ortho-Substituted Polychlorinated Biphenyls Alter Microsomal Calcium Transport by Direct Interaction with Ryanodine Receptors of Mammalian Brain*

A stringent structure-activity relationship among polychlorinated biphenyls (PCBs) possessing two or moreortho-chlorine substituents is observed for activation of ryanodine receptors in mammalian brain, revealing an arylhydrocarbon receptor-independent mechanism through which non-coplanar PCBs disrupt neuronal Ca2+ signaling. Of the congeners assayed, non-coplanar PCB 95 exhibits the highest potency (EC50 = 12–24 μm) toward activating high affinity [3H]ryanodine-binding in rat hippocampus, cerebellum, and cerebral cortex. Coplanar PCB 66 and PCB 126 have no ryanodine receptor activity in all brain regions examined. PCB 95 enhances [3H]ryanodine-binding affinity and capacity by significantly altering modulation by Ca2+ and Mg2+, thereby stabilizing a high affinity conformation of the ryanodine receptor. Ca2+ transport measurements using cortical microsomes reveal that PCB 95 discriminates between inositol 1,4,5-trisphosphate- and ryanodine-sensitive stores. PCB 95 selectively mobilizes Ca2+ from ryanodine-sensitive stores in a dose-dependent manner (EC50 = 3.5 μm) and is completely inhibited by ryanodine receptor blockers, whereas coplanar PCBs are inactive. These data demonstrate that ortho-substituted PCBs disrupt Ca2+transport in central neurons by direct interaction with ryanodine receptors, showing high selectivity and specificity. Alteration of Ca2+ signaling mediated by ryanodine receptors in specific regions of the central nervous system may account, at least in part, for the significant impact of these agents toward neurodevelopment and neuroplasticity in mammals.

els in rat pheochromocytoma cells (PC12 cells) (15). Perinatal exposure of monkeys and rodents to PCBs resulted in behavioral abnormalities including delayed reflex development, altered activity patterns, learning deficits, and impaired memory (16). Rats perinatally exposed to certain ortho-substituted PCB congeners were impaired in learning a delayed spatial alternation task similar to that employed in the behavioral test of the PCB-exposed monkeys (17). Epidemiological studies reveal that children exposed either perinatally or prenatally to PCBs and other HAHs developed long-lasting cognitive function deficits (18 -21). Although the compounds that are responsible for these deficits are unknown, results from animal studies have revealed that ortho-substituted PCBs are probably responsible for the neurotoxicity observed.
In addition to the subtle neurotoxic effects induced by PCBs, the 209 possible congeners in the PCB family that constitute various industrial PCB mixtures has greatly obstructed progress in understanding the cellular and subcellular mechanisms of PCBs action. Studies have demonstrated that certain ortho-substituted PCB congeners are responsible for the neurotoxic action of PCBs, suggesting a non-AhR-mediated pathway. Shain and co-workers performed an extensive structureactivity relationship study on more than 50 PCB congeners (15). ortho-Substituted, non-coplanar PCB congeners were found to be the most active structures in decreasing dopamine level in PC12 cells, whereas coplanar PCBs were found to be inert. However, the underlying molecular mechanisms by which ortho-substituted PCBs induced neurotoxicity are unknown. We have provided evidence of the direct interaction between certain ortho-substituted PCB congeners and the ryanodine-sensitive Ca 2ϩ channel complex (RyRs) localized on sarcoplasmic and endoplasmic reticulum (SR/ER) (22). The studies reported suggest a selective molecular target by which certain ortho-substituted PCB congeners disrupt Ca 2ϩ homeostasis in muscle and within neurons in certain regions of the brain, as RyRs are differentially expressed in the central nervous system.
The present paper demonstrates for the first time a stringent structure-activity relationship among PCBs possessing two or more chlorine substitutions in the ortho positions for activation of RyRs of the mammalian central nervous system, revealing an AhR-independent mechanism through which PCBs disrupt neuronal Ca 2ϩ signaling. The most potent congener at the receptor yet identified, PCB 95 (2,2Ј,3,5Ј,6-pentachlorobiphenyl), is found to alter Ca 2ϩ transport across neuronal microsomal membrane vesicles by a RyR-mediated pathway without affecting the IP 3 R-mediated pathway. These actions of PCB 95 at RyRs may underlie its ability to alter neuronal excitability in rat hippocampal slices in vitro (23), as well as both locomotor activity and spatial learning in an in vivo rat model (24). Taken together, these results demonstrate a RyR-mediated mechanism by which certain ortho-substituted PCBs alter neuronal Ca 2ϩ signaling, through which they may alter neurodevelopment and neurobehavioral function in mammals.

Membrane Preparations
Membrane fractions from cerebellum, cerebral cortex, and hippocampus of rat were prepared as described previously (25). Briefly, tissue from each brain region was homogenized with 10-fold (w/v) ice-cold homogenization buffer consisting of 320 mM sucrose, 5 mM HEPES, pH 7.4, 100 M phenylmethylsulfonyl fluoride (PMSF), and 10 g/ml leupeptin.
Whole Particulate Fractions-The homogenate from each brain region was centrifuged at 110,000 ϫ g for 1 h. The whole particulate fractions were obtained by suspending the pellets in a buffer consisting of 320 mM sucrose, 5 mM HEPES, pH 7.4, at a protein concentration of 8 -12 mg/ml (26). The protein preparations were then rapidly frozen in liquid N 2 , and stored at Ϫ80°C until needed.
Microsomal Fractions Enriched in IP 3 R and RyR-The homogenate from each brain region was centrifuged at 1,000 ϫ g for 10 min. The membrane pellets were collected, suspended in homogenization buffer with a glass Dounce homogenizer, and re-centrifuged at 1,000 ϫ g. The two supernatant fractions were combined and centrifuged at 8,000 ϫ g for 10 min. The resulting supernatants were then centrifuged at 110,000 ϫ g for 1 h. Finally, the crude microsomal pellets were suspended and stored as described above.  (25). Nonspecific binding was obtained by adding a thousandfold excess of cold IP 3 . The samples were permitted to equilibrate either in an ice bath (4°C) for 30 min or at 25°C for 15 min with constant shaking. Binding data were analyzed with twotailed Student's t test, ␣ ϭ 0.05.

Ca 2ϩ Transport Assay
Net Ca 2ϩ flux across brain microsomal vesicles was measured with the long-wavelength Ca 2ϩ indicator fluo-3 using a luminescence spectrometer (Amino Bowman series 2, SLM Aminco, Rochester, NY). Two hundred g of the rat cortical membrane microsome was added into an assay buffer consisting of 40 mM KCl, 62.5 mM KH 2 PO 4 , 1 mM NaN 3 , 10 M K-EGTA, 8 mM MOPS, pH 7.0, 0.5 M fluo-3, 1 mM Mg-ATP, 20 g/ml creatine phosphokinase, 5 mM phosphocreatine, in a final volume of 1.2 ml (28,29). The transport assays were performed in temperaturecontrolled cuvettes at 37°C with constant stirring. The net Ca 2ϩ flux across the vesicles was recorded by measuring the extravesicular free Ca 2ϩ level as fluorescence intensity with excitation and emission wavelengths of 500 and 530 nm, respectively. The free Ca 2ϩ concentration was calculated from fluorescence intensity using the following equation (30).
F is the fluorescence intensity of the response, F max is the fluorescence intensity at saturating Ca 2ϩ , and F min is the fluorescence intensity in the absence of free Ca 2ϩ . The equilibrium dissociation constant, K D , of the Ca 2ϩ ⅐fluo-3 complex was 320 nM (30). F min and F max values for the assay were determined from the fluorescence intensities in the presence of 4-bromo A-23187 (1 g) with the addition of 10 mM EGTA and 1 mM Ca 2ϩ (saturating Ca 2ϩ level), respectively. The initial rates of Ca 2ϩ release induced by 1-10 M PCB 95 were obtained by linear regression analysis of the first 20 -50 s of data that followed the addition of PCB 95. Values of EC 50 and Hill coefficient of initial rate of PCB 95-induced Ca 2ϩ release were calculated from the linear regression analysis of logit-log transformation of the data obtained between 10 and 90% of maximal rate.

RESULTS
ortho-Substituted PCBs Alter Brain RyRs in Vitro-Aroclor 1254 and selected PCB congeners were examined for their ability to modulate the high affinity binding of [ 3 H]ryanodine to microsomes isolated from rat cerebellum, cerebral cortex, and hippocampus in the presence of physiologically relevant concentrations of intracellular K ϩ and Na ϩ . Fig. 1 demonstrated that binding of [ 3 H]ryanodine (10 nM) increased linearly as a function of cerebellar protein concentration and that non-coplanar PCB 95 significantly enhanced occupancy to Ն50 g of protein/assay. In contrast, coplanar PCB 126 did not increase occupancy above control values at any concentrations of protein examined (Fig. 1). Similar enhancement in the occupancy of nanomolar [ 3 H]ryanodine was observed with cortical and hippocampal membrane preparations (data not shown). The influence of ortho-chloro substitutions on the ability of PCB congeners to activate the high affinity binding of [ 3 H]ryanodine was examined with complete dose-response relationships for selected PCB congeners. Coplanar PCB 66 and PCB 126 had no activity toward RyRs in microsomes isolated from rat cerebellum, cerebral cortex, or hippocampus at concentrations up to their solubility limits (200 M) (Fig. 2, Table I). In marked contrast, non-coplanar PCB 95 enhanced binding of [ 3 H]ryanodine, in a dose-dependent manner, to microsomes isolated from all the brain regions examined. PCB 95 gave EC 50 values (in M) of 17, 12, and 25 with microsomes isolated from rat cerebellum, hippocampus, and cerebral cortex, respectively ( Fig. 2, Table I).
An extended structure-activity relationship was performed with cerebellar microsomes. Compared with PCB 95, PCB 4 (2,2Ј-dichlorobiphenyl) and PCB 52 (2,2Ј5,5Ј-tetrachlorobiphenyl) were 2-and 3-fold less potent, respectively. The position of the chlorine substituents about the biphenyl ring structure appeared to be as important as the degree of chlorination for RyR activity, since 2,2Ј,4,5Ј,6-and 2,2Ј,3,4,6-pentachlorobiphenyl (PCB 103 and PCB 88, respectively) were 3-and 5-fold less potent than PCB 95. The stringent structural requirement for receptor activity was further demonstrated by the finding that 2,2Ј,4,4Ј,5,5Ј-hexachlorobiphenyl, 2,2Ј,4,6,6Ј-pentachlorobiphenyl, and 2,3Ј,4Ј,5-tetrachlorobiphenyl (PCB 153, PCB 104, and  PCB 70, respectively) were approximately 10-fold less potent than PCB 95, whereas 2,3Ј,4,4Ј-tetrachlorobiphenyl (PCB 66) was inactive in all of the brain regions tested (Table I). Interestingly, the PCB mixture Aroclor 1254 showed appreciable activity toward RyR (EC 50 ϭ 35 M in cerebellum). The values of EC 50 , maximal occupancy, and Hill coefficient for selected PCB congeners and Aroclor 1254 are summarized in Table I. PCB 95 Induces Ca 2ϩ Release from Cortical Microsomes-Net Ca 2ϩ transport across rat cortical microsomal membrane vesicles was measured with the fluorometric dye fluo-3 under conditions of active Ca 2ϩ loading. In Fig. 3A, a typical trace of the Ca 2ϩ loading phase is shown, and was invariant for each of the experiments shown below. Ca 2ϩ loading was initiated by addition of ATP followed by serial addition of two 2.4-nmol additions of Ca 2ϩ . No change in fluorescence intensity, (i.e. no net Ca 2ϩ transport across the membrane vesicles) was observed before the addition of ATP, suggesting that Ca 2ϩ uptake by the membrane vesicles was ATP-dependent (data not shown). Addition of the Ca 2ϩ ionophore 4-bromo A-23187 after completion of the loading phase demonstrated that the accumulated Ca 2ϩ could be rapidly released from the vesicles. Addition of Ca 2ϩ (2 ϫ 2.4 nmol) at the end of each experiment verified the linearity and the calibration of the dye signal. In all the measurements reported in the present study, none of the drugs at the concentrations used significantly interfered with either the sensitivity or calibration of the fluo-3 dye (data not shown). Under these assay conditions, F min and F max were 0.5 Ϯ 0.1 and 8.0 Ϯ 0.3, respectively (mean of three determinations). In Fig. 3B, addition of coplanar PCB 66 or PCB 126 to cortical microsomes actively loaded with Ca 2ϩ had no effect on the net Ca 2ϩ flux across the vesicles. In marked contrast, addition of PCB 95 induced a rapid release of Ca 2ϩ from the vesicles. The initial rate of Ca 2ϩ release from the vesicles induced by PCB 95 (1-10 M) was dose-dependent ( Fig. 3C) with an apparent EC 50 of 3.5 M (Fig. 3C, inset).
RyR Blockers Inhibit PCB 95-induced Ca 2ϩ Release from Brain Microsomes-The mechanism by which PCB 95 induced Ca 2ϩ release from brain microsomes was further studied using ruthenium red and ryanodine, known RyR blockers. Fig. 4 shows the response of actively loaded cortical microsomes to 5 M PCB 95 (trace A). Addition of ruthenium red (1 M) after loading the vesicles with Ca 2ϩ largely eliminated the response to PCB 95 (ϳ 94% inhibition of the initial Ca 2ϩ release rate) (Fig. 4, trace B). Prior addition of 500 M ryanodine resulted in a typical biphasic response of the receptor, whereas subsequent addition of PCB 95 failed to mobilize Ca 2ϩ from the vesicles (ϳ93% inhibition of the initial Ca 2ϩ release rate) (Fig. 4, trace C). These results suggested that PCB 95 induced Ca 2ϩ release was inhibited by ryanodine-sensitive Ca 2ϩ channel blockers.
PCB 95 Selectively Targets Ryanodine-sensitive Ca 2ϩ Stores-To discriminate which of the Ca 2ϩ stores in the microsomal preparation are sensitive to PCB 95, additional pharmacological studies were performed with IP 3 and ryanodine. Fig. 5 illustrates the ability of D-IP 3 to stereoselectively activate IP 3 Rs in the microsomal preparation, since L-IP 3 is inactive (trace a). Addition of 500 M ryanodine to the preparation caused a biphasic response similar to that seen with junctional SR vesicles isolated from striated muscle (31): initially activating and subsequently inactivating RyRs, resulting in net reuptake of the Ca 2ϩ into the cortical vesicles (Fig. 5, trace b). Addition of D-IP 3 after treating with ryanodine demonstrated that the IP 3 Rs maintained their sensitivity to agonist, as the rate and amount of IP 3 -induced Ca 2ϩ release was similar to that seen with D-IP 3 alone (Fig. 5, compare traces a and b). Addition of ryanodine subsequent to IP 3 -induced Ca 2ϩ release mobilized stored Ca 2ϩ with magnitude and kinetics quantitatively similar to ryanodine-induced Ca 2ϩ release in the absence of IP 3 (Fig. 5, compare traces a and b). Taken together, the results suggested that the IP 3 -and ryanodine-sensitive efflux pathways in the cortical microsomal preparation were on distinct vesicles.
To further determine whether the actions of PCB 95 were selective toward the ryanodine-sensitive store, PCB 95 was added prior or subsequent to addition of D-IP 3 . With either experimental protocol, PCB 95-induced Ca 2ϩ release was quantitatively similar (Fig. 6, compare traces a and b). Addition of 1 M ruthenium red after depleting the IP 3 -sensitive stores completely blocked PCB 95-induced Ca 2ϩ release (Fig. 6, trace c). Similarly, prior treatment of the Ca 2ϩ -loaded vesicles with 1 M ruthenium red completely negated the response to PCB 95 without altering the response to D-IP 3 (Fig. 6, trace d). Addition of 60 M of heparin to Ca 2ϩ -loaded vesicles resulted in an instantaneous jump in the fluo-3 response due to contamination of the drug with 34 pmol of Ca 2ϩ . However, subsequent addition of D-IP 3 failed to induce the release of Ca 2ϩ even though the preparation remained completely responsive to PCB 95 (Fig. 6, compare traces a, b, and e). These results further demonstrated the presence of distinct IP 3 -and ryanodine-sensitive vesicles in the microsomal preparation and that PCB 95 selectively mobilized Ca 2ϩ by directly interacting with vesicles possessing RyRs. PCB 95 (5 M) did not appear to alter the Ca 2ϩ transport properties of the fraction of IP 3 R-containing vesicles. In consonance with this observation, PCB 95 (Յ50 M) did not significantly alter the binding of [ 3 H]IP 3 to cerebellar microsomes, although higher concentrations did produce a statistically significant enhancement of occupancy that was not seen with coplanar PCB 126 (Fig. 7).

PCB 95 Stabilizes a High Affinity Conformation of RyRs-
The mechanism by which PCB 95 enhanced the high affinity binding of [ 3 H]ryanodine to cerebellar microsomes was further elucidated by performing saturation binding measurements and examining changes in modulation by Ca 2ϩ and Mg 2ϩ . In the presence of physiologically relevant concentrations of intracellular Na ϩ (15 mM) and K ϩ (140 mM), negligible specific binding of [ 3 H]ryanodine (0.5-25 nM) was detected (data not shown), whereas inclusion of PCB 95 (50 M) to the same assay medium produced a significant enhancement in the number of high affinity binding sites for [ 3 H]ryanodine exhibiting a B max of 68 Ϯ 5 fmol/mg and a K D of 3.7 Ϯ 0.6 nM (Fig. 10). By contrast, control measurement in the presence of high salt (200 mM KCl) revealed that [ 3 H]ryanodine bound to RyRs on cerebellar microsomes with B max of 52 Ϯ 1 fmol/mg and K D of 24 Ϯ 3 nM (Fig. 8).
In the presence of 50 M PCB 95 and physiological concentrations of Na ϩ and K ϩ , the EC 50 for activation of [ 3 H]ryanodine-binding sites by Ca 2ϩ was 61 nM (Fig. 9A, Table II). Interestingly, no consistent inhibition of [ 3 H]ryanodine binding was observed at Ca 2ϩ as high as 200 mM. In contrast, control measurements performed in the presence of 200 mM KCl exhibited an IC 50 for Ca 2ϩ -activated binding of 1.6 mM and a Hill coefficient of 3, with near-complete inhibition at 10 mM Ca 2ϩ . Furthermore, in the presence of 50 M PCB 95, the IC 50 for Mg 2ϩ was 19 mM and the Hill coefficient was 1.3, and only ϳ50% of the binding sites could be inhibited even with 1 M Mg 2ϩ (Fig. 9B, Table II). Control measurement in the presence of 200 mM KCl revealed an IC 50 of 1 mM and Hill coefficient of 3.7 for Mg 2ϩ , with Ͼ90% inhibition at 10 mM Mg 2ϩ . DISCUSSION

ortho-Substituted PCBs Alter Ca 2ϩ Regulation in Rat Brain by a Novel Mechanism Independent of the AhR-
The present study demonstrates that certain ortho-substituted PCB congeners alter Ca 2ϩ regulation of central neurons by a RyR-mediated mechanism. In neuronal tissues, two types of ligand-gated Ca 2ϩ release channels have been found to be localized to the ER (32, 33): 1) IP 3 R channels and 2) Ca 2ϩ -induced Ca 2ϩ release channels (i.e. RyR). Because of structural homologies shared by IP 3 Rs and RyRs, it is necessary to determine if ortho-substituted PCBs discriminate between these microsomal Ca 2ϩ release channels. In neurons, the ryanodine-sensitive stores have thus far been found mainly localized in cell soma, whereas the IP 3 -sensitive stores appear to be equally distributed in neuronal soma and processes (34,35). Using autoradiography, Synder and co-workers (33) have demonstrated the differential localization of IP 3 Rs and RyRs within the central nervous system. Direct Ca 2ϩ transport measurements are conducted in the present study using microsomal membrane isolated form rat cerebral cortex. The results demonstrate that IP 3 Rs and RyRs do not share the same population of vesicles. Despite the presence of a large fraction of IP 3 -sensitive vesicles in the microsomal preparation used in the present study, the potent actions of PCB 95 are found to be completely selective toward the ryanodine-sensitive microsomes. In addition, PCB 95-induced Ca 2ϩ mobilization from brain microsomes is completely blocked by M ryanodine or ruthenium red, but not heparin. The inability of PCB 95 (Յ50 M) to significantly alter the binding of [ 3 H]IP 3 to cerebellar microsomes offers further support for the hypothesis that PCB 95 selectively targets RyRs. However, at the present time we cannot discount the possibility that other PCB structures possess IP 3 R activity. In light of the current findings, PCB congeners exhibiting potent activity toward RyR proteins in the mammalian brain would be expected to alter Ca 2ϩ signaling and Ca 2ϩ -dependent processes in affected neurons. Although it is currently unclear what the exact role of ryanodine-sensitive Ca 2ϩ channels is in the adult mammalian brain, the distinct heterogeneity in the distribution of neuronal RyRs suggests distinct Ca 2ϩ -associated brain functions for each isoform. Given the potency and specificity of PCB congeners, they may represent a new class of molecular probes to define the function of RyRs in the brain. Because of the wide environmental distribution of PCBs, and our emerging understanding on the role of RyRs in neurodevelopment and neuroplasticity, this newly identified mechanism by which PCBs alter Ca 2ϩ signaling in mammalian brain may underlie the neurotoxicity that has been attributed to non-coplanar PCBs (14).
Three isoforms of RyRs have been shown to be expressed in the mammalian central nervous system, where they are thought to be responsible for Ca 2ϩ -induced Ca 2ϩ release (36). In situ hybridization and immunolocalization studies have revealed a cell type-specific pattern of expression in the different regions of the central nervous system (37)(38)(39)(40). Several recent studies have provided evidence that RyRs are under strict developmental control. First, Ry 1 R and Ry 3 R expression have been shown to be regulated by cytokines and growth factors (41)(42)(43). Second, using polymerase chain reaction analysis, Futatsugi et al. (44) have identified two alternatively spliced regions in mRNA of mouse Ry 1 R, which are characterized by the presence or absence of amino acid sequences that exist within the region where modulatory sites for phosphorylation and binding of Ca 2ϩ , calmodulin, and ATP are postulated to exist. The ratio of Ry 1 R splice variants changes abruptly in the cerebrum between embryonic days 14 and 18. Third, RyRs have been found to be expressed in neural growth cones, where they are thought to play an important role in buffering and releasing Ca 2ϩ during intracellular Ca 2ϩ oscillations (45). Interestingly, RyRs appear to regulate the amplitude of Ca 2ϩ spiking behavior in the growth cone, suggesting a role in signal amplification. Since the homeostatic mechanisms controlling intracellular Ca 2ϩ dynamics of growth cones are likely to be important determinants of growth cone migration during development, it is worthwhile to speculate how prenatal exposure to PCB 95 might alter RyR function and its relationship to reduced motor activity and radial arm maze performance (24). PCB 95 administered perinatally would be expected to alter the sensitivity of Ry 1 R and Ry 2 R (cardiac isoform) to normal stimuli. Indeed, Schantz and co-workers have demonstrated that rats exposed to PCB 95 perinatally exhibit significantly altered locomotor activity and responses to a hippocampal learning  Table II. task. These behavioral changes are correlated with a significant regiospecific changes in [ 3 H]ryanodine-binding (24). If the temporal pattern with which the various isoforms of RyRs expressed are important for normal brain development, then agents that alter the timing or level of their expression could also alter subtle aspects of neurodevelopment.
Studies from Kodavanti and co-workers (46,47) using primary cerebellar granule cell cultures suggest a similar structure-activity relationship of PCBs activity. Inhibition of microsomal and mitochondrial Ca 2ϩ -ATPases by PCBs has been proposed to be the molecular mechanism that perturbs Ca 2ϩ homeostasis in neuronal cells in culture. In support of the hypothesis, recently Kodavanti and co-workers (47) have shown that PCB 4 enhances protein kinase C translocation in neuronal cell culture. However, the present study provides direct evidence for an alternate mechanism which involves mobilization of Ca 2ϩ from neuronal ER. Previously, we have shown that PCB 95 and Aroclor 1254 at Յ10 M exhibited no effect on the activity SERCA1 and SERCA2 pumps (22). Ryanodine-sensitive Ca 2ϩ channels of the brain appear to be a selective molecular target for certain non-coplanar PCBs. Aroclor 1254 is a commercial mixture of PCBs, which is composed of 49% pentachlorinated biphenyls. The high activity of Aroclor 1254 toward RyRs of the brain implies that several of the constituent congeners that have yet to be tested possess significant receptor activity.
Structural Specificity of PCBs-Coplanar PCB 66, a monoortho-substituted tetrachlorobiphenyl, has been shown to possess similar physicochemical properties to PCB 95 based on gas chromatographic retention coefficients (7). Despite their similar hydrophobicity, coplanar PCB 66 is inactive toward modifying RyRs and Ca 2ϩ transport in all the brain regions studied, demonstrating that the number of chlorines in the ortho positions is a major determinant for activity toward RyR. Coplanar PCB 126 has been widely studied for its ability to bind the cytosolic AhR and induce hepatic microsomal enzymes (10). Like PCB 66, PCB 126 does not alter the occupancy of [ 3 H]ryanodine to its high affinity sites in any of the rat brain regions examined, nor does it alter Ca 2ϩ transport across the microsomal vesicles, suggesting that coplanar PCBs cannot directly interact with any of the RyR isoforms present in rat brain. In consonance with these findings, PCB 95, but not PCB 126, alters Ca 2ϩ transport across SR vesicles isolated from either skeletal muscle (enriched in Ry 1 R and SERCA1 isoforms) or cardiac muscle (enriched in Ry 2 R and SERCA2 isoforms) by a RyR-mediated mechanism (22).
The data presented in the present paper demonstrate that PCB congeners with two or three chlorine substituents in the ortho-position confer the highest potency and efficacy toward RyRs of the mammalian brain. Of the congeners assayed, PCB 95 (2,2Ј,3,5Ј,6-pentachlorobiphenyl) and PCB 4 (2,2Ј-dichlorobiphenyl) exhibit the highest potency and efficacy. Therefore, the position of the chlorine substituents on PCBs is more important toward conferring RyR activity than the degree of chlorination. The four ortho-chloro substituents of PCB 104 (2,2Ј,4,6,6Ј-pentachlorobiphenyl) contribute significant steric constraint which severely limit rotation about the biphenyl bond. The finding that PCB 104 has significantly lower receptor potency than PCB 95 and PCB 4 suggests that although a non-coplanar conformation of the biphenyl structure appears to be critical for receptor activity, a certain degree of rotational flexibility about the biphenyl bond seems to be required to produce maximum activation of RyRs. The high receptor activity exhibited by PCB 4 can reflect an induced fit of this congener with its binding domain on the receptor complex. Other than three ortho-chlorine substituents, substitutions at the meta-and para-positions are also important for optimal activity at RyRs. This is exemplified by the stringent structureactivity relationship among pentachlorobiphenyls, which reveals a ranked potency of 2,2Ј,3,5Ј,6-Ͼ 2,2Ј,4,5Ј,6-Ͼ 2,2Ј,3,4,6-pentachlorobiphenyl. PCBs Modulate RyRs by a Novel Mechanism-Typically, studies of the high affinity binding of [ 3 H]ryanodine to RyRs to mammalian brain microsomes have been performed in the presence of high salt (1 M KCl or NaCl) (25, 48 -50). [ 3 H]Ryanodine binding to its high affinity sites was shown to be modulated by Ca 2ϩ , Mg 2ϩ , caffeine, and adenosine nucleotides (25,50), in a manner qualitatively similar to those reported for skeletal and cardiac SR preparations. The requirement for high salt may be important to stabilize an open conformation of the receptors which recognizes [ 3 H]ryanodine with high affinity but the underlying mechanism has remained unclear. Certain ortho-substituted PCB congeners (e.g. PCB 95) effectively eliminate the requirement of high salt in the assay medium. Results from saturation binding of [ 3 H]ryanodine to cerebellar microsomes in the presence of a minimal concentration of salt (200 mM K ϩ ) to permit measurement of high affinity binding of [ 3 H]ryanodine reveal a 6.5-fold lower affinity and 1.5-fold lower capacity compared with an assay medium containing physiological K ϩ and Na ϩ , and PCB 95. Moreover, the value of K D in the presence of PCB 95 is similar to that previously reported by Zimanyi and Pessah (K D ϭ 1-3 nM) and Padua et al. (K D ϭ 2.4 nM), in the presence of 1 M KCl (25,50). Thus PCB 95 appears to significantly stabilize a single high affinity state (Ca 2ϩ conducting state) even in the presence of physiologically relevant K ϩ and Na ϩ .
The mechanism by which PCB 95 favors the high affinity state of brain RyR appears to be related to its ability to dramatically alter the responses of the channel to two important physiological ligands, Ca 2ϩ and Mg 2ϩ . It has been previously demonstrated that micromolar Ca 2ϩ is required to activate high affinity binding of [ 3 H]ryanodine to brain microsomes in the presence of 1 M K ϩ (50). Under the same conditions, millimolar Ca 2ϩ has been shown to fully inhibit the binding of [ 3 H]ryanodine to its high affinity site (50). In the present paper, we report that PCB 95 significantly alters the sensitivity of RyRs to activation and inhibition by Ca 2ϩ , shifting the EC 50 value to 61 nM and essentially eliminating inhibition. Zimanyi and Pessah have reported an IC 50 for Mg 2ϩ of 10 mM in high salt (1 M K ϩ ) (25), and Padua and co-workers (50) have reported IC 50 values for Mg 2ϩ of 2 mM and 5 mM assayed in 200 mM K ϩ and 1 M K ϩ , respectively. In this respect, PCB 95 alters Mg 2ϩ inhibition in two important ways; 1) it shifts the IC 50 for susceptible sites nearly 20-fold compared with control lacking PCB, and 2) it completely eliminates inhibition for approximately 50% of the measurable sites. Therefore, altered modulation by Ca 2ϩ and Mg 2ϩ appears to underlie the ability of PCB 95 to stabilize a high affinity [ 3 H]ryanodine-binding conformation of RyRs of the brain.
In conclusion, ortho-substituted PCBs are shown for the first time to directly and selectively activate the RyR/Ca 2ϩ release channel complex in adult rat brain. Structure-activity relationship studies performed with selected PCBs indicate a stringent structural requirement for activation of RyRs in central nervous system. ortho-Substituted PCB 95, the most potent congener studied, mobilizes microsomal Ca 2ϩ selectively and specifically from ryanodine-sensitive microsomes. Disruption of Ca 2ϩ homeostasis in the affected regions of the brain by certain ortho-substituted PCBs may contribute significantly in altering neurodevelopment and neuroplasticity function in mammals.