Estrogens directly potentiate neuronal L-type Ca2+ channels

L-type voltage-gated Ca2+channels (VGCC) play an important role in dendritic development, neuronal survival, and synaptic plasticity. Recent studies have demonstrated that the gonadal steroid estrogen rapidly induces Ca2+ influx in hippocampal neurons, which is required for neuroprotection and potentiation of LTP. The mechanism by which estrogen rapidly induces this Ca2+ influx is not clearly understood. We show by electrophysiological studies that extremely low concentrations of estrogens acutely potentiate VGCC in hippocampal neurons, hippocampal slices, and HEK-293 cells transfected with neuronal L-type VGCC, in a manner that was estrogen receptor (ER)-independent. Equilibrium, competitive, and whole-cell binding assays indicate that estrogen directly interacts with the VGCC. Furthermore, a L-type VGCC antagonist to the dihydropyridine site displaced estrogen binding to neuronal membranes, and the effects of estrogen were markedly attenuated in a mutant, dihydropyridine-insensitive L-type VGCC, demonstrating a direct interaction of estrogens with L-type VGCC. Thus, estrogen-induced potentiation of calcium influx via L-type VGCC may link electrical events with rapid intracellular signaling seen with estrogen exposure leading to modulation of synaptic plasticity, neuroprotection, and memory formation.

L-type voltage-gated Ca 2؉ channels (VGCC) play an important role in dendritic development, neuronal survival, and synaptic plasticity. Recent studies have demonstrated that the gonadal steroid estrogen rapidly induces Ca 2؉ influx in hippocampal neurons, which is required for neuroprotection and potentiation of LTP. The mechanism by which estrogen rapidly induces this Ca 2؉ influx is not clearly understood. We show by electrophysiological studies that extremely low concentrations of estrogens acutely potentiate VGCC in hippocampal neurons, hippocampal slices, and HEK-293 cells transfected with neuronal L-type VGCC, in a manner that was estrogen receptor (ER)-independent. Equilibrium, competitive, and whole-cell binding assays indicate that estrogen directly interacts with the VGCC. Furthermore, a L-type VGCC antagonist to the dihydropyridine site displaced estrogen binding to neuronal membranes, and the effects of estrogen were markedly attenuated in a mutant, dihydropyridineinsensitive L-type VGCC, demonstrating a direct interaction of estrogens with L-type VGCC. Thus, estrogen-induced potentiation of calcium influx via L-type VGCC may link electrical events with rapid intracellular signaling seen with estrogen exposure leading to modulation of synaptic plasticity, neuroprotection, and memory formation.
estrogen receptors ͉ signaling ͉ estradiol ͉ memory A large body of evidence shows that estrogens exert multiple rapid effects on the structure and function of neurons in a variety of brain regions, including the hippocampus (1). For example, estrogens rapidly potentiate kainite-induced currents in hippocampal neurons from wild-type (2) as well as from estrogenreceptor (ER)-␣ knockout (3) mice and induce rapid spine synapse formation in the CA1 hippocampus of ovariectomized (OVX) rats (4). Furthermore, acute application of estrogens to hippocampal slices increases NMDA and AMPA receptor transmission (5), induces long-term potentiation (LTP) and long-term depression (LTD) (6), and rapidly modulates neuronal excitability in rat medial amygdala (7) and hippocampus (8).
It is well known that estrogens interact with cell membrane components and initiate signaling events leading to a rise in intracellular Ca 2ϩ , and activation of Src kinase, G protein-coupled receptor (GPCR), MAPK, PI3K/AKT, PKA, and adenylyl cyclase (9). The mechanism(s) by which estrogens induce these rapid and diverse effects remains largely unknown. Ca 2ϩ is a second messenger that can trigger the modification of synaptic efficacy. A plasticity-induction protocol like repetitive low-frequency synaptic stimulation (10) induces the elevation of postsynaptic intracellular Ca 2ϩ . The level of intracellular Ca 2ϩ concentration can activate numerous kinases like CAMK, PKA, PKC, MAPK, PI3K, or phosphatases (11)(12)(13)(14)(15), which, respectively, phosphorylate or dephosphorylate ion channels, transcription factors, and other proteins that are involved in synaptic plasticity and memory formation. Because voltage-gated Ca 2ϩ channels (VGCC)-mediated extracellular Ca 2ϩ influx in neurons initiates the activation of these same signaling cascades (16)(17)(18)(19)(20), we hypothesized that estrogens potentiate VGCC. Here, we report that estrogen facilitates L-type VGCC in hippocampal neurons via an ER-independent mechanism, through direct binding with a domain that overlaps the dihydropy-ridine-binding site. Moreover, the capacity of estrogen to potentiate specifically L-type Cav1.2 as shown here may impart a distinctive role of estrogen in modification of synaptic efficacy.

Results
Estrogen Potentiates L-Type VGCC. Initial reports have shown that 17␤-estradiol induces rapid rise of intracellular Ca 2ϩ concentration ([Ca 2ϩ ]i) in hippocampal neurons that was inhibited by an L-type calcium channel blocker (21). Whereas the authors indicated estrogen-mediated activation of signaling events was responsible for the effect, an alternative possibility is that estrogen directly binds to and enhances the activity of the channel.
To test this hypothesis, we measured the effects of estrogens on whole-cell Ba 2ϩ /Ca 2ϩ currents in embryonic day (E)18 primary cultured rat hippocampal neurons. Ca 2ϩ channel currents were isolated by inhibiting Na ϩ currents with extracellular tetrodotoxin and K ϩ channels with intracellular Cs ϩ and extracellular 4-AP, and either Ba 2ϩ or Ca 2ϩ was used as the charge carrier. Fig. 1A shows depolarization-activated Ba 2ϩ currents (I Ba2ϩ ) recorded from this hippocampal neuron in the absence (control) or presence of 100 pM 17␤-E2. The average amplitude of Ba 2ϩ current was 93 Ϯ 13 pA in the control condition from a single hippocampal neuron. In the presence of 100 pM 17␤-E2, the Ba 2ϩ currents were increased to 192 Ϯ 18% of the control (Fig. 1 A). The potentiating effect persisted even after washout of estrogen. In a number of experiments, we observed that the stimulatory effect of a 5-min exposure to estrogen persisted for Ͼ20 minutes (data not shown). In addition, we observed a similar effect of estrogen in a hippocampal slice preparation [supporting information (SI) Fig. S1].
To determine whether the effect of estrogens on Ca 2ϩ channels is concentration dependent, hippocampal neurons were exposed to various concentrations of 17␤-E2 (Fig. 1B). Mean peak I Ba2ϩ revealed a dose-dependent increase in Ba 2ϩ currents with as little as 10 pM 17␤-E2 ( Fig. 1 B and C). Effects of 17␤-E2 were also very rapid. The onset of estrogen action was estimated to be Ͻ550 ms ( Fig. 1D and Methods).
We sought to determine which Ca 2ϩ channel subtype was being modulated by estrogens. Approximately 1/3 of the elicited Ba 2ϩ current was due to activation of L-type Ca 2ϩ channels, as evidenced by the inhibitory effects of the L-type inhibitor nifedipine. We thus assessed whether this channel may be a target of 17␤-E2. As illustrated in Fig. 1 E and F, nifedipine (10 M) nearly completely abolished the 17␤-E2-induced potentiation of Ba 2ϩ current, indi-cating that 17␤-E2's ability to enhance Ca 2ϩ current is due fully to potentiation of L-type channels.

Estrogen Action on L-Type VGCC Does Not Require Classical Estrogen
Receptor. Next, we asked whether the potentiation of VGCC is mediated by an estrogen-generated intracellular signal or by binding to a plasma membrane component. That 17␤-E2 appears to increase Ba 2ϩ currents by acting at the membrane surface rather than through an intracellular receptor was shown by the fact that membrane-impermeable 17␤-E2-BSA was found to potentiate VGCC, although to a lesser extent than 17␤-E2 alone ( Fig. 2 A and  B). To determine the role of ER in this response, we administered the ER antagonist, ICI-82,780, which did not antagonize the 17␤-E2 potentiation of Ca 2ϩ currents ( Fig. 2 C and E). Furthermore, we tested the synthetic derivative of estrogen, ZYC-26 (2-adamantyl-estrone), which does not bind to either ER␣ or ER␤ at concentrations ranging from 1 pM to 10 M (22) and does not stimulate uterine growth in ovariectomized rats (23). As shown in Of the L-type VGCC in the hippocampus, the predominant isoform is Cav1.2 (24)(25)(26). Synaptic and extrasynaptic localizations of Cav1.2 L-type VGCC correspond to putative roles of L-type calcium currents in synaptic modulation and in the propagation of dendritic Ca 2ϩ spikes (27)(28)(29). To investigate the direct action of estrogen on L-type Cav1.2, we assessed the action of 17␤-E2 in HEK293 cells transiently cotransfected with the pore-forming subunit Cav1.2 and the accessory ␤1b and ␣2␦ subunits and GFP expression plasmids. HEK293 cells do not endogenously express either L-type VGCC or estrogen receptors (30,31). Transfected HEK cells (GFP positive) showed the expected Ca 2ϩ current activated in response to the same depolarizing protocol (Fig. S2). However, in the absence of transfection of the neuronal L-type Ca 2ϩ channel, no current could be elicited in these cells in response to depolarization, thus providing a model to assess the dependence of the observed 17␤-E2 response on these two entities. As we observed in neurons, exposure of recombinant L-type Ca 2ϩ channels to 17␤-E2 resulted in a significantly enhanced Ca 2ϩ current (Fig. S2). These data support our contention that the 17␤-E2-potentiated Ba 2ϩ /Ca 2ϩ current in hippocampal neurons is through the L-type Cav1.2 Ca 2ϩ channel via a mechanism independent of ER.

Estrogen Induces Extracellular Calcium Influx Through L-Type VGCC.
To quantify the 17␤-E2-induced potentiation of extracellular Ca 2ϩ influx via VGCC, we measured intracellular Ca 2ϩ transients induced by high-K ϩ -mediated membrane depolarization using the Fura-2 dye and digital imaging microfluorometry. Fig. S3 shows that high K ϩ -induced depolarization was able to activate VGCC. The 17␤-E2 potentiated the extracellular Ca 2ϩ influx in hippocam- Ϫ90 mV to 0 mV (set at 0 min). The current amplitudes were normalized to the control before application with 17␤-E2 (n ϭ 7) or saline (n ϭ 4). Note that 17␤-E2 caused an immediate enhancement of I Ba2ϩ, whereas the saline control had no effect. Furthermore, the estrogen effect was sustained even after application. (E) Effect of L-type Ca 2ϩ channel inhibitor, nifedipine (Nif), on 17␤-E2induced potentiation of Ba 2ϩ current. Bath application of 10 M Nif completely blocked the potentiation of Ba 2ϩ current by 17␤-E2. (F) Summary of Nif blockade of 17␤-E2-induced potentiation of Ba 2ϩ currents. Nif blocked Ϸ1/3 of Ba 2ϩ current, demonstrating the presence of L-type VGCCs. In the presence of Nif, the stimulatory effects of 17␤-E2 were blocked. Current amplitudes were normalized to the control. Each data point represents four cells. * , P Ͻ 0.05, compared with the control; ϩ, P Ͻ 0.05 compared with 17␤-E2 alone, paired t test.
pal neurons cultured in vitro, and nifedipine inhibited the potentiation of Ca 2ϩ influx elicited by 17␤-E2.
Estrogen-Induced L-Type VGCC Potentiation Mechanism. The mechanism by which estrogens potentiate L-type Ca 2ϩ channel was also addressed. One potential mechanism involves estrogen-induced rapid activation of L-type associated kinases such as PKA, PKC, and/or CaMKII. These kinases are known to facilitate L-type Ca 2ϩ channels (32,33). To assess this possibility, we used the pharmacological inhibitors of phospholipase C (PLC)/PKC and CaMKII. As shown in Fig. S4, neither PLC nor CaMKII inhibitors had an effect on estrogen-induced activation of L-type VGCC.
Next, we explored the possibility that potentiation of VGCC occurs as a result of direct binding of estrogen with the channel. It has been shown that estrogen and the estrogen-like compound F90927 directly potentiate Maxi-K (34) and L-type Ca 2ϩ channels (35), respectively. Electrophysiological studies using charged dihydropyridines (DHPs) demonstrate that the binding site is accessible exclusively from the outside of the plasma membrane (36,37). We assessed whole-cell binding of 17␤-E2 to antagonist-binding sites on L-type Ca 2ϩ channels. As shown in Fig. 3, an L-type Ca 2ϩ channel antibody directed against a non-dihydropyridine-binding site region of the channel (Fig. 3, red) and 17␤-E2-FITC-BSA (Fig. 3, green) showed membrane localization in HT-22 cells, an immortalized  murine hippocampal neuronal cell line. This binding pattern resembled that of binding fluorescent (Ϫ) ST-BODIPY-DHP (Fig. 3, blue), an antagonist for L-type Ca 2ϩ channels (35). Additionally, treatment of cells with an excess of nonfluorescent Bay K (1 M), an L-type Ca 2ϩ channel agonist, reduced both 17␤-E2-FITC-BSA and DHP binding but did not affect L-type Ca 2ϩ channel antibody binding (Fig. 3), suggesting that 17␤ E2 may bind to the same DHP agonist/antagonist-binding region of L-type calcium channels. We confirmed this idea by launching a two-pronged experimental approach. A competitive binding assay using membrane preparation from HEK 293 cells transiently transfected with wild-type ␣1c and the accessory subunits showed that estrogen competes with the radioligand [3H] PN200-110 for ␣1c binding with an IC 50 of 0.67 nM (Fig. 4). Because competitive binding is done on membranes that are electrically neutral, and electrophysiological effects are voltage and channel state dependent, it is critical to study the coupling between estrogen binding and potentiation of calcium current. Thus, we used a specific mutant of ␣1c channel that is insensitive to both DHP agonists and antagonists and tested the effects of estrogen on potentiation of calcium current. As shown previously (38), mutant channels composed of Q1070M and T1066Y amino acid residues in motif III S5 of rabbit heart ␣1c subunit of L-type VGCC, are insensitive to DHP agonist and antagonist. Motif IIIS5 of rabbit heart ␣1c is 100% homologous with the rat brain ␣1c used in our experiment (Fig. S5). We therefore chose to test the effects of estrogen on the neuronal ␣1c mutant T1066Y channel. We confirmed that the T1066Yexpressing channel was insensitive to nifedipine (data not shown). Interestingly, the ability of estrogen to potentiate the T1066Y mutant channel was greatly attenuated (Fig. 5 B and C). Furthermore, whole-cell binding assays of transiently expressed wild-type ␣1C channels in HEK 293 cells (Fig. 6) show expression of ␣1C channels as evidenced by binding of a channel-specific antibody (Fig. 6A), binding of FITC-BSA-Estrogen (Fig. 6B), and binding of fluorescently labeled DHP antagonist (Fig. 6C). There was a markedly reduced binding of both FITC-BSA-Estrogen (Fig. 6F) and fluorescently labeled DHP antagonist (Fig. 6G) but not an antibody that recognizes both wild-type and mutant ␣1C channel protein (Fig. 6E), compared with the wild-type channel. This effect, coupled with our finding that estrogen competes with both a radiolabeled and a fluorescently tagged DHP, suggest that the binding domains overlap.

Discussion
The rapid interaction of estrogens with L-type Ca 2ϩ channels and the resulting potentiation of voltage-induced Ca 2ϩ currents could explain the observation that multiple and diverse signaling pathways are rapidly activated by estrogens. Calcium transients, as a result of entry through L-type Ca 2ϩ channels, are known to activate Src kinase, GPCR, MAPK, PI3K/AKT, PKA, and adenylyl cyclase signaling pathways (16)(17)(18)(19)(20).
These rapid actions of estrogens appear to explain the observation that in pre-, peri-, and postmenopausal women, estrogens affect neuronal activity measured by fMRI in a variety of brain regions during the performance of cognitive (39) and sustained attentional (40) tasks. Furthermore, estrogens enhance visual and place memory (41) and working memory performances in rats (42) and facilitate cholinergic neurotransmission in the septal-hippocampal pathways (43). Several mechanisms have been reported by which estrogen acutely and more chronically potentiates memory-related synaptic plasticity in the hippocampus. Estradiol has been shown to increase dendritic spinogenesis in the hippocampus (44,45), increase the expression of NMDA receptor (NMDAR) subunit NR2B (46), and potentiate NMDAR-mediated synaptic activity, including LTP (47,48). The identity of the ER involved in potentiation of synaptic plasticity and memory has not yet been fully confirmed. For example, in one report, ER-␣ but not ER-␤ (49), yet in another report ER-␤ but not ER-␣ (50) regulates hippocampal synaptic plasticity and enhances cognitive ability. Our data suggest a possible mechanism by which estrogen, via L-type Ca 2ϩ channel potentiation, modulates memory-related synaptic plasticity. Recently, it has been shown that activity of L-type Ca 2ϩ channels is essential for generating persistent neural firing in a neural circuit that is involved in working memory (51). Also, theoretical studies have shown that L-type Ca 2ϩ channels enhance persistent firing to variations in synaptic strength or neuronal excitability (52,53). Therefore, we envisage that estrogen by augmenting persistent neural activity may enhance working memory performance seen in various experimental settings.
The lack of involvement of ERs in the observed potentiation by 17␤-E2 of whole-cell hippocampal Ca 2ϩ currents is supported by several observations. First, the potentiation was seen at 10 pM 17␤-E2, a concentration that is 500-fold lower than the EC 50 of 17␤-E2 for either ER␣ or ER␤. Second, the potentiation was seen with ZYC-26, a nonfeminizing estrogen that does not interact in vitro or in vivo with either Ers (22,23). Third, the 17␤-E2 effects were not antagonized by concentrations of ICI-182,780 that are 35-fold higher than the IC 50 for the ERs. Finally, HEK-293 cells transfected with the essential components of the L-type Ca 2ϩ channel but lacking Ers (30,31) also responded potently to 17␤-E2.
We also studied the mechanism underlying the estrogen modulation of VGCC. The lack of effect of CaMK II or PLC/PKC inhibitors does not support that the observed estrogenic action is initiated by these intracellular signaling pathways. However, the very rapid onset of estrogen action, which was estimated to be less than a second, supports a direct interaction of estrogen with the channel protein. We confirmed this idea by (i) whole-cell binding assay using fluorescent ligand in hippocampal-derived neuronal cell line, HT22, where ␣1C channel expresses endogeneously; (ii) whole-cell binding in HEK293 cells transiently expressed mutant and wild-type ␣1C channel; (iii) competitive binding assay using radioligand; and (iv) electrophysiological studies using wild-type and dihydropyridine-insensitive channels. The displacement of a L-type Ca 2ϩ channel agonist Bay K 8644 with estrogen also provides evidence for this direct mechanism. Furthermore, it is noteworthy that a structurally similar estrogen-like compound, F90927, has recently been shown to directly modulate L-type Ca 2ϩ channels in myocytes (35). Our studies indicate that estrogen itself binds with high affinity to the L-type VGCC at a domain that overlaps with the dihydropyridine site.
Functional consequences of estrogen-induced potentiation of L-type Ca 2ϩ channels in hippocampal neurons are now becoming clear. Neuronal activity-dependent potentiation of L-type Ca 2ϩ channels has an important role in synaptic plasticity and in memory  (27,54,55). Recently, the function of L-type Ca 2ϩ channels in spatial learning, synaptic plasticity, and triggering of learning associated biochemical processes were evaluated in a transgenic mouse with an inactivated gene that encodes the Cav1.2 gene in the hippocampus and neocortex (56). This study showed selective loss of protein synthesis-dependent but NMDAR-independent LTP, a severe impairment of hippocampus-dependent spatial memory, loss of activation of MAPK pathway, and repressed cAMP response element-dependent transcription in hippocampal neurons. Also, very recently, it has been shown that the activity of L-type calcium channels is important for spike timing-dependent LTP that is absent in Fragile X syndrome (57). Therefore, we speculate that estrogen-induced, direct potentiation of L-type Ca 2ϩ channels could have implications in modulating synaptic plasticity and memory formation. Inasmuch as the L-type Ca 2ϩ channel potentiation occurs at estradiol concentrations (10 to 1,000 pM) seen in reproductively competent rodents and women, we believe that these observations are relevant to neuronal regulation in both animals and women.

Materials and Methods
Brain Slice Preparation. Transverse hippocampal brain slices (200 m) were preferred for patch clamp recording. Details of the slice preparation procedure are provided in SI Text.

Primary Neuronal Cultures.
We also studied regulation of VGCCs by 17␤-E2 in primary hippocampal cultures. Details of the preparation of the cultures are provided in SI Text.
Whole-Cell Recording. Macroscopic Ca 2ϩ or Ba 2ϩ currents were measured by using whole-cell patch clamp (58). Recording solutions, protocols, and other details are provided in SI Text.

Construction and Properties of Mutant Calcium Channels.
We used a mutant ␣1c subunit containing a threonine-to-tyrosine point mutation at position 1036 of rat brain coding sequence, which was constructed by Michael E. Greenberg [Harvard University (19)]. This mutant is homologous to rabbit heart ␣1c mutant, as shown in Fig. S2. Electrophysiological studies of rabbit heart ␣1c mutant [T1066Y (38)] indicated that this mutant is insensitive to both agonist and antagonist without effecting the basal channel activity.

Measurement of [Ca
Whole-Cell Ligand Binding Assay. Whole-cell ligand binding to HT-22 cells culture is described in detail in SI Text.

Preparation of Membranes.
Transfected HEK 293 cells were washed, scraped, and homogenized by using a glass-Teflon homogenizer in buffer X containing 50 mM Tris, 100 M PNSF, 100 M benzamidine, 1 M pepstatin A, 1 g/ml leupeptin, and 2 g/ml aprotinin (pH 8.0). The homogenate was centrifuged at 1,000 ϫ g for 5 min. The supernatant was collected and centrifuged at 100,000 ϫ g in a Beckman ultra centrifuge using a SW 41 rotor for 1 h at 4°C. The membrane pellet was washed and resuspended in buffer X.
Radioligand Binding. Detailed methods for the equilibrium binding assays culture is described in SI Text.