelated to differences in lipophilicity. In this regard, the free concen tration of BDE49 is likely to be significantly lower than those added to the assays used in this study. This aspect of SAR is similar to that observed with the influence of para chlorines on PCB activity toward RyR1 and RyR2 . In this respect, the larger bromine substitution at the para positions more dramatically reduces PBDE activity than the corresponding chlorine sub stitutions on PCBs and may be because of steric inter ference. The similarity in structureactivities with ortho and para substitutions sug gests a common mechanism for their effects on RyR activity. In support of a convergent mechanism, the activity of PCBs, bastadins, and PBDEs toward RyR1 can be selectively eliminated by using rapamycin to disrupt the RyR1FKBP12 complex (Chen et al. 1999;Pessah et al. 2006;Wong and Pessah 1997), an accessory protein important in finetun ing the gating properties of RyR1 channels (Brillantes et al. 1994).
The results from the present study show ing that 6OHBDE47 produces temporally defined biphasic actions on microsomal Ca 2+ release is consistent with results obtained by Dingemans et al. (2008) showing that acute exposure of PC12 cells to 6OHBDE47 promotes Ca 2+ release from intra cellular Ca 2+ stores. [ 3 H]Ry binding analysis reflects the activity of RyR channels at steadystate condi tions for the radio ligand (3 hr) and apparently fails to detect the initial channel activation elic ited by 6OHBDE47. However Ca 2+ flux measurements clearly show that the immediate actions of this congener are activation of RyR1 and release of Ca 2+ accumulated in microsomal vesicles, whereas prolonged exposure signifi cantly attenuates 4CmC-induced Ca 2+ release. Our results reveal that OHPBDEs have com plex biphasic actions on RyR function and SR/ER Ca 2+ transport properties that depend not only on the local concentration but also on the length of the exposure.
The complex interactions of OHBDEs leading to activation or inhibition of RyR1 channels appear to mirror the complex SAR described with naturally occurring macro cyclic bromo tyrosine toxins from Ianthella basta, termed "bastadins" (Chen et al. 1999;Mack et al. 1994). The reduced toxicophore that confers RyR activity resides within the eastern and western non coplanar bromo catechol ether moieties that resemble OHBDE (Masuno et al. 2006). The bromine and hydroxyl sub stituents about the diphenylether can result in either channel activation or channel inhibition.
Our results with OHBDEs suggest an under lying mechanism for the results reported by Dingemans et al. (2008Dingemans et al. ( , 2010, who iden tified a number of 4OH, 5OH, and 6OH metabolites of BDE47 and BDE49 that at 5-20 µM were more potent at mobilizing Ca 2+ from ER and/or mitochondrial stores than their respective parent structures when applied to unstimulated (resting) PC12 cells. A high concentration of BDE47 (20 µM) enhanced the Ca 2+ transient amplitude trig gered by bolus addition of extra cellular K + , whereas other PBDE congeners lacked sig nificant influence (Dingemans et al. 2010). Interestingly hydroxylation at the 4, 5, and 6 positions of BDE47 or BDE49 (2-20 µM) significantly inhibited depolarizationtriggered Ca 2+ transient amplitude. These previously reported effects of PBDEs and OHPBDEs could be explained, at least in part, by their actions on RyR channels as we describe here. Sensitization of RyRs in resting cells by OHPBDEs could result in chronic Ca 2+ leak age and depletion of ER stores. This could account in large part for both the rapid and delayed rises in cyto plasmic Ca 2+ previously observed by Dingemans et al. (2008Dingemans et al. ( , 2010. The suppression of the Ca 2+ transient ampli tude triggered by depolarization subsequent to 20 min pretreatment with 2 and 20 µM OHPBDE (Dingemans et al. 2010) would be expected to cause partial or complete deple tion of ER stores as a direct consequence of sub maximal or maximal activation of RyR channels, respectively. Interestingly, 4OH BDE49 shows a non monotonic concentra tion-response relationship toward RyR1. In contrast, BDE49 is a very efficacious activator of RyRs in [ 3 H]Ry binding studies, potently sensitizes caffeine or 4CmC-triggered Ca 2+ release in HEK293 cells that express RyRs, and is a potent excito toxicant toward primary cortical neurons. In this regard, 48 hr exposure to BDE47 < 10 µM does not promote loss of neuronal viability, whereas BDE49 does. These observations are apparently at odds with reports that developmental exposure to BDE47 causes developmental neuro toxicity in rodent models (Dingemans et al. 2007;Suvorov et al. 2009). One interpretation of this discrepancy is that RyRindependent mechanisms mediate the adverse effects of BDE47 on the developing nervous system. An alternative suggestion supported by our data is that the neuro toxic effects of BDE47 observed in vivo are mediated by 6OH BDE47 rather than the parent congener; the lack of effect of the parent compound on either spontaneous activity or cell viability in cultured cortical neurons reflects the limited metabolic capacity of this in vitro model.
The present study indicates that [ 3 H]Ry binding studies are a rapid means of defining SARs that predict neuro toxic PBDEs, much like its use to identify neuro toxic PCBs . The toxicological significance of these findings is illustrated by a recent report by (Miller et al. 2009) that BDE49, which is not typically measured in human samples, was detected in gestational tissues from women in southeast Michigan at levels comparable with BDE47. BDE49 comprised 17% of the total PBDE concentration in these tissues. This observation is consistent with reports identifying BDE49 as a major contributor to PBDE load in fish (Mariottini et al. 2008;Roosens et al. 2008), including one study of Great Lakes fish that identified BDE49 as the most abundant congener (ManchesterNeesvig et al. 2001). The mechanisms contributing to this apparent selective enrichment of BDE49 have yet to be determined, but as discussed by Miller et al. (2009), these findings suggest that the majority of human studies under estimate PBDE levels by as much as 14-19%. Our studies suggest that the problem is even greater, in that most exposure studies fail to account for congeners that pose significant risk to the developing nervous system. The SAR defined in our study provides a tool for refining human exposure studies to focus on those PBDE con geners with the greatest neuro toxic potential.

Conclusions
The present study is the first to identify RyR1 and RyR2 as direct targets of both PBDEs and their hydroxylated metabolites. These results are significant because RyRs are broadly expressed in excitable and non excitable cells, where they regulate key physiological and patho physio logical functions . Certain PBDEs and their hydroxylated metabolites have potent (maximum activities < 10 µM) and high efficacies toward altering the activities of RyR1 and RyR2 channels, and these same PBDEs are found in maternal and fetal blood, and human gestational tissues (Miller et al. 2009;Qiu et al. 2009) as well as in animal tissues (Marsh et al. 2004;Qiu et al. 2007;Valters et al. 2005). In addition, the known contribution of RyRs to inherited and acquired disorders ) strongly suggests that Ca 2+ signaling dys regulation mediated by RyR channel mecha nisms should be included in assessments of risk.